ENGINEERING IN THE TROPICS:
EVALUATING A SOLAR-POWERED ANAEROBIC DIGESTION AND HYBRID
CONSTRUCTED TREATMENT WETLAND SYSTEM TO TREAT AGRICULTURAL WASTES IN
COSTA RICA
By
Ronald Esteban Aguilar

A DISSERTATION
Submitted to
Michigan State University
in partial fulfilment of requirements
for the degree of
Biosystems Engineering – Doctor of Philosophy
2018

ABSTRACT
ENGINEERING IN THE TROPICS:
EVALUATING A SOLAR-POWERED ANAEROBIC DIGESTION AND HYBRID CONSTRUCTED
TREATMENT WETLAND SYSTEM TO TREAT AGRICULTURAL WASTES IN COSTA RICA
By
Ronald Esteban Aguilar
In Costa Rica, treatment of biomass residues and wastewater from agro-industrial and
agricultural is often neglected and, consequently, Costa Rica has a severe water pollution
problem. This study evaluated the performance of a solar-powered anaerobic digestion and
hybrid constructed treatment wetland system to treat agricultural wastes in Costa Rica. The
integrated solar thermal collector, anaerobic digester, and hybrid constructed treatment wetland
system was implemented in 2013 at the Fabio Baudrit Experimental Station, Costa Rica and was
proposed as a decentralized self-sufficient, close-loop, organic waste treatment system
technology for rural areas.
A solar thermal collector unit provided sufficient energy to maintain thermophilic
temperatures in a 20 m3 anaerobic digester, which converted agricultural wastes, primarily food
waste and chicken litter, to biogas and nutrient-rich solids. A constructed treatment wetland was
used to treat water from the digestate for use in the digester and/or irrigation. The entire
treatment system was self-sufficient, producing surplus energy. In general, 263 MJ of energy, 28
kg of fertilizer, and 550 kg of treated water were generated from 863 kg of mixed animal and
food wastes. The net revenue considering electricity and fertilizer was $2,146 annually. The
payback period for the system was estimated to be 21 years; however, a sensitivity analysis
demonstrated that through optimization, the payback period could be reduced to 9 years.

The hybrid constructed treatment wetland achieved a treatment performance that allowed
reuse of the treated water for other activities (e.g., irrigation and reuse at the digester). Pollutant
removals were 99.5% for chemical oxygen demand (COD), 94.4% for total solids (TS), 99.8%
for total nitrogen (TN), and 99.1% for total phosphorus (TP) during the rainy season and 96.4%
for COD, 86.5% for TS, 98.9% for TN, and 99.6% for TP during the dry season. The hybrid
configuration, a vertical subsurface flow and a free water surface constructed treatment wetland
in series, was essential to overcome the individual weaknesses of each type of wetland,
especially with regards to water storage and nitrogen. In addition, the vertical subsurface flow
constructed treatment wetland did not become clogged after four years and a constant void space
in the filter media of 20 of 30 m3 was estimated from August 2015 to March 2016.
The system was exergetically sustainable with an environmental exergy efficiency
(Ρenv,ex) of 5.60 and a total pollution rate (Rpol,ex) of -0.821, due to a positive exergy balance in
which food waste and chicken litter were converted into high quality end products (i.e., energy,
fertilizer, and treated water). Precipitation did not impact the exergy-based assessment of
sustainability and more exergetically favorable Ρenv,ex and Rpol,ex values during the dry months
were obtained due to better digestion performance, storage of water for future use, and biomass
production. From the technical point of view, the system can contribute to sustainability of
agricultural systems and communities in Costa Rica. This approach represents an academic effort
toward waste treatment not only in Costa Rica, but also to other rural areas in the tropics.

For God,
for giving me the most precious gift in my life:
my parents.
Love you both

iv

ACKNOWLEDGMENTS

First, I want to thank the entire BAE-MSU faculty, who since 2009 has been in touch
with the University of Costa Rica. A special mention to Dr. Srivastava, Dr. Reinhold, and Dr.
Liao, you kindly opened the door to this marvelous opportunity that I started in 2012. In
addition, to thank to Ing. Aguilar Pereira, from the Department of Agricultural Engineering in
2009, who insisted on the BAE faculty to strengthen ties with The University of Costa Rica.
Special acknowledgment to the University of Costa Rica, which sponsored me. Thanks to
the staff of different research centers at the UCR: Werner Rodriguez, Daniel Baudrit, Alberto
Miranda, Carlos Benavides, and Jorge from the Fabio Baudrit Experimental Station; Paola and
Maria Elena from CELEQ; Mariana, Lorena and Lidieth from CIA. Victor and all from LCA.
You all helped me to achieve my goals, adding the pura vida style to my work at Costa Rica.
Thanks to my lab mates, Rebecca, Niroj, and Khang, you were important in all this
process, you always were generous to share ideas, a good Costa Rican cup of coffee, and an
Imperial. I want to thank all the Latin American friends, especially to Mauricio, Norma, and
Laura. Thanks to my family and friends, with special mention to Fabian, Katia, and Alfredo. You
trusted me.
Finally, to Noilin, who came at the end of this long process, love you.

v

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................................... ix
LIST OF FIGURES ...................................................................................................................... xii
KEY TO ABBREVIATIONS ...................................................................................................... xvi
CHAPTER 1: INTRODUCTION ................................................................................................... 1
1.1. Biomass residues and energy production .............................................................................. 1
1.2. Wastewater treatment and the sanitation problem................................................................. 3
1.3. Self-sufficient energy production and waste treatment system ............................................. 3
1.4. The solar-powered anaerobic digestion and hybrid constructed treatment wetland system . 4
1.5. Overview of the system evaluated in this study .................................................................. 11
1.6. The system studied in this dissertation ................................................................................ 14
1.7. Scope and aims of this research........................................................................................... 15
CHAPTER 2: LITERATURE REVIEW ...................................................................................... 18
2.1. The opportunity: from linear to circular thought ................................................................. 18
2.2. Biomass residues and wastewater from the agro-industry and agricultural activities in
Costa Rica ..................................................................................................................................... 20
2.2.1. Biomass residues from agro-industry and agricultural activities ................................... 21
2.2.2. Wastewater from agro-industry and agricultural activities ............................................ 26
2.2.3. Biomass residues and wastewater: the problem becoming an opportunity ................... 33
2.3. Integration of technologies at the tropics: Costa Rica case ................................................. 34
2.3.1. Waste to energy: optimizing anaerobic digestion through solar thermal collection ...... 35
2.3.2. Constructed treatment wetlands: ending the water pollution problem .......................... 41
2.3.2.1. Horizontal subsurface flow wetlands........................................................................ 41
2.3.2.2. Vertical subsurface flow wetlands ............................................................................ 44
2.3.2.3. Free water surface wetland ....................................................................................... 47
2.3.2.4. Hybrid constructed treatment wetland ...................................................................... 50
2.3.2.5. Selection criteria ....................................................................................................... 51
2.3.3. Closing the loop for water, energy, and food ................................................................. 53
2.4. Technical assessment of a sustainable development ........................................................... 53
CHAPTER 3: TECHNICAL PERFORMANCE OF A SOLAR-POWERED WASTE
UTILIZATION AND TREATMENT SYSTEM IN COSTA RICA ............................................ 56
3.1. Introduction ......................................................................................................................... 56
3.2. Material and methods .......................................................................................................... 61
3.2.1. System description ......................................................................................................... 61
3.2.1.1. Solar thermal collection ............................................................................................ 62
3.2.1.2. Thermophilic CSTR anaerobic digester ................................................................... 63
3.2.1.3. Control unit and data collection ................................................................................ 65
vi

3.2.1.4. Electricity generator.................................................................................................. 65
3.2.1.5. VSSF-CTW............................................................................................................... 66
3.2.2. Mass and energy balance ............................................................................................... 67
3.2.3. Cash flow and payback period ....................................................................................... 68
3.2.4. Analytic method ............................................................................................................. 72
3.3. Results and discussion ......................................................................................................... 74
3.3.1. Solar thermal collection ................................................................................................. 74
3.3.2. Thermophilic CSTR anaerobic digester ......................................................................... 75
3.3.3. VSSF-CTW .................................................................................................................... 77
3.3.4. Mass and energy balance ............................................................................................... 79
3.3.5. Cash flow and payback period ....................................................................................... 81
3.4. Conclusions ......................................................................................................................... 87
3.5. Acknowledgments ............................................................................................................... 88
CHAPTER 4: PERFORMANCE OF THE HYBRID CONSTRUCTED TREATMENT
WETLAND TREATING DIGESTATE EFFLUENT IN COSTA RICA .................................... 89
4.1. Introduction ......................................................................................................................... 90
4.2. Material and Methods .......................................................................................................... 98
4.2.1. System description ......................................................................................................... 98
4.2.2. Hydrological balance ................................................................................................... 103
4.2.3. Treatment performance ................................................................................................ 104
4.2.4. Modeling approach....................................................................................................... 106
4.3. Results and discussion ....................................................................................................... 107
4.3.1. Hydrological balance ................................................................................................... 107
4.3.1.1. Hydrological balance on the VSSF-CTW .............................................................. 109
4.3.1.2. Hydrological balance on the FWS-CTW ................................................................ 111
4.3.2. Treatment performance ................................................................................................ 112
4.3.2.1. The geotextile membrane at the VSSF-CTW ......................................................... 119
4.3.3. Modeling approach....................................................................................................... 123
4.4. Conclusions ....................................................................................................................... 131
4.5. Acknowledgments ............................................................................................................. 132
CHAPTER 5: CLOGGING IN TROPICAL VERTICAL SUBSURFACE FLOW
CONSTRUCTED TREATMENT WETLANDS ....................................................................... 133
5.1. Introduction ....................................................................................................................... 133
5.2. Material and Methods ........................................................................................................ 138
5.2.1. System description ....................................................................................................... 138
5.2.2. Laboratory-scale VSSF-CTW study ............................................................................ 140
5.2.3. Solids accumulation in the filter media ........................................................................ 142
5.2.4. Root development in the filter media ........................................................................... 144
5.2.5. Infiltration in the filter media ....................................................................................... 144
5.2.6. Treatment performance of the VSSF-CTW ................................................................. 145
5.2.7. Statistical analysis ........................................................................................................ 146
5.3. Results and discussion ....................................................................................................... 147
5.3.1. Full-scale VSSF-CTW ................................................................................................. 147
5.3.2. Laboratory-scale VSSF-CTW ...................................................................................... 159
vii

5.4. Conclusion ......................................................................................................................... 167
5.5. Acknowledgment ............................................................................................................... 168
CHAPTER 6: EXERGY-BASED ASSESSMENT OF SUSTAINABILITY OF A SOLARPOWERED ANAEROBIC DIGESTION AND HYBRID CONSTRUCTED TREATMENT
WETLAND SYSTEM TO TREAT AGRICULTURAL WASTES IN COSTA RICA ............. 169
6.1. Introduction ....................................................................................................................... 170
6.2. Material and Methods ........................................................................................................ 175
6.2.1. System description ....................................................................................................... 175
6.2.2. Boundary definitions .................................................................................................... 178
6.2.3. Data collection ............................................................................................................. 179
6.2.3.1. Case 0 – the SPAD ................................................................................................. 180
6.2.3.2. Case 1 – the SPAD with the VSSF-CTW ............................................................... 181
6.2.3.3. Case 2 – SPAD-HCTW .......................................................................................... 182
6.2.4. Analytical methods ...................................................................................................... 183
6.2.5. Exergy calculation ........................................................................................................ 184
6.2.6. Statistical analysis ........................................................................................................ 187
6.3. Results and discussion ....................................................................................................... 187
6.3.1. Case 0 – the SPAD ......................................................................................................... 187
6.3.2. Case 1 – the SPAD with the VSSF-CTW ...................................................................... 194
6.3.3. Case 2 – SPAD-HCTW .................................................................................................. 200
6.4. Conclusions ....................................................................................................................... 205
6.5. Acknowledgment ............................................................................................................... 208
CHAPTER 7: CONCLUSIONS ................................................................................................. 210
7.1. Biomass residues and wastewater: the problem becoming an opportunity for sustainable
development ................................................................................................................................ 210
7.2. Future work ....................................................................................................................... 219
REFERENCES ........................................................................................................................... 223

viii

LIST OF TABLES

Table 1. 1. Description and basis of design. ................................................................................... 7
Table 1. 2. Evaluation of fecal coliforms after digestion. ............................................................. 13
Table 1. 3. Water quality analysis of entire system from Lambda Laboratory (Dec. 2014). ....... 13
Table 2. 1. Characteristics of selected biomass residues in Costa Rica. ....................................... 22
Table 2. 2. Laws and regulations for the management of solid residues in Costa Rica. .............. 23
Table 2. 3. Summary of anaerobic digester projects developed by the Non-conventional
renewable energy program (ICE). ................................................................................................. 25
Table 2. 4. Laws and regulations for the management of liquid residues in Costa Rica. ............. 30
Table 2. 5. Summary of biogarden and constructed treatment wetland projects developed by
academic institutions..................................................................................................................... 32
Table 2. 6. Biomass residues and wastewater: the problem and subsequent opportunities. ......... 34
Table 2. 7. Hybrid constructed treatment wetland efficacy. ......................................................... 51
Table 2. 8. Treatment properties of constructed treatment wetlands. ........................................... 52
Table 3. 1. Characteristics of the feed........................................................................................... 64
Table 3. 2. Energy consumption per piece of equipment in the SPAD-HCTW. .......................... 66
Table 3. 3. Parameters considered for the cash flow evaluation. .................................................. 69
Table 3. 4. Characteristics of the liquid digestate and treated water after the VSSF-CTW during
both low and high precipitation periods........................................................................................ 78
Table 3. 5. Energy balance of the SPAD with the VSSF-CTW and the control system. ............. 81
Table 3. 6. Cash flow analysis. ..................................................................................................... 84
ix

Table 3. 7. Sensitivity analysis of key unit operations on the payback period of the current
system. .......................................................................................................................................... 86
Table 4. 1. Contaminant removal models for designing CTWs. .................................................. 93
Table 4. 2. Hydrological balance at the HCTW.......................................................................... 108
Table 4. 3. Characteristics of studied parameters at the effluent tank, the VSSF-CTW, and the
FWS-CTW for both high (n =16) and low (n = 9) precipitation periods. ................................... 113
Table 4. 4. Sediments collected from the geotextile membrane at the VSSF-CTW. ................. 120
Table 4. 5. Mass removal from the geotextile membrane and from the VSSF-CTW. ............... 122
Table 4. 6. Key water quality parameters for the HCTW during both high and low precipitation
periods. ........................................................................................................................................ 123
Table 4. 7. Contaminant removal models approach for the VSSF-CTW. .................................. 128
Table 4. 8. Contaminant removal models approach for the FWS-CTW. ................................... 130
Table 5. 1. Given parameters for estimating of the void space in the filter media and the time
before clogging at the full-scale VSSF-CTW. ............................................................................ 144
Table 5. 2. Statistical analysis. .................................................................................................... 147
Table 5. 3. Mass removals for COD, TS, TN, and TP at VSSF-CTW *. ................................... 150
Table 5. 4. Statistical p-value and power for solid accumulation analysis in the vertical-flow
wetland. ....................................................................................................................................... 154
Table 5. 5. Statistical p-value and power for infiltration rates in the VSSF-CTW. .................... 156
Table 5. 6. Effluent water quality parameters in the columns (mean Âą S.D., n = 24). ............... 160
Table 5. 7. Total dry root mass per plant treatment after the experiment (mean Âą S.E., n = 9). 163
Table 5. 8. Relative infiltration rates rankings for plant treatments (n = 36). ............................ 165
Table 6. 1. Inputs and outputs chemical characterization. .......................................................... 190

x

Table 6. 2. Energy consumption per piece of equipment in the SPAD-HCTW. ........................ 191
Table 6. 3. SPAD performance during the rainy and dry period. ............................................... 194
Table 6. 4. Exergy-based assessment of sustainability for the SPAD-HCTW. .......................... 197
Table 6. 5. Exergy indexes for sustainability. ............................................................................. 198

xi

LIST OF FIGURES

Figure 1. 1. The SPAC-HCTW located at EEAFBM, Alajuela, Costa Rica. Satellite image was
taken from Google Earth. ................................................................................................................ 5
Figure 1. 2. Removal of COD and TS through the system. Reactor (S1): sampling of the mixture
before feeding. Effluent (S3): sampling of the effluent after digestion. Inlet cell 3 (S4): sampling
of the effluent from cell 2 entering cell 3. Inlet cell 4 (S5): sampling of the effluent from cell 4
entering cell 4. Outlet cell 4 (S6): sampling of the effluent from cell 4. ...................................... 12
Figure 1. 3. Conceptualization of the SPAD-HCTW system. Abbreviations: AD: Anaerobic
digester; S/L separator: Solid/liquid separator; VSSF-CTW: Vertical subsurface flow constructed
treatment wetland; FWS-CTW: Free water surface constructed treatment wetland; HCTW:
Hybrid constructed treatment wetland. ......................................................................................... 15
Figure 2. 1. Dynamics of the water-energy-food nexus in the economy (Own creation)............. 20
Figure 2. 2. A schematic of the system configuration of integrated solar thermal collectors and
anaerobic digesters. a) Alkhamis, El-khazali et al. (2000); b) Axaopoulos, Panagakis et al. (2001)
and Axaopoulos and Panagakis (2003); and c) and d) El-Mashad, van Loon et al. (2004). ....... 40
Figure 2. 3. A schematic of a HSSF wetland. Figure was taken from Kadlec and Wallace (2009).
....................................................................................................................................................... 44
Figure 2. 4. A schematic of a VSSF wetland. Figure was taken from Kadlec and Wallace (2009).
....................................................................................................................................................... 47
Figure 2. 5. A schematic of a FWS wetland. Picture was taken from Kadlec and Wallace (2009).
....................................................................................................................................................... 50
Figure 2. 6. Entropy, exergy and sustainability (Hornbogen 2003)............................................. 55
Figure 3. 1. Flowchart of the SPAD and VSSF-CTW. Mass flow is represented with a continuous
line (-). Energy flow is represented with a hidden line (- -). ........................................................ 62
Figure 3. 2. Individual units of the system at the Fabio Baudrit Agricultural Station. a. Solar
thermal collectors, b. Grinder, c. Feeding tank, d. Thermophilic CSTR-anaerobic digester (silver
tank) and hot water tank (green tank), e. Liquid/solid separator, f. Effluent storage tank, g.
Biogas bag, h. Engines, i. VSSF-CTW. ........................................................................................ 62

xii

Figure 3. 3. Temperature profile of the solar thermal collection unit and thermophilic CSTRanaerobic digester. Data for February 2016 are presented. .......................................................... 75
Figure 3. 4. Change of water quality during the process. ............................................................. 79
Figure 3. 5. Mass balance of the SPAD with the VSSF-CTW. Densities for CH4 and CO2 were
estimated as 0.717 kg/m3 and 1.84 kg/m3. .................................................................................... 80
Figure 3. 6. The cash flow of the SPAD with the VSSF-CTW under the current operational
conditions. Land costs were not considered. ................................................................................ 85
Figure 4. 1. Aerial image of the HCTW. Satellite image was taken from Google Earth. ............ 99
Figure 4. 2. Photo and schematic of VSSF-CTW. Dimensions are in meters. Picture was taken in
August 2015. ............................................................................................................................... 101
Figure 4. 3. Photograph and schematic diagram of FWS-CTW. Dimensions are in meters. Picture
was taken in August 2015. .......................................................................................................... 103
Figure 4. 4. VSSF-CTW during the medium to high precipitation phase. a. A dilution effect
could occur due to inputs such as precipitation and runoff. b. Berms had the capacity to hold
water over the filter media and avoid spillover. Pictures were taken in October 2015. ............. 109
Figure 4. 5. FWS-CTW during the low precipitation period. a. Low HLR, scarce precipitation
events, and high temperatures substantially decreased the water at the FWS-CTW. b. Growth of
algae bloom. Pictures were taken in February 2016. .................................................................. 116
Figure 4. 6. Geotextile membrane placed in the VSSF-CTW inlet area. a. Liquid digestate was
discharged by a PVC pipe (diameter = 0.05 m) into the geotextile membrane. b. Solids were
accumulated on the geotextile while the liquid passed through the membrane and infiltrated in
the filter media of the VSSF-CTW. c. A worker cleaning up the geotextile membrane. Pictures
were taken in October 2014. ....................................................................................................... 120
Figure 4. 7. Relationship between mass loading and outlet concentration at the VSSF-CTW for
A) COD during the high precipitation period, B) TS during the low precipitation period, C) COD
during the low precipitation period, and D) TN during the low precipitation period. ................ 125
Figure 4. 8. Relationship between mass loading and outlet concentration at the FWS-CTW for A)
TS during the high precipitation period, B) NH4 during the high precipitation period, C) FS
during the low precipitation period, D) TS during the low precipitation period, and E) TP during
the low precipitation period. ....................................................................................................... 126

xiii

Figure 5. 1. Full-scale VSSF-CTW: Photos of the location of the inlet and geotextile membrane
and the health of the plants during this study. Schematic of the full-scale vertical-flow wetland
(all lengths are in meters). Pictures were taken in August 2015. ................................................ 140
Figure 5. 2. Division of the VSSF-CTW per zones. Picture was taken in March 2013, when the
VSSF-CTW was recently built. .................................................................................................. 140
Figure 5. 3. Laboratory-scale VSSF-CTW. Picture was taken in July 2015, during establishment
of plants in the columns. ............................................................................................................. 142
Figure 5. 4. Preventive strategies applied for reducing solids in the liquid digestate. a. Rotary
liquid/solid separation unit; and b. effluent storage tank. Pictures were taken in January 2016. 148
Figure 5. 5. The geotextile membrane: a. during liquid digestate influent flow; b. during filtration
of pooled digestate; and c. during removal of sediments. Pictures were taken in October 2014.
..................................................................................................................................................... 149
Figure 5. 6. Mass balance on the VSSF-CTW. Figure is not to scale. ....................................... 151
Figure 5. 7. a. Mass balance of the solids to estimate the void space in the filter media and the
time before clogging; b. Photos of wastewater before and after the VSSF-CTW. ..................... 152
Figure 5. 8. Solid accumulation (as g VS//L) in the VSSF-CTW. T1: January 2016, T2: February
2016, and T3: March 2016. Z1: zone 1, Z2: zone 2, and Z3: zone 3. Statistical analysis applied:
Kruskal-Wallis, using Wilcoxon non-parametric multiple comparison at alpha level of 0.05. N =
9, except for Cyperus (n=6). ....................................................................................................... 154
Figure 5. 9. Total dry root biomass per species (n = 9, except for Cyperus (n = 6)) established at
the VSSF-CTW. Abbreviations: Iris (IG), Canna (CI), Cyperus (CP), and control (C). Position 1
(center of the block); position 2 (0.50 m from the center of the block); and position 3 (1.0 m from
the center of the block). .............................................................................................................. 155
Figure 5. 10. Infiltration rates at the VSSF-CTW (n = 9, except for Cyperus (n = 6)). T1: January
2016, T2: February 2016, and T3: March 2016. Z1: zone 1, Z2: zone 2, and Z3: zone 3.
Statistical analysis applied: two way analysis of variance at alpha level of 0.05. ...................... 156
Figure 5. 11. Infiltration rates per positions at the VSSF-CTW. Position 1 (center of the plant);
position 2 (0.50 m from the center of the plant); and position 3 (1.0 m from the center of the
plant). Letters A, B, and C indicate significant differences between infiltrations per position per
each plant treatment. Statistical analysis applied: Turkey-Kramer HSD at alpha level of 0.05. 158
Figure 5. 12. Infiltration rates in the VSSF-CTW as a function of root biomass for Canna,
Cyperus, and Iris. ........................................................................................................................ 159
xiv

Figure 5. 13. a. VS accumulation unplanted columns and six of the seven plant treatments
(Cyperus, Canna-Cyperus, Canna, Canna-Iris-Cyperus, Canna-Iris, and Iris-Cyperus); b. VS
accumulation for Iris plant treatment. Note. A: from surface to 0.05 m deep, B: from 0.05 to 0.10
m deep, C: from 0.10 to 0.15 m deep, and D: from 0.15 to 0.30 m deep. .................................. 162
Figure 5. 14. a. Infiltration rates (median, n = 9) for Canna, control, Cyperus, and Iris based on
time. b. Infiltration rates (median, n = 9) for Canna-Cyperus, Canna-Iris, Canna-Iris-Cyperus,
and Iris-Cyperus based on time. T1: August 2015, T2: September 2015, T3: November 2015,
and T4: February 2016. ............................................................................................................... 164
Figure 5. 15. a. Root structure for Canna (Picture was taken from www.plantsgrow.com); b.
Root structure for Cyperus (Picture was taken from www.olabrisagardens.com); and c. Root
structure for Iris (Picture was taken from www.grovida.us). ..................................................... 167
Figure 6. 1. Flowchart of the SPAD-HCTW system. Numbers indicate pumps and other
equipment needed to operate the SPAD-HCTW: 1. Solar heating fluid transfer pump, 2. Digester
heating pump, 3. Digester mixing pump, 4. Feeding tank pump, 5. Solid/liquid separator, 6.
Grinder, 7. Effluent pump, 8. Feed preparation pump, 9. SPAD-HCTW exit pump, 10. Biogas
flowmeter, 11. VSSF-CTW recirculation pump, and 12. FWS-CTW recirculation pump. ....... 177
Figure 6. 2. Boundary definitions for conducting the exergy-based calculation of sustainability
for the SPAD-HCTW. Case 0 – the SPAD. Case 1 – the SPAD with the VSSF-CTW. Case 2 –
SPAD with the VSSF-CTW and FWS-CTW. ............................................................................ 179
Figure 7. 1. Linear versus circular economy (Own creation). .................................................... 212
Figure 7. 2. Closing the loop for water, energy, and food (Own creation). ................................ 214
Figure 7. 3 . The SPAD-HCTW as a sustainable approach for the sanitation problem in Costa
Rica. ............................................................................................................................................ 222

xv

KEY TO ABBREVIATIONS

ACEPSA

Central American Association for Economy, Health and Environment

ADREC

Anaerobic Digestion Research and Education Center

AyA

Costa Rican Water and Sewage Institute

CapEx

Capital expenditures

CIA-UCR

Agronomy Research Center, University of Costa Rica

CELEQ

Center for Research in Electrochemistry and Chemical Energy

CHP

Combined heat and power

CICA-LCA

Water Quality Laboratory-Research Center of Environmental Pollution

CSTR

Continuous stirred tank reactor

CTW

Constructed treatment wetland

DAQ

Data acquisition system

ECA

Costa Rican Accreditation Institute

EEAFBM

Fabio Baudrit Experimental Station

FAO

Food and Agriculture Organization of the United Nations

FWS-CTW

Free water surface constructed treatment wetland

GDP

Gross domestic product

GHG

Greenhouse gas

HCTW

Hybrid constructed treatment wetland

HDPE

High-Density-Polyethylene

HSSF-CTW

Horizontal subsurface flow constructed treatment wetland

HLR

Hydraulic loading rate
xvi

HRT

Hydraulic retention time

ICE

Costa Rican Electricity Institute

IRET

Regional Institute for Studies on Toxic Substances

MINAET

Ministry of the Environment and Energy

MINSALUD

Ministry of Health

MACRS

Modified Accelerated Cost Recovery System

MSU

Michigan State University

OLR

Organic loading rate

OpEx

Operational expenditures

SPAD-HCTW

Solar-powered anaerobic digester-hybrid constructed treatment wetland

TEC

Technological Institute of Costa Rica

TIS

Tank in series

UASB

Up-flow anaerobic sludge blanket

UNA

National University

UCR

University of Costa Rica

VSSF-CTW

Vertical subsurface flow constructed treatment wetland

xvii

CHAPTER 1: INTRODUCTION

This chapter introduces the problem regarding biomass residues and wastewater in Costa
Rica, the decentralized self-sufficient, close-loop, organic waste treatment system, and the scope
and aim of this research. The problem regarding wastes in Costa Rica justifies technology
selection: anaerobic digestion, solar thermal collectors, and constructed treatment wetlands. The
integration of these technologies is presented and named as the solar-powered anaerobic
digestion and hybrid constructed treatment wetland system or, simple, the SPAD-HCTW. The
SPAD-HCTW is proposed as a decentralized self-sufficient, close-loop, organic waste treatment
system for agro-industrial and agricultural activities in rural areas of Costa Rica. An overview of
the SPAD-HCTW presents where the system was constructed, basis of design, and preview
performance before this study. Finally, the configuration of the system studied in this dissertation
and the scope and aims of this research are presented.

1.1. Biomass residues and energy production
Costa Rica is advancing toward carbon neutrality, with a goal of being carbon neutral by
2021. One action that reduces carbon emissions is to substitute fossil fuels with renewable
energy sources (Coto 2013). In Costa Rica, it is estimated that agricultural and agro-industrial
activities produce 6,000 metric tons of biomass residues per year. These residues could
potentially be used to generated 600 MW of electricity (MINAET 2011, Coto 2013), or 23% of
the installed electrical generation capacity of 2,600 MW in Costa Rica (Coto 2013). Currently,
only 0.73% (18.9 MW) of Costa Rica’s energy is produced from biomass residues, which
indicates that a vast potential of energy is wasted (EN 2015).

1

Anaerobic digestion is a suitable technology for handling mixtures of biomass residues
with high moisture content (65 – 99.9%), such as animal manures, food wastes, sewage
wastewater, and other industry organic residues (IRENA 2012, Funk, Milford et al. 2013).
Anaerobic digestion is a biological process performed by methanogenic communities that
convert biomass residues into biogas in the absent of oxygen (Gould and Crook 2010) while
generating solid and liquid digestate. Anaerobic digesters typically operate at different
temperatures: psychrophilic (<25°C), mesophilic (25-40°C), and thermophilic (>45°C) (ElMashad, Zeeman et al. 2004, Suryawanshi, Chaudhari et al. 2010). As reaction rates by
microorganisms increase with increasing temperatures, digestion yields higher biogas production
with higher temperatures. Thus, in contrast to mesophilic anaerobic digestion, thermophilic
anaerobic digestion is characterized by higher microbial growth rates, higher solids reduction,
improved solid-liquid separation, higher destruction of pathogens, and improved odor control
(Zabranska, Dohanyos et al. 2002, Suryawanshi, Chaudhari et al. 2010). However, thermophilic
anaerobic digestion requires higher input energy to maintain temperatures greater than 45°C.
To achieve thermophilic temperatures, combined heat and power (CHP) systems use a
portion of the biogas produced to heat the digestate (Zabranska, Dohanyos et al. 2002). However,
for small-scale digesters, biogas utilization for heating can result in negative energy balances
(Vindis, Mursec et al. 2009), and for large-scale digesters, heating greatly reduces surplus energy
(Zabranska, Dohanyos et al. 2002). Solar energy is an alternative, renewable energy source for
heating digestion processes. In tropical regions, solar thermal collectors can satisfy the energy
requirement necessary to maintain thermophilic temperatures year-round without any additional
energy sources. Therefore, this study implemented flat-plate solar thermal collectors to provide
heat for a thermophilic continuous stirred tank reactor (CSTR) anaerobic digester in Costa Rica.

2

1.2. Wastewater treatment and the sanitation problem
In Costa Rica only 5.81% of the 0.086 km3/year of domestic wastewater is treated (Sato,
Qadir et al. 2013). In addition, treatment of industrial wastewater is infrequent; for example, only
175 of 3,500 industries in the Rio Virilla watershed, which represents only 0.3% of Costa Rican
territory, treat their wastewater (UN 2009, Coto 2013, Echeverria and Cantillo 2013). Therefore,
it is clear that wastewater treatment of any source, including agricultural and agro-industrial
activities, is frequently neglected. Consequently, untreated wastewater is likely impacting
aquatic ecosystems, potentially causing health problems and environmental degradation (Dallas,
Scheffe et al. 2004, UN 2015).
Wastewater treatment aims to reduce the level of pollutants in the wastewater before
reuse or disposal into the aquatic ecosystems, thereby alleviating its negative impacts.
Ecologically-based treatment systems tend to be less costly and sophisticated in operation and
maintenance. For example, constructed treatment wetlands (CTWs) are biological systems with
low capital cost and simple operation, with potentially efficient removal of pollutants; however,
CTWs require preliminary treatment of the raw wastewater and demand relatively larger areas
than centralized wastewater treatment facilities (the latter of which is usually not critical for rural
areas) (Dallas, Scheffe et al. 2004). In this study, CTW technology is proposed to treat the liquid
effluent from the thermophilic CSTR anaerobic digester in the tropics.

1.3. Self-sufficient energy production and waste treatment system
Even though biomass residues and wastewater pose an environmental concern to the
aquatic ecosystems, both of them are rich in carbon and nutrients, representing a potential source

3

of energy. Treating and utilizing the biomass residues and wastewater can lead to economic,
social, and environmental benefits in rural areas (UN 2015). As stated by the Report of the
Special Rapporteur on the human right to safe drinking water and sanitation, Catarina de
Albuquerque, in 2013, “development must be sustainable and must protect the environment on
which present and future generation depend” (UN 2013). One action is to constrain resource
consumption and waste generation, to satisfy human needs. Therefore, an integrated biomass and
wastewater treatment system that can provide energy, fertilizers, and reclaimed water could aid
sustainable development for rural areas in the tropics. The solar-powered anaerobic digester and
hybrid constructed treatment wetland system (SPAD-HCTW) combines solar thermal collection,
anaerobic digestion, and CTW technologies to treat biomass residues and wastewater from agroindustrial and agricultural activities, while producing energy, fertilizers, and reclaimed water.
Thus, the SPAD-HCTW can be offered as one approach for solving the problem of water
pollution due to agro-industrial and agricultural waste disposal in rural areas of Costa Rica.

1.4. The solar-powered anaerobic digestion and hybrid constructed treatment
wetland system
In 2011-2012, the SPAD-HCTW was designed and constructed. The design corresponded
to knowledge and expertise developed in temperate regions with the intention of adapting the
technology into the tropics. The system is located at the Fabio Baudrit Experimental Station
(EEAFBM), in the province of Alajuela, Costa Rica (10.00 m N, -84.26 m W). The altitude of
this site is 840 m above the level of the sea. The average annual precipitation is 1,940 mm,
distributed from May to November, and the average annual temperature is 22°C (IMN 2016).
EEAFBM is owned and operated by the University of Costa Rica (UCR) and promotes the

4

holistic development of sustainable agro-industrial and agricultural activities in Costa Rica. One
of EEAFBM’s missions is to research innovative treatment/reutilization methods for biomass
residues and wastewater. Therefore, this site was selected to house the SPAD-HCTW. Moreover,
several organic wastes can be provided on-site (e.g., chicken litter and crop residues). In
addition, EEAFBM is in a rural area, close to agro-industries and farms potentially interested in
the technology (e.g., cow manure, coffee residues, and food wastes).
The SPAD consists of two integrated technologies that cover 180 m2: the solar thermal
collection unit and the anaerobic digestion unit. The HCTW consists of four constructed
treatment wetlands, which covers an effective treatment area of 576 m2 (Figure 1.1).

Figure 1. 1. The SPAC-HCTW located at EEAFBM, Alajuela, Costa Rica. Satellite image was
taken from Google Earth.

The proposed system was designed based on previous studies at the Michigan State
University (MSU) Anaerobic Digestion Research and Education Center (ADREC). The
estimated mass and energy balance for the proposed system in Costa Rica was based on a study

5

of a 20-L bench-scale solar-biopower system at MSU. The design of the HCTW followed Healy,
Rodgers et al.(2007) and Kadlec and Wallace (2009) approaches for intermittent sand filters and
wetlands for wastewater; however, the size of the HCTW was greatly reduced due to space
constraints. The entire system was designed for the flow rate of one cubic meter per day. Table
1.1 shows the technical parameters of individual units of the system.

6

Table 1. 1. Description and basis of design.

Solar thermal collection

Unit

Component

Estimated value

Current value at
EEAFBM

Basis of design

Flat-plate
thermal
collectors

50 m2

36 m2

50 m2 based on an annual irradiance of 10 MJ/ m2. Area was reduced to 36 m2 based on
new estimations on-site and area and budget constraints.

Hot water tank

5 m3

5 m3

Stores heated water from the flat-plate solar thermal collectors.

Hot water pump
(UP 26-99 F
from Grundfos,
Olathe, KS)

n.d.

0.17 hp

Circulates water from the flat-plate thermal collectors to the hot water tank and viceversa. Every day, a timer controls the hot water pump, which works from 9 am to 4 pm
(solar radiation time at EEAFMB).

Hot water pump
(PB 351MA
from Wilo,
Korea)

n.d.

0.46 hp

Circulates hot water from the hot water tank through a High-Density-Polyethylene
(HDPE) tubing heat exchanger to heat the digester and maintain the thermophilic
temperature (45 ¹ 2°C).

HDPE tubing
heat exchanger

n.d.

40 m

Originally, a copper tubing heat exchanger was installed. This material was corroded by
biogas inside the AD, thus HDPE tubing was chosen for replacing the broken copper
tubing.

--

Data acquisition system (DAQ) collects data from thermocouples (type K, probe
ungrounded) within the digester every 20 seconds for feedback control of the digestion
temperature (¹2°C of the set temperature of 45°C). The DAQ sends a digital signal to
power the hot-water pump if the digester temperature was lower than the set
temperature.

Data acquisition
system
(CR1000
Campbell
Scientific,
Logan, UT)

--

7

Table 1. 1. (cont’d)

Electric
generator

Thermophilic CSTR anaerobic digester

Unit

Component

Estimated value

Current value at
EEAFBM

Anaerobic
digester tank

Volume: 20 m3
Flow rate: 1m3/d
HRT: 10-15 days
Mixture
characteristics
TS: 10%
COD: 90 g/L

Volume: 20 m3
Flow rate: 1m3/d
HRT: 20 days

Basis of design

Cylindrical tank with flat bottom made with HDPE. A submersible digester mixing
pump (model 5763 from ATM, Royersford, PA) operates 10 minutes each hour to
assure homogeneous conditions in the digestate. TS was set to 2% based on desired
biogas production of 20 m3/d.

Feeding tank

n.d.

10 m3

Cylindrical tank with flat bottom made with HDPE. Stores the mixture to be fed into the
digester. Weekly, five cubic meters of mixture is prepared with chicken litter and food
waste at an average ratio of 1:12 (dry mass basis) and mixed with five cubic meter of
reclaimed water from the HCTW. This ratio was selected to assure 2.2% of total solids
in the mixture. The feeding tank has an external pump (model AMTP/N 1626-305-00
from ATM, Royersford, PA) for mixing (six minutes before feeding) and feeding the
mixture into the anaerobic digester.

Effluent storage
tank

n.d.

10 m3

Cylindrical tank with flat bottom made with HDPE. Stores liquid digestate. Has
sufficient storage for two-weeks’ worth of digestate.

Biogas bag

60 m3

60 m3

Biogas bag made with HDPE. Stores biogas for engine use to maintain a steady power
and heat flow. Size was selected based on biogas production. Initial calculation
estimated a biogas production of 20 m3/d.

Combined heat
and power
(CHP) engine

One engine
30 kW

Two engines
16 kW each one

Power was chosen based on the initial energy estimation of 684 MJ/d. CHP engine
(BrancoÂŽ B4T-5000 Bioflex, Brazil) powers pumps and other pieces of equipment in
the system to satisfy operational requirements.

8

Table 1. 1. (cont’d)

Constructed treatment wetland (CTW)

Unit

Component

Estimated value

Area

1,100 m2
Inlet flow rate:
1m3/d.
COD inlet
concentration:
45,000 mg/L.

Hybrid
constructed
treatment
wetland

Current value at
EEAFBM

576 m2

Basis of design
Area of 1,100 m2 was calculated using the tank in series (TIS) model (Kadlec and
Wallace 2009) for a free water surface constructed treatment wetland. Maximum
effluent concentrations are the discharge standards for Costa Rica (COD < 150 mg/L,
TS < 50 mg/L, TN < 50 mg/L, TP < 8 mg/L, 5 < pH < 9, and 15 °C < T < 40 °C)
(MINAE-MSP 2007). Final area of 576 m2 was chosen due to space constraints and
research goals at EEAFBM.
Cell 1: Intermittent sand filter. Cell 2: Vertical subsurface flow constructed treatment
wetland (VSSF-CTW). Cells 3 and 4: free water surface constructed treatment wetland
(FWS-CTW) planted with floating plants.

n.d.

4 cells.

Dimensions. Top: 12 x 12 m square. Bottom: 9 x 9 m square. Depth: 1.1 m. Walls’
slope: 27° with respect to the horizontal. All cells are interconnected, and each cell has a
submersible pump (WS V52 from Franklin Electric, Fort Wayne, IN).

Cells 1 and 2

--

--

Media profile, from bottom to top: 0.2 m of stone (particle size of 12-20 mm), 0.2 m of
pea gravel (particle size of 4-8 mm), and 0.7 m of coarse sand (particle size of 0.75-2
mm, 32% porosity). Media profile and particle size were chosen based on Healy,
Rodgers et al. (2007).

Cell 2
planting
selection

Juncus effusus
Coix Lacryma jobi
Cyperus papyrus
Iris graminea
Canna indica

--Cyperus papyrus
Iris graminea
Canna indica

Plants were planted in March 2013. Five emergent plants were chosen based on
availability on-site and literature review of wetlands in the tropics. Only Cyperus
papyrus, Iris graminea, and Canna indica were considered in this study as the other two
plants did not s survive in the wetland.

2:00 am to 4:00
pm

2:00 am to 4:00
pm

Height: 1.60 m

Height: 1.60 m

4 upright
sprinklers

4 upright
sprinklers

Cell 2
recirculation

Recirculation at night to avoid higher losses of water due to higher evaporation rates
during the day.
Height of the recirculation spray was chosen to decrease interference of the
redistribution spray by plants.
Sprinklers were located at each corner of a square (7.00 × 7.00 m) centered at the center
of the wetland to reach maximum wetland surface area.

9

Table 1. 1. (cont’d)

(CTW)

Unit

Component

Estimated value

Current value at
EEAFBM

Cell 2
geotextile
membrane

n.d.

2 x 2 x 0.25 m3
(length x width x
depth)

Dimensions were chosen based on the discharge of 1 m3 into the wetland.

Cell 4
planting
selection

Eichornia
crassipes
Pistia stratiotes
Salvinia

Eichornia
crassipes
Pistia stratiotes
--

Plants were planted in March 2013. Three floating plants were chosen based on
availability on-site and literature review of wetlands in the tropics. Only Eichornia
crassipes and Pistia stratiotes were considered in this study as Salvinia was no longer
present in August 2015.

Basis of design

** TS: total solids
* COD: chemical oxygen demand

10

1.5. Overview of the system evaluated in this study
The SPAD-HCTW began operation in March 2013. One cubic meter of food waste and
animal manure was fed into the anaerobic digester and one cubic meter of digestate was
discharged into the HCTW daily. Initially, the HCTW operated in series, from cell 1 to cell 4,
transferring one cubic meter per day between cells. The liquid effluent was discharged into cell 1
or cell 2 due to high concentration of chemical oxygen demand (COD), total solids (TS), and
ammonium (NH4). After observation of leaking in cell 1 during the first few months of
operation, digestate was only discharged into cell 2. The goal of the first two cells was to remove
large portions of COD and TS while reducing NH4 concentrations. Intermittent sand filters and
vertical subsurface flow constructed treatment wetlands (VSSF-CTWs) provide oxidative
conditions for degradation of solids trapped in the filter media and transformation of ammonium
to nitrate. Treated water from cell 2 was transferred to cell 3, then to cell 4, for polishing. Free
water surface constructed treatment wetland (FWS-CTW) provides anaerobic zones for
denitrification to reduce concentrations of nitrate (NO3).
The SPAD-HCTW has operated for 30 months, from March 2013 to July 2015. During
this period, little data on treatment performance was collected. In 2014, COD and TS were
measured at the HCTW (Figure 1.2) and fecal coliforms were evaluated (Table 1.2). In
December 2014, Lambda Laboratories (certified by the Costa Rican Accreditation Institute
(ECA)) conducted a sampling campaign on the entire system (Table 1.3). Results indicated that a
reduction of solids (56% of suspended solids and 36% of total organic carbon) was occurring in
the anaerobic digester (Table 1.3). These data were inconsistent with the data shown in Figure
1.2, where COD and TS were not removed by the digester. Consistently, Figure 1.2 and Table
1.3 show that cell 2, the VSSF-CTW, removed large portions of TS, COD, total organic carbon,

11

and nutrients (as total nitrogen (TN) and total phosphorus (TP)); while, the cell 3 and cell 4, the
FWS systems, polished the treated water. The evaluation of fecal coliforms indicated that the
treated water was pathogen free, except for a single detectable result at the effluent from the
anaerobic digester in October 2014.

Figure 1. 2. Removal of COD and TS through the system. Reactor (S1): sampling of the mixture
before feeding. Effluent (S3): sampling of the effluent after digestion. Inlet cell 3 (S4): sampling
of the effluent from cell 2 entering cell 3. Inlet cell 4 (S5): sampling of the effluent from cell 4
entering cell 4. Outlet cell 4 (S6): sampling of the effluent from cell 4.

12

Table 1. 2. Evaluation of fecal coliforms after digestion.
Dates
Sampling point

Digester effluent
(S3)
Effluent cell 2/
Inlet cell 3 (S4)
Effluent cell 3/
Inlet cell 4 (S5)
Outlet cell 4 (S6)

22/10/14

3/12/14

10/12/14

MPN

CFU/mL

CFU/mL

Detectable

Non-detectable

Non-detectable

Non-detectable

5.1 x 103

Non-detectable

Non-detectable

3.2 x 102

Non-detectable

Non-detectable

Non-detectable

Non-detectable

Notes. MPN: Most Probable Number. CFU: Colony-forming unit. Effluent (S3):
sampling of the effluent after digestion. Inlet cell 3 (S4): sampling of the effluent from cell 2
entering cell 3. Inlet cell 4 (S5): sampling of the effluent from cell 4 entering cell 4. Outlet cell 4
(S6): sampling of the effluent from cell 4.

Table 1. 3. Water quality analysis of entire system from Lambda Laboratory (Dec. 2014).
TOC

SS

TN

TP

Sampling points
(mg/L)
Allowed by law in Costa Rica

150

50

50

25

Reactor (S1)

6,460 Âą 130

13,840 Âą 6

1,750 Âą 40

107 Âą 1

Digester effluent (S3)
Effluent cell 2/
Inlet cell 3 (S4)
Effluent cell 3/
Inlet cell 4 (S5)
Outlet cell 4 (S6)

4,140 Âą 80
13 Âą 1

6,200 Âą 6
62 Âą 6

1,675 Âą 40
5.06 Âą 0.5

73 Âą 1
0.10 Âą 0.02

8.5 Âą 0.5

26 Âą 6

2.8 Âą 0.5

0.04 Âą 0.02

9.0 Âą 0.5

12 Âą 6

3.5 Âą 0.5

> 0.02

Note. Reactor (S1): sampling of the mixture before feeding. Effluent (S3): sampling of
the effluent after digestion. Inlet cell 3 (S4): sampling of the effluent from cell 2 entering cell 3.
Inlet cell 4 (S5): sampling of the effluent from cell 4 entering cell 4. Outlet cell 4 (S6): sampling
of the effluent from cell 4. TOC: Total organic carbon. SS: Suspended solids. TN: Total
nitrogen. TP: Total phosphorus.

13

1.6. The system studied in this dissertation
This study evaluated the SPAD-HCTW in detail from August 2015 to March 2016. This
study represented the first campaign of continuous (weekly) monitoring of the system. Herein,
the HCTW consisted of the VSSF-CTW, cell 2, and the FWS-CTW, cell 4. Budget constraints
limited the inclusion of the entire system (4 cells). However, the VSSF-CTW working in-series
with the FWS-CTW allowed the evaluation of treatment performance and clogging. A
description of the SPAD-HCTW is shown in Figure 1.3. The CSTR anaerobic digester was
assisted by flat-plate solar thermal collectors to maintain thermophilic conditions during
digestion. Produced biogas was used as a fuel for running two electric generators (16 kW each),
which produced electricity for on-site uses. Solid digestate was composted and used for
fertilizing the crops at EEFBM. Liquid digestate was further treated by the HCTW, which
included two CTWs working in-series, a VSSF-CTW followed by a FWS-CTW. Treated water
was used for irrigation, reused in the anaerobic digester, or discharged into surface water.

14

Figure 1. 3. Conceptualization of the SPAD-HCTW system. Abbreviations: AD: Anaerobic
digester; S/L separator: Solid/liquid separator; VSSF-CTW: Vertical subsurface flow constructed
treatment wetland; FWS-CTW: Free water surface constructed treatment wetland; HCTW:
Hybrid constructed treatment wetland.

1.7. Scope and aims of this research
Treatment of biomass residues and wastewater from agro-industrial and agricultural
activities by the SPAD-HCTW is proposed as a decentralized self-sufficient, close-loop, organic
waste treatment system technology for rural areas at the tropics. Therefore, the scope of this
dissertation is to evaluate the treatment performance and exergetic sustainability of the SPADHCTW. In particular, the following principal research question will be addressed: Does the
SPAD-HCTW sustainably convert biomass residues and wastewater into energy, fertilizers, and
treated water, while meeting standards for waste discharges into the environment? The
following four main questions were developed to answer the principal research question.
First, is the SPAD a self-sufficient energy production and treatment system for providing
energy and fertilizers from biomass residues, and treated water? To answer this question, the
integrated SPAD with a VSSF-CTW was evaluated in terms of technical performance. In
particular, this study evaluated the technical performance of: 1) solar thermal collectors for
15

maintaining thermophilic temperature in the anaerobic digester; 2) thermophilic anaerobic
digestion for converting organic wastes into energy and fertilizers; and 3) VSSF-CTW for
treating wastewater (i.e., liquid digestate). These evaluations were important to demonstrate that
the SPAD is a self-sufficient, close-loop, energy production and waste treatment system. In
addition, a cash flow analysis was conducted to determine the payback period of the current
system. Technical performance of the solar-powered waste utilization and treatment system in
Costa Rica is presented in Chapter Three.
Secondly, how does the performance of the HCTW respond to precipitation conditions at
EEAFBM? Better treatment performance is expected during the precipitations due to dilution of
pollutants in the wetlands. To answer this question, the treatment performance of the HCTW
during rainy and dry periods at EEAFBM was analyzed. In addition, the applicability of loading
charts and contaminant removal models commonly used to describe temperate wetlands to model
treatment performance of the tropical HCTW was evaluated. Performance of the hybrid
constructed treatment wetland is presented in Chapter Four.
Thirdly, are implemented preventive strategies in the VSSF-CTW positively impacting in
the longevity of the filter media by decreasing clogging? Clogging of VSSF-CTWs is the largest
challenging to maintaining a high performing system that can effectively treat wastes with high
solids concentrations. Strategies to prevent clogging that were implemented in the VSSF-CTW
were wastewater pretreatment and inlet distribution of the wastewater. Wastewater pretreatment
included a rotary liquid/solid separation unit, an effluent storage tank that allowed settling, and a
geotextile membrane. Influent was discharged into a geotextile membrane, which both
distributed the influent and further filtered out solids. Finally, recirculation of treated wastewater
distributed the partially treated water over the entire treatment area. To answer this question, the

16

longevity of the sand media in the VSSF-CTW was evaluated for solid accumulation, root
development, and infiltration. In addition, laboratory-scale VSSF-CTW was built to determine
how different root structures affect clogging. These evaluations were important to determine
practices that expand the lifespan of the VFSS-CTW. The clogging in tropical vertical subsurface
flow constructed treatment wetlands is presented in Chapter Five.
Finally, is the SPAD-HCTW exergetically sustainable technology? This research question
was answered by conducting an exergy-based assessment of sustainability using the
environmental exergy efficiency (Ρenv,ex) and the total pollution rate (Rpol,ex) indexes. A system
that converts materials with high entropy into high quality end products with low entropy will be
sustainable if the balance between inputs (e.g.: wastes) and outputs (e.g.: energy, fertilizers, and
treated water) is positive (Wall 2010, Woudstra 2016). Assessments were conducted for the
exergy baseline or the SPAD alone (case 0), the SPAD and the VSSF-CTW (case 1), and the
SPAD-HCTW (case 2) to compare potential improvement of sustainability through inclusion of
wetlands during treatment of agricultural wastes. The exergy-based assessment of sustainability
of a solar-powered anaerobic digestion and hybrid constructed treatment wetland system to treat
agricultural wastes in Costa Rica is presented in Chapter Six.
Conclusions, limitations, and future work, in relation with the SPAD-HCTW as a
decentralized wastewater treatment facility for rural areas in the tropics, are presented in Chapter
Seven.

17

CHAPTER 2: LITERATURE REVIEW

This chapter introduces biomass residues and wastewater production from agro-industrial
and agricultural activities in Costa Rica. This review reveals that the treatment of biomass
residues and wastewater is frequently neglected, even though the existing legal framework in
Costa Rica recognized the need for wastewater treatment. Consequently, improper management
of biomass residues and wastewater causes a severe sanitation and water pollution problem in
Costa Rica. To overcome the pollution problem, a discussion of technologies applicable to treat
biomass residues and wastewater from agro-industrial and agricultural activities is presented
here. Finally, exergy is introduced as a metric to evaluate the sustainability of new systems.

2.1. The opportunity: from linear to circular thought
The Food and Agriculture Organization of the United Nations (FAO) has defined a new
approach for supporting the food security and sustainable agriculture: The Water-Energy-Food
Nexus (FAO 2014). These three resources are essential for the human well-being and a positive
resource balance between the supply from the environment and demand by the human is
essential for sustainability. However, the balance tends to be negative due to the population
growth and climate change, which in turn restrict even more the access to water, energy, and
food. First, there is an unsustainable demand of water, energy, and food by the population based
on a linear economic model (use, make, dispose). Consequently, the disposal of untreated
residues degrades the environment and creates a severe pollution problem. For example, it is
estimated that over 80% of wastewater worldwide goes untreated, so that more than two million
cubic meters of wastewater are discharged into the world’s rivers, lakes, and oceans every day

18

(Otchet 2003, UNESCO 2015). On the other hand, the effects of climate change, such as
droughts, floods, and increasing temperatures, threaten the availability of resources (water,
energy, and food) and restrict access to them for vulnerable populations. FAO (2014) estimates a
world population of 10 billion people for 2050, which will result in a 60% increase in demand
for food. Thus, the agro-industry and agricultural activities will require more water and energy,
which currently account for, on average, consumption of 70% of fresh water and 30% of energy
resources worldwide. Under this paradigm, there is an urgent need to reduce pressure on water,
energy, and food resources. Sustainable development requires recognizing the linkages between
these three resources (Biggs, Bruce et al. 2015). Once the linkages are recognized, integrated
treatment systems can be implemented to optimize resource supply and consumption by the
creation of close-loop systems to convert linear economies into circular economies (Agrocycle
2017). In contrast to a linear economy, a circular economy considers residues as resources,
alleviating the demand of resources and reducing waste generation and disposal (Figure 2.1).

19

Figure 2. 1. Dynamics of the water-energy-food nexus in the economy (Own creation).

2.2. Biomass residues and wastewater from the agro-industry and agricultural
activities in Costa Rica
In general, agro-industry and agricultural activities follow a linear economy model, with
the simple disposal of biomass residues and wastewater adversely impacting the environment. In
this section, the biomass residues and wastewater from agro-industry and agricultural activities
in Costa Rica are characterized. Definition, characteristics, and estimated waste production are
indicated for biomass residues and wastewater. Moisture content and gross calorific value are
indicated for the solid wastes; whereas, water quality parameters are indicated for liquid wastes.
In addition, regulations, efforts, and limitations are established for identifying the problem in
Costa Rica.

20

2.2.1. Biomass residues from agro-industry and agricultural activities
In Costa Rica, biomass residues from agro-industry and agricultural activities are named
RAOs (from the Spanish Residuos Agricolas Organicos) (Coto 2013). Biomass residues from
agricultural crops (53% of total agricultural residues) are predominantly from sugar cane,
pineapple, oil palm, coffee, and banana; while, biomass residues from livestock activities (47%
of agricultural residues) are largely consist of cow manure, chicken litter, and swine manure. In
addition, sawdust, wood chips, and other rejected wood products are sawmill residues, but only
account for a small percentage of residues. Table 2.1 characterizes biomass residues based on the
gross calorific value and moisture content. The gross calorific value indicates the potential
energy (as heat) that can be obtained per kilogram of biomass residues when combusted. For the
selected biomass residues, the gross calorific values are similar; however, moisture contents of
the residues are substantially different. Thus, selection of technologies for treating and extracting
energy from biomass residues should be based on moisture content. Technologies to convert
biomass residues into energy include direct combustion, gasification, and anaerobic digestion
(Zafar 2016). Direct combustion and gasification are appropriate technologies for low-moisture
content feedstocks (IRENA 2012, Funk, Milford et al. 2013). For example, commonly used
biomass residues for direct combustion are sawdust, wood bark, shavings, end cuts, and chips,
hog fuel, bagasse, and rice husks – all materials with moisture contents of 10 – 50% on a wetbasis. Predominant biomass residues for gasification are wood chips, hog fuel, rice hulls, dried
sewage sludge, pellets, wood scrapes, and nut shells, which have moisture contents ranging 15 –
50% (IRENA 2012, Funk, Milford et al. 2013). Alternatively, anaerobic digestion is an
appropriate technology for processing biomass residues with moisture contents of 65 to 99%
(IRENA 2012, Funk, Milford et al. 2013). In this study, animal manure and food waste are the

21

biomass residues of interest. Coto (2013) estimates a production of 2,652,143 dry metric tons per
year of animal manure. Thus, in Costa Rica, a total net energy of 349 MW electricity per year
could be generated from animal manure, demonstrating the great potential Costa Rica has for
electric energy generation from biomass residues. In comparison, the status quo capacity for
electrical generation in Costa Rica is 2,600 MW.

Table 2. 1. Characteristics of selected biomass residues in Costa Rica.
Agricultural
activities

MC*
(%)

Gross
calorific
value
(MJ/kg)

Livestock
activities
(manure)

MC*
(%)

Gross
calorific
value
(MJ/kg)

Coffee pulp
Mucilage

81.0
81.0

15.9
15.9

Pig
Chicken

85.0
36.0

13.8
15.9

Coffee husk

11.0

17.9

Milk

80.0

15.6

Rice husk
Sugar cane
bagasse
Field
residues
from
pineapple
Pineapple
crown

15.0

15.4

Meat

50%

17.5

--

80.0
--

--

--

90%

11.6

78.5

11.6

MC*
(%)

Gross
calorific
value
(MJ/kg)

32.0
50.0

18.5
18.5

55.0

18.5

15.6
--

Sawdust
Woodchips
Rejected
wood (e.g.,
bark)
---

---

---

--

--

--

--

--

--

--

--

--

--

Sawmill

*. MC: moisture content. Modified from Coto (2013).

With regards to regulatory agencies in Costa Rica, the Ministry of Health (MINSALUD)
regulates public health while the Ministry of the Environment and Energy (MINAET) regulates
environmental protection. These two institutions have developed the legal framework for
management of biomass residues. In total, five laws and regulations have been decreed
concerning legal aspects to the management of biomass residues. In general, three laws (General
Health law, Organic Environmental law, and Law of Integrated Waste Management) ensure
public health by the appropriate management of wastes; however, there is not a specific law or
22

regulation for biomass residues from agro-industrial and agricultural activities. Instead,
household residues (garbage and rubbish such as bottles, cans, clothing, compost, disposables,
food packaging, food scraps, newspapers and magazines, and yard trimmings) and hazardous
wastes are specific regulated (Table 2.2).

Table 2. 2. Laws and regulations for the management of solid residues in Costa Rica.
Law

Main goal

Legal aspects with respect to biomass residues

General Health law
(N° 5395, 1973)

To ensure the public health of
the population.

Organic
Environmental law
(N° 7554, 1995)

To ensure proper management of
natural resources.
To ensure environmental
protection.
To regulate management and
disposal of household residues.
To regulate efficient use of
resources and integrated
management of wastes through
the planning and execution of
regulatory, operational,
financial, administrative actions,
and educational, environmental,
and health monitoring and
evaluation.
Regulate management and
disposal of industrial hazard
residues.

Forbid pollution of ground and surface water due to direct
or indirect discharge of solid, liquid and gaseous wastes.
Forbid discharge of industrial residues into the sewage
system or storm sewer system.
Industries should provide treatment of residues to avoid
environmental pollution.
Promote plans for protection of the environment. Includes
appropriate management of solid residues.
Industries are responsible to provide treatment of residues
to avoid environmental pollution.
Note: Does not include biomass residues and wastewater
from agro-industrial and agricultural industries.
Industries should establish programs for efficient use of
resources and management of solid wastes prior to
disposal.
Efficient use of resources includes reduce, recycle, and
reuse strategies for reducing quantity of generated waste.
Management of wastes includes treatment of residues to
convert them into resources with added value.

Landfill regulation
(N° 27387-S, 1995)
Law of Integrated
Waste Management
(N° 8839, 2010)

General Regulation
for the Classification
and Management of
Hazardous Waste (N°
37788-S, 2013)

Note: Does not include biomass residues and wastewater
from agro-industrial and agricultural industries.

Instead of separate laws, biomass residues from agro-industrial and agricultural activities
are regulated through this legal framework. Major efforts have been focused in the production of
energy through the treatment of biomass residues. The Non-conventional renewable energy
program, led by the Costa Rican Electricity Institute (ICE), has been promoting anaerobic
digestion for the treatment of animal manure with the main goal of producing energy. Currently,

23

renewable energy production from biomass represents 0.73% of the total power capacity in Costa
Rica (18.9 of 2,600 MW) (ICE 2016). Based on estimates by Coto (2013), current energy
production from biomass is only 5.44% of the potential energy that can be produced from animal
manure in Costa Rica. Additionally, ICE also promotes anaerobic digestion projects as part of
integrated watershed management for protecting reservoirs from hydroelectric energy
production, which is the main renewable energy source in Costa Rica (75%, 1,950 of 2,600 MW)
(ICE-UEN 2010). Prestigious academic institutions in Costa Rica (e.g., EARTH University,
National University (UNA), University of Costa Rica (UCR), and Technological Institute of
Costa Rica (TEC)) have played an important role in the treatment of biomass residues. These
institutions have been developing research and extension projects promoting anaerobic digestion
(Kinyua 2015). Unfortunately, there are no data bases of anaerobic digestion projects
demonstrating the technical and treatment performance. A survey of biogas production
elaborated by the ICE in 2014 simple indicates that 58% of the digesters are tubular, plug-flow
bags (e.g., tubular polyethylene bag digesters) for small-scale farms; whereas covered lagoon
anaerobic digesters correspond to 42%. The limited information that is available is from projects
developed by ICE at many agro-industries and agricultural farms. Cow manure is the most
utilized feedstock for biogas production, with a variable production (20 – 40 m3/d), whereas,
swine manure is an attractive alternative feedstock due to high energy production. In general, all
these projects target to mesophilic conditions (Table 2.3).

24

Table 2. 3. Summary of anaerobic digester projects developed by the Non-conventional
renewable energy program (ICE).
Project

Type of
reactor

Volume
(m3)

Heads

Feedstock:
Manure
source

Biogas
produced
(m3/d)

CH4
(%)

Power
capacity
(kW)

Robago Farm 1

Covered lagoon

90.0

85

Cow

20

n.a.

30

Julieta Farm 1

Covered lagoon

106.2

50

Cow

30

n.a.

38

Cerro Grande Farm 1

Covered lagoon

44.0

50

Cow

21

66.0

20

Noble Farm 1

Covered lagoon

86.4

120

Cow

40

n.a.

n.a.

Sermide Farm 1

Covered lagoon

988

4,000

Swine

320

n.a.

60

El Cerro Farm 1

Covered lagoon

2,093

5,530

Swine

412

80.0

70

Agro-Turistica Don
Carlos Farm 1

Covered lagoon

180

n.a.

Cow (with
cheese
whey)

29

62.0

n.a.

Porcina America 1

Covered lagoon

3,600

30,000

Swine

1,800

60.0

250

Kafur Farm 1

Covered lagoon

n.a.

4,000

Swine

n.a.

n.a.

70

EARTH University

Tubular
polyethylene
bag digesters

85

5

Cow

27.5

62.6

40

Tubular
polyethylene
bag digesters

85

40

Swine

6.0

76.4

40

Tubular
polyethylene
bag digesters

12

10

Swine

2.83

71.0

n.a.

(Lansing, VĂ­quez et
al. 2008)
EARTH University
(Lansing, VĂ­quez et
al. 2008)
Monteverde,
Puntarenas
(Kinyua 2015)
1

Projects developed by the Non-conventional renewable energy program (ICE).
n.a. represents not assessed

Implementation of anaerobic digestion provides economic benefits to farms while
satisfying legal compliance with Costa Rican regulations. These farms are saving money by the
consumption of self-produced electric energy for operation of the farms or, for smaller projects,
for cooking (Coto 2013, Kinyua 2015, ICE 2016). In addition, implementation of anaerobic
digestion increases the electric energy generation from biomass in Costa Rica. Costa Rica is

25

striving to be the first carbon-neutral country by 2021 and utilization of biomass residues as a
renewable energy has been prioritized to reduce electricity production from fossil fuels.
Through conversion of organic matter into biogas, anaerobic digestion reduces organic
matter and suspended solids within the digester. However, as the influent feedstock has very high
concentrations of organic matter and suspended solids (e.g., >1 g/L concentrations), effluent
digestate still has high solid, carbon, and nutrient contents (Lansing, VĂ­quez et al. 2008, Kinyua
2015). For example, Lansing, VĂ­quez et al. (2008) reported 796 and 189 mg COD/L, 1,440 and
717 mg TS/L, 178 and 177 mg TKN/L, and 16.6 and 19.8 mg TP/L, in the digester effluent from
a dairy and swine farm, respectively. These values are much greater than the discharge standards
in Costa Rica, discussed in the next section. Usually, digestate is applied to agricultural fields as
a fertilizer for enhancing nutrient retention in soils. However, improper management of nutrients
through land application can cause eutrophication (Sutton and Joern 1992, Johnson and Eckert
1995). In addition, mesophilic anaerobic digestion does not eliminate pathogens to
concentrations below the detection limits; thus, there is potential health impact due to the
transmission of pathogens to food and water when digestate from mesophilic reactors is land
applied. For example, Kinyua (2015) detected Giardia lamblia and Cryptosporidium parvum in
the effluent from a swine tubular polyethylene bag digester in Monteverde, Puntarenas.

2.2.2. Wastewater from agro-industry and agricultural activities
The Costa Rican Regulation of Dumping and Reuse of Wastewater law (N° 33601-S,
2007) defines wastewater as water that has been used and, due to the use, it characterized by the
presence of pollutants. In addition, the regulation classifies wastewater as ordinary or special.
Ordinary wastewater is wastewater from households, including grey and black wastewater;

26

whereas, special wastewater results from non-household sources. Thus, wastewater from agroindustry and agricultural activities is classified as special wastewater. In 2013, the MINSALUD
had registered 5,028 industries producing special wastewater. However, only 30% of the
industries presented an operational report, which is requested by the Regulation of Dumping and
Reuse of Wastewater law (N° 33601-S, 2007). And, even worse, a small percentage of these
reports are reliable. Thus, there is not an accurate estimation of the special wastewater produced
in Costa Rica.
The characteristics of special wastewater widely differ based on source. For example,
wastewater from crop production, food processing, slaughterhouses, livestock production, and
anaerobic digesters have entirely different characteristics. Wastewater from crops is
characterized by pesticides and chemical products applied to crops (Espigares and Perez 1985).
Food processing wastewater has high concentrations of COD, TS, and oil; however,
characteristics of food processing wastewater fluctuate substantially depending on what food is
processed. For example, COD concentrations of 660, 1,500, 7,400, 800, and 18,000 mg/L are
typical for beverage production, breweries, oil and fat production, milk and dairy production, and
wheat starch production, respectively (Mori 2003). Slaughterhouse wastewater consists of fats,
proteins, fibers, organic carbon, pathogens, and veterinary pharmaceuticals. Typical
concentrations are 3,000 mg COD/L, 3,000 mg TS/L, 450 mg TN/L, and 50 mg TP/L (BustilloLecompte and Mehrvar 2017). Livestock wastewater is collected water used for removal of
manure from confined spaces where animals are kept, such as milking parlor wastewater. This
wastewater is characterized by high organic content and solids, pathogens, and high nutrient
contents. For example, Lansing, VĂ­quez et al. (2008) reported 5,720 and 2,330 mg COD/L; 4,570
and 2,180 mg TS/L; 218 and 231 mg TKN/L; and 30.0 and 30.7 mg TP/L, in the wastewater

27

from dairy and swine production, respectively. In the case of anaerobic digestion, livestock
wastewater is treated as a semi-solid feedstock. The digestate from anaerobic digesters is then
considered to be a special wastewater.
Wastewater treatment aims to reduce the level of pollutants in wastewater prior to reuse
or disposal, thereby alleviating the negative impacts of waste production. In large cities,
centralized wastewater treatment facilities are common (UN 2015). A centralized wastewater
treatment facility is a large-scale wastewater treatment plant that receives wastewater from
domestic, commercial, and industrial activities in urban areas (Anderson and Sheffield 2015).
This large-scale system manages physical, chemical, and biological processes through a complex
combination of mechanical devices that requires technically skilled manpower for operation and
maintenance; thus, it is a capital-intensive technology (Avila, GarcĂ­a et al. 2016). Instead, a
decentralized wastewater treatment facility is an on-site system that treats wastewater from
individual users or small communities (UN 2015). Decentralized wastewater treatment facilities
tend to be lower in cost and less sophisticated with regards to operation and maintenance than
centralized wastewater treatment facilities.
MINSALUD and MINAET have developed the legal framework in terms of the
management of wastewater residues. The General Health law (N° 5395, 1973) and Organic
Environmental law (N° 7554, 1995) establish legal aspects and regulations with respect to
wastewater residues. The General Water Law (No. 276, 1942) establishes water as a public
resource for all Costa Ricans and penalizes anyone who contaminates the water. However, this
water law is outdated, and fines correspond to 1942. For example, a person would have to pay
180 to 720 colones (~$0.32 to $1.26 in US dollars) if found guilty of contaminating a body of
water. In 2017, a new law, the Integrated Management of Water Resources, was discussed in

28

Congress and, if approved, will replace the outdated water law from 1942. Additionally, in 1961,
the Costa Rican Water and Sewage Institute (AyA) was created as the entity to ensure water for
human consumption and collection and treatment of wastewater. The Regulation of Dumping
and Reuse of Wastewater law (N° 33601-S, 2007) replaced the previous 1997 version (N°
26042-S) and establishes discharge standards for treated wastewater. This wastewater law
obligates all industries that generate special wastewater to implement wastewater treatment prior
to discharging wastewater into any water body. An operational report of the treatment plant in
each industry is requested and if the effluent does not comply standard limits, the MINSALUD
can suspend the economic activity and apply fines based on the Regulation Environmental Tax
for Dumping (N° 31176, 2003) and the General Water law (Table 2.4).

29

Table 2. 4. Laws and regulations for the management of liquid residues in Costa Rica.
Law

Main goal

Legal aspects with respect to wastewater

General Water Law
(No. 276, 1942)
Constitutive Law of
the Costa Rican Water
and Sewage Institute
(N° 2726, 1961)

To protect and manage water as
a public resource in Costa Rica.
To establish policies and norms,
support and develop programs
and plans, and treat water for
human consumption and
wastewater treatment to assure
public health and avoid
environmental degradation.
To ensure the public health of
the population.

Penalize anyone who discharges contaminants into
ground and surface water in Costa Rica.
This law created the Costa Rican National Institute of
Aqueducts and Sewers (AyA) as the national entity for
assuring water for human consumption and wastewater
treatment.

General Health law
(N° 5395, 1973)

Organic
Environmental law
(N° 7554, 1995)

To ensure proper management of
natural resources.
To ensure environmental
protection.

Regulation
Environmental Tax
for Dumping (N°
31176, 2003)

To regulate and tax the discharge
or spills of residues into water
bodies.

Regulation of
Dumping and Reuse
of Wastewater law
(N° 33601-S, 2007)

To protect public health and the
environment by management and
treatment of wastewater.

Forbid pollution of ground and surface water due to
direct or indirect discharge of solid, liquid and gaseous
wastes.
Forbid the discharge of industrial residues into the
sewage system or storm sewer system.
Industries should provide treatment of residues to
avoid environmental pollution.
Promote plans for protection of the environment. This
section includes the implementation of wastewater
treatment systems.
Industries must provide treatment of residues to avoid
environmental pollution.
Industries must pay a tax based on the quantity of
kilograms of COD or TSS discharged into water
bodies.
The tax must be paid if the concentration of the
contaminant (e.g., kg COD/L or kg TSS/L) is greater
than the maximum limits for discharge of treated
water.
Industries must treat the wastewater.
Effluent from wastewater treatment need to comply a
discharge standard (e.g., COD < 150 mg/L, TS < 50
mg/L, TN < 50 mg/L, TP < 8 mg/L, 5 < pH < 9, 15 °C
< T < 40 °C, and others).
Industries can be suspended and taxed if effluents do
not comply with discharge standards (Regulation
Environmental Tax for Dumping (N° 31176, 2003)).

In Costa Rica, the legal framework for wastewater seems powerful. Unfortunately,
regulations and laws are not enforced, and Costa Rica suffers severe water pollution problems.
The clearest example is the AyA, the governmental entity in charge of collection and treatment
of ordinary wastewater. For years, multiple sewage systems have collected ordinary wastewater
in the central valley; however, the collected wastewater has been simply discharged into the
Torres, Rivera, and Maria Aguilar rivers (in the province of San Jose), the Pirro, Burio, and
30

Bermudez rivers in the province of Heredia, the Zopilote river in Cartago, and El Barro river in
Alajuela (Araya, Araya et al. 2003, Angulo 2013). Recently, in 2015, Los Tajos wastewater
treatment plant began receiving ordinary wastewater from the approximately 1,700,000
inhabitants of the province of San Jose; however, this wastewater treatment plant only has
primary treatment and the effluent complies with discharge standards only rarely (Otarola 2015,
Nogarin 2017). Secondary and tertiary treatment units are being planned; however, currently, the
effluent from Los Tajos is simply discharged into Torres river (Ruiz 2014, Nogarin 2017).
Treatment of agro-industrial and agricultural wastewater is also extremely limited in
Costa Rica. As mentioned previously, only 5,028 industries are registered and only 30% of them
presents the operational wastewater treatment report (Angulo 2013). Until responsible execution
of the legal framework in Costa Rica occurs, the best approach for the appropriate management
of wastewaters is environmental education (Angulo 2013, Calvo 2014). Many efforts have been
performed to create decentralized wastewater treatment system within industries, particularly for
those in rural areas, where access to sewage systems is unpractical. Alternative treatment
systems, also called non-conventional treatment systems, have been proposed by academia and
non-governmental organizations. These systems tend to be lower in cost and less sophisticated in
operation and maintenance than septic tanks, Imhoff tanks, up-flow anaerobic sludge bed
reactors, and activated sludge treatment systems. Biogardens or constructed treatment wetlands
(CTWs) have been implemented as alternative treatment systems in Costa Rica (Angulo 2013).
The Central American Association for the Economy, Health and the Environment (ACEPESA)
has built more than 100 biogardens to treat grey water at houses and hotels in rural communities.
In general, this approach seems to be promising for ordinary wastewater in rural areas, where
area constraints are not a problem. However, these two approaches have also be applied in urban

31

areas (Moncada 2011, Alfaro, Perez et al. 2013) (Table 2.5). Biogardens and treatment wetlands
differ by water content, where biogardens are typically unsaturated and treatment wetlands are
continuously saturated.

Table 2. 5. Summary of biogarden and constructed treatment wetland projects developed by
academic institutions.
Project

Barra
Honda,
Guanacaste
Barra
Honda,
Guanacaste
Barra
Honda,
Guanacaste
La Virgen,
Guanacaste
La Virgen,
Guanacaste
Zapote, San
Jose
n.r.

Popular
Cultural
Museum,
Heredia
Monteverde,
Puntarenas

Institution
(Reference)

UNA / ACEPESA
(Cubillo and
Gomez 2017)
UNA / ACEPESA
(Cubillo and
Gomez 2017)
UNA / ACEPESA
(Cubillo and
Gomez 2017)
UNA / ACEPESA
(Cubillo and
Gomez 2017)
UNA / ACEPESA
(Cubillo and
Gomez 2017)
TEC / ACEPESA
(Moncada 2011)
UNA (PĂŠrez,
Alfaro et al. 2013)

Wastewater
type

Treatment
system

Dimensions
(Length x
width x
height, m)

Water quality parameter
(In / Out, mg/L)
COD

TS

TN

TP

Grey water

Biogarden

12 x 4 x n.r.

161 /
132

n.r.

0.16 /
0.029 1

0.137 /
0.246

Grey water

Biogarden

12 x 4 x n.r.

54.0 /
13.0

n.r.

0.089 /
0.116 1

0.088 /
0.05

Grey water

Biogarden

12 x 4 x n.r.

140 /
75.0

n.r.

0.024 /
0.03 1

0.07 /
0.066

Grey water

Biogarden

12 x 4 x n.r.

145 /
22.0

n.r.

0.022 /
0.107 1

0.043 /
0.22

Grey water

Biogarden

12 x 4 x n.r.

54.0 /
26.0

n.r.

0.03 /
0.082 1

0.05 /
0.046

Grey water

Biogarden

5 x 1 x 0.7

HSSFCTW

12 x 3 x 0.6

60 /
6.2
400 /
108

4.5 / 6.7

Black water
from a
septic tank

444 /
62
250 /
50

n.r.

3.5 /
2.5
8/3

UNA (Alfaro,
Perez et al. 2013)

Grey water

HSSFCTW

8 x 2.5 x
0.7

98% 2

99%

n.r.

n.r.

(Dallas, Scheffe et
al. 2004)

Grey water

HSSFCTW

12 x 1.2 x
0.5

167 /
10 3

8.4 /
1.1 4

1.6 /
3.6

2

15 /
6

n.r.: not reported. HSSF-CTW: horizontal subsurface flow constructed treatment wetland
1
reported as ammonium.
2
only treatment performance was reported.
3
reported as BOD.
4
reported as ammonium.

32

The major limitation in terms of wastewater treatment in Costa Rica is the lack of
compliance and enforcement with the current legal framework. Efforts conducted by academic
institutions increase the public’s awareness of water pollution problems in Costa Rica.
Environmental education, until now, has been the best tool for improving sanitation and
decreasing water pollution in Costa Rica. However, more research is needed on design of
biogardens and CTWs. The projects reported in Table 2.5 focus on treatment performance under
specific conditions; general design parameters were not established. Thus, design of biogardens
and CTWs in Costa Rica continues to be empirical. Instead, to promote development of CTW,
more engineering analyses are needed.

2.2.3. Biomass residues and wastewater: the problem becoming an opportunity
In part due to improper management of biomass residues and wastewater, Costa Rica has
a severe water pollution problem. A simple indicator of this problem is poor water conditions in
most surface waters in Costa Rica. For example, the Regional Institute for Studies on Toxic
Substances (IRET) sampled 487 points in 250 rivers across Costa Rica, and found that 71% of
the sampling points have a moderate to extreme pollution condition (IRET 2012). Laws and
regulations exist; however, the legal framework is not enforced or properly followed by the
regulators. Therefore, proper management of wastewater depends on social will and awareness
of agro-industry and agricultural activities. However, this contamination problem can become an
opportunity for much needed change. Environmental education and academic efforts can turn the
current linear economic models into circular, close-loop, economic models. Integrating
decentralized self-sufficient, close-loop, organic waste treatment systems into current agroindustry and agricultural activities is expected to promote a circular economy that considers

33

residues as resources, alleviating the demand of resources and reducing waste generation and
disposal (Figure 2.1). Additionally, Costa Rica has two important related goals, 1) to be the first
carbon neutral country by 2021 and 2) to develop a green economy. Thus, integration of
treatment systems into current agro-industry and agricultural activities can solve the water
pollution problem, while helping Costa Rica achieve carbon neutrality and green development
goals (Table 2.6).

Table 2. 6. Biomass residues and wastewater: the problem and subsequent opportunities.
Costa
Rican goals

Action taken
by Costa
Rica

The problem

The opportunity

Carbon
neutrality by
2021

Generation of
energy using
renewable
resources.

Green
development

Optimization
of resource
utilization.

Current energy production
from biomass is only
5.44% of the potential
energy.
Current linear economic
model does consider
residues.
Biomass residues and
wastewater are not
properly treated.

Integrate
decentralized selfsufficient, closeloop, organic waste
treatment system
to consider
residues as
resources.

Benefits

Energy generation from
biomass resources.
Reduction of fossil fuel use.
Reduction of resource
consumption and waste
disposal.
Recovery of nutrients.
Water reclamation.
Reduction of sanitation
problems.

2.3. Integration of technologies at the tropics: Costa Rica case
A discussion of technologies for treatment of biomass residues and wastewater is
presented here. Integration of technologies can overcome individual disadvantages of each
technology. The main goal of the SPAD-HCTW is implementation of a decentralized selfsufficient, close-loop, organic waste treatment system. The integrated system should provide
energy, fertilizers, and reclaimed water, to contribute to the protection of water resources in
Costa Rica. In addition, this approach can reduce the demand of water, energy, and food
resources (Figure 2.1). This integration can contribute to sustainable development in rural areas
in Costa Rica, in particular with regards to agro-industry and agricultural activities.

34

2.3.1. Waste to energy: optimizing anaerobic digestion through solar thermal
collection
Anaerobic digestion is a solid waste management technology for processing high
moisture content biomass residues into energy (IRENA 2012, Funk, Milford et al. 2013).
Anaerobic digestion starts with hydrolysis of lipids, complex polymers (e.g. cellulose,
polysaccharides), and particulate organic materials into monomers (e.g., sugar and amino acids)
and long chain fatty acids. Then, in acidogenesis, monomers are fermented or anaerobically
oxidized into short chain fatty acids, alcohol, and ammonia. In acetogenesis, alcohol and short
chain fatty acids are converted to acetic acid, hydrogen (H2) and/or carbon dioxide (CO2).
Finally, methanogens, in the absence of oxygen, produce biogas from acetic acid, CO2, and H2
(Gould and Crook 2010). Solid and liquid digestate are also products of anaerobic digestion.
Solid digestate can be used as a fertilizer to enhance nutrient retention in soils (Liedl,
Bombardiere et al. 2006). Then, to prevent environmental impairment and to allow reuse of the
water, liquid digestate can be treated. The treatment of the liquid digestate in discussed further.
Biogas is primarily a mixture of methane (CH4) and CO2. Small amounts of hydrogen
sulfide (H2S), water vapor (H2O), nitrogen (N2), H2, and oxygen (O2) are minor components in
biogas. Methane is a relatively dense and reliable biochemical energy source (Tsalkatidou,
Gratziou et al. 2009). The simplest utilization of CH4 as energy is use of biogas as a cooking fuel
(Kinyua 2015). Similarly, biogas can be burnt for heating fluids (e.g., water, air, and even,
digestate) or buildings. Finally, CH4 can be used to fuel electrical generators, such as combined
heat and power (CHP) systems. Importantly, before using CH4, it is recommended to remove any
H2S from the biogas to prevent formation of sulfuric acid, a corrosive liquid. Iron steel wool can

35

be placed inside biogas effluent pipes for simple use of the biogas, such as a cooking fuel;
activated carbon filters can also be used to remove H2S from biogas prior to CHP (Kinyua 2015,
Noramelya, Shahbudin et al. 2016).
In Costa Rica, the most common anaerobic digester reactors are covered lagoon and
tubular polyethylene bag digesters (Table 2.3). Covered lagoon and tubular polyethylene bag
digesters are attractive to inhabitants in rural areas due to simplicity of installation and operation,
as well as low investment costs and energy savings (Lansing, VĂ­quez et al. 2008). However, due
to the simplicity of these systems, anaerobic digestion performance can be affected by lack of
control over operating temperatures, instability of operating temperatures, and limited mixing.
Due to warmer ambient temperatures in the tropics (15 – 30°C), covered lagoons and tubular
polyethylene bag digesters are not typically heated. Thus, these digesters work at operating
temperatures close to the low mesophilic range. For example, at Earth University, Costa Rica,
Lansing, Víquez et al. (2008) reported temperatures of 26.7 and 25.8°C from a 85 m3 tubular
polyethylene bag digesters receiving cow and swine manure, respectively. Consequently,
microbial growth and organic matter degradation is lower than that reported for higher
thermophilic operating temperatures. For example, Vindis, Mursec et al. (2009) used 0.5 L mini
digesters fed with maize to demonstrate that higher biogas production (494 – 611 NL kg/VS)
was achieved at 55°C, as compared to 315 – 409 NL kg/VS produced at 35°C. In addition, as the
operating temperature depends on the ambient temperature, changes in ambient temperature can
produce instability in the operating temperature. For example, Visser, Gao et al. (1993) indicated
that instability of temperature decreased microbial growth and CH4 production. In fact, EspinosaSolares, Valle-Guadarrama et al. (2009) observed that methane production decreased 11.1% per
°C for small changes in temperature from 52 to 56°C. Moreover, contents in covered lagoons and

36

tubular polyethylene bag digesters are not mixed. Without mixing, localized pockets of
temperature can create heterogeneous operating temperatures due to stratification of solids and
formation of dead zones (Espinosa-Solares, Valle-Guadarrama et al. 2009, Suryawanshi,
Chaudhari et al. 2010). In contrast, through mixing, Bombardieri, Espinosa-Solares et al. (2007)
maintained a 40 m3 continuous stirred tank reactor (CSTR) anaerobic digester temperature at
stable ¹0.1°C from the target temperature, allowing steady-state conditions for biogas
production. Mixing also allows homogeneity of pH, better interaction between microbes and
substrate, and release of generated biogas (Suryawanshi, Chaudhari et al. 2010). In addition,
temperatures close to the psychrophilic (<25°C) and mesophilic (25-40°C) range do not assure
pathogenic destruction. For example, Kinyua (2015) detected pathogens (e.g., Giardia lamblia
and Cryptosporidium parvum) in the effluent of a swine tubular polyethylene bag digester
working at 20.7°C. Instead, at thermophilic (<45°C) conditions, there is thermal destruction of
pathogens (Vindis, Mursec et al. 2009).
Therefore, to improve utilization of energy in biomass residues and eliminate pathogens
in the digestate, a thermophilic CSTR anaerobic digester is recommended. Thermophilic
anaerobic digestion speeds up microbial growth and organic matter degradation, consequently
producing more biogas and discharging pathogen free digestate. A CSTR is characterized by
continuous feeding of feedstock and discharge of digestate, and continuous mixing inside the
digester. Therefore, the CSTR reactor type assures homogeneity of parameters (e.g., temperature,
chemical concentration, pH, alkalinity, and substrate concentration) inside the digester,
promoting steady-state conditions for biogas production (Bharati and Kalamdhad 2016). In
addition, numerous studies have demonstrated that thermophilic cultivation can enhance
anaerobic digestion performance by shortening retention time, and thus requiring smaller vessel

37

sizes, improving odor control, and reducing total solids in waste streams (Aitken, Sobsey et al.
2005, Suryawanshi, Chaudhari et al. 2010, Sharma, Espinosa-Solares et al. 2013, Zarkadas,
Sofikiti et al. 2015).
However, as a disadvantage, a thermophilic CSTR anaerobic digester requires higher
input energy compared to covered lagoon and tubular polyethylene bag digesters working at
ambient temperature. Energy is required for mixing and maintaining thermophilic temperature.
For conducting mixing, energy input is required to operate a blower for recirculating the biogas
or a pump for recirculating the digestate. For example, in a 40 m3 CSTR anaerobic digester,
Espinosa-Solares, Valle-Guadarrama et al. (2009) reported that a blower and a digester
recirculation pump consumed per day 3.18% and 1.28% of the total energy needed to operate the
digester (428 MJ). The major input energy is heat for maintaining thermophilic conditions;
95.5% of the input energy was required to heat the CSTR anaerobic digester (Espinosa-Solares,
Valle-Guadarrama et al. 2009). Thus, heating represents the major disadvantage to thermophilic
digesters.
One strategy for heating thermophilic CSTR anaerobic digesters is using the biogas
generated. However, heating using the biogas may lead to an unfavorable energy balance for
small-scale digesters (Vindis, Mursec et al. 2009). In contrast, higher biogas production in largescale reactors covers heating demand to maintain thermophilic temperature and provides surplus
energy for electric generation. For example, Zabranska, Dohanyos et al. (2002) demonstrated
that a 4,800 m3 thermophilic (55°C) CSTR anaerobic digester treating raw sludge used 167
MWh/d for heating the system and produced 108 MWh/d as a surplus energy. Thus, the biogas
produced was mainly used for heating purposes (Zabranska, Dohanyos et al. 2002).

38

To overcome the higher energy requirement of thermophilic anaerobic digestion, solar
heating systems have been implemented. Low-density and inconsistent solar energy can be
converted into a relatively dense and reliable biochemical energy source – methane. In Jordan,
Alkhamis, El-khazali et al. (2000) installed a 1.54 m2 flat-plate solar thermal collector to heat a
laboratory scale reactor (volume = 0.053 m3). Hot water from the collectors was stored in a
storage tank (volume = 0.096 m3). A heat exchanger passing through the storage tank and the
reactor recirculated water to maintain a constant temperature of 40°C in the digester (Figure
2.2.a). In Greece, Axaopoulos, Panagakis et al. (2001) and Axaopoulos and Panagakis (2003)
installed a 21 m2 flat-plate solar thermal collectors to heat an underground anaerobic digester
(total volume 116 m3, with a useful volume of 40 m3). Hot water was pumped through a heat
exchanger installed on the bottom of the anaerobic digester to maintain a digestion temperature
of 35°C. In 2001, the authors developed a mathematical model to describe the dynamic behavior
of the system, which showed a good agreement with the measured values. Then, in 2003, results
showed sufficient biogas production from the digester to heat a swine nursery during winter
(Figure 2.2.b). In Egypt, El-Mashad, van Loon et al. (2004) targeted a thermophilic (50°C)
process in a 10 m3 completely stirred tank reactor (CSTR) anaerobic digester. The authors
evaluated two different configurations, one with the solar collector separate from the digesters
(Figure 2.2.c) and the other with the solar collector integrated into the digester (Figure 2.2.d).
The area covered by the solar collector was 4.88 m2. Solar-heated water maintained the
thermophilic temperature most of a year except during cold months.

39

Figure 2. 2. A schematic of the system configuration of integrated solar thermal collectors and
anaerobic digesters. a) Alkhamis, El-khazali et al. (2000); b) Axaopoulos, Panagakis et al. (2001)
and Axaopoulos and Panagakis (2003); and c) and d) El-Mashad, van Loon et al. (2004).

These studies have shown that solar thermal collectors can supply the energy required to
maintain a specific temperature during digestion. However, during winter in the temperate
regions (e.g., Egypt) a portion of biogas produced is needed to heat thermophilic anaerobic
digesters (El-Mashad, van Loon et al. 2004). Instead, in the tropics, it is expected that solar
thermal collectors can maintain thermophilic conditions due to warmer ambient temperature (15
– 30°C). However, even with constant, higher temperatures, effluent digestate from thermophilic
digesters still has high solid, carbon, and nutrient contents, and improper disposal can cause
water pollution.

40

2.3.2. Constructed treatment wetlands: ending the water pollution problem
Natural wetlands are ecosystems that have these attributes: 1) saturated soil or soil
covered by shallow water, 2) hydric soils, or 3) presence of macrophytes (Mitsch and Gosselink
2007). Therefore, a CTW is an engineered modification of the landscape to mimic a natural
wetland ecosystem. This man-made system is intended to create an ecosystem that stabilizes,
sequesters, accumulates, degrades, metabolizes, and mineralizes nutrients in the wastewater as
natural wetlands do (Halverson 2004, Vymazal 2007). The longevity of CTWs varies depending
on the system; however, estimations vary from 10 to 15 years (Wallace and Knight 2006). As
mentioned previously, CTWs are a decentralized wastewater treatment technology appropriate
for individual users or small communities in rural areas that tend to be lower in cost and less
sophisticated in operation and maintenance than centralized wastewater treatment facilities
(Table 2.5). In addition, CTWs have been implemented for treating industrial and agricultural
wastewater (Vymazal 2014) and high-strength wastewaters (Paing, Serdobbel et al. 2015).
Based on the hydrologic criteria, CTWs can be categorized as horizontal subsurface flow
(HSSF), vertical subsurface flow (VSSF), and free water surface (FWS) wetlands. Hybrid
constructed treatment wetlands (HCTW) consist of, in any sequence, a combination of HSSF,
VSSF, and FWS wetlands to achieve specific treatment goals.

2.3.2.1. Horizontal subsurface flow wetlands
HSSF wetlands consist of a filter media planted with emergent plants. Filter media,
usually sand or gravel, and roots provide surface area for microorganisms to grow. Roots exude
oxygen into the rhizosphere to support aerobic microorganism metabolism. Wastewater is kept
below the surface and flows, from the inlet to the outlet, horizontally through the filter media and

41

in and around the roots (Kadlec and Wallace 2009). The wastewater is kept at a constant level in
the filter media and continuous saturation promotes anoxic/anaerobic conditions due limited
oxygen transfer capacity. Thus, anoxic conditions prevail in HSSF wetlands and facilitate
chemical reduction processes (Vymazal 2007, Kadlec and Wallace 2009). Limited aerobic zones
occur around roots and rhizomes; thus aerobic chemical oxidation processes in HSSF wetlands
are limited. Because wastewater is not exposed to the atmosphere, HSSF wetlands do not provide
habitat suitable for mosquitoes, provide odor control, and minimize exposure to pathogenic
organisms (Vymazal 2007, Kadlec and Wallace 2009, Stefanakis and Tsihrintzis 2012) (Figure
2.3).
HSSF wetlands have demonstrated to be effective at removing organic matter and
suspended solids. Organic matter can be removed through biological degradation by aerobic and
anaerobic microbial processes and microbial uptake (e.g., organic matter serves as carbon
sources and energy for nitrification and denitrification). Sedimentation and filtration remove
suspended solids. For example, Summerfelt, Adler et al. (1999) applied 30 kg/m2/yr at 0.6 m3/d
of sludge from an aquaculture system into a 4.5-m2 HSSF wetland. The removal of organic
matter and suspended solids were 97% and 74% from an inlet concentration of 6,855 mg COD/L
and 7,860 TSS/L, respectively (Summerfelt, Adler et al. 1999). Clogging is potential concern
during operation of HSSF wetlands. The inlet area can clog with organic matter and suspended
sediment filtered and trapped in the media. For example, De Paoli and Sperling (2013), reported
clogging and surface runoff in a HSSF wetland planted with Typha latofolia as the inlet area
accumulated 35 g VS/L.
The main process for nitrogen removal in wetlands is nitrification followed by
denitrification. Nitrification is restricted due to limited aerobic zones in HSSF wetlands.

42

Nitrification is a two-step process driven by strictly aerobic nitrifying bacteria. First, Nitrosomas
oxidize the ammonium (NH4) to nitrite (NO2-), and then, Nitrobacter oxidize NO2- to nitrate
NO3-. Both sequential processes uses oxygen as an electron acceptor, carbon dioxide as a carbon
source, and either ammonium or nitrite as source of energy (Vymazal 2007) (equations 2.1).
Instead, anaerobic conditions prevail in the HSSF wetland and the NO3- is reduced to nitrogen
gas (N2) by denitrification. For denitrification, organic matter is used as a carbon source by
denitrifiers, while nitrate is used as electron acceptor (Vymazal 2007) (equation 2.2). For
example, for a 54-m2 HSSF wetland treating tilapia production wastewater, Zachritz, Hanson et
al. (2008) reported 76% removal of NO3-, whereas removal of NH4 was low (7.5%). Zachritz,
Hanson et al. (2008) indicated that the HSSF wetland was oxygen limited. Carbon can also be a
limiting factor for denitrification. For example, only 13% of NO3- was removed from a 10-m2
HSSF wetland receiving effluent from a VSSF wetland (Soroko 2007). Soroko (2007) indicated
that carbon was the limiting factor because the VSSF wetland removed from 2,448 mg BOD/L
from 2,452 mg BOD/L in the wastewater. Minor mechanisms that also remove nitrogen from
wastewater are ammonia volatilization, ammonia adsorption, and plant uptake if plants are
harvested (Vymazal 2007).
𝑁𝐻4+ + 2𝑂2 → 𝑁𝑂3− + 2𝐻 + + 𝐻2 𝑂

[Equation 2.1]

6(𝐶𝐻2 𝑂) + 4𝑁𝑂3− → 6𝐶𝑂2 + 2𝑁2 + 6𝐻2 𝑂

[Equation 2.2]

In HSSF wetlands, phosphorus (P) is removed by plant uptake, microbial uptake,
sorption, and precipitation. Phosphorus removal by plant uptake only occurs if vegetation is
harvested, otherwise, when plants decays, phosphorus is released back into the water. Microbial
uptake by bacteria, fungi, algae, and microinvertebrates occurs quickly but phosphorus is
released when the microorganisms decay. Sorption depends on the material of the filter media.

43

For example, sand has more sorption capacity than gravel or crushed rock. Minerals such as
reactive iron, aluminum hydroxide, or oxide groups on the surface area of filter media increase
adsorption capacity. In addition, calcareous materials can promote phosphorus precipitation.
However, filter sorption sites can become saturated, limiting phosphorus sorption. Thus,
phosphorus moves in a sedimentary cycle, and phosphorus removal from wetlands tends to be
low (Vymazal 2005, Kadlec and Wallace 2009, Vymazal 2014). For example, Avila, Salas et al.
(2013) reported that a 229-m2 HSSF receiving effluent from a VSSF wetland treating sewage
wastewater only removed 22% of influent phosphorus. In fact, further experiments with the same
system indicated no retention of P as the effluent concentration remained equal or more than the
inlet concentration to the HSSF wetland (Avila, GarcĂ­a et al. 2016).

Figure 2. 3. A schematic of a HSSF wetland. Figure was taken from Kadlec and Wallace (2009).

2.3.2.2. Vertical subsurface flow wetlands
VSSF wetlands have a similar configuration to HSSF wetlands, with the exception of the
direction of wastewater flow. In VSSF wetlands, wastewater can flow downward or upward and
either intermittently or continuously. In downward-flow wetlands, the wastewater moves from
the top of the surface media to the bottom; whereas, in upward-flow systems, wastewater fills
from bottom to top. In tidal flow wetlands (a specific type of upward-flow wetland), the
44

wastewater is applied from the bottom, moving upward until filling the substrate, and then the
wastewater is drained (Kadlec and Wallace 2009). During intermittent downward-flow
application, wastewater floods wetland surface, then moves downward through the filter media
by gravity. The intermittent application of the wastewater allows for higher levels of oxygen
transfer. Consequently aerobic conditions prevail in the filter media and facilitates chemical
oxidation processes (Vymazal 2007). In VSSF wetlands, the wastewater is only momentarily
exposed to the atmosphere, thus VSSF wetlands do not provide habitat suitable for mosquitoes
and partially provide odor control and minimize exposure to pathogenic organisms (Vymazal
2007, Kadlec and Wallace 2009, Stefanakis and Tsihrintzis 2012) (Figure 5.4).
In VSSF and HSSF wetlands, similar processes account for removal of organic matter
and suspended solids. However, removal is greater due to aerobic conditions in VSSF wetlands.
For example, Summerfelt, Adler et al. (1999) treated sludge from the tilapia aquaculture system
(0.6 m3/d at 30 kg/m2/yr) using a 4.5 m2 VSSF wetland, and, compared to the HSSF wetland, the
authors found superior removals of organic matter (92%) and suspended solids (98%) from an
inlet concentration of 6,855 mg COD/L and 7,860 TSS/L, respectively (Summerfelt, Adler et al.
1999). Filtered and trapped organic matter and suspended sediments can impact the filter media
of VSSF wetlands; however, intermittent application of wastewater minimize clogging due to
degradation of organic matter (i.e., volatile solids) during the resting period (Leverenz,
Tchobanoglous et al. 2009).
In VSSF wetlands, the primary pathway for nitrogen removal is nitrification followed by
denitrification. Aerated conditions in the filter media prevail, thus VSSF wetlands have the
ability to oxidize NH4 to NO3- by nitrification (equation 2.1). For example, in a laboratory scale
VSSF wetland (surface area = 0.13 m2), Xinshan, Qin et al. (2010) treated high concentration

45

nitrogenous domestic wastewater. Due to high dissolved oxygen (DO) concentration (>1.5 mg/L)
the NH4 was mostly converted to NO3- in a hydraulic retention time (HRT) of two days;
however, extended HRT reduced nitrification rates due to lack of carbon source (Xinshan, Qin et
al. 2010). In contrast, in VSSF wetlands, denitrification is limited (equation 2.2). This is reported
in Summerfelt, Adler et al. (1999) by data showing 45 mg NO3-/L from the VSSF wetland
effluent compared to 0.38 mg NO3-/L from the HSSF wetland effluent. Likewise, minor
mechanisms, such as ammonia volatilization, ammonia adsorption, and microbial and plant
uptake, enhance TN removal from the wastewater (Vymazal 2007). For example, NO2- and NO3can be assimilated by microbial and plant biomass. However, nutrients can be released back if
microbes or plants die in the wetland. Thus, total nitrogen removal by plants only occurs if plants
are harvested (Kadlec and Wallace 2009, Avila, Salas et al. 2013).
Plant uptake, microbial uptake, sorption, and precipitation are responsible for phosphorus
removal in VSSF CTW; however, removal is limited (Vymazal 2007, Stefanakis, Akratos et al.
2014, Vymazal 2014). Phosphorus is removed by plant uptake only if plants are harvested.
Microbes uptake phosphorus, which is released back when microbes die. Similar to HSSF
wetlands, sorption is typically the primary removal mechanisms but depends on the material of
the filter media; additionally, phosphorus sorption is limited because of the relatively rapid
movement of wastewater through the filter media (Stefanakis and Tsihrintzis 2012). Compared
to a downward-flow VSSF wetland, sorption of phosphorus was higher in an upward-flow VSSF
wetland due to longer HRT (Campbell and Safferman 2015). Similarly, in HSSF wetlands,
removal capacity of the filter media decreases with time. For example, Campbell and Safferman
(2015) reported that phosphorus removal decreased from 74% at the beginning of the experiment
to 30% at the end of the experiment due to less availability capacity in the filter media. Similarly,

46

after 3 years of operation of a VSSF wetland, no phosphorus removal was reported by Avila,
GarcĂ­a et al. (2016) as effluent concentrations were equal to inlet concentrations.

Figure 2. 4. A schematic of a VSSF wetland. Figure was taken from Kadlec and Wallace (2009).

2.3.2.3. Free water surface wetland
FWS wetlands consist of an area of open water, usually a shallow channel or basin, in
which the wastewater freely flows above the ground surface (Kadlec and Wallace 2009,
Stefanakis and Tsihrintzis 2012). Different types of plants can grow in FWS wetlands. Emergent
plants are rooted in the soil growing beneath the water surface, with leaves, stems, and
reproductive parts above the water, while submerged plants are also rooted in the soil, with all
their parts growing beneath the water surface (Mitsch and Gosselink 2007). Free floating plants
have their roots hanging freely in the water, with leaves and stems floating on the water surface;
whereas, floating-leaved plants have leaves that float on the water surface but roots that are
grounded in the soil (Mitsch and Gosselink 2007). Parts of plants beneath the water surface
(roots, leaves, stems) and roots in the underlying sediments provide surface area for
microorganisms to grow. Roots exude oxygen into the surrounding areas for aerobic microbial
activities. In addition, via photosynthesis, algae supply oxygen in the water column; however,

47

algae activity can decrease if floating plants cover the wetland surface. Because standing
wastewater is exposed to the atmosphere, FWS wetlands provide habitat suitable for mosquitoes
and there is a health risk due to exposure of pathogenic organisms (Vymazal 2007, Kadlec and
Wallace 2009). Thus, to avoid any health risks, FWS are used for secondary or tertiary treatment.
In addition, due to the open water area, the DO concentration varies from high (near the surface)
to low (near the bottom) (Stefanakis, Akratos et al. 2014). Thus, a variety of chemical oxidation
and reduction processes can occur within the water column (Figure 2.5).
FWS wetlands can be effective at removing suspended solids and organic matter.
Sedimentation, filtration, aggregation, and surface adhesion can remove suspended sediments.
Particles settle and filter out as the wastewater flows through the wetland vegetation. Vegetation
enhances sedimentation by reducing water velocities, intercepts, and filters particles so that the
particles are sorbed to surface areas of the plants. El-Sheikh, Saleh et al. (2010) reported that
suspended solids were reduced by 76% by a 12,500 m2 FWS wetland treating 4,300 m3/d of
wastewater from the Bahr El Baqar drain in Egypt. Like suspended sediment removal, organic
matter can be removed through deposition and filtration, particularly the largest and heaviest
particles. In addition, biological degradation removes organic matter. Aerobic degradation can be
performed by microbial growth in the water column, whereas anaerobic degradation occurs
mainly at the bottom of the wetland. Burgoon, Kadlec et al. (1999) reported 87% of COD
removal for a 100,000-m2 FWS wetland receiving potato processing wastewater with 2,528 mg
COD/L. Removal was in part performed by a large population of aerobic and anaerobic
microorganisms, and facilitated by larger wetland size and HRT (Burgoon, Kadlec et al. 1999,
El-Sheikh, Saleh et al. 2010).

48

Nitrification and denitrification is the most effective process for nitrogen removal in FWS
wetlands. As mentioned before, the FWS wetlands provide aerobic zones (water column) to
oxidize NH4+ to NO3- and anoxic zones (sediments) to reduce NO3- to nitrogen gas or nitrous
oxide. For example, Kapellakis, Paranychianakis et al. (2012) reported high removal efficiencies
of NH4+ and NO3- due to paired nitrification and denitrification in a FWS wetland treating olive
mill wastewater. Minor mechanisms that also remove nitrogen from wastewater are ammonia
volatilization, ammonia adsorption, anaerobic ammonium oxidation (ANAMOX), organic
nitrogen burial, and plant uptake if plants are harvested (Vymazal 2007). In general, FWS
provide aerobic and anaerobic environments and carbon for nitrogen removal. Consequently,
larger FWS wetlands with larger HRT have demonstrated high removal efficiencies. However,
this becomes a disadvantage if land is not available or costly (Burgoon, Kadlec et al. 1999, ElSheikh, Saleh et al. 2010).
Phosphorus can be removed by plant uptake, microbial uptake, adsorption, and soil
formation in FWS wetlands. Plant and microbes incorporate phosphorus in their tissues;
however, phosphorus can be released back to the wetland if their organic matrices degrade. By
harvesting plants, phosphorus can be physically removed from the wetland. Adsorption of
phosphorus occurs in the soil surface and increases in soils with high clay content and high
concentrations of aluminum, iron, and calcium. In addition, phosphorus precipitation can also
occur if aluminum, iron, and calcium are in the soil. However, phosphorus can be re-dissolved
under altered conditions. In fact, FWS wetlands can export phosphorus. Long-term phosphorus
removal, which is commonly thought to be limited to ~10% of influent phosphorus, is due to soil
formation based on sediments and litter (Vymazal 2007, Kadlec and Wallace 2009).

49

Figure 2. 5. A schematic of a FWS wetland. Picture was taken from Kadlec and Wallace (2009).

2.3.2.4. Hybrid constructed treatment wetland
HCTWs combine multiple types of wetlands to overcome disadvantages of individual
treatment wetlands types. Depending on wastewater characteristics, type, quantity, and loading,
wetlands can be selected to achieve higher removal efficiencies. The VSSF wetland followed by
a HSSF wetland configuration targets NH4 removal by nitrification in the VSSF wetland
followed by denitrification in the HSSF wetland. Carbon source could be limited for
denitrification by high removal of organic matter at the VSSF wetland; however, this is of
limited concern for treatment of high strength wastewaters. Combining a VSSF wetland with a
FWS wetland also promotes sequential nitrification and denitrification. In this case, carbon can
be supplied by decomposed organic matter at the bottom of the FWS wetland. A third option for
nitrogen removal could be a sequential HSSF wetland with a VSSF wetland for high
concentrations of NO3- in the wastewater. Denitrification is performed in the HSSF wetlands
using organic matter in the wastewater as carbon source. Then, nitrification is performed in the
VSSF wetland, and some portion of the effluent can be recirculated into the HSSF for
denitrifying, enhancing the TN removal. Table 2.7 demonstrates the efficacy of different HCTW
configurations.
50

Table 2. 7. Hybrid constructed treatment wetland efficacy.
Location

Denmark
(Brix, Arias
et al. 2003)
Nepal
(Laber,
Haberl et al.
2003)
Spain
(Serrano, de
la Varga et
al. 2011)
Spain
(Avila,
GarcĂ­a et al.
2016)

Type of
wastewater

Sewage

Sewage

Winery

Sewage

Type of
CTWs

Water quality parameter
(In / Out, mg/L)
COD

TSS

NH4-N

NO3-N

TN

TP

HSSF
+
VSSF
HSSF
+
VSSF

376 / 50

n.r.

17 / 16

2 / 3.1

40.6 / 20.7

n.r.

50 / 36
162 / 45

n. r.
57 / 19

16 / 0.4
32 / 27

3.1 / 6.6
0.2 / 0.4

20.7 / 16.8
n.r.

n.r.
4.4 / 2.6

45 / 10

19 / 1.5

27 / 0.1

0.4 / 27

n.r.

2.6 / 1.4

VSSF
+
HSSF

1,558 / 711

129 / 65

28 / 19

n.r.

52.9 / 26.0

n.r.

711 / 448

65 / 17

19 / 12

n.r.

26.0 / 25.2

n.r.

VSSF
+
HSSF
+
FWS

335 / 147

166 / 47

34 / 17

** / 14

n.r.

3.2 / 3.2

147 / 82

47 / 12

17 / 10

14 / 13

n.r.

3.2 / 3.2

82 / 73

12 / 8

10 / 7.2

13 / 7.7

n.r.

3.2 / 3.2

**: below limit of detection
n.r.: no reported

2.3.2.5. Selection criteria
In general, for the same wastewater flow and characteristics, VSSF wetlands achieve
higher removals than FWS due to predominant aerobic conditions and high specific surface area
of the media. Removal of organic matter, suspended solids, and NH4 are higher in VSSF
wetlands than in FWS wetlands. In contrast, NO3- removal is higher in HSSF wetlands. FWS
wetlands have also demonstrated high removal efficiencies; however, FWS wetlands require
larger areas to achieve the same treatment VSSF or HSSF wetlands. If land is available and
inexpensive, FWS wetlands can be an attractive option. Investment, maintenance, and operation
costs are lower for FWS wetlands than for VSSF and HSSF wetlands. If land is not available,
VSSF and HSSF wetlands can be reliable options. However, both systems are relatively costly
compared to a FWS wetland. The selection of subsurface wetland type would depend on the
51

target pollutant. If wastewater has high NH4 concentration, a VSSF is preferable, but if the target
is the removal of NO3-, a HSSF is suggested. Table 2.8 summarizes CTW treatment properties
for selection criteria.

Table 2. 8. Treatment properties of constructed treatment wetlands.
Treatment properties1

CTW type
FWS

HSSF

VSSF

1

Excellent reduction of organic matter and suspended solids.
Excellent for denitrification of nitrate and final removal of total nitrogen.
Not effective nitrifying ammonium to nitrate.
Excellent reduction of organic matter and suspended solids.
Excellent for denitrification of nitrate and final removal of total nitrogen.
Not effective at nitrifying ammonium to nitrate.
Compared to a FWS: For the same wastewater (characteristic and flow rate), the HSSF
possesses higher investment cost, but requires less treatment area, provides mosquito control,
odor control, and avoid contact with the wastewater.
Excellent reduction of organic matter and suspended solids.
Excellent for nitrification of ammonium.
Not effective denitrifying nitrate to nitrogen.
Compared to HSSF: For the same wastewater (characteristic and flow rate), the VSSF requires
less treatment area, but require more operation and maintenance. Both wetland types have
similar investment cost, provides mosquito control, odor control, and avoid contact with the
wastewater.

(Kadlec and Wallace 2009)

Therefore, in Costa Rica, a HCTW can be implemented to treat the wastewater from a
thermophilic CSTR anaerobic digester and assure water treatment. The implementation of CTWs
has been demonstrated (Table 2.5); however, these systems have been designed for sewage
wastewater. The effectiveness and removal capacity of a CTW treating the effluent from a
thermophilic CSTR anaerobic digester in Costa Rica has not been reported. The design of CTWs,
including the evaluation of the hydrological balance, treatment performance, and modeling
approach, is discussed in Chapter Four. In addition, strategies to prevent clogging need to be
evaluated for treating high strength wastewater to consider the utilization of a VSSF wetland.
Clogging in the tropics is discussed in Chapter Five.

52

2.3.3. Closing the loop for water, energy, and food
Current agro-industrial and agricultural activities are based on linear economies. Under
this model, water, energy, and food resources are unsustainably demanded and improper disposal
of wastes provokes environmental degradation, which restricts resource availability. As an
alternative, in a circular economy, the waste is considered as a resource. Anaerobic digestion can
treat the biomass residues to obtain energy and fertilizers. The digestion process can be enhanced
by promoting thermophilic processes. Solar energy can be capture through solar thermal
collector to provide energy for heating digesters to thermophilic conditions. In Costa Rica, due to
warmer conditions, solar thermal collectors can supply enough energy to maintain thermophilic
digestion, which, when compared to mesophilic digestion, enhances solids degradation, yields
more biogas, and releases pathogen-free digestate. The liquid digestate still has high
concentration of solids, organic matter, and nutrients. To avoid water pollution, a HCTW can be
implemented to reclaim water. Closing the loop by including an integrated system, such the one
described, into current industries can decrease waste disposal into the environment and reduce
demand for water, energy, and food.

2.4. Technical assessment of a sustainable development
Sustainability is a complex and broad concept, which varies in interpretation depending
on the discipline and context. Multiple definitions can be found in literature. The 1987
Brundtland Report of the World Commission on Environment and Development defines
sustainable development as “development that meets the needs of the present without
compromising the ability of future generations to meet their own needs”. The United Nations
defines sustainable development as “the process that constrains resource consumption and

53

waste generation to an acceptable level, makes a positive contribution to the satisfaction of
human needs, and provides enduring economic value to the business enterprise” (UN 2015).
Both definitions imply that any action in present societies should not threaten the satisfaction of
existing and future generations.
Multiple social and technical assessments are available to evaluate the sustainability of
new systems or technologies. Social assessments are complex as they evaluate the interaction
between consumers, society, and the environment. For social assessment of sustainability,
collaboration of stakeholders is needed to determine the economic, social, and environmental
benefits or affectations a new development would yield on the population. Social assessments are
based on sustainable development indices, and interpretations largely vary depending on
economic, social, and environmental evaluator’s point of view.
In contrast, technical assessments of sustainability are based on quantifiable metrics with
relatively objective interpretations. Metrics assess direct interactions between new developments
or systems and society. Customer satisfaction, safety, security, efficiency, lifetime, and cost, are
some metrics that can easily indicate the benefits or affectations a new development would yield
on society. For example, exergy is a thermodynamic metric that describes system performance
according to the second law of thermodynamics. Opposite to entropy, which measures the lowquality energy of materials in thermodynamic disequilibrium, exergy measures the quantity and
quality of energy that a particular material possesses if it is brought into thermodynamic
equilibrium (Jørgensen 2006, Rosen 2012, Querol, Gonzalez-Regueral et al. 2013). A system
that converts materials with high entropy into high quality end products with low entropy will be
sustainable if the balance between inputs (e.g.: wastes) and outputs (e.g.: biogas, fertilizers, and
reclaimed water) is positive (Wall 2010, Woudstra 2016). Therefore, exergy serves as a powerful

54

tool for improving sustainability as identification and implementation of systems that utilize
wastes would reduce resource consumption and waste generation while providing beneficial
outcomes from wastes to satisfy human needs. In thermodynamic terms, Hornbogen (2003)
indicate that waste utilization reduces entropy in the environment at the time high quality end
products (with high exergy) are yielded, increasing sustainability (Figure 2.6). This approach is
used to evaluate sustainability is this dissertation.

Figure 2. 6. Entropy, exergy and sustainability (Hornbogen 2003).

55

CHAPTER 3: TECHNICAL PERFORMANCE OF A SOLAR-POWERED WASTE
UTILIZATION AND TREATMENT SYSTEM IN COSTA RICA

Abstract: Organic wastes are seen as residues, even though energy can be extracted from
them. To simultaneously treat and utilize organic wastes, it is proposed to combine solar thermal
collector, anaerobic digester, and constructed treatment wetland technologies. The goal of this
study was to evaluate the technical performance of: 1) solar thermal collectors for maintaining
thermophilic temperature in the anaerobic digester; 2) thermophilic anaerobic digestion for
converting organic wastes into energy and fertilizers; and 3) VSSF-CTW for treating water from
liquid digestate. In addition, a cash flow analysis was conducted to determine the payback period
of the proposed system. Daily, the system utilizes 863 kg of mixed animal and food wastes to
generate 263 MJ renewable energy, produce 28 kg nitrogen and phosphorus fertilizer, and
reclaim 550 kg water. The net revenue, considering electricity and fertilizer production, is $2,146
annually. The payback period for the system is estimated to be 21 years. A sensitivity analysis
demonstrated that by optimizing the total solids in the feed and reducing the solar thermal
collector area, the payback period can be reduced to 9 years. The implemented system has
successfully demonstrated a self-sufficient and flexible waste utilization and treatment system. It
creates a win-win solution to satisfy the energy needs of the community and address
environmental concerns of organic wastes disposal in the region.

3.1. Introduction
The agriculture sector, as the second largest industry in Central America, has contributed
an average of 9.19% of the total Gross Domestic Product (GDP) of Costa Rica in the past decade

56

(EN 2015). Agricultural and agro-industrial activities generate a vast amount of organic wastes,
such as animal manure, pineapple residues, sugarcane bagasse, rice straw, and coffee residues.
Combustion of dry residues (i.e., sugarcane bagasse) and land application of wet wastes (i.e.,
animal manure), which are the most often used disposal approaches, have unfavorable economic
performance and produce greenhouse gas (GHG) emissions, air pollution, and water pollution.
For example, it is estimated that equipment costs for combustion ranges from 2,500 to 4,000
USD/kw; whereas, anaerobic digestion ranges from 1,650 to 1,850 USD/kw (IRENA 2012). On
the other hand, if land application of wet waste is not properly managed, nutrient and carbon
pollution becomes an environmental risk. Runoff and erosion can transport nutrients into water
bodies and provoke eutrophication (Sutton and Joern 1992, Johnson and Eckert 1995). Then,
volatilization of ammonia and emission of methane can occur within after land application
(Sherlock, Sommer et al. 2002, Huijsmans, Hol et al. 2003, Safferman and Wallace 2015).
Instead, organic wastes, rich in proteins and high-caloric carbohydrates, are potential
renewable resources for clean energy generation. Previous estimates suggest that approximately
600 MW electricity can be generated in Costa Rica from the agricultural residues each year
(Coto 2013). However, only 18.9 MW of electricity is currently generated from organic residues,
which is merely 0.73% of the total power capacity of 2,600 MW in Costa Rica (ICE 2016,
Kohlmann 2016). Development and implementation of integrated technologies to treat
agricultural wastes can create opportunities to alleviate negative environmental impacts of
organic waste streams, increase access to affordable clean energy, and reduce GHG emission in
rural communities in Costa Rica.
Anaerobic digestion is a natural biological conversion process that is proven effective at
converting wet organic wastes into biogas and producing clean electricity while also reducing

57

GHG emissions and nutrients, carbon, and solid loads in the effluent digestate (AgStar 2010).
Based on its operating temperature, anaerobic digestion can be categorized into thermophilic and
mesophilic digestion. Thermophilic digestion occurs at a temperatures >45°C, while mesophilic
digestion occurs at temperatures between 25 – 40°C. Numerous studies have demonstrated that
thermophilic cultivation can enhance anaerobic digestion performance by shortening retention
time, improving odor control, eliminating pathogens, increasing biogas production, and reducing
total solids in waste streams (Aitken, Sobsey et al. 2005, Suryawanshi, Chaudhari et al. 2010,
Sharma, Espinosa-Solares et al. 2013, Zarkadas, Sofikiti et al. 2015). However, thermophilic
digestion requires input of thermal energy to maintain the temperature, which may lead to an
unfavorable energy balance for small-scale operations (Vindis, Mursec et al. 2009). In contrast,
large-scale operations can produce sufficient biogas to cover heating demand and maintain
thermophilic temperatures while providing surplus energy for electricity generation (Zabranska,
Dohanyos et al. 2002). However, the biogas produced is mainly used for heating purposes
(Zabranska, Dohanyos et al. 2002). In order to overcome unfavorable energy balances for smallscale operations, other renewable energy sources, such as solar energy, need to be used
(Suryawanshi, Chaudhari et al. 2010). Solar energy, an abundant renewable energy source in
Costa Rica, is an excellent candidate to combine with small-scale thermophilic anaerobic
digestion systems.
Several solar thermal conversion technologies have been developed, such as flat-plate
solar thermal collectors, evacuated-tube solar thermal collectors, parabolic trough systems,
power tower systems, and dish solar systems (Siva Reddy, Kaushik et al. 2013). Among these
designs, flat-plate solar thermal collectors are simple and economical systems that are capable of
efficiently providing the heat to maintain culture temperature during anaerobic digestion (EPA

58

1978, Alkhamis, El-Khazali et al. 2000). Furthermore, flat-plate solar thermal collectors are
suitable for the tropics, since warm weather reduces heat loss and improves thermal efficiency.
Integrating a simple solar collection method with anaerobic digestion technology may overcome
unfavorable energy balances for small-scale thermophilic operations not only for rural Costa
Rica, but also for other remote communities around Central America. In addition, the anaerobic
digester can also play an important role of storing low-density and inconsistent solar energy (as
heat) into a relatively dense and reliable biochemical energy source – methane (Tsalkatidou,
Gratziou et al. 2009).
Even with the utilization and treatment of wastes provided by the solar-heated
thermophilic anaerobic digester, the digestate still has relatively high levels of chemical oxygen
demand (COD) (more than 10,000 mg/L) and nutrients (e.g., approximately 1,000 mg/L nitrogen
and 200 mg/L phosphorus). Mechanical separation is widely adopted by anaerobic digester
operations to separate the effluent into liquid and solid digestates (Monlau, Sambusiti et al.
2015). Solid digestate, which is rich in fiber and phosphorus, can be used as a fertilizer with
enhanced nutrient retention in soils (Liedl, Bombardiere et al. 2006). As for liquid digestates,
direct land application is common. However, further treatment to reclaim water from liquid
digestate has attracted increasing attention (Carretier, Lesage et al. 2015, Sanyal, Liu et al. 2015),
including treated liquid digestate prior to reutilization of treated water for irrigation, ground and
surface water recharge, or process uses (e.g. washing floors of milking parlour).
Numerous studies have demonstrated that utilizing a constructed treatment wetland
(CTW) to treat liquid digestate is an economically and technically sound approach to reclaim
water (Denny 1997, Ritter and Shirmohammadi 2001, ITCR 2003, Kadlec and Wallace 2009).
Free water surface (FWS) and subsurface flow are two typical CTW configurations. Compared

59

to FWS-CTWs, subsurface flow-CTWs have the advantage of ensuring intensive contact
between the wastewater and microbial biofilms growing on the media (ITCR 2003), thereby
reducing the footprint of the wetland necessary to achieve treatment goals. Vertical subsurface
flow-CTWs (VSSF-CTWs) are more common than FWS-CTWs for intermittent wastewater
influents and, when surface fed, increase the aeration of the media (Kadlec and Wallace 2009).
Therefore, a VSSF-CTW was incorporated into the integrated utilization system to treat the
liquid digestate.
Costa Rica has many opportunities to increase the energy production from organic waste
resources. One potential solution is the integration of anaerobic digestion and solar thermal
technologies to simultaneously generate renewable energy and produce fertilizer. Besides the
management of organic wastes, water reclamation is important. Thus, inclusion of a posttreatment technology is important for handling the liquid digestate, which otherwise could impair
the environment if simple discharged into land or water bodies. Developed in Costa Rica, the
goal of this study was to evaluate the technical performance of an integrated solar-powered
anaerobic digester (SPAD) with a VSSF-CTW. In particular, this study evaluated the technical
performance of: 1) the solar thermal collectors as the energy source for maintaining thermophilic
temperature in the anaerobic digester; 2) the thermophilic anaerobic digestion as a technology for
converting organic wastes into energy and fertilizers; and 3) the VSSF-CTW as a liquid digestate
treatment technology for treating water. Finally, a cash flow compared the current system and a
control system without thermal collectors, and was used to determine the payback period of the
current system.

60

3.2. Material and methods
In 2011-2012, a solar-powered anaerobic digester and hybrid constructed treatment
wetland (SPAD-HCTW) was installed at the University of Costa Rica (UCR) Fabio Baudrit
Experiment Station (EEAFBM) located in Alajuela, Costa Rica (10.00 m N, -84.26 m W). In
March 2013, the SPAD-HCTW started continuous operation. This study was conducted from
August 2015 to March 2016, and constituted the first set of data continuously collected since
operation began in March 2013.

3.2.1. System description
The SPAD included a modified flat-plate solar thermal collector, a thermophilic
continuous stirred tank reactor (CSTR) anaerobic digester, and electrical generators (Figure 3.1).
The modified flat-plate solar thermal collector converted solar energy into thermal energy to heat
the influent of anaerobic digester and maintain the digester at thermophilic condition. A methane
biogas storage bag served as the fuel storage. Solid effluent from the digester was composted and
liquid digestate was post-treated by the HCTW, a VSSF-CTW working in series with a FWSCTW. This study focused on the SPAD, thus only the VSSF-CTW has been included to
demonstrate how these two technologies can create a close-loop system utilization of organic
wastes and liquid wastes. Analysis of the HCTW and clogging status of the VSSF-CTW was not
part of this study. Materials for construction of the SPAD-HCTW were bought at local suppliers
in Costa Rica. The detailed individual units of this system are described as follow and shown in
Figure 3.2.

61

Figure 3. 1. Flowchart of the SPAD and VSSF-CTW. Mass flow is represented with a continuous
line (-). Energy flow is represented with a hidden line (- -).

Figure 3. 2. Individual units of the system at the Fabio Baudrit Agricultural Station. a. Solar
thermal collectors, b. Grinder, c. Feeding tank, d. Thermophilic CSTR-anaerobic digester (silver
tank) and hot water tank (green tank), e. Liquid/solid separator, f. Effluent storage tank, g.
Biogas bag, h. Engines, i. VSSF-CTW.

3.2.1.1. Solar thermal collection
The solar thermal collection unit aimed to provide sufficient thermal energy to maintain
consistent thermophilic temperature in the anaerobic digester in the tropics. Thus, waste heat to
maintain anaerobic digester temperature was not evaluated in this study. The solar thermal
heating module consisted of a circulation pump (Model UP 26-99 F from Grundfos, Olathe, KS),
a heat exchanger, and 36 m2 of flat-plate solar thermal collector. Eighteen 2 m2 thermal
collectors (Termi-solarÂŽ, Costa Rica) were installed in three parallel rows of six collectors each
62

row (Figure 3.2a). The average annual irradiance at the site was 10.2 MJ/m2 (Wright 2008). The
collectors were installed facing south at a 10° angle. Aluminum bronze (90/10) coils were used
as the heat tubes in the solar thermal collectors. Water was the heat transfer fluid. The heated
water was then stored in a 5 m3 hot water tank (green tank in Figure 3.2d). A hot-water pump
(Model PB 351MA from Wilo, Korea) circulated the hot water to heat the digester and maintain
thermophilic temperatures (45 ¹ 2°C) using a 40 m High-Density-Polyethylene (HDPE) tubing
heat exchanger in the digester.

3.2.1.2. Thermophilic CSTR anaerobic digester
The implemented system included an anaerobic digester tank, a feeding tank, and an
effluent storage tank (Figure 3.2c, d, and f). All vessels were cylindrical tanks with flat bottoms
made with HDPE. The effective volume of the digester was 20 m3. The feeding and effluent
tanks were 10 m3 each. Food waste was transported from a nearby food distribution facility and
consisted of non-commercial over-ripe or damaged vegetables and fruits, including cucumbers,
peppers, avocado, papayas, pineapples, and tomatoes. Food wastes were ground by a grinder
(Figure 3.2b) (model Leeson C 184K17FB150 from ICAFE ÂŽ, Costa Rica), mixed with chicken
litter and treated water from the HCTW. This mixture was stored in the feeding tank. The
chicken litter was collected from a chicken farm at EEAFBM. The food wastes (1.93 kg dry
mass per day) and chicken litter (23.1 kg dry mass per day) were mixed at an average ratio of
1:12 (dry mass) with five cubic meters of water from the HCTW to target two percent of total
solids (TS) in the feed. The characteristics of the feed are listed in Table 3.1. In the feeding tank,
the mixture was mixed for 30 minutes per week by an external feeding tank pump (model AMT
P/N 1626-305-00 from AMT, Royersford, PA). From Monday to Friday, one cubic meter with a

63

total of 863 kg of the feed with an organic loading of 0.50 kg VS/m3 digester volume/day was
pumped (model AMT P/N 1626-305-00 from AMT, Royersford, PA) into the anaerobic digester
from the feeding tank. The average hydraulic retention time (HRT) of the digestion was 20 days.
The anaerobic digester was a thermophilic CSTR reactor with a submersible digester mixing
pump (model 5763 from AMT, Royersford, PA) that mixed the anaerobic digester’s contents for
10 minutes each hour. Biogas produced was collected in a biogas bag (HDPE 60 m3 from Viogaz
ÂŽ, Costa Rica). The biogas flow rate was measured using a biogas flowmeter (EKM-PGM 75
from EKM Metering, Santa Cruz, CA) installed on a pipeline connecting the digester to the
biogas bag. A cubic meter of effluent from the anaerobic digester flowed by gravity when the
anaerobic digester was fed. After digestion, a rotary liquid/solid separation unit (ICAFE ÂŽ, Costa
Rica, particle size > 0.5 mm in diameter) was used to separate liquid and solid digestate from the
anaerobic digester effluent (Figure 3.2e). The semi-solid digestate was used as fertilizer for onsite crop applications. The liquid digestate was stored in the effluent storage tank (Figure 3.2f).
Finally, one cubic meter of the liquid effluent was discharged into the VSSF-CTW (Figure 3.2i)
by gravity approximately daily from Monday to Friday.

Table 3. 1. Characteristics of the feed.
Parameters
TS (g/L)
VS (g/L)
COD (g/L)
Total carbon (% TS)
Total nitrogen (% TS)
Total phosphorus (% TS)
pH

Mixture feed
22.00 Âą 3.30
11.60 Âą 1.33
37.99 Âą 2.75
36.40 Âą 1.30
4.50 Âą 0.20
1.20 Âą 0.11
5.49 Âą 0.12

Note. Data are the average of three replicates with standard deviation.

64

3.2.1.3. Control unit and data collection
A data acquisition system (DAQ model CR1000 Campbell Scientific, Logan, UT)
collected data from thermocouples (type K, probe ungrounded) every 20 seconds for the
feedback control to maintain the digestion temperature as ¹2°C of the set temperature (i.e.,
45oC). The DAQ sent a digital signal to power the hot-water pump if the digester temperature
was lower than the set temperature. The DAQ also recorded temperatures of the water at the flatplate solar thermal collectors and the hot water tank every 5 minutes. In addition, the DAQ
controlled a recirculation pump for the VSSF-CTW operation.

3.2.1.4. Electricity generator
The electricity generators were two 16 kW (BrancoÂŽ B4T-5000 Bioflex, Brazil) biogas
engines equipped with two activated carbon filters and a gas burner (Figure 3.2h). The biogas
flowed from the biogas storage bag through the filters into the engines. The electricity generated
from the engines was used to power pumps and other pieces of equipment in the system to
satisfy operational requirements. Electricity usage for each piece of equipment was calculated
based on duration of equipment operation (Table 3.2). Heat waste from electrical generators was
not used for heating purposes as the thermal collection unit provided sufficient thermal energy to
maintain consistent thermophilic temperature in the anaerobic digester.

65

Table 3. 2. Energy consumption per piece of equipment in the SPAD-HCTW.
Power

Time

(hp)

(h/week)

Consumed
energy
(kWh/week)

Solar heating fluid transfer pump

0.17

49

6.0

7 hours/day, 7 days/week

Digester heating pump

0.46

43

15

6.27 hours/day, 7 days/week

Digester mixing pump

3.0

28

62

10 min/hour, 24/7

Feeding tank pump

2.0

0.5

0.75

30 min/week

Solid/liquid separator

1.0

0.85

0.63

10 min/day, 5 days/week

Grinder

5.0

0.50

1.8

30 min/week

Effluent pump

0.50

0.85

0.32

10 min/day, 5 days/week

Equipment

Schedule

Feed preparation pump

0.50

0.21

0.080

12 min/week

SPAD-HCTW exit pump

1.0

1.0

0.75

1 hour/week

VSSF-CTW recirculation pump

0.50

16.6

6.2

2.4 hours/day,7 days/week

3.2.1.5. VSSF-CTW
Liquid digestate usually is classified as a high strength wastewater and post-treatment is
required to treat the water. Due to high concentrations of COD (more than 10,000 mg/L), TS
(more than 5,000 mg/L), and TN (more than 1,000 mg/L) in the liquid digestate, this study
evaluated a VSSF-CTW as the post-treatment unit. The VSSF-CTW (Figure 3.2i) was planted
with Cyperus papyrus, Iris graminea, and Canna indica in 2012. Monthly, plants were trimmed
to avoid expansion. By gravity, the VSSF-CTW received liquid digestate from the effluent
storage tank. On the north side of the wetland, a PVC inlet pipe (diameter = 0.05 m) discharged
liquid digestate into a 2 x 2 m geotextile membrane (GT 131 from Skaps, Athens, GA). The
geotextile membrane has an apparent opening size of 0.30 mm, a flow rate per square meter of
0.102 m3/s, and a permittivity of 2.20 s-1. The cell is an inverse square pyramid and the
dimensions of the VSSF-CTW are: 9 x 9 m of the bottom area, 12 x 12 m of the top area, and 1.1
m depth. The substrate media in the VSSF-CTW from bottom to top are: 0.2 m of stone (particle
size of 12-20 mm), 0.2 m of pea gravel (particle size of 4-8 mm), and 0.7 m of coarse sand
(particle size of 0.75-2 mm, 32% porosity). A recirculation pump (model WS V52 from Franklin
66

Electric, Fort Wayne, IN) was used to recirculate water from the bottom of the VSSF-CTW to
the surface. The recirculation distribution system was 1.60 m above the sand. The recirculation
distribution system consisted of a PVC network (diameter = 0.0127 m) of four upright fire
sprinkler nozzles on each corner of a square (7.00 × 7.00 m) centered in the wetland that were
obtained from a local hardware store (EPA, Alajuela, Costa Rica). The height of the recirculation
spray was chosen to decrease interference of the redistribution spray by plants. The recirculation
was carried out once per day from 2:00 am to 4:00 am to reduce losses of water due
evapotranspiration.

3.2.2. Mass and energy balance
Mass and energy balance analyses were conducted based on the data from the operation
during August 2015 to March 2016. The mass balance was used to describe the mass flow
through the system including solid digestate, CH4, CO2, and water. The energy balance analysis
was based on the mass balance and operational data. Energy input was defined as negative and
energy output was defined as positive. Energy inputs corresponded to daily average electrical
consumption from weekly operation, including pumps, grinder, and mechanical separator (Table
3.2). The SPAD with the VSSF-CTW was compared with a control thermophilic CSTRanaerobic digester system without solar thermal collectors or post treatment. For the control
system, a portion of the biogas produced would be required to maintain the thermophilic
temperature of the digester. A hot water heating unit (replacing one of the engines in the studied
system) is included to use biogas to maintain the digestion temperature. The heat input required
for the system was calculated based on specific heat (equation 3.1), where Q is the energy input
for heating the feed (MJ/d), cp is the specific heat capacity of the feed (kJ/kg °C), ρ is the density

67

of the feed (kg/L), and Δt is the change in temperature (°C). Energy was required to heat an
amount of 1,000 kg of feed per day (one cubic meter per day, considering a feed density of 1,000
kg/m3). The specific heat of the feed was estimated at 4.2 kJ/kg °C (Kosseva and Kent 2013).
The average temperature in Costa Rica was approximately 20°C (IMN 2016). The operational
temperature of the thermophilic CSTR was 45°C. Finally, since biogas was used for heat and
electric energy generation, the heating value (55.5 MJ/kg CH4) of the methane was used to
calculate the energy output. The overall methane utilization efficiency was set at 90%.
𝑄 = 𝑐𝑝 𝜌∆𝑡

[3.1]

3.2.3. Cash flow and payback period
A cash flow evaluation was conducted for the current system and compared with the
control system. The capital expenditure (CapEx) and operational expenditure (OpEx) of the
system operation were used for the cash flow evaluation. In addition, the evaluation included
revenues produced by the operation of the system. The parameters used for the evaluation are
described in Table 3.3.

68

Table 3. 3. Parameters considered for the cash flow evaluation.
System components
Feedstock
Food wastes
Chicken litter
Solar panel
Anaerobic digester
Digester technology
Loading rate (m3/day)
TS of the feed (g/L)
Retention time (day)
Reaction temperature (°C)
Biogas utilization

1,000 kg per day with 2.2% TS
No cost
No cost
36 m2 solar panel
20 m3
Thermophilic digestion
1
22
20
46
Two 16 kw engines

Land use
SPAD system (m2)
VSSF-CTW (m2)
Labor cost (operator)
Maintenance
Bioenergy, water, and
fertilizer
Bioenergy

$113/m2
180
144
20% of a full-time employee
Pumps, chemicals, and filters
--

Control system without solar
thermal collector
1,000 kg per day with 2.2% TS
No cost
No cost
None
20 m3
Thermophilic digestion
1
22
20
46
One 16 kw engine for electric
generation and a biogas burner for hot
water heating unit
$113/m2
106
144
20% of a full-time employee
Pumps, chemicals, and filters
--

On-site electricity uses,
compensating the energy demand
On-site uses, compensating the
fertilizer use
Process uses
-3%
MACRS

On-site electricity uses, compensating
the energy demand
On-site uses, compensating the
fertilizer uses
Process uses
----

Fertilizer
Water
Financial analysis
Inflation rate
Depreciation

SPAD with the VSSF-CTW

The CapEx included the acquisition of assets that would have a useful life beyond the tax
year, and that will bring benefits from the day-to-day operation of the system. In this study, the
solar thermal collection unit (solar collectors, hot water pumps, pipes, and hot water tank), the
feeding unit (feeding tank, grinder, and conveyor), the anaerobic digestion unit (vessel, pump,
and gas meter), biogas utilization unit (gas bag and engines), digestate management unit (effluent
storage tank and solid/liquid separator unit), treatment unit (VSSF-CTW, including media and
pump), and system installation were considered as CapEx. The system installation was estimated
as the 15% of the CapEx (MacDonald 2011). In addition, land utilized for installation of the
system was considered in the CapEx. A total of 74 (solar panels area) of the 180 m2 needed for
69

the solar thermal collection unit were not considered for the control system. EEAFBM is located
in a region where the land cost is $113 per m2 (Hacienda 2016). The Modified Accelerated Cost
Recovery System (MACRS) is a depreciation tool which allows for greater accelerated
depreciation over longer time periods. The MACRS was used to calculate the annual
depreciation of CapEx. The depreciation period was set at 20 years. The depreciation was
estimated for the solar thermal collection unit, the feeding unit, the anaerobic digestion unit,
biogas utilization unit, digestate management unit, and treatment unit. The system installation
and land were not included in depreciation calculations. The annual depreciation rates from
MARCRS were: 0.100, 0.188, 0.144, 0.115, 0.092, 0.074, 0.066, 0.066, 0.065, 0.065, 0.033, and
0.033 (after 10 years). Finally, feedstock (food wastes and chicken litter) was delivered on-site
free of cost and was not included as costs. Thus, no cost of acquisition and transportation for the
feedstock was considered. This case would be typical for farmers that have feedstock onsite.
OpEx included the expenses needed to run the system, such as labor and maintenance of
equipment. The labor cost was based on the current wage rate in San Jose, Costa Rica (year
2015-2016) (UCR 2016). Operation of the SPAD consisted of preparing feedstock (one hour per
week). Feeding the anaerobic digester, collecting solids from the solid/liquid separation, and
other minor activities such as clean up took 1.5 hours per week. An additional two hours per
week consisted of activities of overseeing discharge of liquid digestate effluent into the VSSFCTW and harvesting plant biomass at the VSSF-CTW. In total, labor consisted of 0.9 hours per
day, from Monday to Friday. Extra work time was needed due to unexpected maintenance of the
system and, for safety, a total of 1.6 hours per day (20% of a full-time employee) was considered
in this analysis for both current and control systems. Maintenance costs included repairs,
replacement, and unexpected maintenance, and was estimated as the 3% annually of the CapEx

70

(MacDonald 2011). The system installation and land were not considered in the CapEx for
calculation of maintenance costs. An annual inflation rate of 3% was set for OpEx based on the
five-year (2010-2015) average inflation rate in Costa Rica (EN 2015).
Revenues included electricity offset and savings on fertilizer and water. Energy saving
included the savings from both electricity and heat uses (such as heating the digester). The
energy cost was $0.20/kwh equivalent according to the utility price in Costa Rica (year 2015)
(ICE 2016). Fertilizer savings corresponded to the amount of fertilizer produced in the system.
The fertilizer cost was calculated based on nitrogen and phosphorus contents in the solid
digestate. The nitrogen and phosphorus costs based on the commercial fertilizer (year 2015) were
used for the calculation. The cost of nitrogen fertilizer was $0.95/kg nitrogen and for phosphorus
was $2.12/kg phosphorus (MEIC 2013). Water savings corresponded to treated water from the
wetland that was used to replace the demand of fresh water for dilution, washing, and other uses
for the SPAD. For the current operation, there was a minimum amount of the reclaimed water
released from the system to the irrigation system. Thus, the cost savings for irrigation was not
accounted for in the calculation. Similarly, an annual inflation of 3% was set for revenues.
The net cash flow based on depreciated CapEx, inflated OpEx, and revenues was
calculated to determine the payback period. The payback period method indicates the length of
time that the investment takes to payback costs based on the CapEx, OpEx, and revenues. In
addition, a sensitivity analysis was conducted to elucidate effects of unit operations on the
payback period of the system. Three key parameters, the solar collector area, TS of the feed, and
wetland treatment area, were investigated. All the current values were adjusted by Âą50% of their
base values to elucidate their impact on changes of the payback period. The base payback period
corresponded to the SPAD with the VSSF-CTW under the specified operational conditions. The

71

biogas production was assumed to proportionally change with the TS change in the feed. The
corresponding revenue from the energy savings was used as the base cost. The sensitivity
analysis also assumed that wetland treatment would not be affected by changes in treatment area
enough to prevent reuse of the treated water in the SPAD. Further analysis incorporating changes
in treatment needs to be conducted to conclude more accurate results.

3.2.4. Analytic method
Weekly samples were collected from August 2015 to March 2016. Liquid samples were
taken weekly from the feeding tank, the effluent storage tank, and the VSSF-CTW. Certified
methodologies for collection of samples at the Water Quality Laboratory at the Research Center
of Environmental Pollution, University of Costa Rica (CICA-UCR) were followed. Samples
were collected using 1 L bottles, capped with a lid, and were kept at 4°C until analyses.
Temperature, pH, COD, and TS (including VS and FS) were measured at EEAFBM.
Temperature and pH were measured using a pH meter (model HI-2211 from Hanna Instruments,
UK). Hach method #8000 and Hach method #8276 were followed for COD and TS, respectively.
A DRB 200 reactor (Hach product #LTV082.53.40001) and a DR 900 multiparameter handheld
colorimeter (Hach product #9385100) were used to digest and measure COD digestion vials
(high range digestion vials from 0 to 1,500 mg COD/L, Hach kit). For TS, VS, and FS, samples
were dried in disposable aluminum dishes (VWRÂŽ, catalog number 25433-008) for 24 hours in a
StabilTherm gravity oven (model OV-12A from Blue M, East Troy, WI) at 100°C. Then, dishes
were put in a desiccator and, after cooling for 30 minutes, were weighed in an analytical balance
(Ohaus Corporation, Mexico). Then, samples were put in a StableTemp furnace (model
CBFS516A from Cole-Parmer, Vernon Hills, IL) at 500°C for 30 minutes to determine FS and

72

VS. Additionally, total nitrogen (TN) and (total phosphorus) TP were analyzed at CICA-LCA
laboratory. TN (method MAQA-40) and TP (method MAQA-1) methods followed the Standard
Methods for the Examination of Water and Wastewater (Rice and Bridgewater 2012). Method
5310 B (modified) was followed for TN by a combustion method (TOC-V CSH/CSN from
Shimadzu, Columbia, MD); method 4500-P D was followed for TP by a UV-visual
spectrophotometer (Evolution 600 from Thermo Scientific, Madison, WI).
Feedstock (food waste and chicken litter), solid digestate (collected after the solid/liquid
separator), and sediments (collected at the geotextile membrane) characteristics were measured
at the Agronomy Research Center, at the University of Costa Rica (CIA). First, on-site, total
feedstock, solid digestate, and sediments were weighed by an industrial scale (Romanas Oconi
S.A., Costa Rica). Then, a sample of each one was taken to measure the TS (including VS and
FS) of the feedstock, solid digestate, and sediments following Hach method #8276. Chemical
composition (C and N) of the solids were analyzed at the CIA laboratory. Method SC09-LSFP06 followed the Dunas method to determine N and C using an autoanalyzer (Vario Cube from
Elementar, Philadelphia, PA).
Biogas samples were taken monthly using a sampling pump (SKC ÂŽ Grab Air, Bag
Sampler Cat. No. 222-2301) and stored in gas sampling bags. Bags were kept at 4°C prior to the
analysis at the Center for Research in Electrochemistry and Chemical Energy, University of
Costa Rica (CELEQ). Biogas quality (CH4 and CO2 contents) was quantified using a gas
chromatographic method (Hewlett Packard ÂŽ model HP6890 Plus, Littleton, CO) equipped with
a thermal conductivity detector. The column was maintained at 250°C and argon was used as
carrier gas. The injected sample volume was 100 ÂľL and the syringe was purged three times

73

before injection. All three laboratories, CICA, CIA, and CELEQ, follow standard methodologies
accredited by the Costa Rican Accreditation Institute (ECA).

3.3. Results and discussion
3.3.1. Solar thermal collection
The temperature profile of the SPAD demonstrated that the heat transfer fluid (water) in
the solar thermal collectors reaches a peak temperature of 82°C around noon (Figure 3.3). The
solar thermal energy kept the temperature of the 5 m3 hot water tank (green tank in Figure 3.2d)
in the range of 50 to 78°C, which held enough thermal energy to maintain the anaerobic digester
at a consistent thermophilic temperature of 46 ¹ 2°C (Figure 3.3). The temperature profile clearly
indicated that 36 m2 was sufficient for a\the solar collector to satisfy the thermal energy demand
of a 20 m3 thermophilic CSTR- anaerobic digester in Costa Rica. In contrast, an anaerobic
digester assisted by solar power in a temperate region, still required biogas energy or waste heat
from engines to maintain the digester temperature in winter months (Tsalkatidou, Gratziou et al.
2009). The stable year-round solar radiation and temperature in tropical areas were certainly
beneficial to simplify design and implementation of solar-powered systems. Similarly, there is an
advantage of utilization of solar thermal collection in the tropics if compared to an anaerobic
digester heated by a solar greenhouse. In China, Hassanein, Qiu et al. (2015) evaluated a solar
greenhouse, which basically consists on a greenhouse surrounding the anaerobic digester to keep
warm temperatures, even during winter. The greenhouse maintained high enough temperature to
perform biogas production in the anaerobic digester (28°C, 49°C, 33°C, and 16°C during spring,
summer, autumn, and winter, respectively); however, those conditions were unstable for
methanogens (Hassanein, Qiu et al. 2015).

74

120
Solar collector
Thermal storage
Digester

Temperature ( C)

100
80
60
40
20

3/2/2016

2/29/2016

2/27/2016

2/25/2016

2/23/2016

2/21/2016

2/19/2016

2/17/2016

2/15/2016

2/13/2016

2/11/2016

2/9/2016

2/7/2016

2/5/2016

2/3/2016

2/1/2016

0

Time
Figure 3. 3. Temperature profile of the solar thermal collection unit and thermophilic CSTRanaerobic digester. Data for February 2016 are presented.

3.3.2. Thermophilic CSTR anaerobic digester
As the solar thermal collection unit maintained a temperature of 46 ¹ 2°C in the
anaerobic digester, the CSTR- anaerobic digester was categorized as thermophilic. The
thermophilic CSTR- anaerobic digester was fed, from Monday to Friday, with one cubic meter
(or 863 kg) of feed with an organic loading of 0.50 kg VS/m3 digester volume/day (HRT = 20
days and mixing time of 10 minutes per hour). Under this condition, the thermophilic CSTRanaerobic digester had an average biogas production of 15.1 m3 per day, with a corresponding
CH4 content of 68% (CH4 production of 10.2 m3 per day). The TS and VS were reduced to
0.99% and 0.47% in the anaerobic digester effluent from 2.20% and 1.16% in the feed,
respectively. The TS and VS removal of 55% and 59%. Correspondingly, a CH4 productivity of

75

1.42 m3/kg VS reduced was achieved. After liquid and solid separation, the TS in the liquid
digestate was further reduced to 0.47%. The solid digestate had 15.9% TS and contained 2.62 g
TN/kg TS and 31.0 g TP/kg TS, which was used as a fertilizer for on-site applications. The pH of
the digestion was steady at 7.89 and no external pH adjustment was needed, even though the feed
had a relatively low pH of 5.45. The pH of the liquid digestate after the liquid/solid separation
was at 7.95. The strong buffer capacity of the digestion indicated that stable and robust anaerobic
microbial communities were established in the anaerobic digester. Zarkadas, Sofikiti et al. (2015)
obtained similar VS reductions (53.3 – 73.9%) and pH values (7.8 – 8.2) during the co-digestion
of food waste and cattle manure in a laboratory-scale thermophilic CSTR- anaerobic digester
(reactor volume: 118 mL at 55°C). Biogas production was doubled by our CSTR- anaerobic
digester, which produced 878 mL CH4/g of VS added compared to 370 mL CH4/g of VS added
reported by the Zarkadas, Sofikiti et al. (2015). Feedstock characteristics and frequency of
mixing can be impacting in the methane produced per gram of VS (Zarkadas, Sofikiti et al.
2015). In addition, a thermophilic (56°C) CSTR- anaerobic digester (volume capacity of 40 m3)
with a working volume of 27.4 m3 treated 1.02 m3/d of chicken litter-slurry (Bombardieri,
Espinosa-Solares et al. 2007). Mixing was performed by biogas recycling (bubbling) and a pump
that conveyed the digestate to the heat exchanger at a rate of five minutes each hour. The biogas
production was 20.5 m3/d; however the CH4 content was lower (56.3%) compared to the value
obtained in our study. Thus, CH4 production was similar, 11.48 m3/d for Bombardieri, EspinosaSolares et al. (2007) compared to the 10.2 m3/d in this study.

76

3.3.3. VSSF-CTW
Even though the SPAD converted a significant portion of VS into biogas, the liquid
digestate from the digester still had very high nutrient contents. Liquid digestate concentrations
of COD, TS, TN, and TP were 7,456, 5,200, 1,209, and 124.2 mg/L, respectively (Table 3.4).
The color of the liquid digestate was still black (Figure 3.4). Additional post-treatment was
needed to further reduce the nutrient load of the water and improve its irrigation quality. The
VSSF-CTW was used for treatment as VSSF-CTW are effective for removing TS, COD, and
NH4 (Vymazal 2007, Kadlec and Wallace 2009, Abou-Elela and Hellal 2012, BohĂłrquez,
Paredes et al. 2016). The VSSF-CTW was resilient to fluctuating loads and to low (rains lower
than 1.91 cm/week) and high precipitation periods during this study. On average, from August
2015 to March 2016, the removal of COD, TS, TN, and TP by the VSSF-CTW was 96.7%,
73.1%, 90.8%, and 99.0%, respectively. Concentrations of COD, TS, TN, and TP in the treated
water varied based on the precipitation period. During the high precipitation period, from August
to December 2015, all concentrations except TS (543 mg/L) satisfied the discharge standard in
Costa Rica (COD < 150 mg/L, TS< 50 mg/L, TN < 50 mg/L, and TP < 8 mg/L) (MINAE-MSP
2007) (Table 3.4). In contrast, during the low precipitation period, from January to March 2016,
only TP (1.88 mg/L) met the discharge standard in Costa Rica (Table 3.4). First, the geotextile
membrane removed 4.81% of COD, 27.9% of TS, 13.4% of TN, and 19.5% of TP. This
membrane avoided direct contact of the liquid digestate with the filter media and allowed easy
recovery of sediments. In fact, since March 2013, no crust formation has been observed in the
filter media of the VSSF-CTW. During the high precipitation period, the treated water was used
for primarily the SPAD and, to a lesser extent, for irrigation. For irrigation, the TS concentration
did not represent a risk compared to usual TS concentration of manure when is applied as

77

fertilizer. During the low precipitation period, the treated water was reused in the SPAD to
replace the demand of fresh water for dilution, washing, and other uses. In addition, the VSSFCTW unit demonstrated comparable performance with other studies, even with substantially
higher concentrations of influent nutrients, which indicated that the CTW was an efficient
process to treat the water from the liquid digestate. Barros, Ruiz et al. (2008) reported that using
a CTW to treat the effluent from an up-flow anaerobic sludge blanket digester reduced COD,
TN, and TP from 466, 55, 3.76 mg/L to 28, 37.5, and 2.8 mg/L, respectively (Barros, Ruiz et al.
2008). Comino, Riggio et al. (2013) used a VSSF-CTW to treat the diluted effluent of a CSTR
digester. The concentrations of COD, TN, and TP in the treated water were 194.5, 9.42, and 0.26
mg/L, respectively, with corresponding removal of 76%, 91%, and 80% (Comino, Riggio et al.
2013).

Table 3. 4. Characteristics of the liquid digestate and treated water after the VSSF-CTW during
both low and high precipitation periods.
Parameters

Liquid digestate (n=25)

COD (mg/L)
TS (mg/L)
VS (mg/L)
TN (mg/L)
TP (mg/L)

7,456 Âą 571
5,203 Âą 440
1,814 Âą 261
1,209 Âą 82
124.2 Âą 16

Treated water after the VSSF-CTW
High precipitation (n=16)
Low precipitation (n=9)
66.50 Âą 12.86
564.0 Âą 120.2
543.4 Âą 63.5
2,926 Âą 272
171.2 Âą 35.3
1,087 Âą 154
34.04 Âą 8.55
246.9 Âą 28.2
0.8010 Âą 0.1310
1.880 Âą 0.75

78

Figure 3. 4. Change of water quality during the process.

3.3.4. Mass and energy balance
The mass balance for the SPAD with the VSSF-CTW is depicted in Figure 3.5. A cubic
meter per day was fed to an anaerobic digester of 20 m3. The HRT was 20 days. The anaerobic
digester reduced 45% of TS and correspondingly produced 14.5 kg biogas containing 6.13 kg
CH4 and 8.43 kg CO2 per day from 863 kg wet feed. Biogas was stored in a 60 m3 gas bag for
electricity generation. After liquid and solid separation, 28 kg per day of semi-solid digestate
containing phosphorus and nitrogen were produced, which was used as a fertilizer for crop
farming at EEAFBM. Liquid digestate (820 kg) was further treated by the VSSF-CTW. The
VSSF-CTW reclaimed 550 kg of water per day which was used for irrigation or operation of the
SPAD (Figure 3.5).

79

Figure 3. 5. Mass balance of the SPAD with the VSSF-CTW. Densities for CH4 and CO2 were
estimated as 0.717 kg/m3 and 1.84 kg/m3.

The energy balance analysis for the SPAD with the VSSF-CTW and the control system
without solar thermal collector is presented in Table 3.5. The energy balance analysis indicated
that there was net energy output for the current system. The energy required to heat the feed and
maintain the thermophilic temperature of the anaerobic digester was 126 MJ/day. The thermal
solar collector unit provided sufficient energy to heat the feed and maintain the thermophilic
temperature of the anaerobic digester, thus no external heat input was required for the SPAD.
Instead, in the control system, a portion of the biogas produced would be utilized for the hot
water heating unit to ultimately heat the feed and maintain thermophilic temperatures in the
digester. The electricity input was electricity needed to power all equipment in the system (Table
3.2). In order to maintain a routine operation, 43 MJ/day of electricity was needed. The energy
output from CH4 combustion was 306 MJ/day. The electricity input for equipment and the
energy output from methane were the same for both current and control systems. The net energy
of the current system was 263 MJ/day, which was approximately twice as much energy produced
by the control system without solar thermal collectors (127 MJ/day) (Table 3.5). The energy
balance clearly demonstrated an advantage of use of solar thermal collectors to heat the SPAD.

80

Table 3. 5. Energy balance of the SPAD with the VSSF-CTW and the control system.
System

SPAD with the VSSF-CTW

Heat input (MJ/day)
Electricity input (MJ/day)
Energy output (MJ/day)
Net energy (MJ/day)

0
-43.4
306
262.6

Control system without solar
thermal collector
-126
-43.4
306
136.6

According to the energy balance analysis of the system, it was apparent that the energy
efficiency was enhanced by integrating solar thermal and anaerobic digestion technologies, as
yielded biogas was not used for heating. Instead, Zabranska, Dohanyos et al. (2002) reported
that, of a total 275 MWh/d produced, full-scale thermophilic CSTR- anaerobic digesters (12
anaerobic digesters, 4,800 m3 each) required 60% of biogas produced for heating purposes and
only 40% remained available for electricity production. Implementation of the SPAD and VSSFCTW would also alleviate greenhouse gas emissions and ground/surface water contamination
associated with current practices of organic residues handling in Costa Rica. For instance of the
animal agriculture in Costa Rica, if all animal wastes (approximately 2,652,143 dry metric ton
per year) from farm animals (Coto 2013) were treated by the studied system, 856,377 metric tons
of CH4 could be captured, 623,253 metric tons of dry solid digestate (as fertilizer) can be
produced, and a net energy of 37 petajoule (349 MW electricity, approximately 18 times more
electricity than current electricity produced by biomass in Costa Rica) can be generated each
year.

3.3.5. Cash flow and payback period
A cash flow evaluation was carried out to examine CapEx, OpEx, and revenue of both
current and control systems. As presented in Table 3.6, the CapEx of the current system
implementation was $86,062. Among the CapEx components, land was the most expensive (42%

81

of the total CapEx). Then, among the operation units, the solar thermal collection (16%) was the
most expensive, followed by anaerobic digestion (10%) and wetland treatment (8%). Thus, the
CapEx for the control system was estimated to be 28% less costly than the current system from
exclusion of the solar thermal collection unit and land (74 m2 less). In addition, the system
installation (15% of the CapEx) was less for the control system. Importantly, the heating system
(a hot water heating tank replacing one of the engines) did not represent an additional cost for the
current system.
The OpEx included both maintenance and labor costs, and the control system was
estimated to be 12% less costly than the current system. The maintenance cost was estimated as
the 3% annually of the CapEx. For maintenance, the CapEx did not include system installation
and land costs. Thus, maintenance was estimated to be 31% less for the control system. This
difference corresponded to exclusion of the flat-plate solar thermal collectors in the control
system. In general, maintenance included oil changes of pumps, chemicals for biogas clean-up,
filters for engines, sludge clean-up from the feeding tank and the effluent storage tank, PVC
fittings (e.g., valves), and replacements of mechanical equipment. Life expectancy of mechanical
equipment ranges from 7 to 10 years. Based on reported time operation in Table 3.2, the solar
heating fluid pump, the digester heating pump, the digester mixing pump, and the VSSF-CTW
recirculation pump should be prioritized for maintenance and replacement. Except the digester
mixing pump, all pumps can be checked and maintained every week due to the easy access to
them. In contrast, the digester mixing pump was installed inside the anaerobic digester and the
only way to check it is by monitoring the recorded data at the DAQ. Yearly maintenance is
required for this pump to avoid malfunctioning and alteration of anaerobic conditions in the
anaerobic digester. All this maintenance required $1,290 per year (Table 3.6).

82

Labor costs corresponded to daily operation (from Monday to Friday) of the current
system. The exclusion of the solar thermal collection unit and the inclusion of the heating system
did not represent a difference in labor cost for both the current and control system. Daily routine
focused on feeding and checking the system (1.5 hours per week), in addition to weekly
preparation of the feed (one hour per day, only one day per week). Moreover, two hours per
week consisted of activities such as overseeing discharge of liquid digestate effluent into the
VSSF-CTW and harvesting plant biomass from the VSSF-CTW. These activities totaled 0.9
hours per day, from Monday to Friday. Since the system was relatively simple and automated,
and considering a safety factor for unexpected maintenance, a worker spent 20% of his time
every day (1.6 hours per day) for operation. Thus, the labor cost was $2,000 per year (based on
wage rates in San Jose, Costa Rica, year 2015 (UCR 2016)) for both the current and control
system.
Revenues considered energy offset from electricity produced from CH4 combustion and
savings on fertilizer. It was estimated that the current system generates $5,436 revenue per year,
two times more than the control system. Incorporating solar thermal technologies, the biogas
production can be doubled by promoting a thermophilic environment, without the need of
external energy sources for heating (Suryawanshi, Chaudhari et al. 2010). In fact, the SPAD
provided energy savings of $5,325 per year, more than 2 times the energy savings by the control
system, which need a substantial portion of the biogas produced for heating the anaerobic
digester mixture. Fertilizer savings were the same for both the current system and the control
system. Fertilizer savings totaled to $111 per year, based on annual production of solid digestate
(fertilizer) of 1,631 kg TS with 0.31% P and 0.026% N (mass balance, Figure 3.5). Fertilizers are

83

produced regardless of the anaerobic digestion process. A net positive revenue of $2,146 per year
was realized compared to the negative revenue (-$280 per year) from the control system.

Table 3. 6. Cash flow analysis.
System components

Solar thermal collection unit
Feeding unit
Anaerobic digestion unit
Biogas utilization unit
Digestate management unit
Treatment unit
System installation
Land use
Total CapEx cost
Energy saving
Fertilizer
Net Revenue per year
Maintenance
Labor cost
Total OpEx cost
Total net revenue

SPAD with the
VSSF-CTW
CapEx
$13,500
$4,500
$9,000
$4,500
$4,500
$7,000
$6,450
$36,612
$86,062 / 49,450*
Revenue per year
$5,325
$111
$5,436
OpEx per year
$1,290
$2,000
3,290
Net revenue
$2,146 per year

Control system without
solar thermal collector
$4,500
$9,000
$4,500
$4,500
$7,000
$4,425
$28,250
$62,175 / 33,925*
$2,494
$111
2,605
$885
$2,000
2,885
($280)

*: Total CapEx not including land costs.

Under the conditions presented in Table 3.6, the estimated payback period for the current
system was 30 years. Land value ($113/m2) in the region where EEAFBM is located
substantially influenced the extended payback period obtained for the current system. Land costs
for implementing the SPAD and the VSSF-CTW can be excluded in cases were the farm has
sufficient room, such as dairy farms in rural areas. Excluding land costs, the total CapEx was
$49,450 for the current system and the cash flow analysis indicated a payback period for the
current system of 21 years (Figure 3.6). The payback period was reduced by 30% when the land
cost was not considered.

84

Figure 3. 6. The cash flow of the SPAD with the VSSF-CTW under the current operational
conditions. Land costs were not considered.

Still, a payback period of 21 years is not attractive for commercial applicability. The
system built was considered a full-scale experimental system and further studies on key
parameters need to be conducted for scaling down the system based of client requirements. A
sensitivity analysis was further carried out to delineate the impacts of three key parameters (solar
collector area, TS of the feed, and wetland treatment area) on the payback period (Table 3.7).
Results elucidated that TS of the feed was the most sensitive among the three parameters. A 50%
increase of the TS of the feed could increase the CH4 and fertilizer production to 15 m3/day and
42 kg/day, respectively. The corresponding revenue was increased to $8,154/year. Under these
conditions, the payback period can be reduced by 52% to 10 years. The second most sensitive
parameter was the solar collector area. Removing 50% of the collector area can shorten the
payback period by 12% to 18 years. The wetland area was the least sensitive parameter to

85

influence the payback period. A 50% of area reduction only caused a 7% change of the payback
period. The sensitivity analysis clearly indicated that the economic impact of the wetland was not
as large as other unit operations on the stand-alone system. Consequently, decreasing the size of
VSSF-CTW did not greatly affect the payback period; therefore, risking treatment performance
of the VSSF-CTW by decreasing its size is not economically justified.

Table 3. 7. Sensitivity analysis of key unit operations on the payback period of the current
system.
Unit
operation
Solar thermal
collection
Anaerobic
digestion unit
Treatment
unit

Key parameter
Solar collector
area (m2)
TS of the feed
(%)
Wetland
treatment area
(m2)

Values
Sensitivity
Base value
range

Corresponding base
cost for the unit
operation ($)

Change on
payback period
(%)

36

18-54

13,500

Âą 12

2

1-3

5,436

Âą 52

100

50-150

7,000

Âą7

According to the sensitivity analysis, several approaches can be adopted to improve
overall net revenues. Increasing the TS content of the feed was certainly the best option to
significantly increase the biogas production and reduce the payback period. Meanwhile,
considering the fact that the current solar thermal energy exceeded the thermal need of the
system (more collector area than needed) (Figure 3.3), reducing the collector area was the second
option to improve the economic performance. Combined together, a 50% increase in TS and
50% decrease in solar collector area could lead to a relatively short payback period of 9 years,
which would greatly improve the economic performance of the system.

86

3.4. Conclusions
Since March 2013, the system has accomplished the goal of producing energy, fertilizers,
and reclaimed water from biomass residues; however, no data had been continuously recorded
prior to this study. In order to evaluate the treatment performance of the SPAD-HCTW, this
study analyzed data from August 2015 to March 2016 (28 weeks) to demonstrate how each
separate unit of the system performed, as part of a closed-loop system to yield energy, fertilizers,
and reclaimed water.
The solar thermal collector unit provided more than sufficient heat for maintaining
thermophilic temperature in the anaerobic digester. Thus, external heating sources were not
needed (e.g., biogas, heat waste from engines), even during nights. The solar thermal collectors
simplified implementation of thermophilic anaerobic digestion in the tropics. The thermophilic
CSTR- anaerobic digester performed in concordance to similar laboratory and full-scale
thermophilic CSTR-anaerobic digesters. Importantly, VS reduction during the digestion resulted
in high quality biogas (68% CH4 content by volume) and CH4 production (10.2 m3/d). Solid
digestate, recovered after digestion, was easily manageable for producing fertilizer. The energy
balance demonstrated the benefits of integrating solar and anaerobic digestion technologies.
Compared to a thermophilic CSTR- anaerobic digester heated with a hot water heating unit
utilizing biogas, the SPAD doubled the energy production. Thus, the SPAD was self-sufficient in
terms of energy demand by the system. The VSSF-CTW was resilient to fluctuating loads of
high strength wastewater and, importantly, reclaimed water for reuse. The treatment performance
was comparable to other CTWs that treated effluent from other types of reactors. The treated
water was useful for irrigation and for replacing the demand of fresh water for dilution of the
feed.

87

From the technical point of view, the SPAD-HCTW demonstrated one approach for
treating biomass residues and wastewater in Costa Rica. The sensitivity analysis demonstrated
that optimization of parameters can reduce the payback period. Therefore, more studies to
optimize thermal solar collector area, feedstock characteristics, feeding frequency, mixing
frequency and others would be important, particularly since this study considered a limited time
(August 2015 to March 2016, 28 weeks) outside of which other external factors could impact in
the technical performance. A country, such as Costa Rica, that sells itself as a green country for
ecotourism, needs to be aware of the treatment of organic wastes, and the SPAD-HCTW can turn
an environmental and economic liability into a public and private asset.

3.5. Acknowledgments
This research was supported by the U.S. Department of State West Hemisphere Affairs
(S-LMAQM-11-GR-075). The authors would like to thank Michigan AgBioResearch for funding
this work through faculty salaries. The authors would also like to acknowledge the assistance
and support of the following laboratories at the University of Costa Rica: Agronomy Research
Center (CIA), Water Quality Laboratory at the Research Center of Environmental Pollution
(CICA-LCA), and Center for Research in Electrochemistry and Chemical Energy (CELEQ).

88

CHAPTER 4: PERFORMANCE OF THE HYBRID CONSTRUCTED TREATMENT
WETLAND TREATING DIGESTATE EFFLUENT IN COSTA RICA

Abstract: Integrated into the landscape, constructed treatment wetlands are impacted by
climatological conditions. Constructed treatment wetlands are attractive options for wastewater
treatment because they have low costs of construction, maintenance, and operation; additionally,
in tropical regions, wetlands are expected to have relatively consistent treatment year-round due
to consistently warm temperatures. However, there are limited studies monitoring performance
of constructed treatment wetlands in the tropics. Thus, there is a need to evaluate the treatment
performance of wetlands in the tropics, particularly with respect to rainy and dry seasons. This
study evaluates a hybrid constructed treatment wetland (HCTW) in Alajuela, Costa Rica that
treats high-strength liquid digestate from a thermophilic anaerobic digester. The specific goals
were to 1) evaluate treatment performance of the HCTW during rainy and dry periods and 2)
evaluate the applicability of loading charts and contaminant removal models commonly used to
describe temperate wetlands to model treatment performance of the tropical HCTW. This study
was conducted during a low precipitation period (August to December 2015) and a high
precipitation period (January to March 2016). During these two periods, the HCTW
demonstrated sufficient storage capacity, even when unexpected runoff entered the system
during the high precipitation period. Slightly better treatment of chemical oxygen demand
(COD) and total solids (TS) occurred during the high precipitation period, likely due to dilution
from precipitation. However, removal of total nitrogen (TN) and total phosphorus (TP) were
similar for both the low and high precipitation periods. Effluent from the HCTW was either
reused to dilute feed for the anaerobic digester (during both periods) or used for irrigation

89

(during the rainy period). The performance of the HCTW was good; however, more data would
be needed to narrow central tendencies represented by loading charts and contaminant removal
models.

4.1. Introduction
Worldwide, treatment of wastewater is often neglected in developing countries. It is
estimated that over 80% of wastewater worldwide goes untreated, so that more than two million
cubic meters of wastewater are discharged into the world’s rivers, lakes, and oceans every day
(Otchet 2003, UNESCO 2015). Costa Rica, located in Central America, is not an exception to
this global problem. Limited financial commitments and a lack of enforcement of regulations
endanger fresh water resources in Costa Rica (GWP 2012, Echeverria and Cantillo 2013). Only
5.81% of the 0.0860 km3/year of domestic wastewater is treated in Costa Rica (Sato, Qadir et al.
2013). Additionally, industrial wastewater is rarely treated; for example, only five percent of
3,500 industries at the Rio Virilla watershed (0.3% of the Costa Rican territory) treat industrial
wastewater (UN 2009, Coto 2013, Echeverria and Cantillo 2013). Therefore, there is an urgent
need for innovative approaches for wastewater treatment, not only in Costa Rica, but in
undeveloped and developing countries worldwide.
Constructed treatment wetlands (CTWs) can be used to treat multiple types of
wastewater. Since 1950, CTWs have been used to treat domestic, animal, mining, and industrial
wastewaters, liquid leachates, urban stormwater, and field runoff (Kadlec and Wallace 2009).
Based on the water flow regime, different engineered designs, such as vertical subsurface flow
(VSSF), horizontal subsurface flow (HSSF), free water surface (FWS), and hybrid constructed
treatment wetlands (HCTWs), aim to stabilize, sequester, accumulate, degrade, metabolize, and

90

mineralize nutrients in the wastewater, emulating natural wetlands (Halverson 2004, Vymazal
2007). CTWs are resilient to fluctuating loads, require low technical expertise to operate,
typically require less time for operation and maintenance, and are typically lower in costs than
conventional wastewater treatment (Kadlec and Wallace 2009, Konnerup, Koottatep et al. 2009).
The longevity of CTWs varies depending on the system; however, estimations vary from 10 to
15 years (Wallace and Knight 2006). There are limitations to use of CTWs for decentralized
wastewater treatment. Subsurface-flow CTWs require preliminary treatment of the wastewater to
remove sediments that would clog the wetland; additionally, CTWs typically require larger areas
than centralized wastewater treatment facilities (Dallas, Scheffe et al. 2004, Langergraber 2007).
The hydrological balance is an important component of CTW design, including
estimation of the saturated water depth in the wetland. The saturated zone plays an important role
in wastewater treatment as it conveys the pollutants in the wastewater through the filter media or
vegetation, including the root zone (Kincanon and McAnally 2004). Therefore, a hydrological
balance of the wastewater inflows, outflows, and storage within the CTW is crucial to
determining the capacity of a CTW, the hydraulic loading rate (HLR), and the hydraulic
retention time (HRT). Equation 4.1 shows the inflows and outflows for the hydrological balance
of a CTW. Qi (m3/d) is the influent flow, Qo (m3/d) is the effluent flow, Qc (m3/d) is the influent
flow due to runoff, Qf (m3/d) is the loss of flow due to filtration, and Qgw1 and Qgw2 (m3/d) are the
flow additions and losses due to groundwater discharge or recharge. P (m/d) is the precipitation
depth, which is multiplied by the CTW surface area, A (m2). ET(A) (m3/d) is the
evapotranspiration losses (in height) multiplied by the CTW surface area. Ds (m3/d) is the change
in volume of water stored in the CTW (Kadlec and Wallace 2009).
𝑄𝑖 − 𝑄0 + 𝑄𝑐 − 𝑄𝑓 − 𝑄𝑔𝑤1 + 𝑄𝑔𝑤2 + 𝑃(𝐴) − 𝐸𝑇(𝐴) = 𝐷𝑠

91

[4.1]

Usually, CTWs are lined, either by installing a geomembrane or using a compacted clay
layer. Therefore, Qf, Qgw1 and Qgw2, are zero, and equation 4.1 takes a simplified form (equation
4.2).
𝑄𝑖 − 𝑄0 + 𝑄𝑐 + 𝑃(𝐴) − 𝐸𝑇(𝐴) = 𝐷𝑠

[4.2]

Contaminant removal models (e.g.: black-box models) are commonly used for the design
of CTWs (Reed, Ronald et al. 1995, EPA 2000, Shepherd, Tchobanoglous et al. 2001, Kadlec
and Wallace 2009). These models do not aim to accurately delineate treatment processes, but
serve as a simple tool for designing CTWs based on influent concentrations, HLR (or HRT), and
pseudo-first-order areal rate coefficients (k). Pseudo-first-order areal rate coefficients depend on
temperature, pollutant concentration, and other factors (Kadlec and Wallace 2009). These
coefficients indicate an empirical rate of pollutant removal due to the sum of all biological and
physicochemical processes that occur in the CTW (Stein, Biederman et al. 2006). The Arrhenius
relationship describes the rate coefficient’s dependency on temperature (equation 4.3), where kT
is the rate constant at a specified temperature T (d-1), k20 is the rate constant at 20°C (d-1), θ is a
dimensionless temperature coefficient, and T is the temperature (°C). The dimensionless
temperature coefficient indicates the influence of the water temperature on the removal of
pollutants. If θ = 1.0, the water temperature does not influence in the pollutant removal. But, if
θ>1.0 the pollutant removal rate kT increases with increasing water temperature. In contrast, kT
decreases with increasing water temperature if θ <1.0 (Kadlec and Knight 1996). Reported
values for θ for constructed treatment wetlands range from 0.946 to 0.985 for biological oxygen
demand (BOD), 1.001 to 1.062 for TN, 1.005 to 1.072 for ammonium (NH4), 1.043 to 1.177 for
nitrate (NO3).
𝑘𝑇 = 𝑘20 × 𝜃 𝑇−20°𝐶 [4.3]

92

For design purposes, k values serve as an approximation. Kincanon and McAnally (2004)
suggest to design using conservative (low) k values, as the microbial activity in the proposed
CTW is still unknown. Usually, the k value used for design corresponds to the coldest month.
Four contaminant removal models that rely on k or similar pseudo-first-order rate coefficients are
the plug-flow, the k-C*, tank in series (TIS (P k–C*)), and time-dependent retardation (TDR)
models (Table 4.1).

Table 4. 1. Contaminant removal models for designing CTWs.
Model approach

Governing equation
𝐶
= 𝑒 −𝐾𝑇 × 𝑡 [4.4]

Plug-flow

𝐶𝑖
𝐶−𝐶 ∗

Modified k–C*

𝐶𝑖 −𝐶 ∗
𝐶−𝐶 ∗

TIS (P k–C*)
TDR

𝐶𝑖 −𝐶 ∗
𝐶
𝐶𝑖

= 𝑒𝑥𝑝 [(

=𝑒

=

1
𝑃

(1+𝑘⁄𝑃𝑞 )

−𝑘0
𝑏

−𝐾𝑇 × 𝑡

References
(Reed, Ronald et al. 1995)

[4.5]

(Kadlec and Wallace 2009)

[4.6]

) 𝑙𝑛(𝑏𝜏 + 1)]

(Kadlec and Wallace 2009)
[4.7]

(Shepherd, Tchobanoglous et al. 2001)

Notes. C is effluent concentration (mg/L), Ci is influent concentration (mg/L), C* is
background concentration (mg/L), KT is the rate coefficient at a specified temperature T (d-1), t is
the hydraulic retention time (d), k is the modified first-order areal constant (m/d), P is the
apparent number of tanks, q is the hydraulic loading rate (m/d), k0 is the initial first order
volumetric rate constant (d-1), b is the time based retardation coefficient (d-1), and τ is the
retention time (d).

In the plug-flow model (equation 4.4, Table 4.1), the CTW is considered an attached
growth biological reactor, in which removal is irreversible, first-order, and homogeneous. This
model predicts the pseudo-first-order removal of individual pollutants (e.g., BOD, COD, and TS)
that are not part of multi-step decomposition processes (e.g., nitrogenous compounds) (Kincanon
and McAnally 2004). The plug-flow model assumes that all of each pollutant can be completely
degraded in the CTW. In contrast, the modified k–C* model (equation 4.5, Table 4.1) includes
the C* parameter, which is the residual or background concentration of the pollutant in the CTW.
The background concentration, C*, accounts for nondegradable compounds or minimum values

93

of nutrients due to plant senescence at low nutrient concentrations – both phenomena which
result in a non-zero lower limit that is approached asymptotically (Kadlec and Wallace 2009).
Similarly, the tank in series (TIS) model (equation 4.6, Table 4.1) uses parameters k and C*, but
also includes the parameter P, which is the apparent number of TIS. Since the assumption of a
homogeneous attached growth biological reactor with ideal plug flow is not realistic within a
wetland (Kadlec 2000), the TIS approach accounts for dispersion within the wetland. As the
number of tanks approaches infinity, the CTW acts as a plug flow reactor, while a wetland with
P=1 behaves as a continuous stirred tank reactor (CSTR). The C*, for COD in subsurface CTW,
can be estimated by using an empirical relationship, equation 4.8 (Kadlec and Wallace 2009);
however, Stein, Biederman et al. (2006) indicated that C*changes seasonally as microbial
decomposition changes with the temperature.
𝐶 ∗ = 3.5 + 0.053𝐶𝑖

[4.8]

This unsteady behavior of C* introduces more noise into the contaminant removal
modeling. The time-dependent retardation (TDR) model (equation 4.7, Table 4.1), by Shepherd,
Tchobanoglous et al. (2001), assumes that the removal rates of organic matter decrease with the
percentage of biodegradability of the wastewater, from relatively fast rates for easily
biodegradable constituents to relatively slow rates for less biodegradable constituents; time
dependence of removal rates is used in place of a constant C* value. Therefore, the TDR model
considers changes in the removal rate by replacing the C* in the k–C* model with two
parameters, K0 and b. Shepherd, Tchobanoglous et al. (2001) obtained k0 values from 9 to 12 d-1
and b values from 2 to 5 d-1, which allows a steady decrease in COD rather than an asymptotical
approach to a unique C*value. However, this model was developed for one pilot-scale CTW

94

planted with bulrush and cattails; K0 and b have not been determined or the model reproduced for
full scale systems.
Loading charts, or graphical representations of output concentrations versus input loading
rates, developed from a large number of empirical data sets are an alternative to contaminant
removal models for design purposes. Loading charts based on organic loading rates (OLR,
g/m2/d) and influent concentrations (mg/L) are used to estimate the surface area required for
desired treatment. Empirical data has also been used to prepare graphs of output versus input
concentrations. The slope of output versus input concentrations indicates the potential treatment
performance of a certain type of CTW to reduce a particular contaminant (Kadlec and Wallace
2009). However, estimating surface area with this second approach is not possible due to the lack
of area and water flow information (Kadlec 2009, Kadlec and Wallace 2009). These two
graphical approaches show a central tendency of a treatment performance for a particular type of
CTW; however, graphs should be used with caution because of considerable variability within
and between systems. First, treatment performance within an individual CTW is variable due to
seasonal and year-to-year changes (plant communities, internal hydraulics, and weather
conditions). Second, there is intersystem variability in the treatment performance due to plant
type, cell configuration (e.g., system geometry, depth, and substrate), and climate (Kadlec 2009,
Kadlec and Wallace 2009).
Models and loading graphs have been developed to describe wetland treatment
performance for most standard wastewater parameters. Contaminant removal models have been
developed for TSS, BOD/COD, total phosphorus (TP), total nitrogen (TN), NH4, NO3, and fecal
coliform (Table 4.1). Reputable loading graphs are available for HSSF-CTW of organic N,
ammonia, TN, TP, fecal coliforms, TSS, BOD, and Total Kjeldahl Nitrogen (TKN), while

95

loading graphs for VSSF-CTW are only available for TSS, BOD, and TKN (Kadlec and Wallace
2009). However, all this information was collected from temperate climates and the use of these
data to design CTWs in the tropics might contribute to incorrectly sized CTWs.
Principle component analysis (PCA) is a multivariate statistical tool that has been used to
elucidate if one or more parameters affect the treatment performance of wetlands. For example,
Lee and Cha (2015) conducted a PCA to identify key meteorological parameters that might
impact treatment efficiency of a CTW for stormwater treatment. For one year, water samples
were collected after storm events at the influent and effluent of a FWS-CTW receiving runoff
from agricultural lands in Naju city, Korea. Water quality parameters analyzed were TSS, BOD,
COD, TN, and TP. Results indicated that the first three principal components described 73.5% of
the total variance, in which rainfall depth, rainfall intensity, and antecedent dry days were
identified as key meteorological parameters. In addition, Dong and Reddy (2010) conducted
PCA to identify that, as TKN, phosphate (PO4), and COD concentrations in the wastewater
decreased across the wetland, the diversity and richness of microorganisms also decreased. These
water quality parameters were found in the first two principal components, which explained the
51% of the total variance. Moreover, PCA can be used as a tool to identify which parameters are
more mathematically robust to explain most of the variability in the data set, including
independent and dependent parameters. This second approach simplifies identification of
parameters that explain what is occurring in the treatment performance of the wetland.
CTW in temperate climates do not perform at the same efficiencies year-round due to
winter dormancy of macrophytes and microorganisms (Healy, Rodgers et al. 2007). Temperature
drastically changes due to seasonal variations in temperate regions, and temperature has been
identified by several authors as one of the important parameters determining treatment

96

performance (Bachand and Horne 2000, Kadlec and Reddy 2001, Stein, Biederman et al. 2006).
Microbial communities typically have maximum growth rates in spring and later in midsummer.
Microbial activity is driven by temperature and impacts wetland’s treatment performance. For
example, Bachand and Horne (2000) reported a reduction of nitrification rate from 1.1 to 0.10 g
N/m2d due to a change in the water temperature from 26 to 12°C; however, this is inconsistent
with reported θ values. This discrepancy may be due to changes in environmental conditions
other than temperature, such as shorter light periods and lower humidity in winters; additionally,
temperature may also indirectly affect treatment, as it also influences solubility of oxygen, plant
activity, and flow of partially frozen wastewater.
As the performance of CTWs is impacted by climatological conditions, modification of
temperate-climate design approaches for CTWs in the tropics may be needed to maximize
performance while minimizing costs. Studies indicate greater removal of pollutants in tropical
CTWs than in temperate wetlands, most likely due to steady warm temperatures year-round in
the tropics (Trang, Konnerup et al. 2010, Dan, Quang et al. 2011, Kelvin and Tole 2011, Zhang,
Jinadasa et al. 2015). However, wetland’s treatment performance might be impacted by
prolonged and intense rainfalls in the tropics. Avila, GarcĂ­a et al. (2016) indicated that rains
could impact treatment efficiencies. Initial rains of the rainy season could flush pollutants from
the surrounding area into wetlands; however, other studies have found that water inputs by
runoff and precipitation dilute effluent concentrations (SuĂĄrez and Puertas 2005, Avila, Salas et
al. 2013). This study aims to assess how the performance of a HCTW, which consisted of a
VSSF-CTW and FWS-CTW in-series, responded to periods of high and low precipitation.
Additionally, loading charts and contaminant removal models were evaluated for their
applicability to describe the treatment performance of the HCTW.

97

4.2. Material and Methods
In 2011-2012, a solar-powered anaerobic digester and hybrid constructed treatment
wetland was designed and built at EEAFBM, in Alajuela, Costa Rica (10.00 m N, -84.26 m W).
In March 2013, the system began continuous operation. This study was conducted with data
collected from August 2015 to March 2016, and focused on the hybrid constructed treatment
wetland receiving wastewater from the anaerobic digester. Details and specifications of the solarpowered anaerobic digester can be found at Aguilar Alvarez, Bustamante Roman et al. (2016).

4.2.1. System description
The CTW was fed with liquid digestate from the solar-powered anaerobic digester system
(SPAD) treating chicken litter and food waste. Effluent from the digester was separated
mechanically into liquid and solid streams with a rotary screen separator (Solid/Liquid separator)
(ICAFE ÂŽ, particle size > 0.5 mm in diameter). The solid digestate was used as a fertilizer for
on-site farming uses. The liquid digestate was stored in an effluent tank and, at a rate of one
cubic meter per from Monday to Friday, was discharged by gravity into the HCTW to be further
treated. The HCTW consists of a VSSF-CTW followed by a FWS-CTW operating in-series.
Treated water was used for irrigation or reused in the anaerobic digester (Figure 4.1). The design
of the HCTW considered Healy, Rodgers et al. (2007) and Kadlec and Wallace (2009)
approaches for intermittent sand filters and wetlands for wastewater; however, space constraints
and research goals ultimately determined the size of the HCTW.

98

Figure 4. 1. Aerial image of the HCTW. Satellite image was taken from Google Earth.

The VSSF-CTW dimensions are 12.0 × 12.0 m at the top and 9.00 × 9.00 m at the
bottom, with a median treatment area of 138 m2 (Figure 4.2). The height of the wetland is 1.10 m
and the slope of the walls is 27° with respect to the horizontal. The bottom and walls of the
VSSF-CTW are lined with clay. The media is predominantly coarse sand (particle size of 0.75 –
2.0 mm) with a porosity of 32%. Below 0.70 m of the sand is 0.20 m of pea gravel (particle size
of 4 – 8 mm) and 0.20 m of stone (particle size of 12 – 20 mm) to facilitate collection of the
treated wastewater. The VSSF-CTW has a maximum water storage capacity in the media of 30.9
m3, with an additional 80 m3 of storage above the media. A PVC inlet pipe (diameter = 0.05 m)
is located at the north side of the cell. One cubic meter per day of liquid digestate from the
effluent storage tank flows through the inlet pipe into a square 4-m2, 0.25-m deep trench of that

99

is covered with a geotextile membrane (GT 131, Skaps, Athens, GA). The geotextile membrane
has an apparent opening size of 0.30 mm, a flow rate per square meter of 0.102 m3/s, and a
permittivity of 2.20 s-1. In the bottom stone layer, a distributed grid of drainage pipe (diameter =
0.075 m) collects and conveys the treated effluent to the VSSF-CTW well pump. In the well, a
submersible pump (model WS V52 from Franklin Electric, Fort Wayne, IN) recirculates the
treated effluent to the same cell at a flow rate of 110 L/min. Each weekday, recirculation
occurred from 2:00 am to 4:00 am to avoid high losses of water from evapotranspiration.
Recirculated effluent is sprayed from four sprinklers located 1.60 m above the ground and at
each corner of a square (7.00 × 7.00 m) centered at the center of the wetland. From Monday to
Friday, 1-m3/d of the treated effluent from the VSSF-CTW was transferred to the FWS-CTW.
The VSSF-CTW was planted with Juncus effuses (common rush), Coix lacryma jobi (Job’s
tears), Cyperus papyrus (papyrus), Iris graminea (dwarf iris or grass-leaved iris), and Canna
indica (wild canna lily or African arrowroot) in 2012. After three years, Juncus effuses and Coix
lacryma jobi had not survived, and at the time of this study, only Cyperus papyrus, Iris
graminea, and Canna indica were thriving in the wetland. In August 2015, it was observed that
Cyperus papyrus dominated the wetland (Figure 4.2). To avoid expansion of Cyperus papyrus
across the wetland area and to maintain equal areas of each species, plants were trimmed
monthly.

100

Figure 4. 2. Photo and schematic of VSSF-CTW. Dimensions are in meters. Picture was taken in
August 2015.

The FWS-CTW (Figure 4.3) has the same dimensions as the VSSF-CTW. A layer of clay
was used to line the bottom and walls of the cell. The maximum storage capacity is 152 m3,
which is the total volume in the cell (138 m2 × 1.10 m deep). From Monday to Friday, the FWSCTW receives one cubic meter of treated effluent from the VSSF-CTW. A PVC inlet pipe
(diameter = 0.05 m) is located at the north side of the cell. The effluent pipe is located 2.73 m
from the edge of the HSSF-CTW at a depth of 1.09 m above the bottom. The effluent pipe, a
PVC pipe (diameter = 0.075 m), conveys the treated water to the FWS-CTW well pump. Similar
to the VSSF-CTW, a submersible pump (model WS V52 from Franklin Electric, Fort Wayne,
IN) recirculates the water from 2:00 am to 4:00 am every weekday. The recirculated effluent is
discharged from a PVC pipe (diameter = 0.012 m) located in the East side of the wetland. The
101

exit pump (model WS 102 from Franklin Electric, Fort Wayne, IN) pumps FWS-CTW effluent
to either the digester feed tank or the station’s irrigation system. The FWS-CTW is planted with
two species of floating plants, Eichhornia crassipes (water hyacinth) and Pistia stratiotes (water
lettuce) (Figure 4.3). Floating plants were divided by species by a frame of bamboo (3 m wide
and 9 m length per species), and covered almost 100% of the surface area (27 m2 for each
species) with an approximate density of 100 Eichhornia crassipes plants per m2 and 60 Pistia
stratiotes plants per m2 during the study period. One square meter of each plant was harvested
once per month to prevent overgrowth.

102

Figure 4. 3. Photograph and schematic diagram of FWS-CTW. Dimensions are in meters. Picture
was taken in August 2015.

4.2.2. Hydrological balance
A hydrological balance was conducted for the HCTW (equation 4.2). Flows into and out
of the VSSF-CTW and FWS-CTW were measured in m3/week. Liquid digestate was discharged
into the VSSF-CTW via gravity; time of flow per transfer was determined manually by recording
the time it took the flow to fill a 1 m3 tank. Flow transferred from the VSSF-CTW to the FWSCTW and treated water pumped from the FWS-CTW was measured by recording the time each

103

pump operated, as the flow rate of each pump was known from their specifications (VSSF-CTW
pump = 110 L/minute, FWS-CTW pump = 150 L/minute). A weather station located on-site
(IMN, 10.00 m N, -84.26 m W) recorded precipitation (mm), temperature (°C), relative humidity
(%), wind speed (m/s), and solar radiation (MJ/m2h) every two minutes. Meteorological data
were analyzed on a weekly basis to measure precipitation and estimate evapotranspiration.
Evapotranspiration was estimated using the FAO Penman-Monteith equation (Allen, Pereira et
al. 1998). Runoff was calculated for the drainage area of 1,500 m2, applying the curve number
method (Ward and Trimble 2004). The volume of water stored (stored volume, SV, in m3/week)
in both the VSSF-CTW and the FWS-CTW was directly measured by recording the height of
water in each pumping well on a daily basis, and averaging to find a weekly stored volume.

4.2.3. Treatment performance
Weekly samples were collected from August 2015 to March 2016. Liquid samples were
collected from the digester effluent tank and effluents of both the VSSF-CTW and FWS-CTW.
Samples were collected using 1 L bottles, which were capped with a lid and kept at 4°C until
analysis, following certified methodologies for collection of samples established by the Water
Quality Laboratory at the Research Center of Environmental Pollution (CICA-LCA), University
of Costa Rica). CICA-LCA follows standard methodologies accredited by the Costa Rican
Accreditation Institute (ECA). Measured water quality parameters were temperature, pH, COD,
TS, volatile solids (VS), fixed solids (FS), TN, NO3, NH4, TP, and PO4. Temperature and pH
were measured using a pH meter (model HI-2211 from Hanna Instruments, UK). COD and
solids, including VS and FS, were analyzed following Hach method #8000 and Hach method
#8276, respectively. A DRB 200 reactor (Hach product #LTV082.53.40001) and a DR 900

104

multiparameter handheld colorimeter (Hach product #9385100) were used to digest and measure
COD digestion vials (high range digestion vials from 0 to 1,500 mg COD/L, Hach kit). For TS,
VS, and FS, samples were dried in disposable aluminum dishes (VWRÂŽ, catalog number 25433008) for 24 hours in a StabilTherm gravity oven (model OV-12A from Blue M, East Troy, WI)
at 100°C. Dishes were placed in a desiccator to cool for 30 minutes and were then weighed on an
analytical balance (Ohaus Corporation, Mexico). Next, samples were placed in a StableTemp
furnace (model CBFS516A from Cole-Parmer, Vernon Hills, IL) at 500°C for 30 minutes to
determine FS and VS. TN (method MAQA-40), NO3 (method MAQA-20), NH4 (method
MAQA-38), TP (method MAQA-1), and PO4 (method MAQA-20) analyses were conducted at
CICA-LCA. CICA-LCA analyses are based on the Standard Methods for the Examination of
Water and Wastewater (Rice and Bridgewater 2012). Method 5310 B (modified combustion
method) was used to assess TN (TOC-V CSH/CSN from Shimadzu, Columbia, MD); methods
4500-P D and 4500-NH3 F were used to measure phosphorus and ammonium by UV-visual
spectrophotometry (Evolution 600 from Thermo Scientific, Madison, WI); method 4110 B was
followed for determining NO3 and PO4 by an ion chromatography (model MIC II from Metrohm,
Switzerland).
The geotextile membrane was cleaned when needed and solids were collected and
weighed by an industrial scale (Romanas Oconi S.A., Costa Rica). Moisture content of sediments
collected by the geotextile membrane followed the previously described Hach method #8276.
Carbon, nitrogen, and phosphorus concentrations of the sediments were analyzed at Agronomy
Research Center (CIA) at the University of Costa Rica. Nitrogen and carbon concentrations were
measured with Method SC09-LSF-P06 (Dunas method) using an autoanalyzer (Vario Cube from
Elementar, Philadelphia, PA), while phosphorus concentrations were measured with method

105

SC09-LSF-P10 using an ICP-plasma atomic emission spectroscopy (Optima 8300 ICP-OES
Spectrometer from Perkin Elmer, Spain).
The JMPÂŽ, Version 10.0.0 (SAS Institute Inc., Cary, NC) was used for statistical
analyses. The Mann-Whitney test for two independent samples was conducted for determining
significant differences between effluent concentrations in the HCTW for both low and high
precipitation periods.

4.2.4. Modeling approach
The median weekly precipitation during the study period was 1.91 cm/week. Weeks with
precipitation from 0.00 to 1.91 cm/week, occurring from January to March 2016, were
considered to belong to the low precipitation period, while weeks with precipitations higher than
1.91 cm/week, occurring from August to December 2015, were considered to belong to the high
precipitation period. More mathematically robust parameters that described the treatment
performance of the VSSF-CTW and the FWS-CTW, during both low and high precipitation
periods, were selected using PCA. Parameters included in the PCA were loading (g/m2/week)
and output concentrations (mg/L) of each water quality parameter analyzed, as well as, the
influent flow rate (Qi), precipitation (P(A)), and runoff (Qc). Parameters that showed the high
eigenvalues of principal components were identified as the key parameters; the cumulative
percentage of these parameters accounted for close to 70% of the total variance. Once the key
parameters were identified, graphical representations (loading charts) of those parameters were
generated for each wetland and each precipitation period. Linear regression (R2) was conducted
for each graphical representation. Additionally, attempts to fit the data with contaminant removal
models were made. Contaminant removal models were run on a spreadsheet and parameters were

106

fitted using solver (Excel Microsoft) to minimize the sum of squares between actual and
predicted outlet concentrations.

4.3. Results and discussion
4.3.1. Hydrological balance
Inputs (Qi, Qc and P(A)) and outputs (Qo and ET(A)) of the hydrological balance for both
the VSSF-CTW and FWS-CTW are shown in Table 4.2. The change in volume of water stored
in each cell (Ds), calculated using the equation 4.2, and the average volume of water stored in
each cell (SV) are shown in Table 4.2. Runoff from the surrounding drainage area entered the
cells during October, November, and December 2015. ET(A) was higher during January to
March than the ET(A) from September to December, due to high temperatures and dry conditions
that are typical for January to April in Alajuela, Costa Rica. Consequently, there were periods of
zero transfer of water from the VSSF-CTW to the FWS-CTW and from the FWS-CTW to the
irrigations system during January to March 2016. Instead, effluent water was stored within the
wetlands to support plant growth and was only discharged to provide 5 m3 per week to the
digester. For example, in the FWS-CTW, the hydrological balance was negative and
consequently the stored volume of water substantially decreased from a daily average of 40 m3 in
January 2016 to 4.41 m3 in March 2016.

107

Table 4. 2. Hydrological balance at the HCTW.
Weather
Month
Sept2015
Oct2015
Nov2015
Dec2015
Jan2016
Feb2016
Mar2016

P(A)
1.11
Âą0.28
2.34
Âą0.85
1.97
Âą0.74
0.618
Âą0.663
0.004
Âą0.005
0.003
Âą0.002
0.000

ET(A)
0.142
Âą0.008
0.110
Âą0.099
0.135
Âą0.033
0.233
Âą0.025
0.285
Âą0.018
0.293
Âą0.044
0.319
Âą0.009

Qc
-6.76Âą
3.62
6.95Âą
3.33
1.83Âą
1.62
----

Qi
0.623Âą
0.058
0.568Âą
0.045
0.770Âą
0.001
0.674Âą
0.091
0.639Âą
0.075
0.869Âą
0.060
0.856Âą
0.001

VSSF-CTW
(m3/d)
Qo
Ds
0.244Âą
1.35Âą 0.44
0.091
5.64Âą 4.46

3.91Âą 9.07

8.85Âą 3.97

0.705Âą
8.074

0.417Âą
0.157
0.384Âą
0.152
0.120Âą
0.044
0.0391Âą
0.0001

2.47Âą 2.55
-0.0256Âą
0.2500
0.458Âą
0.150
0.498Âą
0.010

FWS-CTW
SV
20.6Âą
0.2
37.4Âą
2.5
46.7Âą
2.1
21.6Âą
1.6
6.15Âą
0.32
5.77Âą
0.13
6.25Âą
0.46

Qi
0.244Âą
0.091
5.64Âą 4.46

Qo
0.259Âą
0.095
7.91Âą
5.95

8.85Âą 3.97

12.3Âą 5.3

0.417Âą
0.157
0.384Âą
0.152
0.120Âą
0.044
0.0391Âą
0.0001

0.823Âą
0.321
0.886Âą
0.185
0.206Âą
0.170
0.384Âą
0.237

Ds
0.953Âą
0.474
6.72Âą
14.90
5.33Âą
13.40
1.81Âą 2.78
-0.783Âą
0.360
-0.376Âą
0.265
-0.664Âą
0.241

SV
87.2Âą
0.9
115Âą 6
119Âą 4
94.0Âą
1.4
40.3Âą
4.0
13.4Âą
0.6
4.41Âą
4.94

a. Weather columns apply to both VSSF-CTW and FWS-CTW.
b. For each CTW, Qi: influent flow, Qo: effluent flow, Ds: change in volume of water stored in the CTW (a negative sign (-)
indicates that the inputs are lower than the outputs), and SV: stored volume of water in the wetland (directly measured by
assessing the height of water in each cell).

108

4.3.1.1. Hydrological balance on the VSSF-CTW
The VSSF-CTW has a water storage capacity in the media of 30.9 m3, with an additional
80 m3 of storage above the media. The wetland was flooded only during October and November
2015, when 6.5 m3 and 15.8 m3 of water per day was stored above the media, respectively
(Figure 4.4). Under-sizing of diversion structures around the HCTW allowed runoff to enter the
VSSF-CTW; however, the VSSF-CTW had sufficient storage capacity to hold both runoff and
precipitation within its berms. During October and November, transfer of water from the VSSFCTW to the FWS-CTW was increased to avoid spillover of untreated wastewater from the
VSSF-CTW. The additional input of water due to precipitation and runoff diluted the applied
wastewater (Figure 4.4.a). Negatively, runoff introduced uncertainty to the experimental results,
as concentrations of pollutants in the runoff were not measured.

Figure 4. 4. VSSF-CTW during the medium to high precipitation phase. a. A dilution effect
could occur due to inputs such as precipitation and runoff. b. Berms had the capacity to hold
water over the filter media and avoid spillover. Pictures were taken in October 2015.

Little precipitation occurred from January to March 2016. During this dry period,
operation of the VSSF-CTW was modified to maintain saturated, anoxic conditions in the bottom
109

of the cell by gradually decreasing the effluent flow from 1 m3/d to 0.384 – 0.391 m3/d. On
average, the VSSF-CTW was 30% full, with saturated conditions in the bottom 0.20 m and
unsaturated conditions in the top 0.90 m. HLRs (only considering wastewater) averaged 0.00474
Âą 0.00032 m/d during the dry period and 0.00568 Âą 0.00053 m/d during the rainy period. These
values were substantially lower than published values which range from 0.0200 to 0.200 m/d
(Kadlec and Wallace 2009). Wastewater hydraulic loading was low for two reasons. First,
loading rates were determined from pollutant loading rates since influent concentrations of COD
(>6,000 mg/L) and TS (>4,000) were very high. Pollutant loading rates varied from 24.7 – 37.2 g
COD/m2/d and from 16.5 – 24.8 g TS/m2/d. These loads were higher than recommended loads
for COD (< 20 g/m2d) and TS (<5 g /m2d) (Winter and Goetz 2003). Second, the Qi applied to
the VSSF-CTW was almost 50% less than the designed 1 m3/d during the high precipitation
period, whereas during the low precipitation period, the Qi was 15% lesser than the designed
when days that had no discharge were included in the average hydraulic loading rate. Discharge
of liquid digestate in the wetland was conducted only from Monday to Friday and on some
weekdays the liquid digestate was not discharged due to flooding in the wetland, geotextile
cleaning, holidays, and days when equipment was broken in the system. Inclusion of
precipitation and runoff in the hydraulic loading rate substantially increased the HLR; actual
HLR during October and November 2015 was 0.0700 m/d, which was between the suggested
ranges. HRTs, calculated as the stored volume divided by the total flow rate in (P(A) + Qc + Qi),
ranged from 3.87 to 11.9 d during the rainy period and 6.69 to 9.55 d during the dry period.
These HRTs were similar to the ones calculated by Avila, Salas et al. (2013) in a VSSF-CTW
during a rainy period (3.5 d) and dry period (6.88 d) in Spain.

110

4.3.1.2. Hydrological balance on the FWS-CTW
The FWS-CTW has a storage capacity of 152 m3. During the high precipitation period,
additional water was discharged from the FWS-CTW to avoid overtopping the VSSF-CTW cell
after large runoff events. On average, 7.91 and 12.3 m3/d were discharged during October and
November 2015, respectively, as compared to an average of 0.607 m3/d during the dry period. As
previously mentioned, runoff could impact modeling as the concentrations of pollutants in the
runoff were unknown and the volume of runoff was approximated. During the dry period, the
FWS-CTW had sufficient capacity to store water from the VSSF-CTW; the stored value
decreased with time due to higher Qo than Qi, as well as higher ET(A) (Table 4.2). The Qo was
higher that the Qi because the effluent from the FWS-CTW was used to feed the digester at a rate
of 5 m3/week. This rate was higher than the Qo from the VSSF-CTW in order to keep water at
the bottom of the VSSF-CTW.
The FWS-CTW had an HLR of 7.24 × 10-3 m/d, which is within the recommended range
of 7.00 × 10-3 to 5.00 × 10-2 m/d (Kadlec and Wallace 2009). The HRT of the FWS-CTW varied
substantially from 6.72 to 64.4 d during the period of high precipitation and from 103 to 112 d
during the period of low precipitation. Lower HRTs during the high precipitation period were
due to high flow rates (measured as P(A) + Qc + Qi), especially during October and November
2015. Based on HLR suggested by Kadlec and Wallace (2009), recommended HRTs for FWSCTWs range from 22 to 157 d (HLR = 7.00 × 10-3 m/d). The HRT of our FWS-CTW was within
this range during September (64.4 d), December (32.8 d), and from January to March 2016
(>100 d). HRT was shorter than desired for October (7.84 d) and November (6.72 d).

111

4.3.2. Treatment performance
Table 4.3 summarizes the treatment performance of the VSSF-CTW and FWS-CTW. The
effluent from the system is compared to the Costa Rican water quality standards (COD < 150
mg/L, TS < 50 mg/L, TN < 50 mg/L, TP < 8 mg/L, 5 < pH < 9, and 15 °C < T < 40 °C) for
discharging treated effluents into bodies of water or onto crop fields (MINAE-MSP 2007).

112

Table 4. 3. Characteristics of studied parameters at the effluent tank, the VSSF-CTW, and the
FWS-CTW for both high (n =16) and low (n = 9) precipitation periods.

High precipitation period

Parameter

VSSF-CTW

FWS-CTW
Mean Âą S.E.

66.5 Âą 12.8

%
Removal
99.0 Âą 0.3

4,399 Âą 383

543 Âą 63.5

VS (mg/L)

1,462 Âą 210

FS (mg/L)

HCTW

34.3 Âą 2.05

%
Removal
48.4 Âą 0.1

%
Removal
99.5 Âą 0.1

87.6 Âą 0.2

243 Âą 23.7

55.1 Âą 0.2

94.4 Âą 0.2

171 Âą 35.4

88.3 Âą 0.3

92.5 Âą 19.2

45.9 Âą 0.2

93.7 Âą 0.4

2,938 Âą 267

372 Âą 47.5

87.3 Âą 0.2

151 Âą 13.7

59.4 Âą 0.1

NH4 (mg/L)

1,312 Âą 119

13.9 Âą 3.38

1.20 Âą 0.332

91.4 Âą 0.5

NO3 (mg/L)

3.64 Âą 0.550

71.4 Âą 22.8

1.23 Âą 0.320

98.3 Âą 0.6

66.2 Âą 0.3

TN (mg/L)

1,078 Âą 101

34.0 Âą 8.55

98.9 Âą 0.3
(1800 Âą
8)
96.8 Âą 0.3

94.8 Âą 0.2
99.9 Âą 0.4

2.03 Âą 0.458

94.0 Âą 0.4

99.8 Âą 0.3

PO4 (mg/L)

80.0 Âą 48.6

1.36 Âą 0.181

98.3 Âą 0.7

0.600 Âą n.d.

99.2 Âą 0.6

TP (mg/L)

106 Âą 16.3

0.801 Âą 0.131

99.2 Âą 0.3

0.913 Âą 0.145

55.9 Âą 0.1
(14.0 Âą
0.1)

pH

7.95 Âą
0.0434

6.93 Âą 0.102

--

6.63 Âą 0.119

--

--

26.4 Âą 0.530

25.7 Âą 0.550

--

25.4 Âą 0.452

--

--

8,095 Âą 914

564 Âą 120

93.0 Âą 0.3

288 Âą 57.9

48.8 Âą 0.2

96.4 Âą 0.3

Mean Âą S.E.

Mean Âą S.E.

COD (mg/L)

6,841 Âą 681

TS (mg/L)

Temperature
(°C)
COD (mg/L)

Low precipitation period

Effluent
Tank

99.1 Âą 0.3

TS (mg/L)

6,630 Âą 853

2,926 Âą 272

55.8 Âą 0.1

892 Âą 125

69.5 Âą 0.2

86.5 Âą 0.2

VS (mg/L)

2,439 Âą 589

1,087 Âą 155

55.4 Âą 0.2

288 Âą 53.1

73.5 Âą 0.2

88.2 Âą 0.4

FS (mg/L)

4,191 Âą 340

1,839 Âą 172

56.1 Âą 0.1

604 Âą 87.4

67.1 Âą 0.2

85.6 Âą 0.2

NH4 (mg/L)

1,714 Âą 99

78.0 Âą 30.9

15.4 Âą 6.79

80.3 Âą 0.7

NO3 (mg/L)

12.7 Âą 6.23

377 Âą 121

22.4 Âą 10.7

94.1 Âą 0.7

TN (mg/L)

1,443 Âą 108

246 Âą 28.2

95.4 Âą 0.4
(2,800 Âą
9)
82.9 Âą 0.2

15.7 Âą 6.79

93.6 Âą 0.2

99.1 Âą 0.5
(76.1 Âą
0.4)
98.9 Âą 0.3

PO4 (mg/L)

8.50 Âą 2.6

3.74 Âą 0.876

56.0 Âą 0.3

1.92 Âą 0.268

48.7 Âą 0.2

77.4 Âą 0.4

TP (mg/L)

156 Âą 34

1.88 Âą 0.756

98.8 Âą 0.6

0.632 Âą 0.0845

66.4 Âą 0.4

99.6 Âą 0.3

pH
Temperature
(°C)

8.35 Âą 0.031

7.59 Âą 0.0936

--

7.63 Âą 0.0551

--

--

26.8 Âą 0.29

26.8 Âą 0.242

--

26.6 Âą 0.224

--

--

*. Mann-Whitney test for two independent samples comparing the HCTW effluent during
both low and high precipitation periods (alpha = 0.05). Removal rates were significantly
different between the high and low precipitations periods (p<0.05, with the exception of total
phosphorus (p = 0.208).

113

During the high precipitation period, removal for all water quality parameters except NO3
(66.2%) were higher than 90% for the HCTW (Table 4.3). After the VSSF-CTW, water quality
parameters such as COD, TN, TP, pH, and temperature satisfied discharge standards for Costa
Rica. Further treatment at the FWS-CTW reduced the concentration of these parameters. Despite
the substantial reduction of TS, 87.6% after the VSSF-CTW and 94.4% after the HCTW, the
effluent did not meet water discharge standards. Despite this, the effluent from the HCTW was
considered reclaimed for four reasons. For irrigation, Avila, Salas et al. (2013) indicated that TS
of mineral origin cause a minimal impact if is discharged onto land or into aquatic ecosystems.
In our case, the effluent from the HCTW had low COD concentration and low VS to TS ratio
(38.0 Âą 5.1%), and the sediments in the effluent were considered of mineral origin. Secondly,
regulations in Costa Rica allows a discharge of wastewater from livestock activities with
concentrations of 200 mg TS/L (MINAE-MSP 2007), and this concentration was met by the
effluent of the HCTW if considered only the FS (151 Âą 13.7 mg/L). Third, land application of
liquid manures is a common practice in Costa Rica. Liquid manures are defined to have less than
4% TS, with typical measured concentrations up to 23.9 g TS/L (Lorimor, Powers et al. 2004).
Thus, TS concentrations were not considered a barrier to use of the effluent for irrigation, as TS
concentration in the liquid manures is 44 times larger than the concentration from the VSSFCTW effluent in this study. Additionally, the sequence of recirculation and discharge likely
increased TS in the effluent of the FWS-CTW; water within the FWS-CTW was recirculated
from 2 – 4 am and effluent samples were collected prior to 8 am. It is expected that modifying
this schedule would substantially decrease the TS in the effluent of FWS-CTW; however, this
action was not immediately taken during this study so that effluent concentrations during the

114

rainy and dry periods could be directly compared. Finally, the effluent from the HCTW can be
reused for dilution of feed for the digester.
During the low precipitation period, total removal for the HCTW varied from 77.4 to
99.6% (Table 4.3). NO3 effluent concentrations from the HCTW increased with respect to initial
concentrations in the effluent tank. After the VSSF-CTW, only TP, pH, and temperature satisfied
discharge standards for Costa Rica, whereas the FWS-CTW further removed TN to a permitted
concentration. In contrast, COD and TS were above the maximum discharge standard after both
the VSSF-CTW and the FWS-CTW. During this period, the water from the FWS-CTW was used
only for dilution of feed for the digester.
Thus, there was a difference in the treatment performance during the two precipitation
periods. During this study, lower concentrations were obtained during the high precipitation
period (Table 4.3). The Mann-Whitney test for two independent samples indicated that there
were significant differences (p < 0.05) in the HCTW effluent concentrations for each water
quality parameter, except TP (p = 0.208), between the low and high precipitation periods (Table
4.3). Dilution, due to precipitation, and concentration, due to evapotranspiration, likely
contributed to differences in HCTW effluent concentrations between the low and precipitation
periods. Avila, Salas et al. (2013) indicated that inlet concentrations for a VSSF-CTW were
lower during a wet period (59.0 mg/L) than during a dry period (294 mg/L), due to the dilution
effect caused by precipitation. Consequently, outlet COD concentrations were lower during a wet
period (29 mg/L) than during a dry period (46 mg/L) (Avila, Salas et al. 2013). Additionally, in
this study, it was likely that concentration of pollutants occurred from January to March 2016.
Low HLR, scarce precipitation events, and high temperatures substantially decreased the water
levels in the HCTW, thus increasing the concentration of pollutants. For example, the water at

115

the FWS-CTW was green (Figure 4.5), likely due to algae bloom promoted by high temperatures
and availability of nutrients (e.g., nitrate and phosphate). Availability of nutrients occurred due
to resuspension of sediments by the movement of the water during recirculation of water in the
FWS-CTW and during sampling and plant harvest. It was likely that these conditions caused
high concentrations during the low precipitation period in the HCTW.

Figure 4. 5. FWS-CTW during the low precipitation period. a. Low HLR, scarce precipitation
events, and high temperatures substantially decreased the water at the FWS-CTW. b. Growth of
algae bloom. Pictures were taken in February 2016.

Use of multiple types of CTWs in-series aims to increase removal by balancing the
strengths and weakness of each type of CTW and is a common approach to achieve high
removals of pollutants (Tunçsiper 2009, Avila, Salas et al. 2013, Comino, Riggio et al. 2013,
Vymazal and KrĂśpfelovĂĄ 2015, Avila, GarcĂ­a et al. 2016). In this case, a VSSF-CTW was
combined in series with a FWS-CTW. The VSSF-CTW was placed first because the influent had
high concentrations of NH4 (>1,000 mg/L), COD (>6,000 mg/L), and TS (>4,000 mg/L). The

116

aerobic conditions that are predominant in VSSF-CTW facilitate nitrification (Vymazal and
KrĂśpfelovĂĄ 2015). Then, the FWS-CTW, which is predominantly anaerobic, supports
denitrification (Vymazal 2007, Kadlec and Wallace 2009). In our VSSF-CTW, concentrations of
NH4 decreased by more than 95.0% while NO3 concentrations increased substantially. NO3
accounted for less than the 0.120% of the TN in the liquid digestate and 34.6 to 47.6% of the TN
in the effluent from the VSSF-CTW. This increase in NO3 concentrations indicated that
nitrification was a major process responsible for ammonium removal in the VSSF-CTW.
Additionally, >80% of TN in the liquid digestate was removed by the VSSF-CTW and removal
of ammonium was much greater than the net production of nitrate. In contrast, nitrogen removal
in a sand-based VSSF-CTW planted with T. angustifolia L. and Cyperus alternifolius L in
Thailand was limited due to a lack of anoxic conditions and subsequent denitrification
(Kantawanichkul, Kladprasert et al. 2009, Avila, Salas et al. 2013). As our VSSF-CTW was
always at least 30% full, with saturated conditions in the bottom 0.2 m and unsaturated
conditions in the top 0.90 m, it is likely that nitrification did occur. Additional processes that
could have contributed to nitrogen removal, aside from denitrification and nitrification, include
microbial and plant assimilation, anaerobic ammonium oxidation (anammox), and volatilization
of ammonium (Kantawanichkul, Kladprasert et al. 2009, Avila, Salas et al. 2013). However,
neither volatilization nor anammox were expected to occur at high rates as the water pH and
temperature were 6.93 and 25.7°C, respectively, which are not conducive to either process (Lee,
Fletcher et al. 2009, He, Tao et al. 2012). Furthermore, plant uptake is typically assumed to
represent no more than 10% of removal (Vymazal 2007, Vymazal 2011), Additionally, some
removal of TN might occur from filtration of solids by the geotextile placed at the inlet area of
the VSSF-CTW, which is analyzed in the next section.

117

NO3 concentrations decreased by more than 90% in the FWS-CTW (Table 4.3),
indicating that denitrification was likely a major nitrogen removal process in the FWS-CTW
(Kantawanichkul, Kladprasert et al. 2009, Avila, Salas et al. 2013). Plant uptake also likely
contributed to nitrogen removal. The FWS-CTW was important for removal of TN. Table 4.3
shows that percentage removal for almost all of the water quality parameters was low, from 45 to
75%, but higher for NH4, NO3, and TN (higher than 80%), during both low and high
precipitation periods.
Despite the high OLR reported in this study (24.7 – 37.2 g COD/m2/d and 16.5 – 24.8 g
TS/m2/d), the HCTW performed similarly to other wetlands. A VSSF-CTW receiving an organic
loading of 19.6 g BOD/m2/d achieved removal of 92% for BOD, 89% for COD, 95% for TS,
77% for TN, and 62% for TP (Tsihrintzis, Akratos et al. 2007). In our study, the FWS-CTW
served as a polishing step and achieved lower treatment performance than Katsenovich,
Hummel-Batista et al. (2009) and Avila, Salas et al. (2013), who demonstrated that adequate
treatment by FWS-CTW can occur for higher loads. For example, in El Salvador, Katsenovich,
Hummel-Batista et al. (2009) evaluated a FWS-CTW recieving municipal wastewater. Despite
the higher HLR (0.206 m/d), a FWS-CTW planted with Thypa angustifola removed 81% for
BOD (20 mg/L), 65% for COD (73 mg/L), 59% for TN (6.1 mg/L), and 67% for TP (1.9 mg/L)
(Katsenovich, Hummel-Batista et al. 2009). Better performance was obtained by Avila, Salas et
al. (2013); a FWS-CTW receiving municipal wastewater removed 98% of BOD (7.0 mg/L), 91%
of COD (50 mg/L), 98% of TS (6.0 mg/L), 85% of TN (7.9 mg/L), and 34% for TP (5.3 mg/L)
(Avila, Salas et al. 2013). Note that comparisons between systems should be taken with caution
due to different configurations of the cells.

118

4.3.2.1. The geotextile membrane at the VSSF-CTW
The geotextile membrane (2 m x 2 m) was placed in a shallow trench of approximately
0.25 m depth, with the purpose to remove suspended solids still present in the liquid digestate
(Figure 4.6). This area has sufficient volume to hold the one cubic meter discharged each day.
Ponded in the geotextile membrane area, the liquid digestate slowly passed through the
membrane and infiltrated only into the filter media beneath the membrane (4 m2 of the total
median treatment area of 138 m2). The geotextile was cleaned three times during this study
(Table 4.4). The first batch was collected before heavy rains started. The dry conditions during
the day allowed the sediments to lose moisture and be easily collected. In contrast, in October
and November 2015, it was not possible to collect sediments as the wetland was flooded (Figure
4.4). The last two batches were collected after rain events from October to November 2015
(second batch) and after intermittent rains during January 2016 (third batch). Even though the
second batch was after 9 weeks of operation, during which rains and flooding occurred in the
VSSF-CTW, more sediments were collected after 5 weeks of operation in the third batch. Due to
flooding, it was likely that a portion of the sediments, during the second batch, were dispersed
and settled in in the main treatment area, rather than being isolated to the geotextile area (Figure
4.4). Additionally, solids had accumulated in the effluent tank to the height of the outlet pipe by
January 2016, increasing the discharge of solids into the VSSF-CTW. For these two reasons,
there was a substantial difference in the mass of dry sediments collected during the first and third
collection. Consequently, the effluent tank was cleaned out in January 2016.

119

Table 4. 4. Sediments collected from the geotextile membrane at the VSSF-CTW.
Period

Number of weeks

08/31/2015 to 10/06/2015

5

10/06/2015 to 12/07/2015

9

12/07/2015 to 01/12/2016

5

Collected sediments

Moisture content

Collected sediments

(wet, kg)

(%)

(dry, kg /m2/week)

19.2 Âą 0.5

57.3 Âą 0.5

0.41 Âą 0.01

399 Âą 1

90.66 Âą 0.4

1.04 Âą 0.01

855 Âą 1

90.66 Âą 0.6

3.99 Âą 0.03

Figure 4. 6. Geotextile membrane placed in the VSSF-CTW inlet area. a. Liquid digestate was
discharged by a PVC pipe (diameter = 0.05 m) into the geotextile membrane. b. Solids were
accumulated on the geotextile while the liquid passed through the membrane and infiltrated in
the filter media of the VSSF-CTW. c. A worker cleaning up the geotextile membrane. Pictures
were taken in October 2014.

In terms of mass removal, the geotextile contributed to performance of the VSSF-CTW.
Table 4.5 shows the mass and removal loading rates for the geotextile and the overall VSSFCTW for COD (used as a proxy for carbon), TS, TN, and TP. In general terms, the geotextile
membrane accounted for a small portion of the total removal in the VSSF-CTW. Average mass
removal by the geotextile membrane was 4.81% for COD, 27.9% for TS, 13.4% for TN, and
19.5% for TP. The geotextile membrane prevented direct contact of these sediments with the
substrate media (coarse sand), likely reducing clogging, as formation of a crust was not observed
in the treatment area of the VSSF-CTW. Additionally, accumulation of sediments in the
geotextile membrane facilitated the collection of this material for further treatment (composting).
120

In this way, the geotextile membrane decreased the mass load to be treated by the VSSF-CTW,
reduced potential clogging, and allowed recovery of nutrients. Many authors and CTW
guidelines suggest a pretreatment system before discharging wastewater into the wetland
(Garcia-Perez, Harrison et al. 2011, Comino, Riggio et al. 2013, de la Varga, DĂ­az et al. 2013);
however, the implementation of a geotextile membrane in the inlet area provided a low-cost,
low-maintenance approach to improving the performance of the VSSF-CTW.

121

Table 4. 5. Mass removal from the geotextile membrane and from the VSSF-CTW.
Perioda

Mass loading into
the geotextileb

Mass removal by
the geotextilec

TP

TN

TS

COD

kg /m2/week

Removal
by the
geotextile
(%)

Mass loading
after the
geotextile

Mass after the
VSSF

Mass removal by
the VSSF-CTW

kg /m2/week

Removal
by the
VSSF
(%)

1

4.32 Âą 00502

0.109

2.52

4.21 Âą 0.502

(6.99 Âą 3.17) x10-4

4.21 Âą 0.502

99.9 Âą 0.1

2

10.1 Âą 1.39

0.274

2.71

9.81 Âą 1.39

0.0113 Âą 0.00423

9.80 Âą 1.39

99.9 Âą 0.1

3

11.2 Âą 2.22

1.06

9.46

10.1 Âą 2.22

0.110 Âą 0.0775

9.93 Âą 2.16

97.9 Âą 0.7

Avg.

8.92 Âą 1.08

0.429 Âą 0.0871

4.81 Âą 0.01

8.49 Âą 1.05

0.0558 Âą 0.0268

8.44 Âą 1.04

99.4 Âą 0.3

1

3.66 Âą 0.488

0.410

11.2

3.26 Âą 0.489

(7.80 Âą 2.96) x10-3

3.24 Âą 0.486

99.8 Âą 0.1

5.38 Âą 0.772

(9.74 Âą 3.60) x10

-2

5.28 Âą 0.760

98.2 Âą 0.6

-2

2

6.41 Âą 0.722

1.03

16.1

3

6.58 Âą 2.30

3.99

60.6

2.59 Âą 2.31

(4.43 Âą 2.09) x10

2.55 Âą 2.29

99.9 Âą 0.5

Avg.

5.77 Âą 0.734

1.61 Âą 0.328

27.9 Âą 0.1

4.15 Âą 0.737

(6.17 Âą 2.06) x10-2

4.09 Âą 0.727

99.0 Âą 0.4

1

0.706 Âą 0.0906

0.0122

1.73

0.693 Âą 0.0906

(3.61 Âą 1.20) x10-4

2
3

1.53 Âą 0.199
1.81 Âą 0.156

0.0307
0.118

2.00
6.52

1.50 Âą 0.199
1.69 Âą 0.156

0.693 Âą 0.0905

99.9 Âą 0.1

(2.41 Âą 0.61) x10

-3

1.49 Âą 0.198

99.8 Âą 0.1

(4.29 Âą 1.88) x10

-3

1.69 Âą 0.156

99.7 Âą 0.1

-3

1.20 Âą 0.136

99.8 Âą 0.1

Avg.

1.39 Âą 0.145

0.186 Âą 0.0298

13.4 Âą 0.1

1.20 Âą 0.136

(2.37 Âą 0.63) x10

1

0.0567 Âą 0.00924

0.00734

12.9

0.0493 Âą 0.00924

(6.20 Âą 2.88) x10-6

0.0493 Âą 0.00924

99.9 Âą 0.1

2

0.165 Âą 0.0321

0.0185

11.2

0.147 Âą 0.0321

(3.32 Âą 1.81) x10-4

0.146 Âą 0.0319

99.7 Âą 0.1

0.133 Âą 0.0729

-5

0.133 Âą 0.0729

99.9 Âą 0.0

0.119 Âą 0.0254

99.8 Âą 0.1

3
Avg.

0.205 Âą 0.0729
0.148 Âą 0.0269

0.0714
0.0289 Âą 0.00587

34.8
19.5 Âą 0.1

0.119 Âą 0.0254

(1.93 Âą 1.12) x10

(1.72 Âą 0.978) x10

-4

a. Period 1 from 08/31/2015 to 10/06/2015, period 2 from 10/06/2015 to 12/07/2015, and period 3 from 12/07/2015 to
01/12/2016.
b. Formula: Qi*Cin /geotextile area/number of weeks.
c. Sediments collected times the amount of carbo (26.5 Âą 2.66%), nitrogen (2.97 Âą 0.307%), or phosphorus (1.79 Âą 0.275%).
Abbreviation: average (Avg).

122

4.3.3. Modeling approach
PCA was conducted to statistically identify key parameters that described the treatment
performance of the VSSF-CTW and the FWS-CTW, during both low and high precipitation
periods. In general, parameters such as influent flow rate (Qi), precipitation (P(A)), and runoff
(Qc) had low eigenvalues, and these parameters were not directly considered for modeling. Based
on the eigenvalues extracted from the loading matrix principal components, the more
mathematically robust parameters that described the treatment performance were the water
quality parameters (Table 4.6). During the low precipitation period, the treatment performance at
the VSSF-CTW was explained by TS load, COD out, and TN load, with a cumulative percentage
of 76.3%. During the high precipitation period, COD load and COD out explained the treatment
performance with a cumulative percentage of 67.2%. For the FWS-CTW, the treatment
performance during the low precipitation period was explained by FS load, TS out, and TP load
with a cumulative percentage of 74.1%; whereas, TS load, TS out, and NH4 out explained the
treatment performance during the medium high precipitation period with a cumulative
percentage of 77.6%. Loading charts and contaminant removal models were evaluated for their
ability to describe the treatment performance of selected water quality parameters.

Table 4. 6. Key water quality parameters for the HCTW during both high and low precipitation
periods.
VSSF-CTW
Period
High
precipitation
period (n=16)
Low
precipitation
period (n=9)

PC

Parameter

Eigenvalues

1
2
3
1
2
3

COD load
COD out
-TS load
COD out
TN load

0.975
0.800
-0.913
0.849
0.733

FWS-CTW
Cumulative
percentage
(%)
37.8
29.4
-37.8
24.6
13.9

123

PC

Parameter

Eigenvalues

1
2
3
1
2
3

TS load
TS out
NH4 out
FS load
TS out
TP load

0.994
0.936
0.689
0.972
0.934
0.773

Cumulative
percentage
(%)
42.2
21.8
13.6
34.0
28.2
11.9

The measured output concentrations (mg/L) were not correlated to input loading rates
(g/m2/week)) for the key water quality parameters, with R2 values ranging from 0.0115 to
0.381(Figures 4.7 and 4.8). These poor relationships are normal for full-scale systems (Knight,
Payne Jr et al. 2000). The internal variability in the wetlands (plant communities, internal
hydraulics, and weather conditions) likely impacted treatment performance, and more time
evaluating the wetlands could narrow the central tendency of the treatment performance for the
HCTW in this study. Variability in treatment performance is expected, especially when large
data sets are used for representing loading charts. Knight, Payne Jr et al. (2000) analyzed the
Livestock Wastewater Treatment Database (LWDB) for North America. LWDB gathered
information from 68 sites, 278 individual treatment cells, and more than 1,390 indivdual records
with data for multiple parameters. Based on this data, Knight, Payne Jr et al. (2000)
recommended against the use of loading charts (output concentrations vs. input loading rates)
due to the poor fit of data (low R2 values) for BOD (R2=0.74) and TSS (R2=0.30); these loading
charts should be used with caution for NH4 (R2=0.87), TN (R2=0.81), TP (R2=0.70), and COD
(R2=0.89). In addition, Kadlec and Wallace (2009) indicated that loading charts are most usefull
when inlet concentrations are constant; in contrast, inlet concentrations were highly variable in
this study.

124

1
2

A

B

3
4

C

D

Figure 4. 7. Relationship between mass loading and outlet concentration at the VSSF-CTW for
A) COD during the high precipitation period, B) TS during the low precipitation period, C) COD
during the low precipitation period, and D) TN during the low precipitation period.

125

1
2

A

B

3
4

C

D

5
6

E

Figure 4. 8. Relationship between mass loading and outlet concentration at the FWS-CTW for A)
TS during the high precipitation period, B) NH4 during the high precipitation period, C) FS
during the low precipitation period, D) TS during the low precipitation period, and E) TP during
the low precipitation period.

126

Attempts to model the data with contaminant removal models were largely unsuccessful.
Model parameters, such as K20 and θ20 for the plug flow model, K20, θ20, and C* for the modified
k-C* model, P, k, and C* for the TIS (p k-C*) model, and k0 and b for the TDR model, were
fitted, based on the measured inlet and outlet concentration obtained during this study (Tables
4.7 and 4.8). For the VSSF-CTW, the sum of squares between actual and predicted outlet
concentrations and low R2 indicated that the fitted parameters should not be used for design
purposes. Similar poor fits were found for the FWS-CTW, with the exception of modeling
ammonium removal with the TIS model during the high precipitation period (sum of the squares
= 17.3 and R2 = 0.512). However; the fitted K20 and θ20 values were unreasonable compared to
estimated values by Tanaka and Tanaka (2011). In general, it was expected to obtain fitted kvalues for COD that were higher than literature values. Due to the temperature dependence of
this rate, higher removal rates were expected in this study, as the HCTW is located in the tropics.
The steady warmer climate across the year in the tropics is expected to yield constant removal
rate, which allows the reduction in the required area for wetland treatment at the tropics (Zurita,
De Anda et al. 2009). However, due to the poor fitting obtained in this study, the adjusted values
cannot be considered for designing wetlands.
The background concentration (C*) is considered in the modified k-C* and TIS (P k-C*)
models. Temperature plays a role in the background concentration because microbes’
degradation increases or decrease based on the temperature. In this study, C* was initially
estimated by equations or values indicated in the literature. The COD background concentration
at subsurface CTW was estimated by equation 4.8. Similarly, TS at the VSSF-CTW (C* = 7.8 +
0.063Ci) and at the FWS-CTW (C* = 5.1 + 0.16Ci) were based on Kadlec’s background equation
concentration (Kadlec and Wallace 2009). In these cases, the C* substantially varied due to the

127

unsteady influent concentration for COD and TS (Table 4.7 and 4.8). Attempts to fit the value
for C* for the key parameters in each wetland from the data were unsuccessful. Finally, the TDR
model (Shepherd, Tchobanoglous et al. 2001) replaces C* for two parameters (K0 and b).
Shepherd, Tchobanoglous et al. (2001) estimated k0 values from 9 to 12 d-1 and b values from 2
to 5 d-1 for modeling COD removal in a subsurface flow constructed wetland for winery
wastewater treatment. Attempts to fit k0 and b to the data were also unsuccessful, yielding high
sums of squares and low R2 values between the predicted and measured effluent concentrations
(Table 4.7). Shepherd, Tchobanoglous et al. (2001) clearly indicated that these parameters were
obtained for a pilot-scale system and the validation at full-scale has not been done.

Table 4. 7. Contaminant removal models approach for the VSSF-CTW.

High precipitation period

Water
quality
parameter

COD

Model
parameters
K20, d-1
Éľ20
Area, m2
Depth, m
C*, mg/L
k, m/d
P
k0, d-1
b, d-1
Sum of the
square
R2

VSSF-CTW
Contaminant removal models
Plug-flow
Modified k–C*
TIS (P k–C*)
Adjusted value
Adjusted value
Adjusted value
1.29
10,528,454
-2.91
56,908,317
-141
84,060
77,945
0.750
16,544,712
0.700
-----53,687,091
--3.00
-------

TDR
Adjusted value
--167
14.0
---9.49
1.65

108,6711

1,742,815

1,742,815

59,233

0.0707

0.0551

0.0551

0.0137

128

Table 4. 7. (cont’d)
Water
quality
parameter

Low precipitation period

COD

TS

TN

Model
parameters
K20, d-1
Éľ20
Area, m2
Depth, m
C*, mg/L
k, m/d
P
k0, d-1
b, d-1
Sum of the
square
R2
K20, d-1
Éľ20
Area, m2
Depth, m
C*, mg/L
k, m/d
P
k0, d-1
b, d-1
Sum of the
square
R2
K20, d-1
Éľ20
Area, m2
Depth, m
C*, mg/L
k, m/d
P
k0, d-1
b, d-1
Sum of the
square
R2

VSSF-CTW
Contaminant removal models
Plug-flow
Modified k–C*
TIS (P k–C*)
Adjusted value
Adjusted value
Adjusted value
0.890
0.540
-1.91
1.61
-9.26
37.7
138
0.480
0.420
0.700
-----0.200
--3.00
-------

TDR
Adjusted value
--10,331
1,025
---248
1,601

1,938,972

1,474,965

1,802,721

1,838,687

0.0558
--138
0.700
-8.22
----

0.0205
--138
0.700
-8.22
----

0.268
--138
0.700
-8.22
3.00
---

0.306
--138
0.700
---3.00
10.0

82,420,449

71,452,072

71,452,072

32,695,892

0.0994
--334
0.700
-0.179
----

0.147
--1,271
0.700
247
0.681
----

0.147
--156,163
0.700
190
83.6
0.16
---

0.114
--1,232,341
6,251
---2,951
49,588

232,006

57,210

55,959

67,038

0.132

0.464

0.0223

0.0365

R2 values indicate the correlation between the predicted and measured values of effluent
concentrations. Initial values were selected based on reported typical values in the literature.

129

Table 4. 8. Contaminant removal models approach for the FWS-CTW.
Water
quality
parameter

High precipitation period

TS

NH4

Low precipitation period

TS

TP

Model
parameters
K20, d-1
Éľ20
Area, m2
Depth, m
C*, mg/L
k, m/d
P
k0, d-1
b, d-1
Sum of the
square
R2
K20, d-1
Éľ20
Area, m2
Depth, m
C*, mg/L
k, m/d
P
k0, d-1
b, d-1
Sum of the
square
R2
K20, d-1
Éľ20
Area, m2
Depth, m
C*, mg/L
k, m/d
P
k0, d-1
b, d-1
Sum of the
square
R2
K20, d-1
Éľ20
Area, m2
Depth, m
C*, mg/L
k, m/d
P
k0, d-1
b, d-1
Sum of the
square
R2

FWS-CTW
Contaminant removal models
Plug-flow
Modified k–C*
TIS (P k–C*)
Adjusted value
Adjusted value
Adjusted value
------138
47.7
9.21
1.10
1.10
1.10
---0.190
0.950
0.18
--1.06
------385,112

260,364

247,584

262,653

0.0189
0.600
1.12
381
3.04
------

0.00870
0.540
9.92
362
2.89
1.20
-----

3.25x10
0.00387
14.4
2.45
1.10
1.50
-3.00
---

1.86x10-4
--138
1.10
---3.00
10.0

47.5

26.4

17.3

448

0.00867
--138
1.10
-0.210
----

0.0118
--53.0
1.10
-1.05
----

0.512
--138
1.10
-2.74
3.00
---

0.120
--1.67x1011
1.10
---1.61x10x8
5.94x109

6,198,300

2,584,745

2,843,504

1,636,612

0.0628
--286
1.10
-0.0683
----

0.147
--1,289
1.10
0.560
0.310
----

0.147
--338
1.10
0.56
0.0800
7.72
---

0.00110
--49.2
0.390
---2.45
3.14

2.74

0.800

0.800

2.28

0.0399

0.0245

0.0794

0.119

130

-5

TDR
Adjusted value
--72.0
0.570
---2.53
6.30

Table 4. 8. (cont’d)
R2 values indicate the correlation between the predicted and measured values of effluent
concentrations. Initial values were selected based on reported typical values in the literature.

4.4. Conclusions
The HCTW described herein successfully treated wastewater from the thermophilic
anaerobic digester and was resilient to the changes in precipitation that occurred during this
study. The HCTW 1) had sufficient water storage capacity even during the high precipitation
period and 2) achieved high removal efficiencies that allow the reuse of the treated water.
However, due to large variability in the data, loading charts and contaminant removal models
were not able to describe a narrow central tendency in the treatment performance of the HCTW.
Importantly, these data are a base for accumulating knowledge and expertise in the application of
CTWs in Costa Rica, in terms of water storage capacity, treatment performance, and estimation
of design parameters.
The HCTW had better performance during the rainy period, likely due to dilution from
precipitation. In addition, high evapotranspiration during the dry period likely increased effluent
concentration from the wetlands. Recently, Costa Rica has been experiencing unusual increases
in rainfall intensity. The amount of water precipitated per year is similar, but that amount of
water is falling during shorter periods (September to October versus July to December)
according to meteorological records. This aspect is important to consider in the hydrological
balance and treatment performance for future CTWs design.
The hybrid wetland system was important for total treatment performance. The VSSFCTW received high strength liquid digestate and demonstrated high treatment efficiency. The
geotextile membrane provided a low-cost, low-maintenance approach to improving the

131

performance of the VSSF-CTW, as the membrane prevented direct contact of these sediments,
facilitated the collection of sediment, and reduced potential clogging. Then, the FWS-CTW
served as a polishing step for the effluent from the VSSF-CTW. The HCTW-treated water was
considered reclaimed as the use of this water (e.g., irrigation or reuse at the SPAD) did not
strictly have to meet discharge standards for water bodies in Costa Rica.
Natural systems such as the HCTW evaluated here need longer periods of evaluation
(years) for establishing central tendencies in the treatment. The limited time (28 weeks from
August 2015 to March 2016), unsteady weather conditions, and dynamic influent concentrations
likely restricted performance analysis and parameter estimation. Kadlec (2000) emphasized firstorder models are useful for designing, but inadequacies in estimating or selecting parameter
values are highly influenced by the variability of unpredictable events such as fluctuation in
input flows and concentration, changes in internal storages, weather, animal activity, and other
ecosystem factors.

4.5. Acknowledgments
This research was supported by the U.S. Department of State West Hemisphere Affairs
(S-LMAQM-11-GR-075). The authors would like to acknowledge the assistance and support
Water Quality Laboratory at the Research Center of Environmental Pollution (CICA-LCA) of
the University of Costa Rica. The authors would also like to thank the staff at the Fabio Baudrit
Experimental Station (EEAFBM), the project “Electric energy production through anaerobic
digestion optimization of agricultural organic wastes and treatment of its effluents”, the
Department of Eco-physiology at EEAFBM and the Meteorological Institute of Costa Rica.

132

CHAPTER 5: CLOGGING IN TROPICAL VERTICAL SUBSURFACE FLOW
CONSTRUCTED TREATMENT WETLANDS

Abstract: Clogging of media in vertical subsurface flow constructed treatment wetlands
(VSSF-CTWs) can compromise both treatment performance and the wetland’s useable lifespan.
The goal of this study was to evaluate the longevity of sand media in a tropical VSSF-CTW.
Wastewater was pretreated through anaerobic digestion and solid-liquid separation prior to
entering the wetland. A geotextile membrane, placed at the inlet area of the VSSF-CTW, further
reduced solids entering the wetland and mostly prevented direct contact of solids with the sand
media. Infiltrated water, collected at the bottom of the wetland, was recirculated through
sprinklers over the surface treatment area of the VSSF-CTW. Under this operation, the VSSFCTW has not presented signs of clogging after three years. In fact, mass balance of the total
solids indicated a constant void space of 20 of 30 m3 in the filter media. Additionally, the role of
plant growth in clogging was examined. At the field scale, no differences in volatile solid
accumulation, root development, and infiltration rates in the filter media were found based on
plant type. However, in laboratory-scale columns, columns planted with Canna and Cyperus
exhibited different infiltration rates, despite similar volatile solid accumulation and root biomass,
indicating that root morphology of Canna could be favorable to infiltration.

5.1. Introduction
VSSF-CTWs have been commonly utilized for decentralized wastewater treatment
throughout Europe and Asia (Kadlec and Wallace 2009, Wu, Kuschk et al. 2014). VSSF-CTWs
consist of filter media, usually sand or gravel, planted with emergent plants. The filter media

133

provides a structural base for the emergent plants and a fixed surface upon which microbial
communities can grow (Tilley, Luethi et al. 2008). Wastewater flows vertically through the
entire height of the filter media prior to discharge (Kadlec and Wallace 2009). Flow in VSSFCTWs can be either downward or upward, with either intermittent or continuous application of
wastewater. In general, VSSF-CTWs in both tropical and subtropical climates are effective at
removing organic matter, total suspended solids (TSS), and nutrients. In Egypt, a VSSF-CTW
planted with Canna indica, Phragmites australis, and Cyperus papyrus, treated municipal
wastewater with removals of 92% of TSS, 88% of biochemical oxygen demand (BOD), 53% of
total nitrogen (TN), and 62% of total phosphorus (TP) (Abou-Elela and Hellal 2012). In small
communities, VSSF-CTWs have been used as secondary and tertiary treatment facilities for
septic tanks (Garcia-Perez, Harrison et al. 2011), up-flow anaerobic sludge bed (UASB) reactors
(de la Varga, DĂ­az et al. 2013), and anaerobic digesters (Comino, Riggio et al. 2013). For
example, a VSSF-CTW filled with 0.80 m of granitic gravel (3 - 6 mm), operating in series with
a horizontal subsurface flow constructed treatment wetland (HSSF-CTW), was used to treat
effluent from a UASB reactor that contained 1,558 Âą 1,023 mg COD/L and 129 Âą 88 mg TSS/L.
Even with a high surface loading rate (18 Âą 13 g BOD/m2d), the hybrid system of wetlands
achieved effluent concentrations of 448 Âą 541 mg COD/L and 17 Âą 15 mg TSS/L (de la Varga,
DĂ­az et al. 2013).
In intermittently-fed VSSF-CTWs, physicochemical and biological processes occur
mostly under aerobic conditions (Knowles, Dotro et al. 2011). Physical treatment processes, such
as sedimentation, entrapment, and adsorption, remove solids, organic matter, and nutrients from
wastewater (Garcia, Rousseau et al. 2010). Biological processes that contribute to wastewater
treatment include microbial degradation of organic matter and uptake of nutrients by

134

microorganisms and plant roots (Hua, Li et al. 2013). Roots provide habitat, carbon substrates,
and oxygen to rhizospheric microbial communities (Cao, Gregson et al. 1998, Garcia, Rousseau
et al. 2010). Due to predominantly aerobic conditions present in intermittently-fed, VSSF-CTWs,
this type of CTW removes more organic pollutants from wastewater than HSSF-CTWs on a per
area basis, as HSSF-CTWs tend to operate under mostly anoxic or anaerobic conditions (Zhang,
Jinadasa et al. 2015).
Wastewater treatment by subsurface flow (SSF)-CTWs can be adversely impacted by
clogging (Pedescoll, Uggetti et al. 2009), which decreases porosity, limits hydraulic
conductivity, and limits oxygen transport through the filter media (Zurita, De Anda et al. 2009,
Hua, Zhu et al. 2010, Nivala, Knowles et al. 2012). Preventive strategies to minimize clogging,
including modifying operation and design of the wetland, aim to prolong the lifespan of SSFCTWs, which is commonly estimated to be 10 to 15 years (Wallace and Knight 2006). With
time, filtration of solids by the filter media, chemical precipitation, growth of microorganisms,
and root growth can increase clogging (Suliman, French et al. 2006). Hua, Li et al (2013)
modeled pore space reduction as a function of accumulation of fixed and volatile solids,
microbial growth, and emergent plant roots. Accumulation of fixed solids dominated the
clogging process, occupying 70 to 80% of the pore space (Hua, Li et al. 2013). Chemical
precipitates of phosphorus and metals can also occupy void spaces in the filter media and thereby
also contribute to clogging (Suliman, French et al. 2006). Processes contributing to clogging by
organic matter (i.e., volatile solids) are dynamic and complex, as volatile solids are degraded and
generated by microbial growth and death (Leverenz, Tchobanoglous et al. 2009). Intermittent
application of wastewater in SSF-CTWs can promote both growth of microbes and decay of
microbial biomass, thus organic matter and microbial growth are generally considered minor for

135

long-term clogging (Kadlec and Wallace 2009, Hua, Li et al. 2013). Emergent plant roots are
considered by some to contribute to clogging, as roots occupy pore spaces in the filter media
(Pedescoll, Corzo et al. 2011, De Paoli and Sperling 2013). Pedescoll, Corzo et al. (2011)
concluded that Phragmites australis root contributed to clogging, as hydraulic conductivity and
porosity were lower by 35% and 10%, respectively, in planted horizontal subsurface-flow
wetlands than in unplanted wetlands. However, Hua, Zhao et al. (2014) concluded that roots only
contributed to clogging during the first stage of plant growth; subsequently, root growth opened
new pore spaces in the filter media. Therefore, the debate regarding the role of root growth on
clogging is ongoing, as root growth has been reported to open clogged filter media (Wang, Xu et
al. 2008) and create preferential pathways for infiltration (Torrens, Molle et al. 2009) or fill pore
voids and reduce hydraulic conductivity (Pedescoll, Corzo et al. 2011, De Paoli and Sperling
2013). However, previous studies do not classify roots based on their morphology or compare
how different morphologies of roots contribute to clogging. Most studies evaluated how porosity
and hydraulic conductivity, as indicators of clogging, differ between planted and non-planted
wetlands.
Clogging can be modeled as a function of solids, microbial growth, and root growth.
Such models have indicated that reduction of pore spaces in filter media was mostly due to
inorganic (or fixed) solids, which occupied 70 to 80% of the pore space, while microbial growth
and emergent root plants only represented 6 to 8% of clogged pore space (Hua, Li et al. 2013).
Other studies have also found that microbial communities and roots played only a minor role in
clogging (Langergraber, Haberl et al. 2003, Sani, Scholz et al. 2013). For this reason, some
models only consider inorganic solids; however, understimation of solids accumulation can
occur due to incomplete degradation of organic solids (Sani, Scholz et al. 2013). Validation of

136

these models with full-scale VSSF-CTWs has been limited due to difficulty in quantifying
parameters without altering the wetland filter in full-scale systems. For example, microbial
growth and density, root growth and density, and other microbial-root dependant patameters are
typically assumed (Sani, Scholz et al. 2013). The time to clogging model (tc model) considers
void space in the filter media (Vo, cm3), bulk density of accumulated fixed and volatile solids
(ρb,solids,, mg/cm3), moisture content of the accumulated solids (MCwet basis, %), influent flow rate
(Q, L/d), and the inlet and outlet TSS concentration (Ci and Ce, mg/L), as shown in equation 5.1.
While this model simplifies clogging to a solids mass balance and ignores biological or chemical
factors (Nivala, Knowles et al. 2012), this model has ben repeatedly used to approximate
clogging due to filtration (Langergraber, Haberl et al. (2003), Zhao, Sun et al. (2004), Kadlec
and Wallace (2009), and Hua, Zhu et al. (2010)). However, a k factor accounting for the
degradation of volatile solids can be used for adjusting the void space in the filter media before
estimating the time to clogging (Zhao, Sun et al. 2004).
(1−𝑀𝐶)

𝑡𝑐 = 𝑉𝑜 × 𝜌𝑠 × 𝑄×(𝐶 −𝐶
𝑖

𝑒)

[5.1]

The VSSF-CTW described herein received liquid digestate from an anaerobic digester.
The liquid digestate was classified as high strength wastewater due to its high organic content
(24.7 – 37.2 g COD/m2/d) and suspended solid content (16.5 – 24.8 g TS/m2/d). These loads
were higher than recommended for temperate wetlands (e.g., 20 g COD/m2d and 5 g TSS /m2d)
(Winter and Goetz 2003, Kadlec and Wallace 2009). The VSSF-CTW, planted with Cyperus
papyrus, Canna indica, and Iris graminea, removed 87% of TS, 99% of chemical oxygen
demand (COD), 97% of TN, and 99% of TP (Aguilar Alvarez, Bustamante Roman et al. 2016).
The goal of this study was to evaluate longevity of sand media in the tropical VSSF-CTW by
measuring solid accumulation, root development, and infiltration. Additionally, a laboratory137

scale VSSF-CTW study was conducted to complement the field study, with the intent of
identifying how different root structures affect clogging.

5.2. Material and Methods
A solar-powered anaerobic digester and hybrid constructed treatment wetland was built at
the Fabio Baudrit Experimental Station, in Alajuela, Costa Rica (10.00 m N, -84.26 m W) in
2013 (Aguilar Alvarez, Bustamante Roman et al. 2016). The hybrid constructed treatment
wetland consisted of a VSSF-CTW followed by a free water surface wetland (FWS-CTW). This
paper focuses on the VSSF-CTW, which received liquid digestate from the digester starting in
March 2013. This study was conducted from August 2015 to March 2016.

5.2.1. System description
The VSSF-CTW has an area of 12.0 x 12.0 m on the surface and 9.00 x 9.00 m on the
bottom (with a median area of 138 m2). The media is 1.10 m deep. Clay was used as a liner. The
media profile from the bottom to the surface is 0.20 m of 12 – 20 mm stone, 0.20 m of 4 – 8 mm
pea gravel, and 0.70 m of 0.75 – 2.0 mm coarse sand (32% porosity). From Monday to Friday,
one cubic meter of liquid digestate was discharged daily into the VSSF-CTW through an inlet
pipe (diameter = 0.05 m) located at the North side of the cell. Liquid digestate discharged onto a
2.0 x 2.0 m geotextile membrane (GT 131, Skaps, Athens, GA), which was installed on the
surface of the treatment media. The geotextile membrane was surrounded by cement curbs
(height = 0.25 m) that allowed 1 m3 of wastewater to pool (Figure 5.1). The geotextile membrane
had an apparent opening size of 0.30 mm, a rate of 0.102 m3/s/m2, and a permittivity of 2.20 s-1.
After the geotextile membrane, wastewater passed through the filter media by gravity. The

138

wastewater was collected by distributed grid of drainage pipe (0.075 m) in the bottom layer of
the cell and conveyed to a well pump (model WS V52 from Franklin Electric, Fort Wayne, IN).
The collected, partially-treated wastewater was recirculated at a flow rate of 110 L/min from
2:00 am to 4:00 am daily. The recirculation ratio, as the volume recirculated (13.2 m3) to the
volume of water in the cell (6.15 to 37.4 m3), varied from 0.46 to 2.8. Recirculation was
scheduled during the night to reduce loss of water due to evapotranspiration. Recirculated
wastewater was sprayed from four sprinklers located 1.60 m above the ground and at each corner
of a square (7.00 × 7.00 m) centered in the wetland. Daily, the pump transferred one cubic meter
of treated wastewater into the free water surface constructed treatment wetland.
In 2012, the VSSF-CTW was planted with a grid of randomly-assigned 1 m2-blocks, with
10 blocks per plant species and 10 unplanted control plots. Plant types were chosen based on
availability and previous use in tropical treatment wetlands; initial plantings were Juncus effuses
(common rush), Coix lacryma jobi (Job’s tears), Cyperus papyrus (papyrus), Iris graminea
(dwarf iris or grass-leaved iris), and Canna indica (wild canna lily or African arrowroot). After
three years, at the commencement of this study, Cyperus papyrus, Iris graminea, and Canna
indica were thriving in the wetland; the other three species had not survived (Figure 5.1).
Excluding the two-meter strip in which the geotextile was installed, the VSSF-CTW was evenly
divided in three zones, each three meters wide, from North to South (Figure 5.2). In each zone,
one plot for each plant and one control plot was selected for evaluating solid accumulation, root
development, and infiltration. In each planted and control plot, infiltration tests were performed
at multiple positions, at 0.0 m, 0.5 m, and 1.0 m from the center of the plot.

139

1

2

Figure 5. 1. Full-scale VSSF-CTW: Photos of the location of the inlet and geotextile membrane
and the health of the plants during this study. Schematic of the full-scale vertical-flow wetland
(all lengths are in meters). Pictures were taken in August 2015.

Figure 5. 2. Division of the VSSF-CTW per zones. Picture was taken in March 2013, when the
VSSF-CTW was recently built.

5.2.2. Laboratory-scale VSSF-CTW study
Twenty four columns of 0.30 m depth and 0.28 m diameter were filled with coarse sand
from the full-scale VSSF-CTW (Figure 5.3). The column depth was chosen based on studies
140

reporting clogging to normally occur within 0.20 m of the surface (Hua, Zhu et al. 2010).
Columns were planted with Canna, Cyperus, and Iris from the VSSF-CTW. Triplicate columns
were planted with the following treatments: only Canna, only Cyperus, and only Iris; Canna and
Cyperus, Canna and Iris, Iris and Cyperus, Canna and Iris and Cyperus, and control columns
without plants. Daily, from Monday to Friday, liquid digestate was applied at a rate of 230 mL/d,
the same loading as the full-scale system on a per area basis (1 m3/d). A PVC outlet with valve
(diameter = 0.0127 m) was installed at the bottom of each column. Columns were drained daily
by opening the outlet. Columns were constructed and planted in June 2015. Plants were allowed
to grow for two months prior to measurements; during this period plants were irrigated with tap
water. Nutrients in the filter media were sufficient for plant growth, as indicated by
establishment of plants in all columns. This study was conducted in a non-climate-controlled
room and room´s temperature depended on environmental conditions; room temperatures ranged
from 15.0 to 33.0°C. Ambient sunlight from the open sides of the room was supplemented with
fluorescent lights. From 6:00 am to 5:00 pm, supplemental light was provided using fluorescent
lights hung 1.30 m above the table (Figure 5.3).

141

Figure 5. 3. Laboratory-scale VSSF-CTW. Picture was taken in July 2015, during establishment
of plants in the columns.

5.2.3. Solids accumulation in the filter media
Media core samples were collected from both the field-scale and column-scale wetlands.
For the column experiment, to avoid alteration of the filter media during the experiment, core
samples were collected only at the conclusion of the experiment (February 2016). For the VSSFCTW, core samples were collected in November 2015 and January, February, and March 2016.
Sand cores were taken from three positions (0.0 m, 0.5 m, and 1.0 m from the center) in
treatment blocks of the field-scale wetland. Sand cores were collected by inserting a 0.025-m
diameter soil corer to a depth of 0.30 m. Samples were collected in plastic bags and were kept at
4°C until analysis. Cores were divided into multiple layers: the surface to 0.050 m deep, 0.050 to
0.10 m deep, 0.10 to 0.15 m deep, and 0.15 to 0.30 m deep. Roots were removed from each sand
layer. Each sand layer was homogenized and a sample from each layer was analyzed for total,
fixed, and volatile solids (Hach method #8276). Sand samples were weighed and put in
aluminum dishes (VWRÂŽ, catalog number 25433-008) and oven dried in a StabilTherm gravity

142

oven (model OV-12A from Blue M, East Troy, WI) at 100°C for 24 hours to determine total
solids. Sand samples then were put in a StableTemp furnace (model CBFS516A from ColeParmer, Vernon Hills, IL) at 500°C for 30 minutes to determine volatile solids and fixed solids.
Consequently, the measurements for fixed and total solids in the core samples included the sand
media.
A mass balance of fixed solids was conducted to estimate void space in the filter media of
the VSSF-CTW and the time before clogging using the approach by Langergraber, Haberl et al.
(2003), Zhao, Sun et al. (2004), and Hua, Zhu et al. (2010). The initial void space in the filter
media (Vo, m3) was estimated by multiplying the porosity of coarse sand (Ɛ), the depth of filter
media (h), and the median surface area (A) (equation 5.2). Influent flow rate (Q, L/d), inlet and
outlet total solids concentration (Ci and Ce, mg/L), bulk density of accumulating solids (ρs,
mg/cm3), moisture content of the accumulating solids (MC, %), and bed operating time (t, d),
were used to estimate the wet volume of captured solids (equation 5.3). Total solids in the
influent and effluent were measured from weekly samples taken at the VSSF-CTW (Hach
method #8276). The available void space in the filter media (Vt, m3) immediately after
wastewater application was estimated as the difference between the intial void space and the wet
volume of captured fixed and volatile solids. Available void space in the filter media after resting
(V01, m3) assumed that organic solids (e.g., volatile solids) were completely degraded and was
calculated by determinng k (%) (equation 5.4) by measuring the ratio of volatile:total solids in
the influent. Table 5.1 summarizes the parameters used in this study. The time to clogging was
estimated with equation 5.1.
𝑉𝑜 = 𝜀 × ℎ × 𝐴
𝑉𝑡,𝑠 =

𝑄×(𝐶𝑖 −𝐶𝑒 )
𝜌(1−𝑀𝐶)

143

[5.2]
×𝑡

[5.3]

𝑉01 = 𝑉0 − 𝑉𝑡,𝑠 (1 − 𝑘)

[5.4]

Table 5. 1. Given parameters for estimating of the void space in the filter media and the time
before clogging at the full-scale VSSF-CTW.
Parameter

Unit

Values

Source

Porosity of coarse sand, Ɛ

%

32

Measured

Depth of filter media, h

M

0.70

Measured

Median surface area, A

2

m

138

Measured

Flow rate, Q

L/d

660 – 1,100

Measured

Inlet TS concentration, Ci

mg/L

2,101 – 8,155

Measured

Outlet TS concentration, Ce

mg/L

255 – 3,770

Measured

Density of solids, ρs

mg/cm

1050

(Zhao, Sun et al. 2004)

Moisture content of SS, MC

%

98.8 – 99.7

Measured

Bed operating time

d

7.0

Measured

Organic substrate of solids, k

%

8.48 – 55.6

Measured

3

5.2.4. Root development in the filter media
Root biomass was collected to evaluate whether infiltration changed due to presence of
roots. Roots were collected from core media samples. Roots were carefully removed from the
sand, placed on a sieve with 0.05-cm openings, and shaken and rinsed until sand was visibly
removed from the roots. Root samples were oven dried in a StabilTherm gravity oven (model
OV-12A from Blue M, East Troy, WI) at 100°C for 24 hours to determine dry mass.

5.2.5. Infiltration in the filter media
Modified infiltration tests, similar to those by (De Paoli and Sperling 2013), were
performed to evaluate the extent of clogging in the wetland and to compare infiltration rates
based on plant treatment. For the column experiments, infiltration tests were performed in
August, September, and November 2015 and February 2016. First, columns were saturated with
tap water. Once saturated, 0.032 m of tap water was applied to each column. Next, the outlet was
144

opened and the time for infiltrating 0.032 m of tap water was recorded. Three repetitions per
column were conducted. The column experiment was conducted to determine differences
between the plant treatments when digestate was directly applied to the treatment media (i.e.,
without geotextile separation of solids). For the VSSF-CTW, infiltration tests were conducted in
January, February, and March 2016. The VSSF-CTW was not saturated prior to testing; instead,
infiltration tests were performed under ambient moisture content in the media. In each zone,
infiltration was conducted for each plant treatment and control plot. In each selected plot, three
infiltration tests were performed at multiple positions (0.0 m, 0.5 m, and 1.0 m from the center of
the plot). The double ring infiltration method was modified due to space constraints in the VSSFCTW. A 0.30-m long PVC pipe (diameter = 0.10 m) was inserted 0.20-m deep in the sand. Prior
to the test, tap water was poured inside the PVC pipe until a constant infiltration rate was
achieved. The time for infiltrating the 0.25 m of tap water at a constant rate was recorded as an
estimate of the infiltration rate. The experiment was conducted to determine differences based on
time, zone (representing distance from the inlet), plant species, and position (distance from the
center of the block).

5.2.6. Treatment performance of the VSSF-CTW
For the column experiments, COD (Hach method #8000) and TS (Hach method #8276)
were measured from the inlet and outlet of the columns. One measurement per month was
conducted in August, September, November 2012, and February 2016, before performing the
infiltration test. Samples were collected in 50 mL plastic containers, capped with a lid, and stored
at 4°C until analyses. A DRB 200 reactor (Hach product #LTV082.53.40001) and a DR 900
multiparameter handheld colorimeter (Hach product #9385100) were used to digest and quantify

145

COD concentration (low range digestion vials from 0 to 150 mg COD/L, Hach product
#2565025). Treatment performance was compared amongst plant treatments.

5.2.7. Statistical analysis
Statistical analysis of data was conducted using JMPÂŽ, Version 10.0.0 (SAS Institute
Inc., Cary, NC). Normality of the data was evaluated using the Shapiro-Wilk test. Parametric
tests were conducted in case of normality; otherwise, non-parametric tests were conducted
(Table 5.2). All tests were evaluated at an alpha level of 0.05. In addition, multiple linear
regressions were conducted for estimating correlations between infiltration rates, volatile solids,
and root biomass.

146

Table 5. 2. Statistical analysis.
Experiment
Solids
accumulation in
the filter media

Scale
Column

Type of data
Non-normal

Statistical test
Kruskal-Wallis

Field

Non-normal

Friedman´s test

Kruskal-Wallis

Root development
in the filter media

Column

Normal

Field

Non-normal

Turkey-Kramer
HSD
Kruskal-Wallis

Column

Non-normal

Kruskal-Wallis

Field

Normal

Infiltration in the
filter media

Goal
To determine differences of solids
accumulation at different depths and plant
treatment.
To determine differences of solids
accumulation based on time.
To determine differences among the zone,
position, and plant species.
To determine differences in root
development based on treatment.
To determine differences among zone,
root development and plant species.
To determine differences based on plant
treatments.

Friedman´s test

To determine infiltration changes based
on time.

Repeated
measurements
analysis of
variance

To determine differences based on time.

Turkey-Kramer
HSD

To determine differences based on zone
(representing distance from the inlet),
plant treatments, and position (distance
from the center of the block).

Multiple linear
regression

To determine correlations between
infiltration rates and root biomass, and
infiltration as function of root biomass
and volatile solids.

5.3. Results and discussion
5.3.1. Full-scale VSSF-CTW
Multiple solid-liquid separation units were used to remove solids from the digestate to
minimize solid loading on the VSSF-CTW, including screen separation of solids and liquids,
storage, and filtration. After anaerobic digestion, a rotary liquid/solid separation unit (ICAFE ÂŽ,
Costa Rica, particle size > 0.5 mm in diameter) was used to separate out large solid particles
from the digestate. As a result, 28 kg of digestate solids were removed per day; separated solids
were used as a fertilizer after composting (Aguilar Alvarez, Bustamante Roman et al. 2016).
147

Separated solids mainly consisted of undigested fruit seeds (e.g., watermelon and cantaloupe)
and plant fibers (Figure 5.4.a). The liquid digestate was then stored in the effluent storage tank,
which served as a settling tank. It was necessary to clean the tank in January 2016 due to
accumulation of sediments (not measured) to the height of the outlet pipe (Figure 5.4.b).

Figure 5. 4. Preventive strategies applied for reducing solids in the liquid digestate. a. Rotary
liquid/solid separation unit; and b. effluent storage tank. Pictures were taken in January 2016.

Finally, the geotextile membrane removed fine particles from the liquid digestate. The
membrane dimension (2 x 2 x 0.25 m3) had sufficient capacity to hold the one cubic meter
discharged per day into the VSSF-CTW. According to specifications, the geotextile membrane
retained particles larger than 0.30 mm and allowed a flow rate of 0.408 m3/s. Thus, in the
absence of solids, approximately one cubic meter of water should have passed through the
membrane in 2.45 s. However, the liquid digestate pooled for much longer while passing through
the geotextile (not measured, approximately one to 15 hours). Continuous accumulation of solids
in the geotextile membrane reduced the flow rate. Per week, the membrane removed 4.81% of
COD, 27.9% of TS, 13.4% of TN, and 19.5% of TP (Table 5.3), creating a thick layer of
148

sediments (Figure 5.5). The membrane was important because it prevented direct contact of
solids with filter media and, subsequently, formation of a crust was not observed in any zone of
the treatment area of the VSSF-CTW. Moreover, accumulated solids were easily collected from
the membrane for further use as a fertilizer (Figures 5.1 and 5.5).

Figure 5. 5. The geotextile membrane: a. during liquid digestate influent flow; b. during filtration
of pooled digestate; and c. during removal of sediments. Pictures were taken in October 2014.

149

Table 5. 3. Mass removals for COD, TS, TN, and TP at VSSF-CTW *.
Mass loading
into the
geotextile

Mass removal
by the
geotextile

Mass loading
after the
geotextile

kg /m2/week

Mass after the
VSSF-CTW

Mass removal
by the VSSFCTW

kg /m2/week

COD

8.92 Âą 1.08

0.429 Âą 0.0871

8.49 Âą 1.05

TS

5.77 Âą 0.734

1.61 Âą 0.328

4.15 Âą 0.737

TN

1.39 Âą 0.145

0.186 Âą 0.0298

1.20 Âą 0.136

TP

0.148 Âą 0.0269

0.0289 Âą
0.00587

0.119 Âą
0.0254

0.0558 Âą 0.0268
(6.17 Âą 2.06) x102

(2.37 Âą 0.631)
x10-3
(1.72 Âą 0.978)
x10-4

Removal
by the
VSSFCTW**
(%)

8.44 Âą 1.04

99.4 Âą 0.3

4.09 Âą 0.727

99.0 Âą 0.4

1.20 Âą 0.136

99.8 Âą 0.1

0.119 Âą 0.0254

99.8 Âą 0.1

* Modified from Chapter 4
** Includes the geotextile membrane

The expected movement of solids in the VSSF-CTW is shown in Figure 5.6. The liquid
digestate was discharged above the geotextile membrane and 27.9% of TS were removed. The
movement of the wastewater through the filter media was not measured or modeled; however,
the filter media beneath the geotextile received high mass loading and the VSSF-CTW removed
99.0% of TS. A mass balance of the solids was conducted in the portion of the wetland beneath
the geotextile membrane (4 m2 and 0.70 m height of the filter media) to estimate the void space
in the filter media and the time before clogging (equations 5.1 – 5.4). The mass balance
estimated that void space in the media beneath the geotextile would be occupied in a range of
2.93 to 3.34 days (beneath the geotextile series, Figure 5.7.a), based on the loading of solids after
the geotextile membrane. Therefore, in a short period, this portion of the wetland was likely
clogged. In fact, at the end of the experiment, liquid digestate pooled for longer times above the
geotextile membrane, even after the geotextile membrane was cleaned. Based on this estimation,
an extra 40 – 50 m3 of solids entered beneath the geotextile membrane than what it could hold.
These solids could become more compacted and/or also diffuse away from the media beneath the
geotextile membrane. Therefore, if there was no compaction, then the volume available in the
entire treatment area is less than estimated volume of solids. However, due to changes in the
150

operation of the wetland after this study, it was not possible to analyze this portion of the wetland
for solids accumulation or infiltration rate.

Figure 5. 6. Mass balance on the VSSF-CTW. Figure is not to scale.

The surface of the main treatment area in the VSSF-CTW (zones 1, 2, and 3, Figures 5.2
and 5.6) only received wastewater through recirculation; influent wastewater would have
dispersed mostly within the media beneath the geotextile. Consequently, loading of solids in
zones 1, 2, and 3 was considerably less than the initial solid loading in the liquid digestate
(Figures 5.6 and 5.7.b). A mass balance of the solids was conducted to estimate the void space in
the filter media and the time before clogging in zones 1, 2, and 3 (108 m2 and 0.70 m height of
the filter media). The inlet total solids concentration was estimated as the 75th-percentile of the
effluent concentration data (Ci = C75e = 978 mg/L), while the outlet TS concentrations were set as
the effluent concentrations from the VSSF-CTW. Under this condition, void space in the filter
media was estimated to be higher than 20 m3 and relatively constant (main treatment area series,
151

Figure 5.7.a). This simulation could explain why a crust was never observed in the surface area
of the VSSF-CTW. A final estimation of solid accumulation considered the case of an even
distribution over the wetland surface area (138 m2 and 0.70 m height of the filter media) without
a geotextile membrane. The VSSF-CTW would then hypothetically become clogged in 80 days
(main treatment area if no geotextile were used, Figure 5.7.a).

Figure 5. 7. a. Mass balance of the solids to estimate the void space in the filter media and the
time before clogging; b. Photos of wastewater before and after the VSSF-CTW.

After three years of operation, the mass balance indicated that fixed solids have not
impacted the filter media of the VSSF-CTW. Consequently, a further examination of volatile
solids in the media was conducted. In general, volatile solid accumulation for all plant treatments
ranged from 4.46 to 12.0 g VS/L (Figure 5.8). Volatile solids concentrations in the media were
similar for each plant species and independent of zone and sampling month, as indicated by the
p-values >0.05 when the Wilcoxon non-parametric multiple comparisons were conducted (Table

152

5.4). However, the experimental design was not powerful enough to detect differences in volatile
solids accumulation, likely due to limited numbers of replicates. The power of comparisons
based on zone and plant treatment ranged from 0.052 to 0.520 (Table 5.4). For example, in zone
1 at February 2016, the comparison of volatile solid accumulation in the media beneath Iris
plants, as compared to the media beneath Cyperus plants, yielded a p-value of 0.0809. However,
the power of the statistical comparison for these two plant treatments (0.520) was below the
commonly accepted value of 0.8, indicating that it is possible that a difference may have been
detected with a greater sample number. Compared with De Paoli and Sperling (2013), our VSSFCTW was not clogged. In De Paoli and Sperling (2013), a subsurface flow wetland planted with
Typha latofolia accumulated 35 g VS/L in the area where wastewater was discharged, causing
clogging and surface runoff. In addition, De Paoli and Sperling (2013) found 16.8 g VS/L in an
unplanted subsurface flow wetland, in which no surface runoff was reported.

153

Figure 5. 8. Solid accumulation (as g VS//L) in the VSSF-CTW. T1: January 2016, T2: February
2016, and T3: March 2016. Z1: zone 1, Z2: zone 2, and Z3: zone 3. Statistical analysis applied:
Kruskal-Wallis, using Wilcoxon non-parametric multiple comparison at alpha level of 0.05. N =
9, except for Cyperus (n=6).

Table 5. 4. Statistical p-value and power for solid accumulation analysis in the vertical-flow
wetland.
Plant
treatments*
p-value
Power

January 2016
Zone 1
>0.149
0.243

Zone 2
>0.662
0.0678

February 2016
Zone 3
>0.148
0.115

Zone 1
>0.0809
0.520

Zone 2
>0.149
0.0520

March 2016
Zone 3
>0.0809
0.421

Zone 1
>0.382
0.154

Zone 2
>0.386
0.0699

Zone 3
>0.383
0.172

Since March 2013, plants have been trimmed to keep them within the predefined one
square meter blocks. During this study, Cyperus demonstrated the most robust growth. Cyperus
plants formed a dense culm that did not allow the collection of core samples and roots or
conduction of infiltration tests in the center of the blocks without injury to the Cyperus plants. In
general, Cyperus had greater total root biomass than Canna and Iris, even when a sample from
center of the block was not collected for Cyperus. Cyperus had denser and longer roots, as roots
were also found at one meter from the center of the block; in contrast, Canna and Iris roots were
154

not observed at one meter from the center of the block. In addition, Cyperus root biomass was
greater at 0.50 m from the center of the block than Iris (p=0.0006) and Canna (p=0.0013).
Finally, Canna root biomass was greater than Iris root biomass at the center block (p=0.0015)
and at 0.50 m from the center of the block (p=0.0134) (Figure 5.9). Similarly, Cheng, Chen et al.
(2009) reported higher root growth for Cyperus (7,000 g/m2) than Canna (5,500 g/m2). In
addition, Liang, Zhang et al. (2011) demonstrated that Iris developed lees root biomass (3.67
g/plant) than Canna (7.26 g/ plant).

Figure 5. 9. Total dry root biomass per species (n = 9, except for Cyperus (n = 6)) established at
the VSSF-CTW. Abbreviations: Iris (IG), Canna (CI), Cyperus (CP), and control (C). Position 1
(center of the block); position 2 (0.50 m from the center of the block); and position 3 (1.0 m from
the center of the block).

Infiltration rates for all plant treatments in any zone and among zones with plant
treatment were not significantly different, as indicated by the p-values >0.05 when the two way
analysis of variance was conducted at alpha level of 0.05 (Table 5.5). In addition, no differences
were found based on time (Figure 5.10). This result was expected, as the volatile solids
155

accumulation over zones 1, 2, and 3 was similar. Any difference in volatile solids accumulation
in the filter media can impact infiltration rate. For example, De Paoli and Sperling (2013)
reported higher hydraulic conductivity (107 m/d) in treatment areas where the VS concentration
was 1 g/L, in contrast to 11 m/d in treatment areas where the VS concentration was 35 g/L.

Figure 5. 10. Infiltration rates at the VSSF-CTW (n = 9, except for Cyperus (n = 6)). T1: January
2016, T2: February 2016, and T3: March 2016. Z1: zone 1, Z2: zone 2, and Z3: zone 3.
Statistical analysis applied: two way analysis of variance at alpha level of 0.05.

Table 5. 5. Statistical p-value and power for infiltration rates in the VSSF-CTW.
Plant treatments*
Zone, p-value
Zone and treatment, p-value

January 2016
0.084
0.734

February 2016
0.414
0.718

March 2016
0.198
0.514

In general, low correlations were found when infiltration was analyzed as a function of
volatile solids and root biomass. Considering all plant treatments, volatile solids, and root
biomass, the R2 yielded a low value of 0.25. In fact, low correlations for Canna (R2= 0.24),
Cyperus (R2= 0.13), Iris (R2= 0.16), and unplanted blocks (R2= 0.005) were found. For all cases,
156

power ranged from 0.052 to 0.75, below the desired power of 0.80, and it was less likely to
detect higher correlations due to limited numbers of replicates.
Infiltration was dependent on proximity to the main stalk for only Canna. It was noticed
that infiltration rate increased from 1.0 m from the center of the block to the center of the block
for Canna (p = 0.0040), even though more roots were observed at the center of the plot (Figure
5.9). A similar trend was found for Cyperus, as infiltration at 0.50 m from the center of the block
was higher than infiltration at 1.0 m from the center of the block (p = 0.0172), even though there
was higher root biomass at 0.50 m from the center of the block (Figure 5.9). In fact, infiltration
rates increased as root biomass increased for Canna and Cyperus; however, the correlations
between root biomass and infiltration rates were low for Canna (R2=0.24) and Cyperus (R2=
0.24) (Figure 5.12). In contrast, Iris had similar infiltration rates at the center of the block and at
0.50 m from the center of the block; however, infiltration at 0.50 m from the center of the block
was significantly greater than infiltration rate at 1.0 m from the center of the block (p = 0.0162),
where no roots were found (Figure 5.9). In fact, the correlation between root biomass and
infiltration rates was low for Iris (R2= 0.15), and slow and rapid infiltration rates were obtained
with a similar root biomass (Figure 5.12). In general, infiltrations rates increased as root biomass
increased; however the correlation was low (R2= 0.25). Despite these low correlations between
infiltration rates and root biomass, it was likely that roots had impacted infiltration rates. For
example unplanted plots had similar infiltration rates based on position, likely due to uniform
similar VS concentration in the filter media (Figure 5.8) and lack of roots occupying pore spaces
in the filter media (Figure 5.11).

157

Figure 5. 11. Infiltration rates per positions at the VSSF-CTW. Position 1 (center of the plant);
position 2 (0.50 m from the center of the plant); and position 3 (1.0 m from the center of the
plant). Letters A, B, and C indicate significant differences between infiltrations per position per
each plant treatment. Statistical analysis applied: Turkey-Kramer HSD at alpha level of 0.05.

158

Figure 5. 12. Infiltration rates in the VSSF-CTW as a function of root biomass for Canna,
Cyperus, and Iris.

5.3.2. Laboratory-scale VSSF-CTW
The laboratory-scale VSSF-CTWs were built to determine how different root structures
affect clogging. This study was limited to only one experimental replicate of columns due to time
and budget constraints. Thus, to improve representativeness of columns for estimating clogging
would be needed more sets of columns keeping some factors constant (e.g.: column Depth, OLR,
plant density) and varying other factors (e.g.: diameter of columns).
The filter media directly received liquid digestate and recirculation was not performed.
Consequently, clogging in the laboratory-scale wetlands was affected by a combination of fixed
solids, volatile solids, and root growth. In general, COD removal was higher than 98%; however,
effluent COD concentrations increased toward the end of this experiment. Similarly, TS removal
decreased with time. After accounting for time-dependence, treatment performance did not differ
159

amongst plant treatments for COD (p>0.247), for TS (p>0.286), and for FS (p>0.243).
Generally, treatment performance tends to increase during initial accumulation of solids on the
surface; however, filtered and trapped solids will eventually limit oxygen transport and reduce
the ability of the system to treat the wastewater, especially if a crust is formed (Hua, Zhu et al.
2010). Our data indicates that crust was formed after 30 days, and treatment performance
decreased in November 2015. In fact, COD effluent was similar in September and October 2015,
however, effluent significantly increased to 54 mg COD/L in November 2015 (p = 0.0041, with
respect to October 2015), and significantly increased to 140 mg COD/L in February 2016 (p <
0.0001, with respect to November 2015). Similarly, TS and FS effluent significantly increased to
4,323 mg TS/L and 1,592 mg FS/L in November 2015 (p<0.0187, with respect to October 2015).
In our case, it was likely that treatment declined due to the accumulation of solids (Table 5.6).

Table 5. 6. Effluent water quality parameters in the columns (mean Âą S.D., n = 24).
Parameter
COD in (mg/L)
COD out (mg/L)
COD removal (%)
TS in (mg/L)
TS out (mg/L)
TS removal (%)
FS in (mg/L)
FS out (mg/L)
FS removal (%)

Sept-2016
4,090 Âą 221
39 Âą 14 a
99.0 Âą 0.4
3,371 Âą 154
n.d.
n.d.
1,936 Âą 62
n.d.
n.d.

Oct-2015
5,896 Âą 637
40 Âą 11 a
99.3 Âą 0.4
4,228 Âą 581
1,861 Âą 517 a
55.9 Âą 0.2
2,732 Âą 210
1,071 Âą 74 a
60.8 Âą 0.1

Nov-2015
11,313 Âą 889
54 Âą 18 b
99.5 Âą 0.4
6,254 Âą 335
4,323 Âą 1,100 b
30.8 Âą 0.1
4,480 Âą 189
1,592 Âą 60 b
64.4 Âą 0.1

Feb-2016
7,420 Âą 400
140 Âą 19 c
98.1 Âą 0.2
4,251 Âą 1,034
5,000 Âą 1,400 b
(17.6 Âą 0.1)
3,089 Âą 623
2,380 Âą 169 c
22.9 Âą 0.1

Letters indicate statistical significant mean differences of COD, TS, and FS concentration
base on time at Îą = 0.05. No data (n.d.)

The filter media removed from 20 to 65% of fixed solids and 0 to 47% of the volatile
solids. In general, concentrations of volatile solids in the top 0.05 m of media were
approximately 50% greater than volatile solid concentrations in the lower three layers; however,
this difference was only statistically significant for columns planted with only Iris (Figure 5.13).
160

For Iris, solids accumulation in the surface layer was significantly different with respect to
deeper layers (0.0131 < p < 0.0306). It is possible that the experimental design, given the
limitations on replicate number, was not powerful enough to detect differences in volatile solids
accumulations between media depths for other columns. The power of comparisons based on
depth and plant treatment ranged from 0.0950 to 0.358. For example, the comparison of volatile
solid accumulation in the surface layer of Cyperus-planted columns, as compared to the second
layer, yielded a p-value of 0.0656. However, the power of the statistical comparison for these
two layers (0.215) was below the commonly accepted value of 0.8, indicating that it is possible
that a difference may have been detected with a greater sample number. Higher accumulation in
the top layer with respect to deeper layers was expected due to filtration and entrapment of solids
by the media (Hua, Zhu et al. 2010). In fact, a crust formation of solids was continuously
observed each time the columns were irrigated with liquid digestate. Similar trends were reported
by Hua, Zhu et al. (2010), when they observed in a laboratory-scale vertical-flow wetland (0.15
m diameter and 0.40 m depth), that the filter gradually clogged due to the application of 600 mg
total solids/L of wastewater. Overall, more volatile solids were accumulated in columns planted
with only Iris than in the control or other planted columns. For example, concentrations of
volatile solids in the top layer of columns planted with only Iris ranged from 12.9 and 22.7 g/L;
in contrast, accumulation of volatile solids in the top layer of other treatments was significantly
lower at 9.98 to 13.7 g/L.

161

Figure 5. 13. a. VS accumulation unplanted columns and six of the seven plant treatments
(Cyperus, Canna-Cyperus, Canna, Canna-Iris-Cyperus, Canna-Iris, and Iris-Cyperus); b. VS
accumulation for Iris plant treatment. Note. A: from surface to 0.05 m deep, B: from 0.05 to 0.10
m deep, C: from 0.10 to 0.15 m deep, and D: from 0.15 to 0.30 m deep.

Root development was generally less for individual plant treatments (e.g., Iris, Canna,
Cyperus) than for multiple plant treatments (e.g., Canna-Iris-Cyperus). Based on comparisons
for all pairs using Turkey-Kramer HSD, root development per treatment was ranked as low,
medium, or high (Table 5.7). With the exception of the Canna-Iris-Cyperus and Canna-Iris
treatments, all planted treatments were classified as “low” root biomass. The greatest root
biomass occurred in columns planted with all three plant species (Canna-Iris-Cyperus) and
Canna-Iris. Importantly, only one plant was planted per species in the column. For example,
Canna only had one plant, whereas Canna-Iris-Cyperus had three plants. In individual species
treatments, Cyperus developed slightly more root biomass than Canna and Iris; similar to results
obtained by Liang, Zhang et al. (2011), who reported that Iris plants grew 3.67 g of roots per
plant, as compared to 7.26 g/Canna plant and 1.17 g/Cyperus plant. Likewise, Cheng, Chen et al.
(2009) reported that Cyperus and Canna roots grew at rates of 7,000 g/m2 and 5,500 g/m2,
respectively.

162

Table 5. 7. Total dry root mass per plant treatment after the experiment (mean Âą S.E., n = 9).
Treatments

IG a

CI a

CP a

CI-IG a

IG-CP a

CI-CP ab

CI-IG-CP b

Mean (g)

0.926

1.40

1.49

1.90

1.99

2.40

4.50

S.E. (Âą g)

0.246

0.47

0.35

0.64

0.50

0.81

0.73

Rank

Low

Low

Low

Low

Low

Medium

High

pvalues

IG

--

CI

0.99

--

CP

0.99

1.00

--

CI-IG

0.88

0.99

0.99

--

IG-CP

0.83

0.98

0.99

1.00

--

CI-CP

0.53

0.87

0.91

0.99

0.99

--

CI-IG-CP

0.0008

0.0054

0.0074

0.032

0.043

0.14

--

Letters indicate statistical significant mean differences of root development between plant
treatment at Îą = 0.05. Iris (IG), Canna (CI), Cyperus (CP), Canna-Iris (CI-IG), Iris-Cyperus (IGCP), Canna-Cyperus (CI-CP), Canna-Iris-Cyperus (CI-IG-CP), and control (C).

For all treatments, infiltration rates were significantly higher during August 2015 than
during the rest of sampling dates (Figure 5.14, p<0.0001). In general, there was a statistically
significant interaction between time and treatment (p = 0.006). A substantial reduction in the
infiltration rate from August to September 2015 was expected, based on previous studies
showing rapid accumulation of solids in the surface of the filter (Hua, Zhu et al. 2010) and own
observations of crust formation in this study. Then, infiltration rates continued to decrease
slightly in subsequent months for all plant treatments.

163

a

b
Figure 5. 14. a. Infiltration rates (median, n = 9) for Canna, control, Cyperus, and Iris based on
time. b. Infiltration rates (median, n = 9) for Canna-Cyperus, Canna-Iris, Canna-Iris-Cyperus,
and Iris-Cyperus based on time. T1: August 2015, T2: September 2015, T3: November 2015,
and T4: February 2016.

Due to the destructiveness for collecting media core samples and root biomass,
relationships for infiltration rates based on time, root biomass, and volatile solids were collected
164

once at the end on the experiment. Overall infiltration rates (i.e., including all sampling times,
Table 5.8) were the highest for the unplanted columns and columns planted with Canna-Iris and
Canna plants. Remaining infiltrations rates, while lower than those observed in the control
columns, were not significantly different from each other. In individual plant treatments, Canna
plants did not decrease the infiltration rate significantly from the control (p = 0.847); however,
infiltration rates for the Canna columns were significantly faster than those for columns planted
with Iris (p = 0.042). The slowest infiltration rates, amongst all treatments, were observed for
columns planted with only Cyperus (p = 0.001 versus the control).

Table 5. 8. Relative infiltration rates rankings for plant treatments (n = 36).
Treatment

Canna-Iris
Control
Canna
Iris
Iris-Cyperus
CannaCyperus
Canna-IrisCyperus
Cyperus

Mean
infiltration
rate (mm/s)
0.0941 Âą
0.0089
0.0917 Âą
0.0091
0.0908 Âą
0.0063
0.0767 Âą
0.0064
0.0750 Âą
0.0076
0.0729 Âą
0.0016
0.0718 Âą
0.0079
0.0699 Âą
0.0064

Overall

August
2015

September
2015

November 2015

February
2016

High

High

High-med

High-med

Med

High

High

High-med

High-med

Med

High

Med

High

High

Med

Low

Med

High-med-low

Med-Low

Med

Low

High-med

High-med

Med-low

Med

Low

Med

Med-low

Low

Med

Low

High

Med-low

Low

Med

Low

Med

Low

Med-Low

Med

Caption: Based on significance of 0.05, relative infiltration rates rankings for plant
treatments of high, medium (med), and low for each sample time. Some treatments are ranked as
high-med, med-low, or high-med-low to indicate there was no statistical difference between that
specific treatment and other treatments in the high, medium, or low rankings. There was no
statistical difference in infiltration rates in February 2016, therefore, all treatments are ranked as
medium.

In general, control columns without plants were categorized as having high (or high and
medium) infiltration rates at all sampling times; therefore, there was no significant evidence that
165

plants increased infiltration in the column studies. On the contrary, columns with plants
generally exhibited lower infiltration rates than the control columns, especially columns planted
with Cyperus and Canna-Cyperus. However, as the experiment progressed, infiltration rates
from all treatments converged, so that there were no significant differences between infiltration
rates in the unplanted columns and the planted columns by February 2016.
Within the planted columns, columns with Canna plants had the highest infiltration rates.
For example, volatile solid accumulation and root development were similar for Canna, CannaIris, Cyperus, and Iris-Cyperus; however, treatments with Canna had higher infiltration rates.
Canna has a bulb type root with a horizontal propagation method of rhizome division (Cabi
2016), and from each division, shoot and root growth can promote the opening of pores in the
filter media (Wang, Xu et al. 2008) (Figure 15.a). In contrast, lower infiltration rates were
measured for columns with Cyperus (e.g., Cyperus and Canna-Iris-Cyperus), despite similar
volatile solid accumulation and root development. Cyperus has a tough root with numerous
rootlets that forms a dense culm at the surface of the soil (Cabi 2016) along with denser and
longer root development from 5 to 15 cm deep of the filter media (Cheng, Chen et al. 2009).
Thus, Cyperus roots could decrease infiltration (Figure 15.b).
Similar to Cyperus, Iris root has a rhizome apical root, from which fine auxiliary buds
and adventitious roots expand (Iris 2016). However, Iris roots are thinner than Cyperus roots and
occupy less void space in the filter media (Figure 15.c). Infiltration rates for columns planted
with Iris were typically between rates for columns with Canna and Cyperus, particularly when
Iris was combined with other plant, except all three plants.

166

Figure 5. 15. a. Root structure for Canna (Picture was taken from www.plantsgrow.com); b.
Root structure for Cyperus (Picture was taken from www.olabrisagardens.com); and c. Root
structure for Iris (Picture was taken from www.grovida.us).

5.4. Conclusion
In operation since March 2013, the VSSF-CTW, as a part of a decentralized, selfsufficient, close-loop, organic waste treatment system in Costa Rica, has been effectively treating
anaerobic digestate without incidences of clogging. Reduction of organic and solid loads by
anaerobic digestion and mechanical separation, settling within the effluent storage tank, and
filtration of fine solids by the geotextile membrane has likely prevented clogging within the
VSSF-CTW. Additionally, only recirculated wastewater was applied to the main treatment area
with the goals of increasing nutrient removal while maintaining a robust plant community. Under
this operation, accumulation of fixed solids has not impacted the filter media, as demonstrated by
the mass balance of the solids that estimated a constant void space of 20 m3 in the main treatment
area. Therefore, for future research and optimization, the use of a geotextile membrane is
recommended for designs in Costa Rica. Moreover, there is a need to evaluate the dynamics of
the solids during treatment in the filter media due to recirculation and flooding.
The filter media below the geotextile membrane was likely rapidly filled with solids, as
demonstrated by the mass balance of the solids. However, this phenomenon did not appear to
impact wastewater treatment; however, replacement of this portion of the treatment area could be
167

necessary. Planting selection is not only important for nutrient removal but also for reducing
incidence of clogging. The identification of morphological traits of root plants that reduce
clogging is important for extending the lifespan of wetlands. In this study, low correlations were
found when infiltration was analyzed as a function of volatile solids and root biomass in the sand
media. More data is needed to determine impacts of volatile solids and root biomass within the
full-scale wetland. Trends were identified; root growth of Canna plants did not adversely affect
infiltration. In the full-scale wetland, infiltration rates near Canna plants were relatively high
despite massive root development. Within the column study, columns planted with Canna had
higher infiltration rates than columns planted with Cyperus, even though solid accumulation and
root biomass were similar.

5.5. Acknowledgment
This research was supported by the U.S. Department of State West Hemisphere Affairs
(S-LMAQM-11-GR-075). The authors would like to acknowledge the assistance and support
Water Quality Laboratory at the Research Center of Environmental Pollution (CICA-LCA) of
the University of Costa Rica. The authors would also like to thank the staff at EEAFBM as well
as the researchers from the project entitled “Electric energy production through anaerobic
digestion optimization of agricultural organic wastes and treatment of its effluents”. Also, the
Department of Eco-physiology at EEAFBM and the Meteorological Institute of Costa Rica for
providing environmental data.

168

CHAPTER 6: EXERGY-BASED ASSESSMENT OF SUSTAINABILITY OF A SOLARPOWERED ANAEROBIC DIGESTION AND HYBRID CONSTRUCTED TREATMENT
WETLAND SYSTEM TO TREAT AGRICULTURAL WASTES IN COSTA RICA

Abstract: Agriculture and ecological tourism are both crucial to Costa Rica’s economy.
However, agricultural activity can damage ecosystems when residues and wastewater are
discharged onto land and into surface water. Integration of profitable bioenergy systems with
engineered ecological treatment systems, such as the solar-powered anaerobic digestion and
hybrid constructed treatment wetland (SPAD-HCTW), can potentially balance agricultural
production and ecological protection. The SPAD-HCTW integrates solar heating, anaerobic
digestion, and constructed treatment wetland technologies to treat biomass residues and
agricultural wastewaters while producing energy, fertilizer, and treated water. In this study, an
exergy-based assessment of sustainability using two indexes, the environmental exergy
efficiency (Ρenv,ex) and the total pollution rate (Rpol,ex), was conducted. Calculations were
completed for three different cases to determine how inclusion of constructed treatment wetlands
affected exergetic sustainability. The baseline case was the solar-powered anaerobic digester
(SPAD) alone. Two additional cases, considering either treatment of the digestate (1) by a
vertical subsurface flow constructed treatment wetland (VSSF-CTW) or (2) by the VSSF-CTW
in series with a free water surface constructed treatment wetland (FWS-CTW), were analyzed.
Results indicated that the baseline case, which only provided energy and fertilizer, was not
sustainable from an exergetic point of view (Ρenv,ex = 0.469 ¹ 0.094 and Rpol,ex = 1.13 ¹ 042).
More exergetically favorable Ρenv,ex and Rpol,ex values were obtained with the inclusion of a
VSSF-CTW; however, case 1 was still not considered exergetically sustainable (Ρenv,ex = 0.734 ¹

169

0.201 and Rpol,ex = 0.362 Âą 0.373). The entire SPAD-HCTW, or case 2, was exergetically
sustainable with a Ρenv,ex of 5.60 ¹ 1.56 and a Rpol,ex of -0.821 ¹ 0.167, due to a positive exergy
balance in which the food waste and chicken litter were converted into high quality end products
(i.e., energy, fertilizer, and treated water). Through examination of these case studies, multiple
approaches for calculation of the exergy of treated and stored wastewater were evaluated.
Exergetically, this study shows that technical innovation in conversion of agricultural wastewater
and biomass wastes into resources (e.g., fertilizer, biogas, and treated water) can help address the
adverse impacts of agricultural production on ecosystems.

6.1. Introduction
Agriculture is an important economic driver in Costa Rica, contributing 9.19% to the
total Gross Domestic Product (GDP) (EN 2015). However, while contributing to local
economies, agricultural activities also consume 84% of Costa Rica’s annual fresh water demand
(12,320 m3) and produce 6,000 tons per year of biomass wastes (GWP 2012, Coto 2013). Most
agricultural wastewaters and biomass wastes are directly discharged into aquatic systems with
limited or no treatment (Ruiz 2014), adversely affecting ecosystems that positively contribute to
the Costa Rican economy and environment (EN 2015). In 2014, ecosystems contributed to $2.6
billion (5.3% of the GDP) to the Costa Rican economy through tourism alone (ICT 2014).
Environmentally, ecosystems in Costa Rica sequestered 26 tons per hectare of carbon each year
(Salazar 2014). However, while there is an economic need to protect ecosystems from the
discharge of agricultural wastes, limited funding is allocated for water resource management and
protection in Costa Rica. For example, it is estimated that Costa Rica should spend $2.0 billion
to manage its water resources from 2010 to 2020 (MINAET 2010); however, actual financial

170

commitments are substantially less than this (GWP 2012, Echeverria and Cantillo 2013). To
confront these challenges, low-cost, sustainable systems that convert wastewater and biomass
residues into valued resources would greatly benefit agricultural areas in Costa Rica.
Combining anaerobic digestion and constructed treatment wetland (CTW) technologies
can efficiently treat biomass residues and wastewater, generate renewable energy, produce
fertilizers, and reclaim water (Barros, Ruiz et al. 2008, Ruiz, DĂ­az et al. 2010, Comino, Riggio et
al. 2013, de la Varga, DĂ­az et al. 2013). Comino, Riggio et al. (2013) reported that a mesophilic
continuous stirred tank reactor (CSTR) followed by parallel VSSF-CTWs and an in-series
horizontal subsurface flow constructed treatment wetland (HSSF-CTW) removed 88% of
chemical oxygen demand (COD), 73% of nitrate (NO3), 98% of ammonium (NH4), and 99% of
phosphate (PO4) from a discharge of 0.200 m3/d, with a maximum organic load rate of 160 g
COD/m2d. Barros, Ruiz et al. (2008) highlighted the benefits of pretreating municipal
wastewater with an up-flow anaerobic sludge blanket (UASB) reactor receiving 3 m3/d and an
organic load of 4,000 mg COD/L. The UASB reactor reduced 80-90% of TSS, 65% of COD, and
40-50% of BOD, while the CTW removal efficiencies were 32-52% for total suspended solids
(TSS), 83% for COD, and 87% for biological oxygen demand (BOD). Similarly, de la Varga,
DĂ­az et al. (2013) and Ruiz, DĂ­az et al. (2010) found that an UASB reactor provided high TSS
removal, while the CTWs reclaimed the water by reducing COD concentrations, which were still
high after the anaerobic digestion pretreatment (COD > 150 mg/L).
Exergy is a thermodynamic concept that describes system performance according to the
second law of thermodynamics that can be used for assessing the sustainability of systems.
Opposite to entropy, which measures the low-quality energy of materials in thermodynamic
disequilibrium, exergy measures the quantity and quality of energy that a particular material

171

possesses if it is brought into thermodynamic equilibrium (Jørgensen 2006, Rosen 2012, Querol,
Gonzalez-Regueral et al. 2013). This allows exergy to be a tool for resource accounting and a
valuable sustainability metric (Chen, Chen et al. 2011). A system that converts materials with
high entropy into high quality end products with low entropy will be sustainable if the balance
between inputs (e.g.: wastes) and outputs (e.g.: biogas, fertilizers, and reclaimed water) is
positive (Wall 2010, Woudstra 2016). Sustainability is a complex concept that varies broadly
depending on the discipline and the context in which is evaluated. In this regard, a sustainable
assessment based on exergy summarize only part of the environmental portion of sustainability
as it evaluates how potential raw material is upgraded to valuable products, thereby alleviating
negative environmental impacts of wastewater and biomass residues (Parsapour 2012).
Exergy analysis has been used previously to analyze the sustainability benefits of
anaerobic digestion. Dong, Chi et al. (2014) found that implementing anaerobic digestion as a
pretreatment step for sewage sludge composting significantly alleviated the environmental
burden of sludge disposal. The environmental exergy efficiency (Ρenv,ex), or the ratio of the total
exergy outputs of products over the inputs, was 16.9% for the composting process alone;
inclusion of AD prior to composting increased Ρenv,ex to 34.6% due to production of biogas
(Dong, Chi et al. 2014). Likewise, Chatzipaschali and Stamatis (2015) indicated that Ρenv,ex of a
steam production plant increased when the biogas produced by anaerobic digestion of cheese
whey was used for the thermal and electrical needs of the plant (Chatzipaschali and Stamatis
2015). Siefert and Litster (2014) modeled the exergetic effects of including an anaerobic digester
with a solid oxide fuel cell (SOFC) into a conventional wastewater treatment plant. The resulting
Ρenv,ex was 58%, mainly due to the production of biogas to fuel the SOFC (Siefert and Litster
2014).

172

Other authors, e.g., Tang, Fang et al. (2014), Chen, Chen et al. (2011), and Shao and
Chen (2015), have evaluated wetland ecosystems using exergy. Tang, Fang et al. (2014) used
exergy, biomass, and diversity as indicators to evaluate the development and health status of
macrofauna for a wetland planted with Sonneratia apetala (mangrove apple). Exergy increased
from 2 MJ/m2 to 9 MJ/m2 in 1,200 days due to the fast growth of S. apetela (Tang, Fang et al.
2014). Chen, Chen et al. (2011) applied cosmic exergy to analyze three wastewater treatment
systems. Cosmic exergy considers the cosmos as a background, as the thermodynamic
equilibrium environment for all the processes in the Earth (Chen, Chen et al. 2011). Chen, Chen
et al. (2011) determined the natural, construction and operational inputs, and the ecological
service outputs of a CTW, an activated sludge system, and a cyclic activated sludge system.
Authors found that the renewable dependency, or the total renewable input over the ecological
input, for the CTW (67%) was more than double the renewable dependency of the activated
sludge system (38%) and the cyclic activated sludge system (31%), indicating that the CTW was
more sustainable and environmentally friendly (Chen, Chen et al. 2011). Finally, Shao and Chen
(2015) used embodied cosmic exergy, which measures the resource utilization efficiency taking
into consideration the renewability index of all inputs of a system. The renewability index is the
total natural and purchased renewable resources over the total resource use in the studied system.
Therefore, the embodied cosmic exergy analysis can be used to determine which inputs should
be changed to improve a system’s renewability. Based on this tool, Shao and Chen (2015)
determined that use of inputs with higher renewability indexes (e.g. PVC pipes, 1.23% versus
cement pipes, RI = 0.89%) during the construction and operation of a constructed wetland would
increase the renewability index from 2.09% to a higher magnitude that was not indicated by the
authors (Shao and Chen 2015).

173

Exergy analysis has also been used to quantify water quality. Tai and Matsushige (1986)
determined that the exergy of COD in wastewater is 13.6 kJ per g COD (Tai and Matsushige
1986). Hellstrom (1997) applied Tai’s ratio to compare energy and exergy while estimating the
physical consumption of resources during conventional wastewater treatment. MartĂ­nez, Uche et
al. (2010) evaluated the exergy efficiency of desalinization during wastewater treatment. Inputs
to the treatment plant were wastewater, described by its COD and salt concentrations, silica used
for coagulation and flocculation, and electricity. Outputs were COD, fat, salts, and silica in the
discharged effluent and sludge. Authors found that chemical-based desalinization processes
(multi-stage flash distillation, multiple effect distillation, reverse osmosis, and electrodialysis)
were less energy efficient than pumping techniques as the ratio of inputs over outputs was much
greater than 1 for chemical-based desalinization (MartĂ­nez, Uche et al. 2010). Khosravi and
Panjeshahi (2013) used an exergy analysis as an optimization tool to identify losses of exergy at
the different stages of the wastewater treatment by using two sustainability indexes, Ρenv,ex, and
total pollution rate (Rpol,ex, ratio of lost exergy from wastes to total exergy outputs). Khosravi and
Panjeshahi (2013) were able to increase efficiency of wastewater treatment by identifying
inefficient steps, increasing Ρenv,ex from 0.14 (low efficiency) to 0.36 and decreasing Rpol,ex from
6.14 (large pollution rate) to 1.79.
Based on previous technical performance analysis (Aguilar Alvarez, Bustamante Roman
et al. 2016), it was demonstrated how each separate unit of the SPAD-HCTW performed as a
close-loop system to yield energy, fertilizers, and treated water. Exergy metrics have
successfully been used to document the sustainability of systems similar to components of the
SPAD-HCTW; consequently, exergy metrics were chosen to analyze the sustainability of the
SPAD-HCTW since the technical point of view. Therefore, an exergy-based assessment of

174

sustainability using two indexes (Ρenv,ex and Rpol,ex) was conducted. The exergy baseline of the
system was the SPAD (case 0) alone. Further analysis considered the SPAD and the VSSF-CTW
(case 1) and, finally, the SPAD, the VSSF-CTW, and the FWS-CTW in-series (case 2) to
compare the potential improvement of sustainability through inclusion of wetlands during
treatment of agricultural wastes.

6.2. Material and Methods
The SPAD-HCTW was designed and built at the Fabio Baudrit Experimental Station
(EEAFBM), in Alajuela, Costa Rica (10.00 m N, -84.26 m W) during 2011 and 2012. Materials
for construction of the SPAD-HCTW were bought at local suppliers in Costa Rica. The system
began operation in March 2013. This study was conducted from August 2015 to March 2016.

6.2.1. System description
The SPAD-HCTW is shown at Figure 6.1 and is described as follow. Flat-plate solar
thermal collectors heated water to 50-78oC. Heated water was pumped with a solar heating fluid
transfer pump (model UP 26-99 F from Grundfos, Olathe, KS) and stored in a 3-m3 hot water
tank. Then, a digester heating pump (model PB 351MA from Wilo, Korea) circulated the hot
water through a heat exchanger within the anaerobic digester to maintain thermophilic conditions
(i.e., 45 Âą 2oC). The anaerobic digester was fed with a mixture of food wastes and chicken litter
at a rate of one cubic meter per day, from Monday to Friday. Each week, 5 m3 of feed mixture
was prepared using treated water from the FWS-CTW. First, 151 Âą 17.1 kg of dry chicken litter
was mixed with one cubic meter of treated water in a 50-gallon drum, and then pumped by feed
preparation pump (model 50ut2.4s-61 from Tsurumi, Glendale Heights, IL) to the feeding tank.

175

Next, approximately 17.4 Âą 3.26 kg of dry food waste was ground (model Leeson C
184K17FB150 from ICAFE ÂŽ, Costa Rica), mixed with 4 m3 of treated water, and then pumped
using the feed preparation pump to the feeding tank. In the feeding tank, mixture was mixed for
30 minutes per week by an external feeding tank pump (model AMT P/N 1626-305-00 from
AMT, Royersford, PA). From Monday to Friday, one cubic meter (~ 1000 kg) of feed was
conveyed by the external pump to the anaerobic digester. The same volume of liquid digestate
left the anaerobic digester and passed to the solid/liquid separator by gravity. In the anaerobic
digester, a submersible digester mixing pump (model 5763 from AMT, Royersford, PA) mixed
the anaerobic digester’s contents for 10 minutes each hour. Biogas produced from the digester
was stored in a biogas storage bag (HDPE 60 m3 from Viogaz ÂŽ, Costa Rica) prior to
combustion by a generator (model B4T-5000 Bioflex from BrancoÂŽ, Brazil). Electricity was
used to power pumps and other pieces of equipment in the system to satisfy operational
requirements. Solids and liquid digestate were separated mechanically with a rotary screen
separator (ICAFE ÂŽ, Costa Rica, particle size > 0.5 mm in diameter), which operated 10 minutes
per day. The solid digestate was used as a fertilizer for on-site farming uses. The liquid digestate
was stored in the effluent tank. Finally, the liquid digestate, one cubic meter per day from
Monday to Friday was delivered by gravity to the HCTW to be further treated.

176

Figure 6. 1. Flowchart of the SPAD-HCTW system. Numbers indicate pumps and other
equipment needed to operate the SPAD-HCTW: 1. Solar heating fluid transfer pump, 2. Digester
heating pump, 3. Digester mixing pump, 4. Feeding tank pump, 5. Solid/liquid separator, 6.
Grinder, 7. Effluent pump, 8. Feed preparation pump, 9. SPAD-HCTW exit pump, 10. Biogas
flowmeter, 11. VSSF-CTW recirculation pump, and 12. FWS-CTW recirculation pump.

The HCTW consists of two CTWs in-series: a VSSF-CTW followed by a FWS-CTW
(Figure 6.1). By gravity, the liquid digestate was discharged into the VSSF-CTW from the
effluent tank. Then, from Monday to Friday, one cubic meter per day of the treated effluent from
the VSSF-CTW was pumped to the FWS-CTW. The treated water from the FWS-CTW was used
as either irrigation water or processing water for on-site uses. Both cells have the same
dimensions. The top is 12.0 × 12.0 m, while the bottom is 9.00 × 9.00 m (median treatment area
= 138 m2). The height is 1.10 m and the slope of the walls is 27° with respect to the horizontal.
The media profile of the VSSF-CTW is, from the bottom to top, 20 cm of stone (12 – 20 mm),
20 cm of pea gravel (4 – 8 mm), and 70 cm of coarse sand (0.75 – 2.0 mm, 32% porosity). The
VSSF-CTW has a maximum water storage capacity in the media of 30.9 m3, with an additional
80 m3 of storage above the media. A 2.00 × 2.00 m geotextile membrane (GT 131 from Skaps,
Athens, GA) was installed beneath the discharge of the pipe from the effluent tank to collect
177

suspended sediments and prior to infiltration of the liquid digestate effluent into the media. The
VSSF-CTW was planted with randomized blocks of one square meter of Cyperus papyrus, Iris
graminea, or Canna indica. Total area covered by each plant was 10 m2. The FWS-CTW has
maximum water storage capacity of 152 m3. The FWS-CTW was planted with two different
species of floating plants: Eichhornia crassipes and Pistia stratiotes. Floating plants were
divided by species by a frame of bamboo (3 m wide and 9 m length per species). The wetland
unit includes three pumps: one recirculation pump installed at the bottom of each CTW (model
WS V52 from Franklin Electric, Fort Wayne, IN) for recirculation within the same CTW and an
exit pump in the FWS-CTW (model WS 102 from Franklin Electric, Fort Wayne, IN) that
transferred the treated water to either the feeding tank or the irrigation system.

6.2.2. Boundary definitions
An exergy-based assessment of sustainability using two indexes was performed for three
cases (Figure 6.2). Case 0, the baseline of the system, only considered the SPAD, which includes
pieces of equipment from number 1 to 9 (Figure 6.1). The inputs included feed, electricity, and
heat. The outputs were solid and liquid digestate, electricity from methane combustion by the
generator, and carbon dioxide, both from the produced biogas. The produced biogas was used as
the output for all cases, offset by different electrical usage due to additional pumps in Cases 1
and 2. Case 1 considered the SPAD and the VSSF-CTW. Water (rain and runoff) caught by the
VSSF-CTW, and electricity needed for running the VSSF-CTW recirculation pump were added
to the inputs for case 0. Outputs for case 1 were biogas, solid digestate, sediments, plant biomass,
and treated effluent from the VSSF-CTW. Case 2 considered the entire SPAD-HCTW. The
inputs for this case were the ones for case 1 and rain and runoff caught by the FWS-CTW, and

178

electricity demand by the FWS-CTW recirculation pump. The corresponding outputs were
biogas, solid digestate, sediments, plant biomass, and treated water from the FWS-CTW.

Figure 6. 2. Boundary definitions for conducting the exergy-based calculation of sustainability
for the SPAD-HCTW. Case 0 – the SPAD. Case 1 – the SPAD with the VSSF-CTW. Case 2 –
SPAD with the VSSF-CTW and FWS-CTW.

6.2.3. Data collection
The SPAD-HCTW had been running for 30 months as of August 2015. Data collected
from August 2015 to March 2016 were used for this study – a time period which includes both
phases of low and high precipitation. JMPÂŽ, Version 10.0.0 (SAS Institute Inc., Cary, NC) was
used to classify precipitation data from August 2015 to March 2016. Weeks with precipitation
from 0.00 to 1.91 cm/week were considered as absent to low precipitation period. This dry
condition persisted from January to March 2016. On the other hand, weeks with precipitations
higher than 1.91 cm/week were considered as medium to high precipitation period, which
corresponded to a rainy period from August to December 2015. Data collection is detailed below
for each case.

179

6.2.3.1. Case 0 – the SPAD
Feed consisted of chicken litter, food waste, and treated water. The chicken litter came
from approximately 4,000 laying chickens at the experimental station. Food waste was
transported from a nearby food distribution facility. Wastes mainly consisted of non-commercial
over-ripe or damaged vegetables and fruits, including cucumbers, peppers, avocado, papayas,
pineapples, and tomatoes. Water for preparing the feed came from the treated water produced by
the FWS-CTW. Quantities of chicken litter, food wastes, and treated water were measured and
recorded every week. Chicken litter and food wastes were weighed by an industrial scale
(Romanas Oconi S.A., Costa Rica). The volume of treated water was estimated from pump time.
Samples of the mixed feed were collected and analyzed for COD, TS, TN, and TP at the Water
Quality Laboratory at the Research Center of Environmental Pollution at the University of Costa
Rica (CICA-LCA); whereas, the chemical composition (C, N, P, Ca, Mg, K, S, Fe, C, Zn, Mn,
and B) of the feed was analyzed at the Agronomy Research Center at the University of Costa
Rica laboratory (CIA).
Electricity usage for equipment was calculated based on duration of equipment operation.
The equipment included in case 0 was seven pumps, all pumps but the VSSF-CTW and FWSCTW recirculation pumps, a solid/liquid separator, and a grinder (Figure 6.1). Heat exergy input
was calculated by recording, at a data acquisition unit (DAQ, model CR1000 Campbell
Scientific, Logan, UT), the temperature from the hot water tank and the digester. Environmental
temperature (°C) was recorded every 2 minutes by a weather station located on-site (IMN, 10.00
m N, -84.26 m W) and averaged for each hour. The total daily flow rate of hot water was
calculated from the fixed flow rate of the hot-water pumps and pumping time. Heat from

180

electrical generator was not considered as it was negligible when compared to the heat from the
flat-plate solar collectors and would be relatively constant for all cases.
Biogas production was measured by a flowmeter (EKM-PGM 75 from EKM Metering,
Santa Cruz, CA) and the total amount of biogas per week was recorded. Biogas samples were
collected by using a sampling pump (SKCÂŽ Grab Air, Bag Sampler Cat. No. 222-2301) and
analyzed monthly for methane and carbon dioxide by the Center for Research in
Electrochemistry and Chemical Energy at the University of Costa Rica (CELEQ).
Solid digestate was separated by a solid/liquid separator, collected, and weighed by an
industrial scale (Romanas Oconi S.A., Costa Rica) weekly. Particle size higher than 5 mm in
diameter were separated by the screen. Total solids (TS) content of the solid digestate was also
measured weekly. The chemical composition (C, N, P, Ca, Mg, K, S, Fe, C, Zn, Mn, and B) of
the solid digestate was analyzed at the CIA laboratory.
The amount of liquid digestate generated from the digester unit was estimated from pump
time. Weekly samples were analyzed for COD, TS, TN, and TP. Since the liquid digestate still
contains suspended sediments, the liquid digestate samples were dried, and the chemical
composition (C, N, P, Ca, Mg, K, S, Fe, C, Zn, Mn, and B) of the dried sample was analyzed at
the CIA laboratory.

6.2.3.2. Case 1 – the SPAD with the VSSF-CTW
Case 1 included all inputs and outputs of case 0. The electricity used by the VSSF-CTW
recirculation pump, calculated from duration of pump use, and water input by the rain and runoff
were included as additional inputs (Figure 6.2). Electricity usage for the VSSF-CTW
recirculation pump was calculated based on duration of equipment operation. Precipitation data

181

was collected from a weather station located on-site (IMN, 10.00 m N, -84.26 m W). Runoff was
calculated for the drainage area of 1,500 m2, applying the curve number method (Ward and
Trimble 2004). Harvested plants, suspended sediments collected by the geotextile, and treated
effluent from the VSSF-CTW were the additional outputs for case 1. Three plants, Cyperus
papyrus, Iris graminea, and Canna indica, were planted in the VSSF-CTW during installation in
2012. Reported growth rates for each plant are 4.06 Âą 0.33, 6.76 Âą 0.29, and 8.30 Âą 0.69 g/m2/d
on a dry mass basis (DM) for Cyperus papyrus, Iris graminea, and Canna indica , respectively
(Li, Yang et al. 2013). Each month, one square meter of each plant was harvested, and weighed
by an industrial scale (Romanas Oconi S.A.). Moisture content on the tissues of each plant was
measured for obtaining the dry weight of the harvested plants. Plant samples were also analyzed
at CIA for chemical composition (C, N, P, Ca, Mg, K, S, Fe, C, Zn, Mn, and B). Suspended
sediments caught by the geotextile were collected and weighed by an industrial scale (Romanas
Oconi S.A.). Moisture content was measured to obtain the dry weight, and a sample was
analyzed at CIA for determining its chemical composition (C, N, P, Ca, Mg, K, S, Fe, C, Zn, Mn,
and B). The effluent from the VSSF-CTW was analyzed weekly for COD, TS, TN, and TP
concentrations to determine the exergy of the water. The volume of wastewater entering and
exiting the VSSF-CTW was estimated from pump time. Additionally, the change of volume of
the saturated portion of the VSSF-CTW was measured weekly by measuring the water level at
the well pump.

6.2.3.3. Case 2 – SPAD-HCTW
Case 2 considered the entire SPAD-HCTW system. In addition to all inputs and outputs
used for the case 1, the electricity used by the FWS-CTW recirculation pump and rain and runoff

182

were included in the case 2 (Figure 6.2). Two aquatic plants (Eichhornia crassipes and Pistia
stratiotes) grew in the FWS-CTW. Reported growth rates for Eichhornia crassipes and Pistia
stratiotes are 5.25 Âą 0.28 and 1.12 Âą 0.05 kg DM/m2week. A square meter of both plants was
harvested monthly. The harvested plants were air dried and weighed prior to chemical
composition (C, N, P, Ca, Mg, K, S, Fe, C, Zn, Mn, and B) analysis. The treated water from the
FWS-CTW was analyzed following the same procedure for the liquid effluent from the VSSFCTW in case 1. Similar methods to those in case 1 were used to quantify rain and runoff,
harvested plants, flow, and volume of water within the FWS-CTW.

6.2.4. Analytical methods
Total solids (TS) for chicken litter, food waste, solid digestate, liquid digestate, and
effluent from the VSSF-CTW and the FWS-CTW were determined by the Hach method #8276.
Samples for TS were placed in disposable aluminum dishes (VWRÂŽ, catalog number 25433008), dried for 24 hours in an oven (OV-12A from Blue M, East Troy, WI), and dishes were
weighed in an analytical balance (Ohaus Corporation, Mexico). Moisture content for sediments
collected by the geotextile and harvested plants followed previous Hach method #8276. COD
(Hach method #8000), TN (method MAQA-40), and TP (method MAQA-1) were determined for
the substrate, liquid digestate, and treated effluent from the VSSF-CTW and FWS-CTW. A DRB
200 reactor (Hach product #LTV082.53.40001) and a DR 900 multiparameter handheld
colorimeter (Hach product #9385100) were used to digest and measure COD digestion vials
(high range digestion vials from 0 to 1,500 mg COD/L, Hach kit). TN and TP analyses were
conducted at CICA-LCA. CICA-LCA analyses are based on the Standard Methods for the
Examination of Water and Wastewater (Rice and Bridgewater 2012). The method 5310 B

183

(modified) was followed for TN by a combustion method (TOC-V CSH/CSN from Shimadzu,
Columbia, MD), while the method 4500-P D was followed for TP by a UV-visual
spectrophotometer (Evolution 600 from Thermo Scientific, Madison, WI). Chemical
composition (C, N, P, Ca, Mg, K, S, Fe, Cu, Zn, Mn, and B) of the feed, solid digestate, solids
dried from the liquid digestate, and plant biomass from CTWs were analyzed at the CIA
laboratory. Method SC09-LSF-P06 followed the Dunas method to determine N and C using an
autoanalyzer (Vario Cube from Elementar, Philadelphia, PA), while the Method SC09-LSF-P10
determine P, Ca, Mg, K, S, Fe, Cu, Zn, Mn, and B using a IPC-plasma atomic emission
spectroscopy (Optima 8300 ICP-OES Spectrometer from Perkin Elmer, Spain). Chemical
composition (CH4 and CO2) of biogas samples were determined at the CELEQ laboratory by
using a gas chromatograph (model HP6890 Plus from Hewlett Packard, Littleton, CO) equipped
with a thermal conductivity detector. The column temperature was set up at 250°C and argon
was used as carrier gas. CICA-LCA, CIA, and CELEQ are laboratories accredited by the Costa
Rican Accreditation Institute (ECA).

6.2.5. Exergy calculation
To conduct the exergy-based assessment of sustainability, it was necessary to calculate
the chemical exergy of the feed, solid digestate, solids in the liquid digestate, and harvested
plants. The general method used to calculate the chemical exergy was based on the work done by
Parsapour (2012). Standard chemical exergy values of the different substances and compounds
are based on numbers from Szargut, Morris et al. (1988) and Ayres, Masini et al. (2001).
Equation [6.1] defines the exergy for individual substances,
𝐸𝑐ℎ𝑛 = ∆𝐺𝑓 + ∑𝑒 𝑛𝑒 𝐸𝑐ℎ𝑛𝑒

184

[6.1]

where Echn is the exergy of the substance n [J/kg], ΔGf is Gibbs free energy of formation
of the chemical compound [J/kg], ne is the mole fraction of the eth substance, and Echne is the
standard chemical exergy of the element e [J/kg] (Szargut, Morris et al. 1988, Szargut 1989).
However, in our case was needed to calculate the exergy for or a mixture of substances such as
feed, the solid digestate, solids in the liquid digestate, and harvested plants. Equation 6.2 defines
the exergy for a mixture of substances,
𝐸𝑐ℎ𝑛,𝑚𝑖𝑥𝑡𝑢𝑟𝑒 = ∑𝑖 𝑛𝑒 𝐸𝑐ℎ𝑛𝑒 + 𝑅𝑇0 𝑦𝑒 𝑙𝑛𝑦𝑒

[6.2]

where Echne is the exergy of the substance e pure element of the substance [J/kg], R is the
gas law constant [J/kg K], T0 is temperature [k], and ye is the mole fraction of the eth substance
(Parsapour 2012).
The physical exergy of the heat was calculated using equation 6.3,
𝐸𝑝ℎ = (ℎ − ℎ0 ) − 𝑇0 (𝑠 − 𝑠0 )

[6.3]

where Eph is the physical exergy, h is the specific enthalpy [J/kg], h0 is the specific
enthalpy of saturated liquid water at room temperature [J/kg], s is the specific entropy [J/kg K],
s0 is the specific entropy of saturated liquid water at room temperature [J/kg K], and T0 is the
room temperature [K] (Martin and Parsapour 2012).
The exergy of water in the feed, liquid digestate, rain, runoff, treated effluent, and the
reclaimed water were calculated from the standard chemical exergy of the water is 0.05 kJ/g
(Martin and Parsapour 2012). For comparison, the exergy of the feed and liquid effluent was
measured considering the COD, TN, and TP concentrations. The standard chemical exergies
were 13.6 kJ/g for COD, 17.8 kJ/g NH4-N for TN, and 1.40 kJ/g PO42- for TP (Hellstrom 1997,
MartĂ­nez, Uche et al. 2010).

185

The exergy coefficient of electricity is 1, meaning that “1 kJ of electrical energy
corresponds to an exergy flow of 1 kJ” (Ayres, Masini et al. 2001), while the exergy of the
biogas was calculated from the mass of methane and carbon dioxide. Methane was used for
energy generation at a rate of 55.5 kJ per g of methane, and an overall methane utilization
efficiency of 90% (Aguilar Alvarez, Bustamante Roman et al. 2016). The density of carbon
dioxide is 1.84 kg/m3, with a corresponding chemical exergy of 0.451 kJ/g.
The exergy balance was completed for cases 0, 1 and 2. Net exergy outputs and inputs
were used to compare the performance of these three cases based on the environmental exergy
efficiency (Ρenv,ex) and total pollution rate (Rpol,ex) sustainability indexes, which are defined as
𝜂𝑒𝑛𝑣,𝑒𝑥 =
𝑅𝑝𝑜𝑙,𝑒𝑥 =

𝑜𝑢𝑡𝑝𝑢𝑡𝑠𝑒𝑥
𝑖𝑛𝑝𝑢𝑡𝑠𝑒𝑥

>1

𝑖𝑛𝑝𝑢𝑡𝑠𝑒𝑥 −𝑜𝑢𝑡𝑝𝑢𝑡𝑠𝑒𝑥
𝑜𝑢𝑡𝑝𝑢𝑡𝑠𝑒𝑥

[6.4]
<0

[6.5]

Large values of Ρenv,ex indicate little wasted energy and large quantities of high quality
end products relative to inputs; therefore, high Ρenv,ex values indicate little environmental
degradation (Khosravi and Panjeshahi 2013). Large values of Rpol,ex indicate low removals of
pollutants when converting raw material into high quality end products (Khosravi and Panjeshahi
2013).
Labor is calculated in extended exergy accounting analysis as a factor to measure the
exergy required to produce a good or service (Sciubba 2011). For the SPAD-HCTW,
calculations of sustainability with and without exergy due to labor were calculated to examine its
impact on the exergy balance. The exergy equivalent of labor is 78.7 MJ/h for Mexico (Sciubba
2011). This value was selected for the SPAD-HCTW, as Costa Rica and Mexico have similar
GDP/person (Mexico is ranked 67, Costa Rica 77) and similar human development index (HDI),
with Mexico in the 74th position and Costa Rica in the 69th.
186

6.2.6. Statistical analysis
Statistical analysis of data was conducted using JMPÂŽ, Version 10.0.0 (SAS Institute
Inc., Cary, NC). Inputs and outputs consisted of 25 sampling events, 16 from August to
December 2015 and 9 of them from January to March 2016. Data from inputs or output during
the rainy or dry season were statistically compared at an alpha value of 0.05. Normality of the
data was evaluated using the Shapiro-Wilk normality test (p < 0.05). The t-tests were conducted
in case of normality. Otherwise, the Mann-Whitney test was conducted on data that failed the
Shapiro-Wilk test.

6.3. Results and discussion
6.3.1. Case 0 – the SPAD
The total input exergy for the case 0 was 3,730 Âą 402 MJ/week (Table 6.4). The feed
contributed to 70.4% of the total inputs, as compared to heat (21.1%) and electricity (8.51%).
Two approaches to estimate the exergy of the feed were compared. First, the exergy of the feed
was calculated using equation 6.2, the elemental composition of the feed, and the standard
chemical exergy of the water used for preparing the feed. The chemical composition of the feed
consisted of C, N, P, and other minor components (Table 6.1). Per week, 169 Âą 18 kg of dry
matter of chicken litter and food waste were mixed with 4,611 Âą 183 kg/week of water and the
estimated input exergy was 2,351 Âą 218 MJ/week. The second approach for calculating the
exergy of the feed used previously reported chemical exergies of COD, TN, and TP. The
estimated exergy was 2,624 Âą 230 MJ/week, due to the large concentrations of COD, TN, and TP
in the feed (Table 6.1). There was a difference of approximately 10% between the two

187

approaches. This difference was 273 Âą 448 MJ/week, indicating the standard error of the
estimate exceeded the difference between the two methods. The second approach was used in
this study as it had a larger value than the first one and was simpler to calculate. For the first
approach, two important elements, Oxygen (O) and Hydrogen (H), were not included. Compared
to carbon (34.2 kJ/g), oxygen has very low amount of exergy per gram (0.120 kJ/g) (Parsapour
2012), so no substantial changes would be expected if were included. Hydrogen, similar to
Carbon, is important for biogas production and it has high amount of exergy (236.1 kJ/g)
(Szargut, Morris et al. 1988, Ayres, Masini et al. 2001). Thus, hydrogen could substantially
change the exergy values depending on the amount of hydrogen is in the analyzed substance.
Parsapour (2012) estimated 7% of hydrogen in the raw effluent from a brewing process;
whereas, after anaerobic digestion, the liquid digestate contained 4% of hydrogen. Considering
these hydrogen percentages, the total exergy input in case 0 increased more than two times;
however, the sustainable indexes (Ρenv,ex and Rpol,ex) followed the same trend for each case. The
SPAD alone would not be sustainable, as well as the SPAD plus the VSSF-CTW. Finally, the
SPAD-HCTW would be sustainable.
A direct electric input of 88.2 Âą 2.5 kWh/week was needed to run the equipment for case
0 (Table 6.2). The digester mixing pump had higher electricity consumptions than other
equipment. The high electricity demand was caused by the larger size of the pump and longer
running times that were required to fulfill the mixing demand (Table 6.2). The remaining pieces
of equipment only consumed 29.0% of total energy demand. As 1 kJ of electrical energy
corresponds to an exergy flow of 1 kJ, the total physical exergy input by electricity of the case 0
was 317 Âą 8 MJ/week. The exergy due to electricity could be reduced by optimizing the
operation of the different pieces of equipment at the case 0. For example, if the digester mixing

188

pump only operates 5 minutes per hour (instead of 10 minutes per hour), the electricity
consumption would decrease to 56.8 Âą 2.5 kWh/week, and accordingly, the input exergy due to
electricity (204 Âą 2 MJ/week), which is 36% less than the current exergy. However, reducing
mixing time would impact methane production and treatment performance of wastes. Wang and
Larson (2015) observed that digesters with no and intermittent mixing had significantly lower
methane production rate ( < 1.5 L methane per kg VS destroyed) than a continuously mixed
anaerobic digester ( > 2.0 L methane per kg VS destroyed).
The total physical exergy of heat provided by the solar panels was calculated as the sum
of heat transferred from the solar panels to the hot water storage tank and then to the digester.
The solar heating fluid transfer pump operated 7 hours per day, with a flow rate of 0.960 m3/h.
The amount of water used to transfer heat from the solar panels to the hot water storage tank was
47,040 kg/week. The average temperature of the hot water storage tank was 68.6 ¹ 3.2°C. The
digester heating pump ran 43.9 Âą 15.7 h/week at a flow rate of 0.720 m3/h. In total, 31,624
kg/week of hot water at 54.8 ¹ 5.8°C were needed to heat the digester. The corresponding
physical exergy of the heat was calculated by Equation 6.3, considering a local environmental
temperature of 25°C. The total heat exergy input was 787 ¹ 162 MJ/week.

189

Table 6. 1. Inputs and outputs chemical characterization.
Case 0
COD

TN

TP

C

N

P

Ca

Mg

K

(mean Âą SD, n = 28, mg/L)
Feed
Solid
digestate
Liquid
digestate

S

Fe

Cu

Zn

Mn

B

(n = 3, %)

38,000 Âą 2,700

860 Âą 40

279 Âą 25

36

4.5

1.3

5.0

0.92

4.8

0.82

0.060

0.00

0.010

0.020

0.010

--

--

--

41

2.5

2.1

5.4

0.51

0.41

0.59

0.15

0.010

0.020

0.030

0.00

6,841 Âą 681

1,000 Âą 100

106 Âą 16

26

2.9

1.7

5.3

1.3

10

0.70

0.08

0.010

0.030

0.020

0.010

Mg

K

S

Fe

Cu

Zn

Mn

B

Case 1
C

N

P

Ca

(n = 3, %)
Sediments
collected
by
geotextile
C. papyrus

--

--

--

26

2.9

1.7

5.3

1.3

10

0.70

0.08

0.010

0.030

0.020

0.010

--

--

--

42

1.6

0.10

0.61

0.18

2.8

0.16

0.020

0.00

0.00

0.020

0.00

C. indica

--

--

--

34

4.7

0.45

0.45

0.54

6.8

0.25

0.030

0.00

0.010

0.46

0.00

I. graminea

--

--

--

46

0.95

0.15

0.92

0.13

1.8

0.11

0.030

0.00

0.00

0.010

0.00

Case 2
Eichhornia
crassipes
Pistia
stratiotes

--

--

--

39

1.9

0.12

1.8

0.41

3.9

0.15

0.010

0.00

0.00

0.030

0.00

--

--

--

37

2.0

0.16

4.4

0.60

3.3

0.23

0.080

0.00

0.00

0.040

0.010

190

Table 6. 2. Energy consumption per piece of equipment in the SPAD-HCTW.
Power

Time

Energy

(hp)

(h/week)

(kWh/week)

Solar heating fluid transfer pump

0.17

49

6.0

7 hours/day, 7 days/week

Digester heating pump

0.46

43

15

6.27 hours/day, 7 days/week

Digester mixing pump

3.0

28

62

10 min/hour, 24/7

Feeding tank pump

2.0

0.5

0.75

30 min/week

Solid/liquid separator

1.0

0.85

0.63

10 min/day, 5 days/week

Grinder

5.0

0.50

1.8

30 min/week

Effluent pump

0.50

0.85

0.32

10 min/day, 5 days/week

Feed preparation pump

0.50

0.21

0.080

12 min/week

Equipment

Schedule

SPAD-HCTW exit pump

1.0

1.0

0.75

1 hour/week

VSSF-CTW recirculation pump

0.50

16.6

6.2

2.4 hours/day,7 days/week

FWS-CTW recirculation pump

0.50

16.5

6.1

2.3 hours/day,7 days/week

Exergy of the solar-provided heat varied with changes in the environmental temperature
(Hellstrom 1997). During this study, a daily average temperature of 22.9 ¹ 0.1°C, with daily
temperatures ranging from 19.3 to 25.4°C occurred. Hourly, temperatures range from 15.0 to
33.0°C. On the lowest daily temperature, 144.8 ¹ 26.5 MJ/d would be required to heat the
digester, as compared to 112.5 Âą23.2 MJ/week on the highest daily temperature. Much greater
variations in the exergy required to heat a thermophilic digester would be expected in temperate
climates, as large differences in temperatures occur between winter and summer. In this case,
even though the t-test for two independent variables indicated that hourly temperatures from
January to March 2016 (23.5 ¹ 0.1°C) were significantly higher (p = 5.98x10-14) than the ones
from August to December 2015 (22.6 ¹ 0.1°C), the change in exergy input due to heat would not
substantially impact the balance between inputs and outputs. The approach used herein to
calculate the exergy used to maintain the digester temperature did not account for the heat loss
from pipes and the hot water storage tank, as the loss would be offset by solar heat, resulting in a
net zero impact on the exergy balance.

191

After digestion, the feed was upgraded to electricity (from biogas combustion) and solid
digestate, with a total output exergy of 1,749 Âą 161 MJ/week (Table 6.4). The electricity
represented 99.6% of the total outputs as exergy, while solid digestate was only 0.369%.
Average biogas production of the SPAD was 73.2 Âą 6.7 m3/week with methane contents of 66.5
Âą 1.4% (v/v). The remainder of the biogas was mainly carbon dioxide (33.5 Âą 1.4% (v/v)), which
was considered a loss. Solids collected from solid/liquid separation were 0.467 Âą 0.112 kg dry
matter per week, which were upgraded to fertilizer, yielding an exergy gain of 6.46 Âą 1.67
MJ/week for the baseline case, based on the chemical composition of the solid digestate (Table
6.1).
Liquid digestate and carbon dioxide were considered losses, with a total of 607 Âą 76
MJ/week (Table 6.4). The SPAD discharged 4,578 Âą 178 kg/week of liquid digestate that had
very high concentrations of COD, TN, and TP (Table 6.1). Even though land application of
liquid digestate is often used, this practice was avoided in this study as direct discharge to crops
or water bodies could impair the environment (MartĂ­nez, Uche et al. 2010), especially given the
close proximity of crop fields to surface waters at the research site. Thus, the liquid digestate was
considered a loss. In total, the liquid digestate contained 21.5 Âą 0.8 kg/week of dried solids. The
exergy for the liquid digestate totaled to 459 Âą 17 MJ/week when standard exergies based on
elemental composition were considered. On the other hand, exergy of the liquid digestate was
587 Âą 74 MJ/week when literature-values of chemical exergies for COD, TN, and TP were used
to estimate the liquid digestate exergy (Table 6.4). Similar to the feed, this generalized approach
based on COD, TN, and TP exceeded estimate based on chemical composition by 21.8%. The
difference between the approaches was 128 Âą 91 MJ/week, indicating that the difference between
approaches was not substantial. Similar to the feed, the exergy based on COD, TN, and TP

192

standard chemical exergies was slightly larger, and due to the lack of information, especially the
hydrogen percentage in the liquid digestate, the exergy obtained by the COD, TN, and TP
concentration was considered in the exergy balance. Finally, carbon dioxide was considered loss
due to its global warming potential.
In case 0, Ρenv,ex (0.469 ¹ 0.094) was low and the Rpol,ex (1.13 ¹ 0.42) was large (Table
6.5). The low Ρenv,ex was due to the substantially large input exergy of the feed, which was not
completely converted to electricity; instead approximately 22.4% of the exergy in the feed was
discharged as liquid digestate. Likewise, the large Rpol,ex indicated that a relatively small
percentage of the feed was converted into high quality end products. Even if the liquid digestate
and its sediments were considered useful as fertilizer by land application, the Ρenv,ex would only
increase to 0.627 Âą 0.130, which, while improved, does not indicate sustainability from an
exergetic perspective. Labor had a minimal impact on exergy in Case 0 (Table 6.4 and 6.5).
Labor included 2.5 hours per week to operate the equipment. If labor (196 MJ/week) were
considered, the Ρenv,ex and the Rpol,ex (Table 6.5) did not change substantially, decreasing Ρenv,ex
by 5.01% and increasing Rpol,ex by 9.93%.
More exergetically favorable Ρenv,ex and Rpol,ex values were obtained from January to
March 2016 than from August to December 2015 (Table 6.5). The Mann-Whitney test for two
independent samples indicated that there were not significant differences between inputs based
on wet versus dry period. For example, composition of the feed (as COD, TN, and TP) between
the wet and dry period was not significant different (p = 0.329). Table 6.3 shows that from
January to March 2016, volatile solids (VS) were 64.5 Âą 3.2% of the TS in the feed, whereas
from August to December 2015 the VS were 61.5 Âą 3.8% of the TS, indicating no significant
difference based on season in the percent VS in the feed (p = 0.887). Additionally, the carbon to

193

nitrogen ratio (COD/N) in the feed was higher, but not significantly (p = 0.170), from January to
March 2016 (COD/N = 56.4 Âą 8.4) than from August to December 2015 (C/N = 44.3 Âą 3.0).
However, the power of the statistical comparison for each parameter in Table 6.3 was low,
indicating that it is possible that a difference may have been detected with a greater sample
number. Despite the similarity in inputs, higher concentrations of methane were observed in the
biogas during the dry period, as methane composition in the biogas was higher (70.0 Âą 2.5% v/v)
than during the wet period (64.4 Âą 1.0% v/v; p = 0.0314). However, the total methane produced,
and subsequent electricity production, did not statistically differ (p=0.365).

Table 6. 3. SPAD performance during the rainy and dry period.
Parameters
Food waste
C/N ratio (g COD /g TN)
VS/TS in the feed
(%, g VS / g TS)
Biogas composition (% CH4)
Electricity produced (MJ/week)

August to December
2015
Cucumbers, peppers,
avocados, pineapples, and
tomatoes
44.3 Âą 3.0

January to March
2016
Cantaloupes,
watermelons, and
papayas
56.4 Âą 8.4

p value*

Power

--

--

0.170

0.352

61.5 Âą 3.8

64.5 Âą 3.2%

0.952

0.084

64.4 Âą 1.0 (n=5)
1,606 Âą 226

70.0 Âą 2.5 (n=3)
1,925 Âą 214

0.0314
0.365

0.471
0.162

* Mann-Whitney test for two independent samples.
The sample numbers are 16 for August to December 2015 and 9 for January to March
2016, unless otherwise indicated.
6.3.2. Case 1 – the SPAD with the VSSF-CTW
Case 1 included the VSSF-CTW to treat liquid digestate from the SPAD. Total exergy
input for the case 1 was 4,852 Âą 815 MJ/week (Table 6.4). Electricity to run the VSSF-CTW
recirculation pump (Table 6.2) and precipitation and runoff were the additional inputs. Electricity
represented only 0.639%, while precipitation and runoff represented 22.7% of the inputs.
Consequently, optimization of the operation of the VSSF-CTW recirculation pump did not
represent a substantial reduction in exergy input. While input of water by precipitation cannot be

194

avoided, better isolation of the wetland could have excluded more runoff. Runoff only occurred
during the rainy period, and the 0.40 m tall berms excluded runoff from the wetlands 51% of the
wet period days. If all runoff were excluded, the input exergy by rain and runoff would decrease
from 1,100 Âą 423 MJ/week to 260.7 Âą 94.3 MJ/week, consequently, decreasing the total input for
case 1 to 4,013 Âą 486 MJ/week.
Total exergy output for case 1 was 3,561 Âą 377 MJ/week (Table 6.4). Outputs from case
0 (i.e., methane and solid digestate) corresponded to 49.1% of the outputs from case 1. The
liquid digestate, considered loss for case 0, was treated by the VSSF-CTW. The liquid effluent
from the VSSF-CTW accounted for 44.3% of the exergy outputs for case 1.
During the rainy period, the VSSF-CTW removed 87.6% of TS, 99.0% of COD, 96.8%
of TN, and 99.2% of TP from the liquid digestate, with final effluent concentrations of 543 Âą 63
mg TS/L, 66.5 Âą 12.8 mg COD/L, 34.0 Âą 8.5 mg TN/L, and 0.801 Âą 0.131 mg TP/L. With the
exception of TS, the effluent met the surface water discharge limits for Costa Rica (TS < 50
mg/L, COD < 150 mg/L, TN < 50 mg/L, and TP < 8 mg/L) (MINAE-MSP 2007). The COD
concentrations and the VS to TS ratios (29.3 Âą 3.9% g/g) in the effluent were relatively low,
indicating that solids in the effluent from the VSSF-CTW was mostly of mineral origin. Previous
studies have indicated that mineral sediments in wastewater discharged into land or aquatic
ecosystems, have minimal impact (Avila, Salas et al. 2013); consequently, the effluent would be
expected to have minor impacts on ecosystems if discharged to aquatic ecosystems. However,
other uses of the effluent were prioritized over surface discharged. First, effluent from the VSSFCTW could be used for dilution of feed for the digester, as the reductions in COD, TN, and TP
were greater than 95%. Additionally, the remaining effluent from the VSSF-CTW was also
considered an output as it met standards for use as irrigation water with regards to COD, TN, and

195

TP. As land application of liquid manures is a common practice in Costa Rica, TS concentrations
were not considered a barrier to use of the effluent for irrigation. Liquid manure, which is
classified as manure with no more than 4% TS, has been measured to have concentrations of
23.9 g TS/L (Lorimor, Powers et al. 2004). The TS concentration applied is 44 times larger than
the concentration from the VSSF-CTW effluent in this study; therefore, no adverse impacts from
solids in the effluent on agricultural production are expected. Given the mineral nature of the
solids, they would also not likely represent a source of nutrients in runoff from the fields after
irrigation.
In contrast, during the dry period, only TP (1.88 Âą 0.76 mg/L) met discharge limits for
Costa Rica, as final effluent concentrations were 564 Âą 120 mg COD/L, 2,926 Âą 272 mg TS/L,
and 246 Âą 28 mg TN/L. However, due to dry conditions, this effluent was stored for dilution of
the feed for continued operation of the SPAD. Consequently, the treated effluent from the VSSFCTW was considered an upgraded product, as the effluent from the VSSF-CTW can be used for
any purpose during the rainy months (e.g., irrigation or reuse at the SPAD) or during the dry
period (e.g., reuse at the SPAD). For this reason, its exergy value was estimated by considering
the standard chemical exergy of the water. As the effluent from the VSSF-CTW was upgraded,
the estimation of the exergy based on COD, TN, and TP standard chemical exergies was low,
79.1 Âą 15.4 MJ/week, which was only 5.01% of the exergy calculated with the standard chemical
exergy of the water. Because the effluent was meant for irrigation or reuse at the SPAD, the
exergy value due to COD, TN, and TP was considered an output, but only would increase net
exergy due to the treatment of water at the VSSF-CTW to 1,657 Âą 215 MJ/week for case 1.
The geotextile located at the inlet area of the VSSF-CTW enhanced treatment
performance by providing solid-liquid separation. The geotextile, with apparent pore size of 0.30

196

mm, flow-through rate of 0.102 m3/s/m2, and permittivity of 2.2 s-1, retained sediments in the
liquid digestate. Sediments in the liquid digestate, considered losses for case 0, were collected
and upgraded to fertilizer through composting for case 1 (and 2). On average, 6.46 Âą 1.31
kg/week of dry sediments were retained by the geotextile, representing an exergy of 69.2 Âą 1.9
MJ/week, based on chemical composition (Table 6.1).

Table 6. 4. Exergy-based assessment of sustainability for the SPAD-HCTW.
Case 0
Exergy
Input
(MJ/week)
Feed 2,624 Âą 230
Electricity
Heat
Total

317.4 Âą 8.9
787.8 Âą 162.3
3,730 Âą 402

Labor
Total input
Output
Solid digestate
Electricity

196
3,926
Exergy
(MJ/week)
6.461 Âą 1.666
1,743 Âą 159

--

--

--

--

Total output
Losses
Liquid
digestate
CO2

1,749 Âą 161
Exergy
(MJ/week)

Case 1
Exergy
Input
(MJ/week)
Case 0 3,730 Âą 402
Electricity
Precipitation
and runoff
Total
Labor
Total input
Output
Case 0
Water:
Irrigation
Dilution
Stored
Sediments
Plant biomass:
Harvested
Unharvested
Total output
Losses

587.4 Âą 73.8

CO2

20.39 Âą 1.86

--

Total losses 607.7 Âą 75.7
Total losses
n = 28 for each input, output, and loss.

22.24 Âą 1.49
1,100 Âą 423
4,852 Âą 815
353
5,205
Exergy
(MJ/week)
1,749 Âą 161
1,578 Âą 200
--1,578 Âą 200
69.20 Âą 1.9
164.1 Âą 13.5
4.135 Âą 0.470
160.5 Âą 13.4
3,561 Âą 377
Exergy
(MJ/week)
20.39 Âą 1.86
-20.39 Âą 1.86

197

Case 2
Input

Exergy (MJ/week)

Case 1

4,852 Âą 815

Electricity
Precipitation and
runoff
Total

22.13 Âą 1.50

Labor
Total input
Output
Case 1
Water:
Irrigation
Dilution
Stored
-Plant biomass:
Harvested
Unharvested
Total output
Losses
CO2
-Total losses

1,100 Âą 423
5,974 Âą 1,239
432
6,406
Exergy (MJ/week)
3,561 Âą 377
4,480 Âą 601
1,223 Âą 528
230.5 Âą 9.0
3,026 Âą 64
-25,416 Âą 1,411
943.4 Âą 49.2
24,473 Âą 1,362
33,458 Âą 2,389
Exergy (MJ/week)
20.39 Âą 1.86
-20.39 Âą 1.86

Table 6. 5. Exergy indexes for sustainability.
Parameters
Base case
studies
Including labor
Rainy period
Dry period

Case 0
Index

Value

Case 1
Index

Value

Case 2
Index

Value

Ρenv,ex

0.469 Âą 0.094

Ρenv,ex

0.734 Âą 0.201

Ρenv,ex

5.60 Âą 1.56

Rpol,ex

1.13 Âą 0.42

Rpol,ex

0.362 Âą 0.373

Rpol,ex

-0.821 Âą 0.167

Ρenv,ex

0.446 Âą 0.086

Ρenv,ex

0.684 Âą 0.179

Ρenv,ex

5.22 Âą 1.38

Rpol,ex

1.24 Âą 0.43

Rpol,ex

0.462 Âą 0.383

Rpol,ex

-0.809 Âą 0.166

Ρenv,ex

0.433 Âą 0.111

Ρenv,ex

0.722 Âą 0.214

Ρenv,ex

4.65 Âą 1.36

Rpol,ex

1.31 Âą 0.59

Rpol,ex

0.385 Âą 0.411

Rpol,ex

-0.785 Âą 0.171

Ρenv,ex

0.627 Âą 0.144

Ρenv,ex

0.896 Âą 0.225

Ρenv,ex

9.37 Âą 1.81

Rpol,ex

0.596 Âą 0.367

Rpol,ex

0.115 Âą 0.279

Rpol,ex

-0.893 Âą 0.157

CapEx* ($)**
29,500
36,500
40,000
Net energy
1,521
1,499
1,476
(MJ/week)**
*CapEx: Capital expenditure
**CapEx and Net energy values from Aguilar Alvarez, Bustamante Roman et al. (2016)

The VSSF-CTW provided the conditions (e.g., water, nutrients) suitable for plant growth;
subsequently, the harvested biomass accounted for an exergy output. Chemical exergies for
Cyperus papyrus (103,653 kJ/kg), Iris graminea (15,993 kJ/kg), and Canna indica (220,875
kJ/kg) differed due to the chemical composition of their tissues (Table 6.1). It was observed that
Canna indica possessed higher chemical exergy than Cyperus papyrus and Iris graminea due to
the amount of potassium (K) in the tissues. Potassium has a higher standard chemical exergy
(9,400 J/g) than the other analyzed elements. Carbon also has higher standard chemical exergy
(34,188 J/g), but the composition of the plants was similar. Net exergy output from harvested
biomass totaled to 4.13 Âą 0.47 MJ/week (Table 6.4). Non-harvested plants accumulated a
biomass of 277 Âą 23 g/week for Cyperus papyrus, 473 Âą 20 g/week for Iris graminea, and 581 Âą
48 g/week for Canna indica. This mass stored at the VSSF-CTW represented an output exergy of
160 Âą 13 MJ/week. Similarly to Tang, Fang et al. (2014), plants at the VSSF-CTW were
considered outputs as increased habitat for biodiversity and converted nutrients in the wastewater
into an organized structure in the plants. Opio, Jones et al. (2014) and Zhang, Rengel et al.
198

(2007) have reported different growth rates as the ones reported by Li, Yang et al. (2013). In
Uganda, a wetland planted with Cyperus papyrus had a productivity growth rate from 16.7 to
37.4 g/m2d (Opio, Jones et al. 2014), whereas 33.5 Âą 0.3 g/m2d of Canna indica were measured
at wetland microcosms in Australia (Zhang, Rengel et al. 2007). Results by Li, Yang et al.
(2013) were used in this study to be conservative and avoid huge variabilities of growth rates
obtained at field scale. The conservative assumption of plant growth rates in this study did not
substantially impact the exergy output from biomass. When the less conservative rates were used
for biomass growth for the exergy calculations, the relative output exergy of harvested biomass
was 18.1 Âą 6.2 MJ/week and plant biomass was 722 Âą 1 MJ/week, increasing the output exergy
for case 1 by only 16.0%.
In case 1, the Ρenv,ex increased to 0.734 ¹ 0.201 and the Rpol,ex decreased to 0.362 ¹ 0.373
(Table 6.5). Sustainability indexes changed in a positive way with respect to case 0 due to
inclusion of the VSSF-CTW. The VSSF-CTW treated the liquid digestate, considered loss for
case 0, provided conditions for plant growth, and recovered sediments. Overall, the SPAD plus
the VSSF-CTW was not considered a sustainable system from an exergetic perspective, since the
balance between inputs and outputs was negative. Additionally, the SPAD with VSSF-CTW was
not considered sustainable during the rainy or dry period (Table 6.5). The ability of the VSSFCTW to store and export water in the rainy period (45,953 Âą 3,860 kg/week) was significantly
higher than in the dry period (12,371 Âą 2,510 kg/week; p < 0.001), based on the Mann-Whitney
test for two independent samples. However, the higher output of water (2,297 Âą 193 MJ/week)
was partly cancelled out by water inputs by precipitation and runoff (1995 Âą 656 MJ/week) from
August to December 2015, and the net exergy output of stored water in the VSSF-CTW was
higher during the dry period (618.6 Âą 125 MJ/week) than during the rainy period (302.3 Âą 849.4

199

MJ/week). Thus, more exergetically favorable Ρenv,ex and Rpol,ex values during the dry months
were obtained, in which the VSSF-CTW only stored treated liquid digested (Table 6.5).
In case 1, labor consisted of 2.5 hours per week to operate equipment for case 0 and an
additional two hours for case 1. These two hours per week consisted of weekly activities such as
overseeing discharge of liquid digestate effluent into the VSSF-CTW and harvesting plant
biomass. The geotextile was cleaned of sediments only three times during the period of this
study, which represented as 30 minutes of the two hours per week. Considering labor (353
MJ/week) in case 1, the Ρenv,ex and the Rpol,ex would be 0.684 ¹ 0.179 and 0.462 ¹ 0.383,
respectively (Table 6.4 and 6.5). Labor impacted sustainable indexes negatively by a reduction of
6.80% of Ρenv,ex and an increase of 27.5% of Rpol,ex. In fact, 157 MJ/week were employed for
harvesting plants, while harvested biomass produced a net output of only 4.13 MJ/week. In
contrast, sediments collected at the geotextile produced an exergy of 69.2 MJ/week after
considering the 39.3 MJ/week needed for cleaning the geotextile. It is important to note that
determining the effect of plant harvesting or the geotextile on treatment by VSSF-CTW was
beyond the scope of this chapter; consequently, the above exergy comparisons for biomass
harvesting and geotextile use do not reflect changes in the exergy of the VSSF-CTW effluent
that may results from these practices.

6.3.3. Case 2 – SPAD-HCTW
The total input exergy for case 2 was 5,974 Âą 1,239 MJ/week (Table 6.4). Case 2
included the FWS-CTW to further treat effluent from the VSSF-CTW. Additional inputs to case
2 corresponded to the electricity to run the FWS-CTW recirculation pump, and precipitation and
runoff (Table 6.2). This pump only represented 0.371% of the total exergy inputs in case 2.

200

Similarly to case 1, optimization of operation of the FWS-CTW recirculation pump would not
yield a substantial reduction in the exergy inputs. Similarly, input exergy would decrease to
4,295 Âą 582 MJ/week if the wetland were completely isolated from runoff.
Total output exergy for case 2 was 33,458 Âą 2,389 MJ/week (Table 6.4). Outputs from
case 1 corresponded to 10.6% of the exergy output, while 13.4% was provided by treated water
from the FWS-CTW. Approximately 2.82% of exergy output resulted from harvesting of plant
biomass and 73.1% corresponded to non-harvested plants accumulated at the FWS-CTW.
From August to December 2015, water in the FWS-CTW consisted of effluent from the
VSSF-CTW and precipitation and runoff. During that rainy period, all effluent water quality
parameters except TS met discharge limits for Costa Rica as final effluent concentrations were
243 Âą 23 mg TS/L, 34.3 Âą 2.0 mg COD/L, 2.03 Âą 0.45 mg TN/L, and 0.913 Âą 0.145 mg TP/L.
Similarly to the VSSF-CTW, the effluent from the FWS-CTW was considered an output because
the water can be used for any purpose (e.g., irrigation or reuse at the SPAD). The mineral origin
of the TS (low COD concentration and the low VS to TS ratio (34.7 Âą 5.1%)), the fact that the
FS concentration (158.7 mg/L) was below to the 200 mg TS/L allowed for surface discharge
from livestock activities (MINAE-MSP 2007), and the comparison of TS concentration of liquid
manure applied to the land, justify the decision to consider the effluent from the FWS-CTW as
reclaimed. Reclaimed water from the FWS-CTW allowed the use of the effluent for irrigation
and reuse at the SPAD in the rainy period (16 weeks). This was 2,017 Âą 828 MJ/week and 263.7
Âą 15.0 MJ/week for irrigation and reuse at the SPAD, respectively, out of the net exergy of water
stored during the rainy period (6,514 Âą 618 MJ/week).
In contrast, from January to March 2016, the FWS-CTW only received treated effluent
from the VSSF-CTW and final effluent concentrations of 892 Âą 69 mg TS/L, 288 Âą 57 mg

201

COD/L, 15.7 Âą 6.8 mg TN/L, and 0.632 Âą 0.084 mg TP/L indicated that only TN and TP met
discharge standards in Costa Rica. However, as the FWS-CTW effluent was used for operating
the SPAD during the dry period, this water was valuable and considered as an output in the
exergy balance. In the dry period, 208 Âą 28 MJ/week were used from the net exergy of water
stored at the FWS-CTW (1,767 Âą 456 MJ/week).
As the FWS-CTW effluent can be used for any purpose during the rainy months (e.g..,
irrigation or reuse at the SPAD) or during the dry period (e.g.., reuse at the SPAD), the exergy
value was estimated by considering the standard chemical exergy of the water and totaled to
4,480 Âą 601 MJ/week (Table 6.4). Similarly, for water with low concentrations of COD, TN, and
TP, the estimation of the exergy based on COD, TN, and TP standard chemical exergies would
be low, 61.8 Âą 8.17 MJ/week, and the net exergy output due to the treated water at the FWSCTW would increase to 4,541 Âą 609 MJ/week.
Biomass harvested at the FWS-CTW corresponded to 2.82% of the exergy outputs in
case 2. The exergy output from harvesting of floating plants depended plant densities at the time
when plants were harvested. These aquatic plants, especially the Eichhornia crassipes, spread
rapidly across the surface in the water, and the accumulated biomass stored at the FWS-CTW
totaled to 24,473 Âą 1,362 MJ/week, which was the 73.1% of the exergy outputs in case 2. In this
study, reported growth rates for Eichhornia crassipes and Pistia stratiotes were 5.25 Âą 0.28 and
1.12 Âą 0.05 kg DM/m2week, which were 12.5 to 37.5 and 2.7 to 8.00 times higher than the
growth rates reported by Gutierrez, Ruiz et al. (2001) and Gupta, Roy et al. (2012) (from 0.14 to
0.42 kg/m2week). If their maximum growth rate (0.42 kg/m2week) for both Eichhornia crassipes
and Pistia stratiotes were used, the exergy output due harvested biomass at the FWS-CTW
would have decreased to 60.8 and 67.5 MJ/week for Eichhornia crassipes and Pistia stratiotes,

202

respectively. The plant biomass stored at the FWS-CTW would decrease to 1,644 MJ/week for
Eichhornia crassipes and 1,824 MJ/week for Pistia stratiotes. Considering this conservative
growth rate, the total output exergy for case 2 would decrease to 11,637MJ/week. However, the
balance between exergy inputs and outputs would remain positive. In terms of plant growth rate,
the FWS-CTW produced 228 Âą 38 times more exergy than the VSSF-CTW. Reported growth
rates by Opio, Jones et al. (2014) and Zhang, Rengel et al. (2007) for Cyperus papyrus (16.7 to
37.4 g/m2d) and Canna indica (33.5 Âą 0.3 g/m2d) shows that biomass production is lower by 11
times the growth rates reported by Gutierrez, Ruiz et al. (2001) and Gupta, Roy et al. (2012)
(from 0.14 to 0.42 kg/m2week) for plant type, which support the observation that FWS-CTW can
obtain higher exergy output from plant biomass than VSSF-CTW systems.
For case 2, the Ρenv,ex was 5.60 ¹ 1.56 and the Rpol,ex was -0.821 ¹ 0.167 (Table 6.5).
Sustainable indexes were higher than 1 and lower than 0 for the Ρenv,ex and the Rpol,ex,
respectively. This study was conducted to answer the question is the SPAD-HCTW exergetically
sustainable? Thus, for the period of this study, the SPAD-HCTW was sustainable from the
exergetic point of view. The exergy outputs increased by 1,800% and 800% with respect to case
0 and case 1, respectively, because of the capacity of the HCTW to treat and store water and
produce biomass (harvested and stored plants).
Addition of the FWS-CTW to the system positively impacted the exergy balance between
inputs and outputs, mainly because the greater storage capacity of the FWS-CTW allowed for
storage of runoff and precipitation that would otherwise be considered a loss of exergy in the
system to be converted into useful water (an exergy output). Because the filter media (32% of
porosity) occupies storage volume in the VSSF-CTW, the FWS-CTW stored more water, thus
higher output exergy was achieved with the FWS-CTW in terms of water storage useful for any

203

purpose (e.g., irrigation or reuse at the SPAD). The FWS-CTW processed 130,287 Âą 12,367
kg/week of water during the rainy months, in contrast to 35,351 Âą 9,136 kg/week of water during
the dry months, volume of water that were 2.84 and 2.14 times the volume stored at the VSSFCTW. The exergy output of the stored water was significantly higher in the rainy period (6,514 Âą
618 MJ/week) than in the dry period (1,325 Âą 456 MJ/week), with a p value <0.001. However,
the dry period obtained better Ρenv,ex and Rpol,ex values than during the rainy for case 2 (Table 6.5).
Input during the rainy period was higher than during the dry period due to the input of rain and
precipitation (1,995 MJ/week versus zero in the dry period), but with respect to the input, higher
output was obtained during the dry period (input=3,124 MJ/week, output=29,237 MJ/week) than
during the rainy period (input=7,755 MJ/week, output=36,074 MJ/week) due to the constant
growth of plants in both periods.
Labor at the FWS-CTW was one hour per week to harvest plants. In addition to the four
and half hours per week from case 0 and case 1, the total exergy by labor for the case 2 was 432
MJ/week. Labor decreased the Ρenv,ex by 6.74% to 5.22 ¹ 1.38 and increased the Rpol,ex by 1.57%
to -0.809 Âą 0.166 with respect to the case 2 values. Even when exergy due to labor was
considered in case 2, the SPAD-HCTW was considered borderline sustainable, in contrast to
case 0 and 1. In fact, the exergy for harvesting the plants was only 8.34% of the exergy gained by
the harvesting of plants (943 Âą 49 MJ/week).
Table 6.5 includes the capital expenditure (CaPex) and net energy produced for the three
cases (Aguilar Alvarez, Bustamante Roman et al. 2016). Addition of the HCTW to the SPAD
increased costs by 23.0% and 35.0% for case 1 and case 2, respectively, while the net energy
decreased by 1.24% and 2.42% for case 1 and 2, both with respect to case 0. These two negative
impacts in costs and energy consumption can be offset by potential sustainability of the SPAD-

204

HCTW. The SPAD-HCTW had net revenue of $2,436 per year and net energy of 1,476
MJ/week (Aguilar Alvarez, Bustamante Roman et al. 2016). However, this basic energy and
expenditure analysis did not provide insight into the environmental benefits that were observed
from inclusion of the HCTW as a treatment system for the SPAD digestate. By this exergy
assessment of sustainability, it was seen that energy production, reuse of agricultural residues,
and water treatment contributed to environmental benefits accrued by the SPAD-HCTW.

6.4. Conclusions
Sustainability is a complex and broad concept, which varies in interpretation depending
on the discipline and context. The United Nations define sustainable as “the process that
constrains resource consumption and waste generation to an acceptable level, makes a positive
contribution to the satisfaction of human needs, and provides enduring economic value to the
business enterprise”. The SPAD-HCTW constrained resource consumption and waste generation
while producing energy, fertilizer, and treated water. These outputs represent positive
contributions to the satisfaction of the human needs. The exergy-based assessment of
sustainability from two indexes (Ρenv,ex and Rpol,ex) indicated net environmental benefits from the
SPAD-HCTW from a thermodynamic perspective. Future work is needed to quantify additional
economic, social, and environmental aspects of sustainability with regards to the SPAD-HCTW.
From an exergy perspective, the baseline SPAD (case 0) was not sustainable, based on
Ρenv,ex and Rpol,ex. By including the HCTW as treatment for digestate from the SPAD, the system
became thermodynamically sustainable, not only producing energy and fertilizer, but also treated
water - an output that was not achievable with the SPAD alone. Due to a positive exergetic
balance (inputs < outputs), Ρenv,ex and Rpol,ex had values higher than 1 and lower than 0,

205

respectively. This result was achieved by conversion of biomass residues and agricultural
wastewater (e.g., food waste, chicken litter, and treated water) into high quality end products
(energy, fertilizer, and treated water) when combining solar technology with anaerobic digester
and HCTW treatment.
For the entire SPAD-HCTW, energy, treated water, and plant biomass were the main
outputs from the conversion of raw material into high quality end products. The SPAD (case 0)
produced energy from current problematic wastes in Costa Rica and can be offered as an option
for alleviating waste problems and energy demands. As a post-treatment process to treat water
and produce plant biomass, the HCTW treated the liquid digestate and avoided potential surface
and ground water contamination by direct discharge to land or surface water. In addition, the
HCTW, especially the FWS-CTW, became a source of water which had higher storage capacity
than the VSSF-CTW. This storage capacity turned in higher exergy output by water as the FWSCTW had higher holding capacity during the rainy period. Interestingly, the plants at the FWSCTW showed higher growth rate than the plants at the VSSF-CTW, with a consequently higher
exergy output by plant biomass.
The exergy-based assessment of sustainability was not impacted by the precipitation. The
SPAD alone did not show differences between the rainy and dry period as no elements of the
SPAD were affected by rain. However, precipitation and runoff were included as inputs in case 1
and 2. Positively, the VSSF-CTW and FWS-CTW had enough storage capacity to hold water
from precipitation and runoff. However, no differences between the rainy and dry months were
found as the exergy output of water stored in both the VSSF-CTW and the FWS-CTW were
partly cancelled out by water inputs by precipitation and runoff. In general, more exergetically
favorable Ρenv,ex and Rpol,ex values during the dry months were obtained due to better digestion

206

performance (case 0), storage of water for future use, and biomass production (harvested and
non-harvested plants) (case 1 and 2).
Biomass production marked a difference in the exergy balance, because independently of
the period, the growth rate was constant and, in the particular case of Eichhornia crassipes, fast.
Plants converted the nutrients in the wastewater into an organized structure (steam, leaves,
flowers, etc) and provide habitat, which in turn increase biodiversity (Jørgensen 2006, Tang,
Fang et al. 2014). Biomass harvested was valuable as a source of nutrients after composting;
however, the exergy output from nutrients was not remarkable. Other processes could be
explored for utilizing the biomass, especially the Eichhornia crassipes, which had a fast growth
rate. Eichhornia crassipes could be used as a substrate for anaerobic digestion as long as a
pretreatment process removes the lignin (Bharati and Kalamdhad 2016). Recent processes such
as electrohydrolysis pretreatment (Barua, Raju et al. 2017), thermal pretreatment (Barua and
Kalamdhad 2017), microwave pretreatment (Budiyono, Sumardiono et al. 2015), and older
process such as chopping and drying (Moorhead and Nordstedt 1993), could be studied for
creating another the loop in the SPAD-HCTW. For example, Budiyono, Sumardiono et al.
(2015) estimated that the microwave pretreatment of Eichhornia crassipes at 560 W for seven
minutes double the biogas production to 75.12 mL biogas/g TS from fresh when compare to the
non-pretreated Eichhornia crassipes (37.56 mL biogas/g TS).
This study is the base for starting a protocol for determining sustainability of similar solid
and liquid wastes management in exergetic terms. It was identified that for determining the
exergy of the wastewater with high concentrations of contaminants (COD, TN, and TP) the
approach developed by Tai and Matsushige (1986) and applied by Hellstrom (1997) and
MartĂ­nez, Uche et al. (2010) is easier to apply than the calculation by the chemical composition

207

of the wastewater. It is simpler to calculate the COD, TN, and TP concentrations in the
wastewater and use the standard chemical exergy of COD, TN, and TP than to determine the
elemental composition. For effluents with low concentration of COD, TN, and TP, the use of the
COD, TN, and TP standard chemical exergy will be low. For these cases, the exergy of the water
can be calculated using the standard chemical exergy of the water, and depending the purpose of
the treated water (e.g., irrigation, reuse at the SPAD), could include the exergy value due to the
COD, TN, and TP concentration.
In conclusion, the SPAD-HCTW (case 2) was exergetically sustainable. Some economic,
social, and environmental benefits can be deduced from this study. Economically, energy,
fertilizer, and water reclamation could represent savings for families at rural areas. Socially,
savings could cover other important needs (e.g.: education, health, etc.) and the SPAD-HCTW
became another solution to provide water treatment. Environmentally, treatment of biomass
residues and agricultural wastewater could avoid the emission of greenhouse gases, as well as the
impairment groundwater and surface water. However, to better understand the sustainability
benefits of the SPAD-HCTW, more analyses with additional economic, social, and
environmental metrics are needed.

6.5. Acknowledgment
This research was supported by the U.S. Department of State West Hemisphere Affairs
(S-LMAQM-11-GR-075). The authors would like to acknowledge the assistance and support of
the following laboratories at the University of Costa Rica: Agronomy Research Center (CIA),
Water Quality Laboratory at the Research Center of Environmental Pollution (CICA-LCA), and
Center for Research in Electrochemistry and Chemical Energy (CELEQ). The authors would

208

also like to thank the staff at the Fabio Baudrit Experimental Station (EEAFBM) as well as the
researchers from the UCR project entitled “Electric energy production through anaerobic
digestion optimization of agricultural organic wastes and treatment of its effluents”. Also, the
Department of Eco-physiology at EEAFBM and the Meteorological Institute of Costa Rica for
providing environmental data.

209

CHAPTER 7: CONCLUSIONS

7.1. Biomass residues and wastewater: the problem becoming an opportunity for
sustainable development
Costa Rica is striving to be the first neutral carbon country for 2021. One front to carbon
neutrality is generation of electric energy using renewable resources, replacing fossil fuel based
electricity. Excluding transportation, which depends 100% on fossil fuels, Costa Rica generates
mainly electric energy from hydric (75%), geothermic (12%), wind (10%), solar (0.01%), and
biomass (0.73%) resources (ICE 2016). Only 1.8% of electric energy is produced from fossil
fuels. Even though the installed power thermic capacity generation in Costa Rica is 200 MW
(8% of the total power capacity in the country), the country only uses this thermic plant as a
backup (Leandro 2011). Indeed, in 2016, the country produced electric energy from renewable
resources for 271 days in a row (Raedle 2017). Precipitations and rivers fed reservoirs that
allowed a continuous generation of renewable electric energy, even when El NiĂąo-Southern
Oscillation (ENSO) impacted the country with less precipitation (IMN 2016). This is a warning
sign; the reliance of Costa Rica on hydroelectric energy can turn negative as climatic phenomena
(e.g., ENSO) can bring severe droughts. For example, in 2013, the water level of Arenal Lake,
the main artificial pluriannual reservoir for hydric power generation in Costa Rica, almost
depleted to a non-productive point (Aguero 2013). Thus, to prevent fossil utilization for electric
energy production, other renewable electric energy generation approaches need to be adopted
and strengthened. For example, Coto (2013) demonstrated that agricultural organic residues have
a potential electric energy generation of 600 MW; this would be 23% of the total power capacity
in the country.

210

In addition to its carbon neutrality goal, Costa Rica is marketing itself as a green country,
committed to nature. This is an economic strategy for attracting investors to green development,
an essential for Costa Rica. One front is the protection of the environment (e.g. national parks
and private refuges) and the electric energy generation from renewable resources.
In addition, Costa Rica promotes the optimization of resource utilization (e.g., minimum
utilization, reutilization, and recycle) in any economic activity. For example, business should use
less energy, water, and materials in production and operational lines. These efforts are awarded
by the Ecological Blue Flag. However, the majority of the businesses follow a linear economy
model (make, use, dispose), in which there is awareness in the use of resources, but no awareness
of the management of sub-products (e.g., solid and liquid wastes). In fact, as stated by Calvo
(2014), “Costa Rica has invested greater efforts to conserve bio-diversity than to pollution
control and waste disposal”. Consequently, the environment, in particular surface and
groundwater resources, is impaired, and contradictorily, this “green” country has a severe
sanitation and water pollution problem.
Moreover, there is a global challenge. More food, energy, and water are demanded by a
growing population, while current environmental degradation and climate change effects restrict
the access to these resources (FAO 2014). The food-energy-water nexus is a conceptual tool for
balancing the natural resource supply and demand for achieving sustainable development (FAO
2014, Biggs, Bruce et al. 2015). Once the linkage between these three resources is recognized in
a particular economy, integrated systems can be implemented to optimize resource supply and
consumption. The key is to create close-looped systems to convert linear into circular economies
(Agrocycle 2017). In contrast to a linear economy, a circular economy considers residues as
resources. In turn, circular economy alleviates the demand of resources and reduce waste

211

disposal. Figure 7.1 depicts how integrating a treatment system, the linear model can be
converted into a circular one. The treatment system processes the residues to convert them into to
resources, decreasing resource demand from the environment and decreasing waste disposal into
the environment.

Figure 7. 1. Linear versus circular economy (Own creation).

Even with Costa Rican efforts toward carbon neutrality and green development, there are
gaps have not been considered, while environmental degradation and climate change are
restricting even more resources. Thus, there is an opportunity to transform current linear
economies into circular ones by the integration of innovative treatment systems. Innovative
treatment systems can partially solve the water pollution problem at Costa Rica and produce
energy from biomass resources, while protecting the environment and decreasing the demand
from the current electric grid. This clearly will strengthen Costa Rican green development and
212

the progress towards carbon neutrality. In particular, biosystems engineering can focus on agroindustry and agricultural activities in rural areas of Costa Rica. The idea is to provide a
decentralized self-sufficient, close-looped, organic waste treatment system technology as a
solution to linear development in agro-industry and agricultural activities located in rural areas of
Costa Rica.
The present document proposed one approach for solving the problem of water pollution
due to agro-industrial and agricultural waste disposal in rural areas of Costa Rica. The proposed
integrated system consisted of three well-known technologies: solar thermal collection,
anaerobic digestion, and constructed treatment wetland. All together created a decentralized selfsufficient, close-loop, organic waste treatment system, which in turn, provided energy, fertilizers,
and treated water. The solar-powered anaerobic digestion and hybrid constructed treatment
wetland (SPAD-HCTW) utilized the residues from current linear production systems to convert
them into resources and create a circular, close-loop, system. The present dissertation
demonstrated the technical performance and sustainability of the SPAD-HCTW in Costa Rica
(Figure 7.2).

213

Figure 7. 2. Closing the loop for water, energy, and food (Own creation).

Technical performance of a solar-powered waste utilization and treatment system in
Costa Rica, chapter 3, demonstrated that the solar thermal collector unit provided sufficient
energy to maintain thermophilic temperature in a 20 m3 CSTR anaerobic digester. Under
thermophilic conditions, the CSTR anaerobic digester converted organic wastes (e.g., food waste
and chicken litter) into energy and fertilizers. The entire treatment system was self-sufficient and
surplus energy was obtained. In terms of treatment of wastewater, it is well-known that under
thermophilic conditions pathogens are killed; however, the liquid digestate still had high
concentrations of organic matter, solids, and nutrients. Then, to assure wastewater treatment, the

214

vertical flow subsurface (and free water surface) constructed treatment wetland treated the
digestate. Treated water was reutilized either for irrigation or for diluting the feedstock. This
final step assured the recovery of resources (energy, fertilizers, and water) and converted a linear
model into a circular, close-loop, system. In general, 263 MJ renewable energy, 28 kg nitrogen
and phosphorus fertilizer, and 550 kg of treated water were generated from 863 kg of mixed
animal and food wastes. The net revenue considering electricity and fertilizer was $2,146
annually. The payback period for the system was estimated to be 21 years; however, a sensitivity
analysis demonstrated that through optimization, the payback period can be reduced to 9 years.
The proposed system is innovative compared to decentralized systems developed at farms
in Costa Rica. In Costa Rica, the Costa Rican Electricity Institute (ICE) leads the installation of
anaerobic digesters. ICE promotes anaerobic digestion for producing energy from manure (cow,
swine, and chicken litter), while treatment of wastewater is not a priority. In general, anaerobic
digesters are tubular, plug-flow bags (e.g., tubular polyethylene bag digesters) for small-scale
farms, usually 10 to 15 heads. In addition, covered lagoon anaerobic digesters are used for larger
farms (e.g., Noble farm: 120 cows; and Kafur: 4,000 swine). These systems are meant to be
mesophilic (e.g., 35°C); however, mesophilic temperature is rarely reached as the systems
depend on ambient temperatures (Kinyua 2015). Even though the production of biogas is
achieved, the digestion process is neither efficient reducing solids, organic matter nor for killing
pathogens. Then, digestate is not post-treated and it is simple discharged. Usual practice is land
application of the liquid digestate; however, there is a potential health impact due to the
transmission of pathogens to food and water, as well as an impairment to ground and surface
water due to high concentration of nutrients (Kinyua 2015).

215

Performance of the hybrid constructed treatment wetland treating digestate effluent in
Costa Rica, chapter 4, demonstrated that the HCTW had sufficient water storage capacity and
achieved a treatment performance that allowed the reuse of the treated water in other activities
(e.g., irrigation and reuse at the SPAD). The hybrid configuration was important to overcome the
disadvantages each individual wetland had. The VSSF-CTW received high strength liquid
digestate and demonstrated high reduction of solid and organic matter, as well as high
transformation of ammonium to nitrate. Then, the FWS-CTW served as a polishing step for the
effluent from the VSSF-CTW, in particular, in the transformation and removal of nitrate through
denitrification. Finally, the study of natural treatment systems, such as the HCTW, require larger
campaign evaluation (years) for establishing central tendencies in the treatment. Thus, during
this evaluation (from August 2015 to March 2016), it was not possible to represent a narrow
central tendency using loading charts or contaminant removal models, given the variability in the
digestate and environmental conditions.
The HCTW is innovative compared to constructed treatment wetlands installed in Costa
Rica. To our knowledge, this is the first HCTW treating high strength wastewater from
agricultural residues in Costa Rica. Few wetlands have been installed in Costa Rica. These
systems were built to treat grey and black water (Dallas, Scheffe et al. 2004, Alfaro, Perez et al.
2013, PĂŠrez, Alfaro et al. 2013). For rural areas, this dissertation, as well as studies by Dallas,
Scheffe et al. (2004), PĂŠrez, Alfaro et al. (2013), and Alfaro, Perez et al. (2013), demonstrated
that the sanitation and water pollution problem at Costa Rica can be overcome by implementing
decentralized wastewater treatment systems such as CTWs. In fact, the implementation of
decentralized wastewater treatment systems is recommended by Naik (2014). Naik (2014)
indicated that the high capital investment (e.g., collection system and the treatment plant) makes

216

the implementation of centralized system in rural regions unfeasible. In addition, centralized
treatment systems have shown to be unmanageable in some places due to technically skilled
manpower for operation and maintenance (Dallas, Scheffe et al. 2004, Avila, GarcĂ­a et al. 2016).
For example, the Costa Rican Water and Sewage Institute (AyA) reported in 2003 that
centralized wastewater treatment plants at the cities of Heredia, Alajuela, and Cartago, are not
operating, and wastewater is instead directly discharged into rivers (Araya, Araya et al. 2003).
The situation has not changed at all in recent years, and sewage collectors in those cities simple
discharge the wastewater into the rivers (Calvo 2014).
Infiltration measurements, chapter 5, demonstrated that after three years of continues
operation, the VFSS-CTW was not clogged. Even with the high strength characteristics of the
liquid digestate, strategies implemented to prevent clogging have positively impacted in the
longevity of the filter media. The geotextile membrane removed solids, allowed the recovery of
nutrients (solids retained by the membrane), and avoided the direct contact of solids with the
filter media. The implementation of a geotextile membrane in the inlet area provided a low-cost,
low-maintenance approach to improving the performance of the VSSF-CTW. The effective
treatment surface area only received treated water through the recirculation system. Under this
condition, the estimated void space at the filter media was 20 of 30 m3. At the field scale, no
differences in volatile solid accumulation, root development, and infiltration rates in the filter
media were found based on plant type. However, in laboratory-scale columns, columns planted
with Canna and Cyperus exhibited different infiltration rates, despite similar volatile solid
accumulation and root development, indicating that root morphology of Canna could be
favorable to infiltration.

217

Exergy-based assessment of sustainability of a solar-powered anaerobic digestion and
hybrid constructed treatment wetland system to treat agricultural wastes in Costa Rica, chapter 6,
demonstrated that SPAD-HCTW was a tool for sustainable development. Sustainability is a
broad concept, and herein was limited to a system that constrains resource consumption and
waste generation to an acceptable level, makes a positive contribution to the satisfaction of
human needs, and provides enduring economic value to the business enterprise (UN 2015). The
SPAD-HCTW processed residues (e.g.: wastes) and converted then into resources (e.g.: energy,
fertilizers, and treated water). Thus, the integration of the SPAD-HCTW as a decentralized selfsufficient, close-looped, organic waste treatment system into agro-industry and agricultural
activities located in rural areas of Costa Rica would:
•

Constrain waste generation. The SPAD-HCTW treats waste from agro-industrial and
agricultural wastes.

•

Constrain resource consumption. Treating wastes, the SPAD-HCTW generates energy,
fertilizers, and water, that can be reused.

•

Satisfy human needs. The SPAD-HCTW provides wastewater treatment, which in turn
brings environmental and public health benefits. Environmentally, the electric generation
from biomass can increase, reducing the need of fossils and strengthening the neutral
carbon process to 2021. In addition, the SPAD-HCTW alleviates the severe water
pollution problem at Costa Rica.

•

Economic value. Saving from energy generation and fertilizers. Per day, the SPADHCTW generated 263 MJ renewable energy from 863 kg of mixed animal and food
wastes, which produced net revenue of $2,146 annually from electricity and fertilizer.

218

From the technical point of view, the SPAD-HCTW was a sustainable technology that
produced energy, fertilizers, and treated water. Integration of technologies was key to overcome
the disadvantages of the individual technologies. The higher energy requirement of thermophilic
anaerobic digestion can be supplied by solar thermal collectors. By itself, the solar thermal
collector represents an unsteady energy flow for solar power generation. Instead, the
thermophilic anaerobic digestion converts low-density and inconsistent solar energy (as heat)
into a relatively dense and reliable biochemical energy source – methane. Together, anaerobic
digestion and solar thermal collection technology provide energy, fertilizers, and reduce
greenhouse gas emissions; however, water reclamation is not possible. Thus, a post-treatment
system is needed for wastewater treatment. The HCTW demonstrated to the capability of treating
strength wastewater. The treatment performance depends on the wetland status, and preventive
strategies prevented clogging, one of the commonest problems of wetlands.

7.2. Future work
The SPAD-HCTW can become one approach to solving Costa Rica’s water pollution
problem, while advancing Costa Rica toward carbon neutrality and the green development.
However, there are still aspects to overcome in the future related to this technology.
There is a need to optimize the SPAD-HCTW for reducing costs to make this approach
more economically attractive. It is clear that the SPAD-HCT provides environmental benefits;
however, the large payback period is not attractive. Future research should focus on modeling
and validation of the solar thermal collection unit and on minimizing range of temperature
variation inside the digester. Modeling and validation can be based on energy demand by the
anaerobic digester and average annual environmental temperature and irradiance at the site for

219

optimizing solar thermal collector area (Suryawanshi, Chaudhari et al. 2010). In the tropics, this
will be a useful tool for designing solar-powered digesters based on the organic waste and region
where the system will be installed. For example, in China, Hassanein, Qiu et al. (2015) modeled
and validated a solar greenhouse to surround the anaerobic digester with the intention to maintain
minimum temperature for anaerobic digestion. The validated model allowed determination of
geographic zones to house the proposed solar greenhouse unit in China. Then, controls can be
implemented (e.g., PIDs) for promptly turning on and off heat transfer pumps to avoid large
temperature variation inside the digester (Alkhamis, El-khazali et al. 2000).
Moreover, the experience gained in the operation of the SPAD narrowed topics for future
studies. For example, more studies are needed to evaluate parameters such as mixing frequency
and intensity (Zabranska, Dohanyos et al. 2002, Bombardieri, Espinosa-Solares et al. 2007,
Suryawanshi, Chaudhari et al. 2010), feed frequencies (Bombardieri, Espinosa-Solares et al.
2007), and TS content in the feed (Espinosa-Solares, Valle-Guadarrama et al. 2009).
More data collection at the HCTW is needed to narrow central tendencies in treatment
performance of the wetlands. In addition, the geotextile membrane implemented in the VSSFCTW was important, and more studies are needed to characterize its impact on treatment
performance of the VSSF-CTW. Some questions for future research are:
•

Did the geotextile create short circuiting and dead points within the VSSF-CTW?
Treatment of the primary influent was assumed to occur predominantly beneath the
geotextile, leaving most of the treatment area for treatment of the recirculated
wastewater. However, the mass balance indicated that diffusion of solids beneath the
geotextile likely occurred.

220

•

Was there any relationship between sediment accumulation in the geotextile and the
effluent concentration from the VSSF-CTW? As the sediments accumulated in the
geotextile, more sediments were filtered out, likely improving sediment removal.

•

How much could the main treatment area be decreased due to the geotextile?
In addition, the geotextile membrane could prevent clogging of the sand media of the

VSSF-CTW. From this observation, a principal question is proposed for further research and
optimization of the vertical-flow wetland: How do the solids distribute as the wastewater moves
downward through the filter media below the geotextile membrane? A tracer study is proposed
for determining dynamics of the solids across that portion of the wetland. Distribution, retention
of solids, and time to clogging can be impacted by precipitation and flooding events and the real
condition of the filter media beneath the geotextile membrane is currently unknown. In addition,
there is a need to identify morphological traits of root plants, particularly Canna plants which did
not adversely affect infiltration in the full-scale wetland.
Finally, environmental benefits can justify special incentives by the government to
promote this technology in the country. As depicts in Figure 7.3, this dissertation has shown the
environmental benefits from the SPAD-HCTW; however, an extended sustainable study
considering the SPAD-HCTW as part of the surrounding ecosystem and community needs to be
conducted to elucidate how the decentralized self-sufficient, close-looped, organic waste
treatment system also brings economic and social benefits. Holistic approaches such as the
SPAD-HCTW can partly alleviate the water pollution problem; however, as concluded by Calvo
(2014) the country needs to politically prioritize sanitation and water pollution problems. Until
that happens, academic researchers need to promote and encourage the target population to

221

implement this approach by using the best available weapon in Costa Rica: environmental
education.

Economic

SPADHCTW
Environmental

Society

Figure 7. 3 . The SPAD-HCTW as a sustainable approach for the sanitation problem in Costa
Rica.

222

REFERENCES

223

REFERENCES

Abou-Elela, S. I. and M. S. Hellal (2012). "Municipal wastewater treatment using vertical flow
constructed wetlands planted with Canna, Phragmites and Cyprus." Ecological
Engineering 47: 209-213.
Agrocycle. (2017). "A blueprint and EU policy-forming protocol for the recycling and
valorisation of argi-food waste." 2017, from http://www.agrocycle.eu/.
AgStar (2010). Continued growth in digester projects, US-EPA.
Aguero, M. (2013). "Ruinas de pueblos emergen ante baja de embalse Arenal." 2017, from
http://www.nacion.com/nacional/comunidades/Ruinas-pueblos-emergen-embalseArenal_0_1346265404.html.
Aguilar Alvarez, R. E., M. Bustamante Roman, D. Kirk, J. A. Miranda Chavarria, D. Baudrit, J.
F. Aguilar Pereira, W. Rodriguez Montero, D. Reinhold and W. Liao (2016). "Technical
and economic feasibility of a solar-bio-powered waste utilization and treatment system in
Central America." Journal of Environmental Management 184, Part 2: 371-379.
Aitken, M. D., M. D. Sobsey, M. Shehee, K. E. Blauth, V. R. Hill, J. B. Farrell, S. P. Nappier, G.
W. Walters, P. L. Crunk and N. Van Abel (2005). "Laboratory evaluation of
thermophilic-anaerobic digestion to produce class a biosolids. 2. Inactivation of
pathogens and indicator organisms in a continuous flow reactor followed by batch
treatment." Water Environment Research 77(7): 3028-3036.
Alfaro, C., R. Perez and M. Solano (2013). "Wastewater treatment by artificial wetlands in the
Museum Popular Culture National University." Tropical Journal of Environmetal
Sciences 45(1).
Alkhamis, T. M., R. El-khazali, M. M. Kablan and M. A. Alhusein (2000). "Heating of a biogas
reactor using a solar energy system with temperature control unit." Solar Energy 69(3):
239-247.
Alkhamis, T. M., R. El-Khazali, M. M. Kablan and M. A. Alhusein (2000). "Heating of a biogas
reactor using a solar energy system with temperature control unit." Solar Energy 69(3).
Allen, R., L. Pereira, D. Raes and M. Smith (1998). Determination of ETo. Crop
evapotranspiration. Guidelines for computing crop water requirements. FAO irrigation
and drainage paper 56. Rome, FAO.

224

Anderson, K. and S. Sheffield. (2015). "Water reuse." Retrieved 8/22/2016, 2016, from
http://www.sheffy6marketing.com/index.php?page=test-child-page.
Angulo, F. (2013). Decimonoveno informe estado de la nacion en desarrollo humano sostenible.
Manejo, disposicion y desecho de las aguas residuales en Costa Rica. San Jose, Costa
Rica.
Araya, D., A. Araya and S. Trejos (2003). Estudio sobre la situaciĂłn de la tecnologia de
tratamiento de las aguas residuales de tipo ordinario en Costa Rica. San Jose, Costa Rica,
Acueductos y Alcantarillados (AyA).
Avila, C., J. GarcĂ­a and M. GarfĂ­ (2016). "Influence of hydraulic loading rate, simulated storm
events and seasonality on the treatment performance of an experimental three-stage
hybrid constructed wetland system." Ecological Engineering 87: 324-332.
Avila, C., J. Salas, I. Martin, C. Aragon and J. Garcia (2013). "Integrated treatment of combined
sewer wastewater and stormwater in a hybrid constructed wetland system in southern
Spain and its further reuse." Ecological Engineering 50: 13-20.
Axaopoulos, P. and P. Panagakis (2003). "Energy and economic analysis of biogas heated
livestock buildings." Biomass and Bioenergy 24(3): 239-248.
Axaopoulos, P., P. Panagakis, A. Tsavdaris and D. Georgakakis (2001). "Simulation and
experimental performance of a solar-heated anaerobic digester." Solar Energy 70(2): 155164.
Ayres, R. U., A. Masini and L. W. Ayres (2001). Application of Exergy Accounting to Five
Basic Metal Industries. Fontainebleu, France.
Bachand, P. and A. Horne (2000). "Denitrification in costructed free-water surface wetland:II
effects of vegetation and temperature." Ecological Engineering 14: 17-32.
Barros, P., I. Ruiz and M. Soto (2008). "Performance of an anaerobic digester-constructed
wetland system for a small community." Ecological Engineering 33(2): 142-149.
Barua, V. B. and A. S. Kalamdhad (2017). "Effect of various types of thermal pretreatment
techniques on the hydrolysis, compositional analysis and characterization of water
hyacinth." Bioresource Technology 227: 147-154.
Barua, V. B., V. W. Raju, S. Lippold and A. S. Kalamdhad (2017). "Electrohydrolysis
pretreatment of water hyacinth for enhanced hydrolysis." Bioresource Technology 238:
733-737.

225

Bharati, V. and A. Kalamdhad (2016). "Water hyacinth to biogas: a review." Pollution Research
35(3): 491 - 501.
Biggs, E. M., E. Bruce, B. Boruff, J. M. A. Duncan, J. Horsley, N. Pauli, K. McNeill, A. Neef, F.
Van Ogtrop, J. Curnow, B. Haworth, S. Duce and Y. Imanari (2015). "Sustainable
development and the water–energy–food nexus: A perspective on livelihoods."
Environmental Science & Policy 54(Supplement C): 389-397.
BohĂłrquez, E., D. Paredes and C. A. Arias (2016). "Vertical flow-constructed wetlands for
domestic wastewater treatment under tropical conditions: effect of different design and
operational parameters." Environmental Technology: 1-10.
Bombardieri, J., T. Espinosa-Solares, M. Domaschko and M. Chatfield (2007). "Thermophilic
anaerobic digester performance under different feed-oading frequency." Applied
Biochemistry and Biotechnology 136 - 140: 765 - 775.
Brix, H., C. Arias and N. Johansen (2003). Experiments in a two-stage constructed wetland
system: nitrification capacity and effects of recycling on nitrogen removal. Wetlands:
Nutrients, metals and mass cycling. J. Vymazal. Leiden, The Netherlands, Backhuys
Publishers: 237 - 258.
Budiyono, S., D. Sumardiono and D. Mardiani (2015). "Microwave pretreatment of fresh water
hyacinth (Eichhornia crassipes) in batch anaerobic digestion tank." International Journal
of Engineering 28(6): 832-840.
Burgoon, P. S., R. H. Kadlec and M. Henderson (1999). "Treatment of potato processing
wastewater with engineered natural systems." Water Science and Technology 40(3): 211215.
Bustillo-Lecompte, C. and M. Mehrvar (2017). Slaughterhouse wastewater: treatment,
management and resource recovery. Physico-chemical wastewater treatment and resource
recovery. R. Farooq and Z. Ahmad.
Cabi. (2016). "Inasive species compendium." Retrieved July 7, 2016, from
http://www.cabi.org/isc/datasheet/14575.
Calvo, C. (2014). Towards sustainable environmental sanitation in Costa Rica. 3631161 Ph.D.,
University of Kansas.
Campbell, E. and S. Safferman (2015). "Design Criteria for the Treatment of Milking Facility
Wastewater in a Cold Weather Vertical Flow Wetland." Transactions of the ASABE
58(6): 1509.

226

Cao, M., K. Gregson and S. Marshall (1998). "Global methane emission from wetlands and its
sensitivity to climate change." Atmospheric Environment 32(19): 3293–3299.
Carretier, S., G. Lesage, A. Grasmick and M. Heran (2015). "WATER AND NUTRIENTS
RECOVERING FROM LIVESTOCK MANURE BY MEMBRANE PROCESSES."
Canadian Journal of Chemical Engineering 93(2): 225-233.
Chatzipaschali, A. A. and A. G. Stamatis (2015). "Exergetic and economic analysis of a cheese
whey wastewater anaerobic treatment plant with a cogeneration system." Desalination
and Water Treatment 56(5): 1223-1230.
Chen, Z. M., B. Chen and G. Q. Chen (2011). "Cosmic exergy based ecological assessment for a
wetland in Beijing." Ecological Modelling 222(2): 322-329.
Cheng, X.-Y., W.-Y. Chen, B.-H. Gu, X.-C. Liu, F. Chen, Z.-H. Chen, X.-Y. Zhou, Y.-X. Li, H.
Huang and Y.-J. Chen (2009). "Morphology, ecology, and contaminant removal
efficiency of eight wetland plants with differing root systems." Hydrobiologia 623(1): 7785.
Comino, E., V. A. Riggio and M. Rosso (2013). "Constructed wetland treatment of agricultural
effluent from an anaerobic digester." Ecological Engineering 54: 165-172.
Coto, O. (2013). Evaluacion de la generacion de residuos agricolas organicos (RAO) en Costa
Rica e identificacion de sector prioritario. Uso de los residuos agricolas organicos como
fuente de energia. Aprovechamiento de los recursos y reduccion de gases de efecto
invernadero. San Jose, Costa Rica.
Cubillo, M. and W. Gomez (2017). "Biojardineras como alternativa para el tratamiento de aguas
residuales: experiencia en cinco biojardineras de la comunidad de Barra Honda y La
Virgen de Nicoya, Guanacaste." Revista Universidad en Dialogo 7(1).
Dallas, S., B. Scheffe and G. Ho (2004). "Reedbeds for greywater treatment—case study in
Santa Elena-Monteverde, Costa Rica, Central America." Ecological Engineering 23(1):
55-61.
Dan, T., L. Quang, N. Chiem and H. Brix (2011). "Treatment of high-strength wastewater in
tropical constructed wetlands planted with Sesbania sesban: Horizontal subsurface flow
versus vertical downflow." Ecological Engineering 37(5): 711-720.
de la Varga, D., M. A. DĂ­az, I. Ruiz and M. Soto (2013). "Avoiding clogging in constructed
wetlands by using anaerobic digesters as pre-treatment." Ecological Engineering 52(0):
262-269.

227

De Paoli, A. C. and M. v. Sperling (2013). "Evaluation of clogging in planted and unplanted
horizontal subsurface flow constructed wetlands: solids accumulation and hydraulic
conductivity reduction." Water Science & Technology 67(6): 1345-1352.
Denny, P. (1997). Implementation of constructed wetlands in developing countries.
Dong, J., Y. Chi, Y. Tang, F. Wang and Q. Huang (2014). "Combined Life Cycle Environmental
and Exergetic Assessment of Four Typical Sewage Sludge Treatment Techniques in
China." Energy & Fuels 28(3): 2114-2122.
Dong, J., Y. Chi, Y. Tang, F. Wang and Q. Huang (2014). "Combined Life Cycle Environmental
and Exergetic Assessment of Four Typical Sewage Sludge Treatment Techniques in
China." Energy &amp; Fuels 28(3): 2114-2122.
Dong, X. and G. B. Reddy (2010). "Nutrient removal and bacterial communities in swine
wastewater lagoon and constructed wetlands." Journal of Environmental Science and
Health, Part A 45(12): 1526-1535.
Echeverria, J. and B. Cantillo (2013). "Instrumentos economicos para la gestion del agua."
Ambientales 45: 13 - 22.
El-Mashad, H. M., W. K. P. van Loon, G. Zeeman, G. P. A. Bot and G. Lettinga (2004). "Design
of A Solar Thermophilic Anaerobic Reactor for Small Farms." Biosystems Engineering
87(3): 345-353.
El-Mashad, H. M., G. Zeeman, W. K. P. van Loon, G. P. A. Bot and G. Lettinga (2004). "Effect
of temperature and temperature fluctuation on thermophilic anaerobic digestion of cattle
manure." Bioresource Technology 95(2): 191-201.
El-Sheikh, M. A., H. I. Saleh, D. E. El-Quosy and A. A. Mahmoud (2010). "Improving water
quality in polluated drains with free water surface constructed wetlands." Ecological
Engineering 36(10): 1478-1484.
EN. (2015). "Estado de la Nacion." Retrieved 8/19/2016, 2016, from
http://www.estadonacion.or.cr/21/.
EPA (1978). Use of Solar Energy to Heat Anaerobic Digesters.
EPA (2000). Constructed wetalnds treatment of municipal wastewaters. Cincinnati, Ohio, Office
of Research and Development.
Espigares, M. and J. Perez (1985). Aspectos sanitarios del estudio de las aguas. Granada, Spain,
Universidad de Granada.
228

Espinosa-Solares, T., S. Valle-Guadarrama, J. Bombardiere, M. Domaschko and M. Easter
(2009). "Effect of heating strategy on power consumption and performance of a pilot
plant anaerobic digester." Applied Biochemical Biotechnology 156: 465 - 474.
FAO. (2014). "The water-energy-food nexus at FAO." 2017, from
http://www.gwp.org/globalassets/global/toolbox/references/the-water-energy-food-nexusat-fao---concept-note-fao-2014.pdf.
Funk, K., J. Milford and T. Simpkins. (2013). "Waste no, want not: environmental viability of
waste-to-energy (WTE) technology for site-specific optimization of renewable energy
options." Retrieved September 2, 2016, from
http://www.nrel.gov/docs/fy13osti/52829.pdf.
Garcia, J., D. P. L. Rousseau, J. Morato, E. L. S. Lesage, V. Matamoros and J. M. Bayona
(2010). "Contaminant Removal Processes in Subsurface-Flow Constructed Wetlands: A
Review." Critical Reviews in Environmental Science and Technology 40(7): 561-661.
Garcia-Perez, A., M. Harrison and B. Grant (2011). "Recirculating Vertical Flow Constructed
Wetland for On-Site Sewage Treatment: An Approach for a Sustainable Ecosystem."
Journal of Water and Environment Technology 9(1): 39-46.
Gould, C. and M. Crook (2010). Michigan on-farm anaerobic digester operator handbook. East
Lansing, MI. USA, MSU Extension.
Gupta, P., S. Roy and A. Mahindrakar (2012). "Treatment of water using water hyacinth, water
lettuce and vetiver grass - a review." Resources and Enrinment 2: 202 - 215.
Gutierrez, E., E. Ruiz, E. Uribe and J. Martinez (2001). Biomass and productivity of water
hyacinth and their application in control programs. Biological and Integrated Control of
Water hyacinth, Eichhornia crassipes: 109 - 119.
GWP (2012). Global Water Partnership. Agenda del Agua: Costa Rica 2013 - 2030. San Jose,
Costa Rica.
Hacienda, M. d. (2016). Mapa de valores de terrenos por zonas homogeneas Provincia 2 de
Alanjuela, Canton 1 Alajuela, Distrito 13 Garita, Municipalidad de Alajuela.
Halverson, N. V. (2004). "Review of Constructed Subsurface Flow vs. Surface Flow Wetlands."
Technical report. Westinghouse Savannah River Co(509).
Hassanein, A., L. Qiu, P. Junting, G. Yihong and F. Witarsa (2015). "Simulation and validaion of
a model for heating underground biogas digesters by solar energy." Ecological
Engineering 82: 336 - 344.
229

He, Y., W. Tao, Z. Wang and W. Shayya (2012). "Effects of pH and seasonal temperature
variation on simultaneous partial nitrification and anammox in free-water surface
wetlands." Journal of Environmental Management 110: 103-109.
Healy, M. G., M. Rodgers and J. Mulqueen (2007). "Treatment of dairy wastewater using
constructed wetlands and intermittent sand filters." Bioresource Technology 98: 22682281.
Hellstrom, D. (1997). "An Exergy Analysis for a Wastewater Treatment Plant: An Estimation of
the Consumption of Physical Resources." Water Environment Research 69(1): 44-51.
Hornbogen, E. (2003). A definition of sustainability based on entropy production by matter and
energy, Metall. 57: 124 - 128.
Hua, G., Z. Zhao, J. Kong, R. Guo, Y. Zeng, L. Zhao and Q. Zhu (2014). "Effects of plant roots
on the hydraulic performance during the clogging process in mesocosm vertical flow
constructed wetlands." Environmental Science and Pollution Research International
21(22): 13017-13026.
Hua, G. F., L. Li, Y. Q. Zhao, W. Zhu and J. Q. Shen (2013). "An integrated model of substrate
clogging in vertical flow constructed wetlands." Journal of Environmental Management
119(0): 67-75.
Hua, G. F., W. Zhu and Y. H. Zhang (2010). "A conceptual approach based on suspended solids
to estimate clogging time in constructed wetlands." Journal of Environmental Science
and Health, Part A 45(12): 1519-1525.
Hua, G. F., W. Zhu, L. F. Zhao and J. Y. Huang (2010). "Clogging pattern in vertical-flow
constructed wetlands: Insight from a laboratory study." Journal of Hazardous Materials
180(Issues 1–3): 668–674.
Huijsmans, J. F. M., J. M. G. Hol and G. D. Vermeulen (2003). "Effect of application method,
manure characteristics, weather and field conditions on ammonia volatilization from
manure applied to arable land." Atmospheric Environment 37(26): 3669-3680.
ICE (2016). Pograma de biogas Plan de fuentes renovables no convencionales. Biogas as
sustainable energy solution for developing countries. EARTH University, Guacimo,
Limon, Costa Rica, mbvberlin: 60 - 66.
ICE (2016). Tarifas del ICE.

230

ICE-UEN (2010). Planificacion y desarrollo de proyectos de generacion electrica. Costa Rica:
Ejemplo de aprovechamiento de las fuentes energeticas renovables convencionales y no
convencionales. San Jose, Costa Rica.
ICT (2014). Anuario estadistico de turismo 2014. San Jose, Costa Rica.
IMN. (2016). "Instituto Metereologico Nacional de Costa Rica." Retrieved July 7, 2016, from
https://www.imn.ac.cr/web/imn/inicio.
IRENA. (2012). "Biomass for power generation. Renewable energy technologies: cost analysis
series." Retrieved September 2, 2016, from
https://www.irena.org/DocumentDownloads/Publications/RE_Technologies_Cost_Analy
sis-BIOMASS.pdf.
IRET (2012). Indicador de contaminantes y de macroinvertebrados. CGR DFOE-AE-IF-012013. Heredia, Costa Rica, Universidad Nacional de Costa Rica.
Iris. (2016). "The iris flower." Retrieved July 7, 2016, from http://flower-iris.ru/en/knigi-proiridariy/zaglyanut-v-knigu/58/.
ITCR. (2003). "Technical and regulatory guidance document for constructed treatment
wetlands." from http://www.itrcweb.org/GuidanceDocuments/WTLND-1.pdf.
Johnson, J. and D. Eckert. (1995). "Best Management Practices: Land Application of Animal
Manure." Retrieved May 13, 2015, from http://ohioline.osu.edu/agf-fact/0208.html.
Jørgensen, S. E. (2006). Eco-exergy as sustainability. Scitech Book News. Portland, Ringgold
Inc. 30.
Kadlec, R. (2000). "The inadequacy of first-order treatment wetland models." Ecological
Engineering 15: 105-119.
Kadlec, R. and R. Knight (1996). Treatment wetlands. Lewis, Boca Raton.
Kadlec, R. H. (2009). "Comparison of free water and horizontal subsurface treatment wetlands."
Ecological Engineering 35(2): 159-174.
Kadlec, R. H. and K. R. Reddy (2001). "Temperature Effects in Treatment Wetlands." Water
Environment Research 73(5): 543-557.
Kadlec, R. H. and S. D. Wallace (2009). Treatment wetlands. Boca Raton, FL, CRC Press.

231

Kantawanichkul, S., S. Kladprasert and H. Brix (2009). "Treatment of high-strength wastewater
in tropical vertical flow constructed wetlands planted with Typha angustifolia and
Cyperus involucratus." Ecological Engineering 35(2): 238-247.
Kapellakis, I. E., N. Paranychianakis, K. Tsagarakis and A. Angelakis (2012). Treatment of
Olive Mill Wastewater with Constructed Wetlands.
Katsenovich, Y. P., A. Hummel-Batista, A. J. Ravinet and J. F. Miller (2009). "Performance
evaluation of constructed wetlands in a tropical region." Ecological Engineering 35(10):
1529-1537.
Kelvin, K. and M. Tole (2011). "The efficacy of a tropical constructed wetland for treating
wastewater during the dry season: the Kenya experience." Water, air and soil pollution
215.1(4): 137 - 143.
Khosravi, S. and Panjeshahi (2013). "Application of exergy analysis for quantification and
optimisation of the environmental performance in wastewater treatment plants."
International journal of exergy 12(1): 119.
Kincanon, R. and A. S. McAnally (2004). "Enhancing commonly used model predictions for
constructed wetland performance: as-built design considerations." Ecological Modelling
174(3): 309-322.
Kinyua, M. N. (2015). Energy Production and Effluent Quality in Tubular Digesters Treating
Livestock Waste in Rural Costa Rica. 3712534 Ph.D., University of South Florida.
Knight, R. L., V. W. E. Payne Jr, R. E. Borer, R. A. Clarke Jr and J. H. Pries (2000).
"Constructed wetlands for livestock wastewater management." Ecological Engineering
15(1–2): 41-55.
Knowles, P., G. Dotro, J. Nivala and J. GarcĂ­a (2011). "Clogging in subsurface-flow treatment
wetlands: Occurrence and contributing factors." Ecological Engineering 37(2): 99-112.
Kohlmann, B. (2016). Renewable energies in Costa Rica status quo and prospects. Biogas as a
sustainable energy solution for developing countries. EARTH University, Guacimo,
Limon, Costa Rica, mbvberlin: 41 - 45.
Konnerup, D., T. Koottatep and H. Brix (2009). "Treatment of domestic wastewater in tropical,
subsurface flow constructed wetlands planted with Canna and Heliconia." Ecological
Engineering 35(2): 248-257.

232

Kosseva, M. and C. Kent (2013). Thermophilic aerobic bioprocessing technologies for food
industry wastes and wastewater. Food industry wastes: Assessment and recuperation of
commodities. M. Kosseva and C. Webb, Academics Press: 198 - 217.
Laber, J., R. Haberl and G. Langergraber (2003). Treatment of hospital wastewater with a twostage constructed wetland system. Achievments and prospects of phytoremediation in
Europe. Vienna, Austria, Uniersity of Natural Resources ad Applied Life Sciences.
Langergraber, G. (2007). "Modeling of processes in subsurface flow constructed wetlands: a
review." Vadose Zone Journal 7(2): 830 - 842.
Langergraber, G., R. Haberl, J. Laber and A. Pressl (2003). "Evaluation of substrate clogging
processes in vertical flow constructed wetlands." Water Science & Technology 48(5): 2534.
Lansing, S., J. VĂ­quez, H. MartĂ­nez, R. Botero and J. Martin (2008). "Quantifying electricity
generation and waste transformations in a low-cost, plug-flow anaerobic digestion
system." Ecological Engineering 34(4): 332-348.
Leandro, C. (2011). "Arranca mayor planta termica del pais." 2017, from
https://www.larepublica.net/noticia/arranca_mayor_planta_termica_del_pais.
Lee, C. g., T. D. Fletcher and G. Sun (2009). "Nitrogen removal in constructed wetland
systems." Engineering in Life Sciences 9(1): 11-22.
Lee, S. H. and Cha (2015). "Performance Evaluation of the Free Water Surface Constructed
Wetland Treating Nonpoint Source Pollutants in the Agricultural Area." Journal of
environmental & analytical toxicology 5(280).
Leverenz, H., G. Tchobanoglous and J. Darby (2009). "Clogging in intermittently dosed sand
filters used for wastewater treatment." Water Research 43(3): 695–705.
Li, L., Y. Yang, N. F. Y. Tam, L. Yang, X.-Q. Mei and F.-J. Yang (2013). "Growth
characteristics of six wetland plants and their influences on domestic wastewater
treatment efficiency." Ecological Engineering 60: 382-392.
Liang, M.-Q., C.-F. Zhang, C.-L. Peng, Z.-L. Lai, D.-F. Chen and Z.-H. Chen (2011). "Plant
growth, community structure, and nutrient removal in monoculture and mixed
constructed wetlands." Ecological Engineering 37(2): 309-316.
Liedl, B. E., J. Bombardiere and J. M. Chatfield (2006). "Fertilizer potential of liquid and Solid
effluent from thermophilic anaerobic digestion of poultry waste." Water Science and
Technology 53(8): 69-79.
233

Lorimor, J., W. Powers and A. Sutton (2004). Manure characteristics. Ames, IA, MIdWest Plan
Service. MWPS-18-Section 1.
MacDonald, M. (2011). Costs of low-carbon generation technologies. Committee on Climate
Change, London, UK.
Martin, M. and A. Parsapour (2012). Upcycling wastes with biogas production: An exergy and
economic analysis. International Symposium on Energy from Biomass and Waste.
Venice, Italy.
MartĂ­nez, A., J. Uche, C. Rubio and B. Carrasquer (2010). "Exergy cost of water supply and
water treatment technologies." Desalination and Water Treatment 24(1-3): 123-131.
MEIC (2013). Estudio preliminar para determinar la posibilidad de regular el mercado de
fertilizantes en Costa Rica. San Jose, Costa Rica, Direccion de Investigaciones
Economicas y de Mercado.
MINAE-MSP (2007). Reglamento de vertido y reuso de aguas residuales. P. Ejecutivo. San Jose,
Costa Rica. N° 33601-MINAE-S: 36.
MINAET (2010). Evaluacion de flujos de inversion y financiamiento para la adaptacion al
cambio climatico de los sectores biodiversidad y recursos hidrico. PNUD. San Jose,
Costa Rica.
MINAET (2011). Diagnostico Sector Energia de Costa Rica. VI Plan Nacional de Energia 20122030. San Jose, Costa Rica, Ministerio de Ambiente, Energia y Telecomunicaciones.
Mitsch, W. and J. Gosselink (2007). Wetlands.
Moncada, S. (2011). Evaluacion del diseno de una biojardinera de flujo subsuperficial para el
tratamiento de aguas grises en Zapote, San Jose. Licenciatura, Instituto Tecnologico de
Costa Rica.
Monlau, F., C. Sambusiti, E. Ficara, A. Aboulkas, A. Barakat and H. Carrere (2015). "New
opportunities for agricultural digestate valorization: current situation and perspectives."
Energy & Environmental Science 8(9): 2600-2621.
Moorhead, K. and R. Nordstedt (1993). "Batch anaerobic digestion of water hyacinth: effect of
particle size, plant nitrogen and inoculum volume." Bioresource Technology 44: 71-76.
Mori, H. (2003). Technology Transfer Manual of Industrial Wastewater Treatment. Japan,
Ministry of Environment, Government of Japan.
234

Naik, K. S. (2014). Decentralized wastewater management solutions for improved public health
protection and reclamation: Optimization and application. 3614141 Ph.D., University of
California, Los Angeles.
Nivala, J., P. Knowles, G. Dotro, J. GarcĂ­a and S. Wallace (2012). "Clogging in subsurface-flow
treatment wetlands: Measurement, modeling and management." Water Research 46(6):
1625-1640.
Nogarin, M. (2017). Los Tajos advances San Jose's wastewater treatment. Tulsa, OK, Water and
Wastewater International.
Noramelya, N., M. Shahbudin, J. Jahim and E. Herianto (2016). "Overview of H2S removal
technologies from biogas production." International Journal of Applied Engineering
Research 11(20): 10060 - 100666.
Opio, A., M. Jones, F. Kansiime and T. Otiti (2014). "Growth and Development of Cyperus
papyrus in a Tropical WetlandGrowth and Development of Cyperus papyrus in a
Tropical Wetland." Open Journal of Ecology 4: 113-123.
Otarola, S. (2015). Planta de tratamiento Los Tajos inicia operaciones en segundo semestre con
aguas residuales de los colectores Rivera y Torres. San Jose, Costa Rica, GobiernoCR.
Otchet, A. (2003). "UNESCO - Political inertia exacerbates water crisis, says World Water
Development Report First UN system-wide evaluation of global water resources."
Retrieved April 23, 2015, from http://portal.unesco.org/en/ev.phpURL_ID=10064&URL_DO=DO_TOPIC&URL_SECTION=201.html.
Paing, J., V. Serdobbel, M. Welschbillig, M. Calvez, V. Gagnon and F. Chazarenc (2015).
"Treatment of high organic content wastewater from food-processing industry with the
French vertical flow constructed wetland system." Water Science & Technology 72(1):
70-76.
Parsapour, A. (2012). Biogas production system as an upcycler: exergy analysis and economic
evaluation. Master, LinkĂśping University.
Pedescoll, A., A. Corzo, E. Álvarez, J. García and J. Puigagut (2011). "The effect of primary
treatment and flow regime on clogging development in horizontal subsurface flow
constructed wetlands: An experimental evaluation." Water Research 45(12): 3579-3589.
Pedescoll, A., E. Uggetti, E. Llorens, F. GranĂŠs, D. GarcĂ­a and J. GarcĂ­a (2009). "Practical
method based on saturated hydraulic conductivity used to assess clogging in subsurface
flow constructed wetlands." Ecological Engineering 35(8): 1216-1224.

235

PĂŠrez, R., C. Alfaro, J. Sasa and J. AgĂźero (2013). "Performance evaluation of an alternative
artificial wetlands system used as wastewater treatment." Uniciencia 27(1): 9.
Querol, E., B. Gonzalez-Regueral and J. L. Perez-Benedito (2013). Practical Approach to Exergy
and Thermoeconomic Analyses of Industrial Processes. New York and Heidelberg.
Raedle, J. (2017). "Costa Rica, apunto de conseguir el 100% de toda su energia de fuentes
renovables." La Vanguardia, 2017, from
http://www.lavanguardia.com/economia/20170104/413093838923/costa-rica-energiarenovables-2016.html.
Reed, S., W. Ronald and E. Middlebrooks (1995). Natural systems for waste management and
treatment. Ney York, N.Y., McGraw-Hill.
Rice, E. and L. Bridgewater (2012). Standard methods for the examinationof water and
wastewater. Washington, D.C., American Public Health Association.
Ritter, W. F. and A. Shirmohammadi (2001). Agricultural nonpoint source pollution; watershed
management and hydrology, Lewis Publishers, Boca Raton, FL.
Rosen, M. A. (2012). Environment, Ecology and Exergy : Enhanced Approaches to
Environmental and Ecological Management. Hauppauge, NY, USA, Nova Science
Publishers, Inc.
Ruiz, G. (2014). Agua limita crecimiento inmobiliario. El Financiero. Costa Rica.
Ruiz, G. (2014). Agua limita crecimiento inmobiliario. El Financiero. San Jose, Costa Rica: 23.
Ruiz, I., M. A. DĂ­az, B. Crujeiras, J. GarcĂ­a and M. Soto (2010). "Solids hydrolysis and
accumulation in a hybrid anaerobic digester-constructed wetlands system." Ecological
Engineering 36(8): 1007-1016.
Safferman, S. I. and J. M. Wallace (2015). "Cow Manure: Waste or Resource?" IEEE Potentials
34(1): 25-29.
Salazar, M. (2014). "Costa Rica aims for carbon neutrality with payments for ecosystems
services." Retrieved June, 2016, from
http://www.ecosystemmarketplace.com/articles/costa-rica-aims-for-carbon-neutralitywith-payments-for-ecosystem-services/.
Sani, A., M. Scholz, A. Babatunde and Y. Wang (2013). "Impact of Water Quality Parameters on
the Clogging of Vertical-Flow Constructed Wetlands Treating Urban Wastewater."
Water, air, and soil pollution 224(3): 1.
236

Sanyal, O., Z. G. Liu, B. M. Meharg, W. Liao and I. Lee (2015). "Development of
polyelectrolyte multilayer membranes to reduce the COD level of electrocoagulation
treated high-strength wastewater." Journal of Membrane Science 496: 259-266.
Sato, T., M. Qadir, S. Yamamoto, T. Endo and A. Zahoor (2013). "Global, regional, and country
level need for data on wastewater generation, treatment, and use." Agricultural Water
Management 130: 1-13.
Sciubba, E. (2011). "A revised calculation of the econometric factors ι- and β for the Extended
Exergy Accounting method." Ecological Modelling 222(4): 1060-1066.
Serrano, L., D. de la Varga, I. Ruiz and M. Soto (2011). "Winery wastewater treatment in a
hybrid constructed wetland." Ecological Engineering 37(5): 744-753.
Shao, L. and G. Q. Chen (2015). "Exergy based renewability assessment: Case study to
ecological wastewater treatment." Ecological Indicators 58: 392-401.
Sharma, D., T. Espinosa-Solares and D. H. Huber (2013). "Thermophilic anaerobic co-digestion
of poultry litter and thin stillage." Bioresource Technology 136: 251-256.
Shepherd, H. L., G. Tchobanoglous and M. E. Grismer (2001). "Time-Dependent Retardation
Model for Chemical Oxygen Demand Removal in a Subsurface-Flow Constructed
Wetland for Winery Wastewater Treatment." Water Environment Research 73(5): 597606.
Sherlock, R. R., S. G. Sommer, R. Z. Khan, C. W. Wood and et al. (2002). "Ammonia, methane,
and nitrous oxide emission from pig slurry applied to a pasture in New Zealand." Journal
of Environmental Quality 31(5): 1491-1501.
Siefert, N. S. and S. Litster (2014). "Exergy and economic analysis of biogas fueled solid oxide
fuel cell systems." Journal of Power Sources 272: 386-397.
Siva Reddy, V., S. C. Kaushik, K. R. Ranjan and S. K. Tyagi (2013). "State-of-the-art of solar
thermal power plants—A review." Renewable and Sustainable Energy Reviews 27(0):
258-273.
Soroko, M. (2007). "Treatment of wastewater from small slaughterhouse in hybrid constructed
wetlands systems." Ecohydrology & Hydrobiology 7(3): 339-343.
Stefanakis, A., C. Akratos and V. Tsihrintzis (2014). Vertical flow constructed wetlands: ecoengineering systems for wastewater and sludge treatment, Elsevier.

237

Stefanakis, A. I. and V. A. Tsihrintzis (2012). "Effects of loading, resting period, temperature,
porous media, vegetation and aeration on performance of pilot-scale vertical flow
constructed wetlands." Chemical Engineering Journal 181–182(0): 416-430.
Stein, O. R., J. A. Biederman, P. B. Hook and W. C. Allen (2006). "Plant species and
temperature effects on the k–C* first-order model for COD removal in batch-loaded SSF
wetlands." Ecological Engineering 26(2): 100-112.
Suliman, F., H. K. French, L. E. Haugen and A. K. Søvik (2006). "Change in flow and transport
patterns in horizontal subsurface flow constructed wetlands as a result of biological
growth." Ecological Engineering 27(2): 124-133.
Summerfelt, S. T., P. R. Adler, D. M. Glenn and R. N. Kretschmann (1999). "Aquaculture sludge
removal and stabilization within created wetlands." Aquacultural Engineering 19(2): 8192.
Suryawanshi, P. C., A. B. Chaudhari and R. M. Kothari (2010). "Thermophilic anaerobic
digestion: the best option for waste treatment." Critical Reviews in Biotechnology 30(1):
31-40.
Sutton, A. and B. Joern. (1992). "Land Application of Manure." Retrieved May 13, 2015, from
https://www.extension.purdue.edu/extmedia/WQ/WQ-16-W.html.
SuĂĄrez, J. and J. Puertas (2005). "Determination of COD, BOD, and suspended solids loads
during combined sewer overflow (CSO) events in some combined catchments in Spain."
Ecological Engineering 24(3): 199-217.
Szargut, J. (1989). "Chemical exergies of the elements." Applied Energy 32(4): 269-286.
Szargut, J., D. R. Morris and F. R. Steward (1988). Exergy analysis of thermal, chemical, and
metallurgical processes. New York, Hemisphere.
Tai, S. and Matsushige (1986). "Chemical Exergy of Organic Matter in Wastewater."
International Journal of Environmental Studies IJEVA 27: 301 - 315.
Tanaka, K. and N. Tanaka (2011). Wetlands for Tropical Applications : Wastewater Treatment
by Constructed Wetlands. Singapore, SGP, World Scientific Publishing Co.
Tang, Y., Z. Fang, Z. Zhang, Y. Zhong and K. Chen (2014). "Ecological indicators showing the
succession of macrofauna communities is Sonneratia apetala artificial mangrove
wetlands on Qi'ao Island at Zhuhai, South ChinaEcological indicators showing the
succession of macrofauna communities is Sonneratia apetala artificial mangrove wetlands
on Qi'ao Island at Zhuhai, South China." aActa oceanologica sinica 33(8): 62-72.
238

Tilley, E., C. Luethi, A. Morel, C. Zurbruegg and R. Schertenleib (2008). Compendium of
Sanitation Systems and Technologies. Duebendorf and Geneva, Swiss Federal Institute of
Aquatic Science and Technology.
Torrens, A., P. Molle, C. Boutin and M. Salgot (2009). "Impact of design and operation variables
on the performance of vertical-flow constructed wetlands and intermittent sand filters
treating pond effluent." Water Research 43(7): 1851-1858.
Trang, N., D. Konnerup, H. Schierup, N. Chiem, L. Tuan and H. Brix (2010). "Kinetics of
pollutant removal from domestic wastewater in a tropical horizontal subsurface flow
constructed wetland system: Effects of hydraulic loading rate." Ecological Engineering
36(4): 527-535.
Tsalkatidou, M., M. Gratziou and N. Kotsovinos (2009). "Combined stabilization ponds–
constructed wetland system." Desalination 248(1–3): 988-997.
Tsihrintzis, V. A., C. S. Akratos, G. D. Gikas, D. Karamouzis and A. N. Angelakis (2007).
"Performance and Cost Comparison of a FWS and a VSF Constructed Wetland System."
Environmental Technology 28(6): 621-628.
Tunçsiper, B. (2009). "Nitrogen removal in a combined vertical and horizontal subsurface-flow
constructed wetland system." Desalination 247(1–3): 466-475.
UCR, U. d. C. R. (2016). Escala salarial para el sector administrativo, Oficina de Recursos
HUmanos.
UN. (2009). "Report of the independent expert on the issue of human rights obligations related to
access to safe drinking water and sanitation, Catarina de Albuquerque, Addendum:
Mission to Costa Rica." Retrieved March 17, 2015, from
http://www.ohchr.org/EN/Issues/WaterAndSanitation/SRWater/Pages/CountryVWoSts.a
spx.
UN. (2013). "Report of the Special Rapporteur on the human right to safe drinking water and
sanitation, Catarina de Albuquerque." Retrieved March 17, 2015, from
http://www.zaragoza.es/contenidos/medioambiente/onu/1048eng_report_of_special_rapporteur_on_human_right_to_safe_drinking_water_and_sanitat
ion.pdf.
UN. (2015). "Wastewater management. A UN water analytical brief." Retrieved 8/19/2016,
2016, from http://www.unwater.org/fileadmin/user_upload/unwater_new/docs/UNWater_Analytical_Brief_Wastewater_Management.pdf.

239

UNESCO. (2015). "The United Nations world water development report 2015." Retrieved June
10, 2016, from http://www.unesco.org/new/en/naturalsciences/environment/water/wwap/wwdr/2015-water-for-a-sustainable-world/.
Vindis, P., B. Mursec, M. Janzekovic and F. Cus (2009). "The impact of mesophilic and
thermophilic anaerobic digestion on biogas production." Journal of Achievements in
Materials and Manufacturing Engineering 36(2): 192 - 198.
Visser, A., Y. Gao and G. Lettinga (1993). "Effects of short-term temperature increases on the
mesophilic anaerobic breakdown of sulfate containing synthetic wastewater." Water
Research 27(4): 541-550.
Vymazal, J. (2005). "Horizontal sub-surface flow and hybrid constructed wetlands systems for
wastewater treatment." Ecological Engineering 25(5): 478-490.
Vymazal, J. (2007). "Removal of nutrients in various types of constructed wetlands." Science of
The Total Environment 380(1–3): 48-65.
Vymazal, J. (2011). "Plants used in constructed wetlands with horizontal subsurface flow: a
review." Hydrobiologia 674(1): 133-156.
Vymazal, J. (2014). "Constructed wetlands for treatment of industrial wastewaters: A review."
Ecological Engineering 73: 724-751.
Vymazal, J. and L. KrĂśpfelovĂĄ (2015). "Multistage hybrid constructed wetland for enhanced
removal of nitrogen." Ecological Engineering 84: 202-208.
Wall, G. (2010). "On exergy and sustainable development in environmetal engineering." The
Open Environmental Engineering Journal 3: 21 - 32.
Wallace, S. D. and R. L. Knight (2006). Small-scale constructed wetland treatment
system:feasibility, design criteria, O&M requirements. Alexandria, Virginia, Water
Environment Research Foundation (WERF).
Wang, H. and R. Larson. (2015). "Effects of mixing duration on biogas production and
methanogen distribution in dairy manure anaerobic digester." Retrieved April 26, 2017,
from http://articles.extension.org/pages/72955/effects-of-mixing-duration-on-biogasproduction-and-methanogen-distribution-in-dairy-manure-anaerobi.
Wang, S., Z. Xu, H. Li and D. Zhou (2008). "Effect of plant tillering and root development on
hydrodynamics and wastewater purification of vertical down flow wetlands." Tongji
Daxue Xuebao / Journal of Tongji University (Natural Science) 36(4): 519-524.

240

Ward, A. and S. Trimble (2004). Environmental hydrology. Boca Raton, FL, Lewis Publishers.
Winter, K. J. and D. Goetz (2003). "The impact of sewage composition on the soil clogging
phenomena of vertical flow constructed wetlands." Water Science & Technology 48(5):
9-14.
Woudstra, N. (2016). "Exergy: the quality of energy." Retrieved July 2, 2016, from
http://www.tbm.tudelft.nl/fileadmin/Faculteit/TBM/Onderzoek/Onderzoeksportfolio/CPI
RD/Projecten_en_activiteiten/Exergie/doc/Exergy_the_quality_of_energy.pdf.
Wright, J. (2008). "Calculo y mapeo de la radiacion solar directa y difusa en Costa Rica."
UNICIENCIA 22: 15.
Wu, S., P. Kuschk, H. Brix, J. Vymazal and R. Dong (2014). "Development of constructed
wetlands in performance intensifications for wastewater treatment: A nitrogen and
organic matter targeted review." Water Research 57: 40-55.
Xinshan, S., L. Qin and Y. Denghua (2010). "Nutrient Removal by Hybrid Subsurface Flow
Constructed Wetlands for High Concentration Ammonia Nitrogen Wastewater." Procedia
Environmental Sciences 2(0): 1461-1468.
Zabranska, J., M. Dohanyos, P. Jenicek, P. Zaplatilkova and J. Kutil (2002). "The contribution of
thermophilic anaerobic digestion to the stable operation of wastewater sludge treatment."
Water Science and Technology 46(4 -5): 447 - 453.
Zachritz, W. H., A. T. Hanson, J. A. Sauceda and K. M. Fitzsimmons (2008). "Evaluation of
submerged surface flow (SSF) constructed wetlands for recirculating tilapia production
systems." Aquacultural Engineering 39(1): 16-23.
Zafar, S. (2016). "Summary of biomass combustion technologies." Retrieved 8/24/2016, 2016,
from http://www.bioenergyconsult.com/.
Zarkadas, I. S., A. S. Sofikiti, E. A. Voudrias and G. A. Pilidis (2015). "Thermophilic anaerobic
digestion of pasteurised food wastes and dairy cattle manure in batch and large volume
laboratory digesters: Focussing on mixing ratios." Renewable Energy 80: 432-440.
Zhang, D., K. Jinadasa, R. Gersberg, Y. Liu, S. Tan and W. Ng (2015). "Application of
constructed wetlands for wastewater treatment in tropical and subtropical regions (2000–
2013)." Journal of Environmental Sciences 30: 30-46.
Zhang, D.-Q., K. B. S. N. Jinadasa, R. M. Gersberg, Y. Liu, S. K. Tan and W. J. Ng (2015).
"Application of constructed wetlands for wastewater treatment in tropical and subtropical
regions (2000–2013)." Journal of Environmental Sciences 30: 30-46.
241

Zhang, Z., Z. Rengel and K. Meney (2007). "Growth and resource allocation of Canna indica and
Schoenoplectus validus as affected by interspecific competition and nutrient availability."
Hydrobiologia 589(1): 235-248.
Zhao, Y. Q., G. Sun and S. J. Allen (2004). "Anti-sized reed bed system for animal wastewater
treatment: a comparative study." Water Research 38(12): 2907-2917.
Zurita, F., J. De Anda and M. A. Belmont (2009). "Treatment of domestic wastewater and
production of commercial flowers in vertical and horizontal subsurface-flow constructed
wetlands." Ecological Engineering 35(5): 861-869.

242