1 lwlu...il. . . , , V. . _ fig
.u...u.....§........-.z..§ . . p:
u .n
, ...

from»... s.
I: . ~
I.

r: n.
a: a. h...
5;

r duluufu.
.. 2......
£355»...
:i 4L.
‘51:». .4:

.1 V {.24.

. . ..
. . .
0:3; i . . I av‘ é}
3.1.! a... . . . , . . . $5”? 32.4...
. . _:..s,v.......4..i: . waifrbwvfiw
$552.1. to x .t n . . . . :57: 1.: ..
,. , mi...
.55.

, .1
(an!
I... 1

r , 3:
,3: INF

. I “44
x 3.3. 1%... ..

105... i.
.: r13 :31?
.‘ {'rI.

l 01?}.
1!‘ t; 5.
.. ‘4 -

 

 

 

I'll! 1.5.31; ”In. .1... 1 .1

3 LIBRARY
’10 ,3) Co Michigan State
University

This is to certify that the
dissertation entitled

MODLEING THE FIELD WATER BALANCE ON IRRIGATED
CROP FIELDS USING
ROOT ZONE WATER QUALITY MODEL

presented by
MICHAEL I GANGWER

has been accepted towards fulfillment
of the requirements for the

degree in Agriculture Technology and
Ph.D. Systems Management

 

 

£0... 3.5”

Major Professor's Signature

lz/flf

Date

MSU is an Affirmative Action/Equal Opportunity Institution

...-.-.---

an-c-O---n-.-I-O-I-O-O-l-l-o-I-1-C-o-t-l-I-O-I-n-O-I-O-t-D-I-l-a

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE
ALIéflg " 0512 7
9 ‘5»... I4

4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd-p.1

 

MODELING THE FIELD WATER BALANCE ON IRRIGATED CROP FIELDS
USING ROOT ZONE WATER QUALITY MODEL

By

Michael I. Gangwer

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Biosystems and Agricultural Engineering

2005

ABSTRACT

MODELING THE FIELD WATER BALANCE ON IRRIGATED CROP FIELDS
USING ROOT ZONE WATER QUALITY MODEL

By

Michael I. Gangwer

Managing the amount and timing of irrigation water and liquid manure
applications on crop fields is a major requirement for optimizing crop yield and reducing
the potential for soil and water resource degradation.

This study had two objectives. One, develop parameters for evaluating the
performance of a comprehensive water budgeting model, and two, evaluate the
performance of the model by comparing empirically derived field water balance (FWB)
with model simulations.

Ten fields on four dairy farms in the Willamette Valley were instrumented with a
soil moisture sensor, Irrometer Watermark. These sensors measure electrical resistance,
thereby using conductivity as a measure of moisture content. Local rainfall amounts
were obtained by installing a weather station on each of the four farms. Other climate
data was obtained from AgriMet and National Weather Service weather stations.

Bulk soils from these ten fields were evaluated for particle size analysis and moist
bulk density. Soil cores were analyzed for paired matric potential and volumetric water
content. The paired data yielded the soil water characterization curve (SWCC).

The Root Zone Water Quality Model (RZWQM) was parameterized with three

data sets. One, data obtained from NASIS (soil data mart) based upon the soil map unit

for that field. Two, a mix of RZWQM data, field, and laboratory data. And three,
complete data sets for each field based upon the SWCC, and related field data.

The model simulations were evaluated using four statistical measurements,
including Root Mean Square Error, Mean Error, and the Nash Sutcliffe Equation. Results
show that using strictly NASIS data result in poor model performance. This is likely
related to alluvial soils deposited from recent flood events.

Results were largely mixed comparing a combination of default, field, and
laboratory data with the complete field and data set. This study suggests that the large
investment deriving the SWCC in the laboratory did not improve model performance.

The model performed reasonably well with field determinations of particle size
analysis, including sand, silt, and clay content. Once the clay content was known, the
model assigned default values for volumetric water content at saturation and field
capacity. Model performance could have been improved if saturated hydraulic
conductivity was measured using some field technique. RZWQM does not have a robust
crop growth component; FWB simulations could have been improved by knowing
evapotranspiration (ET) from the crop based on growth.

Obtaining bulk soil from a field, then determining particle size analysis, coupled
with field measurements of moist bulk density, volumetric water content at saturation and
FC, saturated hydraulic conductivity and accurate ET for the growing crop could be used

with RZWQM for the determination of the Field Water Balance over one or more years.

© Copyright By
Michael I. Gangwer
2005

ACKNOWLEDGEMENTS

I thank my Guidance Committee for helping me during my journey at Michigan
State University. They are:
Dr. Robert von Bemuth, Chair
Director, School of Planning, Design, and Construction

Dr. Ruth Shaffer
USDA-NRCS, East Lansing, Michigan

Dr. William Northcott
Biosystems and Agricultural Engineering Department

Dr. Tim Harrigan
Biosystems and Agricultural Engineering Department

Dr. Delbert Mokrna
Crops and Soils Department

I am grateful for the Oregon dairy farmers participating in this study. They are
Bernie Faber, Paul Staehely, Kryn Van Loon, and Bob Williamson and their families.

I thank four associates in Oregon, who encouraged me early on in this study to
keep going. They are Dr. James Moore and Dr. John Selker, Oregon State University,
Bruce Wilson and Monte Graham of Oregon NRCS.

I appreciate the encouragement and support from Kevin Wickey (my supervisor),
Vicki Anderson, and Jerry Grigar of Michigan NRCS.

I am most grateful for my wife Sandy, who has supported and encouraged me

from the very beginning of this study, which began in Oregon almost ten years ago.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
Chapter One. Introduction ........................................................................ l
1.1 History .................................................................................... 1
1.2 Hypothesis ................................................................................ 1
1.3 Rationale for Objectives ................................................................ 2
1.4 Research Structure ...................................................................... 4
1.5 Research Benefit ........................................................................ 6
Chapter Two. Literature Review ................................................................. 7
2.1 Introduction .............................................................................. 7
2.2 Irrigation ................................................................................. 8
2.3 The History of irrigation in the Western United States ..... ‘ ....................... 9
2.4 US Bureau of Reclamation today ................................................... 12
2.5 USBR AgriMet weather stations .................................................... 14
2.6 Development of the FWB definition ................................................ 15
2.7 Determining the FWB ................................................................. 18
2.7a Direct measurement ......................................................... 18
2.7b Indirect FWB measurement ................................................ 21
2.8 Evapotranspiration ..................................................................... 24
2.9 Combination models .................................................................. 26
2.10 Penman Monteith model ............................................................ 28
2.11 Crop coefficients ..................................................................... 32
2.12 FWB and ET .......................................................................... 35
2.13 Soil water .............................................................................. 37
2.14 Soil particle analysis ................................................................. 40
2.15 Saturated hydraulic conductivity ................................................... 43
2.16 Unsaturated hydraulic conductivity ................................................ 47
2.17 Empirical infiltration models ....................................................... 48
2.18 Physically based infiltration models ............................................... 52
2.19 Parameter estimation ................................................................. 59
2.20 Layered soil profiles ................................................................. 68
2.21 Root Zone Water Quality Model ................................................... 73
2.22 Literature summary .................................................................. 80
2.23 Research needs ........................................................................ 87
2.24 General scope ......................................................................... 87
2.25 Current scope ......................................................................... 91
2.26 Research gap .......................................................................... 93

vi

Chapter 3. Study Area ............................................................................ 95

3.1 Willamette River Basin ....... i ........................................................ 95
3.2 WRB weather ........................................................................... 96
3.3 WRB Soils .............................................................................. 97
3.4 Soil capability classification ......................................................... 99
3.5 WRB dairy industry .................................................................. 100
3.6 Regulatory oversight ................................................................. 101
3.7 Study introduction .................................................................... 102
3.8 Dairy cooperators in the WRB ...................................................... 103
3.9 Cooperating dairy farm descriptions ............................................... 105
Chapter 4. Materials and Methods ............................................................ 109
4.1 Overview .............................................................................. 109
4.2 Installation and use of weather stations ........................................... 109
4.3 WS software ........................................................................... 113
4.4 Watermark conductivity sensors ................................................... 115
4.5 AgriMet data .......................................................................... 118
4.6 Resolution of WS data ............................................................... 119
4.7 Soil characterization .................................................................. 120
4.8 Soil coring field work ............................................................... 120
4.9 Laboratory work, particle size analysis ........................................... 123
4.10 Laboratory work, soil water characterization ................................... 125
4.11 SWCC equipment .................................................................. 127
4.12 Soil water extraction ............................................................... 128
4.13 Final extraction ..................................................................... 13 1
4.14 RZWQM data requirements ....................................................... 132
4.15 RZWQM data pararneterization ................................................... 136
4.16 RZWQM evaluation ................................................................ 137
Chapter Five. Results and Discussions ...................................................... 139
5.1 Overview .............................................................................. 139
5.2 AgriMet and WS data ............................................................... 139
5.3 Soil moist bulk density and particle size analysis ................................ 141
5.4 Soil water characteristic curve development ..................................... 144
5.5 Volumetric water content at saturation ............................................ 145
5.6 The SWCC complete curve ......................................................... 146
5.7 Volumetric water contents: empirical field data ................................. 162
5.8 RZWQM scenarios .................................................................. 168
5.9 Field 1-1 ............................................................................... 173
5.10 Field 1-2 .............................................................................. 176
5.11 Field 2-3 .............................................................................. 179
5.12 Field 2-4 .............................................................................. 182
5.13 Field 2-7 .............................................................................. 185
5.14 Field 2-8 .............................................................................. 187
5.15 Field 3-1 .............................................................................. 190

vii

5.16 Field 3-6 .............................................................................. 193

5.17 Field 3-7 .............................................................................. 195
5.18Field4-1....................... ....................................................... 198
5.19 Summary ............................................................................. 201
Chapter 6. Conclusions .......................................................................... 202
6.1 RZWQM evaluation and performance ............................................. 202
6.2 Statistical evaluation with Nash Sutcliffe Equation values ..................... 203
6.3 Objectives met ........................................................................ 207
6.4 RZWQM input requirements ....................................................... 208
6.5 Needs for future work ............................................................... 210
Bibliography ....................................................................................... 213

viii

LIST OF TABLES

Table 2.1 USDA textural classes and selected properties including bulk
density, porosity, and saturated hydraulic conductivity ............................... 42

Table 2.2 Saxton textural classes and related properties including bulk
density, porosity, and saturated hydraulic ............................................... 43

Table 2.3 SCS Hydrologic soil group infiltration rate as cm per hour
and description of infiltration ............................................................. 51

Table 2.4 UNSODA and Soil Survey hydraulic parameters including
water retention and saturated hydraulic conductivity values ......................... 66

Table 3.1 Cooperating farm physical descriptions including storage

capacity, animal units per storage, hectares per unit storage,
and hectares per animal unit .............................................................. 108

Table 5.1 Field bulk soil physical analyses, NASIS default, and

RZWQM default values .................................................................. 142
Table 5.2 Field linear and log equations for SWCC, with RSQ, by depth. . . . . . . ....159
Table 5.3 Field saturation, F C, and PWP volumetric water % by depth .................... 160
Table 5.4 Field l-l RZWQM simulation comparison and NSE model...... ................175
Table 5.5 Field 1-2 RZWQM simulation comparison and NSE model ..................... 178
Table 5.6 Field 2-3 RZWQM simulation comparison and NSE model ..................... 180
Table 5.7 Field 2-4 RZWQM simulation comparison and NSE model ..................... 184
Table 5.8 Field 2—7 RZWQM simulation comparison and NSE model ..................... 186
Table 5.9 Field 2-8 RZWQM simulation comparison and NSE model ..................... 189
Table 5.10 Field 3-1 RZWQM simulation comparison and NSE model ................... 192
Table 5.11 Field 3-6 RZWQM simulation comparison and NSE model ................... 194
Table 5.12 Field 3-7 RZWQM simulation comparison and NSE model ................... 197
Table 5.13 Field 4-1 RZWQM simulation comparison and NSE model ................... 200

ix

Table 6.1 Scenario sum, mean value, and standard deviation ranking ..................... 206

LIST OF FIGURES

Figure 2.1 Field Water Balance illustrating water inputs and outputs for a
crop system .................................................................................. 36

Figure 2.2 USDA textural class triangle illustrating location of twelve
classes based upon percent sand, silt, and clay content of bulk soil ................. 41

Figure 2.3 Green Ampt parameters and idealized illustration of water
content profile ............................................................................... 52

Figure 2.4 Idealized water release curve for sand, silt, and clay soils as
matric head versus volumetric water content ............................................ 61

Figure 2.5 UNSODA and Soil Survey fitted water release curves

(VG and BC parameters) for a silt loam soil ............................................ 66
Figure 3.1 Study site farm and weather station location map ................................ 108
Figure 4.1 Weather Station, Farm 4 ............................................................. l 13
Figure 4.2 Watermark sensor and hand held ohmmeter ....................................... 115
Figure 4.3 Soil core shell and soil ............................................................... 123
Figure 4.4 SWCC chamber. .' ..................................................................... 127
Figure 4.5 SWCC chamber and manifold ...................................................... 127
Figure 4.6 Soil cores on ceramic plate in chamber ............................................ 131

Figure 5.1 Volumetric water content (%) at saturation comparing calculated
versus measured, with linear equation and R2 value .................................. 146

Figure 5.2 Field l-l linear scale 0-0.003 MPa matric potential-volumetric water
content ....................................................................................... 148

Figure 5.3 Field l-l log scale 0.003-1.5 MPa matric potential-volumetric water
content ....................................................................................... 148

Figure 5 .4 Field 1-2 linear scale 0-0.003 MPa matric potential-volumetric water
content ....................................................................................... 149

xi

Figure 5.5 Field 1-2 log scale 0.003-1.5 MPa matric potential-volumetric water
content ....................................................................................... 149

Figure 5.6 Field 2-3 linear scale 0-0.003 MPa matric potential-volumetric water
content ...................................................................................... l 50

Figure 5.7 Field 2-3 log scale 0.003-1.5 MPa matric potential-volumetric water
content ....................................................................................... 150

Figure 5.8 Field 2-4 linear scale 0-0.003 MPa matric potential-volumetric water
content ....................................................................................... 151

Figure 5.9 Field 2—4 log scale 0003-1 .5 MPa matric potential-volumetric water
content ...................................................................................... 151

Figure 5.10 Field 2-7 linear scale 0-0.003 MPa matric potential-volumetric water
content ....................................................................................... 152

Figure 5.11 Field 2-7 log scale 0003-1 .5 MPa matric potential-volumetric water
content ...................................................................................... 152

Figure 5.12 Field 2-8 linear scale 0-0.003 MPa matric potential-volumetric water
content ...................................................................................... 153

Figure 5.13 Field 2-8 log scale 0.003-1.5 MPa matric potential-volumetric water
content ...................................................................................... 153

Figure 5.14 Field 3-1 linear scale 0-0.003 MPa matric potential-volumetric water
content ...................................................................................... 154

Figure 5.15 Field 3-1 log scale 0.003-1.5 MPa matric potential-volumetric water
content ...................................................................................... 154

Figure 5.16 Field 3-6 linear scale 0—0.003 MPa matric potential-volumetric water
content ...................................................................................... 155

Figure 5.17 Field 3-6 log scale 0.003-1.5 MPa matric potential-volumetric water
content ...................................................................................... 155

Figure 5.18 Field 3-7 linear scale 0-0.003 MPa matric potential-volumetric water
content ...................................................................................... 156

Figure 5.19 Field 3-7 log scale 0.003-1.5 MPa matric potential-volumetric water
content ...................................................................................... 156

xii

Figure 5.20 Field 4-1 linear scale 0-0.003 MPa matric potential-volumetric water
content ................................ , ....................................................... 157

Figure 5.21 Field 4-1 log scale 0.003-1.5 MPa matric potential-volumetric
water content ............................................................................... 157

Figure 5.22 Field scatterplot of volumetric water content % at saturation, FC,

Figure 5.23 Field l-l Watermark volumetric water content % for study
period, by depth163

Figure 5.24 Field 1-2 Watermark volumetric water content % for study
period, by depth ........................................................................... 163

Figure 5.25 Field 2-3 Watermark volumetric water content % for study
period, by depth ........................................................................... 164

Figure 5.26 Field 2-4 Watermark volumetric water content % for study
period, by depth ........................................................................... 164

Figure 5.27 Field 2-7 Watermark volumetric water content % for study
period, by depth ........................................................................... 165

Figure 5.28 Field 2-8 Watermark volumetric water content % for study
period, by depth ........................................................................... 165

Figure 5.29 Field 3-1 Watermark volumetric water content % for study
period, by depth ........................................................................... 166

Figure 5.30 Field 3-6 Watermark volumetric water content % for study
period, by depth ........................................................................... 166

Figure 5.31 Field 3-7 Watermark volumetric water content % for study
period, by depth ........................................................................... 167

Figure 5.32 Field 4-1 Watermark volumetric water content % for study
period, by depth ........................................................................... 167

Figure 5.33 Field 1-1 two hour empirical versus simulation volumetric
water content .............................................................................. 174

Figure 5.34 Field 1-1 12 hour empirical versus simulation volumetric
water content ............................................................................... 175

xiii

Figure 5.35 Field 1-2 two hour empirical versus simulation volumetric
watercontent........................ ....................................................... 177

Figure 5.36 Field 1-2 12 hour empirical versus simulation volumetric
water content ............................................................................... 177

Figure 5.37 Field 2-3 two hour empirical versus simulation volumetric
water content ............................................................................... 179

Figure 5.38 Field 2-3 12 hour empirical versus simulation volumetric
water content .............................................................................. 180

Figure 5.39 Field 2-4 two hour empirical versus simulation volumetric
water content ............................................................................... 183

Figure 5.40 Field 2-4 12 hour empirical versus simulation volumetric
water content .............................................................................. 183

Figure 5.41 Field 2-7 two hour empirical versus simulation volumetric
water content ............................................................................... 185

Figure 5.42 Field 2-7 12 hour empirical versus simulation volumetric
water content ............................................................................... 186

Figure 5.43 Field 2-8 two hour empirical versus simulation volumetric
water content ......................................................... ' ...................... 1 88

Figure 5.44 Field 2-8 12 hour empirical versus simulation volumetric
water content ............................................................................... 188

Figure 5.45 Field 3-1 two hour empirical versus simulation volumetric
water content ............................................................................... 191

Figure 5.46 Field 3-1 12 hour empirical versus simulation volumetric
water content ............................................................................... 191

Figure 5.47 Field 3-6 two hour empirical versus simulation volumetric
water content ............................................................................... 193

Figure 5.48 Field 3-6 12 hour empirical versus simulation volumetric
water content ................................................................................ 193

Figure 5.49 Field 3-7 two hour empirical versus simulation volumetric
water content ............................................................................... 196

xiv

Figure 5.50 Field 3-7 12 hour empirical versus simulation volumetric
watercontent........................ ...................................................... 196

Figure 5.51 Field 4-1 two hour empirical versus simulation volumetric
water content ............................................................................... 199

Figure 5.52 Field 4-1 12 hour empirical versus simulation volumetric
water content ............................................................................... 199

Figure 6.1 Field comparison Nash Sutcliffe Equation values ................................ 205

XV

Chapter One. Introduction

1.1 History

In what was to become a nine year odyssey, this research paper helps complete a
journey begun in June 1996. During the summer of that year representatives from
Oregon State University Extension Service (OSUE), Oregon Depamnent of Agriculture
(ODA), Oregon Department of Environmental Quality (DEQ), and several livestock
commodity groups met and talked about research needs for the dairy industry in the
Willamette Valley Oregon.

The charge was given to this researcher: develop a tool that would help
landowners estimate the field water balance, utilize water for crop growth that would
improve irrigation water management, and reduce the potential for surface runoff of

manure into surface water. A working hypothesis was developed.

1.2 Hypothesis

Evaluating Root Zone Water Quality Model (RZWQM) using field, laboratory,
and default input values, can lead to predicting the Field Water Balance for a crop field
receiving irrigation water and liquid manure. Model output can provide technicians a
new tool for estimating those periods during the year that may require greater attention to
irrigation water management, therefore improving water utilization on fields, and reduce
the potential for surface runoff, which meet recommended guidelines and enhance

environmental management of the water resource.

1.3 Rationale for Objectives

The rationale for using a model was based upon limited resources (costs of field
instrumentation, labor, and monitoring). The funding for this project, a DEQ 319 Grant
(USEPA Region 10, Seattle, WA, and Oregon DEQ, Salem, OR) was sufficient for the
installation of four weather stations and instrumenting ten fields with soil matric sensors.
The farms and specific fields were carefully identified. Four relatively stable dairy farms
were selected, and then on each farm specific fields were identified for instrumenting.
Campbell Scientific instruments were used for the WS package. The Irrometer
Watermark was selected for the matric potential measurements. The Watermark is a
granular matrix resistance block that mimics the volumetric water content in the soil. A
voltage is expressed across the block. The current flow results in a voltage drop, and the
resistance is measured. The resistance is a measure of moisture content. The drier the
block, the greater the resistance. After some period of data collection a FWB was
determined at four incremental 30 cm depths to 1.2 m. The F WB or volumetric water
content over time was the empirical data set used for this study.

The question for modeling was: what are the data input requirements for
evaluating the modeled FWB with the empirical data obtained in the ten instrumented
fields? Two objectives were proposed so that question could be answered:

1. Develop parameters for evaluating RZWQM, including three approaches:
a. USDA National Soil Information System (NASIS) data, based upon
selected Soil Map Unit identification for each field based on the Soil
Survey.

b. Particle size analysis, bulk density, and RZWQM default values.

c. Determination of the Soil Water Characterization Curve (SWCC) using
full data input of paired Volumetric water content and matric potential
values, in addition to particle size analysis and bulk density (complete soil
hydraulic and physical data).

2. Evaluate empirical FWB with Model FWB simulations using three methods of
model parameterization in objective one.

a. Empirical data are coupled data of field matric potential (Watermark) and
lab derived SWCC.

b. Model data are combinations of NASIS, RZWQM default, and complete
soil hydraulic and physical data

In terms of developing the parameters for evaluating the model, the NASIS data is
readily available on the intemet for public use. A user could identify a field location,
select the predominant soil map unit, and then use NASIS data for input soil hydraulic
and physical values in RZWQM. One advantage is no field work is required.

The second approach was using a combination of model default values and some
field measurements. The field measurements included particle size analysis (textural
class), moist bulk density, and volumetric water content at saturation, each at 30 cm
incremental depths. RZWQM default values included all other required hydraulic and
physical data to run the model simulations.

The third approach required full parameterization. The user overrides all default
values for hydraulic and physical properties with either field or laboratory data. In this
study the SWCC was derived for each 30 cm depths (four per field) to 1.2 m. The model

required volumetric water content at saturation, at 0.01 MPa, at 0.033 MPa (Field

Capacity), and 1.5 MPa (Permanent Wilting Point). The model requires horizon depth by
textural class with defined sand, silt, and clay content by percentage.

Saturated hydraulic conductivity was not measured in this study. The NASIS data
mean values were used for first approach; for the second and third approach the saturated
hydraulic values that gave no runoff were used. Landowners do not purposely apply an
amount of irrigation water or waste water that will cause a runoff event.

Once the model parameters were obtained or measured, the field simulations were
conducted using RZWQM. The model output as volumetric water content over the study
period were compared with the Watermark field data for the purpose of evaluating the
three approaches at determining the FWB in the instrumented field.

Four statistical analyses were used for evaluating model performance. The
goodness to fit analysis Root Mean Square Error and the model efficiency measure
known as the Nash Sutcliffe Equation were particularly useful in evaluating RZWQM for
these irrigated crop fields.

The basic research structure with chronological timeline is presented below. This
list provides the reader with a list of field, laboratory, and modeling activities that are

presented in this paper.

1.4 Research Structure
1. Select experimental study period: October 1997 - September 2002.

2. Identify four dairy farms in the Willamette Valley study area.

. Instnunent each field with Watermark soil moisture sensors at 30 cm depths to 1.2
m. Hard wire to data logger. Include weather instrumentation for a portion of the
study period.

. Record spatial and temporal water input volumes (manure and irrigation) on ten
fields found on the four dairy farms (as supplied by the landowners). Evaluate
delivery equipment in terms of volume flow per unit time.

. Measure physical soil characteristics of typical soils in the study area. These
include particle size distribution, bulk density, and SWCC at 30 cm depths to 1.2
m.

. Obtain climate data for the study area. Use AgriMet, National Weather Station,
and local weather data from instrumentation on each farm.

. Parameterize RZWQM with climate, field, and management data based on the
time period the farm had an operable weather station.

. Determine FWB for all fields. [Completes Objective One.]

. Evaluate RZWQM performance based on three methods of parameterization (Lab
only, Lab and RZWQM combination, and USDA-NASIS data only). Rank
scenarios using rainfall distribution as sensitivity analysis, then suggest input
parameters required for model use in other non instruments fields. [Completes
Objective Two.]

Finally, a discussion of why this research is necessary is presented below.

1.5 Research Benefit

This paper takes a step towards providing a list of input data for determining the
FWB on irrigated crop fields using a model. One further step will help landowners
implement irrigation water management practices during the growing season, and utilize
liquid manure during all periods of the year, based on knowledge of the FWB over a time
period. These improve water utilization and reduce the potential of groundwater and
surface water pollution to acceptable levels. The work fills a part of the research gap by
providing a model or tool for use in determining the volumetric water content over a
period of time without instrumenting a specific field or farm. Further, this work offers a
new approach at combining fragmentary data into a systems analysis in meeting the
research objectives. The effort is based on systems thinking across the SPAC, including
water, bridging the disciplines of soil physics, water flow, climate and environmental
phenomenon, plant physiology, and a comprehensive deterministic research model. The
key is the selection of RZWQM as the model of choice for this systems approach, and
identifying which input parameters are required for model calibration on non-

instrumented fields.

Chapter Two. Literature Review

2.1 Introduction

On Earth, life is dependent upon water. Democritus wrote in 400 BC. that “plant
growth involved the cycling of nutrients through water” (Tindall and Kunkel, 1999).
Later, in the 16th Century Sir Francis Bacon believed that water was the source of
nourishment for plants (Daumus, 195 8). Jan Baptiste van Helmont confirmed Bacon’s
work in his famous experiment, which he concluded water was the sole nutrient for plant
growth (Tindall and Kunkel, 1999). In the early 17th Century Robert Boyle stated that
plant growth was a function of salts and other materials found in water (Tindall and
Kunkel, 1999). Further work by Buckingham of England, Green and Ampt of Australia,
Widstoe of Utah, Gardner, Martin, Bouyoucus, Richards, Kirkham, Keen, Penman,
Bresler, and Hillel all made huge contributions to the understanding of water in soils
(Tindall and Kunkel, 1999). Today, advances in plant physiology have linked water in
the soil to water flow through plants. In photosynthesis, water is oxidized providing a
source of electrons and protons. These drive carboxylation of C02 and ultimately organic
carbon is made into plant tissue. Plants are cooled by water evaporation. The
transpiration ratio of water used versus dry mass retained in plants is 500 to 1,500
(Mengel and Kirkby, 2001). Water carries solutes from the root zone into the plant,
thereby transporting plant essential elements from soil to growing plant tissue.

Soil water is essential for plant growth. The evaporative process for cooling and
nutrient uptake for grth occurs only when soil water is sufficiently moving into root

tissue as a function of hydraulic suction, or negative pressure. In the plant rooting zone

soil moisture moves by the inputs of rain, irrigation, or capillarity, and the outputs of
evaporation, transpiration by plants, and the movement of soil water by gravimetric
potential, or deep leaching. The status of soil water at any point in time is the Field
Water Balance, FWB; it is dynamic both spatially and temporally.

For most of the Earth’s surface, the FWB is largely weather driven. The land use
will alter the FWB. Forests can consume large volumes of water given deep mating
zones of several meters. Perennial grasslands will consume water in a shallower rooting
zone (no greater than one meter), thus timely rainfall is required to maintain the FWB.
On the farm scale, crops are selected by climate zones and distribution of water
throughout the growing season. In rainfed agronomic systems, crop yield is a function of

timely rainfall, thereby maintaining the FWB.

2.2 Irrigation

Farmers can irrigate soils with supplemental water. This practice likely began
when groups of people, or communities, settled in places where rain did not supply water
to growing crops. Surface waters were diverted using hand-dug canals. Some form of
energy was used to draw water from hand-dug wells, thus soil water at depths beyond any
root zone were obtained.

With the development of machines and the movement of people to the Western
United States, entire societies were located in places where rain rarely fell. The darnming
of rivers, the construction of canals and pumping stations and reservoirs supplied water to
the desert. The FWB was maintained by diverted water, not rain. Irrigation turned

barren land into productive farms.

Productive farms are usually profitable farms. As the need to produce food for
humans and feed for livestock increases, the diversion of water to desert soils helps meet
this need. Yet an important question looms: is irrigation sustainable?

At the farm level, the landowner attempts to optimize yield, derive profit, and do
so efficiently. Some understanding of the FWB provides greater irrigation efficiency: the
timing, the location, and the amount of water applied to a crop field are important
decisions. For livestock farmers with animals in confinement, liquid manure may be
irrigated on the field. For this practice, special care is given to avoiding applications
when the F WB is sufficiently saturated. Irri gating water on soils at or greater than field
capacity move liquid manure out of the rooting zone before sufficient treatment can take
place. Surface runoff or deep leaching can occur. The livestock owner with a liquid
manure system, therefore, uses two criteria, which guide irrigation decisions: timing and

depth of application.

2.3 The History of irrigation in the Western United States

In a book dedicated to writing the reclamation history of the West, historian
Michael Robinson (1979) describes the development of irrigation on arid lands. The
manifestation of “agrarian ideals” by Thomas Jefferson set the stage for Federal Laws,
construction, and settlement of a new society. One of the first to do so was the Mormons.
They settled the Salt Lake Valley in 1847, digging irrigation canals by hand, and turning
the “Great American Desert” (Thomas, 1948) into productive fields sustaining an entire
farming community. The great trek West, including the Oregon and Californian Trails

during the 1880’s was brought about in part by the Mormon effort to reclaim the desert.

During this period of time, the reclamation effort was born. Congress sought to
encourage pioneers settlement in the West; the Carey Act of 1894 granted one million
acres to ten states if these lands were developed for agricultural irrigation.

The individual Western States, however, lacked the resources to fully develop the
scope of irrigation projects. State’s boundaries and water flow were different. The
Federal Government, led by President Teddy Roosevelt’s natural resources interest,
passed the Federal Reclamation Act in 1902 (Gates, 1968). During the late 19th Century,
an engineer, pioneer, and explorer, John Wesley Powell, trekked over much the West.
He mapped, made inventory, and conducted crude water flow measurements. He did not
like what he saw.

Powell’s report to Congress in 1878, titled “The Lands of the Arid Regions of the
United States (Smythe, 1911), advocated using science and a thorough inventory to
utilize water and land while avoiding overuse, non sustainable farming practices, and
monopolistic control. In what was to become known as a development plan for the West,
Powell suggested three components: a topographical map, measuring the annual volume
of water for each catchment or drainage areas, and conduct soil surveys (Robinson,
1979). Congress, however, largely ignored his foresight. His request for funding was
almost completely denied.

During the first years of the 20th Century, three men joined in an effort to build
support for Federal Reclamation in the West (Hodge, 1963). Frederick Newell, an
engineer and first head of the Reclamation Service, Francis Newland, a Congressman
from Nevada, and George Maxwell, a California Lawyer and founder of The National

Irrigation Congress in 1987 pushed through Congress a bill which funded the first six

10

reclamation projects in the West (Hollen, 1966). The second appropriations act supplied
the funds for another 25 projects, largely the formation of reservoirs by building Federal
Dams. The first, Roosevelt Dam (appropriately named after Teddy Roosevelt and his
Reclamation Service bill in 1902), was built on the Salt River in Arizona (Robinson,
1979)

The rapid increase of people settling in the West spurred Congress to provide
significant dollars towards reclaiming the West (Smith, 1950). Stegner (1954) coined the
term “Beyond the Hundredth Meridian,” as he described the efforts of the Federal
Reclamation Service to “tame” the Western Rivers and develop irrigation infrastructure.
He claimed the rationale was driven by an “engineering ethos” which simply stated that
humans can conquer Nature if given enough resources, time, and most of all, public will.
Engineers made up the upper hierarchy of the Federal Reclamation Service. They sold
their philosophy to politicians in the Eastern US. in economic terms. The newly
constructed dam would provide hydropower to new settlements in arid lands. The new
reservoir behind the dam would provide flood control and irrigation water. With power
and water, arid land would support agriculture and family farming (Robinson, 1979).
The original Reclamation Act of 1902 had important language which sought to avoid
monopolistic development of irrigated lands (Smith, 1950). The l60-acre farm or quarter
section (one quarter mile by one quarter mile square) was deemed enough irrigated land
to support one family unit.

In 1925 the Reclamation Service was renamed a branch service under the
Department of Interior, called the Bureau of Reclamation. The second, third and fourth

decade of the 20th Century were heady days for dam, reservoir, and irrigation

11

construction by the Bureau, including the Elephant Butte Dam in New Mexico (1920),
the Hoover Dam on then Colorado River (1935) and the Grand Coulee Dam on the
Columbia River (1941). All provided energy and power to rapidly growing settlements
of city and farm people (Robinson, 1979).

After WWII, the emphasis of the Bureau changed. In what is largely thought to
be a revisit of John Wesley Powell’s concern for water management, the Bureau sought
to place greater emphasis upon flood control, recreation, fish management, and improved
irrigation efficiency (Coddington, et al., 1972). Some of the last dams built by the
Bureau were on the Columbia River System. As a youngster in the late 1950’s my father
would take our family to a Columbia River bluff overlooking the construction site of the
McNary Dam. We watched workers manage huge buckets of concrete, lay thousands of
rebar rods, and install magnetic turbines in the bowels of this construction site. Fish
ladders and ship locks were built too, and in the later days of the construction, erections
of steel towers were set upon the expanse, thereby moving electricity hundreds of miles
south across the Eastern Oregon Basin. Soon after hundreds of center pivot irrigation

systems turned the flat brown desert into circles of potatoes, onions, corn, and alfalfa hay.

2.4 U.S Bureau of Reclamation today

The Bureau of Reclamation remade itself during the 50’s, 60’s and 70’s but
retained its theme, which according to Robinson (1979) is “To Conserve and Manage
Water in the Public Interest.” The changes, however, during these years up to the present
are dramatic. Instead of attempting to tame the Western Rivers, the Bureau attempts to

manage them efficiently, using water “more than once” (USBR, 2004). The Bureau

12

attempts to address the complex issues of water use and management in the West,
whereas in earlier years supplying water to reclaimed land was its emphasis. Today, the
Bureau of Reclamation makes the following verbatim statement on their Website (U SBR,

2004):

. Manages, develops, and protects water and related resources in an
environmentally and economically sound manner in the interest of the American
public.

. Serves as the fifth largest electric utility in the 17 Western States and the nation's
largest wholesale water supplier, administering 348 reservoirs with a total storage
capacity of 245 million acre-feet (an acre-foot, 325,851 gallons of water, supplies
enough water for a family of four for one year).

. Provides 1 out of 5 Western farmers (140,000) with irrigation water for 10 million
farmland acres that produce 60% of the nation's vegetables and 25% of its fruits
and nuts.

. Operates 58 hydroelectric power plants averaging 42 billion kilowatt-hours
annually.

. Delivers 10 trillion gallons of water to more than 31 million people each year.

. Manages in partnership 308 recreation sites visited by 90 million people a year.

The third bullet point, that of providing water for irrigation, is coupled with the
water efficiency objective; both were responsible for the Bureau’s development of
weather stations (WS) scattered throughout 17 Western States. In 1983, the Bureau

worked with Bonneville Power Administration (EPA) to develop a satellite based

13

weather station (WS) system for capturing climate and field characteristics (U SBR,
2004). The system is known as Agrimet, a word derived from Agriculture and
Meteorology. Today there are 74 WS in the Pacific Northwest (PNW) including
Washington, Oregon, Idaho, Montana, and parts of California, Nevada, Utah, and
Wyoming. The AgriMet Network (U SBR, 2004) is supported by a consortium of public
and private resources, including the United States Department of Agriculture’s Natural
Resources Conservation Service (NRCS) and Agriculture Research Service (ARS), Land
Grant Universities, the Cooperative Extension Service, and the Northwest Energy

Efficiency Alliance (NWEEA).

2.5 USBR AgriMet weather stations

AgriMet WS are equipped with below and above ground instruments. Data is
stored in a datalogger, and on five—minute intervals logged telemetry is relayed via
satellite to a mainframe computer in Boise Idaho. Complied mainframe data is posted to
USBR Website for each WS on the Network. A time delay of approximately four hours
exists for field data and complied web posted data. One calculation provided is
evapotranspiration, the sum of evaporation from the field surface and transpiration of
water through the growing crop. The value is listed as m or inches of water
consumption daily for a reference crop (ET,) that is well watered and growing. In the
more arid locations, alfalfa is used for the reference crop; in the West side of the Cascade
Mountain range, in Washington and Oregon, a grass crop is used. The calculation is

based upon a Modified Penman Monteith equation (FAO, 1998).

14

Irrigators may make better irrigation decisions using these data. Specifically, the
ETr is multiplied by a crop coefficient, which provides the irrigator with a volume depth
of water consumption on a daily basis. Irri gators can schedule a single irrigation volume
depth based on resupplying the crop field with water that is consumed by the crop ET, or
ETC. The FWB is, therefore, maintained using a mass balance approach: input
(irrigation) equals output (ETC).

The USBR’s AgriMet Network provides a valuable tool for irrigators where WS’s
are located. An estimate of water consumption for a particular crop must be known so
greater efficiency can be obtained. However, the water consumed by a particular crop is

just one component of the F WB.

2.6 Development of the FWB definition

Daniel Hillel, in his twentieth Chapter of Environmental Soil Physics, Water
Balance and Energy Balance in the Field (Hillel, 1998), defined the FWB in these words:
“The field water balance, like a financial statement of income and expenditures, is an
account of all quantities of water added to, subtracted from, and stored within a given
volume of soil during a specified amount of time.” He used the Law of Conservation of
Mass, which states matter (water in this case) cannot be created or destroyed but can only
change from one form to another. Hillel went on to explain the FWB in terms of energy
balance; that the Conservation of Matter and the Conservation of Energy are integrated
since they are involved in the same processes (form change accountability) within the
same physical volume (a volume of soil supporting plant growth). The root zone water

balance can be expressed as follows (Hillel, 1998):

15

S+V=(P+I+U)-(R+D+E+Tc)
Where
S = change in soil water storage per unit time
V = change in volume or mass of water incorporated into plant biomass per unit time
P = precipitation volume depth per unit time
I = irrigation volume depth per unit time
U = upward capillary flow per unit time
R = runoff volume depth per unit time
D = downward drainage out of the root zone (leaching) per unit time
B = evaporation from the soil surface
Tc = transpiration of water from the crop per unit time
Evett and Lascano (1993) suggested a FWB (SV, volumetric water) as follows:
SV=AS+P+R-F-ET
Where
AS = change in soil water storage
P = precipitation or irrigation volume depth
R = the sum of field runoff and runon
F = flux of water across the lower boundary of the profile
ET = water loss to the atmosphere including evaporation and transpiration
Both equations have commonalities. The S factor, the change in soil water is a
volume concept. The difference over a time unit, i.e. 24 hours, is a volumetric change.
The volume can be converted to a weight assuming soil water at 1.0 g cm”. In typical

root zone profile, the volumetric water content of the soil is multiplied by the volume

16

depth to yield a depth of water stored (Howell, 1995). For example, at a rooting depth of
60 em, if the volumetric water content is25%, or 0.25 cm water/cm soil depth, the depth
of water stored is 15 cm. This volume depth constantly changes depending upon the
other components of the F WB. Hillel (1998) refers to a change in water content for
plants in this rooting depth. Canny (1998) described water transport and water retained
over the growing season. He showed that water retention in a plant had several roles,
including stem turgor, solute transfer, leaf extension, and cell wall maintenance. The
whole plant water content of agronomic plants at the end of their vegetative growth
period is at least 80% and can be as high as 95% (Boyer, 1985).

The P component is water added to the soil root zone profile by precipitation or
irrigation. This value is measured and expressed as a volume depth, i.e. 7 mm of rain or
15 mm of irrigation water. The value is always zero or positive. The R component is
runoff or runoff and runon. The likely case is runoff from fields after a rainfall event of
some intensity. Runon can occur too, if surface flow from an adjacent field enters the
surface the area of this study field. The U and the F terms are somewhat similar. The
movement of water into the root zone from a layer at some depth below the zone occurs
due to capillary potential, or from a wet soil into a dry soil, the result of decreasing
negative matric potential (Ritchie, et al., 1999; Mandal, et al., 2002; Starks, et al., 2003).
This value is negative (water inflow) or positive (water outflow, or deep leaching).

The E component in the first equation refers to the amount of thermal energy,
which drives the evaporation of water from the soil. This is an energy term, expressed as
some measure of thermal heat, as in MJ m"2 day", the amount of energy entering the soil

surface one m square in one day (Bowers, et al., 1965; Sepaskhah et al., 1979; Peart, et

17

al., 1998. The E term in the second equation is the mass of water leaving the soil root
zone, thereby it is positive when water is evaporating from the soil surface and
transpiration is occurring.

The components of a F WB can be a derivative of a moment in time, or in integral
form, which describes the FWB flux over a time period. For example, a 24 hour period
in summer when no rain falls, but thermal input as radiant energy is significant, i.e. 26°
C, and a crop is in its vegetative state fully transpiring water from a soil root zone at or
near field capacity. In this case the change during the twenty-four hours is negative; the
change in water storage is decreased by 7 mm. The FWB can be positive the next day, if
10 mm of rain falls. This rainfall depth is balanced against the amount of water leaving
as evaporation and transpiration, and-the amount gained in plant tissue. The FWB can be
a combination of time steps, for one growing season, for one calendar year, or several

calendar years.

2.7 Determining the FWB
2.7a Direct measurement

The F WB can be measured by installing monolithic weighing lysimeters in a field
(Wright, 1991; Phene, et al., 1990). Large lysimeters (10 m diameter) can yield good
results (Howell, et al., 1995). A lysimeter is described in detail in a review paper by
Howell (1991). Work done by Grebet and Cuenka (1991) provided guidance
methodology for installing and maintaining lysimeters, which minimized error. They
report the largest error was differences between the lysimeters bulk soil in situ due to

disturbance and the field soil in its normal state. Differences between tillage primarily

18

and seed planting to some extent influences the error between lysimeters and field
conditions. Other workers report heat flux errors in smaller (30 cm) lysimeters (Black, et
al., 1968) and the high cost (Lourence and Moore, 1991), although Schneider et al.
(1996) described a low cost 1.5 m diameter monolithic lysimeter, which gave nearly as
accurate results. These low cost lysimeters were suggested for use in developing
countries.

There are two principal designs. One design (Marek et al. 1988) offers access to
lysimeters sides and bottom for installation and repair. The datalogger and related
instruments are installed in the cavity beneath the lysimeters, including the load cells
directly beneath the monolith soil container. The volumetric water content is recorded
with Time Domain Reflectometry. The excess water is drained into a sump hole and
removed from the cavity with a pump. A diameter of 3 m is common but the largest
lysimeters are 10 m in diameter located at the University of California — Davis.

The second design (Schneider, 1998) is simplified and less costly. The monolith
of soil (1.5 m diameter) is obtained from the field and installed in a hole slightly larger
than the monolith diameter. Prior to installation, the cavity is dug 0.5 m deeper and filled
with fine gravel. A volume of space is left for water to accumulate when the monolith
drains. This volume is removed by using a hand driven vacuum pump. The scale is
usually a platform beam unit resting on the gravel surface. The scale is wired to a
datalogger at some point above ground from the lysimeter. The soil monolith, once
installed, prohibits users from repairing the instruments and scales beneath the monolith.

Weighing lysimeters measure water mass, and over time, the change in water

content in the soil monolith and plant tissue. Users may assume water bulk density is

19

exactly 1 g cm3, thereby converting mass to volume. The change in water mass, over
time, and if negative, measures the water storage loss over that time. If negative, water
has been lost as evaporation, transpiration, or out of the bottom of the monolith. If
positive, a rainfall volume depth, or irrigation event, increase the mass, and therefore the
soil water storage is increased. The soil water change is known, and can be plotted as
time step (x) and storage status (y) (Evett, 1998).

The other components of a FWB are measured as well. Thermal energy provides
a measure of energy influencing the crop and the surface of the soil monolith. Wind
speed and relative humidity influence the rate of ET for the crop growing in the soil
monolith. The physical characteristics of the soil mineral, the matric, osmotic, and
gravimetric potentials are measured thereby providing knowledge of field capacity and
permanent wilting point, the volumetric water content between these two points is
available water capacity for crop consumption (Hillel, 1998). For example, if the FC is
44%, or 13.2 cm per 30 cm depth, and the PWP is 24%, or 7.2 cm 30 per cm depth, the
AWC is 6 cm per 30 cm depth. Optimal crop growth occurs between these two points;
the 6 cm or 60 mm of soil water storage is maintained by irrigation or rainfall.

Microlysimeters have been used at the field level (Lascano and Hatfield, 1992).
These smaller devices have a 10 cm diameter, can be removed by hand, weighed, and
then replaced in the field core. For a precision of 0.5 mm d], only seven units were
required at any soil wetness. The units are labor intensive, but offer the researcher a low

cost measurement of the FWB.

20

2.7b Indirect FWB measurement

The measurement of F WB can be done indirectly. A methodology for measuring
several of the components in any FWB equation can predict the change in soil water
status over a time period. As early as 1802, John Dalton (Davenport, 1984) suggested a
method for fiee water evaporation. The evaporation rate was calculated using the vapor
pressure gradient, the difference between saturated vapor pressure and actual vapor
pressure, and this factor was multiplied by an empirical coefficient to determine rate.
The Dalton equation follows:

E = C(eW — ea)
Where
E = evaporation rate
C = empirical coefficient
ew = saturation vapor pressure
ea = actual vapor pressure

Dalton proposed the use of an empirical coefficient, C, to refine vapor pressure
gradient (ew — ea). This gradient was influenced by air and water temperature, wind
velocity, atmospheric pressure as a function of altitude, and the orientation and shape
(slope) of the field surface (Tindall et al., 1999). Refinement of Dalton’s Law, as it was
known, and continued by Fitzgerald (1886), Meyer (1915), Horton (1917), Rohwer
(1931), Harbeck, et a1. (1958), Kohelker et a1. (1955), and Dingrnan, et a1. (1968).
Tindall et a1. (1999) provides a list of equations for each of these researchers. The newer
equations suggest models for determining the empirical coefficient, C. For instance,

Fitzgerald (1886) used wind velocity for C, Meyer ( 1915) suggested that E rates should

21

be based upon depth of water, and Horton (1917) in his rather well known work modified
C by incorporating the relationship of turbulent wind time and relative humidity over a
large area, i.e. and field instead of a single location in that field. Harbeck et al. (1958)
incorporated air temperature in his equation, and Kohler et al. (1955) began using pan
evaporation data collected at some site in the field.

The efforts of these researchers yielded some knowledge of evaporation of water
from a water surface and from the wet soil surface. Evaporation data like these were
indirectly influenced by location: if the evaluation site was not vegetated, the values were
truly evaporation estimates. If, however, the site was heavily vegetated, i.e. a forest,
grass or actively growing crop, the estimate certainly was influenced by the presence of
biomass.

Viehmeyer and Brooks (1954) suggested a near saturated soil surface will
evaporate water approximately at the same rate as water from a water surface. Meyer
(1915) assigned the E rate as a function of water depth, but included the corresponding
mean monthly vapor pressure (in inches mercury) and the mean monthly air temperature.
Thus the monthly evaporation rate was known. Meyer (1915) further refined his equation
by fine-tuning the coefficient C, by measuring wind velocity. A monthly mean was
determined using velocity as miles per hour at a height of 10 m above the evaluation site.
In what was to become known as the Meyer Equation (Tindall et al., 1999), the monthly
value could be divided by 30 to approximate daily evaporation assuming equation factors
were daily mean values. Meyer was, therefore, one of the earliest researchers to use
empirical data to calculate monthly and daily evaporation rates from a wet soil or water

surface.

22

One additional note on evaporation should include the use of the U.S. National
Weather Service’s use of Class A Evaporation Pans (James, 1988; Jensen, 1980; Pair et
al., 1983). The National Weather Bureau sought to know something about water flux,
that of an evaporative flux, or the conversion of liquid water into water vapor. This is
known as the latent evaporative flux; it is the amount of energy required to dissociate
water molecules, which are bonded into a matrix, into water molecules, which escape the
bonded matrix and enter the much less dense atmosphere. This reaction requires energy.
For example, at 20° C, the latent heat of evaporation is 2.45 M] kg", or in other words
2.45 M] of energy is required to vaporize 1 kg of water. The expression is usually
expressed as energy per unit area: 2.45 MJ of energy is required to vaporize 0.001 m or 1
mm of water per square m, or m2. The energy is supplied by sunlight through radiant
solar energy.

Class A Evaporation Pans were placed in over five hundred sites in the U.S. with
another several hundred in other countries (Cuenca, 1989). The pan is 1.2 m in diameter
and 250 mm deep. The operating level is 175-200 mm. The pan sits 150 mm above the
soil surface. Birds and debris are kept out of the water by placing a meshed screen on the
pan’s rim. Measurements are taken daily. Excess water from rain is removed from the
pan (and recorded); water is added when evaporation over time exceeds rainfall. An
automatic depth recorder can be used to automatic data collection.

The volume depth removed from a pan is combined with a coefficient, which
estimates the amount of evapotranspiration for a particular crop. Sample cr0p
coefficients are supplied by FAO (1998). For evaporation pans crop coefficients that are

less than one, indicating that less water is removed from the field than that water from a

23

water surface. The coefficient could be greater than one, however, if the soil was wet and
the crop was transpiring more water (a function of rooting depth) than that water fiom a

water surface (FAO. 1998).

2.8 Evapotranspiration

The idea that evapotranspiration could be more accurately calculated began in the
1940’s and was examined by many researchers in the 1950’s (Tindall et al., 1999).
Jensen et al. (1990) reviewed this early research effort and discovered three
methodologies: temperature methods, radiation methods, and combination methods.

The most commonly accepted temperature method is the Blaney-Criddle Method
(Blaney and Criddle, 1950). This method assumes that ET is a firnction of mean monthly
air temperatures and monthly percentage of daylight hours for an actively growing crop
in ‘a well-watered soil. The daylight hour component was determined by knowing the
degrees of North Latitude, i.e., the further away from the equator (larger Latitude) the
differences in daylight hour percentages in a 24-hour cycle. The percentages were
determined monthly. Blaney and Criddle (1962) developed monthly crop coefficients,
which served the same purpose as the pan evaporation coefficients. Three corrective
factors were made to the original BC Method. Pochop et al., (1984) recommended an
elevation factor, thereby in addition to the planimetric location, a vertical input was
needed. They proposed reducing ET by 10% for every elevation gain of 1,000 m, which
accounted for lower temperatures at higher elevations at a given level of solar radiation.
Lastly, Doorenbos and Pruitt (1977) modified the BC Method by using a standard

reference crop, or what is now known as ETC. They used a well watered, actively

24

growing grass crop as their ETC. Their work was the basis for the F A0 24 publication,
which during the 1970’s and 1980’s became the accepted method for determining ETc
using grass, and the development of crop coefficients for all other cr0ps.

The radiation method was originally suggested by Jensen and Haise in the late
1950’s (Jensen and Haise, 1963). The JH Method is based upon knowing solar radiation,
and the term they used is Langley’s, calories per square cm of field surface per day.
They borrowed the planimetric component from the BC Method, and developed a chart
of mean solar radiation in the listed units as a function of 5 degree increments both north
and south of the equator. The values are lower as latitude increases, in approximately the
same manner as daylight hour percentage (Tindall, 1999). Other components include
saturated and actual vapor saturation or humidity, mean air temperature, and a
temperature scale intercept when ET divided by R,, defined as Langley’s at that point
when no evaporation (not ET) occurs. Richardson and Wright (1984) provided a method
for determining solar radiation if local data were not available. One of the original JH
method authors, Jensen and other workers (Jensen et al., 1971), would later define the
crop coefficients, which would be used in almost all combination methods.

This early work by a group of scientists set the stage for development of what
were as combination methods. That is, temperature and solar radiation, both direct
measurements, were used together. In a well written review paper published in 1990,
Jensen, Burman, and Allen (Jensen et al., 1990) would review the work to that time and
suggest what would be called the FAO Penman Model, a combination model, based upon
the previous effort at combining temperature and solar radiation into a more accurate

model.

25

2.9 Combination models

Penman (1948) developed the first ET combination model. Known as the F AO
Penman Model, it combined four primary climatic factors: solar radiation as energy or
thermal input, wind speed in terms of turbulent intensity, vapor flux, as the difference
between saturated vapor pressure, or 100% relative humidity, and actual vapor pressure,
or relative humidity, and mean ambient air temperature. The thermal input term drives
the energy into the soil-atmosphere system. Logically, the greater the thermal input and
the longer the daylight hours (sunlight exposure), the greater the thermal input load per
24-hour period. Wind speed is required given the thermodynamic properties of turbulent
flow. Increased wind velocity drives turbulent flow, thereby removing water vapor from
the soil-atmosphere at a rate faster than still or calm air. The vapor flux, or what is
known as the vapor pressure deficit (VPD), is an evaporative property of the air space
above the soil surface and the soil atmosphere as well. The larger the VPD, the greater
the difference between existing humidity and atmospheric water content at saturation.
The VPD is a function of all climate terms, especially thermal input. The latent heat of
evaporation requires energy; the evaporative capacity of the atmosphere is the VPD.
Finally, ambient air temperature expressed as mean degrees per day is required to
complete the FAO Penman model. Temperature is a measure of energy in a system, and
the degrees to which energy increases or decreases per unit time.

Further refinements of Penman’s original model occurred in the 1950’s. 60’s and
70’s (George et al., 2002). Three modifications sought to incorporate stomatal resistance
(Penman, 1963), modify wind function (Monteith, 1965) and VPD’s (Doorenbos and

Pruitt, 1977). A scientist in France, Perrier (1977) moved the calculation of reference ET

26

towards more emphasis upon energy terms. His model used sensible heat and latent heat
exchange functions to describe the available energy for the actively growing crop canopy.
Bouchet (1963) added another perspective. He suggested that as a soil surface dries from
an initially wet or saturated condition, the potential ET increases while the actual ET
decreases. When the soil is nearly dry, there is no or little water to evaporate, even
thought the potential to evaporate water is high. He suggested the following relationship:
dEc + dETp = 0
Where
dEc = the change in ET caused by a change in soil water in the root zone
dETp = the change in potential ET caused by a change in actual ET
The equation is set to zero, inferring that when one increases the other decreases.
As a soil is rewet, for instance due to rain or irrigation, the actual ET increases due to
greater water availability in the root zone. Because more water is evaporating, i.e., the
latent heat flux is increasing, the increasing water vapor content of the atmosphere above
the actively growing crop reduces the VPD. This work was known as the complimentary
relationship between actual and potential evaporation. Many researchers shifted their
focus towards developing models that referenced a specific crop, as described by ET,.
The work in refining potential ET, ETp was not done yet. Scientists were
beginning to understand the necessity of refining crop characteristics (Pereira et a1, 1999).
Two components stood out that required definition. The first is crop canopy height, or
vertical distance of the uppermost biomass leaf layer from the field soil surface.
Certainly there are differences. Grass is less than one meter, soybeans, potatoes, and

sugar beets are just less than one meter, com is two to three meters, and then forest trees

27

can be many meters high. The actual ET for any given crop, ETC, would be larger as crop
height increases due to a greater vertical resistance of water vapor movement. The
second component is crop density. The term Leaf Area Index is defined as the one sided
green leaf area per unit ground area in broadleaf canopies, or as the projected needleleaf
area per unit ground area in needle canopies. As crops grow and biomass increases, the
LAI increases, thereby increasing the resistance of water vapor flow to the uppermost
distance in the canopy. Once these two components are established, aerodynamic
resistance from wind speed and average surface resistance (stomatal regulation and
canopy dimensional shape) can be measured for a particular crop. Two crops were

selected: grass and alfalfa.

2.10 Penman Monteith model

The original FAO Penman Model, coupled with an understanding of field soil
water status, and the aerodynamic relationships of crop vertical height and crop density,
was modified by Monteith in 1965 into what was know as the Penman Monteith ET
Model, or PM ET Model (Monteith, 1965). Ahnost 15 years would pass before the PM
Model would be used as the reference ETp model by the American Society of Civil
Engineers, ASCE (Itenfisu et al., 2003). The PM Model was also adopted by the Food
and Agriculture Organization of the United Nations in 1998 (FAO, 1998). The model
would have two names: the ASCE Standardized Reference PM ET Model, and the FAO
PM ET Model (Itenfisu et al., 2003). Two separate groups of scientists developed the
reference crop measurements. Doorenbos and Pruitt (1977) earlier had developed grass

crop ET calculations based upon a well-watered clipped cool season grass. Wright

28

(1982) worked at the Kimberly ARS Research Station in Idaho. The climate there is arid,
thereby requiring additional calibration cf the original PM ET Model. In fact much of the
Western U.S. uses a modified PM ET model with alfalfa as a reference crop. The model
was named after Wright’s work (in location terms, not name): the Kimberly Penman ET
Model.
The FAO PM ET Model reference equation for grass follows:
ET, = [0.408A(Rn — G) + MCn/(T + 273)u2(e, — ea)]/[A + M1 + Cdu2)]

Where
ET, = daily reference (grass) ET (mm day")
Rn = net radiation at reference crop surface (MJ rrr'2 day")
G = soil heat flux density (MJ rn'2 day")
T = mean daily temperature 1.5 m height above soil surface (°C)
u; = mean daily wind speed at 1.5 m height above soil surface (m s")
e, = mean daily saturated vapor pressure at 1.5 m above soil surface (kPa)
ea = mean daily actual vapor pressure 1.5 m above the soil surface (kPa)
A = slope of saturation vapor pressure temperature curve (kPa °C")
h = psychrometric constant (kPa °C")
Cn = numerator constant representing reference type and time step
Cd = denominator constant representing reference type and time step
Four assumptions are required for ET, based on grass. They are:

0 grass canopy height is 0.12 m, or 12 cm above the soil surface

0 a fixed surface resistance of 70 s rn'l

o albedo (surface color) of 0.23

29

o LAI greater than one, crop is actively growing, and all fluxes are vertical
For alfalfa as a reference crop two assMptions are different. They are:
0 alfalfa canopy height of 0.5 m, or 50 cm above the soil surface
0 fixed surface resistance of 45 s m'1
The FAO PM ET Model requires climate data (F A0, 1998). They are the same
climate data proposed by Penman in his earlier 1948 work. A measurement of solar
radiation is required as an input, Rn. Minimum and maximum ambient air temperatures
are required as measure of energy status, T. Wind speed, uz, provides a measure of
turbulent flow, thereby air turnover of drier air replacing wetter air. Finally, the
evaporative capacity of that air volume, es - ea, or VPD, is measured. One component of
the model, the heat flux density G, is assumed to be zero when the time step is one day.
A measurement of actual temperature changes in the soil may be used to alter the model
if the time step is many weeks or months.
If climate data were obtained in a field of grass or alfalfa, and the appropriate PM
ET model parameters were used, an accurate estimation of ET would be made. There are
two issues, however with these calculations. The first is the impact and accuracy of
weather data (Itenfisu et al., 2003). The need to provide accurate climate data as near to
the evaluation field is obvious. This is likely one reason why WS Networks are
expanding in the PNW (as well as Midwestern States) (U SBR, 2004). For each WS, the
calibration of sensors is a necessity, to the extent that electronic sensors are exposed to
climate extremes and must be monitored often.
The second challenge is born out of the word “estimate”. In a review article by

Itenfisu et a1. (2003), a group of combination models were used to calculate ET, for grass

30

and alfalfa at 49 locations in the U.S. They found good agreement using the ASCE PM
ET Model at locations across diverse climates in the U.S. However, given the differences
of climate, with the use of several equations to model reference ET, the root mean square
difference values comparing these models ranged from as low as 0.04 mm day'1 to a high
of 0.76 mm day". These results underscore the variations of models to estimate reference
ET. The RMSD as a percentage of mean daily ET ranged from a low of 0.8% to a high
of 22.2%. Of all models, the ASCE (and FAO) PM ET Model performed the best using
grass as a reference crop. The importance of this work suggests that on a field basis, the
calculation of reference ET is challenging and the results are estimates.

If the crop is something other than grass or alfalfa, another adjustment must be
made. The challenge of making this adjustment is huge. For a user to know ET for corn
when the equation is based upon a cool season grass, requires extensive crop
physiological knowledge in terms of water consumption and biomass growth (from
which TR can be modeled), as well as some understanding of the FWB. Recall that the
PM ET Model requires a wet soil so water can be evaporated and water can be transpired
(Bouchet, 1963).

Crop coefficients have been developed by workers in fields the world over (FAO,
1998). These are known as crop ET values, or ETC, and specifically model particular

crop water consumption over the growing season or some evaluation time period.

31

2.11 Crop coefficients

A crop coefficient is defined as a dimensionless number which is multiplied by
the reference ET, ET,, to obtain crop ET, ETc (FAO, 1998). The formula takes on this
flow

ETc = ET, * Kc
Where
ETc = Evapotranspiration for a particular crop c (m day")
ET, = Evapotranspiration for a reference crop (alfalfa or cool season grass) (mm day")
Kc = crop coefficient (dimensionless)
The crop coefficient Kc integrates four crop characteristics that are, at any one point
during crop growth, different from the reference crop, i.e. alfalfa or cool season grass
(FAO, 1998). They are:

0 Crop height, which influences the aerodynamic resistance term U2 in the FAO PM
ET model.

0 Albedo, or the reflectance of the crop soil surface. A wet soil surface is darker,
thereby absorbing more energy into the soil and reflecting less into the
atmosphere. This term influences the Rn —- G terms in the FAO PM ET model.

0 Canopy resistance, which influences the surface resistance term derived by es — e,,
and is affected by LAI, leaf area (stomatal density), and physiological degree of
stomatal control (specifically, the location of the stomata in reference to the leaf
surface i.e.).

o Evaporation from soil especially exposed when the crop canopy is minimal, as a

LAI less that 0.5.

32

Crop coefficients are given as standard and non-standard numbers (F A0, 1998).
Standard conditions include no limitations on crop growth. Soil water is moving through
the plant in sufficient amounts to regulate leaf temperature as a function of solar
radiation. Sufficient crop density, good weed control, minimal insect activity, and no soil
fertility problems (macro or micro nutrient deficiencies, salinity, or soil pH) are apparent
in the field. Non-standard conditions represent a limitation on optimal crop yield,
including the presence of any environmental or characteristic that would limit or
significantly the crop yield in relationship to the crop coefficient.

A good reference source for crop coefficients is found in F AO Irrigation and
Drainage Paper No. 56. Researchers in all parts of the world have developed crop
coefficients for food and feed cr0ps. Other sources include Smith etal., 1991, and Allen
et al., 1998. Pereira et al. (1999) discussed some fine-tuning efforts by workers (Allen et
al., 1996) in adjusting crop coefficients during especially the mid growing season for
agronomic crops. Wright (1982) sought to adopt what he called baseline crop
coefficients by modifying the Kc value as a function of time period after an irrigation or
rainfall event. He suggested the term soil water stress coefficient, Ks, which would
account for a longer period of time after an irrigation or rainfall event. Logically this
term, if higher, was the result of crop stress, and that stress was the result of soil water
depletion in the root zone without replacing it with irrigation or timely rainfall. Allen
(1996) suggested accounting for crop stress as a function of the relative evaporation from
the soil surface as a layer. This approach was derived by knowing a firndamental

property of soil water change over time using the Law of Conservation of Mass.

33

Work by Ritchie (1972) suggested a different approach. He proposed defining a
crop coefficient by measuring crop density and water availability in the root zone. His
formula is:

ET, = f 1(9)f 2(LAI)ETr
Where
ET, = ET for a selected crop
f1(0) = volumetric water content in the crop root zone
f2(LAI) = LAI, as a function of crop density
ET, = reference ET using alfalfa or cool season grass
Function f1 and f 2 are soil specific and crop specific, and both are influenced by climate
factors, farming practices, and crop density. One final approach is the concept of a
composite crop coefficient. Pereira et al. (1999) suggested the following equation:

ET, = a,a,ET,

Where ET, = ET for a selected crop

or, = term expresses the influence of climate upon aerodynamic resistance

or, = term explains the crop specificity when compared to the reference crop, or how
similar or different is the selected cr0p with the reference crop of alfalfa or cool season
grass. Therefore,

Kc =aoac
Where

K, = a selected crop coefficient

34

Pereira et al. (1999) state this final equation as the promise for future work. For now, the
referenced source of crop coefficients is found in FAO Irrigation and Drainage Paper NO.

56 (FAQ, 1998).

2.12 FWB and ET

At this point, the discussion of FWB has included direct measurement using some
form of lysimeter equipment, and an indirect measurement using climate data. However,
in both measurements, there is one unknown: the relative proportions of the solid and
void space in the soil volume. A lysimeter measurement is based upon an understanding
of soil particle size, mass of soil per unit volume, or bulk density, and the relative
proportions of gravimetric, matric, and osmotic potential which can be used to derive the
soil water characteristic curve. Two additional components are required: an
understanding of infiltration as a function of water movement into the soil volume, and
two, the conductivity of water through the soil volume during matric flow and during
gravimetric flow.

The FWB, therefore, is influenced by the climate factors used in determining ET,
but can only be known if the fundamental properties of the soil root zone are known.
Farmers can have access to reference ET values, and based upon their crop can know
something about the adjustments made to ET, by using a crop coefficient. But to what
extent does ET, influence the volume of water in the crop root zone?

The FWB is known only when the parameters of the crop root zone are known;
that volume of soil containing soil particles, void space, some volume of water, some root

tissue, will have a volume of water which is capable of moving into the root tissue, into

35

the vertical xylem canals, and through the stomata via evaporative flux. This volume of
soil will also lose water due to surface evaporative flux, a function of solar radiation and
the presence of crop biomass.

An example of the F WB inflows and outflows are presented in Figure 2.1.

Rainfall Irrigation

Evapou'ansprratron' '

 

a F

 

 

Runoff

I

 

 

 

Chang es In
5011- water a
storage

' . Bottom of
root zone

Deep U ward of mature
percolation ow plant

Figure 2.1 Field Water Balance illustrating water inputs and outputs for a crop system

In figure 2.1, the crop root zone is defined at some depth of soil from the soil
surface. The FWB is known only when the change in soil water storage is quantified.
The values of rainfall and irrigation can be measured with relative ease (in units of
volume depth, or m of water). Crop ET, ET, is calculated from a reference ET, ET,,
and a crop coefficient, K,. Runoff, or water on the soil surface which has not entered the

soil void space, can be calculated by knowing infiltration rate and soil bulk density of the

36

soil surface layer, and the rate of water as an input, i.e., a hydrograph of rainfall intensity
or irrigation rate per unit time.

Below the crop root zone is a leaching zone, in which water may move upward
against the gravimetric potential, or water may enter from the crop root zone in response
to a lower hydraulic potential. The below ground component measurements are
problematic. Defining changes in soil water storage, or volumetric water content per unit
depth, are required, however, if the FWB is to be known. The above ground climate
measurements will predict water moving out of the crop root zone, in units of volume
depth, m day", but this value must be referenced to a unit of volume per volume depth
of soil. If the crop root depth had units of measurement similar to water, i.e. 1 Mg m'3,

then the volume depth removal would be accurately known.

2.13 Soil water

Gravimetric water content of soil can be expressed as the ratio of the mass of
water in the soil sample before drying to the mass of the soil sample after it has been
dried to a static mass at 105°C (Topp et al., 2002). The soil sample must be taken in as
undisturbed manner as possible, thereby minimizing sample compaction. The water
volume is expressed as a percentage of total volume. The mass based gravimetric water
content is related to the volume based water content by measuring the soil bulk density,
p, in units of Mg m3, and the density of water pw, in units of Mg rn'3 (Topp et al., 2002).
The volumetric water content equation is:

0, = (W pw )0".

Where

37

0,, = volumetric water content
0,, = gravimetric water content

The bulk density of soil particles, p, is given as 2.65 Mg rn'3 (Zavattaro and
Grignani, 2001; Logsdon and Cambardella, 2000; and Grossman and Reinsch, 2002). A
recent paper by Flint and Flint (2002) described soil particle bulk density in detail. They
provide a list of soil particle densities; selected materials are presented below:

Apatite 3.2 g cm'3 (or Mg m'3)

Montrnorillonite 2.5

Kaolinite 2.65
Feldspar 2.5 — 2.8
Mica 2.6 — 3.2
Biotite 2.7 — 3.1
Quartz (sand) 2.65

The bulk density of water, pw is given as 1 Mg In3 at 4°C (Hillel, 1998), and
decreases as temperature increases. For instance at a temperature of 25°C, the bulk
density of water is 0.99708 Mg m3, and decrease of 2.92 kg m'3, or about 0.29% less
mass per cubic meter.

Soil bulk density is defined as the mass of soil per unit volume. Typical values
ranges are 0.9 -— 1.3 (top 20 mm) and 1.2 — 1.6 Mg m'3 for agricultural soils (Logsdon and
Cambaredella, 2000), 1.4 to 1.52 (Ap Horizon) and 1.3 Mg m“3 (> 20 cm) for agricultural
soils (Zavattaro and Crignani, 2001), 1.41 — 1.56 Mg rrr'3 on agricultural soils (Mandal, et
al., 2002), and an average of 1.35 Mg rn'3 on tilled Ap horizons (Rawls, et al., 1998).

Soil bulk density was variable given the non-uniform spatial regularities of the soil

38

volume (Grossman and Reinsch, 2002). The authors list a number of studies reporting
variability but no clear range or coefficient of variation around the mean values. They do
report, however, their own study of 5,000 samples moving through the USDA Soil
Survey Laboratory. The standard deviation of duplicate soil samples was 0.04 g cm’3
(0.04 Mg m”).

Grossman and Reinsch (2002) discuss the shrinkage of soils as they undergo
wetting and drying cycles. The shrinkage of sandy soils was negligible; clayey soils will
have measurable shrinkage. Grossman (1990) presented a series of equations that
convert bulk densities to a standard value at 33 kPa soil matric potential (field capacity).
Grossman recommends that from a practical matter, soils should be sampled at or near
field capacity, and the stated bulk density value referenced as such.

Once the bulk density of the soil particles is determined (or assumed to be 2.65
Mg m3) and the bulk density of the soil mass is determined, the void space or porosity, f
is known using the following equation:

f (porosity) = 1 —- (ps/pb)

Where

f = porosity as a dimensionless value less than one

ps = bulk density of soil volume (Mg m")

pb = bulk density of soil particles (2.65 Mg rrr'3 f (porosity) = 1)

For example, if the soil volume bulk density is 1.35 Mg m'3, then the porosity f is equal
to:

f (porosity) = 1 — (1.35/2.65) = 0.4906

39

The porosity value of 0.4906 may be used two ways. One, the soil volume is multiplied
by 0.4906 to yield the void space, i.e. a cubic meter of soil has a void space of 490.6 L,
equivalent to 49.06 % of the total volume. Two, this value is a hypothetical value of the
volume of water, which can replace the void space. The value is never attainable in the
field given the entrapment of air particles as the soil volume is rewet (Hillel, 1998).
From a practical standpoint the porosity of a soil volume in the field tells us something
about its compaction, and the volume of water that could enter the void space.

The measurement of bulk density on agricultural soils has several other useful
purposes. The ease of root penetration (Pierce et al., 1983), prediction of water
conductivity (Rawls et al., 1998; Wildenschild et al., 2001; and Saxton, 1986), and an
indicator of soil quality (Lal et al. 1998; and Larson and Pierce, 1994).

Soil bulk density can be measured by driving a soil core into the soil matrix at the
selected depth. The driving function is done without compacting the soil entering the

core. Field cores can be taken using Soil Moisture® (2004) sampler and brass cores.

2.14 Soil particle analysis

Particle size analysis, PSA is a measurement of the various sizes of particles
comprising the soil solids, or soil matrix (Gee and Or, 2002). The USDA system of
classification (Soil Survey Staff, 1975) for mineral soils divides particles sizes into sand,
silt and clay for particles less than 2000 pm (2 mm) as follows:

0 Sand, < 2000 — 50 pm

0 Silt, < 50 — 2 pm

0 Clay, < 2 um

40

Particles greater than 2000 um are classified as gravel and cobbles.

PSA is used to evaluate textural class of mineral soils based upon the mixture of
sand, silt, and clay. The Soil Survey Staff (1975) developed a textural triangle that
placed 12 textural classes within the triangle based upon relative mixture of particles size.

The USDA textural class triangle is shown in Figure 2.2.

100

’0

80

 

 

 

 

40 cla
A IV clay loam Vsilty clay ’9
30 V V v Ioalm
/\ EA /\ A A A
sand clay loam q?
20 ¢ A
loam _ ‘ c9
sandy loam sutt loam '
10 V ‘ V
QQ

Aw. « v ~
\ \ \ \ \ \ \ \ \ \

A
gee o‘é%%%%6
IErcemsand

Figure 2.2 USDA textural class triangle illustrating location of twelve classes based upon
percent sand, silt, and clay content of bulk soil

The PSA has been used to predict saturated and unsaturated hydraulic
conductivity (Todd, 1964; Bloernmen, 1980). The PSA and bulk density have been used
to predict water retention, or water release curves, and unsaturated hydraulic conductivity
(Arya, 1999; Saxton, 1998; Arya et al., 1981; Arya et al., 1999; Assouline, et al., 1998;

and Abuja, et al., 1980; Todd, 1964; and Bloemmen, 1980). Table 2.1 lists the 12 USDA

41

textural classes, typical bulk densities, total porosity, and typical saturated hydraulic

conductivities (Ks) (Ahuj a, et al., 1998)}

Table 2.1 USDA textural classes and selected properties including bulk density, porosity,
and saturated hydraulic conductivity

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Textural Class bulk density porosity Ks
Mg rn'3 % cm h'
sand 1.49 43.7 21.0
loamy sand 1.49 43.7 6.11
sandy loam 1.45 45.3 2.59
loam 1.42 46.3 1.32
silty loam 1.32 50.1 0.68
silt 1.35 49.1 0.68
sandy clay loam 1.59 39.8 0.43
clay loam 1.42 46.4 0.23
silty clay loam 1.40 47.1 0.15
sandy clay loam 1.51 43.0 0.12
silty clay 1.38 47.9 0.09
clay 1.39 47.5 0.06

 

Saxton (1998) developed a Soil Water Characteristic Tool based upon textural

class. The values he derived are presented in Table 2.2.

42

 

Table 2.2 Saxton textural classes and related properties including bulk density, porosity,
and saturated hydraulic conductivity

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Textural Class bulk density porosity Ks
Mgn'3 % cm h’1

sand 1.68 36.6 5.01
loamy sand 1.63 38.5 3.18
sandy loam 1.53 42.3 1.51
loam 1.44 45.7 0.89
silty loam 1.41 46.8 1.41
silt 1.49 43.8 3.22
sandy clay loam 1.40 47.2 0.27
clay loam 1.32 50.2 0.25
silty clay loam 1.29 51.3 0.39
sandy clay loam 1.33 49.8 0.12
silty clay 1.24 53.2 0.24
clay 1.26 52.5 0.15

 

 

 

 

 

2.15 Saturated hydraulic conductivity

The relationship between soil particle size and bulk density describes the
arrangement of soil particles per unit space, or their distribution in a volume of soil.
From these, porosity is known. Total porosity of a soil volume indicates volumetric
water content if the soil is saturated, except for captured air in prose space. If a soil
volume is saturated, i.e. the pore space is 90-95% filled with water, the volumetric water
content of soil cannot be increased as more water is added to the soil surface. If a
hydraulic head of water exists at the soil surface, then water flow through the soil volume
is saturated flow, or what is known as saturated hydraulic flow.

Flow through a saturated soil volume is driven by gravimetric and pressure
potentials (Tindall and Kunkel, 1999). Gravity exerts a potential energy on a volume
basis in units of Pa or on a mass basis in units of J kg". (Note, 1 Pa = 103 J kg'l). The
gravimetric potential is, therefore, the amount of work required to move a mass of water

to some point from a reference datum (Young and Sissson, 1999). Work done by Darcy

43

(Philips, 1995) sought to define saturated flow in sands, and he proposed the following
expression (Darcy’s Law):

Q = K(AAH)/L
Where
Q = volume of water that passes through a depth of soil per unit time, cm3 h'1
K = hydraulic conductivity in cm sec'l
A = cross section area of soil plane, cm'2
AH = datum reference point change, head plus depth of soil, cm
L = length of soil depth, cm

If the change in head, AH is zero, then there is no flow. The head drop per length
in the distance of flow (AH/L) is the hydraulic gradient (Hillel, 1998). The specific
discharge rate is known as the flux density (Q/A) and is indicated as q (small Q). The
units of q are length per unit time, or cm h”. The units in Tables 1 and 2 list q in these
units. This flux density q, is used to describe saturated conductivity while the large Q
refers to the volume of water moving through the soil depth at a particular flux density
and over a surface area, A. The head, or hydraulic gradient is placed in the numerator
while the length or depth of the soil profile is placed in the denominator. A larger
hydraulic gradient, i.e. one m instead of one cm, significantly increases q as a function of
head.

Saturated flow of soil water through a volume of soil may be vertical and
horizontal from the soil surface (Young and Sissson, 1999). Due to the heterogeneity of
soils and soil particle distribution, which increase or decrease tortuosity per unit length,

water movement through soils is termed anisotropic (Tindell and Kunkel, 1999).

Generally, water flux density q, is greatest when flow is parallel to plate shaped clay
particles, and least when water moves perpendicular to these plates. A third flow may be
described as orthogonal towards the clay particle. Work done by Fetter (1994) showed
that flux density of water flow in saturated soil can vary by more than two orders of
magnitude. His work showed that Q, or volume flow reflects many different q’s, or flux
density flows.

Saturated hydraulic conductivity obeys the Conservation Law of Mass, or what is
known as the continuity law. Simply, it states a water volume entering the soil volume is
equal to the volume leaving the soil volume. If the volume leaving is less, then an
increase in volumetric water content occurs, and therefore water flow in the soil volume
is not saturated flow. If water leaving the soil volume exceeds water entering, then a
decrease of volumetric water content exists. These conditions are based upon a time
period; inflows and outflows are expressed as volumes at some volume depth for a unit of
time. The FWB is based exactly upon the continuity equation; otherwise a change in
volumetric water based upon inflow and outflow of water could not be determined.

The soil volume may be the root zone or depth to a water table or depth to some
soil profile based upon soil morphology or soil physical characteristics. Saturated water
flow may be described as flow of water through multiple layers or soil horizons.
Typically soil layers or horizons may be heterogeneous to the extent different flux
densities occur throughout the complete soil volume. Each layer will have a different
saturated (and unsaturated) hydraulic conductivity. As such, the total soil volume can be
modeled as an anisotropic homogeneous layer or horizon (Freeze and Cherry, 1979). The

entire soil volume is comprised of multiple soil layers jn, each with an individual q (or K),

45

or in the case of a specified surface area, Q, and each concurrent layer has its own
hydraulic resistance R, where

R = dj/Kj
Where
R = hydraulic water resistance for layer j
dj = thickness of layer j
Kj = hydraulic conductivity of layer j

The summation of all layers Ed is the entire soil thickness; the R value is the
resistance for each layer. K2, the effective saturated hydraulic conductivity for the entire
soil thickness is defined as follows (Tindall and Kunkel, 1999):

K. = av Ere/K.)
Where
Kz = saturated hydraulic conductivity of water through the entire soil thickness, cm hr-l
{id = summation of all soil layers
adj/K1) = summation of all R values of each individual soil layer, j

A comprehensive review of determining saturated hydraulic conductivity in soils
by using laboratory and field techniques are given in the recently published Methods of
Soil Analysis, Part 4, Physical Methods (Dane and Topp, 2002). The laboratory methods
include using an undisturbed (as possible) intact soil core, and placing a head of water at
some depth above the core. After equilibrium (constant or steady state flow), the volume
is collected. The result is volume per unit time, or Q, whereas the flux density is known

by removing area A, from the equation, with units as cm h".

46

Field methods include ring infiltrometers (Youngs et al., 1997), concentric ring
infiltrometers (Bouwer, 1986), pressure infiltrometers (Reynolds, 2000), multiple ring
infiltrometers (Reynolds, 1993), constant head well permearneter (also known as the
borehole perrneameter method) (Elrick et al., 1989), and the auger hole method
(Amoozeger and Warrick, 1986). Amoozeger and Warrick refined some early work by
Kirkharn (1945) and suggested what is known as the piezometer method. This method is
based upon a inserting a solid wall tube inside an auger hole at a depth in the soil profile.
Water is added until a steady state is reached, thereby a field estimate of water volume
per unit time which leaves the tube is the saturated hydraulic conductivity. Individual
soil layers or horizons can be fitted with multiple piezometers at the appropriate depths.
The most commonly used methods are the auger hole and piezometer methods. Their
performance is based on assuming the soil water is moving as saturated flow, thus the soil

pore space is almost completely filled with water.

2.16 Unsaturated hydraulic conductivity

Movement of water into the soil volume from the soil surface is known as
infiltration. Water entering the soil volume must replace air space. Infiltration occurs,
therefore, when the soil volume has some air in its void space. As water enters the soil
volume, the movement of water is lmown as unsaturated hydraulic conductivity. For the
FWB, the rate of flow per unit time is important; changes in soil volumetric content are
influenced by the rate (or ease) water moves through the soil. The driving force is the
physical parameter known as pressure gradient (Clothier et al., 1983). Pressure gradients,

or potentials, are the differences in one pressure locale from another. For instance the

47

pressure gradient decreases from a soil root zone at F C (higher) to a tree leaf 10 m above
the soil surface (much lower). The Second Law of Thermodynamics explains this
gradient. Water always moves to a low-pressure from a higher pressure. When water
enters the soil, the pressure potential is lowered.

Unsaturated flow may be influenced by gravity. The downward “pull” of water is
a force that transpiration is always working against. As water flow is moving into the
root, through the plant xylem and ultimately converted to vapor at the stomatal aperture,
every step of the way gravity is slowing this flow. The water potential or transpiration
gradient, therefore, must exert enough force (or greater gradient) to overcome this
constant downward pressure (Taiz and Zeiger, 2002). For the FWB, gravity can work in
favor by exerting an additional force pulling water into the soil volume, thereby

encouraging infiltration.

2.17 Empirical infiltration models

Infiltration can be defined using empirical equations derived from actual
measurements of cumulative water infiltration. Several intake rates, q are measured
using a ponded water depth at the soil surface. Several measurements are made based
upon soil physical characteristics (like textural class), from which a rate constant per unit
time are derived. One of the first empirical equations was suggested by Kostiakov (1932)
as infiltration:
I = at°

Where

I = cumulative depth as the volume of water per unit soil surface area

48

a and b = empirical parameters from field measurement
t = elapsed time

Kostiakov’s equation disregarded the initial water content of the soil and did not
perform well when the intake rate, i was long. Parlang et a1. (1982) examined several
soils’ intake rate using the Kostiakov equation and found it worked well when the
infiltration rate was fast and the soil volume was relatively dry.

The Horton (1940) equation was proposed as an improvement for the Kostiakov
equation. It is also based upon an exponential decay constant.

I = m + [00 - ir)/kl/(1 -e 4“)
And the intake rate as:

i= if+ (iO—it) e 4“
Where
I = cumulative depth as the volume of water per unit soil surface area
t = elapsed time
if = final infiltration rate based upon steady state
i0 = zero time
k = proportionality constant

The values i0, if, and k are related to soil and vegetative properties (Tindall and
Kunkel, 1998). Skaggs and Khaleel (1982) examined the Horton equation and suggested
typical values for agricultural soils: 30 — 90 cm h'1 for i0 (beginning infiltration), less than
1 to 30 cm h'1 for if (steady state infiltration) and up to 50 cm h'1 for k.

The USDA’s Soil Conservation Service (USDA, 1972) proposed another

empirical infiltration equation in what is now known as the SCS Equation. Scientists at

49

SCS (as it was known until 1992) were interested in developing rainfall runoff model,
which could be applied in all parts of the US. Their focus was reducing the impact of
water erosion on soils. Thus the runoff value R was used in the following equation:
R = [(P —- 0.2fw)2[/(P + 0.8a.)
“mae
R=runoffinunitdepthmm
Fw = initial soil water content
P = precipitation in depth m
The SCS equation derived a runoff depth based upon initial soil water content,
and therefore used infiltration as precipitation minus runoff. The SCS equation assumed
a constant rainfall hydro graph with an initial abstraction, or water storage, of 25 percent
of the rainfall event before runoff would occur. Logically, if the soil volume was near FC
or saturated, the FW term would be at or near zero, indicating an earlier and more
pronounced runoff volume (Rawls et al., 1982).
Holton (1961) suggested yet another empirical equation which derived infiltration
completely proportional upon the soil water content of the receiving soil. His equation is:
i, = i, + ab(or— I)”
Where
i, = infiltration rate
if = final infiltration rate
a = constant related to surface conditions varying between 0.25 — 0.8

b = scaling factor

50

or = water deficit as the pore space per unit area of a cross section initially available for
water, cm
I = accumulative water volume at t
t = time

Holton’s work was an improvement over previous equations and models.
Infiltration could be described during periods of little or no rainfall (Tindall and Kunkel,
1999). Holton et al. (1975) firrther modified his equation as follows:

i = GAI,” + i,
Where
G is the growth index of vegetative cover in percent maturity varying from 0.1 to l
A = infiltration capacity of available storage based upon soil porosity and root density
Ip = available water storage in the surface layer, or A horizon
if = final infiltration rate

The final infiltration rate if, or final constant infiltration rate concept was used by
Musgrave (1955) to define the SCS Hydrologic Soil Groups based upon the Holton

Equation. They are used today and appear in Table 2.3..

Table 2.3 SCS Hydrologic soil group infiltration rate as cm per hour and description of
infiltration rate

 

 

 

 

 

SCS Hydrologic Soil
Group if (cm h") Description of Hydrologic Group
A 0.76 low runoff potential, high rates of infiltration
B 0.3 8076 moderate infiltration rates
C 0.13-0.38 low infiltration rates
D 0.0-0. 13 high runoff potential, low rates of infiltration

 

 

 

 

 

51

2.18 Physically based infiltration models

Green and Ampt (1911) used a physically based approach to derive infiltration of

water into soil. Their approach is particularly useful in course-textured soils that are

initially dry soils. Water infiltration using the GA approach is based upon the following

assumptions (Freyberg et al., 1980):

l.

a distinct wetting front exists and the water content behind it remains constant

2. the soil volume is homogenous in terms of conductivity and water volumetric
capacity

3. matric potential (as influenced by soil mineral attraction to water) remains
constant throughout the wetting front

4. the soil volume is uniformly wet behind the wetting front
The basic parameters of the GA method are presented in Figure 2.3 (USEPA

1998).

 

 

 

 

 

 

 

 

 

V 9
115 .° 9s
\_ Soil . '
Surface
Z Actual
\'.A
De th Z - 7...;-
Wetting Front p '0‘ "
hf Pressure Head r' GmmAmP‘
Model
Green-Ampt Parameters Water Content

Profile

Figure 2.3 Green Ampt parameters and idealized illustration of water content profile

52

The piston type profile assumes saturated volumetric water 0,, content down to
the wetting front at depth Z. The ponded water H, assumes a positive water pressure
exerting some force on the water volume entering the soil volume at some depth Z. The
soil water pressure at the wetting front Hf, is less than the surface pressure, H,. The
dotted line in Figure 5 part b is a suggested actual change in water content from dry soil
00 to saturated water content 0,.

For the duration of time water is entering the soil, t, the infiltration of the wetting
front will be depth Z. The relationship of infiltration into a non-saturated soil and
Darcy’s equation produced the following equation (Freyberg et al., 1980):

q = - Ks[(hr- (H, + Z)]/Z
Where
K, = hydraulic conductivity of the surface water volumetric water content (used in
saturated hydraulic conductivity equation
hf = soil water pressure at the wetting front
H, = soil water pressure at the soil surface
Z = depth of wetting front in vertical distance
The GA equation (U SEPA, 1998) is presented as follows:

1 = Kst - (hr — Hs)(9s - 90)108[1 - (I/(h, — Hs)(0s - 00))

Where
I = accumulative water volume at time t
t = time at depth Z for initial volumetric water content 00 to become saturated 9, by the

wetting front

53

The application of the GA equation is widespread. A lengthy review was
conducted by workers at USEPA (1998) providing 20 referenced bibliographies. Tindall
and Kunkel (1999) suggested that the GA equation could be applied if the user knew the
saturated hydraulic conductivity K,, suction at the wetting front hf, initial volumetric
water content 00, and what is known as the transitional volumetric water content, or 0,.
They suggest these values can be determined or estimated using laboratory measurements
of saturated hydraulic conductivity and volumetric water content. Rawls and Brakensiek
(1983) and Rawls et al. (1993) provide regression equations for fitting GA parameters if
only one parameter is unknown.

Several modifications to the GA model have been suggested. Bouwer (1966)
claimed that Ks should be modified by multiplying it by one half. Neuman (1976)
derived expressions for hf based upon three time periods t, small, intermediate, and large.
Several statistical correlations for K, and hf are listed (Brakensiek and Onstad, 1977;
McCuen et al., 1981; Rawls and Brakensiek, 1982).

Richard (1931) took Darcy’s equation and based infiltration into unsaturated soils
upon matric potential driven by unsaturated hydraulic conductivity (Hillel, 1998). He
proposed the following model:

0 = -K(¢)AH
Where
q = water intake rate per unit time, or flux
K = hydraulic conductivity of soil at a given volumetric water content
if = matric potential

AH = hydraulic head gradient

54

If the continuity equation is introduced the Richards equation (Hillel, 1998) is:
d0/dt = A[K(¢) AH]
Where
d0/dt = change in volumetric water content as a function of total time
Philip (1957) coined the term sorptivity, which is the ability of a soil to absorb
water by capillary process. Water is drawn to the soil mineral interface to the extent that
it can flow against the gravity potential, i.e., upwards from a wet soil vertically into a
drier soil (Clothier and Scotter, 2002). Philip (1957) proposed the following infiltration
equation for vertical flow with water ponded on the soil surface:
I = St°’° + Apt
Where
I = cumulative water volume
S = sorptivity
Ap = saturated hydraulic conductivity if ponded
t = time
The vertical infiltration rate can be calculated by the following differential equation:
i = dI/dt
Where
i = infiltration rate
dI/dt = change in cumulative water volume as a function of total time
Youngs (1964) suggested a method for determining S as follows:
3 = [2(0. — 001010105

Where

55

0, = volumetric water content in the transition or wetting zone
0, = initial volumetric water content
K = hydraulic conductivity of soil at a given volumetric water content
1% = matric potential at the wetting fi'ont

The Darcy-Buckingham Law (Wraith and Or, 2001) is the unsaturated soil analog
version of the Darcy’s saturated soil model. The DB Law is expressed as a flux q, or
distance water moves through an unsaturated soil per unit time. The flux equation is:

q = -K(0)z\II(0)
Where
K = unsaturated hydraulic water conductivity of soil cm s'1 as a firnction of volumetric
. water content
q = water flux cm s'1
0 = volumetric water content as a function of location and time
if = total water head as a function of volumetric water content
2 = vertical coordinate

For this equation 2 is positive in the direction of gravity (vertical flow) with z = O
at the soil surface. The differences between the DB Law and Darcy’s Law (unsaturated
versus saturated flow) are the dependence of hydraulic conductivity and the total head on
the volumetric water content. The greater the ponded head, the more rapid the
unsaturated hydraulic conductivity per unit time. A ponded surface for instance will
yield a positive head, with 2 some value above the soil surface completely related to

ponded depth.

56

Practitioners of the DB Law (Hillel, 1998) combine the continuity equation (Q =
vA) into what is known as the Richard equation:
d0/dt = A(K(0)Ah(0)) -— (dK(0))/dz
Where
d0/dt = change in volumetric water content as a function of time unit
Clothier and Smetten (1990) chose to write the formula using another form of the
dependant variable d0/dt adopted by Philip (1969):
d0/dt = [D(0) A(0)] + [dK(0)/d0](d0/dz)
Where
D = soil water diffusivity defined as K(0)dh/d0
t = time
2 = depth
K = hydraulic conductivity
These two analogs describe only vertical flow into unsaturated soils; the two
dependant variable head h and volumetric water content 0 can either be written in terms
of h or 0. However, in all uses of various forms of Richards Equation, there are
limitations. Philip (1969) described these limitations as follows:
0 The representative volume of soil may not be representative based upon varying
preferential pathways and macropores.
o Colloidal shrink and swelling of soils alters flow after soil colloids have shifted
(even slightly).
0 Soil movement as a function of wetting or drying significantly alters soil

infiltration of water.

57

Thermal inputs cause evaporation of water during infiltration and the
redistribution of water flow in the soil volume (beneath the soil surface).

Soil hysteresis may on the one hand increase water infiltration if drying, and
decrease water infiltration if wetting; therefore unsaturated conductivity may be
quite different in the same soil volume. Some soils will have higher volumetric
water content when drying as compared to that soil when wetted at the same
matric potential. Water in soils is less apt to leave a soil pore than water entering
the soil pore, a phenomena known as the ink bottle effect.

Sinks are neglected, i.e. plant roots may alter water infiltration and unsaturated
hydraulic conductivity, and, plant roots extract water from the soil volume
thereby reducing the volumetric water content per unit time.

Flow q, is one dimensional (vertical, downward) and may suffice for rainfall
events and irrigation applications over the field scale.

Wraith and Or (2001) offered three additional components, which helped,

eliminate these limitations. They are:

A boundary condition at the water supply surface, i.e. the surface area defined.
An initial condition of all depths z, i.e. volumetric water content at any 2 at time =
0 equals 00(2).

The given soil volume hydraulic parameters would be known: K(0) and h(0).

Richards Equation models deal almost exclusively with absorption of water

through a volume of soil, which is not saturated. Little attention is given to redistribution

over time and internal drainage (as affected by gravity but against a capillary gradient.

Wraith and Or (2001) offer only that they are complicated for three reasons: infiltration

58

and conductivity models (like GA and Richards Equations) do not explain redistribution
of water once infiltration has ceased. This is important to the extent that after any rainfall
event or irrigation application, there is a period of infiltration that occurs when the soil
volume is not saturated. After the storm or irrigation has stopped, infiltration does not.
That is, the continued movement of water based upon complexation variables (listed
above by Philip (1969) occurs in all soils. Two, the initial conditions for the
redistribution of water once absorbed are complicated and not easily modeled. Field soils
may be and usually are hugely variable. Add to these variability’s the differences in soil
horizon properties, soil organic matter content and living root tissue, redistribution is
extremely complicated. And three, the redistribution of water may be influenced by
hysteresis near the surface over time, and capillary hysteresis at the lower boundary of a
root zone.

Wraith and Or (2001) present a short description of various Richard Equation
Models, which have been developed from Philip (1969) to Hills and Warrack (1993).
However the bulk of work done in measuring infiltration and unsaturated hydraulic

conductivity is based upon parameter estimation.

2.19 Parameter estimation

Characterization of water flow in soils is influenced by soil mineralogy (PSA),
soil hydraulic properties, Ks and K, soil structure, i.e. the cohesive properties of the soil
matrix with organic and inorganic compounds (roots, SOM, calcium carbonate, hydrogen
ion or pH log), and three hydraulic potentials, or tlz energy state.

There are three forms of hydraulic potential (Hillel, 1998). They are:

59

Gravimetric potential, tl/g which represents the influence of gravity exerted on all
water flow in soils, towards the lowest plane of the soil volume boundary from
some depth 2 above this plane (usually soil surface), and through that plane in
what could be called deep leaching. Movement of water against 1.08 requires work
or energy (pumps are required), and movement of water at some point below the
reference plane is assisted by if, (water flows downhill).

Osmotic potential, (0, is a function of the attraction of solutes in the soil volume.
Salts, cations, and SOM exert an attraction to water molecules, thereby reducing
flow per unit time, and may increase the volumetric water content given this
affinity.

Matric potential, rpm is the largest variable potential in soils. Potential in the case
of soil water represents energy status of soil water. If, for instance, the soil water
is at a pressure greater than atmospheric pressure, it is considered positive. The
inverse is true. In soils that are at some volumetric water content less than
saturation, the soil water is at some negative potential, or what is known as
suction or tension. Matric potential exerts a force upon water, which is described
using the same format. In soils that are at some volumetric water content less
than saturation, the matric potential is increasingly less positive, or at a greater
negative potential. Soil particles have a cohesive attraction to liquid and vapor
water. As the volumetric water content decreases (drier soil) the remaining water
in the soil matrix held more strongly to the mineral interface, thereby matric water
has a more negative potential. Conversely as the volumetric water content

increases, more water in the soil matrix is less apt to adsorb to water molecules

60

already physically closer to the mineral interface. In this case water flow

increases because it has greater potential or less suction. Course textured soils

have lower volumetric water holding capacity as the surface area of the bulk soil
is much lower than fine textured soils. Finally, salts in soils tend to increase the
volumetric water holding capacity given their attraction to water.

A reasonable approach at parameter fitting of unsaturated hydraulic conductivity
was first established by deriving or fitting matric potential in units of energy of water
potential to known volumetric water content of a given soil (Kosugi et al., 2002). The
user assumes rpm is the driving independent variable in water flow, thereby establishing
no influence 100 or 0,. One assumes the soil volume is homogeneous in terms of particle
size distribution as a function of uniformity in three-dimensional space. An idealized
water retention curve is usually based upon a log scale of matric potential, Wm and linear
scale of volumetric water content. The volumetric water content is usually listed as a
percent of total volume. A soil with a bulk density of 1.3 has a void space of 49%;
therefore at potential saturation the soil has volumetric water content of 49%. An

idealized water release curve appears in Figure 2.4.

61

 

10000

1000 <

 

 

 

Matn'c Head (-m)

 

 

 

 

 

0.1 . i . ,
0.0 0.1 0.2 0.3 0.4 0.5

Volume Water Content (m3lm3)

Figure 2.4 Idealized water release curve for sand, silt, and clay soils as matric head versus
volumetric water content

Note the log scale of matric head on the y-axis. These values are matric head of
pressure potential, which are negative values. At saturation the idealized soil volume has
a matric potential of zero. As the soil dries the volumetric water content decreases.
Course textured soils dry more readily with less negative water potential than do clayey
soil. Related to this phenomenon is soil mineral surface area. Sand particles at 2m to
0.05mm have a surface area of approximately one m2 per gram of soil, whereas expansive
clay like Smectite (2:1 layered) at 0.002 mm (2 pm) has a surface area of 750 m2 per
gram of soil (Flint and Flint, 2002).

Matric head or energy potential can be expressed in units of head. A water head
of 1 m represents a pressure of 9.807 kPa, 0.0969 atrn, and 0.0981 bar. The relationship
between positive and negative potentials is equivalent. The values listed on the y-axis are

negative potential, or increasing suction.

62

The water release curve, or Soil Water Characteristic Curve relates the energy

state of water at a given volumetric water content. From a practical standpoint irrigators

use these curves to schedule irrigation water to replace the used water through ET. For

instance a sandy soil has a much narrower range of volumetric water content than clayey

soils; the difference between FC and a 50% depletion value of 80 kPa is smaller in sands

than in clays (Hillel, 1998).

Several assumptions are necessary before estimating parameters of a SWCC.

They are:

The volumetric water content 0, is equal to the saturated water content 0, when
soil matric potential rpm is zero. From a practice standpoint entrapped air causes
an actual value of 0.85 to 0.9 0,.

An air entry region on the curve at slightly negative potentials will not change the
matric potential. This slight negative potential of water in a soil volume moves a
slight amount of water out of the volume; this water volume is influenced only by
slight negative potential and not matric potential. Physically, this water is a
volume furthest away from the soil mineral interface and therefore the first to be
moved out of the volume under a slight negative pressure.

As the water potential becomes more negative the matric potential increases. The
mineral interface begins to control movement of water, its cohesive attraction is
slowing down flow and reducing the water potential. The air entry value is that
point on the SWCC when water potential is controlled by matric potential.

As the matric potential increases the volumetric water content decreases to a point

known as residual water content, 0,. From a practical standpoint 0, may be

63

defined as some point on the matric potential curve which any further negative
potential does not measurably reduce the volumetric water content.
0 SWCC may be derived using a range 0, < 0 < 0,.

There are two predominant parameter models in use today (Kosugi et al., 2002). A
review paper by Wang et al., 2002 provided some guidance for users in fitting BC
parameters to derive a SWCC. The BC equation (Brooks and Corey, 1964) is:

S, = (0 — 0,)/(0s — 0,) = (lid/h)“ h > hd

S, = 1 h < hd
Where
S, = effective saturation
0 = volumetric water content
0, = saturated volumetric water content
0, — residual volumetric water content
hd = matric potential at air entry value
h = matric potential at some volumetric water content
n = power function which characterizes the width of the pore size distribution in the soil
volume

Kosugi et al. (2002) examined the BC model and found it represented a good
SWCC estimate for course textured soils with a clearly defined air entry value but not for
finer textured soils and especially undisturbed field soils. Campbell (1974) altered the
BC model by defining the dependent variable as degree of saturation instead of effective

saturation.

64

van Genuchten (1980) provided a parameter estimation model and is today one of
the most commonly used models in fitting data to derive a SWCC. His equation is:

Se = [1 + (-ahm)n]""
VVhen:
S, = effective saturation
a = a parameter to scale the matric potential
hm = matric potential at some volumetric water content
it = value greater than one which establishes the curve slope

Carsel and Parrish (1998) used the VG model and fit two groups of soils. The
first is Unsaturated Soil Database, or UNSODA (derived from the USDA ARS Soil
Salinity Laboratory in San Bemardino California), and the Soil Survey Mechanics
Laboratory in Lincoln Nebraska. The data in Table 2.4 are published in Or and Wraith

( 1999a).

65

Table 2.4 UNSODA and Soil Survey hydraulic parameters including water retention and
saturated hydraulic conductivity values

 

 

 

 

 

 

--- Water Retentlon-m- --- Saturated Hydraullc --
Nt 0r 0s 5 n Conductlvlty
Textural
class 1/cm Nt Ks cmld
UNSODA
Sand 126 0.058 0.37 0.035 3.19 74 505.8
Loamy sand 51 0.074 0.39 0.035 2.39 31 226.5
Sandy loam 78 0.067 0.37 0.021 1.61 50 41.6
Loam 61 0.083 0.46 0.025 1.31 31 38.3
Silt 3 0.123 0.48 0.006 1.53 2 55.7
$111 loam 101 0.061 0.43 0.012 1.39 62 30.5
Sandy clay loam 37 0.086 0.40 0.033 1.49 19 9.69
Clay loam 23 0.129 0.47 0.030 1.37 8 1.84
Silty clay loam 20 0.098 0.55 0.027 1.41 10 7.41
Silty clay 12 0.163 0.47 0.023 1.39 6 8.40
Clay 25 0.102 0.51 0.021 1.20 23 26.0
Sol u e
Sand 246 0.045 0.43 0.145 2.68 246 712.8
Loamy sand 315 0.057 0.41 0.124 2.28 315 350.2
Sandy loam 1183 0.065 0.41 0.075 1.89 1183 106.1
Loam 735 0.078 0.43 0.036 1.56 735 25.0
Silt 82 0.034 0.46 0.016 1.37 88 6.00
Silt loam 1093 0.067 0.45 0.020 1.41 1093 10.8
Sandy clay loam 214 0.100 0.39 0.059 1.48 214 31.4
Clay loam 364 0.095 0.41 0.019 1.31 345 6.24
Silty clay loam 641 0.089 0.43 0.010 1.23 592 1.68
Sandy clay 46 0.100 0.38 0.027 1.23 46 2.88
Silty clay 374 0.070 0.36 0.005 1.09 126 0.48
Clay 400 0.068 0.38 0.008 1.09 114 4.80
1 Approximate sample size for Soil Survey database

 

 

 

Both groups of soils are listed using the 12 textural classes as defined by USDA.
The columns represent values from laboratory analysis, the number of samples listed in
the second column. The 0, values are matric potentials at 1.5 MPa, or what is commonly
known as Permanent Wilting Point. The 0, values are given at a matric potential of zero.
The next two parameters, a and n, are used to shape the SWCC. Listed on the far right

are K,, or the saturated hydraulic conductivity for each textural class.

66

These data are used to fit pairs of volumetric water content at a given matric
potential, or 0,,wm. The USDA Soil Mechanics Laboratory uses the VG model to fit only
two pairs of data: the volumetric water content at 0.033 MPa (FC) and the volumetric
water content at 1.5 MPa (PWP). The UNSODA data are fitted with four or five data

pairs. The data in Table 2.4 are fitted to the VG model and appear in Figure 2.5 (Leij et

 

al.,1999).
1 z . ,

000; all ch-sm

g ‘-\ , ~ loam soil
A 100$
$
v
(U 3
g 10%
.2

E 10 .\ 1

.m .. -.... .-.. ..

0.1‘

O 0.1 0.2 0.3 0.4 0.5
Soil Water Content (m’lm3)

Figure 2.5 UNSODA and Soil Survey fitted water release curves (VG and BC
parameters) for a silt loam soil

Note the volumetric water content is increasing from left to right in both graphs.
The matric potential is listed in log h (head) form. In both figures the course textured
sand and loamy sand SWCC curves dry with slight increases of soil matric potential. At
the other end of the textural classification, the fine textured soils, clay and silty clay have

higher volumetric water contents at higher matric potentials.

67

2.20 Layered soil profiles

The root zone of agricultural soils includes more than one layer of soils. The
characterization of soil horizons is dependant upon textural class, bulk density, soil
organic content, root mass, base cations, depth and temporal aspect of a water table, and
farming practices (Tindall and Kunkel, 1999). Three properties of soil water in the root
zone vary with different soil horizons. They are volumetric water content, matric
potential, and unsaturated hydraulic conductivity. Volumetric water content is larger in
finer textured soils. Matric potential increases (more suction or greater negative
pressure) in finer textured soils. Unsaturated hydraulic conductivity is lower in fine
textured soils. As clay particles increase in a horizon, water flow is slowed, water is held
more tightly to soil particles (increased matric potential), tortuosity increases, and a
volume of soil in this horizon contains more water than a horizon with courser textured
soils (Hillel, 1998).

Such discontinuity of water flow and volumetric water content is problematic
unless the relevant characteristics for each horizon are known. Differential horizonal or
layer scenarios are presented below:

1. In clay over a sand layer, the flux rate is accommodated by greater matric
potential in the clay layer. As more water is added to the clay surface,
only when matric potential is near zero and gravimetric potential increases
will water move into the sand layer. Water will not leave a small pore in
clay due to significantly increased surface area and therefore matric

potential, and enter the larger pore space in sand. Even though the sand

68

layer has higher unsaturated hydraulic conductivity, the clay layer controls
flow until EC is reached.

In a typical agronomic field, a courser texture soil in the Ap horizon is
above the E and Bt horizon. Clay has moved fi'om the Ap through the E
into the Bt. If the Bt is the bottom of the root zone, then water flux is
slower, matric potential is higher, and volumetric water content is higher
per unit volume of soil in this Bt horizon. In conventional tillage fields
the Ap may have a lower bulk density than the E and Bt horizons, as well
as greater soil organic matter and existing root mass. These increase water
flux due to aggregation of soil particles and lower bulk density, but may
increase volumetric water content due to organic molecules attraction to
water. When organic matter is added to a course textured A or Ap
horizon, volumetric water content is increased to the extent that farmers
use this practice to reduce irrigation volume depths over the growing
season. Organic matter additions provide energy for soil microbes as well.
For any given soil layer, as volumetric water increases matric potential
decreases and approaches zero. As volumetric water content decreases
matric potential increases, i.e. less water is more tightly bound to the soil
mineral interface and thus greater negative pressure.

Large variances exist in well-aggregated soils. Water added to these
layers, usually A or Ap horizons, will flow through soil in two processes.
Water moving into large voids and cracks that exist outside of an

aggregate flows rapidly. The attraction of water to the soil mineral

69

0.

interface is negligible given the physical distance or space. Water flows
via the gravimetric potential. Within the soil aggregate itself, however,
water flux is slower, matric potential is greater and volumetric water
content is increased. Soil aggregates may be saturated but overall
unsaturated conductivity is slow.

Unsaturated conductivity is influenced by the presence of swelling clays.
As the volumetric water content decreases the soil shrinks reducing pore
size but creating cracks in the soil matrix. When this clay soil is rewet, the
unsaturated conductivity is very rapid at the onset but as the clay swells it
is reduced to almost zero. Factors that reduce this conductivity are
reduced porosity, increased tortuosity, and reduction of pore conductivity
as water leaves the large pores, and an increased water viscosity at the

mineral interface.

The Green Ampt equation has been modified by Childs and Bybordi (1969) to
describe infiltration in layered soils if the saturated hydraulic conductivity increases with
greater soil depths. Hacham and Alfaro (1980) provided similar modifications using the
GA equation in multilayered soils. Rawls et al. (1983) measured saturated hydraulic
conductivity in multilayered soils and derived a FC value for individual layers based
upon soil texture and bulk density. Starks et al., 2003 compared limited data with
extensive empirical data to model the estimated root zone water content.

The objective of Stark’s root zone water content work was simple: what are the
minimum soil characteristics which must be known to derive root zone water content, or

in the broader sense, the FWB? This work is unique in that similar work is not found

70

anywhere in the literature. To the extent that a combination study involving minimal to
maximal parameterization of both volMetric water content paired with a matric potential
measurement, does provide an estimation of hydraulic conductivity in both saturated and
unsaturated soils. Their premise’s to describe or model infiltration, soil water
conductivity, volumetric water content, and plant water uptake are:
l. The soil water characteristic curve, or SWCC, described as the paired relationship
between the volumetric water content, 0 and the matric potential ‘1’.
2. Soil hydraulic conductivity K as a function of 0 and ‘1’.
Several laboratory and field methods are available for direct measurements on soil cores
(Klute, 1986) but are tedious, time consuming, and costly (Starks et al, 2003). Other
researchers have developed similar techniques to obtain these relationships from soil
properties that are more easily measured, such as bulk density, soil texture, and F C
(Rawls et. al., 1983; Woston and van Genuchten, 1988; Ahuja et. al., 1985; Williams and
Ahuja, 1993; and Vereechen, 1995). Ahuja (1999) reviews many of these methods.
Ahuja and Ma (2002) suggested these methods fit into a hierarchy based upon
simple to more complex. For volumetric water content, they are:
l. estimation from textural class only.
2. estimation fi'om bulk density, textural class, and organic matter content.
3. estimation from (2) and one measured value of volumetric water content at some
matric potential.
4. estimation from measuring the entire hydraulic function over the SWCC.

For estimation of conductivity, they are:

71

0. estimation of saturated hydraulic conductivity, K, from textural class or effective
porosity.
2. estimation of K, from one of the four methods used to derive volumetric water
content.
The Agricultural Research Service, ARS developed a process model called Root
Zone Water Quality Model, RZWQM. This model used the volumetric hierarchy
suggested by Ahuja and Ma (2002), and was therefore chosen for use. The user may
chose any of the four hierarchal approaches (above) to determine the F WB based upon
volumetric water content and the changes to this volume as a firnction of mass water
balance (climate, irrigation, deep leaching, plant uptake, and evaporation) and time.
Starks, et al., (2003) asked the question “how well the limited data set simulates
the hydrologic system and, in particular, gives satisfactory estimates from soil water
profile (or FWB)? If a simpler data set could be used to estimate the FWB, thus time,
energy, and resources could be saved, or they could be distributed over larger areas of
interest (more fields). They further state:
“Considering the time required for certain laboratory analyses, it would be of
practical significance to determine the effect on soil water profile (FWB)
estimates using soil hydraulic input data derived from standard laboratory

analyses versus those obtained by relatively simple in situ techniques.”

72

2.21 Root Zone Water Quality Model

The RZWQM integrates physical processes in time steps that predict the impact
of agricultural practices, climate, and existing soil characteristics on the movement of
water through the root zone (Ahuja, et al., 2000). The hydrologic component of this
model controls the simulation of infiltration, redistribution, and plant uptake of water so
that the change in soil water in the root zone can be estimated. The time period allows
integration of climate events and agricultural practices with plant growth and volumetric
water changes in the root zone in linear time. The shortest time period is five minutes,
and the usual time period for a growing season or multiple growing seasons in a twenty-
four hour period. Rainfall input mirrors hydrograph data; storm intensity and duration
are given as breakpoint data.

The model requires an extensive amount of data. At a minimum RZWQM
requires the driving variables of mass water balance (daily minimum and maximum
temperatures, relative humidity, solar radiation, wind speed, and rainfall hydrograph) and
the site-specific soil profile descriptions (soil horizons, hydraulic properties, reside cover,
and crop physiologic specifications).

The hydrologic component of RZWQM includes a menu item: “soil hydraulics
data input options” where the user may choose “limited data” or “firll description.” The
limited data input include only two parameters, saturated hydraulic conductivity and
volumetric water content at PC. The user may obtain these data from three sources listed
within the model documentation (Hanson et al., 1962; Rawls, et al., 1983; and Ahuja, et
al., 1988). The full description input allows the user to use any source including

measured values and estimated values or any combination of the two.

73

RZWQM uses a modified GA equation for determining infiltration (Ajuha, et. al.,
1993). The Modified GA equation is:

V = Ks[(‘rc + H0 + wa)/wa)]
Where
V = infiltration rate at any given time, cm h'1
K, = effective average saturated hydraulic conductivity of the wetting zone, cm hr'1
1', = capillary drive or suction head (negative pressure) at the wetting front, cm
Ho = depth of the surface ponding (vertical), cm
wa = depth of the wetting front, cm

RZWQM uses a mass conservation numerical solution of Richard’s Equation to
solve the redistribution of water in the soil volume (Abuja, 2000). The Richards
Equation estimates the redistribution of water between rainfall or irrigation events. The
equation is:

d0/dz = d/dz[K(h,z)(dh/dz) — K(h,z)] — S(z,t)

Where

3 3

0 = volumetric soil water content, cm cm'
t = time, h

z = soil depth, cm

h = soil water pressure potential, or head, cm

K = unsaturated hydraulic conductivity, cm h'1

S(z,t) = sink term for water uptake by plant roots, cm h'1

The GA and Richard Equations require saturated and unsaturated hydraulic

conductivities, K5 and K, which can be estimated using the Brooks and Corey (1964)

74

forms (Starks, et al., 2003). The 0h and Kh relationships are modified in RZWQM as
follows (Starks, et al., 2003):

0h = 0, — A1*|h|

0h = o, + 13th

Kb = K,|h|'"1

Kh = K2|h|'"2

Where

A1, B, h, N1, N2, and K2 are constants

3

0, = saturated soil water content, cm cm'3

3 cm-3

0, = residual soil water content, cm
K, = saturated hydraulic conductivity, cm h'l
Starks et al. (2003) used five methods to determine the water in the soil profile, or
F WB. The study area was located at the Little Washita River Experimental Watershed in
Oklahoma. They methdds were:
1. The soil profile depth was 60 cm. Each soil horizon was determined based
upon soil textural class. Data were obtained fi'om the USDA Soil Survey.
RZWQM uses the textural class mean physical and hydraulic default values.
This is the minimum level of input required by the model. These values are
easily obtained with the appropriate Soil Survey; no laboratory or field
investigation is required.
2. The second method includes the actual determination of sand, silt, and clay in

the laboratory using the hydrometer method (Gee and Or, 2002), thereby more

accurately defining textural class and therefore hydraulic parameters. Moist

75

bulk density is measured for each horizon (Grossman and Reinsch, 2002).
Undisturbed (as possible) soil cores are obtained for each soil horizon when
the soil is at or near FC (hence the term moist bulk density). Using PSA and
BD values for each horizon, RZWQM derives soil texture and adjusts the
saturated volumetric water content 0,, based upon porosity. The mineral bulk
density is set at a default value of 2.65 g cm”.

. The third method includes the two measurements in method 2, with an
additional measurement of volumetric water content at F C, or at a matric
potential of — 0.033 MPa. Undisturbed field cores for each horizon are placed
in a pressure tank at an atmospheric pressure of 0.033 MPa according to the
Klute (1986) Method.

. The fourth method is the same as method 3 except the volumetric water
content at PC was measured in situ based upon two-day drainage data. The
instantaneous profile method (Hillel, 1980) was used to determine soil
hydraulic properties. Double ring infiltrometers were used; a 24-hour
prewetting period was followed by periodic measurement of constant loss of
water in the inner ring as the result of water moving into the saturated wetting
zone at a constant rate. Once the vertical flux rate was constant conductivity
was assumed to be in a steady state. Tensiometers were used to define that
point when the soil layers were at PC. Gravimetric soil samples for each layer
were obtained and given the known bulk density measurements, volumetric
water content could be calculated. This method (4) differed from method 3

only in the way PC was determined. Method 3 used a laboratory analysis of

76

field cores, and method 4 the measurement was calculated after cores were
obtained from a field adjusted wetting process using an instantaneous method
of water infiltration from a ponded surface depth.

4. The fifth method included actual field measurements of tensiometers placed at
different depths in the soil profile. The FC value was assumed to be the
volumetric water content two days after saturated conditions with no
additional water input (rainfall or irrigation). The K, value was assumed to be
constant for all horizons. This method (5) was included to mimic data that
might be obtained from remote sensing (Mattikalli et al., 1998).

RZWQM requires vegetation data for determining plant water uptake and
therefore removal of water fi'om the soil root zone as transpired water. A number of
agricultural crops may be selected, but crop growths curves and crop consumptive water
use parameters can be used to build local crop databases. A turf or perennial grass crop
component is based upon specifying rooting depth, LAI, and harvest cycles (Ahuj a,
2000).

Landowner management practices are defined in RZWQM as inputs of tillage,
planting, harvesting, and crop rotation if more than one growing season is simulated.
Tillage depth, planting date with number of seeds per spatial area, harvest biomass, and
length of growing season are examples the user obtains from the landowner. Irrigation
water volume depths are obtained from the landowner and serve as water inputs with both
a temporal and spatial stamp. Liquid manure as wastewater can be used as water inputs;

nutrient components can be used for modeling nitrogen, phosphorus, potassium, and

77

carbon. However, for the hydraulic component of RZWQM, irrigation of water and
wastewater are given as volume depths (Ahuja, 2000).

The Starks et al., (2003) study included the use of TDR measurements for the 30-
day duration of measurements at the south central Oklahoma site. The general vegetation
in the 610 km2 watershed was a mix of 60% rangeland, 20% cropland, and 20% forests,
riparian areas, water bodies, urban areas, and oil waste land. The watershed is
instrumented with 45 weather stations.

Starks et a1. (2003) drew the following conclusions fi'om this work:

1. RZWQM provided satisfactory results given that no hydraulic properties were
calibrated or optimized over a linear time frame (except method 5), although
some site specific measurements were used in methods three and four.

2. The study site was quite different from other studies that used RZWQM to
determine the soil water profile on largely agronomic fields (Hanson, et al., 1999;
Ma et al., 1998; Wu et al., 1996). The time study of 30 days was considerably
shorter than what RZWQM was designed for: at least one growing season up to
many growing seasons.

3. Martin and Watts (1999) stated the importance of correctly simulating plant water
uptake via a crop growth curve, which correctly provide water consumptive
values over the growing season. Not only is plant specie selection important, but
also the root distribution in terms of three dimensional space is important in terms
of water available in the root zone for potential plant water uptake.

4. One interesting yet surprising finding is that the use of hydraulic properties

estimated from textural class show good agreement between predicted and

78

measured volumetric water content in the soil profile. And, in most cases the
results were better than those where laboratory measurements were made
(methods 2, 3, and 4). A similar study using RZWQM (Landa et al., 1999)
showed the same results by using textural class data (USDA Soil Survey) to
derive the soil water profile from predicted versus measured volumetric water
content. Starks et a1. (2003) claim that their study and the Landa study support
the default values used in RZWQM as acceptable input for model applications
which a have limited input data set.

. Method 4, the in situ field measurements showed good agreement between the
predicted and measured soil water profile over the 30 day study period. Starks et
al. (2003) further state that these in situ measurements are preferable to those
obtained by the detailed laboratory measurements in method 3. They base this
conclusion on two factors. One, that there are large spatial variations in soil
properties, and two, for a given textural class, the corresponding range of values
can be large, thus using the average of each parameter range reduces the chance
that a group of soil cores may not accurately reflect the spatial variability across
the field surface and root zone. Fruther, Starks et al. (2003) state that using
USDA Soil Survey mean values, or in situ field measurements require much less
time, less expense, and importantly, may be considered more representative of
actual field conditions.

. Method 5 using TDR instruments that monitored volumetric water in continuous
times steps (15 minutes) showed good agreement between predicted and the

continuous measurements accumulated in a local data logger.

79

7. The most important values in this study were textural class, bulk density,
saturated hydraulic conductivity, and volumetric water content at PC (- 0.033
MPa). The authors (Starks et al., 2003) conclude that limited data, i.e. the USDA
Soil Survey, or the approach of measuring field in situ conductivity (using the
double ring infiltrometer for determining steady state flux water infiltration)
provided the best estimates of soil water profile (FWB) content. RZWQM is
therefore parameterized with default values based upon the USDA Soil
Classification Method that explain the physical and hydraulic properties based

upon 12 textural classes of soils found in agronomic soils.

2.22 Literature summary

The FWB is determined by combining several components. The literature
provides a number of methodologies for measuring these components. Integration of
these measurements into an estimation of FWB requires a process strategy, or model like
RZWQM. The model framework is built upon all components that establish the F WB,
and more importantly, its change with respect to time. The flux of F WB, or the change in
volumetric water content in the soil root zone over a specified time step (day, week, or
growing season) can be known by combining the Mass Conservation Law and the
Continuity Equation.

The FWB consists of two primary sets of components: those above the soil
surface, and those below that surface. The literature includes apt description of both.
The preferred equation for determining the FWB is set to zero; the additions and

subtractions of water in the soil root boundary are characterized over a time step. The

80

Evett and Lascano (1993) equation [0 = AS + P + R - F — ET] includes the additions of
water above the soil surface, rainfall and irrigation as volume depths, and net water
ponding that may occur if run-on exceeds runoff if the volume depth is applied at a rate
faster than the infiltration rate of the receiving soil (P + R). The movement of water as
matric flow upwards into drier soils in the soil root zone, from wetter soils below the root
zone is an input below the soil surface. The converse is true. Water flow out of the root
zone into soils beneath the rooting depth, or deep leaching occurs if the water potential
gradient is downward as. a result of gravitational potentials greater than matric potential.
This below the surface input or output is described as the F term, denoting a flux. The
final term is known by estimating water movement from the soil into the atmosphere as a
vapor flux from the soil surface or from plant leaves. This is the ET term in the equation.
The change in soil water, AS, is a function of the soil parameters discussed earlier.
These are variable and can be difficult to obtain. The determination of water in the root
zone, or soil volume at some boundary, is influenced by the physical matrix of that
volume, and how it behaves when the other additions and subtractions of water are
known over that time step. For example a rainfall event of 5 cm as volume depth on the
soil surface may entirely enter the soil matrix, may do so only in part, or may almost
completely runoff the soil surface. To quantify the influence of this 5 cm rainfall event
requires knowledge of the duration of that rain event, i.e. a thundershower lasting 20
minutes or a less intense storm lasting 20 hours. Course textured surface layers are likely
to absorb water at a faster intake rate than fine textured soils. The relative bulk density of

the surface soil may either increase or decrease this infiltration rate. Rainfall or storm

81

duration and intensity are obtained fi'om a local weather station; these data are published
in almost all newspapers.

The run-on and runoff factor is more problematic to obtain. For agricultural
fields, the assumption may be made that run-on and runoff are equivalent which thereby
do not influence the change in water storage. However the redistribution of water in the
soil profile can be influenced by the lateral or horizontal flow of water. This horizontal
flux of water certainly occurs. If the scale of measurement is a field, the changes in
horizontal flux may be balanced by the redistribution of the 5 cm rainfall event; the
specific site, however, may receive more water (decreased horizontal flux) from another
adjacent site, which has less water and therefore a higher horizontal flux.

The flux of water into and flour the lower root boundary must be included in any
FWB study. Once the rooting depth has been established, then this lower plane exists
only as a depth and not an actual physical barrier. At a minimum the F term requires
knowledge of the water table depth throughout the study period. If the water table is
considerably deeper than the root zone boundary layer, then this saturation zone in soils
will not influence the root zone water content.

When the upper layer of soils dries, water is moved from wetter soils at a lower
depth into this dry layer. The term capillarity denotes the “wicking effect” of dry soils on
water. This may also be termed redistribution of water in the soil root zone. If capillary
water enters the lower boundary of the root zone, i.e. the soil layer beneath the root layer
boundary is wetter, then this is a below surface input which may increase the water

content of the root zone.

82

The change in soil water can be significant if crops are actively transpiring water
to remain cool as water carries thermal energy into the atmosphere as a water vapor flux.
Soils can lose water by evaporation during and outside of the growing season, thereby
changing the soil water content.

Finally, the change in soil water is influenced by the physical parameters of the
soil, including particle size, soil bulk density, and the SWCC. If the 5 cm rainfall event is
moved completely into the soil matrix, the existing or antecedent soil water influences
unsaturated hydraulic flow (conductivity) i.e. the redistribution of water in the soil
matrix, and in turn, the amount of water that may be available to plants for transpiration,
deep leaching, or surface evaporation. The importance of the SWCC cannot be
overstated. The relationship of particular volumetric water content and its paired matric
potential provides some knowledge of the influence of soil particles upon that volumetric
water content. That is, if 5 cm of volume depth enters a soil that is near FC, the entering
volume may cause deep leaching after a period of redistribution occurs. However, if
another 5 cm volume depth enters a soil that is near PWP, then this volume may be
almost entirely available for plant absorption if the growing crop is capable of extracting
water from the root depth.

As a practical mater, the SWCC provides a range of volumetric water contents
from which plants may absorb water, or available water capacity, AWC. Irrigation water
management includes knowledge of this range. Irrigation schedulers use some percent of
volumetric water depletion to schedule irrigation application at some volume depth. For
instance if the average daily ET is 7 mm, and the AWC is 100 m, then 50 % depletion

(known as Maximum Allowable Depletion, or MAD) is 50 mm. An irrigation volume

83

depth of 50 mm (2.2 inches) is applied every seven days. The irrigation scheduler knows
that at 50 % MAD, the existing soil water is still available for plant absorption, but the
hydraulic gradient is steeper to plant leaves from a drier soil. The scheduler knows that
this steeper gradient may slow plant growth if the plant leaf warms and moves into a
wilting stage.

The RZWQM is the state of the art model for determining the FWB. All of the
equation parameters listed above are accounted for this model. Rainfall data are required
in hydrograph form (duration and intensity). The ET components, wind speed, maximum
and minimum ambient temperature, saturation pressure (relative humidity), and thermal
input (solar radiation) are required as daily time steps. The duration of the growing
season with crop growth stage curves are required. Tillage and other landowner activity
are listed as activities which impact or alter the soil matrix (surface layer turnover,
addition of organic matter, or application of fertilizer and herbicides), or addition water in
the form of irrigation as a volume depth.

The model provides three scenarios for entering soil physical characteristics as
mentioned earlier. These scenarios each provide the foundation by which RZWQM
determines the two phases of water flow in soils: infiltration as originally described by
Green and Ampt (1911) and redistribution as originally described by Richards (1931).
RZWQM is the best model for deriving the FWB if enough parameters are known.

The use of RZWQM has included the evaluation of agronomic crops during one
growing season in Midwest and Plain States. The study sites were called Management
System Evaluation Areas, or MSEA. Wu et al. (1999) evaluated soil hydraulic

parameters in Minnesota and found that over two com-soybean rotations RZWQM

84

performed well in predicting the FWB beginning in May of each year. J aynes and Miller
(1999) studied a MSEA site also using a‘two-season corn — soybean rotation in Iowa.
Their particular interest was modeling herbicide transport, nitrate leaching, and crop
yield. They found RZWQM performed reasonably well in predicting crop yield.
Herbicide transport in drainage water and nitrate leaching were not accurately simulated;
the authors state that field dynamics may not be captured in the model. They include
macropore flow characterization, soil microbe activity in nitrogen cycling, and herbicide
adsorption kinetics on the soil mineral interface.

Ghidey et al. (1999) studied the performance of RZWQM at a Missouri MSEA
site. Again, the study was done on two years of com-soybean rotation. The Model
accurately predicted the F WB over the two years, but the soil water measurements were
taken only during the growing season. The authors concluded that the model should have
a parameter for cracking soils, thereby more accurately characterizing water flow and
solute flow past the lower root boundary. In particular the authors had investigated the
impact of runoff events that result from high intensity but short duration storm events.
They concluded the model over predicted runoff events after long periods of dry weather;
the presence of large cracks in the soil reduced runoff. The model has a macropore
option that the author used to more accurately predict chemical losses due to seepage.

Workers in Nebraska (Martin and Watts, 1999) studied RZWQM on a MSEA site
using two years of continuous com. This site was irrigated instead of rain-fed as in the
other MSEA sites. The authors used two irrigation rates and three fertilizer rates. The

model accurately predicted soil water in the 150 cm profile during the growing season but

85

simulated lower soil water content both before the growing season in spring and after the
growing season in fall.

A study to integrate over winter analysis of soil water in a profile was conducted
by Flerchinger et al. (2000) for a study site in Pullman, Washington and Akron,
Colorado. Their interest was evaluating northern latitudes where soils freeze during
some of the winter period. The authors combined a Simultaneous Heat and Water Model
(SHAW) with RZWQM to modify the heat balance in the surface layer. RZWQM had
assumed the soil surface temperature is the same as the ambient air temperature.
Coupling SHAW output to RZWQM correctly accounted for the colder climates
especially when the soil surface was frozen. However the authors state that in those soil
layers that are frozen, the prediction of soil water was less accurate. The soil water
content in soil layers beneath these frozen layers were accurately predicted.

Specific applications of RZWQM include soil water transport and drainage
studies in Oklahoma (Ahuja, et al., 1993), North Carolina (Johnson et al., 1995), Illinois
(Singh and Kanwar, 1995; Johnson et al., 1995), Minnesota (Cook, 1996), and Iowa .
(Walker, 1996; Kumar et al., 1998). Evapotranspiration studies were conducted by
Farahani and Baush (1995) and Ma et al., 1998, on agronomic crops in Colorado and
Nebraska. Organic matter and nitrogen cycling were simulated using RZWQM by
Hanson et al., 1995 and Ma et al., 1998. Nokes et al., 1996, examined plant growth in an
attempt to fine tune carbon deposition in plants and thus improve the crop grth curves
for corn, soybeans, and wheat. Azevedo et al., 1997, simulated pesticide half-life and

Ahuja et al., 1996 examined pesticide processes in soils.

86

Several papers were published which explained RZWQM itself. Hansen et al.
(1999) provided calibration background processes that integrated the six components of
the model into a research driven process-based tool for research work. Watts et a1.
(1999) discussed the effort to incorporate the MSEA project into RZWQM by studying
corn and soybean production in several states. Ma et al., 1998, provided a comprehensive

review of RZWQM applications.

2.23 Research needs

Knowledge of the F WB by landowners, including livestock farmers, is not well
known. Yet the integration of this knowledge is required if irrigation water management
is to be used, and some understanding about the FWB is required when livestock
operators apply liquid manure to receiving soils in a field.

Livestock farmers may be required to install waste storage facilities on their
farms. Technicians do not have a site specific tool for predicting the storage period
length based upon meeting application criteria in performance standards (590 and 633
NRCS Standards). Any site specific tool will require knowledge of the F WB over a

period of time, usually at least one year, both in and out of the crop growing period.

2.24 General scope

According to the U.S. Department of Agriculture’s National Agricultural
Statistics Service 2002 Census (USDA, 2004d)) there were slightly more than one
million livestock owners in the United States. The farm gate value of livestock sold in

1997 was 98.8 billion dollars, accounting for 50.2% of all agricultural products sold. On

87

these livestock farms, there were approximately 98.9 million cattle and calves, 18 million
dairy animals, 61 million hogs and pigs, 7.8 million sheep and lambs, 2.4 million horses
and ponies, 367 layers and pullets, and 125 million turkeys. The farm gate value of
livestock production sold in 1997 was 101 billion dollars. The average size of farm with
more than $10,000 of sales was 342 hectares, and the average age of farm owners was
54.3 years old. Finally, there were 375 million hectares of land in the US, with 41
percent of these ha in crop production. Of all land use in the US, 19 percent is cropland
and another 22 percent is pastureland. The scope of farming activity across the U.S.
landscape is significant, both in terms of spatial area and economic strength.

The distribution of production agriculture, however, has changed significantly
during the mechanization and labor efficiency gains of this country. Beginning with the
invention of the tractor, the plow, the seed drill, and the automatic milking machine in the
first half of the 20th Century, farmers have moved towards automation and increased
productive units. In the second half of the Century, herds were consolidated into larger
herd sizes, small farms were combined into larger farms, and many farmers sold out or
left the business resulting in fewer numbers of farmers across the rural landscape. These
consequences are the result of economies of scale, machinery size, a well developed
transportation infrastructure, a farm banking industry, advances in technical assistance
through Land Grant Universities, and a population discovering other livelihoods besides
farming (Gardner, 2002).

The U.S. Environmental Protection Agency reports (U SEPA, 2003) a figure of
1.3 million livestock farmers in 2002. These livestock farms have dairy, swine, poultry,

beef, and to a lesser extent other domestic species grown for meat, milk, and fiber. Some

88

livestock owners grow agronomic crops for sale or harvest them as livestock feed for
their own herd. For this reason, “farmer” will be used in this thesis to describe a
livestock owner, a livestock rancher, or a landowner.

Larger herds are confined in barns for most or all of their productive lives. For
dairy farmers, the drive through fi'ee stall barn is now a common housing and feeding
facility. Many cows live their entire productive lives in these barns. The free stall is
bedded with sand or organic bedding. The floor is concrete, and farmers clean these
floors, or alleys, by scraping or flushing the manure volrune into a reception tank or pit.
Depending upon the facility, manure volumes are handled in one of three ways. One,
manure is transported to a crop field for land application. The delivery system may be
over the road hauling with a honey or slurry wagon, or pumped through buried or
overland pipelines to an irrigation system. Two, manure may enter a treatment facility,
such as an anaerobic digester, a solids separator, or settling basin. Three, the manure
volume is discharged by gravity or pump to a manure storage pond. The storage pond
usually contains milk house effluent and rainwater volumes; the rainwater can contribute
a large volume of water into the storage pond if the drainage area is not diverted (roof,
driveway, or any other non-manured surface).

A challenge for farmers and Technical Service Providers, TSP’s (engineers,
agronomists, nutrient management specialists, and contractors) is deciding when storage
is necessary and if so, at what capacity? If a farmer with livestock in confinement and
the TSP’s decide that some storage is necessary, thereby shifting management of the

manure volume from daily haul (option one above) to some form of manure storage, two

89

questions arise. One, how much capacity is enough, and’two, when is the storage period
during the calendar year?

The Federal Clean Water Act of 1972 (U SEPA, 1972), followed by the Federal
Clean Water Action Plan of 1997 (U SEPA, 1997) set the stage for the Environmental
Protection Agency’s effort to monitor and regulate point source and non-point source
waste streams, including livestock manure, through a comprehensive program for
protecting our Nation’s water. Larger livestock farms have moved from NPS to Point
Source designations as they have installed manure systems, especially manure storage
ponds (USEPA, 2004). At the state level, Departments of Agriculture, Environmental
Quality, or Natural Resources operate statewide monitoring and regulatory activity within
the guidelines of the EPA General National Pollution and Discharge System Permit
(U SEPA, 2004). At the local level, governmental agencies further refine livestock
manure activity with site-specific requirements, including buffer zones, storage pond
location from wellheads, and air quality regulations.

Livestock owners can receive help from TSP’s, who assist them with the technical
components, system design, management, and regulatory permitting aspects of answering
these questions correctly (U SDA-NRCS 2004a). Manure systems are designed by the
Natural Resources Conservation Service, or by private engineering firms. Some livestock
owners participate in Federal Cost Share Programs, such as the Environmental Quality
Incentive Program, or they chose to pay for the construction costs themselves. USDA-
NRCS provides technical standards in the form of Engineering Construction Standards,
Conservation Practice Standards, and various Field Office Technical Guide Sheets

(U SDA-NRCS 2004b). Public and private TSP’s as well as livestock owners use these

90

standards. The management criteria of manure systems are constantly changing. The
USDA-ARS, USDA-NRCS, and Land Grant Universities are the usual providers. The
components of a manure system, including technical, system design, management, and
regulatory permit are compiled in a document known as a Conservation Plan (U SDA-
NRCS 2004c).

Most livestock owners have a conservation plan. Some are formal documents on
file with the USDA Farm Service Agency Office and NRCS, as well as the state and local
agency. Such conservation plans are part of the landowners RMS, or Resource
Management System, in which Soil, Water, Air, Plant, and Animal (including Human
Considerations) resources are considered in detail (USDA-NRCS, 2004c). Some are less
formal. A conservation plan may consist of a Comprehensive Nutrient Management Plan
(USDA-NRCS 2004c) that accounts for the spatial application of dairy manure nutrients
on cropland at agronomic rates. Still other landowners may have an informal set of

instructions as to how the farm manure system is managed.

2.25 Current scope

The Federal Environmental Protection Agency issued its Final Rule (U SEPA,
2003) establishing permitting guidelines for livestock owners. EPA objectives include a
robust monitoring program of Concentrated Animal Feeding Operation’s, defined as
large or medium size herds confined for at least 45 days with a waste-handling facility
(U SEPA, 2003). By definition, CAFO’s are point sources of potential manure discharge.
EPA estimates that in the U.S. there are 15,500 CAFO’s that will require participation in

the National Pollution Discharge and Elimination System (U SEPA, 2003). The agency

91

reports a national manure volume of 500 million tons of manure on all livestock farms,
and for CAFO’s, 300 million tons (USEPA, 2003). The impact of confining and
concentrating animals into barns on concrete on these large CAFO’s is clear: CAFO’s
make up approximately 1.5% of all livestock farms but produce 60% of the total manure
volume in the country. EPA’s goals include having farmers eliminate surface discharge
of livestock manures into surface water bodies, applying manure nutrients at an
agronomic rate, and reducing the impact of deep leaching of nutrients in groundwater
(U SEAP, 2003).

As owners of Concentrated Animal Feeding Operations, livestock owners are
updating their conservation plans, CNMP’s, and if necessary, adding or updating manure
handling, storage, and application components. CNMP providers (TSP’s) are being
trained to meet this demand (U SDA-NRCS, 2004a). The CNMP is a document based on
the accounting and flowpath of manure nutrients, specifically nitrogen, phosphorus, and
potassium based upon statewide standards, known as the NRCS 590 Nutrient
Management Standard (U SDA-NRCS, 2004b). Manure nutrients are applied to crop
fields in amounts that meet expected crop uptake, similar to applying fertilizer to a non-
manured crop field. Depending upon location and soil fertility, manure nutrients may be
restricted to the first limiting nutrient, generally phosphorus or potassium. On fields
where manure is applied, a field-specific risk assessment tool will be completed by the
TSP and discussed with the livestock owner.

The importance of the 590 standard as a CNMP component cannot be overstated.
The standard is maintained by the USDA-NRCS at the Federal and State Level (U SDA-

NRCS, 2004b). It is a practice standard that provides guidance to the farmer for manure

92

nutrient accounting and flowpath in both a spatial and temporal framework. However,

one deficiency in the standard exists and becomes the foundation of this research work.

2.26 Research gap

The technical knowledge for nutrient management is based upon a reasonable
understanding of nutrient accounting and flowpath through the farm system. The
technical knowledge for designing and building manure storage ponds to date is extensive
(USDA-NRCS, 2004b). Primarily driven by NRCS specifications, farmers, regulatory
personnel, and the public can be reasonably assured that if built correctly, manure storage
ponds do not represent a hazard for leaching in excess of design criteria.

The technical assistance for properly applying liquid manure and irrigation water
management as a function of field water status both in and out of the growing season are
not sufficient. This deficiency may be stated as a research gap. Here is the wording of
the National 590 Nutrient Management Standard:

0 “Nutrients shall not be applied to frozen, snow-covered, or saturated soil if the

potential risk for runoff exists.
0 Nutrient applications associated with irrigation systems shall be applied in
accordance with the requirements of Irrigation Water Management (NRCS
Standard 449 Irrigation Water Management).

0 The application rate (in/hr) for material applied through irrigation shall not
exceed the soil intake/infiltration rate. The field capacity of the soil shall not
be exceeded at any time.” (Page 3).

Here is the wording of National 449 Irrigation Water Management Standard:

93

“The following principles shall be applied for various crop growth stages:

0 The volume of water needed for irrigation shall be based on plant available
water holding capacity of the soil for the crop rooting depth, management
allowed soil water depletion, irrigation efficiency, and water table
contribution.

o The irrigation fi‘equency shall be based on the volume of irrigation water
needed and/or available, the rate of crop evapotranspiration, and effective
precipitation.

o The application rate shall be based on the volume of water to be applied, the
frequency of irrigation applications, soil infiltration and permeability
characteristics, and the capacity of the irrigation system.”

Currently, these standard criteria are unanswerable. Farmers and TSP’s lack the
knowledge of the FWB to understand water flow as functions of infiltration,
redistribution, drainage, and the relationships of volumetric water content and pressure
potentials. In normal conservation planning activities, TSP’s have avoided this step due
to inadequate research, the lack of a model framework for extending the data set beyond
the experimental field, and the use of a reductionist approach instead of a systems

approach.

94

Chapter Three. Study Area

3.1 Willamette River Basin

A starting point for a systems research study is definition of macro scale, further
defined by spatial and temporal boundary, and then comprised of field units. The macro
scale component is the Willamette River Basin (WRB) [USGS Hydrologic Unit Codes:
HUC 1709005 (North Santiam), and HUC 1709007 (Middle Willamette)] in Oregon
(Hulse et al., 2002).

This geographical area is 160 km north to south and 161 km east to west (29,728
kmz), situated between the Coast Range to the west and the Cascade Range to the east.
The latitude boundaries are approximately 121 .38° W and 123.51° W. The longitude
boundaries are 43.24° N and 45.70° N. Elevation varies from 3 m above mean sea level
(confluence of the Columbia River) to a snow capped peak, Mt Jefferson, on the west
side at 3199 m.

The WRB recent topography was derived from volcanic eruption, lava flows, and
mudflows along with glacial outwash, and river alluvium. The northern end of the WRB
(and the study site for this work) is underlain by basalt erosion over the last 15 million
years. 15,500 to 13,000 years ago the ice dams formed large lakes in Montana that once
broke free, caused sedimentary deposits to settle on the valley floor. Closer to the river
systems, primarily the Willamette River, sediments were slightly coarser (larger) thereby
giving rise to better-drained soils that become the primarily agricultural region of the

basin.

95

Humans have lived in the Basin for about 10,000 years. Several Native American
tribes, including the Calapooia, Luckiamute, Yamhill, Clackarnas, and Kalapuyan lived
on the landscape especially near the river systems. Euro-American Settlers moved into
the Basin about 1820-1840. They were drawn here by fertile soils, moderate climate, and
a river transportation system. The Oregon Trail is the journey of settlers arriving at the
Basin; pioneers left the Midwest in search of a new life in the wilderness. Oregon was
organized as a territory in 1848 and admitted as the thirty-third state in 1859 with a
population of 13,294 settlers as they were known. One hundred years later the population
of the WRB had grown to 1,521,341. The population of the Basin surpassed 2 million in
the late 1980’s, and the present population (2003) is approximately 2.3 million people.
The three largest cities in the Basin are located on the Willamette River. Portland is
Oregon largest city, Eugene is Oregon’s second largest city, and Salem is Oregon’s

Capital City.

3.2 WRB weather

The WRB has a modified marine climate (Hulse et al., 2002). Influenced by the
Pacific Ocean, the Coast range and the Cascade Range, the weather pathway is generally
from west to east. The marine air is generally nearly saturated with water. The western
Basin boundary, the coast range, cools the air as it rises in elevation. Due to a lower
saturation vapor pressure at higher altitudes, the moisture condenses as water droplets
and falls as rain or snow. Most of this precipitation falls on the west slope of the
mountain range. A drier air flows into the Basin, although enough moisture exists for

rainfall especially in late fall, winter, and early spring. As clouds near the Cascade

96

Range, the elevation lift once again condenses water as rain or snow. Land (foothills)
near the Cascade Range can receive significantly more precipitation than more central
valley locations.

The agricultural farming areas are found near the central valley locations. Here,
about 70% of the annual precipitation of 1,120 mm occurs during November through
March (150 days). During most of the growing season for agricultural crops, less than
seven percent of annual precipitation occurs in June, July, and August. Accumulative
snowfall is rare and if so, last no more than four to five days.

Average maximum temperatures are 178°C and average minimum temperatures
are 5.6°C. Only an average of 15 days annually do temperatures drop to 322°C.
Temperatures reach below freezing about 60 days annually and when they do, only
slightly below freezing. In July, the driest month, RH averages 40%. The average for the
year is 89%. The last freezing temperature day can fall in March or April and June in
extreme cases, and the first fall freeze at the end of the growing season is usually in late
October or early November. Therefore the growing season can be as long as almost nine
months to as short as six and a half to seven. Typical growing seasons are 225 days or

seven and a half months: April through the middle of October.

3.3 WRB Soils
An in-depth discussion of soils sampled for this research project is found in
Chapter Three. The discussion in this chapter pertains to a general morphology and

inventory of soils in the WRB (Hulse et al., 2002).

97

The soils in this Basin have formed from eight major kinds of parent material.

Each is described below:

1.

Recent Alluvium. These are the youngest soils in the Basin. They are flood
remnants found near surface waters, streams and creeks, washed from basic
igneous rock.

Gravelly Alluvium. These coarse soils were deposited as outwash from the
Cascade mountain range during the late Pleistocene glaciation (15,000 —10,000
years ago). The predorrrinant soils are gravel and sand.

Young Silty Terrace Alluvium. These soils are found in the best agricultural
fields of the Basin, including the soils in this study. They are the result of floods
settling stratified sand, silt and clay sediment that contains large amount of quartz,
in regions above the flood plain (do not flood in modern times). These soils are
somewhat modified by windblown material and slight weathering, but are no
older than 20,000 years.

Weak Consolidated, Old Gravelly Alluvium. These are older, more highly
weathered soils derived from basalt. They are found in elevations above the
valley floor flood plain, in what are known as foothills. Primarily comprised of
silt, they are easily eroded and therefore are usually planted to perennial crops like
Fescue or Ryegrass seed.

Colluviums from Basalt and Massive Tuffs. These soils are derived from the
Miocene and Pliocene, two to 20 million years ago, and are usually highly
weathered. They are foothill soils and higher, populated by timber and brush.

Agriculture crop production is minimal.

98

6. Sedimentary Alluviurn and Colluviums derived from Tuffaceous Sandstone and
Shale that are made from sandstone are quite old, 20-50 nrillion years ago. Soils
are suitable for timber and brush and are not usually farmed.

7. Glacial Till. These are recent soil laid down during the latterrnost Glacial Period
of 18,000 years ago during the Wisconsin Period. These soils are course textured
or made up of gravel or cobblestones. Soils are suitable for timber and brush
production and are not usually farmed.

8. Organic Material. These muck or peat soils were deposited in former shallow
lakes during the Wisconsin Period. Little development of the soil profile has

taken place; however these soils are quite productive for some agricultural crops.

3.4 Soil capability classification

Soil Capability Classification is a ranking system that describes a soils’ suitability
for agricultural crop production. The system ranks soils from I to VIII, for eight
classification classes. A Class I soil has no limitations and therefore is the most adapted
towards all kinds of agricultural crop production. A Class VIII has many limitations and
is generally not suitable for any crop production. A graduating scale from productive to
non-productive is further refined with lower case letters that designate what kind of
restrictions make a particular Class of soil fit in that rank. Examples are e — erosion
potential, w — wet soil, and s — shallow soil).

For soils in the WRB, 27% of all soils are found in Class 1, Class II, and Class III,
generally considered the soils that can be tilled, farmed, and used to support agricultural

production. Classes IV though VII comprise 71% of all soils. These are soils suitable

99

from timber, brush, Christmas trees, woodlots, and open meadow grass production. Class
VIII soils, at the remaining 2%, are unsuitable for any plant growth other than a remote
tree or brush plant having found a pocket of soil to grow. The soils examined for this
study are Class II or Class III, and in a majority of fields, carry the restriction of w -— wet

soil.

3.5 WRB dairy industry

In 2002 according to Oregon Department of Agriculture, 94 dairy farms produce
Grade A rrrilk in the WRB (Oregon Department of Agriculture, 2004a). Herd size ranges
to as few as 24 cows to as large 2,800. Dairy farms are predominately owned by Dutch
and Swiss owners, having migrated from other parts of the U.S. or overseas. Many dairy
farms are second and third generation; a fourth generation farmer, usually a graduate of a
Land Grant University or technical college that offers dairy training, operate several
farms. Latino workers are commonplace on these dairies.

The majority of milk enters the fluid or manufacturing pool and is transported
interstate as finished dairy products. Few shipper handlers exist, but for those producing
and processing milk, the end product is usually fluid milk. There is an increased interest
in organic dairy farming in the WRB. These farmers produce for a growing market, and
while finding organic feed is a challenge, for those smaller herds (<200 cows) this niche
market helps them maintain profitability.

Dairy farmers in the Basin have built free stall barns that have individual stalls for
cows at rest, and feeding mangers for handling the total mix ration volume of feed ration.

Cows live their entire productive lives in these barns and the milking facility, except

100

during the dry period when they may be grazed if the grazing paddock is dry enough.
Floors are entirely concrete at slopes of One half to three percent; the predominant slope
for alleys is one and a half percent, providing direction for the flush volume and ease of
manure scraping flow. In both cases, the manure volume is directed to a reception pit or
tank. From here, the manure volume flowpath may be one of three directions: directly
hauled to a crop field, pumped or gravity flowed to a settling basin or treatment process,
or pumped into a waste storage pond. Generally, the manure flowpath includes the
manure storage pond. This pond also receives rainwater and wastewater from the milk
house.

The four dairy farmers in this study own and operate typical facilities. All have
manure storage ponds designed by USDA-NRCS and built by private contractor under
the engineering guidelines of NRCS Engineering Staff. All systems were part of a cost

share program through the U.S. Government.

3.6 Regulatory oversight

Oregon is an EPA Region 10 designated state (ODA, 2004b). Historically,
Oregon Departrrrent of Agriculture maintained a General Water Pollution Control
Facilities Permit that provided a regulatory permitting mechanism for dairy farmers
across the state. A recent shift has occurred (2001) towards an EPA National Pollution
Discharge Elimination System General Permit for the dairy industry in the state. The
permit is jointly owned by Oregon Department of Environmental Quality, but is housed
in and regulated by ODA. EPA has released its Final Rule in 2003. Oregon is adopting

the necessary strategy towards maintenance of the General Permit, with the additional

101

criteria of requiring a large Concentrated Animal Feeding Operation to obtain an
individual Federal NPDES Permit (outside the boundary of the General Permit), or an
Oregon Water Pollution Control Facilities Permit.

Federal Law 40 CFR Part 122 and Oregon Law ORS 468B.050 (Oregon
Department of Agriculture, 2004c) prohibit the discharge of pollutants to waters of the
state without a permit. ODA describes pollutants as: biochemical oxygen demand
(BOD), total suspended solids (TSS), organics, bacteria, and the nutrients nitrogen and
phosphorus. All dairy farms will have an updated permit and meet performance criteria
by December 31, 2006. Two new criteria include requiring the farmer to keep a more
robust set of records, such as soil test values, nutrient accounting and flowpath, and
weather conditions, and an Annual Report (AR). The AR will provide ODA with a
historical record of nutrient and volume fate. Additional inventory requirements track
animal numbers, acres receiving manure, crop yield, and if performance criteria were

actually met. An ODA employee will inspect dairy farms at least twice annually.

3.7 Study introduction

In October 1995, the investigator and several NRCS staff met in Salem Oregon
for a technical transfer meeting. The end result of that meeting was the initial planning
for this study. The author agreed to take the lead in a long-term study attempting to
determine the F WB on four typical dairy farms in the WRB. The group of approximately
20 scientists and engineers identified this one objective as the most critical research gap
for the dairy industry, the CAFO program in Western Oregon and specifically manure

application on cropland as an acceptable agronomic and hydraulic practice in the WRB.

102

A proposal was submitted to the Oregon Dairy Farmers Association - Portland,
Oregon Department of Agriculture - Salem, Oregon Department of Environmental
Quality - Portland, Oregon NRCS - Portland, and Oregon State University — Corvallis.
The proposal requested funding for the study and a 0.5 F TE by the author to conduct it.
Both were approved. The majority of the funding came from a DEQ 319 Grant
administered by Oregon DEQ and Region 10 EPA in Seattle. All other groups supported
this study with additional dollars, material, field time, and equipment. During the first
half of 1997, the 0.5 FTE of time (primary investigator) and the funding and additional
support were ready to begin. The next section describes in some detail the materials and
methods used for the initial implementation and the strategies used for maintaining the

study for a five-year period.

3.8 Dairy cooperators in the WRB

The task of selecting cooperators was straightforward. Very large herds and very
small herds were considered outliers and therefore not representative of a typical WRB
dairy. The target herd size (range) was 250 to 500 cows.

Three selection criteria were used:

1. Cooperators would agree to a long-term study of five years. Access to the farm at
any time for measuring any parameter, such as irrigation flow rate, obtain soil
samples, or recording pump hours by the investigator or technician was an
absolute requirement. Cooperators would agree to any examination and scrutiny
of management practices, farming techniques, and records accumulated for this

study. In other words, access to all records with nothing kept from the record

103

keeping process. One significant aspect of this particular requirement is the study
duration of five years. During the initial selection period, the investigator
interviewed 14 dairy farmers, and every one of them mentioned that this study
duration was longer than any other study they had participated with.

2. Cooperators operated businesses that were likely to remain in business for the
study period. The attrition rate of dairy farmers leaving the dairy business in
Oregon is approximately five percent annually. A consideration was given
towards the family dynamics. That is, an assessment of a stable farm family
environment, among family members on the farm and away from the farm, as
well as a labor force that exhibited enthusiasm for working on the farm with
relatively low turnover. The investigator interpreted these assessments subtly.
No written or formal evaluation was conducted.

3. Cooperators would agree to acknowledge, accept, and change any management
practice and farming techniques during the study period based upon the recorded
results and the recommendations of the investigator and other technicians. The
objective of this requirement was the basis of all planning activity: records are
used to monitor and evaluate performance. When the evaluation review suggests
corrective change, then the farmer considered an alternative strategy.

Four dairy farms and cooperating farmers were selected during the summer 1997. The
author made the selections. The agreement between the investigator and the farmer
included the following items:

1. The investigator will provide the cooperators with individual record analysis

during the study period. Specifically, soil test analysis, animal inventory, a

104

summary of water input for irrigation and liquid manure volumes, and a summary
of crop yield. All of these were done on a field basis annually.

2. The investigator would cover all expenses incurred on each farm though the DEQ
319 Grant or other public funds. The investigator or SWCD or NRCS personnel
would provide labor requirement.

The individual farm records are the property of the cooperator and the investigator.
Cooperating farms are identified by numbers one though four. Analytical data derived
from all laboratory analysis is the property of the cooperator and investigator, and

identified only by farm number. Individual fields are numbered one through 11.

3.9 Cooperating dairy farm descriptions

A description of each selected farm follows. All four owners and their families
agreed to the selection criteria requirements, and the record keeping protocol. Each
farmer had a good working relationship with NRCS and SWCD, the regulatory
community, ODA, DEQ, and Oregon State University Extension Service. Each dairy had
been in business at least 25 years and was now operated by at least the second generation.
All have plans for change, primarily in improving efficiency not increasing herd size. In
each case, the second generation is college educated with at least a two-year degree.
There are young children on each farm (third generation), and in three out four farms,
retired family members living on or near the farm. All farm owners were eager to
participate and through the five— year period, endured the record-keeping requirement

long after the newness worn off.

105

1.

F arm 1. This dairy farm is located (facility headquarters) at 45.44°N and
123.31°W. Approximately 320 dairy cows are housed in total confinement fiee
stall barns. Manure is scraped into two reception tanks and subsequently pumped
directly into one of two storage ponds. Storage pond capacity is 30.8 x 103 m3 (25
acre feet). Manure is applied using a traveling irrigator (big gun on a hard hose
reel) with an average output of 1.211 m3 s'1 (320 gpm). All animals are confined
365 days per year. Cropland includes 58 hectares (144 acres). Crop rotation is
corn silage for a summer crop and annual ryegrass for a winter crop. Both are
mechanically harvested as silage, stored in bunker silos on the farm, and fed to the
dairy herd. This is a first, second, and third generation farm.

Farm 2. This dairy farm is located (facility headquarters) at 45.71°N and
122.93°W. Approximately 220 dairy cows and 200 dairy replacement heifers are
housed in partial confinement free stall barns. Manure is scraped in two barns
(heifers) and flushed in one barn (cows) into one reception tank. Manure is
subsequently pumped over a sidehill manure separator before entering one of two
storage ponds. Storage pond capacity is 7.4 x 103 m3 (six acre feet (AF )).
Manure is applied using a traveling irrigator with an average output of 1.022 m3
sec’1 (270 gpm). Heifers are grazed about seven months per year, and cows are
grazed about six months per year. Cropland includes 104 ha (258 acres). Crop
rotation includes permanent pasture, corn silage and perennial grass silage for
summer crops, and annual ryegrass following corn for winter crops. Most feed is
mechanically harvested; some feed is grazed. All feed is fed to the dairy herd.

This is a first, second, and third generation farm.

106

3. F arm 3. This dairy farm is located (facility headquarters) at 44.87°N and
123.02°W. Approximately 330 dairy cows and 380 dairy heifers are housed in
total (cows) or partial (heifers) confinement free stall barns. Manure is scraped in
three barns (heifers) and flushed in two barns (cows) into three gravity basins that
remove a portion of the sand bedding. Liquid manure is subsequently gavity
flowed via weir box to one of three storage ponds. Storage pond capacity is 39.4
x 103 m3 (32 AF). Manure is applied using three traveling irrigators with an
average output of 1.173 m3 s'1 (310 gpm). Heifers are grazed about seven months
year". Cows are confined 365 days per year. Cropland includes 81 ha (201
acres). Crop rotation includes permanent pasture, corn silage and perennial gass
silage for summer crops, and annual ryegass following corn for winter crops.
Most feed is mechanically harvested; some feed is grazed. This is a first, second,
and third generation farm.

4. Farm 4. This dairy farm is located (facility headquarters) at 45.22°N and
122.51°W. Approximately 360 dairy cows and 400 dairy heifers are housed in
total confinement free stall barns. Manure is flushed in all barns into one
reception tank. Manure is subsequently pumped to a sidehill manure separator
before entering one of two storage ponds. Storage pond capacity is 51.5 x 103 m3
(41.6 AF). Manure is applied using two traveling irrigators with an average
output of 0.946 m3 s”1 (250 gpm). All animals are confined 365 days per year.
Cropland includes 113 ha (280 acres). Crop rotation includes primarily alfalfa for
silage and hay, but oats and pea silage is gown as a winter cr0p, and perennial

and annual ryegass is gown as well. This is a first and second generation farm.

107

A physical descriptive summary of each farm appears in Table 3.1.

Table 3.1 Cooperating farm physical descriptions including storage capacity, animal units
per unit storage, hectares per unit storage, and hectares per animal unit

 

 

 

 

 

 

 

 

 

 

 

 

 

Farm AU ha Crops storag:31,000 AUQSOOO ha/I1r;§)00 ha/ AU
1 221 58 com/gass double crop 30.8 7.2 1.9 0.26
2 211 104 com/gass, perennial gass 7.4 28.5 14.1 0.47
3 378 81 com, perennial gass 39.2 9.6 2.1 0.21
corn, perennial gass,
4 307 113 oat/peas, alfalfa 51.5 6.0 2.2 0.37

 

 

A map (Figure 3.1) shows the location of each farm, the location of two AgiMet
WS, the location of the National Weather Service WS (N 0AA), and the Willamette —

Columbia River systems.

N Willamette Valley

w$n sr dy Site

,- ..§q \u r' .

'Z i:-

.‘b ,r .l

1.. ;“"I.' gr, :1 ‘
. Washington ‘

     

r \ h, ‘
5 R"-

   

‘w- .m « ,.
(<v‘. t . . 1.. .* ."-. r
., _ Columbia River

.1...

 

to o to .. 100 mm
Figure 3.1 Study site farm and weather station location map

108

Chapter 4. Materials and Methods

4.1 Overview
There are three parts in this chapter that describe the materials and methods used
in the study area:

0 Installation and use of Weather Stations (W S) and Watermark conductivity
sensors placed in each of ten fields in the study area. Data for RZWQM model
parameterization includes AgiMet data.

0 Site selection, soil core sampling, and determination of soil characteristics
including bulk density, particle size analysis, and soil water characterization curve
at specific depths for each field.

0 The parameterization of RZWQM, including strategies for calibration and

simulations of FWB derivation.

4.2 Installation and use of weather stations
There was one WS installed on each of the four farms in the study area. A
description of each follows:
Farm One
This farm is located in Polk County, Oregon. The WS is located at:
o WSl 45.0453° N l23.0768° W USGS Mission Bottom Quadrangle, OR
The instruments include a rain gauge, an ambient air temperature gauge, a soil

thermocouple at 10 cm depth, and a solar radiation gauge. Four Watermarks were

109

installed at 30, 60, 90, and 120 cm depths in two fields. Field numbers and coordinate
locations were:

0 1-1 45.0455° N 123.0773° W USGS Mission Bottom Quadrangle, OR

o 1-2 45.0442° N 123.0770° W USGS Mission Bottom Quadrangle, OR

All instruments including climate and below gound were hard wired to a
Campbell Scientific Data Logger CX30 using 18 gauge copper wire wrapped in tensile
fiber encased in plastic. The data logger was set to record climate and conductivity
measurements on lS-minute intervals.

The WS itself was built on a 15 cm by 15 cm square treated post placed in the soil
approximately 1 m deep and encased in concrete. Climate instruments were placed about
2.5 m above the soil surface, while the data logger was installed in a weatherproof epoxy
coated electrical box. A 20 cm by 45 cm solar panel was installed on the post, facing
approximately south at a 45-degee angle above a horizontal plane.

The WS was hard wired to a telephone pedestal placed within 1 m of the WS post.
The pedestal was hard wired to a roadside pedestal, which was connected to the local
telephone line that was oriented alongside the roadway. The telephone company
assigred a specific telephone number to the WS, and the landowner was responsible for
maintaining the payment for this telephone number. A Campbell Scientific COM300
Modern was installed in the WS. This modern was a voice synthesizer that converted
CX30 telemetry into a digitized voice. This function allowed users to dial the WS using
a standard landline or cellular phone, and obtain real time telemetry (given a 15 minute
telemetry dump into the data logger) of climate and soil conductivity data for every field.

The landowner liked this feature and made use of it every day. They were particularly

110

interested in accumulated rainfall since midnight (the daily cycle function began at
midnight), air temperatures, and the soil temperature especially in spring. The station
WSl became operational May 14, 1999. The landowner did not purchase telemetry
software (MeasureTekO) for farm use.
Farm Two
This farm is located in Clackamas County, Oregon. The WS is located at:
e WS2 45.2983° N 122.6296° W USGS Canby Quadrangle, OR
The instruments include a rain gauge, an ambient air temperature gauge, a soil
thermocouple at 10 cm depth, and a solar radiation gauge. Four Watermarks were
installed at 30, 60, 90, and 120 cm depths in four fields. Field numbers and coordinate
locations were:
0 2-3 45.2985° N 123.6320° W USGS Canby Quadrangle, OR
o 2-4 45.2987° N 123.6300° W USGS Canby Quadrangle, OR
o 2-7 45.3004° N 123.6294° W USGS Canby Quadrangle, OR
o 2-8 45.3009° N 123.6325° W USGS Canby Quadrangle, OR
The instrument package on WS2 is exactly the same as WS 1 , with the exception
of an additional data storage unit and different telephone number. The station W82
became operational June 6, 1999. The landowner (Farm 2) purchased telemetry software
(MeasureTeko) for farm use.
Farm Three
This farm is located in Marion County, Oregon. The WS is located at:

e WS3 44.7663° N 122.0358° W USGS Sidney Quadrangle, OR

111

The instruments include a rain gauge, an ambient air temperature gauge, a soil
thermocouple at 10 cm depth, and a solar radiation gauge. Four Watermarks were
installed at 30, 60, 90, and 120 cm depths in three fields. Field numbers and coordinate
locations were:

0 3-1 44.7670° N 123.0342° W USGS Sidney Quadrangle, OR
o 3-6 44.7649° N 123.0385° W USGS Sidney Quadrangle, OR
o 3-7 44.7658° N 123.0419° W USGS Sidney Quadrangle, OR

The instrument package on WS3 is exactly the same as WSl, with the exception
of an additional data storage unit and different telephone number. The station WS3
became operational July 25, 1999. The landowner (Farm 3) did not obtain telemetry
software (MeasureTek©).

Farm Four
This farm is located in Marion County, Oregon. The WS is located at:
0 W84 45.2019° N 122.8764° W USGS St. Paul Quadrangle, OR

The instruments include a rain gauge, an ambient air temperature gauge, a soil
thermocouple at 10 cm depth, and a solar radiation gauge. Four Watermarks were
installed at 30, 60, 90, and 120 cm depths in one field. Field number and coordinate
location was:

0 4-1 45.2004° N 122.8783° W USGS St. Paul Quadrangle, OR

The instrument package on W84 is exactly the same as WSl, with the exception
of a different telephone number. The station WS4 became operational November 22,
1998. The landowner (Farm 4) purchased telemetry software (MeasureTek©) for farm

use.

112

A photogaph of WS4 is shown in Figure 4.1.

 

 

Figure 4.1 Weather Station, Farm 4

4.3 WS software

Two software packages were required to obtain and analyze telemetry. They
were developed by MeasureTek° of Corvallis Oregon. A description of each package
follows:

1. Connect+ This software was loaded on a remote laptop computer with a dialup
modem. Each WS used the voice synthesizer modern as default. Upon laptop
command, however each station could be called using the dialup routine, with an
additional several second delay in the dialup routine so the voice modem would

defer to the telephone modem. Once the modem linkage was established between

113

remote laptop modem and WS data logger, a downloading sequence began. Data
were placed in a Microsoft Access database named for each station. At each
dialup new data was added to existing data, creating a new file while storing the
previous file as a backup. Two landowners (Farm 2 and 4) chose to purchase
Connect+ for farm use on local computers. Each downloaded the WS on their
farms only. The remote laptop used in this study by the author was the only
machine capable of downloading all four stations. As a precaution, each WS was
configured with a local USBS port cable directly out of the data logger. This
became usefirl if the telephone system failed (this occurred several time at each
WS) and a data dump was required. The data logger system had a capacity of
approximately three months before data was overridden. Connect+ is a standalone
Windows based software.

. ResultX+ This software was loaded on the same remote laptop as Connect+.
ResultX+ compiles each WS access file into user gaphic and table format. Data
from ResultX+ can be directly exported to Microsoft Excel, and this was the
preferred method for this study. This software was desigred to work in
conjunction with Connect+. Once the retrieving software (Connect+) obtained
new data via telephone delivery modems, and then ResultX+ imported new data
and placed it in an updated MS access file, thereby creating another backup of
telemetry. Two landowners (F arm 2 and 4) chose to purchase ResultX+ for farm

use on local computers. ResultX+ is stand alone Windows based software.

114

4.4 Watermark conductivity sensors

Watermark sensors were being installed in fields by many landowners as part of
an irrigation water management progarn in the Willamette Valley Oregon, during the
late 1990’s and early 2000’s. The rationale for their use was simple: they were
inexpensive ($25-30 per sensor), durable (years of use), easy to install using a soil coring
tool (T-Handle, 30 mm diameter auger, 1.5 m in length), and relatively easy to use once
landowners purchased a relatively inexpensive ohmmeter. A photo of a handheld

ohmmeter attached to one Watermark is shown in Figure 4.2.

 

Figure 4.2 Watermark sensor and hand held ohmmeter

The Watermark sensor is made of gypsum crystals encased inside a permeable
membrane and perforated stainless steel. Two 18 gauge wires enter the cap and of the
sensor. These wires each have a stainless steel electrode embedded in the gypsum matrix
approximately 10 mm apart parallel to each other. The wire cap and base end are geen
plastic resin that contain the gypsum, the membrane, and surround the stainless steel

shell. The sensor can be purchased with wire lengths of 60 to 150 cm length.

115

For this study, the soil coring tool (mentioned above) was used to bore a hole into
the soil at the desired depth. At each field location, the sensor was installed so the mid
point of the sensor was approximately at the desired depth. The bore hole was at some
angle based upon the coring direction, but for each sensor in the profile sat at
approximately a 45 degee angle, with the soil surface at 0 degees and a vertical
direction at 90 degees. Once the borehole was completed, a saturated sensor (wet in a
five gallon bucket for at least 48 hours) was placed in the hole. A 6 mm steel rod was
placed on the sensor cap. Carefully the sensor was shoved to the bottom of the borehole.
The rod was removed. Soil was repacked into the borehole by taking the approximate
soil at that point in the profile, thereby attempting to minimize textural changes for each
depth. The repacking was facilitated by a wooden dowel approximately 25 mm in
diameter. For each sensor, the wires were carefully kept unscarred as soil was repacked,
and on several occasions an entire sensor was replaced because insulation was lost. For
each field, the sensors at 30, 60, 60, and 120 cm depth required two hours installation
time.

The Watermark sensor is based upon electrical resistance and variable
conductivity in the soil based upon water content. The gypsum provides a soil buffer of
ionic activity in soils. As the gypsum matrix in the soil is wetted, the conductivity of
electrical current, i.e. voltage, increases. A wet soil will conduct an electrical current
more readily than a drier soil. The hand held ohmmeter uses a 9-volt DC battery. A
current of five volts is sent to one electrode in the sensor; some voltage is lost due to
voltage resistance, or ohms. The twin electrode receives the remaining voltage and that is

recorded in the ohmmeter as a function of the original voltage. The Watermark has a

116

range of 0-200 centibars, or 0-2 Bars. An algorithm in the ohmmeter progamming
converts this change of voltage at each reading into an estimation of matric potential in
the soil.

Generally, the understanding by technicians in the field is that a reading of zero
means a fully wet soil, or as near to saturation as that part of the soil surrounding the
sensor can get. As the soil dries, the relative change fiom one reading to another
indicates that the soil is drying or wetting, but only in a qualitative sense. This is an
important concept. For purposes of irrigation water management, technicians would
instruct landowners that when the sensor yielded a reading of 70 to 80 centibars, an
irrigation application was required. The irrigation application should rewet the sensor to
20-30 centibars, at or near field capacity. Many landowners checked their sensors arrays,
a goup of sensors placed at 15, 30, and 45 cm depths (approximating a root zone depth)
every two or three days. Those using an irrigation water management strategy recorded
these data on gaph paper, and based upon the drying rate scheduled irrigation
applications based upon sensor readings. However, landowners did not have these
measurements quantified directly to some volumetric measurement in the soil. They
knew, however, that for a particular field if the sensor was to be rewet to 30 centibars
from 80 centibars, 40 mm of water was required if the uppermost sensor in the profile
was at least 15 cm deep.

For this study, Watermark sensors were used to accomplish this same strategy,
except that two additional criteria were added. One, a record of multiple depths and their
relative wetting or drying trends would be recorded over a long interval in and out of the

gowing season. Two, some understanding of when the soil was geater than field

117

capacity, i.e., at or less than 20 to 30 centibars, and when the soil was near or at
saturation would be known. This second point was important for scheduling limited
applications of liquid dairy manure outside of the gowing season, thereby assuming that
if the sensor had some reading geater than zero, then some void space existed and
infiltration could occur.

The sensor array in each field was hardwired to the farm WS. A trencher was
rented and used to dig a 10 cm wide by 60-90 cm deep trench between the array and the
WS. Bundled 18-gauge wire was laid in the trench. At the array end, silicone connectors
were used to seal the trenched wire and sensor wire end. This entire bundle was placed in
a plastic boot approximately one half liter in size, sealed with silicone resin, and buried at
a depth of 60 cm about one meter from the sensor array. The objective during trenching
and sensor burying was avoiding tillage equipment plowing the uppermost sensor (30cm)
out. One rationale for using these sensors was their low cost. Several 30 cm sensors
were plowed out and actually lost over the study period. At the next dialup, the telemetry
was dramatically different for these sensors; they were replaced once the trench wire was

found.

4.5 AgriMet data

As described earlier in this thesis, AgiMet WS offer landowners a regional
weather network for assessing above gound climate, including a calculated ET,. For this
study, two regional AgiMet WS were used. They are located in Aurora and Corvallis at
the following coordinate location:

0 Aurora ARAO 45.2811° N 123.7501° W USGS Canby Quadrangle, OR

118

0 Corvallis CRVO 44.6339° N 123.1929° W USGS Lewisburg Quadrangle, OR

4.6 Resolution of WS data

RZWQM model requirements include two climate files: rainfall and climate data
for determining ET,. For this study, rainfall data from the local WS’s were used when
each station was operable during the five-year study period. For time periods outside
these windows, NOAA data was obtained fiom the Salem, Oregon NOAA weather
station based at the regional airport. The NOAA WS is based at the following coordinate
location:

0 Salem NOAA 45.9108° N 122.9956° W USGS Salem Quadrangle, OR

This WS was centrally located in the study site, with two farms north, one
farm west, and one farm south of Salem. The rainfall climate file for RZWQM consisted,
therefore, of regional data taken from the NOAA WS if the local farm based WS does not
have these data.

The climate data file in RZWQM requires five parameters: minimum air
temperature (C°), maximum air temperature (C°), wind run (km day'l), relative humidity
(0-100%), and solar radiation (MJ m'2 day”). From these, RZWQM calculates pan
evaporation, and then ET,. The AgiMet WS data from both stations was used in this
study; The CRVO WS was used for all climate files beginning October 1997 until
October 1998, when that station’s data were used to populate the RZWQM for all four
farms.

The resolution of weather data is complicated by the fact three different sets of

WS systems are used: four local MeasureTek local WS’s, two regional AgiMet WS’s,

119

and one regional NOAA station were used to populate RZWQM. The challenge was
obtaining enough data over a multiple year study period, and the study period beginning

before local and one AgriMet station became operational.

4.7 Soil characterization
The objectives of this soil characterization work were threefold:

1. Determine the bulk density and particle size analysis [PSA] (Gee and Or, 2002)
for soils in ten fields (listed above) at four incremental depths (30, 60, 80, and
120 cm).

2. Using the Klute (1986) method, derive the soil water characterization curve
[SWCC] based upon paired data of matric potential and volumetric water content
over the range of saturation to 1.5 MPa for soil cores taken from these fields and
depths.

3. Populate RZWQM soil database with data derived from each set of

characterization data for model parameterization.

4.8 Soil coring field work

A soil-coring tool was purchased from Soil Moisture Corporation in Santa
Barbara, California (see figure 4.3). This is a state of the art tool for obtaining as
undisturbed as possible soil cores. The tool consists of a stainless steel shell that houses a
set of internal brass rings. The tool has a handle for shoving the core shell into the soil
matrix at the desired depth. An open bucket auger was used to auger vertically to the

desired depth, then the coring tool was placed in the soil and extracted. Once out of the

120

gound, the shell is removed from the handle, the brass rings are carefully removed from
the shell, and using a putty knife, the cores or leveled at both ends. Two identical cores
are taken. One was placed on a field scale and weighed. The other was wrapped in saran
wrap, placed in a cooler, and eventually was used in the soil extractor method (Klute,
1986).

The soil removed during the open bucket auger coring was placed in a five gallon
plastic bucket. A sample of this volume of soil (approximately 6 kg) was placed in a
large plastic container and stored in another cooler. This disturbed bulk soil sample at a
known depth was used to determine particle size distribution (Gee and Or, 2002).

The location in each field was determined by the placement of Watermark
Sensors in each field. An assessment of field topogaphy and the USDA Soil Survey,
with input from each landowner was used. The objective was placing the sensors as
typical as possible part in of the field. Once the site was determined and the Watermarks
installed, the site was identified using a Survey Grade GPS supplied by USDA-NRCS
technicians. The soil coring locations were based upon three replicates at an approximate
distance of one m from the Watermark location at one third of a circumference, or 120°
fi'om zero degees, or North. These assigrments were not exact; several times during the
coring process a large rock obstructed the bore hole and another holes was began a few
cm near the abandoned hole.

The wetness of the soil in each field at the time of soil sampling was considered.
Soil coring was done at or near FC, thereby reflecting some similarity between bulk
density measurements in the field and those used by USDA-NRCS Soil Laboratory in

Lincoln Nebraska. Also, removing wetter soil as compared to drier soil enhanced core

121

integity. Users of this tool quickly discovered that sampling in dry soils is more difficult
in terms of driving the stainless steel shell into dry soils, and once that was done, the soil
literally fell out of the brass rings upon extraction. An additional challenge was avoiding
compaction of the soil in the brass ring assembly. Driving the tool into the soil matrix at
a depth even slightly beyond the total vertical depth of the tool shell will compact soil
and therefore increase bulk density.

The measurement taken in the field, i.e. the mass of one of the rings containing a
soil core was recorded. At every depth, the same brass ring was used to obtain a weight
of soil, which included the brass ring. This particular brass ring, therefore, had a known
mass and known volume. The bulk density measurement was immediately known given
the total mass minus the mass of the ring itself, divided by the volume of the ring
assuming water is 1 g cm".

At the end of each field sampling, usually one half day of time, the undisturbed
soil in plastic baggies and the soil cores in brass rings were taken back to the laboratory
at Oregon State University. The bulk soil was spread on a newspaper for air-drying.
After 48 to 72 hours at room temperature, the entire air-dry soil was placed back in its
plastic container for storage at room temperature. The brass rings containing the soil
cores remained in their plastic saran wrap and were stored in a refrigerator unit transport.

At this point the bulk density measurements were known, the bulk soil for PSA
(Gee and Or, 2002) was air dried and stored, and the soil cores for the Klute (1986)
method analysis were safely wrapped in plastic, stored in an environment minimizing

water loss, and kept at a temperature that minimized soil rrricrobial activity. A total of

122

120 bulk soil samples for PSA and 120 soil cores in brass rings for SWCC were
inventoried and the fieldwork was completed.
Figure 4.3 shows an example of a soil core shell and soil just removed from the

field bore hole.

 

Figure 4.3 Soil core shell and soil

4.9 Laboratory work, particle size analysis

The PSA is a measure of the distribution of particles in bulk soil less than 2 mm
in size. The relative amounts of these particles, once known, can be used to accurately
place the bulk soil on the USDA Textural Class Triangle mentioned earlier in this thesis.
The PSA was used to evaluate soil texture, a term that describes the solid phase or
mineral phase of bulk soil. It does not refer to bulk density, soil organic matter, soil
fertility, or aggegation of soils, but PSA can be used to describe typical water
conductivity through the matrix and the amount of water that might be in a profile. For
instance, a course textured soil is expected to have a high rate of water conductivity,
whereas a fine textured soil is expected to have low conductivity but relative to course

textured soils, a high affinity for water, therefore a high water content on a volume basis.

123

There are four steps using the hydrometer method (Gee and Or, 2002) for

determining PSA in this study. The work was done at MSU’s Plant and Soil Testing

Laboratory after transporting the bulk soil to Michigan from Oregon. The steps are:

1.

Grinding and sieving. At the time of soil sampling the soils were near or at FC.
Upon drying aggegation occurred and therefore required this physical step of
ginding. This step was accomplished by ginding all bulk soil through a bulk
ginder. Breaking aggegate soils into small, dispersed minerals is required in
order to differentiate particle sizes. This physical process was done using a Soil
Grinder under a hood; the ginder has two inversely rotating carbide drums that
destroy aggegation but usually do not crack or destroy soil minerals in the sand
size range. The second part of this step was sieving though a 2 mm sieve. This
step removes any particles that were larger than 2 mm in the bulk soil that could
be noted as fine gavel, coarse gavel, or cobbles. Upon completion of this step
the bulk soil resembles a dry powder with sand particles dispersed in the soil.

A 100 ml solution of 0.1 M Na-hexarnetaphosphate was added to 50 gams of
bulk soil, placed in a 500 ml bottle, secured with a lid, and placed in a rocking
table for 12-15 hours. Typically the sample bottle is prepared the previous
afternoon, and shaken overnight on the vibrating table.

Using the principal of Stokes Law (Gee and Or, 2002) larger particles of the same
density will move through a liquid faster than a smaller particle in response to
gavity. By extension, a sand particle is larger than a silt particle and these are
both larger than a clay particle. Each particle will have a specific density of

approximately 2.65 g cm'3 . The hydrometer method is begun using a 250 ml

124

gaduated cylinder placed on a flat table surface. The soil samples were removed
from the mixing table and while completely dispersed, the entire soil solution is
added to a gaduated cylinder. The sample immediately began to settle out, with
the sand particles very quickly moving to the bottom of the cylinder. A mixing
probe was immersed in the cylinder and twenty up and down movements were
made to resuspend the particles. A timer was started, and 40 seconds later a
hydrometer (AST M 152 H-Type) was carefully dropped into the liquid. A
reading was taken off the gauge scale on the neck of the hydrometer. At 7 hours
fi'om time zero the process was repeated. After this step a thermometer was
placed in the liquid and a reading was taken and recorded. The typical approach
was taking the first reading in the early morning and then the seven-hour reading
later in the afternoon. At the completion of the final reading the cylinders were

emptied and cleaned.

4.10 Laboratory work, soil water characterization

The development of the SWCC using soil cores is time consuming, demanding in
terms of laboratory monitoring, and based upon a set of assumptions that must be used in
order to bring value to a data set. The procedure is well established in soils laboratories;
the author visited five Universities and sought out assistance before setting up a
laboratory at MSU. Klute (1986) described in detail a pressure plate apparatus method,
using soil cores, ceramic plates, a pressure chamber, and a manifold system for pressure

regulation. Before describing this procedure, a list of assumptions follows:

125

1.

Soil cores are representative of field conditions, i.e. the core of approximately 70

3 is as nearly as possible the same mineralogy, compaction, and aggegation as

cm'
field soil. We know that field variation is large so this assumption can be easily
discounted. Nevertheless this method has been used by many researchers and the
same assumption is given but is problematic of using a small soil core to describe
a field specific process.

The soil in the core is always drying in response to pressure, rather than wetting in
response to infiltration and redistribution. Given the wetting and drying hysteretic
effect, the volumetric water content at a given matric potential, i.e. the end
product of the SWCC, is always based upon the wetting cycle. Generally, a soil
that is drying has slightly higher volumetric water content than a wetting soil at
the same matric potential (hysteresis). For this reason the SWCC curves is always
listed as desorption curve (drying as in this study) or a sorption curve (wetting).
The physics of air flow pressure and negative suction or the inverse of pressure is
exactly equal. This seems reasonable, given that if a soil is pressured in a tank, or
the tank is under a vacuum at some negative pressure, then the removal of water
as desorption should be exactly the same. This is a laboratory procedure and
cannot be used in the field.

The soil cores gadually lost some structural integity over the wetting and drying

cycles, especially for the final three matric potential points of 0.5, 1.0, and 1.5

MPa. This statement was based on visual observation

126

4.11 SWCC equipment

The pressure chamber used in this study was built as a replica of a 15 bar
extractor marketed by Soil Moisture Corporation, Santa Barbara, California. The
chamber is large enough to support five ceramic plates stacked vertically and separated
by 8 cm. The chamber lid was scored to fit on a rubber O-ring inset on the chamber wall.
The lid is held in place with eight gade 8 bolts. Two intake and five exhaust ports were
built into the sidewall; the exhaust ports were at the same vertical elevation as the
ceramic plates when stacked in the tank and loaded with soil cores.

A 150 psi air compressor was used to supply positive air pressure. A two stage
manifold unit was built using two pressure regulators, a combination of valves and
several pressure gauges. The first step was built to pressurize the chamber from zero to
0.3 bars. A water filled manometer was used for the first four air pressure extractions,
0.003, 0.01, 0.02, and 0.03 MPa (FC), thereby removing any gauge error at these very
low pressures. A second stage series of gauges were used for the next set of pressures:
0.04, 0.05, 0.1, 0.2, 0.3, 0.5, and 1.0 MPa. Figures 4.4 and 4.5 shows the chamber and

manifold in use in the laboratory.

   

Figure 4.4 SWCC chamber Figure 4.5 SWCC chamber and manifold

127

The ceramic plates were obtained from Soil Moisture Corporation, Santa Barbara,
California, and MSU Soils Department (courtesy Dr. Alvin Smucker, MSU Crops and
Soils Department). Plates were constructed out of ceramic material at specific bubbling
pressures at which they will conduct water. The larger the ceramic plate porosity, the
lower the bubbling pressure. The use of somewhat uniform ceramic plate bubbling
pressures facilitates more uniform flow of water from soil cores into the plate at
equilibrium, then movement out of the pressure tank at a given pressure.

The laboratory setup at MSU (Farrall Hall) was built for a maximum pressure of
150 psi, or slightly more than the 145-psi for the 1.0 MPa pressure. The final extraction
at 1.5 MPa was done by transporting all soil cores back to Oregon State University and
using a membrane extractor. This procedure will be explained following the MSU

Laboratory description.

4.12 Soil water extraction

The principle of soil water extraction using the ceramic plate and pressure
chamber method requires a hydraulic gadient between the two. The procedure therefore
requires that for every extraction, both the cerarrric plate and the soils cores are rewet to
as near saturation as possible. The soil cores were removed from their transport canisters,
unwrapped from the saran wrap, with one end of the soil core ring wrapped in
cheesecloth held with a rubber band. The cheesecloth helped stabilize the soil particles
on the core bottom thereby maintaining soil integity over the duration of the extraction
process. The cores were placed in a bath of distilled water with a 0.005 M concentration

of calcium sulfate (CaSO4) for 48 hours. The soil core is 33 mm thick with the addition

128

of another two to three mm of cheesecloth material. The soil cores were placed in large
pan type plastic containers and filled to a depth of approximately 25 mm. Special care
was taken in avoiding overtopping the soil core. The objective was allowing water intake
to occur from the bottom upwards as capillary flow. Some cores rewet sooner than
others, but 48 hours appeared to be adequate for total rewetting.

During the same 48 hour wetting period five ceramic plates and one backup plate
was immersed in distilled water. This step ensured the pores in each ceramic plate were
saturated.

The pressure chamber was loaded by first installing a set of four legs, or plastic
dowels, on the chamber floor. These legs were 15 mm high, and allowed the bottom —
most plate to rest slightly above the floor, thereby assuring the ceramic rubber bladder
enough space to slightly inflate at higher pressures. Just prior to placement in the
chamber, each soil core was weighed. This step proved to be insightful; at each pressure
interval, the wetted soil core had a saturation value, which could be compared to the
previous values. Once the plate was installed, 12 soil cores were placed on the plate
surface. Special care was taken to ensure the cheesecloth made complete contact with the
wet surface of the plate, thereby assuring a hydraulic gadient. The next plate was added
after installing four more plastic legs set in open spaces of the lower plate. Plates and
cores were added until the fifth plate had been completed and the chamber held 60 cores.
The total number of cores for extraction were 120, therefore one half of the total were
extracted at any one time.

Once the lid was in place by sequentially tightening up the bolts, the manifold

valves were adjusted admitting the desired airflow into the chamber. Generally, the soil

129

cores and plates immediately lost some water that occupied the largest pores, thereby

requiring very little airflow. The next step involved the variable of time to equilibrium,

that point where air flow at some pressure had pushed all the water out of the soil core

and ceramic plate but no more could be evacuated unless additional pressure was added.

This step was accomplished by observation. Generally, the lower pressures required less

time. The time period ranged from three days to 21 days.

When equilibrium had been reached, the chamber was depressurized. The lid was

removed, and each soil core was weighed. Soil cores were placed in the plastic tray for

rewetting, and the entire process was repeated. The sequence, in summary, was:

1.

2.

Saturation of soil cores and ceramic plates for 48 hours

Soil cores weighed at saturation just prior to chamber loading

Chamber loading of soil cores on ceramic plates: hydraulic gadient made
Pressurized for three to 21 days at some positive air flow (pressure) until
equilibrium had been reached

Soil cores removed and immediately weighed

Step one repeated and the next higher pressure used in the process

For one set of soil cores, 60 samples, five months were required for the laboratory

extraction work. The steps above required daily attention, monitoring, and safety

concerns given the higher pressures that were used further along the process.

Figure 4.6 shows the soil cores on the fifth ceramic plate just prior to placing the

lid on the pressure chamber.

130

 

Figure 4.6 Soil cores on cerarrric plate in chamber

4.13 Final extraction

The 1.5 MPa required the use of a membrane extractor. Samples were transported
to Oregon State University Soils Laboratory. The membrane process differs slightly
from the ceramic plate method in two ways. First, the original soil cores are destroyed
and remade into smaller cores approximately 10 cm deep. These cores were placed on a
carbon membrane in a pressure chamber. The lid was fitted with rubber bladder, which
at closing completely comes into contact with the top of each new soil core. The system
was pressurized at 218 psi, or 1.5 MPa, and equilibrium was reached in three to four
days. Soil core samples were removed and weighed.

The final step was converting the data into volumetric water content. This was
possible only after the cheesecloth mass values were subtracted fiom the soil cores. This
was accomplished by using four blanks, cores having no soil but having cheesecloth
attached to them. The assumption was made that the average difference between
saturation and dryness at each extraction pressure of these four blanks was the same for

cheesecloth on the 60 cores having soil in them.

131

The paired values that make up the SWCC, i.e., matric potential (in this case
assuming the inverse of positive air flow, or negative pressure) at a given volumetric
water content (mass in the soil core at saturation minus mass in the soil core at that

specific pressure) provided the SWCC over the range of soil drying.

4.14 RZWQM data requirements
The RZWQM requires four data sets. They are:

1. Weather data. Two meteorology files were built for each field.

a. The first is a .brk file, with data obtained from the individual farm weather
stations. This file was built by using daily rainfall in units of 0.01 inches
of rain per 24-hour period. The study period for each farm was the time
period these local farm WS were operable.

b. The second weather file is a .met file, with data obtained from AgiMet
WS in Corvallis and Aurora Oregon. This file was built using five sets of
data: minimum air temperature (C°), maximum air temperature (C°), wind
run (km (1"), shortwave radiation (MJ m'2 d'l), and relative hurrridity (%).
There were data for every day in the five-year study period, i.e. 1,826
days.

2. Management data. There are seven goups of data that can be used to describe
farming activities. All of these data were obtained from landowner records and
personal observation.

a. Crop selection, the name of either an annual or perennial crop type gown

onafidd

132

. Planting characteristics, a description of the crop planting date, with
descriptions of data planted, planting density, row spacing and planting
depth, harvest date, stubble height at harvest and residue left on the soil

surface, and harvest type.

. Manure applications, which include descriptions of application location

(field), source, timing, method (liquid irrigation or solid spreading), mass
per field of NH4-N, mass per field of total solids, C:N ratio of bedding,
C:N ratio of manure, and carbon content of manure volume. These data
were obtained from landowner records, and represent some of the most
important data obtained over the five-year study period.

. Irrigation applications, which include field location, type of application
(sprinkler), and total application volume depth (cm). This goup of data
included irrigation volume depth of liquid manure and irrigation volume
depths of freshwater. Data were referenced with a time and spatial stamp
assuming the volume depth was applied over the entire field at the same

uniformity.

. Fertilization, which users may place the applications of inorganic fertilizer

if they are made. For this study, no inorganic fertilizer was applied to any
of the ten fields over the five-year study period.

Pesticides, which users may populate with pesticide data. No pesticide
data were collected from landowners for this study.

. Tillage, which users can define in terms of timing of tillage before

planting (preplant) date of tillage, tillage implement type, tillage depth

133

(cm) tillage intensity (a measure of soil surface disturbance), and a
description of the tillage Operation performed (primary, compaction, or
secondary). These data were obtained from landowners and personal
observation.
3. Initial State. There are three initial state requirements that set the field parameters
so the model can begin from a reference state. They are:

a. Volumetric water content. The model must have a starting point for
adjusting the FWB based upon hydrologic parameters that change every
24-hour period. For all ten fields, the value used was the first Watermark
sensor reading in cbars, and then converted to volumetric water content
using the SWCC for each individual field at the surface layer of 30 cm.

b. Soil chemistry state. This component includes many different soil
chemistry inputs of soil chemistry speciation, salt status, ion status,
carbonates status and CO2 flux are desired model outcomes. These were
not used for this study. Two input fields were populated, however, soil pH
and soil CEC.

c. Pesticide state, which was not used in this study, so all input fields
remained empty.

4. Site description. Several components of the field site are described in this model
component. They are:

a. General information, which includes where the field was located

(spatially) and its elevation above mean sea level. A soil map unit name

134

was identified based upon a USDA Soil Survey. A number identifies the
field.

. Horizon description was an important part of this component; the soil
texture as defined by the USDA Textural Class triangle, horizon depth
(cm) particle size density (2.65 g cm”), bulk density, porosity, and
fractions of sand, silt and clay as defined by the particle size analysis. The
horizon depths were in increments of 30 cm depths to 1.2 m. The bulk
density values were obtained from the field data derived at the time of soil
core sampling. The particle fractions were derived from the soil
laboratory work done at MSU Soil and Plant Nutrient Laboratory.

Soil hydraulics. There are three levels of parameterization. They match
the objectives of this thesis. They are:

i. Default value using USDA NASIS soil default values. This was
the minimum input without any data input from the fieldwork and
laboratory work.

ii. Minimum input, using particle size analysis and bulk density data;
the model uses these data and derives slightly different soil
hydraulic control based on textural class.

iii. Full description, using incremental soil depth volumetric water
content at saturation, at 0.01 MPa, at 0.033 MPa (FC), and at 1.5
MPa (PWP). These data were derived fiom the laboratory work in

deriving the SWCC for each incremental soil depth.

135

d. Hydraulic control. This component was used to further define field
conditions, i.e. the presence of surface crusting, presence of subsurface
drainage, presence of a water table, and the presence of a bottom boundary
during internal soil water distribution. These parameters were populated if

they were a component of each field.

4.15 RZWQM data parameterization

This study examined the parameters required to accurately determine the F WB

over a long period of time, five years, in ten fields. Two model components are

particularly important to achieve this determination. They are:

1.

In the site description component, characterizing the horizons. The model is first
populated with a soil survey description of the particular site based upon field
map unit. The depths will be based on different criteria than the incremental 30
cm soil depths used when fieldwork was done. This approach reflects the least
amount of parameterization necessary to run the model. This assumes that the
user has access to soil survey data through USDA’s NASIS Soil Database, found
in eFOTG Section II on the Internet. The next level of parameterization is
shifting to the incremental description of the soil profile using the 30 cm depths
from which the particle size analysis and bulk density values were derived. All
other fields in this component remain the same. This represents the second level
of parameterization. The site descriptions for the second and third scenarios are

the same.

136

2. The soil hydraulics component, which controls the rate of water flux in the soil
profile as a function of the default or determined SWCC. The model was first
parameterized using default values, based upon the actual soil map unit
characterizations as derived by USDA’s Soil Survey. Again, this represents
minimal effort by the user for data input. The second scenario also required no
changes in this component, other than the horizons will be based upon 30 cm
incremental changes and the particle size analysis parameter based upon
laboratory values at each depth. The hydraulic control of each layer was based
upon default values as influenced by the changes from USDA’s default values, to
new empirical values based upon particle size analysis and bulk density. The
third scenario included the volumetric water content at the four potentials listed
earlier. The third scenario, therefore, represented the geatest effort required to
populate the model and the geatest deviation from established default values

found in the Soil Survey through NASIS.

4.15 RZWQM evaluation

Once the four database requirements were completed, the model provided a FWB
for each of three scenarios. The three scenarios were compared to the Watermark data
using Excel Statistical Software. The objective was finding the best fit in terms of
matching model performance with field data. This evaluation step utilized the empirical
evidence obtained in the field with model simulations over the time period when weather

stations are operating on each farm. A sensitivity analysis was done using two rainfall

137

distributions periods, two hours and 12 hours. The Nash Sutcliffe Efficiency values were
used to rank the three methods of model parameterization.
The final step suggested the input parameters required for model parameterization

so that RZWQM can be used to model the F WB on non-instrumented fields.

Note: imags in this dissertation are presented in color

138

Chapter 5. Results and Discussion

5.1 Overview
There are three parts in this chapter that describe the results and discussions:

1. Describe the AgiMet and four WS data that were used in the RZWQM
parameterization.

2. Describe the bulk density, particle size determination of bulk soils, and derivation
of the SWCC for each of the ten field soils.

3. Describe the empirical versus model results based upon RZWQM
parameterization. This third part will answer the question in hypothesis one: what
degee of parameterization is necessary to model the soil water profile in soils on
a field basis, and answer the question in hypothesis two: what are the suggested

management options offered by model results?

5.2 AgriMet and WS data

Two AgiMet WS data sets were used in building the *.met file. One set was
derived in Aurora Oregon, and the other set was derived in Corvallis Oregon. Both
facilities are owned by Oregon State University.

RZWQM requires five kinds of weather data for determining ET: minimum air
temperature, maximum air temperature both in degees Celsius, wind run in km d'l,
shortwave radiation in MJ m'2 d", and relative humidity expressed as a percentage. An
alternative to these measurements is pan evaporation, but this alternative was not used in

this study.

139

The farm based WS could not supply these measurement as they lacked the

instrumentation. AgiMet, however did supply all of these measurements in the

following format with conversions listed:

1.

Minimum air temperature as degees f (Tf), converted to degees C (T, = (Tf—
32)*5/9)

Maximum air temperature as degees f (Tf), converted to degees C (Tc = (T,—
32)*5/9)

Wind run as miles d" converted to km d" (km = miles*l.609)

Solar radiation as Langleys (d'l) converted to MJ m'2 d'1
(Langley*697.3*0.00006)

Relative humidity is expressed as percentage in both data sets

The RZWQM *.met file was built using these conversions from both Agimet

stations. Farms one, two, and four used the Aurora data set and farm three used the

Corvallis data set. The distance to Agimet WS were as follows: farm one, 33.6 km, farm

two, 6.1 km, farm three, 25.6 km, and farm four, 14.4 km.

The rainfall climate data were obtained from the farm WS. Daily data were

recorded. Two data files were built using breakpoint data in the RZWQM *.brk file

system. The farm WS data was recorded in units of one hundredths of an inch volume

depth as rainfall. The timer was based on midnight to midnight. The RZWQM *.brk file

was built using increments of storm duration; therefore, the intensity and duration of a

particular storm in hyetogaph form. The daily farm WS data were distributed into two

storm durations: two hours and 12 hours. For instance if the one days’ rainfall totaled

140

0.77 inches of rainfall, one file would have that storm duration as 0.385 inches of rain per
hour for two hours, and the other data set would have the same volume depth distributed
over 12 hours, or 0.0642 inches per hour. In both cases the uniformity of the storm was
the same over two hours or 12 hour, but the intensity during the two-hour model was six
times as intense as the 12-hour model.
The rainfall data were local data, with the farm WS located within one km of the

field sites for each farm. For each farm three climate files were created as follows:

1. *.met file based on nearness to AgiMet WS

2. *.brk2 file based upon local farm WS and a storm duration of two hours

3. *.brk12 file based upon local farm WS and a storm duration of 12 hours

5.3 Soil moist bulk density and particle size analysis

Soil bulk density was determined in the field at the time of soil coring. The soil
cores were taken at or as near as possible field capacity, thus are defined as moist bulk
density. The data are listed in Table 5.1.

RZWQM requires moist bulk density as one of its important soil physical
characteristics. The model uses 2.65 g cm'3 as particle bulk density as default. The data
are listed in Table 5.1. The PSA provides an estimation of the relative percentages of
sand silt and clay based upon a bulk sample obtained at 30 cm increments in the ten study
fields. Three replicates were analyzed, providing a mean value with standard deviation
listed in Table 5.1. Note in column one Soil Map unit name, hydrologic goup class, and
crop at the time of sampling for each of the ten instrumented fields. Bulk density units

are g cm'3.

141

Table 5.1 Field bulk soil physical analyses, NASIS default, and RZWQM default values

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Bulk Bulk Bulk
Field lD Field laboratory NASIS density density density
cm Textural Silt Clay Textural RZWQM

Field 10 depth class Sand % °/e % class Mean s.d. default
Field 14 30 Si C Loam 9 60 31 Si Loam 1.35 0.048 1.40
Amity 60 Si Clay 1 58 41 Si C Loam 1.29 0.020 1.38
HGC C 90 Si Clay 3 51 45 Si C Loam 1.28 0.023 1.38
rye/com 120 Si Clay 1 57 41 Si C Loam 1.25 0.023 1.38
Field 1-2 30 Si C Loam 9 60 31 Si Loam 1.31 0.029 1.40
Amity 60 Si Clay 4 55 41 Si C Loam 1.27 0.056 1.38
HGC C 90 Si Clay 0 55 45 Si C Loam 1.35 0.060 1.38
e/com 120 Si Clay 2 57 41 Si C Loam 1.26 0.026 1.38
Field 2-3 30 Si C Loam 17 55 28 Si Loam 1.36 0.044 1.40
Powell 8 60 Si C Loam 11 49 39 Si C Loam 1.33 0.041 1.40
HG C 90 Si C Loam 9 51 39 Si C Loam 1.24 0.010 1.40
perjrass 120 Si C Loam 11 53 35 Si C Loam 1.28 0.019 1.40
Field 2-4 30 Si Loam 17 57 25 Si Loam 1.35 0.050 1.32
Powell C 60 Si C Loam 10 53 37 Si C Loam 1.37 0.018 1.40
HG C 90 Si C Loam 8 53 39 Si C Loam 1.35 0.073 1.40
grass 120 Si C Loam 10 57 33 Si C Loam 1.36 0.049 1.40
Field 2-7 30 Si C Loam 13 59 28 Si Loam 1.31 0.016 1.40
Aloha 60 Si C Loam 7 53 39 Loam 1.24 0.070 1.40
HG C 90 Si C Loam 5 55 39 Loam 1.26 0.022 1.40
per grass 120 Si C Loam 8 57 35 Loam 1.28 0.016 1.40
Field 2.8 30 Si C Loam 12 59 29 Si Loam 1.30 0.019 1.40
Aloha 60 Si C Loam 9 55 35 Loam 1.22 0.015 1.40
HG C 90 Si C Loam 7 55 37 Loam 1.17 0.115 1.40
wrass 120 Si C Loam 8 55 37 Learn 1.20 0.018 1.40
Field 3-1 30 Si C Loam 11 61 28 Si Loam 1.37 0.022 1.40
Amity 60 Si Clay 6 52 41 Si C Loam 1.28 0.036 1.38
HG C 90 Si Clay 6 41 53 Si C Loam 1.27 0.013 1.38
‘ _Ee_r9rass 120 Si C Loam 8 57 35 Si Loam 1.27 0.030 1.38
Field 3-6 30 Si C Loam 9 59 32 Si Loam 1.35 0.035 1.40
Willamette 60 Si C Loam 10 61 29 Si C Loam 1.44 0.021 1.40
HG B 90 Si C Loam 10 53 37 Si C Loam 1.33 0.056 1.40
grass 120 Si C Loam 12 49 39 Si C Loam 1.30 0.025 1.40
Field 3-7 30 Si C Loam 7 65 28 Si Loam 1.41 0.052 1.40
Holcomb 60 Si Clay 7 49 43 Si C Loam 1.25 0.024 1.30
HG D 90 Si C Loam 10 53 37 Clay 1.22 0.041 1.40
mass 120 Si C Loam 12 61 27 Clay 1.21 0.025 1.40
Field 4-1 30 C Loam 21 47 32 Si Loam 1.31 0.019 1.42
Amity 60 C Loam 21 45 33 Si Loam 1.25 0.064 1.42
HG C 90 Si C Loam 19 45 35 Si C Loam 1.22 0.047 1.40
alfalfa 120 Si C Loam 19 45 35 Si C Loam 1.38 0.043 1.40

 

 

 

 

 

 

 

 

 

 

 

 

142

There is a strong relationship between moist bulk density and textural class. The
RZWQM default moist bulk density values are a function of textural class. If the model
is parameterized using Soil Survey map data, two approaches may be used for moist bulk
density values. For this study, one approach included only empirical field data, the use of
field derived moist bulk density and laboratory PSA using field bulk soils. The bulk
samples were taken within one meter of the Watermark array of soil moisture sensors,
therefore an attempt was made to describe local soils (spatially) rather than field soils.
The second approach used a method with minimal input: for a particular field find the
predominant soil map unit in the Soil Survey, then find the textural class in the NASIS
database. When these textural classes are used in RZWQM, each has its own moist bulk
density; in Table 9 these are listed in the right in the right hand column under the heading
default RZWQM. This second approach requires no field visits; attribute data can be
based upon the accuracy of the Soil Survey, but more importantly, negate the local effects
if a field is instrumented with some type of soil moisture sensor.

The particle size analysis (PSA) data appear in two columns. The left column is
derived using the hydrometer method in the laboratory, derived from field bulk soil
samples. Using the USDA —NRCS Soil Survey, the soil map units are selected based
upon the water measurement site within each field. These soil map units appear in Figure
15. The parameterization of RZWQM requires a robust set of soil physical data,
including PSA. Based on this study the selection of NASIS data would change the model
parameters, thereby influencing the soil hydraulic characteristics. Of the 50 field bulk
soil samples measured, 17 samples had the same textural class. Two fields, 2-3 and 2-4

had identical textural classes between field samples and NASIS. Four fields, 2-7, 2-8, 3-

143

1, and 3-7 had no similar textural classes, and two more, 1-1 and 1-2 had just one similar
textural class.

Obtaining PSA data is relatively easy. Bulk soil is taken using an open bucket
auger, which does not require the careful core sampling used in the SWCC laboratory
procedure. The bulk soil requires no in-field measurements, except for depth (30 cm in
this study). The laboratory procedure can be done in two days at relatively low cost.
Based upon this study, RZWQM will require empirical field PSA rather than using

NASIS data.

5.4 Soil water characteristic curve development

The development of the SWCC for each depth in ten fields was the major field
effort of this study. As undisturbed as possible soil samples were obtained at or near
field capacity, within one meter of the soil moisture Watermark arrays, then placed in the
soil pressure chamber for repeated wetting and dying cycles at several different pressures.

The SWCC is derived from paired data: volumetric water content at some matric
potential. The critical points are saturation, field capacity, and permanent wilting point.
Volumetric water content at saturation may be measured based upon actual measurement,
or calculated by assuming a theoretical saturation at complete filling of the void space.
Given air entrapment, the void space is never completely filled with water, so the
comparison between the two values will provide estimation between calculated and

measured volumetric water contents at saturation.

144

5.5 Volumetric water content at saturation

The volumetric water at saturation data was compared using the f statistic for
unequal variances (two-tail). This statistic showed the variances between calculated and
measured were not sigrificantly different (P < 0.05).

What is interesting about the two data sets is the calculated mean is slightly higher
than the measured mean. One explanation is the measured samples were saturated; yet
there was a geater amount of air entrapped in the soil matrix as compared to the
calculated value, which assumes a theoretical complete filling with no air present. The
calculated approach assumes a particle density of 2.65 g cm"; this value is placed in the
denominator with the moist bulk density in the numerator. The product of this division is
subtracted from one yielding pore space with the assumption pore space can be filled
with water.

The two data sets were gaphed using linear regession thus defining the

coefficient of determination, or RSQ value. The data are shown in Figure 5.1.

145

 

 

 

 

 

 

 

 

 

 

Volumetric water content at saturation
0‘55 ”"3; = 0.9809x + 0.0047" W“)!
R2 = 0.9724 ’
0.53 , e"?
Qty-a.
2".»
E 0.51 a" e
5 .. ’
8 M
g 0 49 (”g i
'3: 5
‘1 w.
A 1
0.47 xi 4'
. i
0.45 . . . . 4'
0.45 0.47 0.49 0.51 0.53 0.55
calculated

 

 

Figure 5.1 Volumetric water content (%) at saturation comparing calculated versus
measured, with linear equation and R value
Given that the two data sets were not sigrificantly different the average value of

measured volumetric water contents at saturation was used in RZWQM.

5.6 The SWCC complete curve

The SWCC is usually expressed as a log linear relationship. The matric potential
is usually on the x-axis, as a logarithmic scale, and the volumetric water content is on the
yards as a linear scale. The curve has a starting point at zero matric potential, denoting
volumetric water content at saturation, and then as volumetric water content decreases,
matric potential increases. An equation of this curve, due to its log-linear scale, cannot
have a zero on the y-axis, corresponding to volumetric water at saturation.

One approach at defining the curve is dividing it into two equations. The first and
second data pair of matric potential at 0 and 0.003 MPa can be written as a linear-linear

equation. From 0.003 to 1.5 MPa the equation can be defined as log-linear. The

146

justification for this approach is simple: the change in matric potential between these two
points is small enough that it represents a slight change in volumetric water content. One
additional rationale is the population in RZWQM. The values required for the complete
defining of the SWCC are few: volumetric water at saturation, at 0.001 MPa, at 0.0033
MPa (F C), and at 1.5 MPa (PWP). The model does not use an equation. However the
value of defining an equation is obvious: at any matric potential measurement the
volumetric water content is known. For the irrigator installing some kind of matric
potential measuring device, fitting data values to an equation is critical so that irrigation
scheduling be accomplished correctly.

The SWCC figures below (5.2-5.21) illustrate specific matric potential-volumetric
water content points at each sampling depth. The gaphs are paired for each field: the
linear x-axis for the 0 to 0.003 MPa and the log x-axis for the 0.003 to 1.5 MPa of matric
potential. Each sampling depth is shown. The volumetric water content at measured

saturation is shown in the linear-linear gaphs on the x-axis point 0.

147

 

0.53

0.54 ----___....- ---- __ _.
1

Field 1-1 SWCC
0-0.003 MPa

 

”if ‘

0.52

0.51

 

 

O .50

0.49

 

0 .48

. -. .-__ _ .11...” ..JL-

 

volumetric water content %

0.47

 

 

l
l

 

 

0.46
0

 

0.001
linear scale: matric potential MP3

I

0.002 0.003

fl

0.004

 

 

e30cm

360cm
90cm

x120cm

 

 

Figure 5.2 Field 1-1 linear scale 0-0.003 MPa matric potential—volumetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

e 30 cm
Field 1-1 SWCC 2: 60 cm
0.003 -1.5 MPa
90 cm
x 120 cm
0.55 - m:
0.50 T ‘
32 e 53 j
’E 0.45 A
8 e i
c I
3 0.40 a ’ f
t i
g 0.35 ’ 4
° {3. 3
u 0 i
g 0.25 ‘ L E
-6 . f.‘ l
’ 0.20 e “ 4
. t
0.15 . . . i
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.3 Field 1-1 log scale 0.003-1.5 MPa matric potential-volumetric water content

148

 

 

 

 

 

 

 

 

 

 

 

Field 1-2 SWCC
0-0.003 MPa . 30 cm
u60cm
0.53 .1” W.-. _~_m,,_____ _._ m" . __-_ ..__._*_-.__--.,_._..H #7 90 cm
J 3 x 120 cm
L: i
. 0.52 ‘
°\ !
‘5 l
3 4» w :
8 a
i ' l
O 1
3 0.50 i
3 !
i
E 0.49 l
2 a
9 I
0.48 . . . 1
0 0.001 0.002 0.003 0.004
linear scale: matric potential MPa

 

 

 

Figure 5.4 Field 1-2 linear scale O-0.003 MPa matric potential-volumetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 1-2 swcc t 28 2:
ii
0.003 - 1.5 MPa 90 cm
x 120 cm

0.55 ———~-- -- *————-——1

0.50 ,.
33 e
E 0.45 .
3 e
C 7‘
8 0.40 . A ..

r

g 0.35 -
o I} i
- 0.30 c a
e .
S 0.25 ”—7.; *4
0 ~"' .«
’ 0.20 c».

0.15 . . .

0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.5 Field 1-2 log scale 0.003-1.5 MPa matric potential-volumetric water content

149

 

 

 

 

 

 

 

 

 

 

Field 2-3 SWCC
0-0.003 MPa

0.55 T~-—-~ ____-_, ... H W n“..- __ 7
g 0.54 .L
E
g 0.53 ..
lg 0.52 4
3 j
u 0.51 4. I
§ e
g 0.50 i
"' 1i ‘
g 0.49 ‘ 5

0.48 . . . ’

0 0.001 0.002 0.003
linear scale: matric potential MPa

 

e300m
e60 cm

90 cm
x120 cm

 

 

 

0.004

 

Figure 5 .6 Field 2-3 linear scale 0-0.003 MPa matric potential-volumetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 2-3 swcc 9 30 cm
0.003 - 1.5 MPa ' 5° 0'“
90 cm
x 120 cm
0.60 -_--.--, .. 5...}
0.55 g:
g In J
*5 0.50 3 j
‘3 0 45 A '
g 0‘40 e I: '
9 $.21," l
0.35 "f
i
0.30 ' i.
E . 0 ‘m .1: ;
E 0.25 6 L. 1
o ;.; g
0.20 9 ¢ 5
0.15 . . . i
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

Figure 5.7 Field 2—3 log scale 0.003-1.5 MPa matric potential-volumetric water content

150

 

 

 

 

 

 

 

 

 

 

Field 2-4 SWCC

0-0.003 MPa 0 30 cm

I 60 cm

90 cm

0-50 ' *—' "'*“"‘—"“'"" ‘H” "A“ j x 120 cm
i
.\. i
a . ‘
s l: . l
g 0.49 I
o l
i
2 a 5
g 0.48 i .
E i
_ . .
2 . ,
0.47 . . . i
0 0.001 0.002 0.003 0.004
linear scale: matric potential MPa

 

 

 

Figure 5.8 Field 2-4 linear scale 0-0.003 MPa matric potential-volumetric water content

 

 

 

 

 

 

 

 

 

 

 

Field 24 swcc ° :8 cm
0.003 -1.5 MPa ' cm
90 cm
x 120 cm
0_55 _._._. .H. A.._-_._.,. .. /._., . . . ..._.._ :T",T—‘
0.50 g
32 I
*5 0.45 .
g 3 g ‘
3 0.40 _
. ..I
g 0.35 ' 3,.
° 0 30 ——— a 5?
§ ' I .
E | ._
_: 0.25 I . ’I
> :-
l
0120 I
0.15 1 . . .
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.9 Field 2-4 log scale 0.003-1.5 MPa matric potential-volumetric water content

151

 

 

 

 

 

 

 

 

 

Field 2-7 SWCC
00.003 MPa 0 30 cm
I 60 cm
90 cm
0'55 ' "W ' W" WW " W" " x 120 cm
°\° ll
§ 0.54 "
E
8
'- I
g 0.53 MVMH w__.fl.d-____:
.2
’g 1
2 0.52 r
g i
a .
‘ i
051 . . . ;
0 0.001 0.002 0.003 0.004
linear scale: matic potential MPa

 

 

 

Figure 5.10 Field 2-7 linear scale 0-0.003 MPa matric potential-volumetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 2-7 swcc 0 30 cm
0.003 - 1.5 MPa " 60 cm
90 cm
x 120 cm
0.60 —~—— — —--— _ , ——-~ 77.:
0.55 .5
3 g i
E 0.50 -— ‘ r m
= 0.45 ‘ a .
8 I I i
I- - A
3 0.40 , ‘l.
S 0.35 __~——__wu +~~~--»———_~4
.2 "f ‘I 1
g 0.30 M‘— A‘ L! E .3
E 7 I l i
% 0.25 t. ‘ T. j:
> 0.20 .. ~
0.15 . . . A
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.11 Field 2-7 log scale 0.003-1.5 MPa matric potential-volumetric water content

152

 

Field 2-8 SWCC
043.003 MPa

 

030cm

ill 60 cm
0.60 “rm—""'“"—“"'—“-“ ~-~——~~-———~-~-r~~~w 90 cm

 

 

0.59 - ._r x 120 cm

 

 

0.58
0.57
0.56 _
0.55
0.54 1 1f
0.53 i
0.52 . i

0.51 . i r .
0 0.001 0.002 0.003 0.004

linear scale: matric potential MPa

 

 

 

_- -____..JL._____i_

 

 

volumetric water content %

 

 

 

 

 

 

Figure 5.12 Field 2-8 linear scale 0-0.003 MPa matric potential-volumetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30
Field 2-8 swcc :60 g
0.003 - 1.5 MPa
90 cm
x 120 cm
0.65 —r -- ,
l
0.60 '
Bf 0.55 __
g 0.50 0 a;
C
3 0.45 i?
o L“: ,__
g 0.40 . 1 .
.1. 1%: i
.2 0.35 ’3
43 33‘: ,
E 0.30 9 .
2 9? :4;
g 0.25 3
0.20 .=
0.15 . . . i
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.13 Field 2-8 log scale 0.003-1.5 MPa matric potential-volumetric water content

153

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 3-1 SWCC . 30 cm

0-0.003 MPa 3 60 cm

90 cm

x 120 cm
0.53 7
. 5
°: 0.52 a; i
r: " 3
8 i
g 0.51 i
s , f
0 50 i
g i
.2 «.
0.49 4
g 4. i
:3 I
" 0.48 i
9 l
9 ,
0.47 . . . E
0 0.001 0.002 0.003 0.004
linear scale: matric potential MPa

 

 

 

Figure 5.14 Field 3-1 linear scale 0-0.003 MPa matric potential-volumetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 3-1 swcc ’ :3 W
0.003 -1.5 MPa ‘ cm
90 cm
x 120 cm
0.55
.50 9'4"
a: 0 ,
g 0.45 f
c
8 0.40
E 0.35 4?:
.2 0.30
‘3 " ,4. 9
g 0.25 Pa
3 .
> 0.20 ’5
0.15 . . .
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.15 Field 3-1 log scale 0.003-1.5 MPa matric potential-volumetric water content

154

 

volumetric water content %

Field 3-6 SWCC

 

 

 

 

 

 

 

 

 

 

 

 

o-o.003 MPa , 30 cm

a 60 cm

0.52 T».— -__-____-___. _______..-_-.____ '1 90 cm

, ; x 120 cm
0.51 . I
0.50 .
0.49 “ J;
l
0.48 I
0.47 '
15 i
0.46
0.45 . f
o 0.001 0.002 0.003 0.004

linear scale: matric potential MPa

 

 

 

Figure 5.16 Field 3-6 linear scale 0—0.003 MPa matric potential-volumetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 3-6 SWCC O 30 cm
0.003 -1.5 MPa 5! 60 cm
90 cm
x 120 cm
0.55 T--~~—— ——-. -....-._ --.__-_..._---_..---__---.__.
0.50 5
.\°
2 0.45 a; ’
3 0.40 ' 9
b l :2: .0
g 0.35 "1.5:.
i 9
g 0.30 9
“n 9
_ , .
g 0.25 “—3 % ’
2 ‘5‘ .9
0.20
0.15 . . .
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.17 Field 3-6 log scale 0.003-1.5 MPa matric potential-volumetric water content

155

 

030cm
:60 cm

90 cm
x 120 cm

Field 3-7 SWCC
0-0.003 MPa

 

 

0.55 ---- -,__
0.54 i' 3
0.53 ,. I
0.52 ' f
0.51 i

 

 

 

 

 

 

 

0.50 a
0.49
0.48
0.47 1k ’

0.46 I . 3 i
0 0.001 0.002 0.003 0.004

linear scale: matric potential MPa

 

volumetric water content

 

_- -.lr_._ m.:___..-

 

 

 

 

 

Figure 5.18 Field 3-7 linear scale 0-0.003 MPa matric potential-volumetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 3-7 swcc ’ 638 :"m‘
0.003 -1.5 MPa 2
90 cm
x 120 cm
0.60 l— —_-. ,___.--___ _ ,- -- ____-_3
0.55 ,
f 0.50 _. '
c . g
g 0.45 ,__‘
8 °
g 0.40 5 I" i; .
3 . .3
3 0.35 ,: H
E 0.30 9;:
E . .
3 o 25 ’ i i
o ‘ ' as . . E
> s- i
0.20 4 s:
0.15 . . . 9
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.19 Field 3-7 log scale 0.003-1.5 MPa matric potential-volumetric water content

156

 

 

Field 4.1 swcc 30

o-o.oos MPa ’ cm
a60cm
0'55 ‘“ M 3 90 cm

0.54 f Q x 120 cm

u 3.,

 

 

 

 

 

0.53

 

0.52

 

1
0.51 e

l

 

0.50

 

0.49

volumetric water content %

__\._JL. V . . -.- .

0.48

 

 

l
0.47 r . . i
0 0.001 0.002 0.003 0.004

linear scale: matric potential MPa

 

 

 

 

Figure 5.20 Field 4-1 linear scale 0-0.003 MPa matric potential-volmnetric water content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 44 swcc 9 30 cm
0.003 - 1.5 MPa I 60 cm
90 cm
x 120 cm
0.60
0\. 0.55 2::
a 0.50 . T1
3 ,, 2:" i
c 0.45 9 :g .4
8 - . ,
'- 0.40 ’ 54 g]
a
3 0.35 :T
92 .f; __
E 0.30 9 49 ,
3 0.25 ‘-" g F- j:
° 6 a
> 0.20 T:
0.15 I . . 1
0.001 0.01 0.1 1 10
log scale: matric potential MPa

 

 

 

Figure 5.21 Field 4-1 log scale 0.003-1.5 MPa matric potential-volumetric water content

157

Table 5.2 contains equations for O-0.003 MPa portion of the SWCC (linear-
linear), equations for 0.003-1.5 MPa (log-linear), and the RSQ coefficient of
determination values (R2). The RSQ values indicate how well paired data, matric
potential and volumetric water content fit the equation line. Because the x-axis data, as
matric potential, has been converted to logarithmic values, the RSQ can be interpreted as
a linear transformation. The average RSQ value is 0.98 with a standard deviation of
0.006. The logarithmic equations can therefore be used in RZW QM for determining
volumetric water content for any matric potential data at or greater than 0.003 MPa.
Because the linear equation has a RSQ value of 1 (only two data points, 0 and 0.003
MPa), this equation can be used in RZWQM for any matric potential data at O to 0.003
MPa. The final term in the linear equation is the volumetric water content at saturation

for each field at the sampled depth (y—intercept).

158

Table 5.2 Field linear and log equations for SWCC, with RSQ, by depth

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 - 0.003 MPa 0.003 MPa - 1.5 MPa

Field ID cm depth Linear equation Logarithmic equation RSQ
Field 1-1 30 y = -4.4678(x) + 0.4858 y = -0.0474Ln(x) + 0.1828 0.981
60 y = -0.9551(x) + 0.5227 y = -0.0496Ln(x) + 0.2124 0.977

90 y = -2.2501(x) + 0.5325 y = -0.0475Ln(x) + 0.223 0.972

120 y = -2.4605(x) + 0.5338 y = -0.0479Ln(x) + 0.2153 0.972

Field 1-2 30 y = -3.4803(x) + 0.512 y = -0.0491Ln(x) + 0.2082 0.980
60 y = -3.5610(x) + 0.5236 y = -0.049Ln(x) + 0.2118 0.981

90 y = -3.1242(x) + 0.4973 y = -0.0469Ln(x) + 0.2087 0.988

120 y = 4.8239(x) + 0.519 y = -0.0482Ln(x) + 0.2197 0.987

Field 2-3 30 y = -0.2972(x) + 0.5054 y = -0.0478Ln(x) + 0.2009 0.985
60 y = -0.1457(x) + 0.4929 y = —0.0444Ln(x) + 0.2355 0.985

90 y = 2.6062(x) + 0.5439 y = -0.0486Ln(x) + 0.2363 0.980

120 y = 4.5325(x) + 0.5253 y = -0.0556Ln(x) + 0.1803 0.982

Field 2-4 30 y = -1.9096(x) + 0.482 y = -0.0452Ln(x) + 0.1988 0.984
60 y = -2.2501(x) + 0.4812 y = ~0.0433Ln(x) + 0.1986 0.984

90 y = 0437100 + 0.4907 y = -0.0483Ln(x) + 0.1917 0.986

120 y = 0.0162(x) + 0.492 y = -0.0444Ln(x) + 0.2208 0.985

Field 2-7 30 y = -9.2593(x) + 0.5413 y = -0.0463Ln(x) + 0.2344 0.987
60 y = -3.9174(x) + 0.5441 y = -0.0497Ln(x) + 0.2383 0.981

90 y = -7.4139(x) + 0.5348 y = -0.0459Ln(x) + 0.233 0.975

120 y = -2.6709(x) + 0.5217 y = -0.0509Ln(x) + 0.1921 0.973

Field 2-8 30 y = -1.6835(x) + 0.5212 y = -0.0452Ln(x) + 0.2212 0.963
60 y = -1.8292(x) + 0.5418 y = —0.0507Ln(x) + 0.2254 0.980

90 y = -2.8004(x) + 0.589 y = -0.0459Ln(x) + 0.2813 0.974

120 y = -4.9210(x) + 0.556 y = -0.0478Ln(x) + 0.2433 0.974

Field 3-1 30 y = 4.0469(x) + 0.4867 y = -0.0427Ln(x) + 0.2169 0.988
60 y = -6.5883(x) + 0.5208 y = -0.0476Ln(x) + 0.2025 0.978

90 y = -7.2358(x)+ 0.5164 y = —0.0484Ln(x) + 0.1978 0.982

120 y = -5.7466(x) + 0.5181 y = -0.0489Ln(x) + 0.1922 0.977

Field 3-6 30 y = -1.9487(x) + 0.4985 y = -0.0456Ln(x) + 0.2207 0.988
60 y = -3.7393(x) + 0.465 y = -0.0434Ln(x) + 0.1937 0.987

90 y = -3.6908(x) + 0.5025 y = -0.0449Ln(x) + 0.2024 0.968

120 y = -1.7321(x) + 0.5139 y = -0.048Ln(x) + 0.2107 0.973

Field 3-7 30 y = 0.2428(x) + 0.4712 y = -0.0432Ln(x) + 0.2052 0.988
60 y = -1.4893(x) + 0.5291 y = -0.0507Ln(x) + 0.2115 0.983

90 y = -3.626(x) + 0.5394 y = -0.049Ln(x) + 0.223 0.981

120 y = -2.8328(x) + 0.5424 y = -0.0501Ln(x) + 0.2321 0.974

Field 4-1 30 y = -2.0490(x) + 0.5141 y = -0.0459Ln(x) + 0.2199 0.978
60 y = -0.4047(x) + 0.5371 y = -0.048Ln(x) + 0.2341 0.979

90 y = —0.0809(x) + 0.5422 y = -0.0461Ln(x) + 0.250 0.982

120 y = -1.5864(x) + 0.4838 y = —0.0441Ln(x) + 0.2138 0.979

 

 

The important values of the SWCC are volumetric water content at saturation, FC,

and PWP (expressed as a percentage). These values are shown in Table 5.3.

159

Table 5.3 Field saturation, FC, and PWP volumetric water % by depth

 

Saturation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Soil % FC % PWP%
Field ID digit" 0 MPa fig? '3";
Field1-1 30 0.49 0.34 0.16
60 0.52 0.38 0.19
90 0.53 0.39 0.20
120 0.53 0.38 0.20
Field 1-2 30 0.51 0.38 0.19
60 0.52 0.38 0.19
90 0.50 0.37 0.19
120 0.52 0.38 0.20
Field 2-3 30 0.51 0.36 0.18
60 0.49 0.39 0.22
90 0.54 0.40 0.22
120 0.53 0.37 0.16
Field 24 30 0.48 0.35 0.18
60 0.48 0.35 0.18
90 0.49 0.36 0.17
120 0.49 0.37 0.20
Field 2-7 30 0.54 0.39 0.22
60 0.54 0.41 0.22
90 0.53 0.39 0.21
120 0.52 0.37 0.17
Field 2-8 30 0.52 0.38 0.20
60 0.54 0.40 0.20
90 0.59 0.44 0.26
120 0.56 0.41 0.22
Field 3-1 30 0.49 0.36 0.20
60 0.52 0.36 0.13
90 0.52 0.36 0.18
120 0.52 0.36 0.17
Field3-6 30 0.50 0.38 0.20
60 0.47 0.34 0.18
90 0.50 0.36 0.18
120 0.51 0.37 0.19
Field 3-7 30 0.47 0.35 0.19
60 0.53 0.38 0.19
90 0.54 0.39 0.20
120 0.54 0.40 0.21
Field 4-1 30 0.51 0.38 0.20
60 0.54 0.40 0.21
90 0.54 0.41 0.23
120 0.48 0.36 0.20

 

160

 

The scatterplot of these data are shown in Figure 5.22. The fields shown on the x-
axis begin with field 1-1 at 30 cm depth, and sequentially match the progression of fields

and depths in Table 5 .3, ending with field 4-1, 120 cm depth.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

General plot of saturation. FC, and PWP ° saturation .
a field capaCIty FC
permanent wilting point PWP
0.70 3~-——-—-—---—--—-—3— h-.. «~--~ --- —~------—~~~-——- -- - 9.
e\°
E 0.60 . .
9 09 9 O O 9
§ 0.50 ....*.¢‘96 ...ei ... . 6...¢.¢’.. Q. 6
E 0'40 5...}: ;; a C, r, : fl r... :5 '1 .-~. _ __ ..-.| 1"} ‘. "iii—,TfJLILL-r; ... Lil . 3 _ :71). Ciffi—Li—fl ; .
g e ' .J r- “ .. " «3' 1:, ‘ '1 ~
3 0.30
.2
g 0.20 -«-——-— --—————- — ——- —— -—-—-—-——— - — ———~———-—
5 0.10
o
> 0.00 . . . . . . . .
0 5 10 15 20 25 30 35 40 45
SWCC numbers 1-40

 

 

 

Figure 5.22 Field scatterplot of volumetric water content % at saturation, FC, PWP, by
depth

The average volumetric water content at saturation for all soil depths and fields is
0.5168 with a standard deviation of 0.0252. In a 30 cm soil profile depth, this volume is
15.5 cm of water, with a standard deviation of 0.756 cm of water. The average
volumetric water content at PC is 0.3774, with a standard deviation of 0.0199. In a 30
cm soil profile depth, this volume is 11.3 cm of water, with a standard deviation of 0.6
cm of water. The difference, 4.2 cm of water, is 27.1% of saturation. This is commonly
known as gravitational water that during internal drainage is moved downward in the soil
profile. The average volumetric water content at PWP is 0.1969, with a standard

deviation of 0.0196. In a 30 cm soil profile depth, this volume is 5.9 cm of water, with a

161

standard deviation of 0.59 cm of water. The difference, 5.4 cm of water, is commonly
known as available water capacity (AWC) that is capable of meeting plant transpiration
and soil evaporation. This difference is 35% of saturation.

In summary, the field soils in this study at all soil profile sampling depths
averaged almost 52% saturation, almost 38% field capacity, and just less than 20% at
permanent wilting point. As a function of volume depth, at saturation there is 15.5 cm of
water, 11.3 cm of water at PC, and 5.4 cm of water at PWP. Gravitational water accounts
for 27.1%, AWC accounts for 35%, and the remaining 37.9% is unavailable for plant

transpiration.

5.7 Volumetric water contents: empirical field data

Given the volumetric water equations were built for all matric potential data from
saturation to PWP, the next step was defining the field water profiles at each sampling
depth based on the Watermark data. The Watermark data was compiled as matric
potentials in units of 0-200 centibars, which are converted to units of MPa. All data
points less than 0.003 MPa were converted to volumetric water content using the linear
equation, and all data points to 0.2 MPa from 0.003 MPa were converted using the
logarithmic equation.

The following figures 5.23-5.32 illustrate the volumetric water content for each of

the ten fields at each sampling depth.

162

 

field 1-1 volumetric water —-——60 cm

 

1/1/2000 7/1/2000 1/1/2001 7/1/2001 1/1/2002 7/1/2002
date

 

 

 

Figure 5.23 Field 1-1 Watermark volumetric water content % for study period, by depth

 

 

 

—30 cm
field 1-2 volumetrlc water —60 cm
—90 cm
120 cm
0.55
0.50

volumetric water content %
o p
8 8 8

     

0.20. ,- . . .
1/1/2 7/1/2000 1/1/2001 7/1/2001 1/1/2002 7/1/2002
date

 

 

Figure 5.24 Field 1-2 Watermark volumetric water content % for study period, by depth

163

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Field 2-3 volumetric water :638 gm
cm
— 120 cm
0.60
0.55 M-
* r‘r'e ‘ 1 Hr: m “‘I‘ A
E 0.50 _ I I i ‘ l” [J l w
‘3 W I ll ll
3 0.45
l l ”I i
go..- ~
0
§ 0.35 1W V“ ‘ v
E 0.30
9
0.25
0.20 I l I I I
10/2712000 2/27/2001 6/27/2001 10/27/2001 2/27/2002 6/27/2002
date

 

 

 

Figure 5.25 Field 2-3 Watermark volumetric water content % for study period, by depth

 

Field 24 volumetric water ——-60 cm

  

volumetric water content %
.0 .0
w h-
o: O

0.20 ‘ ‘ " " -‘ ‘
10/27/2000 2/27/2001 6/27/2001 10/27/2001 2/27/2002 6/27/2002
date

 

 

 

Figure 5.26 Field 2-4 Watermark volumetric water content % for study period, by depth

164

 

Field 2-7 volumetric water

 

 

 

0. 55

0.50

0.45

0.40

0.35

0.30

volumetric water content %

0.25

 

0.20
10/27/2000 2/27/2001 6/27/2001 10/27/2001 2/27/2002 6/27/2002
date

 

 

 

Figure 5.27 Field 2-7 Watermark volumetric water content % for study period, by depth

Field 2-8 volumetric water

0.60
0.55
0.50
0.45
0.40
0.35
0.30

volumetric water content %

0.25

0.20
10/27/2000 2/27/2001 6/27/2001 10/27/2001 2/27/2002 6/27/2002
date

 

Figure 5.28 Field 2-8 Watermark volumetric water content % for study period, by depth

165

 

Field 3-1 volumetric water _ 60 cm

 

p
8

p
&

volumetric water content %
_o o o
8 8.’ 8

.°
N
or

 

o
8

6/27/2000 10/27/2000 2/27/2001 6/27/2001 10I27I2001 2/27/2002 6/27/2002
date

 

 

 

Figure 5.29 Field 3-1 Watermark volumetric water content % for study period, by depth

 

Field 3-6 volumetric water —-638 crmn
m

.09
8%

'2;

volumetric water content %
.° ° .°
8 8’. 8

 

10I6l2000 2/6/2001 6/6/2001 10/6/2001 2/6/2002 6/6/2002
date

 

 

 

Figure 5.30 Field 3-6 Watermark volumetric water content % for study period, by depth

166

 

Field 3-7 volumetric water ——60 cm

 

 

 

0.60
0.55
0.50
0.45
0.40
0.35

0.30

volumetric water content %

0.25

 

0.20
10/6/2000 2/6/2001 6/6/2001 10/6/2001 2/6/2002 6/6/2002

date

 

 

 

Figure 5.31 Field 3-7 Watermark volumetric water content % for study period, by depth

 

   

Field 4-1 volumetric water

0.60

 

0.55
0.50
0.45
0.40
0.35
0.30
0.25

volumetrlc water content “I.

0.20
12/3/1998 12/3/1999 12/3/2000 12/3/2001

date

 

 

Figure 5.32 Field 4-1 Watermark volumetric water content % for study period, by depth

167

 

A general discussion of the graphs follows. The x-axis denotes the times during
which the farm WS were operable. The ”data loggers were recording Watermark
telemetry, and the rain gauges were recording rainfall. Each farm has different starting
points, and these time fi'ames are carried into RZWQM. The length of time periods
differs; farm four has the longest time frame, and farm two the shortest. The volumetric
water content on the y-axis denotes the results of matric potential data and the SWCC
equations at each depth.

The more sensitive depth is 30 cm. The flux density of water into and out of this
depth was a function of its adjacency to the soil surface, the atmospheric conditions
causing evaporation and rainfall, and plant roots, responsible for plant transpiration
during the growing season. More careful examination of each graph showed that during
winter the soil profile was wetter, and in some fields at saturation for long periods of
time. The soil water profiles were based upon landowner irrigation of freshwater and
liquid manure wastewater as well. Examination of spikes of increased volumetric water
content during the summer periods were the results of irrigation applications, however

some were due to rainfall events albeit to a lesser degree.

5.8 RZWQM scenarios

The next step was calibrating RZWQM based upon the above water profiles.
RZWQM was parameterized using the soil physical and hydraulic properties, climate
data, crop data, and management data for each farm and field. This was an extensive

requirement for modeling.

168

RZWQM uses the entire measurement depth of 120 cm. The physical and

hydraulic parameters are based upon the Watermark depths of 30, 60, 90, and 120 cm

depths. The model is also parameterized using NASIS data base default values, which

alter the depth of each horizon.

Two versions of RZWQM were run. Fitting the model curves to the empirical

(field) curves was the objective of calibration. A combination of measured (laboratory)

and default (RZWQM, NASIS) values were required to fit the curves. The first and

second version used the same parameters with the exception of the rainfall data,

explained in the *.brk file:

1.

The *.brk file used two hour rainfall data and 12 hour rainfall data. In version
one, the total amount of rainfall that fell in the 24-hour period daily period was
distributed over two hours. In version two, the same volume depth of rainfall was
distributed over 12 hours.
The *.met file used Agrimet data
The management files were based upon landowner records of irrigation and
wastewater application over the duration of the study period for each farm
The cr0p files were based upon each landowner cropping rotation, harvest cycles,
and crop yield
Three physical and hydraulic parameters were modeled:

a. Laboratory data, or full parameterization of data using bulk density, PSA,

and SWCC data (field specific). The represent actual measurements.
b. A combination of lab data and RZWQM default values. The laboratory

values for textural class at each 30 cm increment was retained, but the

169

RZW QM data replaced the laboratory for all other parameters. The

objective here was using the PSA analyses for each field site, and then

using model values, thereby avoiding the need to determine the SWCC.

The remaining *.brk, *.met, and management files remained the same.

c. NASIS data without the use of any laboratory data, using the following

parameters:

i.

ii.

iii.

iv.

Soil horizons based upon Soil Survey identification

Soil hydraulic values based on mean values if a range are given or
exact values if an exact data point is given. For instance the Ks is
given as a range, whereas the bulk density if given as a data point,
both conditional to textural class.

Soil physical values based upon default values in NASIS,
including PSA, and soil horizon depth.

The remaining *.brk, *.met, and management files remained the

$31116.

RZWQM version one and two models were built by field using the study period

for which local rainfall data was available, therefore the starting date for each farm was

different.

The data are presented in two forms: two figures illustrating each of the

volumetric soil water profiles for version one (two hours) and version two (12 hours). A

summary table is next, followed by explanatory text. The figures show four profiles:

1. Volumetric water content to 119 cm as measured Watermark data, i.e.,

measured or observed data: “Field”

170

Volumetric water content to 119 cm as modeled using full laboratory
parameterization: “Lab” '

Volumetric water content to 119 cm as modeled using PSA laboratory
data, retaining 30 cm horizon increments, and RZWQM default values for
all other physical and hydraulic parameters: “Lab RZ”

Volumetric water content to 119 cm as modeled using only NASIS Soil

Survey data and thus no laboratory values: “NASIS”

The table for each field has both attribute data and statistical data. They were:

1.

Attribute data:

. Field ID and model parameter, listing each field identification and model

scenario as Lab, RZ, and NASIS.

. Textural class, which lists the PSA measured (laboratory) abbreviation for

Silt Clay Loam (SiCL), Silt Loam (SiL), and Clay Loam (CL).
Clay content, which lists the clay content as a decimal of one (if multiplied

by 100 would be percentage of one).

. Volumetric water content that was actually measured in the Lab data, or

used as a default value in RZWQM as a function of measured PSA, or the

default value in NASIS as a function of the Soil Survey textural class.

. Rainfall in hours per day is version one at two hours and version 2 at 12

hours of volume depth distribution over time.
R2, or the coefficient of determination. This statistical function compares
the measured or observed data as the independent variable (X) and the

modeled data as the dependent variable (Y i). The RSQ function can be

171

interpreted as the variation on the modeled data that can be explained by
the variation in the measured data. Values nearer one indicate greater
similarity, whereas values nearer zero indicate a poor model fit with
measured values. The R2 formula is:

R2 = i[nmxn-(zxxzm/[(n2x2-(XX)2)(n2Y2-(z302)] "'512

. Root Mean Square Error, or RMSE, defined as the standard deviation of
the regression line. The greater this number the larger the variation of
values either greater than or less than the regression line values. This
statistic helps explain the goodness of fit, i.e., lower values show greater
model agreement with measured or observed values. If this value is
doubled, then approximately 95% of all values are found in the confidence
interval. For these model scenarios the values are cm of volumetric water
in the soil profile to 119 cm depth. The RMSE formula is:

RMSE = [l/nZ(meanYi-Xi)2]°'5

. The Mean Error, or ME, is simply the mean value of the observed or
measured data subtracted from the mean value of the modeled data. The
ME is also described as the mean bias error, or MBE, which describes
model predictions compared to observed or measured values. Positive ME
values indicate the model tends to predict higher values than observed data
and negative ME values tend to under predict model values. The two
terms are determined using the same formula:

ME or MBE = meanY-meanX

172

5.9 Field l-l

Nash Sutcliffe Coefficient, or NSE (Efficiency). The NSE is used by
modelers to explain model performance in terms of goodness of fit. The
numerator is the sum of every paired Xg-Yi squared. The denominator is
the sum of each Xi-meanX squared. This value is subtracted from unity,
or one. If the differences between the observed and simulated values are
small, the fraction will be small, and the NSE will be nearer to one,
indicating good model calibration. If the fraction value is greater than
one, then the subtraction from one will yield a negative number, thus a
poor model fit. The larger the negative number the less calibrated the
model is at simulating the volumetric water content as compared to
observed or measured values. A perfect model fit is one; the fraction is
zero, indicating no variance between observed and modeled data. The
NSE is unitless. Its value is relative to similar simulations so the model
can be calibrated with the greatest NSE. However if all values are
negative even the smallest negative value still suggests a poor model
calibration. The NSE formula is:

NSE = 1-[<z(xi-Y.)2)/2(xi-meanX)21

Model rank is simply the rank of the six scenarios based upon the NSE

value. The largest value is one; the most negative value is six.

This field was conventionally tilled with corn silage grown in rotation with annual

ryegrass as a cover crop. Both crops were harvested as dairy cattle feed (silage).

Irrigation water and manure storage water were applied using a traveler big gun sprinkler

173

system. The Watermark array of sensors, at 30, 60, 90, and 120 cm of soil profile depth
were hard wired to the datalogger at the WS at the edge of the field. This field was

located in Polk County Oregon, and had a NASIS Soil Map Unit of Amity Silt Loam.

 

- - -— Field 1-1
Field 1-1 versron 1, 2 Hour Lab 1

 

—Lab RZ
—NASIS

 

 

cm water in 119 cm soil profile

 

1/1/00 5/1/00 9/1/00 1/1/01 5/1/01 9/1/01 1/1/02 5/1/02
time

 

 

 

Figure 5.33 Field 1-1 two hour empirical versus simulation volumetric water content

174

 

- _ - ——Field 14
Field 1 1 versron 2, 12 Hour —Lab

Lab RZ
— NASIS

 

 

 

 

cm water in 119 cm soil profile
8 8 8 8 8 a 8 8 8 8

 

1/1/00 5/1/00 9/1/00 1/1/01 5/1/01 9/1/01 1/1/02 5/1/02

time

 

 

 

Figure 5.34 Field 1-1 12 hour empirical versus simulation volumetric water content

Table 5.4 Field 1-1 RZWQM simulation comparison and NSE model rank

 

 

 

 

 

 

 

Cla Volumetric rain
Field 1-1 Tlilm'a cont); water FC hours R2 RMS ME NSE mwe'
ass 0 .1 E cm rank
nt A: % (1
Lab SiCL 0.31 0.34 2 0.64 4.48 -1.45 0.59 3
Lab RZ SiCL 0.35 0.34 2 0.63 4.37 -0.20 0.61 1
NASIS SiL 0.15 0.37 2 0.61 13.98 13.26 -2.96 6
Lab SiCL 0.31 0.34 12 0.62 4.37 -0.45 0.61 2
Lab RZ SiCL 0.35 0.34 12 0.61 4.56 0.82 0.58 4
NASIS SiL 0.15 0.37 12 0.60 12.29 12.20 -2.42 5

 

 

 

 

 

 

 

 

 

 

 

 

The best model scenario were the two-hour lab PSA increment horizon with
RZWQM default physical and hydraulic data, however the second ranked scenario had
the same NSE but greater RMSE and ME. For scenario rank one, the FC volumetric
water content was 48.4 cm of water; the RMSE value of 4.37 is 9.03% of FC. For both

versions of rainfall time periods, the rankings were lowest. The negative NSE values

175

indicate poor model performance. One explanation was decreased clay content as the
NASIS Soil Map Unit, as a Silt Loam, had only 43% of the clay as the Silt Clay Loam
soil (15% versus 35%). These data bear out the importance of using PSA. The increased
clay content provides a greater water holding capacity as shown by the volumetric water
content at FC, and the model performance is enhanced as shown by the smaller negative
ME values. Model performance was not influenced by two versions of rainfall
distribution.

The soil water profile, for local PSA scenarios, shows greater water content for
the first eight or nine months, then for the end of the growing season and early winter an
understated water content. The model simulations appear to closely resemble the
Watermark data during the second year, but for the winter period tend to underestimate
the water content in the profile. The default saturated hydraulic values were used in all
scenarios: 1.5 cm h'1 for the surface layer, and 0.68 cm h'1 for the lower depths (< 30 cm).
These values negated any surface runoff. This rationale was justified as the landowner
reported and the researcher observed no surface runoff of any volume of irrigated water
on the field at or immediately after an irrigation event. No field Ks data were obtained;

this is one area of refinement that may improve the precision of the model.

5.10 Field 1-2

Field 1-2 was a field adjacent to field 1-1. The cropping and management
program were the same. The Watermark array of sensors was hard wired to the same
datalogger as field 1-1. Local rainfall data were the same for both fields. Like field 1-1,

this field had a NASIS Soil Map Unit of Amity Silt Loam.

176

 

. — Field 1-2
Field 1 2 version 1, 2 Hour —Lab

——LabRZ
—NAS|S

 

 

 

  

cm water in 119 cm soil profile

 

20 ‘ .-'. . . f " g ‘ _‘£:‘r___.
1l1l00 5/1/00 9/1/00 1/1/01 5/1/01 9/1/01 1/1/02 51102
time

 

 

 

Figure 5.35 Field 1-2 two hour empirical versus simulation volumetric water content

 

Fi ld1-2 ' 2.12 h —Fie|d 1-2
O VOI'SIOI'I our ———Lab

CD
0|

cm water in 119 cm soil profile
8 8 8 8 8 8 8

    

1/1/00 5/1/00 9/1/00 1/1/01 5/1/01 1/1/02 5/1/02
time

 

 

 

Figure 5.36 Field 1-2 12 hour empirical versus simulation volumetric water content

177

Table 5.5 Field 1-2 RZWQM simulation comparison and NSE model rank

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Cla Volumetric rain
Frelzd 1' Tzlxtural contgnt water FC hours R2 RMSE ME NSE model
ass % % d" cm rank
Lab SiCL 0.31 0.38 2 0.63 4.05 -1.45 0.57 4
Lab RZ SiCL 0.35 0.34 2 0.63 3.90 -0.84 0.61 3
NASIS SiL 0.15 0.37 2 0.59 13.42 12.83 -3.68 6
Lab SiCL 0.31 0.38 12 0.64 3.77 -O.54 0.63 2
Lab RZ SiCL 0.35 0.34 12 0.64 3.73 0.06 0.64 1
NASIS SiL 0.15 0.37 12 0.62 12.49 1 1.89 -3.05 5

 

Given the similarities between fields 1-1 and 1-2 the NSE would be expected to

be nearly the same. Scenarios for field 1-2 showed the best model ranking for the 12-

hour rainfall distribution, however the NSE values are only slightly better than the two-

 

hour distribution. The highest NSE, at 0.64 for the lab PSA and RZ default values,
compares to 0.612 for field 1-1. For both fields, the number one rank was PSA Lab data
and RZ default values. For scenario rank one, the FC volumetric water content is 48.4
cm of water; the RMSE value of 3.73 is 7.71% of FC. The NASIS scenarios, like field 1-
1, had negative NSE values indicating poor model performance. Again, the explanation
might be the increased clay content in the PSA Lab data, as Silt Clay Loams, suggesting
the importance of local field data rather than using a NASIS Soil Map Unit. The large
negative ME values indicated that overall, the NASIS scenarios under predicted water
content by as much as 12 to 13 cm of water over the model period. The highest ranked
scenario had a MB of 0.06, indicating a slight over prediction of water content in the
profile. The RMSE values were slightly less than in this field as compared to field 1-1,
indicating slightly better model performance.

Model scenarios did not appear to fit Watermark data, as well, for the second year

in contrast to field 1-1. The model more accurately predicted the water content during

178

the second winter; unlike the same period in field 1-1, the Watermark data do not indicate
the field is at saturation. The same default values of Ks were used in both fields on this

farm. No observed runoff occurred on this field; the slope gradient is < 0.2%.

5.11 Field 2-3

Field 2-3 was a perennial ryegrass field that is used for grass silage production in
spring and then grazed by dairy cattle during the summer and fall. The field was not
conventionally tilled. The Watermark array was hard wired to a datalogger that was built
on the WS frame. Local rainfall data were obtained from this WS. This farm was

located in Clackamas County Oregon and had a NASIS Soil Map Unit of Powell Silt

 

 

 

Loam.
Field 2-3 version 1, 2 Hour :3?!" 2'3

—- Lab RZ
—- NASIS

2 ' . .

c

2

c

3

a

I-

v-

.5

3

E

0

10/27/00 1/27/01 4/27/01 7/27/01 10/27/01 1/27/02 4/27/02
time

 

 

 

Figure 5.37 Field 2-3 two hour empirical versus simulation volumetric water content

179

Field 2-3 version 2, 12 Hour

65
60
55
50
45
40
35
30
25

20
10/27/00 1/27/01 4/27/01 7/27/01 10/27/01 1/27/02 4/27/02

time

2
i:
O
I-
Q
'6
a
E
0
a:
1-
1-
.E
h
2
I!
3
E
0

 

Figure 5.38 Field 2-3 12 hour empirical versus simulation volumetric water content

Table 5.6 Field 2-3 RZWQM simulation comparison and NSE model rank

 

 

 

 

 

 

 

. Cla Volumetric rain
FBI: 2' Telxtural conteynt water FC hours R2 RMSE ME NSE model
c ass % % d" cm rank
Lab SiCL 0.28 0.36 2 0.69 7.35 —6.29 -0.23 4
Lab RZ SiCL 0.35 0.34 2 0.64 5.80 -4.44 0.64 1
NASIS SiL 0.15 0.37 2 0.67 11.34 10.66 -1.93 6
Lab SiCL 0.28 0.36 12 0.66 6.53 -5.16 0.03 3
Lab RZ SiCL 0.35 0.34 12 0.68 4.93 -2.99 0.45 2
NASIS SiL 0.15 0.37 12 0.64 10.24 -9.40 -1.39 5

 

 

 

 

 

 

 

 

 

 

 

 

The highest ranked scenario was the lab PSA data with RZ default data; the NSE
value of 0.64 for the two-hour rainfall distribution exceeds the second ranked scenario
using the same attributes but with rainfall distributed over 12 hours. All scenarios had
under predicted water contents in the 119 cm soil profile as shown in the ME data. The

mean value of the highest ranked scenario was 4.44 cm of water less than the mean value

180

of the Watermark data. The second ranked scenario had a smaller MB of 2.99 cm. As
with both fields 1-1 and 1-2, the NASIS data was ranked at the lowest rankings with NSE
values that are negative indicating poor model performance. Similarly, the importance of
the PSA were found in these scenarios, as the NASIS Soil Map Unit, a Silt Loam with a
clay content of 15%, has a lower water holding capacity than the measured PSA, a Silt
Clay Loam with a clay content of 28%, an increase of 87%. However the volumetric
water content at PC between the lab data and the NASIS data was similar even though the
textural classes are different. The importance of the volumetric water content at PC was
shown by the poorer model performance of the lab data. The lower water content at F C
shifted the water content profile lower in the figure, thereby decreasing the ME and NSE
and increasing the RMSE. This was not the case in fields 1-1 and 1-2; the lab and RZ
water contents at PC were closer in agreement.

For both versions, the data appear to more correctly predict the water content
during the growing season. There was only one full growing season for the modeled time
period but the beginning of the second growing season was reflective of the first. One
improvement was found in the second winter season period. The model did not under
predict the water content as much as the first winter period. The Ks for this field were the
same default values used throughout the modeling exercise. Field 2-3 was a B slope. At
a rolling topographical slope of 2-6%, the landowner was concerned with runoff events.
The risk was decreased by applying smaller irrigation volumes more often. The
landowner mitigated water erosion by keeping this field in permanent vegetation;
however, the Ks values may overstate hydraulic conductivity given what the landowner

calls animal compaction. Decreasing Ks would increase the profile water content for

181

much of the model period, thereby shifting the water profile upwards, reducing the
negative ME, decreasing the RMSE, and possibly increasing the NSE. Increasing Ks

would not increase the water content at FC.

5.12 Field 2-4

Field 2-4 was adjacent to field 2-4. This field was in conventional tillage for
many years, a corn silage (summer) and annual ryegrass rotation (winter). The
landowner, however, sought to mitigate water erosion in this C slope field (6-12%) by
planting permanent vegetation. For the modeled time period the field was in perennial
grass. The field was harvested as grass silage during the early part of the growing season
and then grazed by dairy cattle for the rest of the year. The Watermark array was
hardwired to the datalogger on the WS. Rainfall data were obtained from this local WS.
This farm was located in Clackamas County and had a NASIS Soil Map Unit of Powell

Silt Loam.

182

 

Field 24 version 1, 2 Hour —Elaet:d 2'4

   
   

cm water in 119 cm soil profile
8

20 . ,. . . .. ~ ~ .. .. , . W W“,
10/27/00 1/27/01 4/27/01 7/27/01 10/27I01 1/27/02 4/27/02

 

 

 

Figure 5.39 Field 2-4 two hour empirical versus simulation volumetric water content

 

Field 24 version 2. 12 hour __

 

 

 

cm water in 119 cm soil profile
8

    

20 ‘ ' ' -
10/27/00 1/27/01 4/27/01 7/27/01 10/27/01 1/27/02 4/27/02

time

 

 

 

Figure 5.40 Field 2-4 12 hour empirical versus simulation volumetric water content

183

Table 5.7 Field 2-4 RZWQM simulation comparison and NSE model rank

 

Clay

Volumetric

rain

 

 

 

 

 

 

FBI: 2' Tagggal cars/tent was/E FC hogs R2 RESE ME NSE '22:?
Lab SiL 0.25 0.35 2 0.56 11.67 11.12 -6.85 6
Lab RZ SiL 0.25 0.29 2 0.58 11.56 10.51 —6.70 5
NASIS SiL 0.15 0.37 2 0.57 10.91 -9.89 -5.86 4
Lab SiL 0.25 0.35 12 0.58 10.66 10.05 -5.55 2
Lab RZ SiL 0.25 0.29 12 0.59 10.67 -9.52 -5.56 3
NASIS SiL 0.15 0.37 12 0.59 10.04 -8.91 -4.81 1

 

 

 

 

 

 

 

 

 

 

 

 

Model performance in all scenarios was poor. The relatively high NSE values
exceeded NSE in all other fields. In contrast to the above listed fields, this PSA analysis
and the NASIS data both place this field as a Silt Loam. However the clay contents were
different. The measured PSA lab content were 25% and the value used in the NASIS
database is 15%. The increased water holding capacity in the Lab and Lab RZ scenarios
did not improve model performance. The relatively high negative ME values indicated
under prediction of water content in the soil profile. The RMSE for the Lab and Lab RZ
scenarios were all greater than the first three field RMSE’s.

This may be a field where the bulk soil surrounding the Watermark array and that
sampled for use in the laboratory (PSA) had larger physical differences than the other
fields. The bulk soil was collected approximately one meter from the array (in all fields),
but this may be the one where the array soils do not match the bulk soils taken to the lab.
The profile trend lines do track reasonably well as indicated by the R2 values of 0.57 to
0.58. The second winter period may be slightly better in terms of matching modeled data
with Watermark data, but the very end of the profile, i.e. June and July, show a steep

decrease in water content when compared to Watermark data.

184

If there was one field in the study that would discourage a modeling effort, this is

5.13 Field 2-7

Field 2-7 was the third of four fields on farm two. This field was perennial grass,
and resembles field 2-3 in terms of cropping and management practices. This was one
two most productive fields on this farm. The watermark array was wired to the
datalogger as part of the WS. Rainfall data were obtained from this local WS. This farm

was located in Clackamas County and had a NASIS Soil Map Unit of Aloha Silt Loam.

 

— Field 2-7
Lab

Field 2-7 version 1, 2 Hour

 

 

 

cm water in 119 cm soil profile

25

20
10/27/2000 2/ 27/ 2001 6/ 27/ 2001 1 0/27/ 2001 2/27/2002 6/27/2002

time

 

 

 

 

Figure 5.41 Field 2-7 two hour empirical versus simulation volumetric water content

185

 

cm water in 119 cm soil profile

 

65
60
55
50
45
40
35
30
25

20
10/27/00

1/27/01

Field 2-7 version 2, 12 Hour

4/2 7/01

7/2 7/01

time

10/27/01

 

—— Field 2-7
—— Lab 2 data
— Lab R2 2
— NASIS 2

 

 

1/27/02

 

4/27/02

 

 

Figure 5.42 Field 2-7 12 hour empirical versus simulation volumetric water content

Table 5.8 Field 2-7 RZWQM simulation comparison and NSE model rank

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Cla Volumetric rain
F'e'7d 2' Tzlxtural conteynt water FC hours R2 RMSE ME NSE model
ass % % d" cm rank
Lab SiCL 0.28 0.39 2 0.71 6.11 4.84 0.23 3
Lab RZ SiCL 0.35 0.34 2 0.71 6.62 -5.42 0.09 4
NASIS SiL 0.15 0.37 2 0.60 16.97 16.39 -4.96 6
Lab SiCL 0.28 0.39 12 0.71 5.13 -3.27 0.46 1
Lab RZ SiCL 0.35 0.34 12 0.70 5.52 —3.79 0.37 2
NASIS SiL 0.15 0.37 12 0.59 15.51 14.85 -3.98 5

 

 

 

 

Two 12-hour scenarios ranked one and two with positive NSE values: 0.46 and

0.37. The two scenarios had similar ME, RMSE, and R2 values. Both ME values

indicated the model under predicts water content in the soil profile; the figures illustrate

the typical curve at a lower volumetric water content than the Watermark data. The

modeled scenarios fit the growing season portion of the model period better than the non-

186

growing season. The second year modeled scenarios during the fall and winter period,
November through March, predict the Watermark data more accurately, although at the
very end of the curve, in June, the two curves depart from one another.

As with the previous fields the NASIS data performed poorly. The NSE values
are negative. The modeled scenarios under predict the Watermark data by 14.85 and
16.39 cm of water; the RMSE of the NASIS data are 15.51 and 16.97 cm of water. One
explanation was the difference between textural class, influenced by clay content of 28%
for the Lab PSA soil and 15% for the Silt Loam soil in the NASIS database. The large
negative ME values can be attributed to the lower volumetric water content at FC, which

is a function of clay content and available water holding capacity.

5.14 Field 2-8

This was the fourth of four fields on farm two. This was the second of two most
productive fields on the farm. The management and cropping program was similar to
field 2-3 and 2-7. The watermark array was wired to the datalogger as part of the WS.
Rainfall data were obtained from this local WS. This farm was located in Clackamas

County and had a NASIS Soil Map Unit of Aloha Silt Loam.

187

Field 2-8 version 1, 2 Hour

70
65
60
55
50
45
40
35
30
25

20
10/27/2000 2/27/ 2001 6/ 27/ 2001 1 0/ 27/ 2001 2/27/2002 6/27/2002

time

cm water in 119 cm soil profile

 

Figure 5.43 Field 2-8 two hour empirical versus simulation volumetric water content

Field 2-8 version 1, 12 Hour

70
65
60
55
50
45

40
35
30
25

20
10/27/2000 2/27/ 2001 6/ 27/ 2001 1 O/ 27/ 2001 2/27/2002 6/27/2002

time

cm water in 119 cm soil profile

 

Figure 5.44 Field 2-8 12 hour empirical versus simulation volumetric water content

188

Table 5.9 Field 2-8 RZWQM simulation comparison and NSE model rank

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Cla Volumetric rain
F'elad 2' Tzlxtural conteynt water FC hours R2 RMSE ME NSE model
ass % % d" cm rank
Lab SiCL 0.29 0.38 2 0.55 6.15 -3.47 0.33 4
Lab RZ SiCL 0.35 0.34 2 0.55 5.71 -2.44 0.42 2
NASIS SiL 0.15 0.37 2 0.54 17.50 16.72 -4.42 6
Lab SiCL 0.29 0.38 12 0.52 5.76 -2.35 0.41 3
Lab RZ SiCL 0.35 0.34 12 0.52 5.52 -1.30 0.46 1
NASIS SiL 0.15 0.37 12 0.50 16.50 15.60 -3.81 5

 

The Lab PSA and RZ default scenario values ranked highest for field 2-8. As

with previous fields with the exception of field 2-4, the Lab PSA placed the sampled bulk

soil in a different textural class than the USDA Soil Survey. The clay content in the PSA

data was almost twice the content than the NASIS data. The similarity between the

volumetric water content at PC is striking; the Lab data is a measured value and the

RZW QM data is a default value as a function of identifying textural class. Model

performance confirmed this positive relationship. The third and fourth ranked fields had

slightly lower NSE values. The negative ME values indicate the model does under

predict the water content by 1.3 cm in the first ranked scenario to 3.47 cm in the fourth

ranked scenario. Respectively, these are 2.7% and 7.0% of F C, assuming a field capacity

of 0.41 cm of water per cm of soil, or 48.4 cm of water at FC. The NASIS data reflect

every previous field; the difference in clay content reduced the volumetric water content,

thereby lowering the available water holding capacity and increasing the negative ME

values. The under prediction of NASIS scenarios was 15.6 to 16.72 cm of water, and the

RMSE values for these two poor performing scenarios was slightly great than the mean

error values.

189

 

The two hour and 12 hours figures almost rrrirror each other. No precision in
model performance was gained by changing the rainfall distribution. However one trend
reflected in previous fields occurs in field 2-8. The second year of modeled data more
accurately reflected Watermark data, although in both versions the final two months, May
and June, show modeled data containing more water than Watermark data. This trend

was found during the first growing year as well.

5.15 Field 3-1

Field 3-1 was the first of three fields on farm three. This field was perennial grass
and has been for over two decades. The field was harvested for grass silage in early
spring and then grazed by dairy cattle for the rest of the growing season. Fertility in this
field was optimal. The landowner applied adequate water and liquid manure for ET,
therefore there were no limitations for this cool season grass. The Watermark array was
wired to the datalogger as part of the local WS located adjacent to this field. Rainfall
data were obtained from this local WS. This farm was located in Marion County and had

a NASIS Soil Map Unit of Amity Silt Loam.

190

1‘

 

 

— Field 3-1
—— Lab

Field 3-1 version 1, 2 Hour

cm water In 119 cm soil profile

 

6127100 10127100 2127/01 6127101 10127101 2127/02 6127/02

time

 

Figure 5.45 Field 3-1 two hour empirical versus simulation volumetric water content

 

cm water in 119 cm soil profile

 

Field 3-1
.ab

 

.ab RZ
VASIS

  

25 . .
20 1 .. _. . , . . .. . . ~ , ~( man-.121... .44....
61 27100 9127100 1 2127100 3127101 61 27101 9127101 12127101 3127102 6127102

time

 

Figure 5.46 Field 3-1 12 hour empirical versus simulation volumetric water content

191

 

 

Table 5.10 Field 3-1 RZWQM simulation comparison and NSE model rank

 

Clay

Volumetric

rain

 

 

 

 

 

 

“9'1" 3' Teftu'a' content water FC hours R2 RMSE ME NSE m°de'

c ass % % d" cm rank
Lab SiCL 0.28 0.36 2 0.39 8.48 -5.81 -O.16 2
Lab RZ SiCL 0.35 0.34 2 0.17 14.57 12.66 -2.41 5
NASIS SiL 0.15 0.37 2 0.38 14.79 13.42 -2.52 6
Lab SiCL 0.28 0.36 12 0.38 8.04 -5.12 -0.04 1
Lab RZ SiCL 0.35 0.34 12 0.16 13.85 11.80 -2.09 3
NASIS SiL 0.15 0.37 12 0.37 14.06 12.58 -2.18 4

 

 

 

 

 

 

 

 

 

 

 

 

The NSE values for Field 3-1 were negative indicating all model scenarios
performed poorly. These data show a departure from previous fields, in that the Lab data
had a slightly better NSE value than the Lab-RZ data and NASIS data. The ME values
for all scenarios were negative, thereby under predicting volumetric water content in the
soil profile by a small as 5.12 cm to as much as 13.42 cm. The RMSE values range from
a low of 8.04 cm to a high of 14.79 cm. The volumetric water content at PC was 45.8 cm
using Lab data.

The figures illustrate especially poor model performance during the winter period
in both versions for both years. For the months of November through March, the
Watermark data determined the volumetric water content at just slightly over 60 cm of
water, whereas the Lab are as much as 10 cm less. The second winter period show
slightly better model performance. The model Lab scenarios did perform well during the
first complete growing year during the months of July through September. One other
interesting feature was the wide variation of water profiles between the three scenarios;
this has not been a component of the previous fields. The Watermark data showed a
drying period during the fall of the second growing period (October 2001). The lab data

over predicted the water content in the profile. Taken with the winter periods, in this

192

 

field the model scenarios were less responsive to water content changes than the

Watermark data.

5.16 Field 3-6

Field 3-6 was the second of three fields on farm three. This field was perennial
grass with the same cropping and management program as field 3-1. The watermark
array was wired to the datalogger as part of the WS. Rainfall data were obtained from
this local WS. This farm was located in Marion County and had a NASIS Soil Map Unit

of Willamette Silt Loam.

Field 3-6 version 1, 2 Hour

65
60
55
50
45
40
35
30
25

20
1016100 116/01 416/01 716101 1016101 116/02 416/02 716/02

time

2
fl:
2
D.
'5
(0
E
O
O)
\-
1-
.E
a“:
16
B
E
0

 

Figure 5.47 Field 3-6 two hour empirical versus simulation volumetric water content

193

 

cm water in 119 cm soil profile

 

65
60
55
50
45
40
35
30
25

20
1 01 6100

116/01

Field 3-6 version 2, 12 Hour

416/01

716/01

time

 

—— Field 3-6
—— Lab

—— Lab RZ
—NASlS

 

 

1016101

116102

 

416/02

716/02

 

 

Figure 5.48 Field 3-6 12 hour empirical versus simulation volumetric water content

Table 5.11 Field 3-6 RZWQM simulation comparison and NSE model rank

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Clay Volumetric rain
F'eld 3' Textural content water FC hours R2 RMSE ME NSE model
6 class 0 o .1 cm rank
A A: d
Lab SiCL 0.32 0.38 2 0.22 8.46 1 :45 0.19 1
Lab RZ SiCL 0.35 0.34 2 0.18 8.48 0.21 0.18 3
NASIS SiL 0.15 0.37 2 0.35 9.58 5.64 —0.04 6
Lab SiCL 0.32 0.38 12 0.19 8.51 0.81 0.18 4
Lab RZ SiCL 0.35 0.34 12 0.19 8.38 0.20 0.18 2
NASIS SiL 0.15 0.37 12 0.34 9.21 4.84 0.04 5

 

 

 

Field 3-6 had many of the same attributes as previous fields, specifically the

difference in textural class between measured Lab PSA and NASIS data. The increased

clay content in the Lab PSA data increases the water holding capacity of the soil, thereby

increasing model performance. What was different was the positive ME values for the

194

two rainfall distributions using the Lab-RZ scenarios. A positive mean error indicated
the modeled data values have a mean slightly greater than the measured Watermark data.
One explanation was the greater clay content of the Lab soils. The volumetric water
holding capacity at F C for the Lab data, however, was less than several previous fields
that had lower clay contents. The RMSE errors were, in all scenarios, greater than 8 cm
of water, which are greater than 16.5% of total water content at FC, 48.4 cm.

A similar pattern as previous fields was found in previous figures; the second year
of modeled data during the winter period fit the Watermark data. However the modeled
scenarios and Watermark data showed an inverse relationship during May through July,
2002. For the model period, the water content was less during the growing the season
and greater during the non growing season for all scenarios, indicating the model was less
sensitive to water content changes as measured by the Watermark data. The NASIS data
provided the poorest model performance in both rainfall distributions, ranking five and
six.

5.17 Field 3-7

Field 3-7 was the third of three fields on farm three. This field was perennial
grass with the same cropping and management program as field 3-1. The watermark
array was wired to the datalogger as part of the WS. Rainfall data were obtained from
this local WS. This farm was located in Marion County and had a NASIS Soil Map Unit

of Holcomb Silt Loam.

195

 

 

 

Field 3-7 version 1. 2 Hour : :35” 3'7
65 —
'9 60
G
2
g 55
8 50
E 45
5‘: 4o
5 35
g 30
5 25
20
1016100 116101 416101 716101 1016101 116102 416102 716102
time

 

 

 

Figure 5.49 Field 3-7 two hour empirical versus simulation volumetric water content

 

Field 3-7 version 2, 12 Hour :11?" 3'7

cm water In 119 cm soil profile

 

1016100 116/01 416/01 716101 1016101 116102 416102 716102
time

 

 

 

Figure 5.50 Field 3-7 12 hour empirical versus simulation volumetric water content

196

Table 5.12 Field 3-7 RZWQM simulation comparison and NSE model rank

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Cla Volumetric rain
Rel; 3' Tzlxtural conteynt water F C hours R2 RMSE ME NSE model
ass % % d“ cm rank
Lab SiCL 0.28 0.35 2 0.07 9.26 -6.86 -1.53 4
Lab RZ SiCL 0.35 0.34 2 0.07 8.50 -5.67 -1.12 2
NASIS SiL 0.15 0.37 2 0.03 13.97 12.40 -4.71 6
Lab SiCL 0.28 0.35 12 0.08 8.65 -6.13 -1.19 3
Lab RZ SiCL 0.35 0.34 12 0.08 7.92 -4.95 -0.89 1
NASIS SiL 0.15 0.37 12 0.05 13.19 11.66 -4.09 5

 

 

Field 3-7 was the second of two fields that had all NSE scenarios with negative
NSE values, indicating a poor model fit with Watermark data. All ME values were
negative; the model scenarios under predicted the Watermark data from a 4.95 cm to 12.4
cm of water, or 10.2% and 25.6% of the volumetric water content at FC, 48.4 cm. RMSE
values were slightly higher than Mean Error values. One difference in Field 3-7 data was
the relatively low R2 values. These values are the lowest of all field scenarios. One
explanation is evident in the figures. For much of the entire model period the field is at
saturation, at about 62 cm of water, or 52% volumetric water content. The other fields on
this farm did not show nearly as long a period of wetness. From field observation these
data are accurate; this is a HGC D soil without subsurface drainage. This is the only field
of the modeled fields with a group D soil, and the Watermark data captured the lack of
drainage. The ME values should have been in the positive range given higher clay
content than previous fields if the logic of increasing clay content increases the water
holding capacity. The Lab data, however, showed a volumetric water content of only
35% at FC, which does not match that logic. The difference between 35% in the Lab and
40.67% as a default value in RZ was great enough that the R2 values were lowest for this

field. The Lab-RZ scenarios should have performed better as they did use higher water

197

content at FC. This soil however, likely had more water at PC than the listed default
value of 40.67% that would suggest model performance could be improved if clay
content were included in the model parameters. The Lab-RZ scenarios were based on
PSA data as measured in the lab, but the water content at F C was a default value based on
textural class description. For field 3-7 the shift towards more clay content in the Lab
PSA data was not captured in the water content FC value, and this caused poor model

performance for all scenarios.

5.18 Field 4-1

This was the last of ten modeled fields. This field had the longest model time
period, about three and a half years of Watermark data. On this farm the first WS was
installed and tested before the remaining three were installed. Field 4-1 was the only
alfalfa field in the study sites. It was the only field that had subsurface drainage
(patterned tile on 12 m intervals). The landowner had used the Watermark telemetry for
almost eight years as a tool for irrigation water management, and received a Water
Quality and Quantity Award fi'om the Oregon Department of Environmental Quality in
2000. The Watermark array was hardwired to the WS, and the rainfall data for this farm

were obtained from this WS. This farm was located in Marion County and had a NASIS

198

Soil Map Unit of Amity Silt Loam. This was the fourth of four fields with Amity SiL.

 

—— F' I 4-1
Field 4-1 version 1. 2 Hour _ Aid

 

 

.ab RZ
VASIS

 

   

 

cm water in 119 cm soil profile

 

w _ . 4 " r ‘ , . v . ‘ ' ‘ .
1213198 613/99 1213199 613100 1213100 613101 1213101 613102

time

 

 

 

Figure 5.51 Field 4-1 two hour empirical versus simulation volumetric water content

 

— Field 4-1
— Lab

Field 4-1 version 2, 12 Hour

cm water in 119 cm soil profile

 

1213198 613199 1213199 613100 1213100 613/01 1213101 613102

time

 

 

 

Figure 5.52 Field 4-1 12 hour empirical versus simulation volumetric water content

199

Table 5.13 Field 4-1 RZWQM simulation comparison and NSE model rank

 

 

 

 

 

 

 

. Cla Volumetric r in
F'e'1d 4' Textural conteynt water FC hoaurs R2 RMSE ME NSE model
class % % dn cm rank
Lab CL 0.32 0.38 2 0.45 4.85 1.66 0.38 2
Lab RZ CL 0.35 0.31 2 0.39 6.41 4.22 -0.09 5
NASIS SiL 0.15 0.37 2 0.34 6.42 4.01 -0.09 6
Lab CL 0.32 0.38 12 0.44 4.64 0.41 0.43 1
Lab RZ CL 0.35 0.31 12 0.34 5.67 2.68 0.15 3
NASIS SiL 0.15 0.37 12 0.35 5.77 2.93 0.12 4

 

 

 

 

 

 

 

 

 

 

 

 

Field 4—1 was a Clay Loam soil, with 21% sand, 47% silt, and 32% clay, which
shifts the textural class to CL fi'om SiCL due to a sand content greater than 20%. The
other three Amity SiL fields as defined by NASIS had Lab PSA sand values less than
20% and slightly larger silt values. One consequence of slightly greater sand content was
lower water content at FC. The CL default value is 35.33% water content for CL and
40.67% water content for the SiCL. The Lab data scenarios had the best NSE values.
One explanation was the inverse relationship between Lab data and Lab-RZ data. In
previous fields the water content at PC was greatest in the default RZ values and the Lab
data had lower values. Field 4—1 data were reversed. In fact the NSE values for both
rainfall distributions were well into the positive range indicating good model
performance. The Mean Error values were negative indicating the model under predicts
the Watermark data. The ME values for the Lab data -0.41 and -l .66 cm, which are 0.9%
and 3.9% of water content at FC, 42.04 cm of water.

With the exception of the first growing year period the Lab and Lab-RZ data fit
well for both rainfall distribution periods. Only during the second winter period in 1999
did the data fit the Watermark data. For the other four winter periods the modeled data

showed less water in the profile. The water profiles did not appear better further into the

200

study period (years 2000-2002). If the logic is that with time the modeled data more
closely resemble that of the measured Watermark data, the final year for this field did not
support that logic. In previous fields this was suggested that during the second year of

the model period the measured and modeled profiles showed better agreement.

5.19 Summary

Model performance was described based on the NSE factor, as a fimction of
comparing six modeled scenarios with the measured Watermark data in ten fields. The
Textural class was defined by PSA analysis in the laboratory and compared to the USDA
NASIS Soil Map Unit for each field. The clay content was listed as measured or default
NASIS values. The volumetric water content was listed as measured or default RZWQM
or NASIS data. The mean error values are the mean differences in all water content
values between the modeled data and the measured data. In almost all cases the modeled
data under predicted the measured data. The RMSE values indicate the magnitude of
variation from the linear regression lines; the greater the values as cm of water the greater
the variance from that line. High values indicate wide variances. The model rank values
are simply the rankings of NSE values fi'om highest positive value to greatest negative

value.

201

Chapter 6. Conclusions

6.1 RZWQM evaluation and performance

The modeling of volumetric water content over a time period on a field basis
requires some knowledge of soil texture, soil physical characteristics, and soil hydraulic
characteristics. RZWQM was selected to model the volumetric water content on ten
fields in the Willamette Valley Oregon, all on dairy farms with cropping and
management attribute data. Three methods were utilized, patterned after Starks et al.
(2003) work in Oklahoma. Logically, the first approach was total parameterization of
RZWQM using field soils and deriving PSA, bulk density, and the SWCC for each
increment depth. The second was using just the PSA data and using RZ default values
for all other parameters. The third was no field data at all; the model was parameterized
using USDA Soil Survey NASIS data, easily obtainable by knowing where the field is
located and selected the appropriate Soil Map Unit. These three approaches, identified as
Lab, Lab RZ, and NASIS, were each used as modeled data.

Two versions of these models were run. RZWQM uses breakpoint rainfall data.
Rainfall data, as volume depths, are listed over a period time defined by the volume depth
per unit time per storm event. The parameterization of the breakpoint data is rigorous.
One logical correction was using the 24-hour rainfall volume depth, and assigning that
depth for a length of time. In this study version one is two hours, and version two is 12
hours. The difference between the two versions is rainfall intensity only; the rainfall
volume in that 24ohour time period remained the same. Adjusting the intensity of rainfall

for the 24-hour period served as a check of model sensitivity.

202

The change of volumetric water content per unit time is controlled by ET,
capillarity, leaching > 119 cm soil depth, surface runoff, or lateral flow out of the field
boundary. In this study the Ks parameters were 1.5 cm h'1 at 30 cm depth, 0.68 cm h'1 for
lower depths (< 30 cm). Field studies were not done. At these Ks values, there was
minimal runoff volume off the field. There was minimal deep leaching of water out of
the root zone. The Ks value in RZWQM is one area that should be thoroughly
investigated because this flux gradient drives volumetric water content flux per unit time.
This study was not designed to model surface flow or solute leaching that might occur
through deep leaching, but this next step is measuring Ks values defined by field
measurement. The instantaneous method is suggested (Hillel, 1998). This study could
have been improved by having these values.

The crop modeling subcomponent of RZWQM was not extensively
parameterized. Input data included a rudimentary crop growth curve based upon harvest
yield of biomass and N uptake rates. The four landowners supplied crop harvest yield
data with nutrient composition. However the importance of defining ETc for the crops
grown on these or any other farm cannot be overstated. One additional crop input was
depth of root zone. These were assigned book values. Field measurements would

improve the relationship between ETC, crop yield, and growth curve.

6.2 Statistical evaluation with Nash Sutcliffe Equation values
The analysis of model performance was accomplished using four statistical
measures. The RMSE provided the standard deviation estimate comparing the

independent variable, Watermark data (field or observed), with the dependent variable,

203

model data (predicted or simulated). These units are in cm of water. The second, ME, is
the mean error of predicted minus observed. The negative values in almost all scenarios
gave some measure of model performance; the predicted values were less than the
observed values. The figures provided visual confirmation of these negative ME values.
Model performance was measured using one of the common hydrological model tools,
the Nash Sutcliffe Efficiency value (dimensionless). A perfect fit is one; any negative
value is poor. And the relative differences between scenarios can be compared using this
NSE value.

In the previous chapter the interpretation of each of these measures was done
using a table for each field. The tables included textural class, clay content, and
volumetric water content at FC. After completing the study these three parameters
seemed most important in defining the physical and hydraulic parameters required in this
study.

Figure 6.1 illustrates the comparison of NSE values over the ten fields used in this

study.

204

 

NSE Comparisons

NSE value

 

Field ID

 

 

 

Figure 6.1 Field comparison Nash Sutcliffe Equation values

Three visual interpretations are evident. Scenarios using only NASIS data do not
reflect Watermark data in either rainfall distributions, with the exception of field 3-6 (12
hour) and field 4-1 (12 hour). These two scenarios are slightly greater than zero.
Secondly, field 2-4 and field 3-7 have no positive NSE values. Field 2-4 is especially
poorly modeled. Third, farm one soils modeled particularly well with the exception of
the NASIS data.

Table 6.1 presents the additive NSE values, the mean values, and their population

standard deviation.

205

Table 6.1 Scenario sum, mean value, and standard deviation ranking

 

Scenario NSE sum mean sd

Lab 2 Hour -6.48 -0.65 2.15
Lab R2 2 Hour -7.76 -0.78 2.17
iNASIS 2 Hour -31.17 -3.12 1.89
Lab 12 Hour -4.04 -0.40 1.79
Lab RZ 12 Hour -5.71 -0.57 1.85
NASIS 12 Hour -25.58 -2.56 1.63

 

 

 

 

 

 

 

 

 

 

 

 

The NSE sum values provide some measure of model performance across all ten
fields. The smallest number, Lab 12 hour, rank first of all scenarios at predicting the
volumetric water content over the study period for each farm. The Lab RZ 12 hour is
next, followed by the same scenarios using the two-hour rainfall distribution. This seems
reasonable. Watermark data provide a field estimate of matric potential. If the SWCC
curve is measured on these soils, then model performance should improve. However the
minimal data input of PSA values only, as reflected in the Lab RZ scenarios is only
slightly less able to provide a good model fit.

The mean NSE values are grouped as ——0.4 to —0.78, with the better fit using the
12-hour rainfall distribution. This also seems reasonable. The reason for using this
particular sensitivity analysis was based on typical Mediterranean Climate; long-term
storm events with low intensity volume depths. Thus 0.77 inches of rainfall falling over
twelve hours is more likely in the Willamette Valley than that same volume depth falling
in two hours.

Confirming the NASIS data scenarios in the previous chapter, the mean values for
the two rainfall distributions are —25.58 and —3 1.17. NASIS data requires no field work,
however, and the model user risks the accuracy of instrument readings in a different

textural class.

206

6.3 Objectives met
The following two overall conclusions are made by answering the two objectives
stated in Chapter One.
1. Several input parameters were obtained or measured using field and laboratory
analysis. They are:
a. NASIS soils data available on the Internet
b. Moist bulk density from field cores
c. Particle size analysis from bulk field soil
(1. Volumetric water content at specific matric potential, defining the SWCC
and deriving paired data equations for these fields soil
2. RZW QM was evaluated using three methods of model parameterization. The
minimal input method, using NASIS data default values clearly did not fit the
measured field values using Watermark soil matric potential sensors. Further,
with the exception of one field (2-4) the identified Soil Map Unit in the County
Soil Survey did not match the textural class of the surface layer. The formation of
these soils is ongoing; they are alluvial floodplain soils constantly changing as
new flood events deposit new soils on these fields. In these nine mismatched
fields, the clay content of the laboratory derived PSA was higher. The ME values
of the NASIS data were largely negative, indicating the difference in textural
class played a driving role in under estimating the water holding capacity of the
profile at 119 cm depth. This study concludes that some field measurements are
required. The extent of field measurement is less clear. However the model

performed reasonable well using both the Lab and Lab RZ scenarios. The

207

derivation of the SWCC is time consuming and expensive. The problems of
measurement are expanded given the multiple wetting and drying cycles required
for complete description of the SWCC. The time period for this study was 18
months, including field sampling at or near FC, preparation, and extractor time.
The NSE values for Lab and Lab RZ scenarios were similar; therefore this study
concludes the derivation of the SWCC is not a necessary input for RZWQM

performance. Starks et al. (2003) reached the same conclusion.

6.4 RZWQM input requirements

What then is the minimum input required for RZWQM as examined in this study?
The answer to this question is one of most interest for the user of this model for
predicting the volumetric water content over a period of time. The inputs are:

a. PSA. Clearly an estimation of the textural class, measured in this study
using the hydrometer method, provided some refinement from using
strictly NASIS data. The PSA provides users with sand, silt, and clay
percentages, totaling 100%, which can be used to define textural class in
RZWQM.

b. Clay content. This PSA component can help adjust the volumetric water
content within a given textural class. For instance SiCL has a clay content
range of 28-40%, and in this study almost all surface layer field soils used
this textural class. However a volumetric water content curve as a
function of this clay content range of 12 % (28% to 40% range) can refine

the water holding capacity at FC.

208

c. Volumetric water content at FC. If the above curve is built then scenarios
should better fit observed or measured data. Adjusting water content
upwards as clay content increases is a reasonable adjustment for
RZWQM.

d. Development of K, values will refine water flux gradient per unit time,
thereby refining water in and out of the three-dimensional field space.

The use of a K, value so that surface runoff is rrrinimal is reasonable based
upon field evidence as supplied by the landowner and the researcher.
However less well known is water loss below the root zone as deep
leaching. The K, value can be used that minimize this loss, i.e. this study,
but a more valid approach of field determination would refine this loss.
This would be more important if RZWQM were used for simulating solute
transport (nitrate, atrazine) below the root zone or vadose zone. Further,
the use of K, data would be required for modeling the application of liquid
manure during the non growing season, whereas the surface soil horizon
might contain some void space but the irrigation event would replace that
void space with water. At greater than PC but less than saturation, if the
intake rate exceeded infiltration the potential for surface runoff would
increase. If the surface soil layer were saturated, then runoff would occur
if irrigation rate continued and K, was less than the intake rate.

e. Development of crop growth curves that help define ETc will improve
water flux gradient as transpiration over the growing season. Refinement

of water flow through plants to a more robust model will help simulate

209

different crop responses to irrigation volume depths, and the potential for
solute transport through the root zone and beyond. In this study, the crop
model parameters were minimal. Additional work might have helped fine
tune the predicted water content in the profile especially during the
growing season.

The climate parameters used in this study were reasonable. Some
refinement of the rainfall distribution was done using a 12-hour period
from a two-hour period. If runoff over a shorter period of time is
simulated, then specific breakpoint data, i.e. hyetograph knowledge, is
required. For the duration of this study daily data is reasonable, and the
landowner certainly does not want runoff of liquid manure. This comment
is addressed by the importance of knowing the irrigation volume depth
(amount), the surface soil layer intake rate as infiltrability or permeability,

and then knowledge of saturated hydraulic conductivity.

6.5 Needs for future work

a field basis.

1.

These points are provided as insight for the next attempt at modeling the FWB on

Placement of multiple soil moisture sensors in the field. Done again, this
researcher would have instrumented just the surface layer at several locations in
the field, thus comparing within field variation at the same depth. The lab PSA

could be done for determination of in-field variability (explaining, possibly, the

210

layering of alluvial soils after repeated flood events after the soils were originally
mapped).

. Obtaining moist bulk density soil samples near the sensors at the time each sensor
recorded 0.033 MPa potential, or FC. This one step could have helped define two
necessary parameters: moist bulk density and gravimetric water content, which
could have been converted to volumetric water content, and PSA. In this study
the timetable was not as convenient; the field sampling of the nearly undisturbed
soil cores (for SWCC determination) was not at the specific time the surface or
subsurface Watermark sensors gave a reading of 33 cbars. This is an oversight
that could easily be corrected. The point is that when a Watermark or other soil
moisture sensor is planted in the soil at some depth, taking a bulk soil sample at
the moment the instrument reads 0.033MPa will define the volumetric water
content by multiplying the gravimetric water weight by the bulk density (in this
case moist bulk density). In addition the saturation volumetric water content
could be measured when the sensor recorded zero matric potential. As in this
study the measured value can be compared with the calculated value of assuming
pore space completely filled with water.

. Some mention must be made of trying to define the FWB based upon an array of
four soil matric potential sensors at some location in a field. Given the variation
of the soil matrix over the entire field, the particular site selection can be typical
or some location that is not typical. How do we decide the answer to this
question: what is typical? In this study the measured data, or empirical observed

data, was considered correct; the model simulations are compared to these values.

211

In modeling, who can say the model scenarios are correct and that it is the
empirical data that are not representative of the field attributes? For this reason
one might conclude that using multiple samplings within field of bulk soil to
obtain volumetric water content at F C, then PSA, then measure K, in multiple
sites is correct methodology. In this study that seemed logical. This the
following statement summarizes this study and what could have been done better:
Rather than spend 18 months deriving the SWCC on selected soil cores from one
relatively small (within one square meter) spatial area, a method that sampled
more sites within that field using bulk soil taken with an open bucket auger in
each specific horizon, and determining PSA, gravimetric water content,
volumetric water content, at that moment in time when the soil moisture sensor
read at F C, would be the right input parameters required to simulate the F WB on a
field scale. And as stated earlier an assessment of K, could be done in situ using
the instantaneous profile method.

This study was required to reach that summary statement, and to that end,

accomplished its two primary objectives.

212

Bibliography

Ahuja, L. R. and L. MA. 1999. Parameterization of agricultural systems models: Current
issues and techniques. In: Ahuja, L. R., L. Ma., and T. A. Howell, eds. Agricultural

system models infield research and technology transfer. CRC Press, Boca Raton,
Florida.

Ahuja, L. R. and L. Ma. 2002. Parameterization of agricultural system models: current
approaches and future needs. In: Ahuja, L. R., L. Ma, and T. A. Howell (eds.)
Agricultural System Models in Field Research and Technology Transfer. Lewis
Publishers, Boca Raton, Florida.

Ahuja, L. R., L. Ma, J. T. Boesten, and H. J. Farahani. 1996. A field test of RZWQM
simulation model for predicting pesticide and bromide behavior. .1. Pesticide Sci. 48:101-
108.

Ahuja, L. R., J. W. Nancy, and R. D. Williams. 1985. Estimating soil water
characteristics from simpler properties or limited data. Soil Sci. Soc. Am. J. 49:1100-
1105.

Ahuja, L. R., R. E. Green, S. K. Chong, and D. R. Nielsen. 1980. A simplified functions
approach for determining soil hydraulic conductivities and water characteristics in situ.
Water Resour. Res. 16:947-953.

Ahuja, L. R., B. B. Barnes, D. K. Cassel, R. R. Bruce, and D. L. Nofziger. 1988. Effect
of assumed unit gradient during drainage on the determination of unsaturated hydraulic
conductivity and infiltration parameters. Soil Sci. 1451235-243.

Ahuja, L. R., D. G. DeCoursey, B. B. Barnes, and K. W. Rojas. 1993. Characteristics of
macropore transport studied with the ARS RZWQM. Trans. ASAE 36:369-380.

Ahuja, L. R., Fielder, G. H., G. H. Gunn, J. G. Benjamin, and A Garrison. 1998.
Cahnges in soil water retention curve due to tillage and natural reconsolidation. Soil Sci.
Soc. Am. J. 62:1228-1233.

Ahuja, L. R., K. W. Rojas, J. D. Hanson, M. J. Shaffer, and L. Ma (eds). 2000. Root
Zone Water Quality Model: Modeling Management Effects on Water Quality and Crop
Production. Water Resources Publications, Highland Ranch, Colorado.

Allen, R. G. 1996. Assessing the integrity of weather data used for use in reference
evapotranspiration estimation. .1. Irrig. And Drain. Eng. Div. ASCE 122:97-106.

Allen, R. G., M. Smith, A. Perrier, and L. S. Pereira. 1996. An update for the definition
of reference evapotranspiration. [CID Bulletin 45:1-34.

213

Allen, R. G., L. S. Pereira, D. Raes, and M. Smith. 1998. Crop evapotranspiration:
guidelines for computing water requirements. FAO Irrigation and Drainage Paper 56,
Rome.

Amoozeger, A. and A. W. Warrick. 1996. Hydraulic conductivity of saturated soils:
field methods. In: Klute, A., ed., Methods of Soil Analysis, Part 1: Physical and
Mineralogical Methods, Monograph No. 9. Am. Soc. Agron., Madison, Wisconsin.

Arya, L. M. and J. F. Paris. 1981. A physioemperical model to predict soil moisture
characteristics from particle size distribution and bulk density data. Soil Sci Am. J.
45:1023-1030.

Arya, L. M., F. J. Leij, P. J. Shouse, and M. Th. Van Genuchten. 1999. Relationship
between the hydraulic conductivity function and the particle size distribution. Soil Sci,
Soc Am. J. 63:1063-1070.

Arya, L. M., F. J. Leij, M. Th. Van Genuchten, and P. J. Shouse. 1999. Scaling
parameter to predict the soil water characteristic fiom particle size distribution data. Soil
Sci. Soc. Am. J. 63:510-519.

Assouline, S., D. Tessier, and A. Bruand. 1998. A conceptual model of the soil water
retention curve. Water Resour. Res. 34:223-231.

Azevedo, A. S., R. S. Kanwar, P. Singh, L. R. Ajuha, and L. S. Pereira. 1997.
Simulating Atrazine transport using RZWQM for Iowa soil. J. Environ. Qual. 26: 1 53-
164.

Black, T. A., W. R. Gardner, and G. W. Thurtell. 1968. The prediction of
evapotranspiration, drainage, and soil water storage in bare soil. Soil Sci. Soc. Am. Proc.
32:655-660.

Blaney, H. F. and W. D. Criddle. 1962. Determining consumptive use and irrigation
water requirements. Technical Bulletin 1275. Soil Conservation Service, U.S.
Department of Agriculture, Washington DC.

Blaney, H. F. and W. D. Criddle. 1950. Determining water requirements in irrigated
areas fiom climatological and irrigation data. United States Department of Agriculture
Soil Conservation Service Rep. SCS-T P 96, Washington DC.

Bloemmen, G. W. 1980. Calculation of hydraulic conductivities of soils from texture
and organic matter content. Z. Pflanzenemaehr. Bodenkd. 143:581-605.

Bosecker, R. R. 1997. Census of Agricluture: Agricultural Maps.

hgpzflwwwnasgusdigov/cegsus/itl997. U.S Department of Agriculture, Nataional
Agricultural Statistics Service. (February 2003).

214

Bouchet, R. J. 1963. Evapotranspiration potential. International Association of
Scientific Hydrology. Publication 62. Gentbrugge, Belgium

Bouwer, H. 1966. Rapid field measurement of air entry value and hydraulic

conductivity of soil as significant parameters in flow system analysis. Water Resour.
Res. 2:729-738.

Bouwer, H. 1986. Intake rate. In: Klute, A., ed., Methods of Soil Analysis, Part I:
Physical and Mineralogical Methods, Monograph No. 9. Am. Soc. Agron., Madison,
Wisconsin.

Bowers, S. A. and R. J. Hanks. 1965. Reflection of radiant energy from soils. Soil Sci
100:130-138.

Boyer, J. S. 1995. Measuring the water status of plant and soils. Academic Press, San
Diego, CA.

Brakensiek, D. L. and C. A. Onstad. 1977. Parameter estimation of the Green Ampt
infiltration equation. Water Resour. Res. 13:1009-1012.

Brooks, R. H. and Corey, A. T. 1964. Hydraulic properties of porous media. Hydrology
Paper no. 3. Colorado State University, Ft. Collins, Colorado.

Campbell, G. S. 1974. A simple method for determining unsaturated conductivity from
moisture retention data. Soil Sci. 117:311-314.

Canny, M. J. 1998. Transporting water in plants. American Scientist, 86:152-159.

Carsel. R. F. and R. S. Parrish. 1998. Developing joint probability distributions of soil
water retention characteristics. Water Resour. Res. 24:755-769.

Childs, E. C. and M. Bybordi. 1969. Ther vertical movement of water in stratified
porous media, 1. Infiltration. Water Resour. Res. 5:2.

Clothier, B. 13., and K. R. J. Smetten. 1990. Combining laboratory and field
measurements to define the hydraulic properties of the soil. Soil Sci. Soc. Am. J. 54:299-
304.

Clothier, B. E., and D. Scooter. 2002. Unsaturated water transmission parameters
obtained from infiltration. In: Klute, A., ed., Methods of Soil Analysis, Part 1: Physical
and Mineralogical Methods, Monograph No. 9. Am. Soc. Agron., Madison, Wisconsin.

Clothier, B. E., D. R. Scotter, and A. E. Green. 1983. Diffusivity and one dimensional
absorption experiments. Soil Sci. Soc. Am. J. 47:641-644.

215

Coddington, D. C., H. L. Havis, and J. G. Milliken. 1972. The Federal Reclamation
Program: Its impact, issues, and fixture considerations. Denver, Colorado. In: M.
Robinson, Water for the West: The Bureau of Reclamation, Public Works Historical
Society, Chicago, Illinois.

Cook, S. M. F. 1996. Tillage and macropore structure: preferential transport of atrazine
and rhodamine WT. Ph.D. Dissertation. University of Minnesota, St. Paul, Minnesota.

Conant, J. B. and L. K. Nash. 1957. Harvard Case Histories in Experimental Science.
Harvard Publishing, Inc. Boston, Massachusetts.

Cuenca, R. H. 1989. Irrigation System Design, An Engineering Approach. Prentice
Hall, Englewood Cliffs, New Jersey.

Dane, J. H. and C. G. Topp, eds. Methods of Soil Analysis, Part 4. Physical Methods.
SSSA Book Series 5. Soil Sci. Soc. Am., Madison, Wisconsin.

Daumus, M. 1958. The chemistry of principles. In: R. Taton, ed., The beginnings of
modern science. Thames and Hudson, London.

Davenport, D. A. 1984. John Dalton’s first paper and last experiment. Chem Matters,
April, 1984, page 14.

Dingrnan, S. L., W. F. Weeks, and Y. C. Yen. 1968. The effects of thermal pollution on
river ice conditions. Water Resour. Res. 4:347-364.

Doorenbos, J. and Pruitt, W. O. 1977. Crop water requirements, Irrig. Drain. Paper 24.
U. N. Food and Agriculture Organization, Rome.

Elrick, D. E., W. D. Reynolds, and K. A. Tan. 1989. Hydraulic conductivity
measurements in the unsaturated zone using improved well analysis. Ground Water
Monit. Rev. 9:184—193.

Evett, S. R. 1998. Energy and water balances at soil plant atmosphere interfaces. In:
CRC Handbook of Soil Science. Chapter 5. CRC Publishers, Boca Raton, Florida.

Evett, S. R. and R. J. Lascano. 1993. ENWATERBALBAS: A mechanistic
evapotranspiration model written in compiled BASIC. Agron. J. 85:763-772.

Evett, S. R. Coaxial multiplyer for time domain reflectomety measurement of soil water
content and bulk electrical conductivity. Trans. ASAE 41:361-369.

FAO. 1998. Crop evapotranspiration — guidelines for computing crop water

requirements. Allen, R. G., L. S. Pereira, D. Reas, and M. Simth. Food and Agriculture
Organization: Irrigation and Drainage Paper 56. Rome, Italy.

216

FarahaniH. J. and W. C. Bausch. 1995. Performance of evapotranspiration models for
maize — bare soil to closed canopy. Trans. ASAE 38: 1049-1059.

Fetter, C. W. 1994. Applied Hydrogeology. Third Edition. Prentice Hall College
Division, Upper Saddle River, New Jersey.

Fitzgerald, D. 1886. Evaporation. Trans. ASCE. 15:581-646.

Flerchinger, G. N., R. M. Aiken, K. W. Rojas, and L. R. Ahuja. 2000. Development of
RZWQM for over winter conditions. Trans. ASAE 43:59-68.

Flint, A. L. and L. E. Flint. 2002. Particle density. In: J. H. Dane and G. C. Topp, eds.
Methods of Soil Analysis, Part 4. Physical Methods. SSSA Book Series 5. Soil Sci. Soc.
Am., Madison, Wisconsin.

Freeze, R. A. and Cherry, J. A. 1979. Groundwater. Prentice Hall, Englewood Cliffs,
New Jersey.

Freyberg. D. L., J. W. Reeder, J. B. Franzini, and I. Remson. 1980. Applications of the
Green Ampt model to infiltration under time dependent surface water depths. Water
Resour. Res. 16:517-528.

Gardner, B. L. 2002. American agriculture in the twentieth century: How it flourished
and what it cost. Harvard University Press, Cambridge, Massachusetts.

Gates, P. W. 1968. History of public land law development. Washington DC. In: M.
Robinson, Water for the West: The Bureau of Reclamation, Public Works Historical
Society, Chicago, Illinois.

Gee G. W. and D. Or. 2002. Particle size analysis. In: J. H. Dane and G. C. Topp, eds.
Methods of Soil Analysis, Part 4. Physical Methods. SSSA Book Series 5. Soil Sci. Soc.
Am., Madison, Wisconsin.

George, B. A., B. R. S. Reddy, N. S. Raghuwanshi, and W. W. Wallender. 2002.
Decision support system for estimating reference evapotranspiration. J. Irrig. Drain.
Eng. 128:1-10.

Ghidey, F., E. E. Alberts, and N. R. Kitchen. 1999. Evaluation of the RZWQM using
field measured data from the Missouri MSEA. Agron. J. 91:183-192.

Grebet, R. and R. H. Cuenca. 1991. History of lysimeter design and effects of
environmental disturbances. In: Lysimeters for evapotranspiration and environmental
measurements. Proceedings of the International Symposium of Lysimetry, Honolulu,
Hawaii. ASCE 345 E. 47th St., New York, New York.

217

Green, W. H. and G Ampt. 1911. Studies of soil physics, Part I. The flow of air and
water through soils. J. Agric. Sci. 4(1):1_-14.

Grossman, R. B., F. B. Pringle, R. J. Bigler, W. D. Broderson, and T. M. Sobecki. 1990.
Systematics and field morphology for a use-dependant temporal soil data base for
edaphological and environmental applications. In: International symposium on advanced
technology in natural resources management. Georgetown University, Washington DC.

Grossman, R. B. and T. G. Reinsch. 2002. Bulk density and linear extensibility. In: J. H.
Dane and G. C. Topp, eds. Methods of Soil Analysis, Part 4. Physical Methods. SSSA
Book Series 5. Soil Sci. Soc. Am., Madison, Wisconsin.

Hacham, A. Y. and J. F. Alfaro. 1980. Rain infiltration into layered soils: Prediction. J.
Irrig. Drain. Eng. 106:311-321.

Hansen, 8., M. J. Shaffer, and H. E. Jensen. 1995. Developments in modeling nitrogen
transformations in soil. In P. E. Bacon, ed., Nitrogen Fertilization and the Environment.
Marcel Dekker, Inc. Ames, Iowa.

Hansen, V. E., O. A. Israelason, and G. Stingham. 1979. Irrigation Principles and
Practice. John Wiley and Sons, Inc. New York, New York.

Hanson, J. D., K. W. Rojas, and M. J. Shaffer. 1999. Calibration and evaluation of the
RZWQM. Agron. J. 91:171-177.

Harbeck, G. E., Jr., M. A. Kohler, and G. E. Koberg. 1958. Water loss investigations,
Lake Mead studies. U.S. Geological Survey Professional Paper 28. Washington DC.

Hillel, D. 1980. Fundamentals of Soil Physics. Academic Press, New York.
Hillel, D. 1998. Environmental Soil Physics. Academic Press, San Diego, California.

Hills, R. G. and A. W. Warrick. 1993. Burger’s equation: A solution for soil water flow
in a finite length. Water Resour. Res. 29:1179-1185.

Hodge, C. 1963. Aridity and Man: The challenge of the arid lands in the United States.
Washington DC. In: M. Robinson, Water for the West: The Bureau of Reclamation,
Public Works Historical Society, Chicago, Illinois.

Hollen, E. W. 1966. The great Amercian desert: Then and now. New York, New York.
In: M. Robinson, Water for the West: The Bureau of Reclamation, Public Works
Historical Society, Chicago, Illinois.

Holton, H. N. 1961. A concept of infiltration estimates in watershed engineering.
USDA-ARS I112” 45-51. Washington, DC.

218

Holton, H. N., G. J. Stiltner, W. H. Hensen, and N. C. Lopez. 1975. USDAHL-74
revised model of watershed hydrology. Agricultural Research Service Technical Bulletin
1518. U.S. Department of Agriculture, Washington DC.

Horton, R. E. 1917. A new evaporation formula developed. Eng. News Rec. 78:196-
199.

Horton, R. E. 1940. An approach toward a physical interpretation of infiltration
capacity. Soil Sci Soc. Am. Proc. 5:399-417.

Howell, T. A., J. L. Steiner, S. R. Evett, A. D. Schneider, K. S. Copeland, D. A. Dusek,
and A. Tunick. 1991. History of lysimeter design for use and for evapotranspiration
measurements. In: Lysimeters for evapotranspiration and environmental measurements.
Proceedings of the International Symposium of Lysimetry, Honolulu, Hawaii. ASCE 345
E. 47th St., New York, New York.

Howell, T. A., J. L. Steiner, A. D. Schneider, S. R. Evett, and J. A. Tolk. 1995.
Evapotranspiration of irrigated winter wheat, sorghum, and corn. ASAE paper no. 95-
2081.

Hulse, D. S. Gregory, and J. Baker. 2002. Willamette river basin planning atals:
Trajectories of environmental and ecological change. Oregon State University Press,
Corvallis, Oregon.

Itenfisu, D., R. L. Elliot, R. G. Allen, and I. A. Walter. 2003. Comparison of reference
evapotranspiration calculations as part of the ASCE standardization effort. J. Irrig.
Drain. Eng. 129:440-448.

James, L. G. 1988. Principles of Farm Irrigation System Design John Wiley and Sons,
New York, New York.

Jaynes, D. B., and J. G. Miller. 1999. Evaluation of RZWQM using field measured data
from Iowa MSEA. Agron. J. 91 :192-200.

Jensen, M. E., R. D. Burman, and R. G. Allen. 1990. Evapotranspiration and irrigation
water requirements. ASCE Manuals and Reports on Engineering Practices No. 70. New
York.

Jensen, M. E. and H. R. Haise. 1963. Estimating evapotranspiration from solar radiation.
J. Irrig. Drain. Div. ASCE 89:15-41.

Jensen, M. E. 1980. Design and operation of farm irrigation systems. Monograph no. 3,
ASAE, St. Joseph, Michigan.

Jensen, M. E, J. L. Wright, and B. J. Pratt. 1971. Estimating soil moisture depletion for
climate, crops and soil data. Trans. ASAE 14:954-959.

219

Jensen, M. E., R. D. Burman, and R. G. Allen. 1990. Evapotranspiration and irrigation
water requirements. Irrigation and Drainage Div. ASCE. ASCE Manuals and Reports on
Engineering Practice No. 70. St. Joseph, Michigan.

Johnson, K. E., H. H. Liu, J. H. Dane, L. R. Ahuja, and S. R. Workman. 1995.

Simulating fluctuating water tables and tile drainage with the RZWQM and a new model
WAFLOWM. Trans. ASAE 38:75-83.

Kirkham, D., and C. H. M. van Bavel. 1948. Theroy of seepage into auger holes. Soil
Sci. Soc. Am. Proc. 13:75-82.

Klute, A. 1986. Water Retention: Laboratory Methods. In: Klute, A., ed., Methods of
Soil Analysis, Part 1: Physical and Mineralogical Methods, Monograph No. 9. Am. Soc.
Agron., Madison, Wisconsin.

Kosugi, K., J. W. Hopmans, and J. H. Dane. 2002. Parametric models. In: In: J. H.
Dane and G. C. Topp, eds. Methods of Soil Analysis, Part 4. Physical Methods. SSSA
Book Series 5. Soil Sci. Soc. Am., Madison, Wisconsin.

Kohler, M. A., T. J. Nordenson, and W. E. Fox. 1955. Evaporation fi'om pans and lakes.
Research paper no. 38. United States Department of Commerce, Weather Bureau,
Washngton D.C.

Kohelker, M. A., T. J. Nordenson, and W. E. Fox. 1955. Evaporation from pans and
Lakes. Research Paper 38. U.S. Weather Bureau, Washington DC.

Kostiakov, A. N. 1932. On the dynamics of the coefficient of water percolation in soils
and on the necessity of studying it from a dynamic point of view for purposes of
amelioration. Trans. Com. Int. Soc. Soil Sci, Moscow, Part A:l7-21.

Kosugi, K., and J. W. Hopmans. 2002. Scaling water retention curves for soils with
lognormal pore size distribution. Soil Sci. Soc Am. J. 66: 1496-1505.
Krikham 1945.

Kumar, A., R. S. Kanwar, and L. R. Ajuha. 1998. RZWQM simulation of nitrate
concentrations in subsurface drainage from manured plots. Trans. ASAE 41 :587-596.

La], R., Kinble, J. M., Follett, R. F., and Stewart, B. A., eds. 1998. Management of
Carbon Sequestration in Soil. CRC Press, Boca Raton, Florida.

Landa, F. M., N. R. Fausey, S. E. Nokes, and J. D. Hanson. 1999. Plant production
model evaluation for the RZWQM in Ohio. Agonom. J. 91:220-227.

Larson, W. E., and F. J. Pierce. 1994. The dynamics of soil quality as a measure of
sustainable agriculture. In: J. W. Doran, D. C. Coleman, D. F. Bezdicek, and B. A.

220

Stewart, eds. Defining soil quality for a sustainable development. Soil Sci. Am. Spec.
Pub. No. 35. Madison, Wisconsin.

Lascano, R. J. and J. L. Hatfield. 1992. Spatial variability of evaporation along two
transects of a bare soil. Soil Sci. Soc. Am. J. 56:341-346.

Leij, F. J., W. J. Alves, M. Th. Van Genuchten, and J. R. Williams. 1999. The
UNSODA unsaturated soil hydraulic database. In: M. Th. Van Gneuchten, F. J. Leij, and
L. Wu, eds., Proceeding International workshop characterizing and measurement of
hydraulic properties of unsaturated porous media. University of California, Riversdie,
California.

Logsdon, S. D. and C. A. Cambardella. 2000. Temporal changes in small depth
incremental soil bulk density. Soil Sci. Soc. Am. J. 64:710-714.

Lourence, F. and R. Moore. 1991. Prefabricated weighing lysimeters for remote weather
stations. In: Lysimeters for evapotranspiration and environmental measurements.
Proceedings of the International Symposium of Lysimetry, Honolulu, Hawaii. ASCE 345
E. 47th St., New York, New York.

Ma, L., H. D. Scott, M. D. Shaffer, and L. R. Ahuja. 1998. RZWQM simulations of
water and nitrate movement into a tall fescue field. Soil Sci. 163:259-270.

Mandal, U. K, K. S. Sundara Sarrna, U. S. Victor, and N. H. Rao. 2002. Profile water
balance water model under irrigated and rainfed systems. Agron. J. 94: 1204-121 1.

Marek, T. H., A. D. Schneider, T. A. Howell, and L. L. Ebeling. 1988. Design and
construction of large weighing monolithic lysimeters. Trans. ASAE 31 :477-484.

Martin, D. L. and D. G. Watts. 1999. Evaluation of RZWQM for conditions in Central
Nebraska. Agron. J. 91 :201-21 1.

Mattikalli, N. M., E. T. Engrnan, L. R. Ahuja, and T. J. Jackson. 1998. Microwave
remote sensing of soil moisture for estimation of profile soil property. Int. J. Remote
Sensing. 19:1751-1767.

McCuen, R. H., W. J. Rawls, and D. L. Brakensiek. 1981. Statistical analysis of the
Brooks Corey and the Green Ampt parameters across soil textures. Water Resour. Res.
17:1005-1013.

Mengal, K. and E. A. Kirkby. 2001. Principles of Plant Nutrition. Fifih edition. Kluwer
Academic Publishers, Norwell, Massachusetts.

Meyer, A. F. 1915. Computing runoff fi'om rainfall and other physical data. Trans.
ASCE 79:1056-1155.

221

Monteith, J. L. 1965. Evaporation and environment. In: Nineteenth Symp. Soc Exp.
Biol. 19:205-234.

Musgrave, G. W. 1955. How much of the rain enters the soil? In: Water Yearbook of
Agriculture. U.S. Department of Agriculture, Washington DC.

Neuman, S. P. 1976. Wetting front pressure head in the infiltration model of Green and
Ampt. Water Resour. Res 12: 546-566.

Nokes, S. E., F. M. Landa, and J. D. Hanson. 1996. Evaluation of the crop component of
the RZWQM for corn in Ohio. Trans. ASAE 39:1177-1184.

Or, D. and J. M. Wraith. 1999a. Soil water content and water potential relationships. In:
M. Summer, ed., Handbook of Soil Science. CRC Press, Boca Raton, Florida.

Or, D., and J. M. Wraith. 1999b. A new soil matric potential sensors based on time
domain reflectometry. Water Resour. Res. 11:3399-3497.

Oregon Department of Agriculture. 2004a. Registrations, licenses, and permits issued by
the Oregon Department of Agriculture. http://www.oda._state.or.u§/regs.html. (March,
2004)

Oregon Department of Agriculture. 2004b. Overview: Confined Animal Feeding
Operations Program. http://www.oda.state.or.us/nrd/cafo/prg.html. (March 2004).

Oregon Department of Agriculture. 2004c. Water Quality Management Program.
http://www.oda.state.or.us/nrd/water quaflWmanprachtml. (March 2004).

Pair, C.H., W. H. Hintz, K. R. Frost, R. E. Sneed, and T. J. Schultz (Eds). 1983.
Irrigation. The Irrigation Association. Silver Spring, Maryland.

Parlang, J. Y., I. Lisle, R. D. Braddock, and R. E. Smith. 1982. The three parameter
infiltration equation. Soil Sci. 133:337-341.

Penman, H. L. 1948. Natural evaporation from open water, bare soil and grass. Proc. R.
Soc. London Ser. A 193:120-146.

Penman, H. L. 1963. Vegetation and hydrology. Technical communication no. 53,
Commonwealth Bureau of Soils, Harpenden, England.

Pereira, L. S., A. Perrier, R. G. Allen, and I. Alves. 1999. Evapotranspiration: concepts
and future trends. J. Irrig. Drain. Eng. 125245-51.

Perrier, A. 1977. Evapotranspiration measurement using hydrologic factors. The
Meteorologic, 627-16.

222

Peart, R. M., and R. B. Curry. 1998. Agricultural Systems Modeling and Simulation.
Marcel Dekker, Inc. New York, New York.

Phene, C. J ., R. J. Reginato, B. Itier, and B. R. Tanner. 1990. Sensing irrigation needs.
In: G. J. Hoffman, T. A. Howell, and K. H. Solomon, eds. Management of Farm
Systems. Amer. Soc. Agr. Eng., St. Joseph, Michigan.

Philip, J. R. 1957. The theory of infiltration. 1. The infiltration equation and its
solution. Soil Sci. 83:345-357.

Philip, J. R. 1969. Theory of infiltration. Adv. Hydrosci. 52215-290.

Philip, J. R. 1995. Desperately seeking Darcy in Dijon. Soil Sci. Soc. Am. J. 59:319-
324.

Pochop, L. O., J. Borrell, and R. D.Burman. 1984. Elevation A bias in SCS Blaney
Criddle ET estimates. Trans. ASCE. 27:125-128.

Rawls, W. J. and D. L. Brakensiek. 1982. Estimating soil water retention from soil
properties. J. Irrig. Drain. Div. Am. Soc. Civ. Eng. 108:166-171.

Rawls, W. D., D. L. Brakensiek, and S. D. Logsdon. 1993. Predicting saturated
hydraulic conductivities utilizing fractal principles. Soil Sci. Soc. Am. J. 57:1193-1197.

Rawls, W. J. and D. L. Brakensiek. 1982. Estimating soil water retention from soil
properties. Div., ASCE 108 (IR2):166-171.

Rawls, W. J., D. L. Brakensiek, and B. Soni. 1983. Agricultural management effects on

soil water processes. Part 1. Soil water retention and Green Ampt infiltration parameters.
Trans. ASAE 26:1747-1752.

Rawls, W. J ., D. Gimenez, and R. Grossman. 1998. Use of soil texture, bulk density,

and slope of water retention curve to predict saturated hydraulic conductivity. Trans.
ASAE 41 :983-988.

Reynolds, W. D. 1993. Saturated hydraulic conductivity: field measurement. In: M. R.
Carter, ed. Soil Sampling and Methods of Analysis. Lewis Publ., Boca Raton, Florida.

Reynolds, W. D., B. T. Bowman, R. R. Brunke, C. F. Drury, and C. S. Tan. 2000.
Comparison of tension infiltrometer, pressure infiltrometer, and soil core estimates of
saturated hydraulic conductivity. Soil Sci. Soc. Am. J. 64:478-484.

Richards, L. A. 1931. Capillary conduction of liquids in porous mediums. Physics
12318-333.

223

Richardson, C. W. and D. A. Wright. 1984. WGEN: A model for generating daily
weather variables. ARS-8, Agricultural Research Service, U.S. Department of
Agriculture.

Ritchie, J. T. 1972. Model for prediciting evaporation from a row crop with incomplete
cover. Water Resour. Res. 8: 1204-1213.

Ritchie, J. T., T. A. Howell, W. S. Meyer, and J. L. Wright. 1999. Sources of biased
errors in evaluating evapotranspiration equations. In: C. R. Camp, E. J. Sadler, and R. E.
Yoder, eds. Evapotranspiration and irrigation scheduling. Proceedings of the
International Conference, 1999. San Antonio, Texas.

Robinson, M. 1979. Water for the West: The Bureau of Reclamation. Public Works
Historical Society Press, Chicago, Illinois.

Rower, C. 1931. Evaporation from free water surfaces. USDA Technical Bulletin 27,
U.S. Department of Agriculture, Washington DC.

Saxton, K. E., W. J Rawls, J. S. Romberger, and R. I. Papendick. 1986. Estimating
generalized soil water characteristics from texture. Soil Sci. Soc. Am. J. 50:1031-1036.

Saxton, K. E. 1998. Estimating generalized soil water characteristics from texture.
http://www.bsvse.wsu.edu/saxton/soilwater/Articlehtm. Washington State University
Biosystems Engineering Department, Pullman, Washington. (January, 2000).

Schneider, A. D., T. A. Howell, T. A. Moustafa, S. R. Evett, and W. S. Abou-Zeid. 1996.
A simplified weighing lysimeter for developing countries. In: C. R. Camp, E. J. Sadler,
and R. E. Yoder, eds. Evapotranspiration and irrigation scheduling. Proceedings of the
International Conference, 1999. San Antonio, Texas.

Schneider, A. D. 1998. Personal communication. In: S. R. Evett, ed., Energy and water
balances at soil plant atmosphere interfaces. CRC Handbook of Soil Science. CRC
Publishers, Boca Raton, Florida.

Sepaskhah, A. R. and L. Boersma. 1979. Thermal conductivity of soils as a function of
temperature and water content. Soil Sci. Soc. Am. J. 43:439-444.

Singh, P. and R. S. Kanwar. 1995. Modification of RZWQM for simulating subsurface
drainage by adding a tile flow component. Trans. ASAE 38:489-498.

Skaggs, R. W., and R. Khaleel. 1982. Infiltration. In: Hydrologic modeling of small
watersheds. Monograph 5, C. T. Haan ed. Am. Soc. Ag. Eng, St. Joseph, Michigan.

Smith, H. N. 1950. The American west as symbol and myth. Cambridge,

Massachusetts. In: M. Robinson, Water for the West: The Bureau of Reclamation, Public
Works Historical Society, Chicago, Illinois.

224

 

Smythe, W. E. 1911. The Conquest of Arid America. New York, New York. In: M.
Robinson, Water for the West: The Bureau of Reclamation, Public Works Historical
Society, Chicago, Illinois.

Soil Survey Staff. 1975. Soil taxonomy. A basic system of soil classification for
making and interpreting soil surveys. United States Department of Agriculture — Soil
Conservation Service Agric. Handb. 436. U. S. Government Printing Officde,
Washington DC.

Starks, P. J ., G. C. J eathman, L. R. Ahuja, and L. Ma. 2003. Use of limited soil property
data and modeling to estimate root zone soil water content. J. Hydro]. 27 2:13 1-147.

Stegner, W. 1954. Beyond the Hundredth Meridian. Boston, Massachusetts. In: M.
Robinson, Water for the West: The Bureau of Reclamation, Public Works Historical
Society, Chicago, Illinois.

Taiz, F. G. and L. M. Zeiger. 2002. Plant Physiology. Sinauer Associates, Inc.
Publishers. Sunderland, Massachusetts.

Thomas, G. 1948. Early Irrigation in the Western States. Salt Lake City, Utah. In: M.
Robinson, Water for the West: The Bureau of Reclamation, Public Works Historical
Society, Chicago, Illinois.

Tindall, J. A. and J. R. Kunkel. 1999. Unsaturated Zone Hydrology for Scientists and
Engineers. Prentice Hall, New Jersey.

Tindall, J. A., K. Hemmen, and J. F. Dowd. 1999. An improved method for field
extraction and laboratory analysis of large, intact soil cores. J. Environ. Qual. 28:259-
263.

Todd, D. K. 1964. Groundwater. In: V. T. Chow, ed., Handbook of applied hydrology.
McGraw Hill, New York, New York.

Topp, G. C., and P. A. Ferre. 2002. The soil solution phase. In: J. H. Dane and G. C.
Topp, eds. Methods of Soil Analysis, Part 4. Physical Methods. SSSA Book Series 5.
Soil Sci. Soc. Am., Madison, Wisconsin.

USBR. 2002. The Pacific Northwest Cooperative Agriculture Weather Network. U.S.
Bureau of Reclamation. http://www.r;sg)r.gov/pn/agg'met/proginfohtml (October 2002).

USBR. 2004. Bureau of Reclamation: Facts and figures. U. S. Bureau of Reclamation.
http://www.usbrggl/main/about/facthtml (February, 2004)

USDA. 1972. National Engineering Handbook. Hydrology Section, Chapters 4-10.
Soil Conservation Service, U. S. Department of Agriculture, Washington D. C.

225

USDA. 2004a. Technical Service Provider Registry. http://techregusdwvh U.S.
Department of Agriculture. (July 2003).

USDA. 2004b. Technical Resources. http://www.nrcs.usda.gov/technical/.
U.S. Department of Agriculture. (July 2003).

USDA. 2004c. NRCS National Procedures and Planning Handbook: Draft
Comprehensive Nutrient Management Planning Technical Guidance.
httn://www.nrcs.usdagov/prows/afo/cnmp guideiindexhtml. U.S. Department of
Agriculture. (July 2003).

USEPA. 1972. Clean Water Wct History. httn://www.epagov/regionS/warter/cwzfitm.
U. S. Department of Environmental Quality. (November, 2002).

USEPA. 1997. Clean Water Action Plan.
http://www.epa.govmistow/topics/cwa/O3.htm. U. S. Department of Environmental
Quality. (December, 2002).

USEPA. 1998. Estimation of infiltration rate in the vadose zone: Compilation of simple
mathematical models. EPA/600/R-1 7/128a. National Risk Management Laboratory,
Ada, Oklahoma.

USEPA. 2002. National Pollution Discharge and Elimination Permit: Animal Feeding
Operations. http://cfpub2.eoagov/npdesflromecfm. U. S. Department of Environmental
Quality. (December, 2002).

USEPA. 2003. Concentrated Animal Feeding Operation: Final Rule.
http://cfpub2.gpagov/npdes/afo/cafofinfialrule.cfrn. U. S. Department of Environmental
Quality. (December, 2003).

van Genuchten, M. Th. 1980. A closed from equation for predicting the hydraulic
conductivity of unsaturated soils. Soil Sci. Soc. Am. J. 44:892-898.

Vereechen, H. 1995. Estimating the unsaturated hydraulic conductivity from theoretical
models using simple soil properties. Geoderma 65:81-92.

Viehmeyer, F. J. and F. A. Brooks. 1954. Measurements of cumulative
evapotranspiration fiom bare soil. Trans. Am. Geophysical Un. 35:601-607.

Walker, S. E. 1996. Modeling nitratye in tile drained watersheds in east central Illinois.
Unpublished Ph.D. Doctoral Dissertation, Agricultural Engineering Department,
University of Illinois, Urbana Charnpaign, Illinois.

Watts, D. G., N. R. Fausey, and D. A. Bucks. 1999. RZWQM: Background of the
MSEA-RZWQM modeling project. J. Environ. Qual. 91:169-170.

226

Wildenschild, D., J. W. Hopmans, and J . Simunek. 2001. Flow rate dependence of soil
hydraulic characteristics. Soil Sci. Soc. Am. J. 65:35-48.

Williams, R. D., and L. R. Ahuja. 1993. Using available water content with the one
parameter model to estimate soil water retention. Soil Sci. 156:380-388.

Woston, J. H. M. and M. Th. Van Genuchten. 1988. Using soil texture and other soil
properties to predict unsaturated soil hydraulic functions. Soil Sci. Am. J. 52:1762-1770.

Wraith, J. M., and D. Or. 2001. Soil water characteristic determination from concurrent
water content measurements in reference porous media. Soil Sci. Soc. Am. J. 65:1659-
1666.

Wright, J. L. 1982. New evapotranspiration crop coefficients. J. Irrig. Drain. Eng. Div.
108:57-74.

Wright, J. L. 1991. Using weighing lysimeters to develop evapotranspiration crop
coefficients. In: Lysimeters for evapotranspiration and environmental measurements.
Proceedings of the International Symposium of Lysimetty, Honolulu, Hawaii. ASCE 345
E. 47th St., New York, New York.

Wu, L., L. Pan, J. Mitchell, and B. Sanden. 1999. Measuring saturated hydraulic
conductivity using generalized solution for single ring infiltrometers. Soil Sci. Soc. Am.
J. 63:788-792.

Wu, L. R. R. Allmaras, J. B. Lamb, and K. E. Johnson. 1996. Model sensitivity to
measured and estimated hydraulic properties of a Zimmerman fine sand. Soil Sci. Soc
Am. J. 60:1283-1290.

Young, M. H., and J. B. Sisson. 2002. Tensiometry. In: J. H. Dane and G. C. T opp, eds.

Methods of Soil Analysis, Part 4. Physical Methods. SSSA Book Series 5. Soil Sci. Soc.
Am., Madison, Wisconsin.

Youngs, E. G., G. Spoor, and G. R. Goodall. 1997. Infiltration from surface ponds into
soils overlaying a very permeable substratum. J. Hydrol. 186;327-334.

Zavattaro, L. and C. Grignani. 2001. Deriving hydrological parameters for modeling
water flow under field conditions. Soil Sci. Soc. Am J. 65:655-667.

227