 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProj/Acc&Pres/ClRC/DaIeDuemdd

CHARACTERIZATION OF DISSOLVED ORGANIC CARBON:
ASSESSMENT OF COPPER COMPLEXATION AND EXPORT OF CARBON FROM
WATERSHEDS AS A FUNCTION OF LAND USE
By

Shawn P. McElmurry

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTORATE OF PHILOSOPHY
Environmental Engineering

2008

ABSTRACT

CHARACTERIZATION OF DISSOLVED ORGANIC CARBON:
ASSESSMENT OF COPPER COMPLEXATION AND EXPORT OF CARBON
FROM WATERSHEDS AS A FUNCTION OF LAND USE
By

Shawn P. McEImurry

Dissolved organic carbon (DOC) is a critical component of freshwater ecosystems
and inﬂuences the transport of many pollutants. The aim of this work is to assess DOC
characteristics, to determine if these characteristics inﬂuence the complexation of copper,
and to identify the inﬂuence of land use on DOC characteristics. A method utilizing
solid-phase extraction is developed to simultaneously quantify DOC fractions and
copper-DOC complexes. Fractions are based on speciﬁc bonding mechanisms (hydrogen,
donor—acceptor, London dispersion, and ionic bonding) thought to be responsible for
stabilizing humic substances in aqueous solutions. This method produces different
fractions for a range of natural DOC and standardized humic materials. Using this
method, complexation constants for copper are derived for individual fractions of DOC
and these are found to be similar to those reported elsewhere for bulk DOC using more
intensive analytical techniques. Additionally, the complexation of copper by high and
low molecular weight fractions of DOC is related to the amount of aromatic structure and
oxygen functional groups present in DOC samples.

The solid-phase extraction method is then used to investigate possible

relationships between DOC characteristics and land use. By collecting surface water

samples from sub-watersheds with unique land uses, statistically significant differences
in DOC characteristics are observed. Sub-watersheds containing agricultural and forested
land are found to produce DOC higher in molecular weight and aromatic structure than
urban catchments. DOC from urban landscapes is found to be more hydrophobic than
from other types of land use. Forested catchments produced DOC that is consistently
different from that derived from the other types of land uses. In addition to land use, a
limited number of environmental variables explain variations in DOC aromaticity,
molecular weight and hydrophobicity. When combined with land use, the amount of solar
radiation, precipitation and water temperature explain up to 80% of the variability
observed in DOC characteristics. Results of this work suggest qualitative differences in
DOC can inﬂuence copper complexation and the nature of DOC may vary Signiﬁcantly

between surface waters as a result of land use.

 

 

Copywrite by
SHAWN P. MCELMURRY
2008

This work is dedicated to my wife, Kim.

 

ACKNOWLEDGEMENTS

I would like to acknowledge some of the people that made this dissertation
possible. First, I would like to thank my wife, Kim, for her unending love and support.
During the 8 years I spent at Michigan State University I have relied on her both
emotionally and ﬁnancially. Without her, I would never have attempted such an
audacious task. I know she will be happy to no longer accompany me to the research
complex to shut off instruments late in the evening! To my family, thank you for your
support. Although many of you found it hard to understand how any degree could take so
long, you always believed in me. I would like to thank the undergraduate students who
helped me collect and analyze samples for this work, particularly Brad Detj en, Jonathan
Long and Matt Witter; this project would have taken far longer without your diligent
assistance. While I have made many ﬁiends during my time in graduate school who will
be with me for the remainder of my life, there are a few that deserve Special mention as
they have aided me in the developing this dissertation: Irfan Aslam, Mathew Parsons,
Chris Saffron, Liyan Song, and Dave Szymanski; thank you for your insight and
companionship.

Finally, I would like to thank my committee — David T. Long, Thomas C. Voice,
Syed A. Hashsham, and Phanikumar S, Mantha - for their assistance in developing this
research. With your assistance I gained an understanding of far more than the narrow
subject of this dissertation. I would especially like to thank the co—chairs of my
committee, David T. Long and Thomas C. Voice, for their detailed tutelage and

incredible patience.

vi

TABLE OF CONTENTS
LIST OF TABLES .................................................................................................... ix
LIST OF FIGURES ................................................................................................... x
ABBREVIATIONS ................................................................................................. xii
CHAPTER 1. INTRODUCTION ............................................................................ 1
References .................................................................................................................... 4
CHAPTER 2. LITERATURE REVIEW ................................................................ 7
Dissolved Organic Carbon ........................................................................................... 7
Land Use and Seasonal Inﬂuences ............................................................................ 12
Dissolved Organic Carbon Characterization ............................................................. 15
References .................................................................................................................. l 8

CHAPTER 3. SIMULTANEOUS QUANTIFICATION OF DISSOLVED
ORGANIC CARBON FRACTIONS AND COPPER COMPLEXATION

USING SOLID PHASE EXTRACTION .............................................................. 36
Abstract ....................................................................................................................... 36
Introduction ................................................................................................................ 38
Materials and Methods ............................................................................................... 40
Results and Discussion ............................................................................................... 49
Conclusions ................................................................................................................ 63
References .................................................................................................................. 80

CHAPTER 4. INFLUENCE OF AROMATICITY ON COPPER

COMPLEXATION BY DISSOLVED ORGANIC CARBON .......................... 87
Abstract ....................................................................................................................... 87
Introduction ................................................................................................................ 89
Materials and Methods ............................................................................................... 92
Results and Discussion ............................................................................................... 95
Conclusions .............................................................................................................. 101
References ................................................................................................................ 1 10

CHAPTER 5. VARIATIONS IN DISSOLVED ORGANIC CARBON

CHARACTERSITICS BASED ON LAND USE ............................................... 115
Abstract ..................................................................................................................... 1 15
Introduction .............................................................................................................. 1 17
Materials and Methods ............................................................................................. 120
Results and Discussion ............................................................................................. 127
Conclusions .............................................................................................................. 135
References ................................................................................................................ 144

CHAPTER 6. INFLUENCE OF ENVIRONMENTAL FACTORS ON

DISSOLVED ORGANIC CARBON CHARACTERISTICS ......................... 150
Abstract ..................................................................................................................... 150
Introduction .............................................................................................................. 152
Materials and Methods ............................................................................................. 155
Results and Discussion ............................................................................................. 160
Conclusions .............................................................................................................. 166
References ................................................................................................................ 174
CHAPTER 7. CONCLUDING REMARKS ....................................................... 181
APPENDIX A. METHODS .................................................................................. 185
Alkalinity Analysis ................................................................................................... 186
Cartridge Preparation ............................................................................................... 190
DOC-Trace Metal Fractionation .............................................................................. 200
Identifying Unique Land Use Micro-Watersheds ................................................... 205
Molecular Weight Characterization ......................................................................... 211
Sample Collection Procedure ................................................................................... 218
Sample Watershed Cleanup Script .......................................................................... 227
Total Organic Carbon Analysis ............................................................................... 230
Ultra-Clean Procedure .............................................................................................. 237
APPENDIX B. SAMPLE LOCATIONS ............................................................. 242
APPENDIX C. DATA ............................................................................................ 250

viii

 

LIST OF TABLES

Table 3-1. Properties of SPE cartridges used to isolate ﬁ'actions of DOC and Cu
complexed with DOC. .............................................................................................. 65

Table 3-2. General water chemistry parameter for Michigan State University
tap water, artiﬁcial solution, and natural Red Cedar River (RCR) water. ............. 66

Table 3-3. Parameters used in PHREEQCi to model Cu-DOC complexation
based on values reported in the literature and found to describe experimental
Observations (optimized) ........................................................................................ 67

Table 3-4. Concentrations of Cu and DOC found to leach from SPE
cartridges used to isolate DOC fractions. .............................................................. 68

Table 3-5. Conditional binding constants and partitioning coefﬁcients for Cu
complexation with DOC fractions (N A—not available). ........................................ 69

Table 3-6. Theoretical (based on values reported in the literature) and
observed (optimized) conditional stability constants produced by PHREEQCi
modeling results for Cu-DOC complexation. ....................................................... 70

Table 4-1. General chemical composition of the artiﬁcial river water solutions
investigated .......................................................................................................... 103

Table 5-1. Number of samples collected from each type of land use. Samples
were grouped according to the land use identiﬁcation (LUID). ......................... 136

Table 5-2. Average water chemistry for different land use (grouped) sampling
locations. .............................................................................................................. 137

Table 6-1. Regression coefﬁcients (B) for the GLMS describing DOC
characteristics ....................................................................................................... 168

Table 6-2. Differences determined by Tukey’s HSD Test in NUVA, MW and
polydispersity resulting from the interaction between land use and parameters

describing weather and water chemistry. ............................................................. 169
Table A-1. Example data for Gran titration. ....................................................... 188
Table A-2. Cartridge ID key ............................................................................... 202

Table A-3. Master conversion ﬁle used to rectify data and the resulting
bounding coordinates ........................................................................................... 207

ix

Table A-4. MW standards for calibration ........................................................... 213
Table A-5. Typical arrangement of calibration standards ................................... 215
Table A-6. Sequence of TOC standards .............................................................. 234
Table B-1. Sample location key .......................................................................... 243
Table B-2. Samples collected .............................................................................. 245
Table C-l. Molecular characterization data ........................................................ 251
Table C-2. DOC fraction data ............................................................................. 257

LIST OF FIGURES

Figure 3-1. The retention of Aldrich Humic Acid in MSU tap water at different

ﬂow rates for each of the SPE cartridges used to fraction DOC .............................. 71
Figure 3-2. Relative amount of donor-acceptor interactions based on the

fraction of DOC retained ....................................................................................... 72
Figure 3-3. Ratio of DOC retained by ionic/hydrophobic mechanisms .................. 73
Figure 3-4. Relative amount of London dispersion forces based on the fraction

of DOC retained ......................................................................................................... 74
Figure 3-5. Relative amount of hydrogen bonding based on the ﬁ'action of

DOC retained ............................................................................................................. 75
Figure 3-6. The fraction of DOC less than 1 kDa in molecular weight. ................. 76
Figure 3-7. The ratio of molecular weight binding site densities ............................ 77

Figure 3-8. The relative amount of Cu retained by SPE cartridges in different
types of solutions ....................................................................................................... 78

Figure 3-9. The fraction of freeCu2+, Cu—hydroxide complexes, Cu-carbonate
complexes, and Cu—organic complexes following (a) the ﬁrst modeling step

(equilibrium with C02, without organic complexation) and (b) the second
modeling step (with optimized organic complexation) using PI-IREEQCi .............. 79

Figure 4-1. The LS ratio and the MW BSD ratio for Cu and a set of
standardized DOC samples ...................................................................................... 104

Figure 4-2. The correlation between MW BSD ratio for Cu and aromaticity ...... 105

Figure 4-3. The MW BSD ratio as a function of the concentration of various
elements: (a) oxygen, (b) nitrogen, (c) sulfur and (d) phosphorus. ........................ 106

Figure 4-4. Copper removed from Cu-DOC complexes by a Chelex 100 resin
as a function (a) aromatic structure and (b) oxygen content of DOC .................... 107

Figure 4-5. Organic carbon retained by ionic mechanisms greater than 1 kDa in

size versus the amount of (a) heteroaliphatic and (b) aromatic structure in SOC
and DOC ................................................................................................................... 108

xi

Figure 4-6. The inﬂuence of aromatic structure on hydrophobicity, measured as
the percentage of organic carbon retained by hydrophobic mechanisms, DOC
and SOC ................................................................................................................... 109

Figure 5-1. Box plots of the total concentration of DOC (mg L") in surface
water rtmoff from catchments with speciﬁc types of land cover ........................... 138

Figure 5-2. The average molecular weight and polydispersity of DOC in
Stuface water runoff from catchments with speciﬁc types of land cover ............... 139

Figure 5-3. Box plots of the NUVA (L mgC m '1) in surface water runoff from
catchments with speciﬁc types of land cover .......................................................... 140

Figure 5-4. The percentage of DOC retained on extended hydrophobic,
hydrophobic and H-bonding cartridges for samples from catchments with
speciﬁc types of land cover ..................................................................................... 141

Figure 5-5. NUVA (L mgC m '1) versus molecular weight (Da) for samples
collected from catchments with speciﬁc and mixed land uses ............................... 142

Figure 5-6. The amount of DOC retained on the hydrophobic cartridge versus

NUVA (L mgC m '1) for DOC collected from catchments with speciﬁc and
mixed land uses: agricultural, forested, golf course, mixed, Michigan State
University, automobile parking lot and urban ......................................................... 143

Figure 6-1. The inﬂuence of solar radiation, measured as logio weekly mean
solar ﬂux density (W mg), on the aromaticity, measured as NUVA (L mgC

m ‘1) .......................................................................................................................... 179
Figure 6-2. Ability of generalized linear models (predicted) used to describe

observed DOC characteristics ................................................................................. 171
Figure 6-3. Seasonal inﬂuence on the hydrophobicity of DOC (%). .................... 172
Figure 6-4. The inﬂuence of conductivity on the MW of DOC. ........................... 173
Figure A-l. Attachment of adapter to PTFE tubing ............................................... 192

Figure A-2. Arrangement of syringes and cartridges on for sample collection 201

Figure A-3. Differences in land use for traditional and new watershed
boundaries. ............................................................................................................... 205

xii

ABBREVIATIONS

AG -— agricultural

AHA — Aldrich HA

au — absorbance units

Cu — copper

Da — Daltons (grams mole'l)

DDI — deionized distilled water

DOC — dissolved organic carbon

F -— ﬂuoride

FA — ﬁilvic acid

FOR — forested

GC — golf course

GLM — general linear model

HA -— humic acid

I-IMW —— high molecular weight

HSD -— Tukey’s honestly signiﬁcant difference test
IHSS — International Humic Substances Society
LMW- low molecular weight

LS - ligand size

M — molarity (moles liter'l)

MSU — Michigan State University

MW — molecular weight

MWBSDR - Molecular Weight Binding Site Density Ratio
NOM — natural organic matter (RO isolated DOC)
NRN — Nordic Reservoir NOM

NUVA -— normalized ultraviolet absorbance at 280nm
PL - parking lot

RCR —- Red Cedar River

RO — reverse osmosis

SPE — solid-phase extraction

SRF — Suwannee River FA

SRH — Suwannee River HA

SRN - Suwannee River NOM

URB — urban

xiii

CHAPTER 1
INTRODUCTION

Dissolved organic carbon (DOC) is known to alter the fate of both organic (Chin,
Aiken et al. 1997; Raber, Kogel-Knabner et al. 1998; Burkhard 2000; Lee, Kuo et al.
2003) and inorganic pollutants (Kimball, Callender et al. 1995; Santschi, Lenhart et al.
1997; Linnik 2003) and has been shown to inﬂuence redox reactions in surface waters
(Cory and McKnight 2005). In addition to inﬂuencing geochemistry, DOC plays a
Si gniﬁcant role in aquatic ecology as the primary food source for heterotrophic bacteria in
surface water systems (Ghabbour and Davies 2004; Perga, Kainz et al. 2006). Despite
being a ﬁmdamental component of surface water ecosystems, scientiﬁc understanding of
DOC evolution and behavior is limited.

Previous research has primarily focused on DOC quantity, rather than the quality
(Hope, Billett et al. 1994). Hydrologic conditions and land cover are two factors that have
been shown to inﬂuence the amount of terrestrial DOC exported from the watershed
system to surface waters (Moore and Jackson 1989; Dillon and Molot 1997; Gergel,
Turner et al. 1999; McGlynn and McDonnell 2003). However, concentrations of DOC
alone are unable to explain DOC-pollutant interactions (Lee, Gan et a1. 2003). The
composition, or quality, of DOC also inﬂuences the biological growth (Rosenstock,
Zwisler et al. 2005; McCallister, Bauer et al. 2006; Christian and Lind 2007). Based on
observed variations in the chemical and biological behavior of DOC, there is a need to

assess the variability of DOC characteristics.

The focus of this work is on variations in DOC characteristics, the inﬂuence of
DOC characteristics on copper complexation and the parameters related to DOC
characteristics. Due to the importance of land cover and catchment characteristics in
inﬂuencing the release of DOC from watershed surfaces to aquatic systems, the
hypothesis that DOC characteristics are inﬂuenced by land use and other environmental
factors is investigated. This work was part of a larger project, Michigan State University -
Watershed Action Through Education and Research (MSU-WATER), which, among
other objectives, aimed to determine the inﬂuence of land use on surface water quality.

Results from this work are signiﬁcant on many levels. At the molecular level,
improved understanding of how DOC structure inﬂuence DOC behavior (i.e.
complexation of trace-metals) allows for a more accurate description of chemical
interactions in surface water systems. This provides an improved description of the fate
and transport of trace-metals and organic pollutants. Furthermore, by determining
environmental variables related to DOC characteristics, possible processes responsible
for DOC formation and transformation can be identiﬁed. At a systems level, results also
help to complete gaps in the carbon cycle. By ﬁlling these gaps it is possible to assess
interactions beyond chemistry; ecology, geology, biology, etc. Overall, this research
provides a better understanding of DOC characteristics and identiﬁes parameters related
to these characteristics.

Each proceeding chapter presents original research addressing gaps in
understanding related to DOC; with the exception of the Chapter 2, which provides a
brief review of pertinent literature that forms the foundation of this work, and Chapter 7,

which brieﬂy highlights some of the important results of this work, describes some of the

 

implications of this work and suggests future research needs. Chapter 3 presents a novel
method utilizing solid phase extraction (SPE) for simultaneously quantifying DOC
characteristics as well as trace metal complexation. In Chapter 4, this method is used to
assess the inﬂuence of molecular structure, speciﬁcally aromaticity, on the complexation
Of copper. In Chapter 5, inﬂuence of land use on DOC characteristics is evaluated.
Chapter 6 explores factors other than land use, such as weather and solution chemistry,

which play a role in determining DOC quality.

References

Burkhard, L. P. (2000). "Estimating dissolved organic carbon partition coefﬁcients for
nonionic organic chemicals." Environmental Science & Technology 34(22):
4663-4668.

Chin, Y. P., G. R. Aiken, et al. (1997). "Binding of pyrene to aquatic and commercial

humic substances: The role of molecular weight and aromaticity." Environmental

 

Science & Technology 31(6): 1630-1635.

Christian, B. W. and O. T. Lind (2007). "Multiple carbon substrate utilization by bacteria
at the sediment-water interface: seasonal patterns in a stratiﬁed eutrOphic
reservoir." Hydrobiologja 586: 43-56.

Cory, R. M. and D. M. McKnight (2005). "Fluorescence Spectroscopy Reveals
Ubiquitous Presence of Oxidized and Reduced Quinones in Dissolved Organic
Matter." Environmental Science & Technology 39(21): 8142-8149.

Dillon, P. J. and L. A. Molot (1997). "Effect of landscape form on export of dissolved
organic carbon, iron, and phosphorus from forested stream catchments." Mgr
Resources Research 33(11): 2591-2600.

Gergel, S. E., M. G. Turner, et al. (1999). "Dissolved organic carbon as an indicator of
the scale of watershed inﬂuence on lakes and rivers." Ecological Applications
9(4): 1377-1390.

Ghabbour, E. and G. Davies, Eds. (2004). Humic Substances: Nature's most versatile
materials. New York, Taylor and Francis.

Hope, D., M. F. Billett, et al. (1994). "A Review of the Export Of Carbon in River Water -

Fluxes and Processes." Environmental Pollution 84(3): 301-324.

Kimball, B. A., E. Callender, et al. (1995). "Effects of Colloids on Metal Transport in a
River Receiving Acid-Mine Drainage, Upper Arkansas River, Colorado, USA."
Applied Geochemistry 10(3): 285-306.

Lee, C. L., L. J. Kuo, et al. (2003). "Effects of ionic strength on the binding of
phenanthrene and pyrene to humic substances: three-stage variation model."
Water Research 37(17): 425 0-4258.

Lee, S., J. Gan, et a1. (2003). "Evaluation of Kd Underestimation Using Solid Phase
Microextraction." Environmental Science & Technology 37(24): 5597-5602.

Linnik, P. N. (2003). "Complexation as the most important factor in the fate and transport
Of heavy metals in the Dnieper water bodies." Analytical and Bioanalytical
Chemism 376(3): 405-412.

McCallister, S. L., J. E. Bauer, et al. (2006). "Bioreactivity of estuarine dissolved organic
matter: A combined geochemical and microbiological approach." Limnology and
Oceanography 51(1): 94-100.

McGlynn, B. L. and J. J. McDonnell (2003). "Role of discrete landscape units in
controlling catchment dissolved organic carbon dynamics." Water Resources
Research 39(4): SWC 3 1-18.

Moore, T. R. and R. J. Jackson (1989). "Dynamics of Dissolved Organic-Carbon in
Forested and Disturbed Catchments, Westland, New-Zealand .2. Larry River."
_Vﬂter Resources Resear_ch 25(6): 1331-1339.

Perga, M. E., M. Kainz, et al. (2006). "Carbon pathways to zooplankton: insights from
the combined use of stable isotope and fatty acid biomarkers." Freshwater

Biology 51(11): 2041-2051.

Raber, B., I. Kogel-Knabner, et a1. (1998). "Partitioning of polycyclic aromatic
hydrocarbons to dissolved organic matter from different soils." Chemosphere
36(1): 79-97.

Rosenstock, B., W. Zwisler, et al. (2005). "Bacterial consumption of humic and non-
humic low and high molecular weight DOM and the effect of solar irradiation on
the turnover of labile DOM in the Southern Ocean." Microbial Ecology 50(1): 90-
101.

Santschi, P. H., J. J. Lenhart, et al. (1997). "Heterogeneous processes affecting trace
contaminant distribution in estuaries: The role of natural organic matter." Mgr};

Chemistg 58(1-2): 99-125.

CHAPTER 2

LITERATURE REVIEW

Dissolved Organic Carbon

Dissolved organic carbon (DOC) is a heterogeneous collection of diverse,
relatively low molecular weight, organic constituents that aggregate through hydrophobic
and hydrogen bonding (Sutton and Sposito 2005). DOC is operationally deﬁned as the
amount of organic carbon passed through a 0.45 um pore—size ﬁlter (Leenheer and Croue
2003). Typically, greater than 90% of the organic carbon present in surface waters is
DOC (Thurman 1985). The composition of DOC includes biologically active compounds
at various stages of microbial and physico-chemical transformation (Ghabbour and
Davies 2004; Wickland, Neff et al. 2007). While there is not a consensus on the dominant
source of DOC to surface water systems (Hemond 1990), photosynthetic plants and
autotrophic bacteria are identiﬁed as two major sources (Hedges 1981; Thurman 1985).
Based on the relative inﬂuence of terrestrial and aquatic processes, DOC is often
classiﬁed as terrestrial, derived from terrestrial sources, or aquatic, derived ﬁ'om aqueous
biotic processes. Allochthonous, or terrestrial, sources of DOC include plant exudates,
organic compounds resulting from litter decomposition, the leaching of soil organic
matter, as well as enzymes and biomass produced by microbial grth and decay
(Thurman 1985; Wickland, Neff et al. 2007). Autochthonous sources of DOC are the
result of the exudation and excretion of biomolecules and the decay of organisms from all
trophic levels (Tranvik 1993). The main source of DOC in aquatic systems are aquatic

plants, bacteria, phytoplankton, algae, invertebrate excretion and micro-ﬂora associated

with detritus (Otsuki and Wetzel 1974; Naiman, Melillo et al. 1987; Wetzel 1992;
McKnight, Andrews et al. 1994; Kaiser, Arscott et al. 2004). In both systems, terrestrial
and aquatic, DOC under goes signiﬁcant biotic and abiotic transformations (Hedges
1981)

The amount of DOC exported ﬁom temperate and boreal catchments is 10 to 100
kgC ha-l yr-l (Hope, Billett et al. 1994). In most streams, terrestrial sources of DOC are
signiﬁcant (Fisher and Likens 1972). The inﬂuence of terrestrial sources are supported by
carbon isotope studies that indicated DOC in surface water systems are typically less than
40 years old, while carbon in ground waters and soil organic matter are substantially
older (Thurman and Malcolm 1981; Schiff, Aravena et al. 1990). For instance, soil
organic matter has an average age of 550-700 years (Campbell, Paul et al. 1967). Based
solely on age, the carbon present in groundwater and soil is likely to be highly degraded
and altered. DOC on the other hand is relatively less degraded.

Catchment hydrology inﬂuences the export of DOC from watersheds (Fisher and
Likens 1972; Schlesinger and Melack 1981). The concentration and type of DOC
discharged to surface water bodies varies with different stages of the stormwater
hydrograph (Schlesinger and Melack 1981; McGlynn and McDonnell 2003; Hood,
Williams et al. 2005). The ﬂow path which water travels through the catchment before
being discharged to a surface water body is very important in determining organic carbon
discharges (Edwards and Cresser 1987; Billett and Cresser 1992). By avoiding potential
sorption sites present in deeper soils, laboratory experiments have shown that interﬂow
produces more DOC than water transported through deep soils (Edwards 1984; Moore

and Jackson 1989).

Absorption of DOC to soil plays a signiﬁcant role in regulating the amount and
composition of DOC that is exported to surface waters (McDowell and Wood 1984;
Hope, Billett et al. 1994). Concentrations and ﬂuxes of DOC from forested and grassland
soils has been found to decrease with increasing soil depth (Dawson, Ugolini et al. 1978;
Sollins and McCorison 1981; McDowell and Wood 1984; Cronan and Aiken 1985;
Homung, Stevens et al. 1986; McDowell and Likens 1988; Moore and Jackson 1989;
Meier, Chin et a1. 2004). This decrease in DOC concentration is attributed to the sorption
of organic acids to soils (McDowell and Wood 1984; Cronan and Aiken 1985; Thurman
1985). Soils also exhibit chromatographic retention of hydrophobic fractions of DOC
while preferentially releasing hydrophilic fractions when ﬂushed (J ardine, Weber et al.
1989; Kaiser and Zech 1998; Meier, Chin et al. 2004). DOC with indirect ﬂow paths
through leaf litter and soil is likely to be stripped of hydrophobic DOC (Kawahigashi,
Kaiser et al. 2006).

Generally, biotic processes in riverine environments result in a marked decrease
in DOC transported downstream, particularly in headwater streams (Wallace, Ross et al.
1982; Naiman, Melillo et al. 1987). Based on this observation the “size-reactivity
continuum model” was proposed by Amon and Benner (1996). The model hypothesized
that aquatic organisms are arranged in an ordered continuum based on the selective
degradation of DOC from terrestrial sources, regularly decreasing the MW of DOC at
every step. In support of this model, between 12 and 44% of the DOC released from
forest ﬂoors is degraded microbially (Yano, McDowell et al. 1998). Heterotrophic
bacteria preferentially consume carbohydrates, proteins and other nutrient rich sources

leaving behind aromatic structures (Marschner and Kalbitz 2003). While conceptually

relevant, this model is questionable for some surface water systems (Winterboum,
Rounick et al. 1981). Rather than a stepwise degradation, multiple biological pathways
are more likely, including biotic production of DOC from autotrophs (Pomeroy 1974).
Another important biota are algae, which produce DOC high in MW and low in
aromaticity (McKnight, Andrews et al. 1994). DOC from terrestrial sources and algae
serves as a primary food sources for heterotrophic microorganisms (McCallister, Bauer et
al. 2006).

DOC biodegradability is related to DOC elemental composition (Sun, Perdue et
a1. 1997) and aromatic structure (Gilbert 1988; Zoungrana, Desjardins et al. 1998).
Aromaticity, molecular weight and hydrophobicity are all indictors of bioavailability and
previous transformation processes (Cabaniss, Zhou et al. 2000; Marschner and Kalbitz
2003; Wickland, Neff et al. 2007). Hydrophobic fractions of DOC are less accessible to
microorganisms and are assumed to be less susceptible to microbial decomposition than
hydrophilic components (Qualls and Haines 1992; Piccolo 1998).

There are four main physico-chemical processes of importance to DOC in surface
water systems: adsorption, precipitation, oxidation-reduction reactions and complexation
(Thurman 1985). Adsorption of DOC by iron and aluminum oxides present in immobile
or settle able solids plays a signiﬁcant role in removing DOC from the water column
(Dahm 1981). This is also likely one of the main processes responsible for the
sequestration of DOC in soil regimes (McDowell 1985). As a result, the concentration of
DOC and iron and aluminum in upstream segments of surface waters are related
(J ohnsen, Martinsen et al. 1987). DOC precipitates from solution when the pH of water

decreases or the ionic strength Of solution increases, particularly for polyvalent cations

10

(Hope, Billett et al. 1994). While this may occur at the conﬂuence of freshwater streams
(Thurman 1985), estuaries are more signiﬁcant (Sholkovitz 1976; Sholkovitz, Boyle et al.
1978; Mantoura and Woodward 1983). Oxidation of DOC can be induced
photochemically by solar radiation (Daﬂrer and Wangersky 2002; Waiser and Robarts
2004; Lou and Xie 2006; Shiller, Duan et al. 2006) or chemically (Richey, Brock et al.
1980; Thurman 1985). While it is possible the mobility of DOC can be altered via metal
complexation (McDowell 1985; Moore, Desouza et al. 1992; Wang, Chen et al. 1997),
complexation greatly alters the fate and transport of many trace metals (Florence 1982).
The inﬂuence of DOC on the behavior of trace metals is signiﬁcant (Florence
1982). Partitioning coefﬁcients between dissolved trace metals and colloids have been
found to be greater than for particles (Wen, Santschi et al. 1999; Shafer, Hoffmann et al.
2004). DOC has been found to be the dominant factor in determining the speciation Of
metals (Mn, Cu, Zn, Pb, Cr, Cd) in some surface waters (Linnik 2003) and complexation
greatly alters their mobility (Drever 1997). Irving-Williams Order describes the
adsorption tendency of divalent transition metals by hydrous oxides
(Cu>Zn~Ni>Co>Fe>Mn>Ca) and overall adsorption tendencies are suggested
Pb>Cu>Cd>Zn>Ni>Ca (Langmuir 1997). Similar orders have been described for metal
complexation with humic acid (Kemdorff and Schnitzer 1980). However, humic
substances are, by deﬁnition, more diverse than speciﬁc binding sites, such as hydrous
oxides. Understanding of DOC-trace metal complexation would be greatly enhanced by
simultaneously measuring metal complexation and assessing the characteristics of the

organic ligand.

11

Land Use and Seasonal Inﬂuences

The inﬂuence of land use on surface water chemistry is well documented. (Rhodes,
Newton et al. 2001; Wayland, Long et al. 2003; Xie, Norra et al. 2005). Urban and
agricultural land uses are known to produce non-point source pollutants, such as nitrogen
and phosphorus which result in signiﬁcant eutrophication of rivers, lakes, and coastal
waters (Carpenter, Caraco et al. 1998). In northern climates, road salt and other de-icing
chemicals alter surface water chemistry (Amrhein, Mosher et al. 1993; Buttle and
Labadia 1999; Mason, Norton et al. 1999). The loading rates to surface waters of many
trace metals and organic contaminants, such as pesticides, to surface waters are
dependent on land use (Stangroom, Collins et al. 1998; Gardner and Carey 2004; Leu,
Singer et al. 2004). While DOC itself is not a pollutant, through complexation, ion
exchange and sorption reactions, it can alter the fate and transport of many pollutants of
interest (Cho, Park et al. 2002; Merritt and Erich 2003; Yamamoto and Liljestrand 2003).

The concentration of DOC in surface water is found to vary spatially depending on
the ﬂux from terrestrial sources (Eckard, Hemes et a1. 2007). Alterations to land use are
also known to modify the amount of DOC exported from the watershed to surface waters
(Bemardes, Martinelli et al. 2004; Stedmon, Markager et al. 2006). The amount of
wetlands correlates highly with the ﬂux of DOC from watersheds (Canham, Pace et al.
2004; Frost, Larson et al. 2006). However, most investigations into the inﬂuence of land
use, primarily agricultural and forested (Kalbitz, Solinger et al. 2000), on DOC tend to
focus on production, rather than DOC composition or quality (Richey, Brock et al. 1980).
Due to the close relationship between DOC and biotic processes, Hedges (1980)

proposed unique chemical differences exist for DOC from different landscapes.

12

Variations in parameters indicative of aromaticity have been related to watershed metrics
(Larson, Frost et al. 2007) and physico-chemical characteristics of DOC are found to vary
(Thacker, Tipping et al. 2005). This is important because functional and structural
characteristics of DOC (1) alter the transport of pollutants (Chiou 2002) and (2) impact
aquatic food webs (Qualls and Haines 1992; Lennon and Pfaff 2005).

Part of the difﬁculty in assessing the relationship between land use and DOC is
separating this inﬂuence from other processes, especially in larger systems where
competing processes are inevitable. Additionally, in large watersheds where the transport
of DOC is slow and ﬂow paths are long, degradation weakens relationships between
DOC composition and terrestrial processes (Frost, Larson et al. 2006). However, using
small scale watersheds it is possible to show differences in DOC quality that are not
visible in larger watersheds (Dalzell, F illey et al. 2007). Samples collected from lower
order stream segments, the upstream sections of streams that are close to the stream
headwater, Show progressively stronger correlations to changes in land use (Bemardes,
Martinelli et al. 2004).

When attempting to isolate land use at the micro-watershed scale, anthropogenic
alterations to hydrology must be considered. In agricultural areas, drain tiles alter
traditional ﬂow paths and can enhance the transport of organic contaminants to surface
waters (Leu, Singer et a1. 2004). In urban areas, impervious surfaces and sewer networks
greatly alter watershed hydrology by increasing the amount and speed at which
stormwater is transported to surface water systems (Sheeder, Ross et a1. 2002). As a
result, connections formed between urban catchments and surface waters through storm

sewer networks are strongly correlated to pollutant and DOC loading rates (Hatt, Fletcher

13

et al. 2004). Furthermore, the delineation of micro-watersheds in urban areas is not based
solely on topography because inlets to storm sewers can artiﬁcially extend watershed
boundaries. Taking these factors into consideration and collecting DOC samples directly
from micro-watersheds with unique land uses are required to effectively evaluate the
inﬂuence of land use on DOC characteristics.

In addition to variations attributed to land use, seasonal variations in DOC
characteristics have also been observed (Tipping and Woof 1983; Lindell, Graneli et a1.
2000; Saliot, Derieux et al. 2002; Meier, Chin et al. 2004; Porcal, Hejzlar et al. 2004;
Yano, Lajtha et al. 2004; Guo and Macdonald 2006; Minor, Simjouw et al. 2006).
Seasonal variations in DOC characteristics from terrestrial sources inﬂuence
degradability (Porcal, Hejzlar et al. 2004).The amount of aromatic moieties present in
DOC has been found to be seasonally dependent (Kortelainen 1993; Molot and Dillon
1997; Larson, Frost et al. 2007; Rodriguez-Zirfriga, Milori et al. 2008). Some of the
seasonal variations are also reported to be dependent on land use (Aitkenhead-Peterson,
Smart et al. 2007).

Despite some connections between land use and seasonal inﬂuences, factors
inﬂuencing DOC characteristics, utilization and microbial transformation are poorly
understood (Christian and Lind 2007). The amount of precipitation and the extent of
solar radiation are two factors that inﬂuence the amount of light absorbed by DOC
(Curtis and Schindler 1997; Molot and Dillon 1997 ; Lindell, Graneli et al. 2000; Reche
and Pace 2002). Despite the variability observed in terrestrial sources of DOC, the factors
controlling the chemical characteristics of DOC remains largely unknown (Canham, Pace

et al. 2004).

14

Dissolved Organic Carbon Characterization

Early efforts to characterize DOC were based on procedures developed for the
soil organic matter (Stevenson 1982). The humic fractions of soil organic matter consist
of: humic acids (HAS), which are the fraction of organic matter soluble under alkaline
conditions but not acidic conditions (generally pH < 2); fulvic acids (F As), the ﬁaction
soluble under all pH conditions; and humin, the insoluble fraction of humic substances.
Consistent with soil chemistry, methods were developed using XAD resins to isolate HA
and FA ﬁ'actions of dissolved organic matter (Thurman and Malcolm 1981). While HA
and FA ﬁactions do represent isolates with similar chemical and physical properties,
these properties are selected by the method of isolation and do not represent distinct types
of organic molecules (Hayes, MacCarthy et al. 1989; Sutton and Sposito 2005).

An array of analytical techniques have been used to characterize DOC based on
chemical structure, molecular weight, and fractions isolated using different SPE media
(McDonald, Bishop et a1. 2004). Excitation-emission matrix (McKnight, Boyer et al.
2001; Chen, Westerhoff et al. 2003; Cory and McKnight 2005) and nuclear magnetic
resonance spectrometry (Gauthier, Seitz et al. 1987; Grasso, Chin et a1. 1990; Swift,
Leonard et al. 1992; Christl and Kretzschmar 2001; Kaiser, Simpson et a1. 2003) have
also been used to determine structural features of DOC. While NMR based techniques
can elucidate complex chemical structures, the application of this technology is difﬁcult
and expensive (Leenheer and Croue 2003). Other less costly and intensive techniques
have been shown to describe structurally unique forms of DOC (Hayes and Swift 1978;

Swiﬁ, Leonard et al. 1992; Chin, Aiken et al. 1994; Peuravuori and Pihlaja 1997; Everett,

15

Chin et al. 1999; Pelekani, Newcombe et al. 1999; Christl, Knicker et al. 2000; Muller,
Schmitt et al. 2000; Croue, Benedetti et al. 2003).

Protocols have been developed using SPE to separate bulk DOC into fractions
with signiﬁcantly different chemical structures and similar chemical properties (Croue,
Benedetti et al. 2003; Janos 2003). One of the more extensive fractionation methods
derives six groups of DOC (Aiken, McKnight et al. 1992) that have been shown to differ
signiﬁcantly depending on the sample origin (Irnai, Fukushima et al. 2001).

One of the most common methods used to fraction DOC into hydrophobic and
hydIOphilic fractions employs the use of XAD-8 and XAD-4 resins (Thurman and
Malcohn 1981). Like all analytical techniques, there are some negative attributes for
these methods. Isolation of DOC by XAD resins typically requires a Signiﬁcant change in
pH. Changes in pH may result in hydrolysis and alter the molecular conﬁguration of
DOC through bond breakage (Mace, Lin et al. 2001). Large humic molecules (>30kDa)
are not absorbed by XAD resins (Town and Powell 1993). Additionally, XAD-8 resins
are primarily marketed for industrial purposes. As a result, production of these resins are
performed with low quality solvents resulting in signiﬁcant impurities (Thurman, Ferrer
et al. 2001). In order to remove impurities present in XAD-8 resins, extensive puriﬁcatiOn
procedures are required (Thurman and Malcolm 1981; Louchouarn, Opsahl et al. 2000).

Recently, methods utilizing C13 SPE media to isolate DOC have received
increased use. C18 preferentially retains aromatic moieties and the chemical composition
of DOC retained differs signiﬁcantly from DOC isolated using ultra-ﬁltration (Schwede-

Thomas, Chin et al. 2005). DOC retained by C13 appears enriched in aromatic compounds

16

while ultra-ﬁltration produces DOC enriched in short chained polysaccharides along with
amino sugars (Simjouw, Minor et al. 2005).

Most SPE reactions are rapid, with the limiting factor being diffusion (Wu and
Gschwend 1986), allowing SPE media to behave similar to partitioning phases.
Theoretically, SPE media provides chemical binding sites in which individual
compounds can achieve equilibrium at a lower thermodynamic energy state. In reality,
there are two primary factors which prevent true equilibrium from being achieved:
competition reactions and mass transfer limitations. In order to ensure competing
reactions do not interfere with thermodynamic equilibrium, a large ratio of potential
binding sites to chemical constituents should be maintained. This is achieved by using a
high surface area media and utilizing appropriate quantities of SPE media relative to the
amount of analyte to be retained. Styrene-divinylbenzene (SDVB) cartridges are capable
of providing high surface areas; potentially double that of common C13 media. Mass
transfer limitations are combated by using a SPE media with high number of theoretical
plates (high surface area and low particle size) and keeping ﬂow rates sufﬁciently low to
allow difﬁrsion. With extremely high surface area and low particle size of modern SPE
material, DOC extraction efﬁciencies for some SPE cartridges are reported to be nearly

independent of ﬂow rates.

17

References
Aiken, G. R., D. M. McKnight, et al. (1992). "Isolation of Hydrophilic Organic-Acids

from Water Using Nonionic Macroporous Resins." Organic Geochemistry 18(4):

 

567-5 73.

Aitkenhead-Peterson, J. A., R. P. Smart, et al. (2007). "Spatial and temporal variation of
dissolved organic carbon export from gauged and ungauged watersheds of Dee
Valley, Scotland: Effect Of land cover and C : N." Water Resources Research
43(5).

Amon, R. M. W. and R. Benner (1996). "Bacterial utilization of different size classes of
dissolved organic matter." Limnolggy and Oceanography 41(1): 41-51.

Amrhein, C., P. A. Mosher, et al. (1993). "Colloid-Assisted Transport of Trace-Metals in
Roadside Soils Receiving Deicing Salts." Soil Science Society of AmeﬁcLJomgl
57(5): 1212-1217.

Bemardes, M. C., L. A. Martinelli, et al. (2004). "Riverine organic matter composition as
a function of land use changes, Southwest Amazon." Ecological Applicﬁons
14(4): 8263-8279.

Billett, M. F. and M. S. Cresser (1992). "Predicting Stream-Water Quality Using
Catchment and Soil Chemical Characteristics." Environmean Pollution 77(2-3):
263-268.

Buttle, J. M. and C. F. Labadia (1999). "Deicing salt accumulation and loss in highway

snowbanks." Journal of Environmental Quality 28(1): 155-164.

18

Cabaniss, S. E., Q. Zhou, et al. (2000). "A Log-Norrnal Distribution Model for the
Molecular Weight of Aquatic Fulvic Acids." Environmental Science &
Technology 34(6): 1 103-1109.

Campbell, C. A., E. A. Paul, et al. (1967). "Applicability of Carbon-Dating Method of
Analysis to Soil Humus Studies." Soil Science 104(3): 217-&.

Canharn, C. D., M. L. Pace, et al. (2004). "A Spatially explicit watershed-scale analysis of
dissolved organic carbon in Adirondack lakes." Ecological Applications 14(3):
839-854.

Carpenter, S. R., N. F . Caraco, et a1. (1998). "Nonpoint pollution of surface waters with
phosphorus and nitrogen." Ecological Applications 8(3): 559-568.

Chen, W., P. Westerhoff, et al. (2003). "Fluorescence excitation - Emission matrix
regional integration to quantify spectra for dissolved organic matter."
Environmental Science & Technology 37(24): 5701-5710.

Chin, Y. P., G. Aiken, et al. (1994). "Molecular-Weight, Polydispersity, and
Spectroscopic Properties of Aquatic Humic Substances." Environmentaﬁcience
& Technology 28(11): 1853-1858.

Chiou, C. T. (2002). Partition and Adsorption of Organic Contaminants in Environmental
gym. Hoboken, NJ, Wiley-Interscience.

Cho, H. H., J. W. Park, et al. (2002). "Effect of molecular structures on the solubility
enhancement of hydrophobic organic compounds by environmental amphiphiles."

Environmental Toxicology and Chemistry 21(5): 999-1003.

19

Christian, B. W. and O. T. Lind (2007). "Multiple carbon substrate utilization by bacteria
at the sediment-water interface: seasonal patterns in a stratiﬁed eutrophic

reservoir." Hydrobiologia 586: 43—56.

 

Christl, 1., H. Knicker, et al. (2000). "Chemical heterogeneity of humic substances:
characterization of size fractions obtained by hollow-ﬁbre ultraﬁltration."
European Journal of Soil Science 51(4): 617-625.

Christl, I. and R. Kretzschmar (2001). "Relating ion binding by fulvic and humic acids to

chemical composition and molecular size. 1. Proton binding." Environmental

 

Science & Technology 35(12): 2505-2511.

Cory, R. M. and D. M. McKnight (2005). "Fluorescence Spectroscopy Reveals
Ubiquitous Presence of Oxidized and Reduced Quinones in Dissolved Organic
Matter." Environmental Science & Technology 39(21): 8142-8149.

Cronan, C. S. and G. R. Aiken (1985). "Chemistry and Transport of Soluble Humic
Substances in Forested Watersheds of the Adirondack Park, New-York."
Geochimica Et Cosmochimica Acta 49(8): 1697-1705.

Croue, J. P., M. F. Benedetti, et al. (2003). "Characterization and copper binding of
humic and nonhumic organic matter isolated from the South Platte River:
Evidence for the presence of nitrogenous binding site." Environmental Science &
Technology 37(2): 328-336.

Curtis, P. J. and D. W. Schindler (1997). "Hydrologic control of dissolved organic matter

in low-order Precambrian Shield Lakes." Biogeochemistry 36(1): 125-138.

20

Daiher, E. V. and P. J. Wangersky (2002). "A brief overview of modern directions in
marine DOC studies - Part II - Recent progress in marine DOC studies." Ml
of Environmental Monitoring 4(1): 55-69.

Dahm, C. N. (1981). "Pathways and Mechanisms for Removal of Dissolved Organic-
Carbon from Leaf Leachate in Streams." Canzgian Journal of Fisheries and
Aquatic Sciences 38(1): 68-76.

Dalzell, B. J ., T. R. Filley, et al. (2007). "The role of hydrology in annual organic carbon
loads and terrestrial organic matter export from a midwestem agricultural
watershed." Geochimica Et Cosmochimica Acta 71(6): 1448-1462.

Dawson, H. J ., F. C. Ugolini, et al. (1978). "Role of Soluble Organics in Soil Processes of
a Podzol, Central Cascades, Washington." Soil Science 126(5): 290-296.

Drever, J. (1997). The Geochemistry of Natural Waters: Surface and Groundwate;
Environments. Upper Saddle River, NJ, Prentice Hall.

Eckard, R. S., P. J. Hemes, et al. (2007). "Landscape scale controls on the vascular plant
component of dissolved organic carbon across a freshwater delta." Geochimica Et
Cosmochimica Acta 71 (24): 5968-5984.

Edwards, A. C. (1984). Some factors inﬂuencing elemental mobilities in an upland
catchment in the Grarnpian Region. Computing Science. Aberdeen University of
Aberdeen. PhD.

Edwards, A. C. and M. S. Cresser (1987). "Relationships between Ultraviolet

Absorbency and Total Organic-Carbon in 2 Upland Catchments." Water Research

 

21(1): 49-56.

21

Everett, C. R., Y. P. Chin, et al. (1999). "High-pressure size exclusion chromatography
analysis of dissolved organic matter isolated by tangential-ﬂow ultraﬁltration."
Limnology and Oceanography 44(5): 1316-1322.

Fisher, S. G. and G. E. Likens (1972). "Stream Ecosystem - Organic Energy Budget."
Bioscience 22(1): 33-&.

Florence, T. M. (1982). "The Speciation of Trace-Elements in Waters." Ialarlta 29(5):
345-364.

Frost, P. C., J. H. Larson, et al. (2006). "Landscape predictors of stream dissolved organic
matter concentration and physicochemistry in a Lake Superior river watershed."
Aquatic Sciences 68(1): 40-51.

Gardner, C. B. and A. E. Carey (2004). "Trace metal and major ion inputs into the
Olentangy River from an urban storm sewer." Environmental Science &
Technology 38(20): 5319-5326.

Gauthier, T. D., W. R. Seitz, et a1. (1987). "Effects of Structural and Compositional
Variations of Dissolved Humic Materials on Pyrene Koc Values." Environmental
Science & Technology 21(3): 243-248.

Ghabbour, E. and G. Davies, Eds. (2004). Humic Substances: Nature's most versatile
materials. New York, Taylor and Francis.

Gilbert, E. (1988). "Biodegradability of Ozonation Products as a Function of Cod and
Doc Elimination by the Example of Humic Acids." Water Research 22(1): 123-

126.

22

Grasso, D., Y. P. Chin, et al. (1990). "Structural and Behavioral-Characteristics of a
Commercial Humic-Acid and Natural Dissolved Aquatic Organic-Matter."
Chemosphere 21(10-11): 1181-1197.

Guo, L. D. and R. W. Macdonald (2006). "Source and transport of terrigenous organic
matter in the upper Yukon River: Evidence from isotope (delta C-13, Delta C-14,
and delta N—15) composition of dissolved, colloidal, and particulate phases."
Global Biogeochemical Cycles 20(2).

Hatt, B. E., T. D. Fletcher, et al. (2004). "The inﬂuence of urban density and drainage
infrastructure on the concentrations and loads of pollutants in small streams."
Environmental Management 34(1): 112-124.

Hayes, M. H. B., P. MacCarthy, et al. (1989). The search for structure: Setting the scene.
Humic Substances II: In Search of Structure. M. H. B. Hayes, P. MacCarthy, R.
L. Malcolm and R. S. Swift. New York, John Wiley & Sons: 3-31.

Hayes, M. H. B. and R. S. Swift (1978). The chemistry of soil organic colloids. 1119
Chemistry Of Soil Constituents. D. J. Greenland and M. H. B. Hayes. Chichester,
Wiley: 197-198.

Hedges, J. I. (1980). Flux of organic carbon by rivers to the oceans : report of a
workshop. . Woods Hole, Massachusetts, Division of Biological Sciences,
National Research Council; work supported by US. Department of Energy,
Ofﬁce of Energy Research: 109.

Hedges, J. I. (1981). Chemical Indicators of Organic River Sources in Rivers and

Estuaries. Flux of organic carbon by rivers to the ocean_s: report of a worlﬂop,

Woods Hole, Massachusetts, September 21-25, 1980 / prapared by Committee on

 

23

Flux of Organic Carbon to the Ocean, Division of Biolggical Sciences. National
Research Council. U. S. D. 0. Energy. Washington, DC, US. Dept. of Energy,
Ofﬁce of Energy Research, Ofﬁce of Health and Environmental Research: 109-
141.

Hemond, H. F. (1990). Wetlands as the Source of Dissolved Organic Carbon to Surface
Waters. Organic Acids in Aquatic Ecosystems; report of the Dahlem Workshop
on Organic Acids in Aquatic Ecosystems, Berlin 1989, May 7-12. B. M. Perdue
and E. T. Gjessing. New York, John Wiley & Sons Ltd: 301-313.

Hood, E., M. W. Williams, et a1. (2005). "Sources of dissolved organic matter (DOM) in
a Rocky Mountain stream using chemical fractionation and stable isotopes."
Biogeochemistﬂ 74(2): 231-255.

Hope, D., M. F. Billett, et al. (1994). "A Review of the Export of Carbon in River Water -
Fluxes and Processes." Environmental Pollution 84(3): 301-324.

Homung, M., P. A. Stevens, et al. (1986). "The Impact of Pasture Improvement on the
Soil Solution Chemistry of Some Stagnopodzols in Mid-Wales." Soil Use and
Maggement 2(1): 18-26.

Imai, A., T. Fukushima, et al. (2001). "Fractionation and characterization of dissolved
organic matter in a shallow eutrophic lake, its inﬂowing rivers, and other organic
matter sources." Water Research 35(17): 4019-4028. ‘

Janos, P. (2003). "Separation methods in the chemistry of humic substances." Journal of
Chromatography A 983(1-2): 1-18.

Jardine, P. M., N. L. Weber, et al. (1989). "Mechanisms of Dissolved Organic-Carbon

Adsorption on Soil." Soil Science Society of America Joumal 53(5): 1378-1385.

24

J ohnsen, S., K. Martinsen, et al. (1987). "Seasonal variation in composition and
properties of aquatic humic substances." Science of the Total Environment 62:
13-25.

Kaiser, E., D. B. Arscott, et a1. (2004). "Sources and distribution of organic carbon and
nitrogen in the Tagliamento River, Italy." Aquatic Sciences 66(1): 103-116.

Kaiser, E., A. J. Simpson, et al. (2003). "Solid-state and multidimensional solution-state
NMR of solid phase extracted and ultraﬁltered riverine dissolved organic matter."
Environmental Science & Technology 37(13): 2929-2935.

Kaiser, K. and W. Zech (1998). "Rates of dissolved organic matter release and sorption in
forest soils." Soil Science 163(9): 714-725.

Kalbitz, K., S. Solinger, et al. (2000). "Controls on the dynamics of dissolved organic
matter in soils: A review." Soil Science 165(4): 277-304.

Kawahigashi, M., K. Kaiser, et a1. (2006). "Sorption of dissolved organic matter by
mineral soils of the Siberian forest tundra." Global Change Biology 12(10): 1868-
1877.

Kerndorff, H. and M. Schnitzer (1980). "Sorption of Metals on Humic-Acid."
Geochimica Et Cosmochimica Acta 44(11): 1701 -1708.

Kortelainen, P. (1993). "Content of Total Organic-Carbon in Finnish Lakes and Its
Relationship to Catchment Characteristics." Canadian J ourn_al of Fisheries and
Aquatic Sciences 50(7): 1477-1483.

Langmuir, D. (1997). Aqueous Environmental Geochemistry. Upper Saddle River, NJ,

Prentice Hall.

25

Larson, J. H., P. C. Frost, et al. (2007). "Effects of upstream lakes on dissolved organic
matter in streams." Limnology and Oceanoggaphy 52(1): 60-69.

Leenheer, J. A. and J. P. Croue (2003). "Characterizing aquatic dissolved organic matter."
Environmental Science & Technology 37(1): 18A-26A.

Lennon, J. T. and L. E. Pfaff (2005). "Source and supply of terrestrial organic matter
affects aquatic microbial metabolism." Aquatic Microbial Ecology 39(2): 107-
119.

Leu, C., H. Singer, et al. (2004). "Simultaneous assessment of sources, processes, and
factors inﬂuencing herbicide losses to surface waters in a small agricultural
catchment." Environmental Science & Technology 38(14): 3827-3834.

Lindell, M. J ., H. Graneli, et al. (2000). "Seasonal photoreactivity of dissolved organic
matter from lakes with contrasting humic content." Canadian Journal of Fisheries
and Aquatic Sciences 57(5): 875-885.

Linnik, P. N. (2003). "Complexation as the most important factor in the fate and transport
of heavy metals in the Dnieper water bodies." Analytici and Bioararlytical
Chemist_ry 376(3): 405-412.

Lou, T. and H. X. Xie (2006). "Photochemical alteration Of the molecular weight of
dissolved organic matter. " Chemosphere 65(11): 2333-2342.

Louchouam, P., S. Opsahl, et al. (2000). "Isolation and quantiﬁcation of dissolved lignin
ﬁ'om natural waters using solid-phase extraction and GC/MS." Analﬂical
Chemistg 72(13): 2780-2787.

Mace, J. E., C. H. Lin, et al. (2001). The effect of an XAD-8 resin fractionation scheme

for natural DOM on the sorption of hydrophobic organic chemicals.

26

Understanding and Managiag Orggic Maﬂr in Soils. Sediments and Water:
Proceedings of the 9th International Humic Substances Society. R. S. Swiﬁ and K.
M. Spark. Adelaide, Australia. September 21-25, 1998: 581-587.

Mantoura, R. F. C. and E. M. S. Woodward (1983). "Conservative Behavior of Riverine
Dissolved Organic-Carbon in the Severn Estuary - Chemical and Geochemical
Implications." Geochimica Et Cosmochimica Acta 47(7): 1293-1309.

Marschner, B. and K. Kalbitz (2003). "Controls of bioavailability and biodegradability of
dissolved organic matter in soils." Geoderrna 113(3-4): 211-235.

Mason, C. F., S. A. Norton, et al. (1999). "Deconstruction of the chemical effects of road
salt on stream water chemistry." Journal of Environmental Cum 28(1): 82-91.

McCallister, S. L., J. E. Bauer, et al. (2006). "Sources Of estuarine dissolved and
particulate organic matter: A multi-tracer approach." Organic Geochemistry 37:
454-468.

McDonald, S., A. G. Bishop, et al. (2004). "Analytical chemistry of freshwater humic
substances." Analytica Chimica Acta 527(2): 105-124.

McDowell, W. H. (1985). "Kinetics and Mechanisms of Dissolved Organic-Carbon
Retention in a Headwater Stream." Biogeochemistry 1(4): 329-352.

McDowell, W. H. and G. E. Likens (1988). "Origin, Composition, and Flux of Dissolved
Organic-Carbon in the Hubbard Brook Valley." Ecological Monogram 58(3):

1 77-195.
McDowell, W. H. and T. Wood (1984). "Podzolization - Soil Processes Control

Dissolved Organic-Carbon Concentrations in Stream Water." Soil Science 137(1):

 

23-32.

27

McGlynn, B. L. and J. J. McDonnell (2003). "Role of discrete landscape units in
controlling catchment dissolved organic carbon dynamics." Water Resources
Research 39(4): SWC 3 1-18.

McKnight, D. M., E. D. Andrews, et al. (1994). "Aquatic Fulvic-Acids in Algal-Rich
Antarctic Ponds." Limnology and Oceanography 39(8): 1972-1979.

McKnight, D. M., E. W. Boyer, et al. (2001). "Spectroﬂuorometric characterization of
dissolved organic matter for indication of precursor organic material and
aromaticity." Limnolggg and Oceanography 46(1): 38-48.

Meier, M., Y. P. Chin, et al. (2004). "Variations in the composition and adsorption
behavior of dissolved organic matter at a small, forested watershed."
Biogeochemistry 67(1): 39-56.

Merritt, K. A. and M. S. Erich (2003). "Inﬂuence of Organic Matter Decomposition on
Soluble Carbon and Its Copper-Binding Capacity." J Environ Oual 32(6): 2122-
21 3 1 .

Minor, E. C., J .-P. Simjouw, et al. (2006). "Seasonal variations in dissolved organic
carbon concentrations and characteristics in a shallow coastal bay. " Lanna
Chemist_ry 101(3-4): 166-179.

Molot, L. A. and P. J. Dillon (1997). "Colour - mass balances and colour - dissolved
organic carbon relationships in lakes and streams in central Ontario." Canadian
Journal of Fisheries and Aquatic Sciences 54(12): 2789-2795.

Molot, L. A. and P. J. Dillon (1997). "Photolytic regulation of dissolved organic carbon

in northern lakes." Global Biogeochemical Cycles 11(3): 357-365.

28

Moore, T. R., W. Desouza, et a1. (1992). "Controls on the Sorption of Dissolved Organic-
Carbon by Soils." Soil Science 154(2): 120-129.

Moore, T. R. and R. J. Jackson (1989). "Dynamics of Dissolved Organic-Carbon in
Forested and Disturbed Catchments, Westland, New-Zealand .2. Larry River."
Water Resources Research 25(6): 1331-1339.

Muller, M. B., D. Schmitt, et al. (2000). "Fractionation of natural organic matter by size
exclusion chromatography - Properties and stability of fractions." Environmental
Science & Technology 34(23): 4867-4872.

Naiman, R. J ., J. M. Melillo, et al. (1987). "Longitudinal Patterns of Ecosystem Processes
and Community Structure in a Sub-Arctic River Continuum." My 68(5):

1 139-1 156.

Otsuki, A. and R. G. Wetzel (1974). "Release Of Dissolved Organic-Matter by Autolysis
of a Submersed Macrophyte, Scirpus-Subterminalis." Limnology and
Oceanography 19(5): 842-845.

Pelekani, C., G. Newcombe, et al. (1999). "Characterization of natural organic matter
using high performance size exclusion chromatography." Environmentﬂ Science
& Technology 33(16): 2807-2813.

Peuravuori, J. and K. Pihlaja (1997). "Molecular size distribution and spectroscopic
pr0perties of aquatic humic substances." Analﬂica Chimica Acta 337(2): 133-
149.

Piccolo, A. (1998). Hydrophobic interactions controlling molecular sizes of humic
molecules in soils. Effects on the accumulation and decomposition of soil organic

matter. 16‘h World Congress of Soil Science, Montpellier, France.

29

Pomeroy, L. R. (1974). "Oceans Food Web, a Changing Paradigm." Bioscience 24(9):.
499-504.

Porcal, P., J. Hejzlar, et al. (2004). "Seasonal and photochemical changes of DOM in an
acidiﬁed forest lake and its tributaries." Aapatic Sciences 66(2): 211-222.

Qualls, R. G. and B. L. Haines (1992). "Biodegradability of Dissolved Organic-Matter in
Forest Throughfall, Soil Solution, and Stream Water." Soil Science Society of
America Journal 56(2): 578-586.

Reche, I. and M. L. Pace (2002). "Linking dynamics of dissolved organic carbon in a

forested lake with environmental factors." Biogeochemistry 61(1): 21-36.

 

Rhodes, A. L., R. M. Newton, et al. (2001). "Inﬂuences of land use on water quality of a
diverse new England watershed." Environmental Science & Technology 35(18):
3640-3645.

Richey, J. E., J. T. Brock, et al. (1980). "Organic-Carbon - Oxidation and Transport in the

Amazon River." Science 207(4437): 1348-1351.

 

Rodriguez-Zt'miga, U. E, D. M. B. P. Milori, et al. (2008). "Changes in Optical Properties
Caused by UV-Irradiation of Aquatic Humic Substances from the Amazon River
Basin: Seasonal Variability Evaluation." Environmental Science & Technology
42(6): 1948-1953.

Saliot, A., S. Derieux, et al. (2002). "Winter and spring characterization of particulate and
dissolved organic matter in the Danube-Black Sea mixing zone." Estuarine

Coastal and Shelf Science 54(3): 355-367.

30

Schiff, S. L., R. Aravena, et a1. (1990). "Dissolved Organic-Carbon Cycling in Forested
Watersheds - a Carbon Isotope Approach." Water Resources Research 26(12):
2949-2957.

Schlesinger, W. H. and J. M. Melack (1981). "Transport of Organic-Carbon in the
Worlds Rivers." Te_llu_s 33(2): 172-187.

Schwede-Thomas, S. B., Y. P. Chin, et al. (2005). "Characterizing the properties of
dissolved organic matter isolated by XAD and C-18 solid phase extraction and
ultraﬁltration." Aquatic Sciences 67(1): 61-71.

Shafer, M. M., S. R. Hoffmann, et al. (2004). "Physical and kinetic speciation of copper

and zinc in three geochemically contrasting marine estuaries." Environmental

 

Science & Technology 38(14): 3810-3819.

Sheeder, S. A., J. D. Ross, et al. (2002). "Dual urban and rural hydrograph Signals in
three small watersheds." Journal of the American Water Resources Association
38(4): 1027-1040.

Shiller, A. M., S. W. Duan, et al. (2006). "Photo-oxidation of dissolved organic matter in
river water and its effect on trace element speciation." Limnology and
Oceanography 51(4): 1716-1728.

Sholkovitz, E. R. (1976). "Flocculation of Dissolved Organic and Inorganic Matter
During Mixing of River Water and Seawater." Geochimica Et Cosmochimica
tha 40(7): 831-845.

Sholkovitz, E. R., E. A. Boyle, et al. (1978). "Removal of Dissolved Humic Acids and
Iron During Estuarine Mixing." ﬁrth and Planetary Science Letters 40(1): 130-

136.

31

Simjouw, J. P., E. C. Minor, et al. (2005). "Isolation and characterization of estuarine
dissolved organic matter: Comparison of ultraﬁltration and C-1 8 solid-phase
extraction techniques." Marine Chemistﬂ 96(3-4): 219-235.

Sollins, P. and F. M. McCorison (1981). "Nitrogen and Carbon Solution Chemistry of an
Old Growth Coniferous Forest Watershed before and after Cutting." Mg
Resources Research 17(5): 1409-1418.

Stangroom, S. J ., C. D. Collins, et al. (1998). "Sources of organic rrricropollutants to
lowland rivers." Environmental Technolgy 19(7): 643-666.

Stedmon, C. A., S. Markager, et al. (2006). "Dissolved organic matter (DOM) export to a
temperate estuary: Seasonal variations and implications of land use." Estuaries
and Coasts 29(3): 388-400.

Stevenson, F. J. (1982). Humus Chemistry - genesis, compostion, reactiona. New York,
Wiley.

Sun, L., E. M. Perdue, et al. (1997). "Use of elemental composition to predict
bioavailability of dissolved organic matter in a Georgia river." Limnology and
Oceanography 42(4): 714-721.

Sutton, R. and G. Sposito (2005). "Molecular Structure in Soil Humic Substances: The
New View." Environmental Science & Technolggy 39(23): 9009-9015.

Swift, R. S., R. L. Leonard, et al. (1992). "Changes in Humic-Acid Composition with
Molecular-Weight as Detected by C-13-Nuclear Magnetic-Resonance

Spectroscopy." Science of the Total Environment 118: 53-61.

32

Thacker, S. A., E. Tipping, et al. (2005). "Development and application of functional
assays for freshwater dissolved organic matter." Water Research 39(18): 4559-
4573.

Thurman, E. M. (1985). Organic geochemistry of natural waters. Boston, Kluwer
Academic.

Thurman, E. M., I. F errer, et al. (2001). "Choosing between Atmospheric Pressure
Chemical Ionization and Electrospray Ionization Interfaces for the HPLC/MS
Analysis of Pesticides." A_nalytical Chemistm 73(22): 5441-5449.

Thurman, E. M. and R. L. Malcohn (1981). "Preparative Isolation Of Aquatic Humic
Substances." Environmental Science & Technolagy 15(4): 463-466.

Tipping, E. and C. Woof (1983). "Seasonal-Variations in the Concentrations of Hmnic
Substances in a Soft-Water Lake." Limnology and Oceanography 28(1): 168-172.

Town, R. M. and H. K. J. Powell (1993). "Limitations of Xad Resins for the Isolation of
the Noncolloidal Humic Fraction in Soil Extracts and Aquatic Samples."
Analytica Chimica Acta 271(2): 195-202.

Tranvik, L. J. (1993). "Microbial transformation of labile dissolved organic matter into
humic-like matter in seawater." FEMS Microbiology Ecology 12(3): 177-183.

Waiser, M. J. and R. D. Robarts (2004). "Photodegradation of DOC in a shallow prairie
wetland: evidence from seasonal changes in DOC optical properties and chemical
characteristics." Biogeochemistry 69(2): 263-284.

Wallace, J. B., D. H. Ross, et al. (1982). "Seston and Dissolved Organic-Carbon

Dynamics in a Southern Appalachian Stream." Ecology 63(3): 824-838.

33

Wang, F. Y., J. S. Chen, et al. (1997). "Surface properties of natural aquatic sediments."
Water Research 31(7): 1796-1800.

Wayland, K. G., D. T. Long, et al. (2003). "Identifying relationships between baseﬂow
geochemistry and land use with synoptic sampling and R-mode factor analysis."
J oumal_of Environmental Quality 32(1): 180-190.

Wen, L. S., P. Santschi, et al. (1999). "Estuarine trace metal distributions in Galveston
Bay: importance of colloidal forms in the speciation of the dissolved phase."
Marine Chemistry 63(3-4): 185-212.

Wetzel, R. G. (1992). "Gradient-Dominated Ecosystems - Sources and Regulatory
Functions of Dissolved Organic-Matter in Fresh-Water Ecosystems."
Hydrobiologia 229: 181-198.

Wickland, K. P., J. C. Neff, et a1. (2007). "Dissolved organic carbon in Alaskan boreal
forest: Sources, chemical characteristics, and biodegradability." Ecosystems
10(8): 1323-1340.

Winterbourn, M. J., J. S. Rounick, et al. (1981). "Are New-Zealand Stream Ecosystems
Really Different." New Zealand Journal of Marine and Freshwater Research
15(3): 321-328.

Wu, S. C. and P. M. Gschwend (1986). "Sorption Kinetics of Hydrophobic Organic-
Compounds to Natural Sediments and Soils." Environmental Science &
Technology 20(7): 717-725.

Xie, X. D., S. Norra, et al. (2005). "A GIS-supported multivariate statistical analysis of
relationships among stream water chemistry, geology and land use in Baden-

Wurttemberg, Germany." Water Air and Soil Pollution 167(1-4): 39-57.

34

 

 

Yamamoto, H. and H. M. Liljestrand (2003). "The fate of estrogenic compounds in the
aquatic environment: sorption onto organic colloids." Water Science and
Technology 47(9): 77-84.

Yano, Y., K. Lajtha, et al. (2004). "Chemical and seasonal controls on the dynamics of
dissolved organic matter in a coniferous old-growth stand in the Paciﬁc
Northwest, USA." Biogeochemistry 71(2): 197-223.

Yano, Y., W. H. McDowell, et al. (1998). "Quantiﬁcation of biodegradable dissolved
organic carbon in soil solution with ﬂow-through bioreactors." Soil Science
SociethOf America Joum_a_l 62(6): 1556-1564.

Zoungrana, C. J ., R. Desj ardins, et al. (1998). "Inﬂuence of remineralization on the
evolution of the biodegradability of natural organic matter during ozonation."

_Water Research 32(6): 1743-1752.

35

CHAPTER 3
SIMULTANEOUS QUANTIFICATION OF DISSOLVED ORGANIC CARBON
FRACTIONS AND COPPER COMPLEXATION

USING SOLID-PHASE EXTRACTION

Abstract

Trace metal cycling in natural waters is highly inﬂuenced by the amount and type
of dissolved organic carbon (DOC). Although determining individual Species of DOC is
unrealistic, there has been success in classifying DOC by determining operationally
deﬁned fractions. However, current fractionation schemes do not allow for the
simultaneous quantiﬁcation of associated trace metals. Using operational classiﬁcations,
a scheme was developed to fractionate DOC based on a set of seven solid-phase
extraction (SPE) cartridges. The cartridges isolated fractions based on a range of speciﬁc
mechanisms thought to be responsible for DOC aggregation in solution, as well as
molecular weight. The method was evaluated to determine if it can identify differences in
DOC characteristics, including differences in Cu-DOC complexation. Results are that: (1)
cartridge blanks were low for both DOC and Cu, (2) differences are observed in the
distribution of DOC amongst the fractions from various sources that are consistent with
what is known about the DOC materials and the mechanisms operative for each cartridge,
(3) when present as a free cation, Cu was not retained by non-cationic cartridges allowing
the method to be used to assess Cu binding, (4) the capability of the method to provide
quantitative assessment of Cu-DOC complexation was demonstrated for a variety of

DOC standards, (5) Cu was found to preferentially complex with high molecular weight

36

 

 

ﬁactions of DOC, and (6) estimated partitioning coefﬁcients and conditional binding
constants for Cu were similar to those reported elsewhere. The method developed
describes DOC characteristics based on speciﬁc bonding mechanisms (hydrogen, donor-
acceptor, London dispersion, and ionic bonding) while simultaneously quantifying Cu-
DOC complexation. The method provides researchers a means of describing not only the

extent of DOC complexation but also how that complex will be behave in natural waters.

37

Introduction

Dissolved organic carbon (DOC) is known to Signiﬁcantly inﬂuence aquatic
biology and chemistry (Ghabbour and Davies 2004). Biologically, DOC serves as one of
the primary food sources for microbes and can be considered the foundation of aquatic
food webs in surface water systems (Perga, Kainz et al. 2006). Chemically, DOC alters
the fate and transport of both organic (Chin, Aiken et al. 1997; Raber, Kogel-Knabner et
al. 1998; Burkhard 2000; Lee, Gan et al. 2003) and inorganic pollutants (Kimball,
Callender et al. 1995; Santschi, Lenhart et al. 1997; Icopini and Long 2002). DOC is
often the dominant factor in determining the speciation and partitioning behavior of trace
metals in surface waters (Wen, Santschi et al. 1999; Linnik 2003). Trace metal binding
by DOC derived from various sources, or by broadly isolated fractions of DOC (e. g.
humic or ﬁrlvic acid fractions), has been investigated and general orders of binding
affinities have been documented (Kemdorff and Schnitzer 1980; Langmuir 1997).

DOC is best described as heterogeneous aggregates of low-molecular weight
organic molecules, including recognizable biomolecules, held together by hydrophobic
and hydrogen bonding (Sutton and Sposito 2005). An assortment of analytical techniques
has been developed to characterize DOC for a variety of purposes. Of particular interest
related to environmental mobility are techniques based on retention by an immobile solid
phase, such as liquid chromatography (LC) and solid-phase extraction (SPE), as these
provide information that can be directly related to the physical-chemical behavior and
chemical structure (Janos 2003; McDonald, Bishop et al. 2004). The most commonly
used fractionation techniques involve XAD resins which can be used to separate DOC

fractions based principally on polarity (Aiken, McKnight et al. 1992). Previous

38

 

 

 

investigations have evaluated the extent to which various fractions produced using XAD
resins bind trace metals (Croue, Benedetti et al. 2003). There are two primary limitations
to using this approach to assess the binding that occurs under natural conditions. First, it
has not been established that binding to broadly isolated fractions of DOC are
representative of a larger mixture (Hayes, MacCarthy et al. 1989). Second, many of these
techniques alter the nature of the DOC, and it is likely this affects binding. For example,
the XAD fractionation methods rely on substantial changes in pH, which may result in
hydrolysis of the DOC (Mace, Lin et al. 2001).

A limited number of studies have attempted to evaluate trace metals bound to
different DOC components (Groschner and Appriou 1994; Appelblad, Baxter et al. 1999;
Yamini and Tamaddon 1999; Abollino, Aceto et al. 2000; Icopini and Long 2002;
Gardner and van Veen 2004; Turner, Le Roux et al. 2004). Most of these methods utilize

a single, and often unspeciﬁed, mechanism for retaining DOC-trace metal complexes. A

few utilize octadecyl-bonded silica, or C18, solid phases which may contain open silanols

that provide uncomplexed trace metals the possibility of bonding directly to the support
phase (Donat, Statham et al. 1986). Because of these limitations, previous methods offer
only a limited characterization of the DOC responsible for complexation.

The aim of this study was to develop a technique to quantify DOC fractions based
on multiple characteristics that would likely affect environmental transport, and to
simultaneously quantify the amount of trace metal that had been bound by each fraction.
This method attempts to address some of the major limitations of previous LC and SPE
techniques which could not be employed for one or more of the following reasons a) they

required pH adjustment or chemical changes that would alter binding, b) the leaching of

39

 

 

 

trace metals or DOC from the solid phase was either excessive or required extensive
cleanup procedures, and c) the solid phase retained free metals which would preclude the

quantiﬁcation of binding.

Materials and Methods

The isolation of DOC and trace metals was achieved by passing aqueous samples
through a set of seven SPE cartridges (Table 3-1). Cartridges were run in parallel and
utilized a range of retention mechanisms: four were prepared to isolate fractions based on
ionic interactions, two primarily on the basis of hydrophobic mechanisms, and one that
utilized hydrogen bonding. DOC and trace metal concentrations were determined before
and after samples were passed through the cartridges to identify the amount retained on
the solid phase, allowing the quantiﬁcation of binding.

Solid phases which utilize styrene-divinylbenzene (SDVB) copolymer structure

were employed to avoid problems reported for other bonded phases media (e. g. direct

interaction between trace metals and unreacted silanols on C13 media). SDVB is

particularly well suited for the retention of hydrophobic organic compounds but, unlike

C13, does not contain silica that may retain uncomplexed trace metals. Additionally, the

capacity of SDVB media is typically greater than C13 due to increased surface area
(Thurman and Mills 1998). However, because of the heterogeneous composition of DOC
it is nearly impossible to calculate a single capacity coefﬁcient for each type of cartridge.
By identifying fractions of DOC retained by each cartridge, this method essentially
evaluates the extent of organic constituents with a range of afﬁnities (capacity factors) for

each type of cartridge media. The Argonaut Isolute 101 (200 mg) SPE cartridge was one

40

 

of the cartridges selected to retain DOC primarily through hydrophobic mechanisms and
is referred to as the hydrophobic cartridge. This cartridge is a highly cross-linked SDVB
copolymer that facilitates strong pi-bonding with aromatic organic compounds. The other
cartridge selected to retain analytes through hydrophobic mechanisms was the Waters
Oasis HLB (200 mg), identiﬁed as the extended hydrophobic cartridge. While this
cartridge is similar to the Isolute 101 cartridge, it is expected to retain more hydrophilic
compounds (i.e. some semi-polar compounds) due to additional N-vinylpyrrolidone
functionality (Waters 2003). The third type of non-ionic SPE cartridge used was the the
H-bonding cartridge, Supelco Discovery DPA-6S (250 mg). The Discovery DPA-6S
media consists of a polyamide resin designed to participate in hydrogen bonding. It is
reported that the Discovery DPA-6S media extracts tannins, chlorophyll, humic acid and
other compounds with hydroxyl groups, especially aromatic carboxylic acids (Supelco
2005)

In addition to the non-ionic mechanisms, four cartridges utilizing ion-exchange

mechanisms were also employed. Only one of these four cartridges was designed to

isolate cations, the Bio-Rad Chelex 100 (50—100 mesh, Na+ counterion). This resin

consists of SDVB copolymers containing paired iminodiacetate ions that act as chelating
groups in binding polyvalent metal ions (Bio-Rad 1998). Two basic types of arrion-
exchange cartridges were used, both retaining negatively charged species through
bonding with quaternary ammonium functional groups. One type of anion-exchange
cartridge utilized Bio-Rad AG MP-l (50—100 mesh) resin, the other Bio-Rad AG-X8

resin. The AG MP-l resin is designed to retain a wide size range of constituents within a

large effective surface area (~23 m2 g1) and porosity (20%), while the AG-X8 resin

41

 

contains a lkDa molecular weight (MW) cutoff, a copolymer lattice with an 8% cross
linkage limiting the size of analytes that can access the anion-exchange sites to those with
a MW of less than approximately 1 kDa (Bio-Rad 1998).

One cartridge utilizing the AG-X8 resin (anion-1 1(0) and two cartridges utilizing

the AG MP-l are employed in this method. One of the two AG MP-l cartridges (anion-

F) used a F - counter ion while the other (anion-I) had an I- counterion. By using different

counterions it is possible to gauge the ionic afﬁnity of negatively charged analytes (DOC

and trace metal complexes) that are retained by the resins. The F- counterion has a very

low afﬁnity for the anion-exchange sites, second only to OH" ions, while an I- counterion

has a high affinity (Bio-Rad 1998). Fluoride counterions were used for all anion-lkD
cartridges.

By comparing the retention of DOC on each of the seven cartridges it is possible
to characterize DOC generally, as hydrophobic and hydrophilic, and speciﬁcally, on the
basis of ionic bonding, donor-acceptor interactions, hydrogen bonding and London
dispersion forces. The hydrophilic fraction of DOC can be deﬁned as the additional
amount of DOC retained by the extended hydrophobic cartridge due to the more
hydrophilic functionality (i.e. DOC retained by the extended hydrophobic cartridge minus
DOC retained by the hydrophobic cartridge). Speciﬁcally, the amount of DOC retained
by ionic bonding is approximated by the amount of DOC retained by the anion-F
cartridge. Hydrophobic bonding accounts for both donor-acceptor interactions and
London-dispersion forces. Based on the dominate retention mechanism of the cartridges

used, it is possible to determine the amount of DOC retained through hydrophobic

42

 

 

bonding, as well as, a relative assessment of donor-acceptor interactions and London
dispersion forces. Retention by the H-bonding cartridge describes the propensity of DOC
to undergo hydrogen bonding, while retention by the hydrophobic cartridge estimate the
extent of donor-acceptor interactions. The proportion of DOC retention due to London
dispersion forces is approximated by dividing the amount of DOC retained on the
extended hydrophobic cartridge, which facilitates donor-acceptor, London dispersion,
and hydrogen bonding interactions, by the retention of the hydrophobic and H-bonding
cartridges.

The anion and cation-exchange cartridges used to isolate DOC fractions were
prepared in the laboratory. Resins were slurried (1.005 i 0.005 g for anion exchange,
2.00521: 0.005 g for cation exchange) into 8mL Ultra-CleanTM polypropylene cartridge
tubes (Alltech) with a 20 um Extract-CleanTM polyethylene ﬁit (Alltech). The terms
Ultra-Clean and Extract-Clean referred to are product names and should not be confused
with the ultra-clean acid-washing procedure described later. Once cartridges were ﬁlled
with resin, they were tapped to eliminate air bubbles and sonicated for 30 minutes.
Another 20 um ﬁit was then placed on top of the resin. Care was taken to prevent resins
from becoming dry at any time during assembly and cartridges were stored with a small

amount of water remaining above the top ﬁit. Cartridges were prepared with the desired
counter ion (F- or I-) by ﬂushing with appropriate amounts of 1N NaOH and NaF or NaI

solutions, as recommended by the manufacturer (Bio-Rad 1998). Before use, all
cartridges were conditioned. First, non-ionic manufactured cartridges received lmL of
methanol which was allowed to soak for 2 minutes. Second, all cartridges, with the

exception of the cation-exchange cartridges, were rinsed with 1 mL of 0.02% Trace

43

SELECTTM hydroﬂuoric acid (HF) (Fluka). Finally, all cartridges were rinsed with 80
mL of ultra-pure water (>18 MQ).

All laboratory experiments were conducted by passing samples through SPE
cartridges (in parallel) using ultra-clean 60 mL syringes arranged on a syringe pump
(Harvard Apparatus) to control the sample ﬂow rate. Syringes were connected to SPE
tubes with the use of cartridge adapters (Alltech). With the bottom of the SPE tube
capped, syringes were ﬁlled with 60 mL of sample ﬁom the top. Once ﬁlled, the SPE
tube cap was removed, the syringe plunger was inserted and ~10 mL of sample was
forced through the SPE cartridge and discarded (without the plunger contacting the
aqueous sample). Flushing 10 mL of sample effectively replaced the ~2 mL of ultra-pure
water remaining from cartridge conditioning and allowed the SPE media to equilibrate
with the sample solution. Eluent from cartridges was collected in 40 mL borosilicate
glass vials for DOC analysis. After the DOC samples were collected, SPE cartridges were
temporarily removed from the syringes to be reﬁlled. At all times the SPE media
remained ﬁrlly saturated with excess sample. Syringes were then reloaded with sample
solution to 40 mL. Again plungers were inserted and placed back into the syringe pump.
Eluent was collected in ultra-clean 30 mL polypropylene bottles for trace metals analysis.

DOC samples were capped with TEE/silicone liners and reﬁigerated until
analysis, typically within 1-2 days. The concentration of DOC was determined by
automated analysis based on the Heated-Persulfate Oxidation Method (Clesceri,
Greenberg et al. 1998) using an 01 Analytical Model 1010 Wet Oxidation Total Organic
Carbon Analyzer. Prior to conducting experiments, all glassware used during DOC

analysis was acid (18% HCl) washed, rinsed with ultra-pure water and placed in a 550 °F

44

oven for more than 2 hours to ensure cleanliness. Trace metal samples were preserved by
adding 180 uL of Optima nitric acid (Fisher Scientiﬁc) and refrigerated until analysis.
Concentrations of Cu were determined by inductively coupled plasma/mass spectrometry

(ICP/MS) as outlined by Standard Methods method 3125 B (Clesceri, Greenberg et al.

1998). Indium and bismuth internal standards (20 ug L'l) were used and samples were

analyzed using a Micromass Platform ICP/MS with a hexapole collision cell. All
calibration standards were within 15% of the calibrated value. Syringes and sample
bottles were subjected to ultra-clean procedures to eliminate the possibility of trace metal
contamination, with the exception of the rubber caps on the syringe plungers which were
soaked in ultra-pure water for >24hrs. Special care was taken (maintaining an air gap)
during experiments to prevent possible contamination from the rinsed rubber caps. The
ultra-clean process consisted of soaking syringes and bottles in 18% hydrochloric acid,
35% nitric acid at 20 °C, 35% nitric acid at 45 °C and ultra-pure water for ~24 hours. All
acids were trace element grade (Fisher Scientiﬁc) and cleaning was done inside an EPA
class 100 clean room. All experiments Were conducted within a clean hood at 25°C.
Major anions and cations in solutions were measured by ion chromatography and ﬂame
atomic adsorption (Clesceri, Greenberg et al. 1998).

Aldrich humic acid (HA) (cat. no. H1,675-2; lot no. MV01816HH), a water
sample containing natural DOC from the Red Cedar River (RCR) in East Lansing,
Michigan, and seven reference standards from the International Humic Substances
Society (IHSS) were used to evaluate DOC fractionation and behavior. The IHSS
standards used were: Nordic Reservoir natural organic matter (NOM) (cat. no. 1R108N);

Suwannee River NOM (cat. no. 1R101N), fulvic acid (FA) (cat. no. 2S101F) and HA

45

(cat. no. ZSIOIH); Waskish Peat FA (cat. no. 1R107F) and HA (cat. no. 1R107H); and
Pahokee Peat HA (cat. no. 1S103H). It should be noted that NOM samples are isolated by
reverse osmosis while HAS and F As are isolated using XAD resins. These reference
standards were included because they are well characterized and widely used. Differences
in DOC composition and structure have been reported for the reference standards and are
available elsewhere (IHSS, 2007). Dissolved organic matter (DOM) from the RCR was

prepared by passing the natural water sample through an acid washed 0.45 pm glass-ﬁber

ﬁlter. Samples containing copper were spiked using 1000 mg Cu L-1 aqueous standards

of copper nitrate in dilute nitric acid (Fischer Scientiﬁc).
To simulate complexation and ionic interactions likely to occur in natural surface
waters, an artiﬁcial solution was made using the solution chemistry of a sample from the

RCR as a template (Table 3-2). The artiﬁcial solution was constructed by dissolving the

salts CaCO3, CaSO4, MgClz, KCl, KNOg, NaI-ICO3, NaCl and NazHPO4 in 20 L of

ultra-pure water. The resulting artiﬁcial solution was similar to the RCR solution used as
a except for a lower Ca concentration. While the RCR sample was supersaturated with
respect to calcite, the artiﬁcial solution was in equilibrium with calcite. The chemistry of
the solution remained stable for the duration of experimentation (one week) and, in fact,
maintained stability for more than one year.

Solution chemistry was modeled in PHREEQCi (Parkhurst and Appelo 2005).
For solutions without DOC present, the MINTEQ.v4 database was used. However, to
study Cu-DOC complexation, a database containing a PHREEQCi compatible version of
Tipping and Hurley’s WHAM model was used (Appelo and Postrna 2005). Solutions

containing DOC were modeled in two steps and in each step Na was allowed to vary to

46

 

 

 

achieve charge balance. While Na was not expected to vary Signiﬁcantly during
experiments, it was used to attain charge balance for two reasons: (1) changes in Na
concentration have a relatively minor impact on complexation and (2) the error associated
with its measurement is typically larger than the error observed in the charge balance.
Allowing other parameters to vary, such as pH, could greatly alter solution chemistry and
not accurately represent conditions observed. The ﬁrst step modeled solutions without
organic complexation. For solutions containing reference DOC standards, the initial pH
was set to 8.0 (measured pH of the artiﬁcial solution), carbonate equilibrium was
established (pCOz = -2.84). This resulted in a 2.7% decrease in Na, a pH of 8.00, and a
pE of 11.56. For the RCR sample, the initial pH was set to the pH measured (7.8),
equilibrium carbonate was included (pCOz = -2.32). This resulted in a 0.3% increase in
Na, a pH of 7.78, and a pE of 13.23. The second modeling step described organic
complexation by equilibrating a surface layer with the solution and gas phase ﬁom the
ﬁrst step; resulting in a pH of 7.77 to 7.99 and pE of 12.73 to 13.28. The surface layer
approximated Site speciﬁc binding to 4 carboxylic-like, 4 phenolic-like and 12 bidentate
sites according to the model developed by Tipping and Hurley (1992). The fraction of
bidentate Sites (0.4) remained constant for all types of DOC modeled.

For solutions containing DOC, PHREEQCi model variables based on those
reported in the literature (Table 3-3) were adjusted in order to effectively model observed
complexation, based on conditional stability constants. First, estimates for the number of
binding Sites were derived ﬁom the literature. For Suwannee River HA, F A and NOM,
the total number of binding sites (nHA + nHB) was based on the sum of the maximum

charge density of carboxylic and phenolic groups, assuming a single charge per binding

47

site (1 eq mol'l) (IHSS 2007). This resulted in total binding-site densities (mmol gC-l)

for Suwannee River samples of 9.81 for HA, 11.58 for FA, and 3.82 for NOM. For both

NOM samples, the total number of binding sites was set equal to the total number of Cu

binding sites observed for DOM isolated by reverse osmosis, 4.55 mmol gC.1 (Lu and

Allen 2002). A charge density of 3.8 mmol gC-1 was estimated for Aldrich HA, based on

data from Saito et al. (2005) who used a puriﬁed form of the acid at pH~8. Because
information was not available on the composition of RCR DOM, it was assumed that

25% of DOC was composed of HA and FA fractions. Using an average charge density of

HA (8.0 meq gC-l) and FA (5.2 meq gC-l) at a pH~8 (Higgo and Rees 1986), an overall

charge density of 3.3 mmol gC-1 was estimated.

Second, the distribution of carboxylic—like (type A) to phenolic-like (type B) sites
was estimated in two ways. For DOM samples with the amount of carboxylic and
phenolic groups reported (IHSS 2007), the same ratio was used for type A and B sites.
For DOM samples lacking compositional data, the amount of type A binding sites were

assumed to be double the amount of type B binding sites (Thurman 1985).
Partitioning coefﬁcients (KDOC) and conditional stability constants (KC) were
calculated using a short Basic program within the USER_PUNCH feature of PHREEQCi.

The PHREEQCi model calculated KDOC as follows:

[M bound ]

M
KDOC (MODEL) 2 J—K’EL (Equation 3-1)
[0C surface I

48

where [Mboundl is the amount of metal (moles) associated with surface binding sites and
[Mtoral] is the total amount of metal (moles) present in solution, and the ratio is
normalized to the concentration of organic carbon (kgC L-l) modeled as a surface layer
[OCsm-face]. Ions calculated to be present within the double layer were assumed to be
retained by the SPE cartridge and therefore are accounted for in [Mboundl- The

PHREEQCi model calculated KC_MODEL as follows:

[M org _ complex]
[M non __ org 110C]

 

KC_MODEL 2 (Equation 3-2)

where [Morg_comp1ex] is the amount of metal (moles) complexed with DOC, [Mnon_org]
is the amount of metal (moles) not complexed with DOC, and [0C] is the concentration

of organic carbon (gC L-l). Ions calculated to be present within the double layer were

assumed to be retained by the SPE cartridge and are included for in [Morg_comp1exl-

Results and Discussion

Five basic experiments were conducted: (1) the levels of Cu and DOC that
leached from conditioned cartridges were established, (2) the retention of a model DOC
(Aldrich HA) was measured under different ﬂow rates, (3) the retention of different types

of DOC (i.e. HA, FA, NOM, and a non-puriﬁed surface water sample) obtained from

49

multiple sources (Suwannee River, Waskish Peat, Nordic Reservoir, etc.) was assessed,
(4) the removal of uncomplexed Cu by cartridges was evaluated, and (5) the retention of
Cu in aqueous solutions containing different forms of DOC was determined.

The ﬁrst experiment conducted was to determine the cleanliness of the SPE
cartridges and sampling procedure. With this aim, ultra-pure water was run through each
cartridge according to the method procedure described above. Cartridge eluent was
analyzed to determine the extent of Cu and organic carbon leaching (Table 3-4). After

conditioning, cartridge blanks were below the limit of quantiﬁcation (standard deviation

+ 10 sigma, n=3) for Cu (0.5 ug LII) with the exception of one cartridge. The anion-I

cartridge produced a blank eluent concentration of 2.5 d: 0.6 ugCu L-l. DOC

concentrations observed in the eluent of conditioned cartridges ranged from 0.06 :t 0.11

mg L.1 for the anion-I cartridge to 0.6 i 0.1 mg L.1 for the anion-lkD cartridge.

While other methods tend to focus on either trace metal or DOC retention, this
method is one of only a few that is designed to simultaneously produce both, requiring

clean blanks for Cu and DOC. Cartridge blanks for Cu were typically less than a few u g

L-1 and DOC concentrations were well below those found in natural samples,

establishing these cartridges to be relatively free of contaminants. Once cartridges were
made, sample preparation and cartridge conditioning required less than ten minutes per
sample using a multi-port vacuum manifold. This is favorable compared to other methods
which isolate DOC fractions and require up to 120 hr of resin clean-up prior to analysis
(Louchouam, Opsahl et al. 2000). Overall, the method proved to be simple to implement

and required very little pre-cleaning.

50

 

The second experiment was designed to determine the optimum ﬂow rate to
isolate DOC while minimizing sample processing time. Water samples for this
experiment were prepared by dissolving Aldrich HA in unsoftened groundwater (tap

water; Table 3-2), and passing the samples through cartridges at a series of ﬁve ﬂow rates

ranging from 0.5 to 10 mL min-1. For all cartridges studied, the retention of DOC was

found to decrease with increasing ﬂow rate (Figure 3-1). Retention by anion exchange

. . -l . . .
was relatively constant up to 2.0 mL min , regardless of the counter ion, With the anion-I
. . . . . -l . . .
and arnon-F resrns showrng only a slight decrease at 4 mL min . Non-ionic cartridges

. . . . -1
showed Similar retention for ﬂow rates up to 1.0 mL min .

In order to decrease sample processing time while maximizing DOC retention, a

. -l . . .
ﬂow rate of 1.0 mL min was selected as the most appropriate to obtain DOC fractions.

This ﬂow rate is lower than has been reported by other investigators for the isolation of
lignin and humic substance by C18 and trace metals by cation exchange (Abollino, Aceto

et al. 2000; Louchouam, Opsahl et al. 2000; Gardner and van Veen 2004; Shafer,
Hofﬁnann et al. 2004). The ﬂow rate selected was also deemed prudent to assess the
dissociation of Cu—DOC complexes toward the competing iminodiacetic acid ligand
immobilized on SDVB polymers (cation cartridge) (Shafer, Hoffrnann et al. 2004).

The amount of Cu retained by cation exchange approximates the portion of “free”
species, or labile Cu, that can quickly dissociate from natural ligands (Shafer, Hoffmann
et al. 2004). Differentiating between labile and non-labile fractions of Cu using SPE

depends on (1) the ability of the cation-exchange resin to compete with other ligands in

51

solution and (2) the amount of time Cu-DOC complexes are exposed to competing
iminodiacetic acid ligands. The ﬂow rate used in this study results in a retention time of
greater than 3 minutes for the cation-exchange cartridge. This retention time is greater
than the retention time previously used to determine labile Cu, ~8 seconds (Shafer,
Hofﬁnann et al. 2004), and quantify “free” Cu species, 2 minutes (Gardner and van Veen
2004), using similar cation cartridges with smaller mesh sizes. While a smaller mesh size
may increase the retention of Cu by cation exchange, it is unclear if information obtained
using a smaller mesh size resin provides a better assessment of Cu mobility or toxicity.
Two factors support the use of a larger mesh size resin: (1) the amount of Cu retained by
cation exchange has been found to vary depending on DOC ligands (Shafer, Hoffrnann et
al. 2004); therefore, a resin that is capable of allowing a wider range of ligands to access
iminodiacetic binding sites may provide a more comprehensive assessment of metal-
ligand interaction, and (2) labile fractions determined using cartridges with smaller mesh-
size have been found to overestimate toxicity from free Cu (Florence, Morrison et al.
1992). Ultimately, any method using cation exchange to measure the labile fractions of a
trace metal relies on an arbitrary deﬁnition of labile and these measurements provide only
a relative approximation of kinetic limitations.

The ability of the selected SPE cartridges to isolate fractions of DOC based on
different mechanisms was tested using a variety of DOC standards and one natural water
sample. Seven solutions were made by spiking approximately one liter of artiﬁcial
solution with a single humic sample. The natural water sample was ﬁltered through an
acid-washed Type A/E glass-ﬁber ﬁlter (Pall) and otherwise untreated. These eight

solutions were passed through the selected cartridges as described previously. Given the

52

 

 

 

 

 

known mechanisms of retention for each cartridge, the amount retained by each provides
a powerful tool for assessing the functional characteristics of DOC. For example, the
propensity of DOC to undergo ionic versus hydrophobic interactions can be assessed by
comparing the ratio of DOC retained by the anion-F cartridge versus the extended
hydrophobic cartridge. Another DOC characteristic, the fraction of DOC less than lkDa
in molecular size, can be obtained by comparing the amount of DOC retained on the
anion-lkDa cartridge versus the anion-F cartridge. Additionally, comparing the ratio of
binding-site densities for Cu retained by the anion-F cartridge relative to the anion-lkD
cartridge provides a measure of the tendency of Cu to bind with high versus low MW

ligands. The MW binding-site density (BSD) ratio is described as

 

MF—Mx
MW BSD ' — F E u ' 3 3
t — t -
raio (MX/CX) ( q aion )

where Mp is the amount of metal retained by the anion-F cartridge, Mx is the amount of

metal retained by the anion-lkD cartridge, and CF and Cx are the amounts of carbon

retained by each of the cartridges. A MW BSD ratio equal to one means binding
constants are the same for high and low MW DOC. As this value increases, so does the
afﬁnity of that metal to complex with higher MW fractions of DOC.

Generally, it was expected that DOC standards isolated by reverse osmosis
(NOM) would contain HA and FA constituents and therefore would Show greater

retention on all cartridges. It was also expected that DOC retention by individual

53

 

 

 

 

 

cartridges would follow the order of bond strength: hydrogen bonding < hydrophobic
bonding < ionic bonding. Based on the pH of the artiﬁcial solution (~8), limited
quantities of these constituents were expected to be retained by non-anion exchange
mechanism (Wells, Smith et a1. 2000). No DOC was expected to be isolated by cation
exchange, with the possible exception of a small, likely negligible, amount retained
through cation bridge formation (Sutton and Sposito 2006). Additionally, it was expected
that little HA would be retained on the anion-exchange cartridge containing a MW cutoff
(Anion-1kD) since the MW range of HA is reported to be 1 to 85 kDa (Appelo and
Postma 2005).

Signiﬁcant differences were Observed in the ability of DOC samples to participate
in various types of bonding (Figures 3-2 to 3-7). For example, when comparing Waskish
Peat and Suwarmee River samples, HAS were more likely than FAS to participate in
donor-acceptor and hydrogen bonding interactions (Figures 3-2 and 3-5), while FAS were
retained more effectively by London dispersion forces than HAS (Figure 3-4). Generally,
HAS were more likely than other types of DOC to participate in donor-acceptor
interactions (Figure 3-2) and hydrogen bonding (Figure 3-5), which results in
hydrophobic interactions being relatively more important than ionic interactions for HAS
when compared to other forms of DOC (Figure 3-3). As expected, FA had a greater
amount of DOC less than 1 kDa in size than HA, with the possible exception of Pahokee
Peat HA (Figure 3-6). DOC from the RCR and DOC isolated by reverse osmosis (NOM)
showed the greatest fraction of low MW components (Figure 3-6). The cation exchanger
did not retain quantiﬁable amounts of DOC, regardless of the type (i.e. HA, FA, and

reverse osmosis isolate).

54

Differences Observed in the retention of various types of DOC ﬁom the same
source material are likely attributed to the techniques initially used to isolate and purify
the standardized material. Suwannee River HA samples were obtained by precipitation of
the hydrophobic fraction initially isolated by XAD-8 resins and extracted under alkaline
conditions (IHSS 2007). As a result Suwannee River HA is enriched in hydrophobic
constituents. NOM samples are obtained by reverse osmosis. Suwannee River NOM (RO
isolate) consists of greater amounts of hydrogen, nitrogen and oxygen functional groups
than other forms Suwannee River standards isolated by IHSS resulting in a more polar
aliphatic substance (Serkiz and Perdue 1990). Aliphatic structure with high functionality
does not offer good surfaces for hydrophobic interactions; therefore, NOM is expected to
be less hydrophobic than DOC present in natural water samples. The biases of each
isolation technique are consistent with observations of the ionic to hydrophobic ratio (%
DOC retained by the anion-F cartridge versus the extended hydrophobic) which show
Suwannee River NOM being greater than Suwannee River HA (Figure 3-3).

Of the samples tested, the proportion of DOC less than 1 kDa in size is larger for
FA than for HAS (Figure 3-6). This is consistent with known properties of HAS and FAS.
HAS typically consist of large molecules (1-85 kDa), while F As are typically composed
of much smaller constituents (0.5-1.5 kDa) (Benedetti, Van Riemsdijk et al. 1996;
Appelo and Postrna 2005). Of the three types of DOC tested, NOM is more highly
retained by Anion-1kD cartridges, which can be attributed to a large assortment of small
molecules preferentially isolated by reverse osmosis.

The fourth experiment was designed to determine if uncomplexed Cu present in

solution would be retained by the SPE media used. This was achieved by passing an

55

ultra-pure water solution spiked with 50 ugCu L-1 through each of the cartridges as

described previously. In a solution of ultrapure water, the only complexation available to

“free” ions is through hydroxides and carbonates. The unbuffered nature of ultra-pure

water results in a solution pH that rapidly decreases when exposed to air due to C02

dissolution. Both the lack of complexing ligands and low pH were expected to result in

primarily free Cu2+ ions and positively charged complexes. Therefore, it was expected

that positively charged ions would be effectively retained by cation exchange, not by
anion exchange or through non-ionic mechanisms.
When present in the ultrapure water little Cu was retained by non-cation exchange

cartridges (Figure 3-8). As anticipated, most Cu was found to exist as primarily free
species (e. g. Cu2+) in ultra-pure water, also supported by PHREEQCi modeling. The

amount of Cu retained by most cartridges was within the range of analytical error.
Retention was greatest (12%) on the anion-F cartridge. Retention by cation exchange
approached unity (91%). Surface complexation between Cu in solution and the F' counter

ion attached to the anion-F cartridge may be responsible for retention of Cu by these

cartridges. Cu is known to complex with F- (NIST 1997). However, when PHREEQCi

modeling included a hypothetical addition of ﬂuoride, the resulting aqueous
complexation did not help to explain retention and there is no clear mechanism for this
attachment.

The next objective was to determine the amount of Cu that was bound to each
DOC fraction for various types of DOC (i.e. HA, FA, reverse osmosis isolate, and non-

puriﬁed). Using the artiﬁcial solution as the solvent to simulate complexation likely to

56

occur in natural surface waters, a group of aqueous samples were prepared with a single

type of DOC and spiked to achieve a total concentration Of~100 ugCu L-l. One artiﬁcial

solution was prepared without DOC to assess the impact of inorganic complexation.
Once made, samples were stirred for greater than 2 hr to ensure Cu-binding equilibrium
(Kemdorff and Schnitzer 1980), and subjected to the ﬁ'actionation scheme described
previously. It was expected that inorganic complexation alone, mainly carbonate and
hydroxide complexes, would not result in signiﬁcant retention, while organic
complexation would dominate resulting in Cu being removed as a Cu-DOC complex
when DOC was present.

In the artiﬁcial solution without DOC present, the retention of Cu by non-ion
exchange mechanisms remained low (less than 8%), similar to Cu retention in ultra-pure
water (Figure 3-8). However, under these conditions Cu was retained by anion-exchange
cartridges. Retention ranged from 73% by the anion-F cartridge to 97% by the anion-lkD
cartridge. Speciation modeling in PHREEQCi predicted 55-98% of the Cu present in the
artiﬁcial solution existed as a neutral species, mainly as hydroxide and carbonate
complexes (Figure 3-9a). The lack of negatively charged complexes (<0.4%), does not
account for this behavior. While it is unclear why Cu was retained in these instances, one
possible explanation could be a minor mechanism of retention; cation-pi bonding through
aromatic moieties present in SDVB (Dougherty 1996) or a kinetic separation process
related to the aqueous diffusion of the metals, the hydrated ionic radius of the metals,
and the porosity/ geometry of the SPE media. Additionally, inorganically complexed Cu
was retained by both anion and cation exchange (both greater than 75%). The seemingly

simultaneous presence of anion and cation species can be explained by the following

57

possibilities: (1) the complex formed is arnphoteric allowing for both positive and
negatively charged regions of the molecule to interact with the ion-exchange cartridges,
(2) the microenvironment at the resin surface changed in pH which resulted in different
complexation conditions within each cartridge, or (3) the inorganic complexation was
relatively ﬂexible and the speciation shifted depending on the resin environment.
Multiple factors may have been involved and they suggest limitations of the current
model.

When DOC was present in solution, increased removal of Cu by SPE was
observed for all cartridges except the cation and anion-1kD cartridges, which showed
very little retention (Figure 3—8). The decrease in Cu retention by these cartridges when
DOC was present signiﬁes the dominance of organic complexation. For example, HA is
known to have a MW greater than the MW cutoff of the anion-lkD cartridge and when
DOC was present in solution, Cu retention by this cartridge was less than 6%. The
decrease in Cu retention by this cartridge suggests that Cu—DOC complexes were not able
to access the anion exchange sites. Similarly, based on the lirrrited removal of Cu by
cation exchange from solutions containing DOC, Cu appears to form stronger organic
complexes than inorganic complexes, with HA forming the strongest complex. Based on
differences in Cu retention by the anion-exchange cartridge with a MW cutoff and cation-
exchange cartridge, the primary form of retention for Cu by all cartridges is through Cu-
DOC complexation when sufﬁcient amounts of DOC are present. Like the complexation
observed with inorganic species, there are multiple reasons for why Cu complexes may
be retained as neutral, negative and positive charged species. Unlike inorganic

complexes, organic ligands commonly form amphoteric molecules, especially amines.

58

For Nordic Reservoir NOM and all forms of Suwannee River DOC, HA, FA and

NOM (reverse osmosis isolate), the following equation described the relationship

between DOC and Cu retained by SPE (R2=0.96):

Complexed Cummol) = 1.30 DOC(mm01) + 0.18 (Equation 3-4)

A Similar correlation was observed for Adrich HA, but with a different slope (R2=0.99):

Complexed Cummol) = 0.63 DOC(mm01) + 0.02 (Equation 3-5)

Binding ratios ranged from 0.63 umol Cu mmol-1 0C for Aldrich HA to 1.6 umol Cu

mmol-1 0C for Suwannee River FA.

High MW DOC was found to preferentially complex Cu (Figure 3-7). All DOC
isolates tested showed a MW BSD ratio greater than two. The only sample which showed
similar complexation for high and low MW DOC was the natural water solution. This
was the only DOC not subjected to the isolation or puriﬁcation techniques used by IHSS
(2007) necessary to create standards.

Assuming all Cu retained by SPE in the presence of DOC was complexed with
DOC; partitioning coefﬁcients for DOC (KDQC) can be calculated for each organic

fraction:

59

[Mret]
[hdtouﬂ]

KDOC (SPE) =
I()(:ret]

(Equation 3-6)

where [Mret] and [Mtotal] are the mass ( u g) of metal retained and the total mass in
solution, respectively, which is normalized to the concentration (kgC L4) of the organic
carbon fraction simultaneously retained [OCret]. Partitioning coefﬁcients for Cu ranged
from 1.4 x 104 L kgC-1, for the organic fraction of SRHA retained by the anion-1kD

cartridge, to 5.3 x 10‘5 L kgC.1 for the hydrophilic fraction of NRN.

Partitioning coefﬁcients and conditional binding constants determined during this

study (Table 3-5) are similar to values reported by others. For example, Shafer et al.

(2004) computed Cu partition coefﬁcients for colloids (KD) in samples from estuarine,

oceanic and riverine environments by taking amount of Cu associated with DOC in lkDa

permeate and dividing it by the total amount of dissolved Cu (<0.4 um) and reported log

values ranging from 5.3 to 6.4 L kg.1 DOC. Similar to the K0 calculated by Shafer et al.

(2004), Kooc was calculated by taking the ratio of Cu and DOC retained by SPE and

dividing it by the total amount Of dissolved Cu (<0.45 run) ([Mtotal]: Equation 3-6).

Using this analogous method, similar partitioning coefﬁcients were observed in this

study: log KDOC = 4.2-5.7 L kg'1 DOC. Thacker et al. (2005) reported a conditional

binding constant (log KC) of 4.0 L gC.1 for Suwannee River FA at a lower pH (6.3) by

60

directly measuring free Cuer and assuming hydrolysis and inorganic complexation were

negligible. Similarly, we are able to calculate an observed conditional stability constant

based on the observed retention of Cu and DOC by SPE (KC_SpE):

M
KC_SPE = [ m] (Equation 3-7)

[M non _ org IiOC ret]

 

where [Mm] is the mass (ug) of metal retained as an organic complex, [Mnon_org] is the

concentration (rig L-l) of metal in solution that is not complexed with DOC, and [OCm]

is the mass (g) of the organic carbon fraction simultaneously retained. Using Equation 3-

7 we produced a log KC__SPE of 2.72 L gC.1 for the same DOC sample using the anion-F

cartridge, which is less than the 4.0 L gC-1 reported by Thacker et a1 (2005). However,

Thacker et a1 (2005) assumed that hydrolysis and inorganic complexation was negligible.
While nearly all DOC was removed from solution using this cartridge (98.7%), the

assumption that all Cu not retained during these experiments was in the form of free

2+ . . . 2+ . . .
Cu likely overestimates the true concentration of Cu due to inorganic complexation.

A hybrid conditional stability constant, KC_HYBR]D (L gC-l), is produced by substituting
[Mnon_org] in Equation 3-7 with the concentration (rig L-l) of uncomplexed Cu2+

predicted by PHREEQCi as [Mﬁee].

61

[M ret]
[M free JIOC ret]

 

KC_HYBRID = (Equation 3-8)

The anion-F fraction of Suwannee River FA was found to have a log KC_HYBRID of 6.06
L gC-1 , which is two orders of magnitude larger than the Kc reported Thacker et

al.(2005). There are two possibilities which explain the why the KC_HYBRID values
reported here do not match those reported by Thacker et al. (2005): (1) uncomplexed
Cuer is not the only Cu species capable of complexing with DOC, and (2) it is not valid
to assume inorganic complexation is negligible. Regardless of the assumptions used, the
KC values reported within this paper are found to bracket the value reported by Thacker

et al. (2005).

Based on previously published data, the estimated number of binding sites

resulted in theoretical KC_MODEL values that were generally within an order of

magnitude of those observed (optimized. values), with the exception of Suwannee River

FA which resulted in KC_MODEL values that deviated signiﬁcantly from those observed

(Table 3-6). Theoretical KC_MODEL values produced using parameters reported in the
literature (i.e. the number and distribution of binding sites) showed greater variability
than KC_SpE values observed experimentally amongst the different types of DOC

employed in this study (Table 3-5). Additionally; it was not possible to account for
variations in complexation observed for NOM from different sources. Even though DOC

samples used in this study are extremely well characterized and models describing their

62

complexation exist, this information alone was not sufﬁcient to accurately describe
complexation. To accurately describe observed behavior, the number and distribution of
binding sites were adjusted until model stability constants (Equation 3-2) matched
average stability constants observed (Equation 3-7) for cartridge fractions (with the
exception of anion-lkD and cation cartridges). After optimizing binding-site densities,
the amount of Cu bound by DOM ranged from 68.7%, for the solution containing Aldrich

HA, to 93.2%, for the solution containing Nordic Reservoir NOM (Figure 3-9).

Conclusions

Results Show the method described in this paper offers a simple and rapid
approach to characterizing DOC fractions based on functional behavior and Cu binding
by these fractions. The method produces a greater number of DOC fractions than
previous fractionation techniques and retention is based on speciﬁc binding mechanisms
and molecular size. The mechanisms of retention used to isolate fractions include those
that are responsible for stabilizing humic substances (i.e. hydrogen and hydrophobic
bonding). Using this method, differences are observed in fractions of DOC from various
sources (e. g. Suwannee River and Nordic Reservoir). DOC fractions were also found to
differ based on type (i.e. HA, FA, reverse osmosis isolate, and non-puriﬁed), possibly as
a result of the techniques used to isolate sample material.

This method employs SPE cartridges which require little preparation and do not
leach signiﬁcant amounts of Cu or DOC. When Cu was present in ultra-pure water as a
free species it was not retained effectively by SPE. DOC was effectively retained and the

observed retention is consistent with characteristics known for each type of DOC used. In

63

aqueous solutions designed to mimic natural waters Cu was shown to preferentially
complex with DOC. Complexation constants were derived for individual fractions and
these values are similar to those reported elsewhere for bulk DOC using more intensive
analytical techniques.

This method provides researchers a means of rapidly describing DOC
characteristics based on speciﬁc bonding mechanisms while simultaneously quantifying
Cu-DOC complexation for each fraction. Because aqueous samples do not require
chemical treatment before analysis, complexation is expected to be representative of that
found under natural conditions. Complexation observed using this method can be directly
compared to chemical models and provides researchers a way to test theoretical

predictions with in-Situ behavior.

64

Table 3-1. Properties of SPE cartridges used to isolate fractions of DOC and Cu

 

 

complexed with DOC.
. . . Retention Molecular

Cartridge ID Cartridge Type SPE Media Mechanisms Weight Cutoff
Iminodiacetic acid exchange -

Cation BioRad Chelex 100 styrene divinylbenzene Cation Exchange none
copolymer

H-bonding Supelco DPA-6S Polyamide Resin Hydrogen Bonding none

Hydrophobic Biotage lsolute 101 Polystyrene-dryinylbenzene Hydrophobic none
copolymer

Extended . Poly(divinylbenzene-co-N- .

Hydrophobic Waters Oasrs HLB vinylpyrroli done) Hydrophobic none
Quaternary ammonium -

. . styrene divinylbenzene .
Anion-F BioRad AG MP-l . Anion Exchange none
. copolymer (ﬂuoride

counterion)
Quaternary ammonium -

Anion-I BioRad AG MP-l styrene divinylbenzene Anion Exchange none
copolymer (iodide counterion)
Quaternary ammonium -

Anion-lkD BioRad AG-l xs Styrene d’vmylbeilzene Anion Exchange lkDa
copolymer (ﬂuoride
counterion)

 

65

Table 3-2. General water chemistry parameter for Michigan State University tap water,
artiﬁcial solution, and natural Red Cedar River (RCR) water.

 

 

Parameter Tap Water @3333: RCR Units
pH 7.27.4 7.99 7.78
Conductivity 050.7 0.967 0.618 mS cm"
Hardness 404 186.1 337.2 mg L.1 as CaCO3,
Alkalinity 136.9 273.5 mg L‘1 as CaCO3
Ca 25.5 86.0 mg L"
C1 14 93.7 53.2 mg L"
M8 ' 29.8 29.8 mg L"
NO3' <0.4 0.48 2.5 mg N L"1
PO43” 5.69 1.0 mg L"
K 4.6 4.7 mg L"
Na 12 62.7 33.2 mg L"
SO42‘ . 76 49.0 55.5 mg L'1

 

66

Table 3-3. Parameters used in PHREEQCi to model Cu-DOC complexation based on
values reported in the literature and found to describe experimental observations
(optimized).

 

Sample ID: AHA NRN RCR SRF SRH SRN

 

Literature Values

 

-5d,,h,' -51:
8.93x10 8' _5dj 4.24x10
-5 -5b -5 _ - 7.43x10 _ -
nHMmoleS) 5.69x10 3 4.23xlO 2.20x10 c 7.95x105e" 56,: 3561:1056"
1.06x10 3.72xlO ’
2.271(10-5 d’g’h" '_5 d, 2.12::10'5 b
-5 -5b -5 _ - 3.29xl0 ’ _ -
nHB(moleS) 2.84x10 a 2.12x10 l.lelO C 3.98x1056" -565 1.78x1056’l
2.70x10 1.49x10 ’
0.631
r
1.09
PKHA 0.63 ‘ 0.63t 0.63t g 0.63t 0.63 t
1.19
0.9h
3.75t
r
3.03
PKHB 3.75t 3.75t 3.75t 3.75t 3.75t
g
1.17
3.80h
Optimized Values
HA 1 -5 -5 -5 -5 -s -5
n (mo es) 3.13xlO 7.78x10 7.15xlO 4.62x10 4.96xlO 7.06x10
nHB(moles) 1.56::10'5 3.89x10'5 3.58::10'5 2.31::10'5 2.48::10'5 3.53:110'S
t
PKHA 0.63 0.63 0.63 0.63 0.63 0.63
t
PKHB 3.75 3.75 3.75 3.75 3.75 3.75

 

(a) Koopal et al. (2005)

(b) Lu and Allen (2002)

(c) Higgo and Rees (1986)

((1) Ratio of binding sites based on ratio of carboxylic and phenolic content (IHSS, 2007)
(e) Ratio of carboxylic to phenolic sites based on Thurman (1985)

(f) Cabaniss and Shuman (1988)

(g) Beneditti et al. (1996)

(h) McKnight et al. (1983)

(i) Number of binding sites derived from IHSS (2007)

(t) Tipping and Hurley (1992)

67

Table 3-4. Concentrations of Cu and DOC found to leach from SPE cartridges used to
isolate DOC fractions.

 

 

Cartridge C“-1 DOC.1
ug L mg L
None (ultra pure water blank) 4.8 0.01
Cation 2.3 0.1 1
H-bonding 2.5 0.57
Hydrophobic 2.5 0.28
Extended Hydrophobic 5.0 0.34
Anion-I 2.4 0.07
Anion-F 3.1 0.06
Anion-lkD 3.9 0.63

 

68

Table 3-5. Conditional binding constants and partitioning coefﬁcients for Cu
complexation with DOC fractions (NA—not available).

 

 

Sample ID: AHA NRN RCR SRF SRH SRN
DOC Source: Aldrich Nordic Res. Red Cedar R. Suwannee R. Suwannee R. Suwannee R.
DOC Type: Humic Acid NOM DOM Fulvic Acid Humic Acid NOM
Metal: Cu Cu Cu Cu Cu Cu
Doc (mg L") 20.20 12.56 8.99 9.28 9.96 12.57
Total (3., (pg L") 95.4 95.0 65.6 95.0 93.0 95.2
Fraction SPE Conditional Binding Constant (log KC SPE. L gC'l)

Anion-F 2.04 2.80 3.02 2.72 2.72 2.80
Anion-l 2.03 2.82 3.09 2.77 2.76 2.84
H-bonding 2.09 3.17 NA 3.02 3.08 3.13
Extended Hydrophobic 2.03 2.93 3.16 2.84 2.78 2.99
Hydrophobic 2 .03 2 .87 3.13 2.81 2.71 2.98
Hydrophilic 2.03 3.65 3.41 2.93 3.08 3.06
Anion- lkD 1 .67 2.49 3.03 2.26 1.94 2.30
Fraction Hybrid Conditional Binding Constant (log KC_HYBRID , L gC'l)
Anion-F 5 .01 6.44 6.18 6.06 6.10 6.3 8
Anion-l 5.00 6.47 6.25 6.1 l 6.14 6.41
H-bonding 5 .05 6.81 NA 6.36 6.46 6.71
Extended Hydrophobic 4.99 6.57 6.32 6.18 6.16 6.57
Hydrophobic 4 .99 6.51 6.28 6.1 5 6.09 6.55
Hydrophilic 4.99 7.30 6.56 6.27 6.46 6.64
Anion-lkD 4.64 6.13 6.18 5.59 5.31 5.88
Fraction Partitioning Coefﬁcient (log K DOC, L kgC-l)

Anion-F 4.55 4.87 5.03 4.96 4.93 4.86
Anion-I 4.54 4.89 5.10 5.01 4.97 4.90
H-bonding 4.60 5.24 NA 5.25 5.30 5.19
Extended Hydrophobic 4.54 5.00 5.17 5.08 4.99 5.05
Hydrophobic 4 .54 4.94 5. 13 5.05 4.92 5.04
Hydrophilic 4.54 5.72 5.42 5.17 5.29 5.12
Anion-lkD 4.19 4.56 5.03 4.49 4.15 4.36

 

69

Table 3-6. Theoretical (based on values reported in the literature) and observed
(optimized) conditional stability constants produced by PHREEQCi modeling results for

Cu-DOC complexation.

 

 

 

Sample ID: AHA NRN RCR SRF SRH SRN
3.32 ‘1’”
Theoretical 324 a,e,t 3 15 (1,1,1 259 bieJ
d,t',i . -
log KCfMODEL 2.50 2.59 3.79 4,99 (Lg; 3 13 e,” 2.45 :14
(L 2C- ) 6.75 h,’ ' 2.45 M
4.42 d’ "
Observed
log KC MODEL 2.04 3.04 3.16 2.85 _ 2.86 2.97
-1_
ILEC )
(a) Koopal et al. (2005)
(b) Lu and Allen (2002)
(c) Higgo and Rees (1986)

((1) Ratio of binding sites based on ratio of carboxylic and phenolic content (IHSS, 2007)

(e) Ratio of carboxylic to phenolic sites based on Thurman (1985)

(f) Cabaniss and Shuman (1988)
(g) Beneditti et al. (1996)
(h) McKnight et al. (1983)

(i) Number of binding sites derived from IHSS (2007)

(t) Tipping and Hurley (1992)

70

Figure 3-1. The retention of Aldrich Humic Acid in MSU tap water at different ﬂow
rates for each of the SPE cartridges used to fraction DOC: Anion-F (A), Anion-I (El),
Anion-lkD (A), Extended Hydrophobic (I), Hydrophobic (O), and H-bonding (O).

100 l

90 ‘1

  
  

Aldrich Humic Acid Retained (%:

0 2 4 6 8 10 12
Cartridge Flow Rate (mL min")

71

 

20G mom

202
.M 8:535

202
mom 0%qu

    

<:
.m 5.233

Sm
.M vegan—m

ed
.a aaaoﬁa

<1 £222....

E
.m 5383

<m
.M oongm

 

 

_ L L _ P 1 ﬂ L1 r

a”
m —
0 0 0 0
6 4 2
Howmooo<18=oﬂ

Figure 3-2. Relative amount of donor-acceptor interactions based on the fraction of DOC
80 a

retained.
l 00

72

Figure 3-3. Ratio of DOC retained by ionic/hydrophobic mechanisms.

 

 

 

3 2 1
23M omnoameggoﬁow

ECO MUM

202
:m oogksm

EOZ
.mom 3302

Sr
.m £233

<5
2m ooﬁazﬁm

<m
.m aaxoﬁm

<3 32.634

E
.m 5283

<m
.M oo§>>=m

73

0.83

Figure 3-4. Relative amount of London dispersion forces based on the fraction of DOC

retained.

 

ECO Mum

 

202
.M conga—m

202
.mom 06.52

<:
.a 5.333

<3
.M 005.35

<:
.m aaxoﬁa

<3 now—2&4

 

. <m
.a £6.63

<m
.M 3535

 

 

r T. _
0 5
J 0.
0 0
commeonmQ .8954

0.15
0.00 ~

74

Figure 3-5. Relative amount of hydrogen bonding based on the fraction of DOC retained.

 

 

60

_
.

_
.
4 u

 

 

 

SOD 303

202
.M 35835

202
.mom £202

 

<3
.m 5333

<3
.M 85835

<:
.a 68.6%..

<3 :28:

E .a 5:83

<3
.M 8:533

_ » .
L.

5 O

3 .I.
. chcom comp—33

 

75

Figure 3-6. The fraction of DOC less than 1 kDa in molecular weight.

 

 

 

1.0

8 5 3

0. 0. 0.
«303 v DOD :ouoﬁm

EOQ 303

202
:m Doc—835m

202
.33 220 Z

<:
.m 5383

<3
.3 oogazm

E
.m 86.3

<3 not2<

E
.m 5383

<3
.3 oogasm

76

Figure 3-7. The ratio of molecular weight binding site densities.

 

 

203 303

202
.3 3535

202
.83 06qu

 

E
.a 5283

. .
. . . , .
. — . A . _
. . . . . .
. . . . _ .
. . _ . _ .
_ . _ _
2 . . . . ,
. . . L . .
_ . . . _

NA

 

A Sr
N .m 686.63

I Ea...

 

 

 

 

A E
N .a 53...;
E
.3 “anagram
6 5 1.4
6:2 Own .3:

 

 

77

Figure 3-8. The relative amount of Cu retained by SPE cartridges in different types of

solutions: artiﬁcial solution with Suwannee River FA (I), artiﬁcial solution with
Suwannee River HA (%), artiﬁcial solution with Suwannee River NOM (§), artiﬁcial

solution without DOC (El), and ultra-pure water without DOC (E).

 

 

 

2222.2???

§§§§§§§§.
Egg/2%???

 

§§§§x
7/////////////////////////////////A

   

wu.u.u.u.n.u.n.u.3.va.u.n.u.u.u.uouo"on.”.nouonon.no»o"on."ono».no"snow"...nonououououonouroro.

 

                       

Cation

 

00 6 4
Aﬁv 35863 $380

 

 

100 -——~~ — - - - , 7' 3-3--L.__-
0
0

20
0

Anion- 1 kD

Anion-F

I

Anion-

Extended
Hydrophobic

H-Bonding Hydrophobic

 

78

Figure 3-9. The fraction of free Cu2+ (I, percentage shown), Cu—hydroxide complexes

(%), Cu-carbonate complexes (1]), and Cu-organic complexes (E) following (a) the ﬁrst
modeling step (equilibrium with C02, without organic complexation) and (b) the second
modeling step (with optimized organic complexation) using PHREEQCi.

    

 

 

 

 

100 - ,
(a)

A 80 ..................
ex:
5 60 ................
§ 40 — ............
a?

20 - .....

0 ~ 1.55—r——L1.05_.

Model RCR
Solution Solution

100

hexoo
COO

Percent Cu (%)

N
O

 

O

AHA NRN RCR SRFA SRHA SRN

79

 

 

 

References

Abollino, 0., M. Aceto, et al. (2000). "The retention of metal species by different solid
sorbents - Mechanisms for heavy metal speciation by sequential three column
uptake." Analytica Chimica Acta 411(1-2): 223-23 7.

Aiken, G. R., D. M. McKnight, et al. (1992). "Isolation of Hydrophilic Organic-Acids
from Water Using Nonionic Macroporous Resins." Organic Geochemistry 18(4):
567-573.

Appelblad, P. K., D. C. Baxter, et al. (1999). "Determination of metal-humic complexes,
free metal ions and total concentrations in natural waters." Journal of
Environmental Monitoring 1(3): 21 1-217.

Appelo, C. A. J. and D. Postrna (2005). Geochemistry. groundwater and pollution.
Amsterdam, Netherlands, Taylor & Francis.

Benedetti, M. F., W. H. Van Riemsdijk, et al. (1996). "Metal ion binding by natural
organic matter: From the model to the ﬁeld." Geochimica Et Cosmochimica Acta
60(14): 2503-2513.

Bio-Rad (1998). AG50W and AGMP-50 Cation Exchange Resins; Instructions Manual.
Hercules, CA, Bio-Rad Laboratories: 30.

Burkhard, L. P. (2000). "Estimating dissolved organic carbon partition coefﬁcients for
nonionic organic chemicals." Environmental Science & Technology 34(22):
4663-4668.

Chin, Y. P., G. R. Aiken, et al. (1997). "Binding of pyrene to aquatic and commercial

humic substances: The role of molecular weight and aromaticity." Environmental

Science & Technology 31(6): 1630-1635.

80

Clesceri, L. S., A. E. Greenberg, et al., Eds. (1998). Marci Methods for the
Examination of Water and Wastewater. Washington, DC, American Public Health
Association, American Water Works Association, and Water Environment
Federation.

Croue, J. P., M. F. Benedetti, et al. (2003). "Characterization and copper binding of
humic and nonhumic organic matter isolated from the South Platte River:
Evidence for the presence of nitrogenous binding site." EnvironmentaLScience &
Technology 37(2): 328-336.

Donat, J. R., P. J. Statham, et al. (1986). "An Evaluation of a C-18 Solid-Phase
Extraction Technique for Isolating Metal Orgarric-Complexes from Central North
Paciﬁc-Ocean Waters." Marine Chemistry 18(1): 85-99.

Dougherty, D. A. (1996). "Cation-Pi Interactions in Chemistry and Biology: A New
View of Benzene, Phe, Tyr, and Trp." Sim 271(5246): 163-168.

Florence, T. M., G. M. Morrison, et al. (1992). "Determination of Trace-Element
Speciation and the Role of Speciation in Aquatic Toxicity." Science of the Tota_l
Environment 125: 1-13.

Gardner, M. and E. van Veen (2004). "Comparability of copper complexation capacity
determination by absorption by chelating resin column and cathodic stripping
voltammetry." Analﬂica Chimica Acta 501(1): 113-117.

Ghabbour, E. and G. Davies, Eds. (2004). Humic Substances: Nature's most versaLilp

materials. New York, Taylor and Francis.

81

 

 

 

Groschner, M. and P. Appriou (1994). "3-Column System for Preconcentration and
Speciation Determination of Trace-Metals in Natural-Waters." Analﬂica Chimica
{Leg 297(3): 369-376.

Hayes, M. H. B., P. MacCarthy, et al. (1989). The search for structure: Setting the scene.
Humic Substances II: In Search of Structure. M. H. B. Hayes, P. MacCarthy, R.
L. Malcohn and R. S. Swift. New York, John Wiley & Sons: 3-31.

Higgo, J. J. W. and L. V. C. Rees (1986). "Adsorption of Actinides by Marine-Sediments
- Effect of the Sediment Seawater Ratio on the Measured Distribution Ratio."
Environmental Science & Technology 20(5): 483-490.

Icopini, G. A. and D. T. Long (2002). "Speciation of aqueous chromium by use of solid-
phase extractions in the ﬁeld." Environmental Science & Technolggy 36(13):
2994-2999.

International Humic Substances Society (2007) "IHSS Product Information."
http://www.ihss.gatech.edu

Janos, P. (2003). "Separation methods in the chemistry of humic substances." Journal of
Chromatography A 983(1-2): 1-18.

Kemdorff, H. and M. Schnitzer (1980). "Sorption of Metals on Humic-Acid."
Geochimica Et Cosmochimica Acta 44(11): 1701-1708.

Kimball, B. A., E. Callender, et a1. (1995). "Effects of Colloids on Metal Transport in a
River Receiving Acid-Mine Drainage, Upper Arkansas River, Colorado, USA."

Applied Geochemism 10(3): 285-306. ‘

82

Koopal, L. K., T. Saito, et al. (2005). "Ion binding to natural organic matter: General
considerations and the NICA-Donnan model." Colloids and Surfaces &
Physicochemical and Engineering Aspects 265(1-3): 40-54.

Langmuir, D. (1997). Agiieous Environmental Geochemistry. Upper Saddle River, NJ,
Prentice Hall.

Lee, S., J. Gan, et a1. (2003). "Evaluation of Kd Underestimation Using Solid Phase
Microextraction." Environmental Science & Technology 37(24): 5597-5602.

Linnik, P. N. (2003). "Complexation as the most important factor in the fate and transport
of heavy metals in the Dnieper water bodies." Analytical and Bioanalﬂical
Chemisﬂ 376(3): 405-412.

Louchouam, P., S. Opsahl, et a1. (2000). "Isolation and quantiﬁcation of dissolved lignin
from natural waters using solid—phase extraction and GC/MS." Analy_tica1
Chemistg 72(13): 2780-2787.

Lu, Y. and H. E. Allen (2002). "Characterization of copper complexation with natural
dissolved organic matter (DOM}—link to acidic moieties of DOM
andcompetition by Ca and Mg." Water Research 36(20): 5083-5101.

Mace, J. E., C. H. Lin, et a1. (2001). The effect of an XAD-8 resin fractionation scheme
for natural DOM on the" sorption of hydrophobic organic chemicals.
Understanding and Managing Organic Matter in Soils, Sediments and Watei
Proceedings of the 9th International Humic Substances Societv. R. S. Swift and K.
M. Spark. Adelaide, Australia. September 21-25, 1998: 581-587.

McDonald, 8., A. G. Bishop, et al. (2004). "Analytical chemistry of freshwater humic

substances." Analﬁica Chimica Acta 527(2): 105-124.

83

National Institute of Standards and Technology (1997). NIST Critically selected stability
constants of metal complexes Gaithersburg, MD, US. Dept. of Commerce,
National Institute of Standards and Technology, Standard Reference Data
Program.

Parkhurst, D. L. and C. A. J. Appelo (2005). PHREEQC Interactive, US. Geological
Survey.

Perga, M. E., M. Kainz, et al. (2006). "Carbon pathways to zooplankton: insights from
the combined use of stable isotope and fatty acid biomarkers." Freshwater
Mggy 51(11): 2041-2051.

Raber, B., I. Kogel-Knabner, et a1. (1998). "Partitioning of polycyclic aromatic
hydrocarbons to dissolved organic matter from different soils." Chemosphere
36(1): 79-97.

Santschi, P. H., J. J. Lenhart, et a]. (1997). "Heterogeneous processes affecting trace
contaminant distribution in estuaries: The role of natural organic matter." Marga
Chemism 58(1-2): 99-125.

Serkiz, S. M. and E. M. Perdue (1990). "Isolation of Dissolved Organic-Matter from the
Suwannee River Using Reverse-Osmosis." Water Research 24(7): 911-916.

Shafer, M. M., S. R. Hoffmann, et al. (2004). "Physical and kinetic speciation of copper

and zinc in three geochemically contrasting marine estuaries." Environmental

 

Science & Technologv 38(14): 3810-3819.
Supelco. (2005). "Discovery DPA-6S SPE Tubes." http://www.sigmaaldrich.com.
Sutton, R. and G. Sposito (2005). "Molecular Structure in Soil Humic Substances: The

New View." Environmental Science & Technology 39(23): 9009-9015.

84

Sutton, R. and G. Sposito (2006). "Molecular simulation of humic substance-Ca-
montrnorillonite complexes." Geochimica et Cosmochimica Acta 70(14): 3566-
3581.

Thacker, S. A., E. Tipping, et al. (2005). "Development and application of functional
assays for freshwater dissolved organic matter." Water Research 39(18): 4559-
45 73.

Thurman, E. M. (1985). MC geochemistry ofnatural waters. Boston, Kluwer
Academic.

Thurman, E. M. and M. S. Mills (1998). Solidmse Extraction: Principles and Prﬁtice.
New York, Wiley-Interscience.

Tipping, E. and M. A. Hurley (1992). "A unifying model of cation binding by humic
substances." Geochimica Et Cosmochimica Acta 56: 3627-3641.

Turner, A., S. M. Le Roux, et al. (2004). "Speciation and partitioning of cadmium and
zinc in two contrasting estuaries: The role of hydrophobic organic matter."
Limnology and Oceanography 49(1): 11-19.

Waters (2003). Oasis Applications Notebook. Milford, MA, Waters Coroporation.

Wells, M. L., G. J. Smith, et a1. (2000). "The distribution of colloidal and particulate
bioactive metals in Narragansett Bay, RI." Marine Chemistry 71(1-2): 143-163.

Wen, L. S., P. Santschi, et a1. (1999). "Estuarine trace metal distributions in Galveston
Bay: importance of colloidal forms in the speciation of the dissolved phase."

Marine Chemistry 63(3-4): 185-212.

85

Yamini, Y. and A. Tamaddon (1999). "Solid-phase extraction and spectrophotometric
determination of trace amounts of copper in water samples." Talanta 49(1): 119-

124.

86

CHAPTER 4
INFLUENCE OF AROMATICITY ON COPPER COMPLEXATION BY

DISSOLVED ORGANIC CARBON

Abstract

Dissolved organic carbon (DOC) plays a signiﬁcant role in cycling of copper in
surface water systems. Previous investigations into how the molecular weight (MW) and
molecular structure of DOC inﬂuence complexation have yielded inconclusive and
contradictory results. Experiments were conducted to determine what role molecular
structure had in inﬂuencing copper complexation with high (>1 kDa) or low MW DOC.
Potential differences in how molecular structure inﬂuenced DOC verses soil organic
carbon (SOC) were also explored. The hypotheses that copper complexation with low
MW fractions of DOC is dependent on oxygen functional groups and that aromatic
structures provide strong binding sites for copper complexation were tested. An
assortment of humic standards, with known structural composition, were used to complex
copper. Copper-DOC complexes were then isolated by solid phase extraction (SPE) to
quantify complexation with high and low MW DOC. The relative strengths of complexes
were evaluated by subjecting samples to cation exchange to determining the amount of
liable. High MW fractions of DOC were found to have greater binding site densities
(BSDs) than low MW fractions. The BSD of high MW DOC was found to increase with
increasing aromaticity, while the BSD of low MW DOC was found to increase with
oxygen content. The strength of Cu—DOC complexes was found to increase with

aromaticity. Signiﬁcant differences were observed in the structural composition between

87

low MW fractions of DOC and SOC. Results demonstrate the importance of aromatic
structure in Cu—DOC complexation and suggest mechanisms for the preferential

complexation of copper by selected size fractions.

88

Introduction

In surface water systems, copper is greatly inﬂuenced by complexation with
dissolved organic carbon (DOC) (Linnik 2003). Complexation by DOC can reduce
copper toxicity (Di Toro, Allen et al. 2001) and, through the sorption of Cu—DOC
complexes to solid surfaces, can inﬂuences copper mobility (O'Day, Carroll et al. 1998).
In surface waters, up to 99.99% of copper is complexed with strong binding sites present
in DOC (Hofﬁnann, Shafer et a1. 2007). Investigations into the DOC ligands responsible
for Cu—DOC complexation have primarin focused on marine systems, where DOC
concentrations are low and solution ionic strength is high; surface water systems can
contain relatively high concentrations DOC and solution ionic strength is low (Hoffrnann,
Shafer et al. 2007). In surface water systems DOC has been found to vary in molecular
weight (MW) and molecular structure depending on the source of DOC and season
(Leenheer 1994; Penninova, F rimmel et al. 2003; Mash, Westerhoff et al. 2004). An
assessment of how MW and molecular structure inﬂuence Cu-DOC complexation would
greatly enhance understanding of Cu cycling in surface water systems.

In general, aquatic sources of DOC are reported to have greater binding afﬁnities
for copper than terrestrial sources (Mantoura and Riley 1975; Bresnahan, Grant et al.
1978; Mantoura, Dickson et al. 1978). DOC from aquatic sources are reported to contain
greater heteroaliphatic structure (Jackson 1975). Functional groups containing nitrogen,
oxygen, phosphorous and sulfur have all been suggested as possible heteroaliphatic
binding sites (Tang, Warnken et a1. 2001; Croue, Benedetti et al. 2003; Vachet and
Callaway 2003; Karlsson, Persson et al. 2006). Oxygen and nitrogen functional groups

are suspected because of their prevalence in DOC and evidence showing they form strong

89

bonds with copper inner-sphere electrons through multi-dentate bonding arrangements
(Korshin,AFrenkel et a1. 1998; Karlsson, Persson et al. 2006).

In freshwater systems, attempts to quantify the size fraction of organic ligands
responsible for copper complexation have produced mixed of results. Due to differences
in sampling techniques (e. g. membrane separation, dialysis and high pressure-size
exclusion chromatography), directly comparing data describing the MW of DOC ligands
is difﬁcult. However, the majority of copper in surface water samples is generally found
in the colloidal phase (1-10 kDa in size)(Wen, Santschi et a1. 1999), with the average
MW of DOC associated with copper reported to be around 1.3 kDa (Wu, Evans et al.
2003). The prevalence of copper in the colloidal fraction may be due to a preponderance
of DOC ligands within this size range (Sigg, Xue et al. 2000; Hoffrnann, Shafer et al.
2007). The ability of speciﬁc binding sites present within fractions of DOC to complex
copper may also be responsible (i.e. binding site afﬁnities for speciﬁc size fractions).
Generally, DOC fractions with MWs less than a few thousand Daltons are reported to
have strong binding afﬁnities for copper. Superior binding has been reported for
freshwater DOC ligands less than 3 kDa (Hoffrnann, Shafer et al. 2007), 1 kDa (V ulkan,
Mingelgrin et al. 2002; Menitt and Erich 2003) and 0.5 kDa in size (Wu, Evans et a1.
2003)

In Chapter 3, binding site densities for low MW (LMW) DOC (<1 kDa) were
found to be greater than high MW (HMW) DOC (>1 kDa) for a natural water sample;
while binding site densities (BSDs) for HMW DOC were greater than LMW DOC for
puriﬁed standards of fulvic acid (FA), humic acid (HA) and DOC isolated by reverse

osmosis, referred to as natural organic matter (NOM). The ability of speciﬁc size

90

fractions of DOC to bind copper appear to be related to the molecular structure. As MW
increases, the percentage of aromatic structure present in aquatic DOC has been found to
increase (Chin, Aiken et a1. 1994) and LMW DOC is reported to contain high amounts of
aliphatic constituents (Her, Amy et a1. 2003). Additionally, with decreasing apparent
molecular size, the proportion of phenols in humic fractions tends to increase (Christl,
Knicker et a1. 2000; Scheinost, Kretzschmar et a1. 2001).

Since LMW fractions of aquatic DOC appear to preferentially complex copper
and the presence of aliphatic structure is associated with LMW fractions (Chin, Aiken et
al. 1994; Her, Amy et al. 2003), it is reasonable to infer: (1) aliphatic portions of DOC
with Cu binding functional groups (i.e. heteroaliphatic structure) are primarily
responsible for the preferential complexation observed, and (2) as the amount of
heteroaliphatic structure increases, so too will Cu complexation. However, DOC
containing different amounts of heteroaliphatic structure (e.g. Suwannee River HA and
FA) had similar abilities to complex copper and HMW fractions of several types of DOC
have been shown to preferentially complex copper (Chapter 3; (Kogut and Voelker
2001). In the absence of clear evidence identifying the cause of the reported preferential
complexation, others must be considered. Likewise, multiple mechanisms for Cu
complexation are also possible. In addition to complexation with heteroaliphatic binding
sites, such as oxygen and nitrogen, aromatic constituents have also been found to
complex cations through strong cation-1t bonding (Dougherty 1996). Complexation via
aromatic moieties is consistent with observations that binding site densities for strong
ligands are found to degrade when exposed to UV light, while weaker ligands are

unaffected (Moffett, Zika et al. 1990; Gordon 1992). Base solely on these observations it

91

would be reasonable to suggest that not only do UV absorbing aromatic moieties
complex copper, they also offer stronger binding sites. To clarify possible mechanisms of
Cu complexation and to better understand how MW may impact Cu complexation, there
is a need to investigate the inﬂuence of DOC structure on Cu complexation.

To determine the inﬂuence of DOC structure on Cu complexation, Cu was
allowed to complex with humic standards of known elemental and structural
composition. LMW and HMW fractions of Cu-DOC complexes were then quantiﬁed.
Hypotheses tested were: (1) as the amount of oxygen present in LMW fractions of DOC
increases, so will Cu complexation, and (2) aromatic moieties in DOC provide stronger
binding for Cu complexation than functional groups containing oxygen. Additionally,
differences in the aqueous behavior of humic samples derived from aqueous (i.e. DOC)

and terrestrial sources (i.e. soil organic carbon) were also evaluated.

Materials and Methods

A variety of humic samples were characterized by the same SPE method
presented in Chapter 3. Concentrations of copper and DOC were measured before and

after SPE to determine the amount, or fraction, retained by each type of SPE media. This

method is unique because uncomplexed copper (free Cu2+) is not retained on the SPE

cartridges used to fraction DOC and speciﬁc retention mechanisms can be identiﬁed for
each fraction (e. g. hydrogen bonding), effectively characterizing DOC behavior. Two of
the SPE cartridges used utilize anion exchange resins. Nearly 100% of DOC, including
DOC that is complexed with copper, is retained by anion exchange (Chapter 3).

Additionally, one of the two anion exchange cartridges employed by this method utilizes

92

a 1 kDa molecular cutoff, essentially making it possible to determine the amount of
HMW (0.45 um-l kDa) and LMW (<1 kDa) DOC. This MW cutoff is appropriate for
studying copper complexation because it is near the average MW of organic ligands most
associated with copper in natural water solutions (Wu, Evans et al. 2003).

The strength of Cu-DOC complexes was evaluated by passing samples through a
SPE cartridge consisting of a strong cation exchanger (BioRad Chelex 100). When no
DOC is present in solution, nearly 100% of copper present in solution is retained
(Chapter 3). The strong iminodiacetate bonding site present in the cation exchange
cartridge offer a strong ionic potential capable of liberating copper from weak complexes,
effectively discriminating between labile and nonlabile copper species (Shafer, Hoffrnann
et a1. 2004). Due to the strong ionic bonding potential of the cation exchange resin, it can
be assumed that copper not retained by cation exchange would be less likely to
participate in other geochemical reactions, due to thermodynamic or kinetic limitations,
when complexed by the DOC ligand. Based on this, in solutions with DOC available for
complexation, if the amount of copper retained by the Chelex 100 resin is high, then Cu-
DOC complexation is relatively weak. Similarly, if the amount of copper retained by the
Chelex 100 resin is low, then Cu-DOC complexation is relatively strong. It should be
stressed that this assessment technique is only relative and similar methods using nearly
the same SPE media have been found to over-estimate toxicity (Florence, Morrison et al.
1992)

In order to maintain consistent experimental conditions that were representative of
natural waters while investigating DOC with an array of composition and structure,

samples were prepared using of a variety of well characterized DOC standards in aliquots

93

of a single artiﬁcial solution with inorganic chemistry similar to a freshwater sample

(Table 4-1). To study Cu-DOC complexation, appropriate amounts of 1000 mg L-1

aqueous copper standard (Fischer Scientiﬁc) were added to ﬁve of the eight aqueous

samples to achieving a ﬁnal concentration of approximately 95 ugCu L4. Once made,

samples were stirred for more than two hours to ensure copper-binding equilibrium
(Kemdorff and Schnitzer 1980).

Aldrich HA (cat. no. H1,675-2; lot no. MV01816HH) and standards from the
International Humic Substances Society (IHSS) and were selected as DOC samples
because they are well characterized and offer a range of elemental and structural
composition. The IHSS standards used were: Nordic Reservoir natural organic matter
(NOM) (cat. no. 1R108N); Suwannee River NOM (cat. no. 1R101N), fulvic acid (FA)
(cat. no. 2S101F) and HA (cat. no. 28101H); Waskish Peat FA (cat. no. 1R107F) and HA
(cat. no. 1R107H); and Pahokee Peat HA (cat. no. 18103H). NOM samples are isolated
by reverse osmosis (RO) while HAs and FAs are isolated using XAD resins (IHSS 2006).
Rigorously, Aldrich HA, Pahokee Peat HA, and Wakish Peat FA and HA are soil organic
carbon standards that have been dissolved in the aqueous phase, therefore becoming
DOC. Throughout this paper all standards will be regarded as DOC, since they are being
evaluated in the aqueous phase, unless distinction is required to improve understanding.

The percentage of aliphatic (60-0 ppm), heteroaliphatic (90-60 ppm) and aromatic

(165-110 ppm) structure for IHSS standards was determined by others by integrating

peak areas for selected ranges of chemical shifts using solution state 13C and 1H nuclear

magnetic resonance spectrometry (Aiken, McKnight et al. 1992). The composition (%0,

94

%N, %P, %S) of IHSS standards were detennined by others and are reported on the IHSS
website (http://www.ihss.gatech.edu). The chemical structure of Aldrich HA used in this
study was assumed to be the same as a sample used in a previous study (Lot no.
LE3601KE) which was characterized using methods similar to those used to characterize
IHSS standards (Malcolm and MacCarthy 1986).

Concentrations of DOC were determined by automated analysis based on the
Heated-Persulfate Oxidation Method (Clesceri, Greenberg et al. 1998) using an OI
Analytical Model 1010 Wet Oxidation Total Organic Carbon Analyzer. Concentrations
of copper were determined by inductively coupled plasma/mass spectrometry (ICP/MS)

using a Micromass Platform ICP/MS with a hexapole collision cell and an In internal

-1
standard (20 pg L ) as outlined by Standard Methods method 3125 B (Clesceri,

Greenberg et al. 1998).

Results and Discussion

The ratio of copper (as Cu-DOC) retained on anion exchange cartridges with and

without a 1 kDa MW cutoff was used to identify the fraction of DOC, HMW (0.45 um-1

kDa) or LMW (<1 kDa), that was predominately associated with copper. The DOC

ligand size (LS) ratio was calculated by the following equation:

DOC LS ratio =£M—F1\;£X—) (Equation 4-1)
F

95

where, M]: is the amount of metal, Cu (as Cu-DOC), retained by anion exchange on a

SPE cartridge without a MW cutoff (Anion-F) and Mx is the amount of metal, copper (as

Cu-DOC), retained by anion exchange on a SPE cartridge with a 1 kDa MW cutoff
(Anion-lkD). A ratio equal to one would indicate that copper is distributed evenly
between high and low MW fractions of DOC. If this ratio is above one, more copper is
associated with the HMW fraction of DOC. If this ratio is below one, copper is associated
with the LMW fraction of DOC.

Another indicator of copper complexation with different size fractions of DOC is
the ratio of binding site densities for HMW (0.45 um-l kDa) and LMW (<1 kDa) DOC.

The MW Binding Site Density (BSD) ratio is described as

 

[MP-Mx]

C — C

MW BSD ratio = F x (Equation 4-2)
(M x /Cx)

where Mp is the amount of metal retained by the Anion-F cartridge, Mx is the amount of

metal retained by the Anion-lkD cartridge, and CF and Cx are the amounts of carbon

retained by each of the cartridges. If the MW BSD ratio equals one, then binding is equal
in each fraction of DOC. As this ratio increases, BSDs become greater for HMW DOC.
As this ratio decreases, copper binding site densities become greater for LMWDOC.

Of the ﬁve solutions prepared to examine Cu-DOC complexation, all samples
showed more copper was complexed by HMW fractions of DOC than LMW fractions of

DOC (Figure 4-1). MW BSD ratios were found to be greater than one for all samples,

96

indicating the ability of HMW DOC ligands to complex copper was greater than for
LMW DOC. Nordic Reservoir NOM was found to have only slightly more copper
complexed with HMW DOC, suggesting the MW cutoff used was near the average MW
of DOC ligands for this type of DOC. A nearly equal distribution of copper between the
two MW fractions of DOC was observed despite BSDs being 138 times greater for HMW
DOC than LMW DOC. This indicated preferential complexation by HMW fractions of
DOC while the sample contained a greater proportion of LMW DOC (Chapter 3). The
importance of the quantity of DOC ligands and the ability of those ligands to complex
copper is highlighted by HA and FA samples. Both HA samples had the largest ratio of
copper complexed by HMW versus LMW DOC, despite the smallest MW BSD ratios.
The predominance of Cu in the HMW fraction of DOC for HA is not surprising since HA
is larger in MW than FA (Appelo and Postrna 2005). However, it would be inaccurate to
assume that copper complexation by FA is more strongly inﬂuenced by the LMW
fraction of DOC; rather, the small portion of HMW DOC present in FA was found to
have binding site densities nearly 300 times greater than the LMW fraction (Figure 4-1).
One reason for this may be the presence of oxygen functional groups, since FAs tend to
contain slightly more phenolic groups than HAS (Christl, Knicker et al. 2000). Overall,
both the quantity and quality of DOC ligands were found to be important in identifying
the MW ﬁ'action of DOC that complexes copper.

MW BSD ratios were found to increase with the aromatic content of DOC (Figure
4-2). This produces a relatively strong (r2=0.70) correlation described by the following

equation:

97

MW BSD ratio = 0.30 (%Aromaticity) - 3.9 (Equation 4-3)

When Aldrich HA, which is derived from terrestrial a source, was removed, leaving only

DOC obtained from aquatic sources, the trend was even stronger (r2=0.997). Because

BSDs increase for higher MW fractions of DOC when DOC aromaticity increases, and
aromaticity is correlated with MW (Chin, Aiken et al. 1994), these observations suggest
that the aromatic structure is responsible for increased Cu binding. This is consistent
with previous reports of strong cation-7t bonding capable of outcompeting inorganic
complexation in aqueous solutions (Dougherty 1996). Based on the observed linear
regression for aromaticity, the MW BSD ratio appear to be equal when the aromatic
structure of DOC is around 16%. Although untested, this may suggest that when
aromaticity falls below 16%, other structural components within the DOC molecule, such
as oxygen and nitrogen functional groups, may provide the dominant mechanism of
binding.

As the amount of oxygen present in DOC increased, the MW BSD ratio was
found to decrease (Figure 4-3). This offers ﬁrrther evidence that oxygen functional
groups play a signiﬁcant role in complexation for the LMW fractions of DOC. A similar
trend was not observed for nitrogen, or any of the other elementals previously cited as
important to copper complexation (Tang, Warnken et al. 2001; Croue, Benedetti et al.
2003; Vachet and Callaway 2003; Karlsson, Persson et a1. 2006). While nitrogenous
groups may have a higher capacity to complex copper than those of oxygen (Merritt and
Erich 2003), dramatically greater amounts of oxygen functional groups are likely

responsible for oxygen’s dominance. The strong complexation offered in multi-dentate

98

arrangements, such as a J ahn-Teller distorted octahedron, requires a large number of
binding sites (Korshin, Frenkel et al. 1998; Karlsson, Persson et al. 2006); binding sites
that are most likely to come from oxygen (42-44% of DOC mass) rather than nitrogen
species (<2% of DOC mass).

Based on the observed importance of aromaticity and oxygen composition, the

following equation was found to describe the MW fraction of DOC that preferentially

complexes copper (r2=0.99):

MW BSD ratio = 0.3(%Aromaticit y) — 2.7(%Oxygen) + 57.6 (Equation 4-4)

This equation was developed using by combining the slopes of linear regression lines
used to describe the relationship between the MW BSD ratios and aromaticity (Figure 4-
2) and oxygen content (Figure 4-3a).

As the aromatic content of DOC increased the amount of copper retained by the
Chelex 100 resin was found to decrease (Figure 4-4). No similar trend was observed for
changes in the amount of oxygen present in the same DOC samples. This suggests
bonding between copper and the n—orbitals present in aromatic moieties is stronger than
the bonding between cooper and oxygen functional groups. It also suggests the aromatic
content of DOC may be more important in determining copper availability and toxicity
than oxygen functional groups.

The inﬂuence of aromaticity and heteroaliphaticity on the amount of organic
carbon (OC) retained by ionic mechanisms from solution was different for DOC and

SOC samples (Figure 4-5). As the heteroaliphatic content of SOC increased, the

99

percentage of SOC greater than 1 kDa in MW increased. As the heteroaliphatic content
for truly aquatic DOC increased, the percentage of DOC in solution greater than 1 kDa in
MW decreased. As the aromatic content increased for bulk DOC, the percentage of DOC
greater than 1 kDa increased. As the aromatic content increased for SOC, the percentage
of SOC greater than 1 kDa decreased. The trend in aromatic content and the percentage
of DOC greater than 1 kDa is consistent with previous reports of increasing aromaticity
with increasing MW for DOC from natural waters (Chin, Aiken et al. 1994). The amount
of DOC less than 1 kDa that contained a high percentage of heteroaliphatic structure is
also consistent with reports of increased aliphatic content of LMWDOC ﬁom aquatic
sources (Her, Amy et al. 2003). Opposite trends for DOC and SOC are unambiguous. It
appears as though HMW fractions of DOC and SOC are similar in molecular structure,
while the LMW fractions of differ considerably.

The amount of DOC removed ﬁom solution by hydrophobic mechanisms was
found to increase with aromaticity for aquatic DOC (Figure 4-6). The same trend was not
observed for SOC. Donor-acceptor interactions, largely responsible for hydrophobic
bonding, are greatly enhanced through the presence of n—orbitals in aromatic rings. The
aromatic structure present within SOC molecules may not be available for external
bonding based on fact that the removal of SOC from aqueous solutions did not increase
with aromatic structure. However, other factors contributing to hydrophobic bonding,
such as the molar volume of the DOC molecule, may also be responsible.

Aromaticity was found to have a signiﬁcant inﬂuence on DOC. As the amount of
aromatic structure increased: (1) copper BSDs were greater for HMW DOC than for

LMW DOC (Figure 4-2); (2) the ability to remove copper from solutions when

100

 

complexed with DOC decreased (Figure 4-4a); (3) HMW fractions of DOC increased
(Figure 4-5b), which favors increased removal of Cu-DOC complexes from solution
through coagulation and settling; and (4) more DOC is able to be removed from solution
through hydrophobic mechanisms (Figure 4-6), which increases the removal of Cu-DOC
complexes from solution through sorption to immobile or settling solids. When the
relationships between the aromatic structures of DOC and (a) copper complexation and
(b) DOC characteristics are taken together, results suggest that copper removal from

aquatic systems will increase with increasing aromatic structure of DOC.

Conclusions

Using an assortment of standardized humic samples in solutions of the same
chemical composition, the inﬂuence of molecular structure on DOC characteristics was
investigated. Both the quantity of DOC ligands present in and the BSD of HMW and
LMW fractions of DOC inﬂuenced which MW fraction of DOC copper associated with.
Aromatic moieties and oxygen functional groups were found to be important for Cu—DOC
complexation, with the bonding between copper and aromatic moieties (cation-1t orbital
bonding) proving to be stronger than bonding with oxygen functional groups.
Aromaticity was found to be the most important structural characteristic for copper
complexation with HMW fractions of DOC, while the oxygen content appeared to be
more important for LMW fractions of DOC. Based on the aromatic and oxygen content
of bulk DOC, it was possible to describe which fraction of DOC, the HMW fraction or

the LMW fraction, would have greater BSDs for copper.

101

The percentage of heteroaliphatic and aromatic structure found in bulk DOC was
also found to correlate with the fraction of DOC greater than 1 kDa, and opposing
correlations were found for SOC versus DOC. This suggests that there are substantial
differences in molecular structure for LMW fractions of humic substance from terrestrial
and aquatic sources. As the percentage of aromatic content present in aquatic DOC
increased, the amount of DOC retained by SPE media by hydrophobic mechanisms also
increased. Overall, the aromatic structure of DOC was found to play a critical role in

copper complexation and likely inﬂuences Cu—DOC mobility.

102

Table 4-1. General chemical composition of the artiﬁcial river water solutions
investigated.

 

Artiﬁcial River Water Solution Chemistry

 

 

 

 

pH 7.99
Conductivity 0.967 mS cm.1
Hardness 186.1 mg L.1 as CaCO3
Alkalinity 136.9 mg L" as (:aco3
Ca 25.5 mg L-1
Cl 93.7 mg L"
Mg 29.8 mg L"
- -1
N03 0.48 mg N L
- -1
P043 5.69 mg L
K 4.6 mg L'1
Na 62.7 mg L"
- -1
S042 49.0 mg L
Solution Source of DOC Type of DOC Concentration DOC Concentration Cu
mg L'1 ug L.1
1 Nordic Reservoir NOM 12.6 95.0
2 Suwannee River Fulvic Acid 9.28 95.0
3 Suwannee River Hmnic Acid 10.0 93.0
4 Suwannee River NOM 12.6 95.2
5 Aldrich Humic Acid 20.2 95.4
6 Pahokee Peat Humic Acid 4.23
7 Waskish Peat Fulvic Acid 15.9
8 Waskish Peat Humic Acid 13.6

 

103

 

 

21 1- - 300 .g

r 4 (3

ﬂ a:

o

p ‘. CI)

.. . CD

16 t 1 O 3

. . 2
— ‘i
o 1
E 11
U) i

_J

l
6 i
L D *4
1 . I: 12*: i . . L i

Aldrich HA Nord. Res. Suwanee R. Suwanee R. Suwanee R.
NOM FA HA NOM

Figure 4-1. The LS ratio and the MW BSD ratio for Cu and a set of standardized DOC
samples.

104

 

MW BSD Ratio
A

 

 

 

 

l 5 20 25 3O 35
Aromaticity (%)

Figure 4-2. The correlation between MW BSD ratio for Cu and aromaticity of Nordic
Reservoir NOM (El), Suwannee River F A (O), Suwannee River HA (A), and Suwannee

R NOM (O), and Aldrich HA (I).

105

MW BSD Ratio

MW BSD Ratio

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 ‘ . a a
- ; (b)
6 - 2 6 LT .................................... ................ A. ..................
_ I *5 5 E
— 2 M . i z
4 ~ g Q 4 _ .................. g ..................
’ i m ’ i E
i m o O z
. a s 3 I? i
2 2 2 .................................... ............ am: ..................
0 t 1 ‘ ‘ i O V i ‘.
41 42 43 4 45 0.0 0.8 1.2 1.6
Oxygen (%) Nitrogen (%)
8 .. a 8 .
: (c) :(d)
....... A W
6 E .2 6 * A
E ‘5
g a:
4 1., .................. Q 4 , ..........................................................................
3 m
<>O m i o O
a I 3 - I
2 ......................................................... .................. E 2 L: .........................................................................
ow A 1 .0
0.0 1.2 .4 3.6 4.8 0.000 0.015 0.030 0.045 0.060
Sulﬁrr (%) Phosphorus (%)

Figure 4-3. The MW BSD ratio as a function of the concentration of various elements:
(a) oxygen, (b) nitrogen, (c) sulfur and (d) phosphorus. DOC samples include: Nordic
Reservoir NOM (El), Suwannee River FA (0), Suwannee River HA (A), and Suwannee

River NOM (O), and Aldrich HA (I).

106

.o .o
W A

% Cu Retained by Chelex 100
O
N

.o
.p.

.o
w

% Cu Retained by Chelex 100
p o
-- N

0.0 7 - A - L
41.6 42.0 42.4 42.8 43.2 43.6

 

 

 

 

 

15 20 25 30 35

Aromatic Structure (%)

 

l

...........................................................................

..uco-o-oo-u-u.........-oo.-~I- ...........................................

 

 

 

Oxygen (%)

Figure 4-4. Copper removed from Cu-DOC complexes by a Chelex 100 resin as a
function (a) aromatic structure and (b) oxygen content of DOC. DOC samples include:
Nordic Reservoir NOM (El), Suwannee River FA (0) Suwannee River HA (A), and
Suwannee River NOM (O), and Aldrich HA (I).

107

 

DOC >lkDa (0/2
4:. oo
o o o

N
O

 

 

 

 

0 5 .10 _ 15 20
Heteroalrphatrc (%)

 

8O
- (b)

(0/2
C

' I
a
.
.
a
.
.
u
u
.
o
.
.
.
.
.
0
u
.
u
.
.
v
.
t
.
.
.
.
.
.
.
o
o
.
o
.
.
u
.
.
o
n
s
n
o
n
.
I
.

I .
.
o
o
.
.
.
.
.
.
.
.
.
o
o
.
.
0
u
u

..........................................................................

DOC >lkDa

N
O

 

 

 

 

0 20 40 60
Aromaticity (%)

Figure 4-5. Organic carbon retained by ionic mechanisms greater than 1 kDa in size
versus the amount of (a) heteroaliphatic and (b) aromatic structure in SOC (black, oval l)
and DOC (white, oval 2). DOC samples include: Aldrich HA (I), Nordic Reservoir
NOM (E1), Pahokee Peat HA (A), Suwannee River FA (0), Suwannee River HA (A),
Suwannee River NOM (O), Waskish Peat FA (9) and Waskish Peat HA (O).

108

80

 

Hydrophobicity(%)
A O\
o o

N
O

 

 

 

 

0 l 0 20 30 40 50
Aromatic Structure (%)

Figure 4-6. The inﬂuence of aromatic structure on hydrophobicity, measured as the
percentage of organic carbon retained by hydrophobic mechanisms, DOC (white, oval)
and SOC (black). DOC samples include: Aldrich HA (I), Nordic Reservoir NOM (E1),
Pahokee Peat HA (A), Suwannee River FA (0), Suwannee River HA (A), Suwannee
River NOM (O), Waskish Peat FA (0) and Waskish Peat HA (0).

109

References

Aiken, G. R., D. M. McKnight, et al. (1992). "Isolation of Hydrophilic Organic-Acids
from Water Using Nonionic Macroporous Resins." Organic Geochemist_ry 18(4):
567-5 73.

Appelo, C. A. J. and D. Postrna (2005). Geochemistry. groundwater and pollution.
Amsterdam, Netherlands, Taylor & Francis.

Bresnahan, W. T., C. L. Grant, et al. (1978). "Stability constants for the complexation of
copper(II) ions with water and soil fulvic acids measured by an ion selective
electrode." Anal. Chem. 50(12): 1675-1679.

Chin, Y. P., G. Aiken, et al. (1994). "Molecular-Weight, Polydispersity, and
Spectroscopic Properties of Aquatic Humic Substances." Environmental Science
& Technology 28(11): 1853-1858.

Christl, I., H. Knicker, et al. (2000). "Chemical heterogeneity of humic substances:
characterization of size fractions obtained by hollow-ﬁbre ultraﬁltration."
European Journal of Soil Science 51(4): 617-625.

Clesceri, L. S., A. E. Greenberg, et al., Eds. (1998). StaLdard Methods for the
Nation of Water and Wastewater. Washington, DC., American Public
Health Association, American Water Works Association and Water Environment
Federation.

Croue, J. P., M. F. Benedetti, et al. (2003). "Characterization and copper binding of
humic and nonhumic organic matter isolated from the South Platte River:
Evidence for the presence of nitrogenous binding site." Environmental Science &

Technology 37(2): 328-336.

110

Di Toro, D. M., H. E. Allen, et a1. (2001). "Biotic ligand model of the acute toxicity of
metals. 1. Technical basis." Environmerﬂrl Toxicology and Chemism 20(10):
2383-2396.

Dougherty, D. A. (1996). "Cation—Pi Interactions in Chemistry and Biology: A New

View of Benzene, Phe, Tyr, and Trp." Science 271(5246): 163-168.

 

Florence, T. M., G. M. Monison, et al. (1992). "Determination of Trace-Element
Speciation and the Role of Speciation in Aquatic Toxicity." Science of the Total
Environment 125: 1-13.

Gordon, A. S. (1992). "Isolation of Compounds with Afﬁnity for Copper from Seawater
Using Immobilized Copper-Ion Afﬁnity-Chromatography." _M_arine Chemistry
38(1-2): 1-12.

Her, N., G. Amy, et a1. (2003). "Characterization of DOM as a function of MW by
ﬂuorescence EEM and HPLC-SEC using UVA, DOC, and ﬂuorescence
detection." Water Research 37(17): 4295-4303.

Hofﬁnann, S. R., M. M. Shafer, et al. (2007). "Strong Colloidal and Dissolved Organic
Ligands Binding Copper and Zinc in Rivers." Environmental Science &
Technology.

International Humic Substances Society. (2006). "International Humic Substances
Society." http://www.ihss.gatech.edu/.

Jackson, T. A. (1975). "Humic matter in natural waters and sediments." Soil Science

119(1): 56-64.

111

Karlsson, T., P. Persson, et al. (2006). "Complexation of copper(II) in organic soils and in
dissolved organic matter - EXAF S evidence for chelate ring structures."
Environmental Science & Technology 40(8): 2623-2628.

Kemdorff, H. and M. Schnitzer (1980). "Sorption of Metals on Humic-Acid."
Geochimica Et Cosmochimica Acta 44(11): 1701-1708.

Kogut, M. B. and B. M. Voelker (2001). "Strong copper-binding behavior of terrestrial
humic substances in seawater." Environmental Science & Technology 35(6):

1 149-1 1 56.

Korshin, G. V., A. I. Frenkel, et al. (1998). "EXAF S Study of the Inner Shell Structure in
Copper(II) Complexes with Humic Substances." Environmeﬂrl Science &
Technology 32(18): 2699-2705.

Leenheer, J. A. (1994). Chemistry of Dissolved Organic-Matter in Rivers, Lakes, and
Reservoirs. Environmental Chemistry of Lakes and Reservoirs. Washington,
AMER CHEMICAL SOC. 237: 195-221.

Linnik, P. N. (2003). "Complexation as the most important factor in the fate and transport
of heavy metals in the Dnieper water bodies." Analytical and Biogrlﬁical
Chemism 376(3): 405-412.

Malcolm, R. L. and P. MacCarthy (1986). "Limitations in the use of commercial humic
acids in water and soil research." Environ. Sci. Technol. 20(9): 904-911.

Mantoura, R. F. C., A. Dickson, et al. (1978). "Complexation of Metals with Humic
Materials in Natural-Waters." Estuarine and Coastal Marine Science 6(4): 387-

408.

112

 

 

Mantoura, R. F. C. and J. P. Riley (1975). "Use of Gel-Filtration in Study of Metal
Binding by Humic Acids and Related Compounds." Analﬂica Chimica Acta
78(1): 193-200.

Mash, H., P. K. Westerhoff, et a1. (2004). "Dissolved organic matter in Arizona
reservoirs: assessment of carbonaceous sources." Organic Geochemistry 35(7):
83 1-843.

Merritt, K. A. and M. S. Erich (2003). "Inﬂuence of Organic Matter Decomposition on
Soluble Carbon and Its Copper-Binding Capacity. " J Environ Oua_l 32(6): 2122-
213 1.

Moffett, J. W., R. G. Zika, et a1. (1990). "Distribution and Potential Sources and Sinks of
Copper Chelators in the Sargasso Sea." Deep-Seﬂesearch Part atOceanogaphic
Research Papers 37(1): 27-36.

O'Day, P. A., S. A. Carroll, et al. (1998). "Rock-water interactions controlling zinc,
cadmium, and lead concentrations in surface waters and sediments, US Tri-State
Mining District. 1. Molecular identiﬁcation using X-ray absorption spectroscopy."
Environmental Science & Technology 32(7): 943-955.

Penninova, I. V., F. H. Frimmel, et al. (2003). "Molecular weight characteristics of
humic substances from different environments as determined by size exclusion
chromatography and their statistical evaluation." Environmental Science &
Technology 37(1 1): 2477-2485.

Scheinost, A. C., R. Kretzschmar, et al. (2001). Carbon group chemistry of humic and

fulvic acid: A comparison of C-ls NEXAFS and l3C-NMR spectroscopies. Humic

113

Substances: Structures, Models and Functions. E. A. Ghabbour and G. Davies.
Cambridge, UK, Royal Society of Chemistry: 39-47.

Shafer, M. M., S. R. Hoffrnann, et al. (2004). "Physical and kinetic speciation of copper
and zinc in three geochemically contrasting marine estuaries." Environmental
Science & Technology 38(14): 3810—3819.

Sigg, L., H. B. Xue, et a1. (2000). "Size fractionation (dissolved, colloidal and particulate)
of trace metals in the Thur River, Switzerland." Aquatic Geochemistry 6(4): 413-
434.

Tang, D. G., K. W. Warnken, et a1. (2001). "Organic complexation of copper in surface
waters of Galveston Bay." Limnology and Oceanography 46(2): 321-330.

Vachet, R. W. and M. B. Callaway (2003). "Characterization of Cu(ID-binding ligands
from the Chesapeake Bay using high-performance size-exclusion chromatography
and mass spectrometry." Marine Chemistry 82(1-2): 31-45.

Vulkan, R., U. Mingelgrin, et al. (2002). "Copper and Zinc Speciation in the Solution of a
Soil-Sludge Mixture." J Environ Oual 31(1): 193-203.

Wen, L. S., P. Santschi, et al. (1999). "Estuarine trace metal distributions in Galveston
Bay: importance of colloidal forms in the speciation of the dissolved phase."
Marine Chemistm 63(3-4): 185-212.

Wu, F. C., R. D. Evans, et a1. (2003). "Separation and characterization of NOM by high-
perforrnance liquid chromatography and on—line three-dimensional excitation

emission matrix ﬂuorescence detection." Environmental Science & Technology

37(16): 3687-3693.

114

CHAPTER 5
VARIATIONS IN DISSOLVED ORGANIC CARBON

CHARACTERSITICS BASED ON LAND USE

Abstract

Despite the importance of DOC in surface water systems, little is known about the
factors responsible for determining its composition and character. Due to large ﬂuxes of
terrestrial DOC and the presence of plant biomolecules in DOC, land use/cover is
suspected of being a signiﬁcant factor. To investigate this connection, two hypotheses
were evaluated: (1) DOC from forested and agricultural land uses are larger in molecular
weight and greater in aromaticity than DOC from other sources which contained less
vegetative cover (i.e. urban land uses), and (2) DOC produced from watersheds
dominated by impervious surfaces contain greater amounts of hydrophobic constituents
since landscapes with vegetation preferentially retain hydrophobic fractions of DOC;
DOC produced in watersheds with ﬂow paths over impervious surfaces (urban) produce
hydrophobic DOC. Aromaticity (measured as normalized ultraviolet absorbance at 280
nm, NUVA), molecular weight, polydispersity and the fraction of DOC retained by
hydrophobic and H-bonding mechanisms were used to evaluate DOC characteristics. The
inﬂuence of individual land uses was assessed by sampling from sub-watersheds with
only a single type of land use (>95%) present within the catchment. Multiple sub-
watersheds containing an array of urban (e. g. industrial, high and low density residential)
or agricultural (e. g. different types of row crop, sod) land uses were sampled. Results
show (1) molecular characteristics of DOC differ as a function of land use, (2) DOC

characteristics produced by forested land uses were consistently different from other

115

types of DOC, and (3) agricultural and forest land uses appear to preferentially retain
hydrophobic fractions. Finding indicated that DOC derived from watershed with

dissimilar land uses will likely have different characteristics.

116

Introduction

In surface water systems dissolved organic carbon (DOC) constitutes the base of
aquatic food chains (Lennon and Pfaff 2005) and plays a critical role in the transport of
many organic and inorganic molecules (Chin, Aiken et al. 1997; Santschi, Lenhart et al.
1997; Ghabbour and Davies 2004). Despite the importance of DOC in surface water
systems, the processes responsible for the production of DOC as well as its behavior and
characteristics are not well deﬁned. This limited understanding of DOC formation and
transformation make it difﬁcult to accurately predict the role of DOC across a variety of
surface water systems. To resolve this lack of understanding, there is a need to assess the
environmental processes responsible for controlling DOC quantity and quality, especially
within a range of land use inﬂuences (Kalbitz, Solinger et al. 2000).

Based upon differences in composition, DOC can be broadly classiﬁed as derived
from terrestrial and aquatic processes. The production of DOC by aquatic biota are the
result of the exudation and excretion of biomolecules and the decay of organisms from all
trophic levels (Tranvik 1993). Terrestrial sources of DOC include exudates from
vegetation, litter decomposition, soil leachates, microbial enzymes and biomass
(Wickland, Neff et a1. 2007). Because peat lands provide all of these processes, they have
long been recognized as a signiﬁcant source of DOC to surface waters and much of the
information concerning terrestrial DOC is based on this source (Hemond 1990,
Mullholland et al. 1990, Sachse et al. 2000).

Due to the inﬂuence of algae and bacteria present in surface water ecosystems,
aquatic DOC can be expected to contain larger amounts of carboxyl functional groups

and are more aliphatic than terrestrial sources (Zumstein and Bufﬂe 1989; Croue,

117

Benedetti et al. 2003). While DOC from terrestrial sources is generally lower in
molecular weight (MW) than that from aquatic sources (Sachse, Henrion et a1. 2005),
very high MW (>100 kDa) DOC has been attributed to production of organic molecules
by algal during blooms freshwater lakes (Cole, McDowell et al. 1984). Algal-derived
DOC is also reported to be low in aromaticity (McKnight, Andrews et al. 1994).
Compared to aquatic sources, DOC from terrestrial sources contain a greater amount of
aromatic structure with phenolic functional groups, and generally higher in MW (Jackson
1975; Zurnstein and Bufﬂe 1989; Croue, Benedetti et al. 2003; Linnik 2003). Similarly,
soil pore water DOC is known to be highly aromatic in structure, have a high in MW and
can be characterized as having a large hydrophobic fraction (Chin, Traina et al. 1998;
Wickland, Neff et al. 2007). Additionally, a correlation between the aromatic content,
approximated by measuring the absorbance of ultraviolet light at 280nm normalized to
the concentration of DOC, and the average molecular weight (MW) of DOC has been
found for multiple samples from surface water bodies (Sachse, Henrion et al. 2005).
The concentration and quality of DOC derived from terrestrial sources are
inﬂuenced by hydrologic factors, sorption reactions and microbial processes
(Kawahigashi, Kaiser et al. 2004). The ﬂuxes of terrestrial DOC ﬁ'om soils are highly
dependent on source strength and the amount of water moving through soils to surface
water (Hope, Billett et al. 1994). Microorganisms selectively degrade carbohydrates,
organic acids and proteins (Marschner and Kalbitz 2003). Due to microbial metabolism,
relatively hydrophilic (aliphatic) DOC is transformed to more hydrophobic (aromatic)
DOC. The hydrophobic fraction of DOC is selectively sorbed by lignin degradation

products (Guggenberger, Zech et al. 1994; Kaiser, Arscott et al. 2004) and microbial

118

processes generally enhance the removal of DOC from the aqueous phase. As a result
DOC present in peat has been found to increase in aromaticity with long-term agricultural
practices (Kalbitz, Geyer et al. 1999). Overall, adsorption is thought to inhibit the
transport of DOC from terrestrial sources more than biodegradation (Qualls and Haines
1992)

Due to the integral relationship between biological and geochemical processes
and DOC production, the type of land cover, or land use, present within watershed
catchments is likely to control DOC quantity and quality. Surface water chemistry is
dependent on the catchment characteristics (Molot and Dillon 1997) and the
concentration of both DOC and 1i gnin have been found to vary signiﬁcantly depending
watershed characteristics (Eckard, Hemes et al. 2007). The chemical composition of
DOC derived from Boreal forests has been found to vary signiﬁcantly depending on the
type of vegetation present (Wickland, Neff et al. 2007). Furthermore, DOC
characteristics, measured by the isolation of DOC fractions and UV absorbance at various
wavelengths, have been shown to vary depending on the origin of DOC samples, whether
from forested or agricultural land, or raw or treated sewage (Imai, Fukushima et al.
2001)

Previous studies by others on the inﬂuence of land use tend to focus on DOC
production, not DOC composition or quality (e. g. the ability to bind trace-metals and
serve as a substrate for microbial growth) (Richey, Brock et a1. 1980), most of which
have focused on agricultural or forestry practices rather than the inﬂuence of urbanization
(Kalbitz, Solinger et al. 2000). Part of the difﬁculty in assessing the impact of land use

and biological processes on DOC formation and transformation is identifying unique

119

signals characteristic for each in large watersheds where their inﬂuence is mixed.
Additionally, in surface waters with large watersheds, degradation processes weaken
relationships observed between DOC composition and terrestrial processes (Frost, Larson
et al. 2006). However, it is possible to show differences in DOC from small scale
watersheds that are not visible in larger watersheds (Dalzell, Filley et al. 2007).

Hedges (1980) originally proposed that unique chemical differences exist for
DOC from different landscapes. To our knowledge, no researcher has evaluated this
hypothesis across a wide range of agricultural and urban landscapes. To further test this
hypothesis, surface water runoff was collected from sub-watersheds which contained
unique land use signatures and DOC quality was assessed through a variety of analytical
techniques. In addition to the broad hypothesis proposed by Hedges (1980), two
additional. hypotheses were tested: (1) sub-watershed with agricultural and forested land
uses, which are dominated by vegetation, produce DOC with larger molecular weight and
with greater aromaticity than sub-watersheds with urban land use, and (2) sub-watersheds
with indirect ﬂow paths through soil and leaf litter (agriculture, forested, etc.)
preferentially remove hydrophobic fractions of DOC; sub-watersheds with direct ﬂow

paths over impervious surfaces (urban) produce hydrophobic DOC.

Materials and Methods

The concentration of DOC was determined by automated analysis based on the
Heated-Persulfate Oxidation Method (Clesceri et al., 1998) using an OI Analytical Model
1010 Wet Oxidation Total Organic Carbon Analyzer after passing through a 0.45pm
glass ﬁber ﬁlter that was double acid washed. Prior to conducting experiments, all

glassware used during DOC analysis was acid (18% HCl) washed, rinsed with DDI water

120

and placed in a 550°F oven for more than 2 hours to ensure cleanliness. DOC
characteristics were measured by assessing the amount of aromaticity, the molecular
weight, polydispersity and the amount of DOC retained by hydrophobic and hydrogen
bonding (H-bonding). The aromaticity of DOC was approximated by normalized
ultraviolet (UV) absorbance at 280 nm, NUVA (Chin, Aiken et al. 1994). The extent of
UV absorbance was determined using a Shimadzu UV-l60 spectrophotometer by
measuring the amount of UV light at a wavelength of 280nm absorbed and normalizing it
by the concentration of DOC resulting in units of L gC'l cm". A wavelength of 280nm is
a more effective surrogate for measuring the aromatic content of DOC than 254nm for
two reasons: (1) the transfer of electrons between overlapping n-orbitals occurs at this
wavelength for phenolic and other humic like organ substances(Traina, Novak et al.
1990), and (2) nitrate, which also absorbs UV light and is ubiquitous in natural waters,
does not absorb UV light at 280nm (Chin, Aiken et al. 1994).

The molecular weight (MW) and polydispersity of DOC was determined by size

exclusion chromatography (HPSEC) (Chin, Aiken et al. 1994; Zhou, Cabaniss et al.

2000). The number-averaged MW (Mm) and weight-averaged MW (MW) were calculated

by the following equations:

N N
Mn = Zia/2114M» (Equation 5.1)
i=1 i=1
and
N N
Mw = Emma/211, (Equation 52)

i=1 i=1

121

where hi is the height and M, is the molecular mass of the sample HPSEC eluted at

volume 1'. The MW is commonly referenced as the average MW and this custom will be

maintained throughout the rest of the paper. Polydisperity (p) is the ratio of the weight-

averaged MW and the number-averaged MW:

 

(Equation 5-3)

Low polydispersity indicates a DOC with a relatively narrow range of molecular weights.
The HPSEC system employed utilized a Gilson Model 303 pump (Middleton, WI), a
Waters Protein-Pak 125 modiﬁed silica column (Milford, MA) and UV detection at 254
nm on a Dionex Variable Wavelength Detector (Sunnyvale, CA) (Chin, Aiken et al.

1994; Zhou, Cabaniss et al. 2000). The mobile phase consisted of 0.1 M NaCl, 0.002 M

KHzPO4 and 0.002 M NazHPO4 solutions buffered to an approximate pH of 7 and

calibration was performed using random coil sodium polystyrene sulfonates
(Polysciences, Inc.) (1.8, 5.4, 8 and 18 kDa) and acetone (58 Da) (Zhou, Cabaniss et al.
2000). The amount of DOC retained by hydrophobic and H-bonding was determined by
passing ﬁltered DOC samples through solid phase extraction (SPE) cartridges and
measuring the concentration before and after to obtain the fraction, or percentage, of
DOC retained. Two types of hydrophobic cartridges, one designed to retained organic
constituents primarily through donor-acceptor binding (identiﬁed as hydrophobic) and

another with a more hydrophilic SPE media (identiﬁed as extended hydrophobic) were

122

used to asses the hydrophobicity of DOC. These fractions have been shown to relate to
DOC structure (Chapter 4) and details of the method used to isolate these fractions can be
found in Chapter 3.

Samples were collected from 48 different locations within the Grand River
watershed in central Michigan. Of the 48 locations, 29 were from sub-watersheds
comprised of only one type of land use (>95% of area) present within the tributaries of
the Red Cedar and Looking Glass Rivers. For these 29 sub-watersheds, land use was
identiﬁed based on the Michigan Land Cover/Use Classiﬁcation System (MDNR 2001)
(Table 5-1). Potential sampling locations were initially identiﬁed with the aid of
geographic information systems (GIS) by overlaying hydrologic information on land use
classiﬁcations to determine rough boundaries of sub-watersheds that would contain one
type of land use in ArcMAP 9.1 (ESRI 2003). Sub-watersheds with unique land use
characteristics were identiﬁed by conceptually moving hypothetical sampling locations
from the farthest point upstream (i.e. water source or headwater) of the smallest
hydrologic units (intermittent streams) down stream until just before the catchment would
contain more than one land use. This point, where the sub-watershed still consisted of a
single type of land use, was identiﬁed as a potential sampling location. Potential sub-
watersheds were then inspected to determine if it was logistically possible to collect
samples. Ideally, sample locations were selected where stormwater runoff could be
collected from outfalls or in open channels where water was free ﬂowing so that no
backwater effects would result in mixing from downstream sources. Once sampling
locations were conﬁrmed, sub-watersheds were manually delineated based on topography

in GIS to ascertain the true sub-watershed boundaries. County digital elevation maps

123

were used to determine watershed topography (MDEQ 2005). For urban systems where
storm sewer networks alter natural watershed boundaries, sewer maps from the City of
East Lansing and Michigan State University were used to accurately describe catchment
storm sewer networks. Land use data was based on the IFMAP/GAP Lower Peninsula
Land Cover raster data set (MDNR 2001). Land use within the Ramey Chandler drain in
Ingham County was updated to include the type of land use present at the time of
sampling. Details of the method used to delineate sub-watersheds are available in the
appendices.

The inﬂuence of terrestrial processes on DOC was isolated by collecting samples
directly from storm sewer outlets and ﬁ'om ephemeral streams and ponds within 24 hours
of the start of runoff events. While signiﬁcant effort was made to eliminate the inﬂuence
of aquatic process on DOC, it is impossible to completely isolate only terrestrial
processes. Because samples were collected from surface water at golf courses and
ephemeral ponds that formed after heavy rains in forested areas, the DOC collected from
these two locations are undoubtedly inﬂuenced in someway by aquatic processes (i.e.
rapid microbial and algal growth). Both sample locations contained signiﬁcant algal mats
which remained present at the sampling location even during the longest stretches of dry
weather. Runoff generated from agricultural ﬁelds also requires some pooling before
generating overland ﬂow. As a result of these limitations in sampling logistics, the
inﬂuence of aquatic processes on DOC characteristics is assumed to follow the order
urban << agricultural < forested ~ recreational golf course for samples collected within
unique sub-watersheds. The 19 samples collected from the Grand and Red Cedar rivers,

where multiple land use types are present upstream from the sampling location, aquatic

124

processes were also assumed to be a major factor in determining DOC quantity and
quality.

SYSTAT (version 12.02.00; San Jose, CA) was used for all statistical analysis.
Unless otherwise noted, an a-level of 0.05 was used to determine signiﬁcance.
Questionable data was not included in statistical analysis if it was deemed an outlier by
Dixon’s Q-test at the 95% conﬁdence level (Rorabacher 1991). The standard analysis of
variance (AN OVA) was used to evaluate if differences existed in sample groups and
Tukey’s Honestly Signiﬁcant Difference (HSD) test was used to evaluate if differences
existed between speciﬁc land uses. For the purpose of statistical analysis, similar samples
were grouped according their primary land use classiﬁcation: urban (URB), agricultural
(AG), forested (FOR) or mixed (MD(). Samples collected from Michigan State
University storm sewer outfalls (MSU), automobile parking lot (PL) and recreational golf
course (GC) were analyzed separately from all other types of urban samples because of
their sample size and shared characteristics (i.e. MSU has parking lots and manicured
lawns that are similar to PL and GC). An AN OVA was conducted on the replicate
samples from sub-watersheds with urban land uses other than those speciﬁcally ﬁom
MSU campus, parking lots and golf courses and no statistically signiﬁcant difference was
observed in DOC concentration and the parameters used to assess DOC characteristics
for each sampling location. As a result, these samples were grouped as urban (Table 5-1).
Similar to urban samples, an ANOVA was performed on all samples from sub-
watersheds composed of only agricultural land uses. Samples from sub-classes of
agricultural land use were found to differ based on the concentration of DOC (p-value =

0.041) and the amount DOC retained by H-bonding (p-value = 0.03 8). A difference in

125

DOC concentration was found between sub-watersheds planted with corn versus sugar
beets (p-value = 0.027) and a difference in the amount of DOC retained by H-bonding
was found between sub-watersheds with corn versus sod (p-value = 0.032). However,
samples from all agricultural land uses were grouped together as AG because: (1) MW,
polydispersity, hydrophobicity were not found to be signiﬁcantly different and (2) the
number of samples were too small to obtain a representative population for individual
types of agricultural land use; there as only a maximum of four replicates for speciﬁc
subgroups of agricultural land use.

Based on land use classiﬁcations, sets of data for each land use type were further
grouped for statistical analysis to determine if there were statistically signiﬁcant
differences that could be attributed to land use characteristics or aquatic transformation
processes. Group 1 (URB, FOR, MSU and AG) was intended to determine if there was a
signiﬁcant difference in DOC characteristics between land uses. Group 2 (GC, PL and
URB) was evaluated to determine which of the primary components of URB landscapes
inﬂuence DOC characteristics. Group 3 (AG, FOR, GC and MIX) was used to determine
if there was a signiﬁcant difference between samples likely inﬂuenced by aquatic
processes. To provide context for solution chemistry, the general water chemistry for
each of the grouped land uses is described in Table 5.2. Alkalinity was measured by Gran
titration and the pH, conductivity and DO were measured using a Horiba U-10 water

quality analyzer.

126

Results and Discussion

The greatest concentrations of DOC found in surface water samples were from

sub-watersheds with forested and golf course land uses, around 22 mg D1 (Figure 5-1). A

high amount of variability in DOC concentrations was observed at each sampling
location, with the exception of DOC derived from AG and MIX samples. Among the four
primary types of land use, samples from forested sub-watersheds were found to have a
DOC concentration greater (p-values < 0.001) than from agricultural, MSU and urban
land uses. The concentration of DOC from GC samples was found to be greater than from
PL (p-value = 0.002) and URB (p-value < 0.001). Among the four types of samples
collected from sub-watersheds where terrestrial processes were suspected, DOC
concentrations from forests and golf courses were found to be signiﬁcantly different from
agricultural and mixed land use (p-values < 0.001).

The MW of DOC from forests was found to be around 2 kDa in size, while all
other sample were found to have MWs less than 1.5 kDa, which was statistically
signiﬁcant for all contrasts evaluated (Figure 5-2). The average MW of DOC from
primary land uses followed the order forested > agricultural > urban ~ MSU, with only
urban and MSU samples not being statistically different. The average MW of DOC ﬁ'om
urban sub-watershed, 1.22 kDa, was found to be signiﬁcantly different from, although
between, the MW of DOC from both golf courses, 1.43 kDa (p-value = 0.011), and
parking lots, 0.98 kDa (p-value = 0.004). No statistically signiﬁcant difference was
observed between the average MW of DOC from agricultural, golf course or mixed land

uses.

127

Based on a mean polydispersity of 2.0, samples from forested sub—watersheds
were signiﬁcantly different from all other sources of DOC; with the exception of parking
lots which had a polydispersity of 1.6 (Figure 5-2). There was considerable variation in
polydispesity observed in both forested and parking lot samples. No statistical difference
was observed between DOC from agricultural, MSU or urban land uses and between
DOC from urban and golf course or parking lot samples.

Sub-watersheds assumed to have the highest amount of vegetation (AG, FOR,
CG) were found to contain DOC with higher values of NUVA (Figure 5-3). The NUVA
was found to follow the order forested > agricultural > urban > MSU, and all differences

were found to be statistically signiﬁcant with the exception of differences between AG,

2.22 L gC-1 cm-l, and MSU, 1.58 L gC-1 cm-l. No statistically signiﬁcant difference was

observed between samples collected from golf course, parking lot and urban sub-
watersheds and the only signiﬁcant difference between samples potentially inﬂuenced by

aquatic processes was between samples from FOR and mixed land uses (p-value =

0.003).
A strong correlation (r2 = 0.89) was observed between the average MW and

NUVA for DOC from sub-watersheds with forested land cover (Figure 5-4). This

relationship can be described by the equation:

NUVA = 0.0015 x MW — 0.035 (Equation 5-4)

128

Although DOC from other sampling locations appears to also cluster around this line, a
clear trend was not observed.

For the primary land use types, DOC hydrophobicity followed the general order
MSU > urban >> agricultural > forested, with DOC from MSU and urban samples being
signiﬁcantly (p-values < 0.02) greater than agricultural and forested samples (Figure 5).
Differences in hydrophobicity between sample types were approximately the same for
both types of hydrophobic cartridges used. The amount of DOC retained through H-
bonding tended to follow the same trend as hydrophobicity, with the exception of DOC
from forested sub-watersheds. Based on H-bonding, forest samples were signiﬁcantly (p-
values < 0.015) greater the other primary types of land use (URB, MSU and AG) as well
as MIX samples (p-value = 0.001). For both measures of hydrophobicity and the extent
of DOC retained by H-bonding, no statistical difference was observed between DOC

from urban and parking lot samples.
A strong correlation (r2=0.89) was observed between the NUVA and the

hydrophobicity of DOC from sub-watersheds with urban land use (highlighted by oval l

in Figure 5-6). This trend can be described by the equation:
Hydrophobic DOC = 22.7 x (NUV A) — 2.3. DOC (Equation 5-5)
DOC collected from sub-watersheds intended to account for two of the primary types of

land cover, manicured lawns (GC) and paved surfaces (PL), present in urban landscapes

do not show similar trends, although DOC from parking lot runoff does plot within the

129

same region as urban samples. DOC samples from golf course (highlighted by oval 2 in
Figure 5-6) did not increase in hydrophobicity despite increasing in NUVA.

The high concentration of DOC from forested watersheds (Figure 5-1) is
consistent with previous reports of DOC generated from deciduous forests (Moore and
Jackson 1989; Park and Matzner 2003). The high concentration of DOC in both forests
and golf courses may be due to the inﬂuence of aquatic processes since samples collected
ﬁom these locations were collected from pooling surface water rather than directly from
stormwater runoff. However, according to the size-reactivity continuum model proposed
by Amon and Benner (1996), DOC altered by heterotrophic bacteria are expected to
decrease in MW. Samples from sub-watersheds with forested land were found to have the
highest MW of all locations sampled (Figure 5-2). Furthermore, ﬁrlvic acids derived ﬁom
algae are marked by low aromaticity (McKnight, Andrews et al. 1994) and DOC ﬁ'om
forested sub-watersheds were found to have the highest molar absorbitivity, which is
directly related to the amount of aromatic structure in DOC (Chin, Aiken et al. 1994).
This molecular characterization does not support the supposition that aquatic processes
are responsible for the high DOC concentration observed in water samples from sub-
watersheds with forests and golf courses. Regardless of the cause, the concentration of
DOC in surface water runoff was found to vary depending on land use and the
concentration was generally consistent with the assumed amount of vegetative cover:
forested >agricultural > urban.

Differences in the MW and aromaticity of DOC were also observed between the
primary types of land use, supporting the hypothesis originally proposed by Hedges

(1980) that unique chemical differences exist for DOC from different landscapes. Sub-

130

watersheds with land uses comprised primarily of vegetative cover produced DOC higher
in molecular weight than urban and parking lot runoff (Figure 5-2). The MW of DOC
observed from other locations was within the range observed in other surface water
systems (Chin, Traina et al. 1998; Sachse, Babenzien et al. 2001; Maurice, Pullin et al.
2002; Frost, Larson et al. 2006). The distinctly larger MW of DOC from FOR sub-
watersheds is consistent with an expectedly large input of plant exudates, such as
phytosterols and triterpenoids (J affe, Rushdi et al. 2006). The relatively high retention of
DOC via H-bonding from sub-watersheds with forested land cover (Figure 5-4) is
consistent with large amounts of plant exudates, such as terpenoids and ﬂavonoids which
can be effectively isolated by the H-bonding cartridge (Supelco 2005), are known to be
biomakers for higher plants (J affe, Rushdi et al. 2006) and are major precursors of
Suwannee River fulvic acids (Leenheer and Rostad 2004). Like the concentration of
DOC observed in surface water samples, the MW was found to follow the presumed
order of vegetative cover: FOR>AG>URB.

The larger polydispersity values observed for DOC from sub-watersheds with
forested land cover indicates a more diverse MW assemblage than for the other
landscapes evaluated (Figure 5-2). The majority of polydispersity values observed are
similar to those reported for surface waters in other studies (Chin, Traina et al. 1998;
Zhou, Cabaniss et al. 2000). While one might expect low polydispersity for samples from
parking lot runoff, given the low diversity of land cover, and larger polydispersity for
samples from mixed watersheds, given the heterogeneous watershed characteristics, the
opposite was observed. One explanation could be due to the wavelength used for DOC

detection using HPSEC. At a wavelength of 254nm, -C=C- bonds are preferentially

131

detected and non-chromophoric DOC components (e.g., polysaccharides) are not
included in the molecular weight determination and thus are not reﬂected in the
calculated polydispersity values. (O'Loughlin and Chin 2004). This may be particularly
relevant for this study because the sample collection scheme was designed to minimize
the inﬂuence of aquatic processes and the preferential degradation of polysaccharides and
other non-saturated components of biomolecules. The high polydispersity observed for
parking lot runoff may be attributed to an anthropogenic source of DOC, rather than plant
derived biomolecules.

The aromaticity of DOC, measured as NUVA, generally appeared to increase
with MW for all samples, although DOC from sub-watersheds with forested land uses
were found to have a strong correlation (Figure 5-4). This trend is consistent with the

observations reported by Sachse et al. (2005) for speciﬁc ultraviolet absorbance at 254

nm (SUVA254). It is important to note that the trend observed by Sachse et al. (2005) was

for DOC collected from surface water bodies where aquatic processes were likely.
Additionally, the MW of DOC was greater (3.5—5.5 kDa) and a greater amount of C-

bonds absorb the ultraviolet light with a wavelength of 254 nm than 280 nm used to
determine the NUVA in this study, hence a larger UV absorbance (SUVAZ 54 = 2.3-4.3 L
mgC.1 m-l) was observed by Sachse et a1 (2005).

The amount of hydrophobic DOC present in samples from locations believed to
be inﬂuenced only by terrestrial processes was greater than from locations where some

aquatic processes were likely (Figure 5-5). The linear relationship (highlighted by oval)

132

between average molecular weight and NUVA for DOC from sub—watersheds with

forested land use can be described by the equation (r2=0.89):

NUVA = 0.0015 x MW — 0.035. (Equation 5-6)

Observed differences in hydrophobicity are not explained by the inﬂuence of pH on DOC
protonation (Table 5-2). One plausible explanation for the observed difference in
hydrophobicity is due to the hydraulic connectivity present within urban land uses.
Samples collected from MSU, parking lots and urban land uses were transported over
impervious surfaces and through sewer networks, whereas the ﬂow path for other types
of land use included extensive contact soil, vegetation and detritus. Forested soils have
been shown to remove hydrophobic fraction of DOC from aqueous the phase (Meier,
Chin et al. 2004) and low amounts of hydrophobic DOC are retained by soils with low
microbial activity result (Kawahigashi, Kaiser et al. 2004). Further evidence to support
the hypothesis that sub-watersheds with indirect ﬂow paths preferentially retain
hydrophobic ﬁactions of DOC is the lack of a connection between aromaticity, measured
as NUVA, and the hydrophobic content of DOC (Figure 5-6). The presence of aromatic
structure generally increases hydrophobicity (Chapter 4). Hydrophobicity was only found
to increase with NUVA for urban and parking lot samples. However, the hydrophobic
fraction remained relatively constant for DOC from other types of land use (e. g. GC)
despite a wide range of aromaticity.

Statistical analysis performed on data when grouped to determine the inﬂuence of

land use and aquatic processes revealed: (1) DOC characteristics were different among

133

the main types of land use investigated (URB, FOR, MSU and AG), (2) DOC
characteristics from urban land use more closely resembled that from parking lots rather
than golf course, and (3) there were signiﬁcant differences between samples that may
have been inﬂuenced by aquatic processes. DOC from sub-watersheds with the four main
types of land use were found to be statistically different (p-values S 0.001) for each of
the parameters used to measure DOC characteristics. As a result, it is important to note
that DOC ﬁ'om forested sub-watersheds is not representative of DOC from other types of
land use. No statistical difference (a = 0.05) was observed between DOC from urban and
parking lot samples in the total concentration of DOC concentration, NUVA,
hydrophobicity (based on the retention on two types of hydrophobic SPE cartridges) and
the amount of DOC retained by H-bonding. Similarities in DOC characteristics between
urban and parking lot runoff suggest the inﬂuence of paved surfaces dominate urban
processes. Unlike other studies where the signal produced from terrestrial process was
muted by aquatic process (Frost, Larson et al. 2006), all measures (NUV A, MW,
polydispersity, total concentration, percent hydrophobic and percent H-bonding) showed
a statistically signiﬁcant difference between urban, forested, golf course and mixed land
uses, with the exception of the amount of DOC retained via the extended hydrophobic
cartridge (F-ratio=2.78, p=0.052). The consistent differences observed were undoubtedly
due in part to a sampling scheme which isolated sub-watershed with unique land uses and
obtained samples directly from overland ﬂow before reaching surface water bodies where

aquatic transformation processes were likely.

134

Conclusion

Statistically signiﬁcant differences in DOC concentrations and characteristics
(MW, polydispersity, aromaticity, and fractions of DOC retained by H-bonding and
hydrophobic mechanisms) were observed based on land use. These differences were
identiﬁed by isolating sub-watersheds with unique land uses and sampling stormwater
runoff to minimize the impact of aquatic transformation processes (i.e. microbial and
physicochemical). Sub-watersheds dominated by vegetation, such as agricultural and
forested land uses, were found to produce DOC higher in MW and aromaticity than sub-
watersheds with urban land uses. Variations in the hydrophobicity observed support the
hypothesis that hydrophobic fractions of DOC are preferentially removed during the
transport of surface water runoff from sub-watersheds with vegetative cover, such as
agricultural and forested land. Results of this study support the hypothesis proposed by

Hedges (1980) that unique chemical differences exist for DOC from different land uses.

135

 

E N.

 

N .N. N. S v. 5 N. N. 5 N. .352 E:
N a a N N E N .NV a 2.: as 86%;: 2%: .Na
. . . . . _ . $832.... E6562 :4
N N. N. N N 8 N .N. N. :38. 3.83... N. :0...
a a a N N N .. Ea... 8N .NNN
. . . N o o . 23.32.... 35.5.5 NN
E N E N N E N N N .Nv N 98.. 3am .36 33. 3523:? : .N
N N N N N N N 58 ago 33. 233:? N. .N
8 N. .N. e. N. E N N N 5 v. .38. 3.3.3.5... N u<
.1: N. N. N. N N N. N. 62.39 :5 .2835:— .NNN. 0..
N 8 .. N. N N E N. N. .3 9.1:... 6.38.33. as. 4..
N N.. 8 N.. .N. N NN N.. am 5.22.5 3.3 53.5.2 «EN. 2N2
o N o . _ N N NooesﬁezEesam .N.
N N N N N N N Naemaﬁztaﬁo NcaaoNN NN.
. _ . o o . . 4:». 656.. 3:335: N :
N N N N N .4 E N 5.993%". 235 N:
o . _ o o _ _ 88-33 .NNENNNEE N:
a a a a a N 3 a 955$? .N:
E . . . o o . . 3:828”. :
. _ o _ . . _ a: 23m .N :35 _
E N. a. N. N. N. .N .N. NN :33. 535 . $5
632382: 2235
- a
9.6.3.. = £2.35... 3.5:... £225.56.— 53262 <>az co: .3: Eu: 8: 2.3 9:32 95..

 

doom #2sz 5385820 39.960 v.84 53:32 - 0301:). .65 no woman commas—o 89$ 8388 =< .AQHDAV
couwomcEE 9»... can. on. o. maﬁooom 39.8w 80>» 338$ .02. was. mo 0%. :08 Bow 3.00:8 $1.88 No .38.. Z .Tm «ENE

136

Table 5-2. Average water chemistry for different land use (grouped) sampling locations.

 

 

LUID Land Use Description pH Conductivity DO Alkalinity
-1 -1 -1
mS cm mgO L mgCaCO3 L

Mean 7.1 0.72 5.7 90.3

URB Urban Std. Dev. 1.2 0.91 3.3 65.6
n 22 22 17 ll

Michi an State Mean 7.9 0.69 6.7 121.0

MSU Univefsity Std. Dev. 0.4 0.53 2.6 118.0
11 44 8 7 15

Mean 8.4 1.34 4.3 137.6

PL Automobile Parking Lot Std. Dev. 1.5 1.67 3.3 197.5
n 10 10 8 9

. Mean 7.4 0.54 7.7 276.0

GC Recmat'ona] Go” Std. Dev. 0.3 0.14 1.9 55.4

Course

11 13 13 ll 11

Mean 7.0 1.70 3.3 337.2

AG Agricultural Std. Dev. 0.3 1.55 3.4 203.5
n 13 12 10 8

Mean 5.8 0.10 2.3 41.5

FOR Forested Std. Dev. 1.0 0.12 1.1 51.9
n 6 6 6 6

Mean 8.0 0.55 7.9 84.6

MIX Mixed Std. Dev. 0.5 0.19 2.6 136.2
n 19 19 ll 10

 

137

4:.
O
I

‘1’
1 J—N

 

 

 

 

 

1
——1
[1» -)l-

DOC Concentration (mg/L)
N
O
1

 

Figure 5-1. Box plots of the total concentration of DOC (mg L") in surface water runoff
from sub-watersheds with speciﬁc types of land cover: agricultural (AG), forested (FOR),
golf course (GC), mixed (MIX), Michigan State University (MSU), automobile parking
lot (PL) and urban (URB). Values plotted as asterisks are 1.5 times beyond the range
where the central 50% of the observations fall.

138

 

 

 

 

 

 

3,000—
(a)
,‘Ebzpoo— *
5
B
.8. [T1 [11
:3 $ $
0
2
01,000—
2 E
0'— c 4. o
<5 <1- x»
N go o x9 yea 8 3.3”
3.—
(b) l
_ *
+22- -
m _
5 >1:
3‘ - *
'13 T [:1
>4. .—
331— E 1 T
Oh c A. 6 e
o 8- s»

Figure 5-2. The average molecular weight (a) and polydispersity (b) of DOC in surface
water runoff from sub-watersheds with speciﬁc types of land cover: agricultural (AG),
forested (FOR), golf course (GC), mixed (MIX), Michigan State University (MSU),
automobile parking lot (PL) and urban (URB). Values plotted as asterisks are 1.5 times
beyond the range where the central 50% of the observations fall.

139

NUVA
N
1
. ~10
{1:1

 

r9 aoq' <50 ‘9‘} $9 8" 313’

Figure 5-3. Box plots of the NUVA (L mgC m '1) in surface water runoff from sub-
watersheds with speciﬁc types of land cover: agricultural (AG), forested (FOR), golf
course (GC), mixed (MIX), Michigan State University (MSU), automobile parking lot
(PL) and urban (U RB). Values plotted as asterisks are 1.5 times beyond the range where
the central 50% of the observations fall.

“140

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5-4. The percentage of DOC retained on extended hydrophobic (El), hydrophobic
(I) and H-bonding (8) cartridges for samples from sub-watersheds with speciﬁc types of
land cover: agricultural (AG), forested (FOR), golf course (GC), mixed (MIX), Michigan
State University (MSU), automobile parking lot (PL) and urban (URB). The standard
error is plotted as error bars from the mean.

141

 

4.0

 

 

 

 

3.5 —
3.0 ~
2.5; e e. _ .00,
<2 1 o" .
E200 , ,, ., Bog: ,,,,,,
D O .
1 6’ i g
1.5 1 (%0 60%0
. ‘ I .
1.0 ‘ O y .
05. _ _ _. V ______ _ ......
0.0 . ,
0 500 1,000 1,500 2,000 2,500
Molecular Weight

Figure 5-5. NUVA (L mgC m -1) versus molecular weight (Da) for samples collected
from sub-watersheds with speciﬁc and mixed land uses: agricultural (O), forested (A),
golf course (<>), mixed (0), Michigan State University (0), automobile parking lot (:1) and
urban (I).

142

 

100

Hydrophobic DOC (%)

 

 

 

 

0.0 1.0 2.0 3.0 4.0
NUVA

Figure 5-6. The amount of DOC retained on the hydrophobic cartridge versus NUVA (L

mgC m '1) for DOC collected from sub-watersheds with speciﬁc and mixed land uses:
agricultural (O), forested (A), golf course (<>), mixed (0), Michigan State University (0),
automobile parking lot (E1) and urban (I).

143

References

Amon, R. M. W. and R. Benner (1996). "Bacterial utilization of different size classes of
dissolved organic matter." Limnology and Oceanography 41(1): 41-51.

Chin, Y. P., G. Aiken, et al. (1994). "Molecular-Weight, Polydispersity, and
Spectroscopic Properties of Aquatic Humic Substances." Environmental Science
& Technology 28(11): 1853-1858.

Chin, Y. P., G. R. Aiken, et al. (1997). "Binding of pyrene to aquatic and commercial

humic substances: The role of molecular weight and aromaticity." Environmental

 

Science & Technology 31(6): 1630-1635.

Chin, Y. P., S. J. Traina, et al. (1998). "Abundance and properties of dissolved organic
matter in pore waters of a freshwater wetland." Limnology and Oceanography
43(6): 1287-1296.

Cole, J. J ., W. H. McDowell, et al. (1984). "Sources and Molecular-Weight of Dissolved
Organic-Carbon in an Oligotrophic Lake." m 42(1): 1-9.

Croue, J. P., M. F. Benedetti, et al. (2003). "Characterization and copper binding of
humic and nonhumic organic matter isolated from the South Platte River:
Evidence for the presence of nitrogenous binding site." Environmental Science &
Technology 37(2): 328-336.

Dalzell, B. J ., T. R. F illey, et al. (2007). "The role of hydrology in annual organic carbon
loads and terrestrial organic matter export from a midwestem agricultural

watershed." Geochimica Et Cosmochimica Acta 71(6): 1448-1462.

144

Eckard, R. S., P. J. Hemes, et al. (2007). "Landscape scale controls on the vascular plant
component of dissolved organic carbon across a freshwater delta." Geochimica Et
Cosmochimica Acta 71(24): 5968-5984.

ESRI (2003). ArcMAPTM 9.1. ArcEditor. Redlands, CA, ESRI.

Frost, P. C ., J. H. Larson, et a1. (2006). "Landscape predictors of stream dissolved organic
matter concentration and physicochemistry in a Lake Superior river watershed."
Aquatic Sciences 68(1): 40-51.

Ghabbour, E. and G. Davies, Eds. (2004). Humic Substances: Nature's most versatile
materials. New York, Taylor and Francis.

Guggenberger, G., W. Zech, et al. (1994). "Formation and Mobilization Pathways of
Dissolved Organic-Matter - Evidence from Chemical Structural Studies of
Organic-Matter Fractions in Acid Forest Floor Solutions." gganic Geochemistg
21(1): 51-66.

Hedges, J. I. (1980). Flux of organic carbon by rivers to the oceans : report of a
workshop. . Woods Hole, Massachusetts, Division of Biological Sciences,
National Research Council; work supported by US. Department of Energy,
Ofﬁce of Energy Research: 109.

Hope, D., M. F. Billett, et al. (1994). "A Review of the Export of Carbon in River Water -
Fluxes and Processes." Environmental Pollution 84(3): 301-324.

Imai, A., T. Fukushima, et al. (2001). "Fractionation and characterization of dissolved
organic matter in a shallow eutrophic lake, its inﬂowing rivers, and other organic

matter sources." Water Research 35(17): 4019-4028.

145

Jackson, T. A. (1975). "Humic matter in natural waters and sediments." Soil Science
119(1): 56-64.

J affe, R., A. I. Rushdi, et al. (2006). "Natural product biomarkers as indicators of sources
and transport of sedimentary organic matter in a subtropical river." Chemosphere
64(11): 1870-1884.

Kaiser, E., D. B. Arscott, et a1. (2004). "Sources and distribution of organic carbon and
nitrogen in the Tagliamento River, Italy." Aquatic Sciences 66(1): 103-116.

Kalbitz, K., W. Geyer, et al. (1999). "Spectroscopic properties of dissolved humic
substances - a reﬂection of land use history in a fen area." Biogeochemistry 47(2):
219-238.

Kalbitz, K., S. Solinger, et al. (2000). "Controls on the dynamics of dissolved organic
matter in soils: A review." Soil Science 165(4): 277-304.

Kawahigashi, M., K. Kaiser, et al. (2004). "Dissolved organic matter in small streams
along a gradient from discontinuous to continuous permaﬁost." Global Change
May 10(9): 1576-1586.

Leenheer, J. A. and C. Rostad (2004). Tannins and Terpenoids as Major Precursors of
Suwannee River Fulvic Acid. U. S. D. o. t. Interior and U. S. G. Survey, US.
Geological Survey.

Lennon, J. T. and L. E. Pfaff (2005). "Source and supply of terrestrial organic matter
affects aquatic microbial metabolism." Aquatic Microbial Ecology 39(2): 107-

119.

146

Linnik, P. N. (2003). "Complexation as the most important factor in the fate and transport
of heavy metals in the Dnieper water bodies." Analytical and Bioanalytical
Chemisml 376(3): 405-412.

Marschner, B. and K. Kalbitz (2003). "Controls of bioavailability and biodegradability of
dissolved organic matter in soils." Geoderma 113(3—4): 211-235.

Maurice, P. A., M. J. Pullin, et al. (2002). "A comparison of surface water natural organic
matter in raw ﬁltered water samples, XAD, and reverse osmosis isolates." Wale;
Research 36(9): 2357-2371.

McKnight, D. M., E. D. Andrews, et al. (1994). "Aquatic Fulvic-Acids in Algal-Rich
Antarctic Ponds." Limnology and Oceanogaphy 39(8): 1972-1979.

MDEQ (2005). Digital Elevation Model, Michigan Geographic Data Library, Michigan
Department of Environmental Quality.

MDNR (2001). IF MAP Souther Michigan Land Cover, Michigan Geographic Data
Library, Michigan Department of Natural Resources, Forest, Mineral and Fire
Managrnent Division.

MDNR (2001). Michigan Land Cover/Use Classiﬁcation System - 2000: DRAFT.
Lansing, MI, Michigan Department of Natural Resources: 56.

Meier, M., Y. P. Chin, et al. (2004). "Variations in the composition and adsorption
behavior of dissolved organic matter at a small, forested watershed."
Biogeochemistry 67(1): 39-56.

Molot, L. A. and P. J. Dillon (1997). "Colour - mass balances and colour - dissolved
organic carbon relationships in lakes and streams in central Ontario." Canadian

Journal of Fisﬁeries and Aquatic Sciences 54(12): 2789-2795.

147

Moore, T. R. and R. J. Jackson (1989). "Dynamics of Dissolved Organic-Carbon in
Forested and Disturbed Catchments, Westland, New-Zealand .2. Larry River."
Water Resources Research 25(6): 1331-1339.

O'Loughlin, E. J. and Y. P. Chin (2004). "Quantiﬁcation and characterization of
dissolved organic carbon and iron in sedimentary porewater from Green Bay, WI,
USA." Biogeochemistry 71(3): 371-386.

Park, J. H. and E. Matzner (2003). "Controls on the release of dissolved organic carbon
and nitrogen from a deciduous forest ﬂoor investigated by manipulations of
aboveground litter inputs and water ﬂux." Biogeochemistry 66(3): 265-286.

Qualls, R. G. and B. L. Haines (1992). "Biodegradability of Dissolved Organic-Matter in
Forest Throughfall, Soil Solution, and Stream Water." Soil Science Society of
America Journal 56(2): 578-586.

Richey, J. E., J. T. Brock, et al. (1980). "Organic-Carbon - Oxidation and Transport in the
Amazon River." m 207(4437): 1348-1351.

Rorabacher, D. B. (1991). "Statistical treatment for rejection of deviant values: critical
values of Dixon's "Q" parameter and related subrange ratios at the 95%
conﬁdence level." Anal. Chem. 63(2): 139-146.

Sachse, A., D. Babenzien, et al. (2001). "Characterization of dissolved organic carbon
(DOC) in a dystrophic lake and an adjacent fen." Biogeochemistg 54(3): 279-
296.

Sachse, A., R. Henrion, et al. (2005). "Classiﬁcation of dissolved organic carbon (DOC)
in river systems: Inﬂuence of catchment characteristics and autochthonous

processes." Organic Geochemistry 36(6): 923-935.

148

Santschi, P. H., J. J. Lenhart, et a1. (1997). "Heterogeneous processes affecting trace
contaminant distribution in estuaries: The role of natural organic matter." _Man'arg
Chemistﬂ 58(1-2): 99-125.

Supelco. (2005). "Discovery DPA-6S SPE Tubes." Retrieved May 5, 2005, from
http://www.sigmaaldrich.com.

Traina, S. J ., J. Novak, et al. (1990). "An Ultraviolet Absorbance Method of Estimating
the Percent Aromatic Carbon Content of Humic Acids." J Environ Oual 19(1):

1 5 1 - l 53.

Tranvik, L. J. (1993). "Microbial transformation of labile dissolved organic matter into
humic-like matter in seawater." FEMS Microbiology Ecology 12(3): 177-183.

Wickland, K. P., J. C. Neff, et a1. (2007). "Dissolved organic carbon in Alaskan boreal
forest: Sources, chemical characteristics, and biodegradability." Ecosystems
10(8): 1323-1340.

Zhou, Q. H., S. E. Cabaniss, et al. (2000). "Considerations in the use of hi gh-pressure
size exclusion chromatography (HPSEC) for determining molecular weights of
aquatic humic substances." Water Research 34(14): 3505-3514.

Zumstein, J. and J. Bufﬂe (1989). "Circulation of pedogenic and aquagenic organic

matter in a eutrophic lake." Water Research 23: 229-239.

149

CHAPTER 6
INFLUENCE OF ENVIRONMENTAL FACTORS ON

DISSOLVED ORGANIC CARBON CHARACTERISTICS

Abstract

Signiﬁcant temporal variability has been observed in DOC characteristics, even
when from a constant source. Seasonal variations are known to alter the type of DOC
found in many watersheds; however, speciﬁc factors responsible for these differences in
DOC characteristics remain unclear. Environmental factors (e.g. solar radiation, water
temperature) that may inﬂuence the production and transformation of organic carbon
from terrestrial sources are investigated to determine if they are related to DOC
characteristics. Samples were collected in a manor to minimize the inﬂuence of aquatic
processes, such as microbial and algal growth, on DOC characteristics, however, results
suggest photochemical and biological alteration of DOC may have occurred. General
linear models (GLMS) incorporating precipitation, solar radiation and some aqueous
chemical parameters were found to successfully explain variations observed in DOC
aromaticity (measured as normalized UV absorbance at 280nm), molecular weight and
hydrophobicity. GLMs incorporating multiple environmental factors and land use
accounted for up to 86% of the variability observed in DOC characteristics. Interactions
between land use - solar radiation, water temperature and water conductivity were found
to be statistically signiﬁcant (p-values < 0.05). Of all DOC characteristics investigated,
hydrophobicity was found to be the most difﬁcult to correlate with a speciﬁc

environmental parameter; however, the hydrophobicity of DOC from sub-watersheds

150

with forested land use did appear to oscillate during the year. This work demonstrates
strong correlations between land use, environmental factors and DOC characteristics
which can be explored to further elucidate cause and effect relationships responsible for

observed variations in DOC characteristics.

151

Introduction

Dissolved organic carbon (DOC) is composed of a heterogeneous mixture of
biological organic compounds at various stages of microbial and physico—chemical
transformation (Ghabbour and Davies 2004; Wickland, Neff et al. 2007). In surface water
systems, DOC represents a critical ecological component by inﬂuencing chemical and
biological reactions (Qualls and Haines 1992; Chin, Aiken et al. 1997 ; Santschi, Lenhart
et al. 1997; Lennon and Pfaff 2005). Due to the important role of DOC in food webs and
chemical cycles, there is a need to determine the factors responsible for DOC
characteristics and production (Kalbitz, Solinger et al. 2000).

Previous investigations by others into possible variations in DOC quality have
revealed inconclusive results, while Chapter 5 identiﬁed correlations between land use
and DOC characteristics. Little variation in DOC characteristics has been observed
between watersheds with similar types of land use (Schumacher, Christl et al. 2006). The
likely cause of the ambiguity observed by others are sampling schemes which do not
collected DOC samples immediately from the terrestrial environment before aquatic
processes inﬂuence DOC characteristics. In large surface water systems where aquatic
processes are dominant and residence times for DOC are high - Oligotrophic lakes, large
rivers or oceans — DOC characteristics consist of relatively stable pools of aliphatic
molecules with low abilities to attenuate light (Tipping, Hilton et al. 1988; McKnight,
Andrews et a1. 1994; Sun, Perdue et al. 1997). However, the characteristics of DOC
found in stable surface water environments are different from the characteristics of DOC
entering them from terrestrial landscapes (Cole, McDowell et a1. 1984; Loh, Bauer et a1.

2006). DOC produced from different types of vegetation are also found to vary in

152

biodegradability (Wickland, Neff et al. 2007). Chapter 5 described statistically signiﬁcant
differences in aromaticity, molecular weight, polydispersity and the fraction of DOC
retained by hydrophobic and H-bonding mechanisms based on land use; however,
considerable variability still existed among some types of samples.

In order to address the variability observed between samples collected from sub-
watersheds with the same land use, environmental factors must also be considered. The
cause for variability observed in DOC characteristics from terrestrial sources remains
largely unknown (Canham, Pace et a1. 2004). Some DOC characteristics, such as the
aromatic content and photo-degradation of DOC are thought to be seasonally dependent
(Kortelainen 1993; Molot and Dillon 1997; Larson, Frost et al. 2007; Rodriguez-zuﬁiga,
Milori et al. 2008). For instance, the rate at which DOC from terrestrial sources is
chemically altered has been found to vary seasonally (Porcal, Hejzlar et al. 2004).
Additionally, clear seasonal trends have been observed in the structural composition (i.e.
aromatic content) of DOC (Clair and Sayer 1997). Environmental factors that are
seasonally dependent, precipitation and solar radiation, have been found to inﬂuence the
amount of light absorbed by DOC (Curtis and Schindler 1997; Molot and Dillon 1997;
Lindell, Graneli et al. 2000; Reche and Pace 2002). Due to these limited observations,
environmental factors responsible may be responsible for variations in DOC
characteristics and requires ﬁrrther investigation (Christian and Lind 2007).

In order to determine the processes responsible for DOC variation, it is necessary
to collect and evaluate DOC immediately downstream from its source. By carefully
isolating DOC produced from unique catchments, a range of DOC characteristic have

been shown to vary depending on the land cover present in watersheds (Imai, F ukushima

153

et al. 2001; Page, van Leeuwen et al. 2001). As a result, the farther upstream DOC
samples are collected, or closer to the terrestrial source of organic carbon, the smaller the
inﬂuence of aquatic processes, such as chemical transformations from UV light, and
microbial and algal growth (Vannote, Minshall et al. 1980).

DOC aromaticity, molecular weight (MW) and hydrophobicity offer effective
measures DOC characteristics since they have been suggested as indictors of
bioavailability and transformation processes (Cabaniss, Zhou et al. 2000; Marschner and
Kalbitz 2003; Wickland, Neff et al. 2007). The aromaticity of DOC can be approximated
by measuring the extent of ultraviolet (UV) light absorbed at a wavelength of 280nm
when normalized by the concentration of DOC (Chin, Aiken et al. 1994). Differences in
the UV absorption of DOC have been found to depend on watershed characteristics
(Larson, Frost et al. 2007 ). It has been proposed that the MW of DOC indicates the extent
to which it is susceptible to microbial degradation (V annote, Minshall et a1. 1980; Amon
and Benner 1996). Initially, it was suggested that microorganisms in the headwater of
riverine systems preferentially degrade smaller organic molecules (Vannote, Minshall et
al. 1980). As DOC moved downstream the low MW ﬁactions of DOC would be
preferentially utilized producing recalcitrant pools of high MW DOC in oceans, lakes and
other receiving waters. However, a new size-reactivity continuum model has challenged
this assumption and suggests that high MW DOC may be more reactive and a better
indicator of microbial diagenesis (Amon and Benner 1996; Fischer, Sachse et al. 2002).
Additionally, the hydrophobicity of DOC may also be an indicator of microbial utility.
Polysaccharides and other high energy constituents are rapidly consumed by

microorganisms (Fischer, Sachse et al. 2002). As these energy rich constituents, which

154

are more hydrophilic than hydrophobic, are consumed by microorganism DOC becomes
increasingly hydrophobic. Hence, hydrophobicity may serve as an indicator of the extent
that DOC has been processed by microorganisms.

The focus of this chapter is to investigate environmental variables associated with
variations in DOC characteristics. These factors are in addition to the inﬂuence of land
use on DOC characteristics described in Chapter 5. Based on seasonal ﬂuctuations in
DOC characteristics and the presences of plant biomarkers present in DOC, variations in
DOC characteristics from catchments composed of the same land uses are hypothesized
to be due to the growth of terrestrial plants. The amount of solar radiation was used as a
surrogate for vegetative growth. Possible correlations between multiple environmental
factors and the aromaticity, MW and hydrophobicity of DOC from catchments with one
of seven types of land use were used to evaluate this hypothesis. Other weather related
and water chemistry parameters were investigated with the use of General Linear Models
(GLMS) to determine if they were responsible for variations observed in DOC

characteristics from similar land uses (Chapter 5).

Methods and Materials

Samples were collected from 48 different locations within the Grand River
watershed in central Michigan. Of the 48 locations, 29 were from catchments with only
one type of land use present within the Red Cedar and Looking Glass River tributaries.
Based in part on the Michigan Land Cover/Use Classiﬁcation System (MDNR 2001),
samples were classiﬁed into one of seven land use types: urban (URB), agricultural (AG),

forested (FOR), Michigan State University’s campus (MSU), automobile parking lot

155

(PL), recreational golf course (GC) or mixed (MIX). Samples classiﬁed as MIX were
collected from the Grand and Red Cedar Rivers at points where it would be impossible to
differentiate the inﬂuence of the different land uses responsible for DOC characteristics.
General chemistry and a more detailed description of sampling locations can be found in
the previous chapter (Chapter 5) and the appendix.

The concentration of DOC was determined by automated analysis based on the
Heated-Persulfate Oxidation Method (Clesceri et al., 1998) using an 01 Analytical Model
1010 Wet Oxidation Total Organic Carbon Analyzer after passing though a 0.45 pm glass
ﬁber ﬁlter that was double acid washed. Prior to conducting experiments, all glassware
used during DOC analysis was acid (18% HCl) washed, rinsed with DDI water and
placed in a 550°F oven for more than 2 hours to ensure cleanliness. DOC characteristics
were measured by assessing the amount of aromaticity, the molecular weight,
polydispersity and DOC hydrophobicity.

The aromaticity of DOC was approximated by the normalized ultraviolet
absorbance at a wavelength of 280 nm (NUV A) (Chin, Aiken et al. 1994). The NUVA

was determined using a Shimadzu UV-160 spectrophotometer and normalizing by the

concentration of DOC, resulting in units of L gC.1 cm-1. A wavelength of 280 nm is a

more effective surrogate for measuring the aromatic content of DOC than 254 nm for two
reasons: (1) the transfer of electrons between overlapping n—orbitals occurs at this
wavelength for phenolic and other humic like organ substances (Traina, Novak et a1.
1990), and (2) nitrate, which also absorbs UV light and is ubiquitous in natural waters,

does not absorb UV light at 280 nm (Chin, Aiken et al. 1994).

156

The molecular weight (MW) and polydispersity of DOC was determined by size

exclusion chromatography (HPSEC). The number-averaged MW (Mn) and weight-

averaged MW (MW) were calculated by the following equations:

N N
Mn =Zhi/Zhi(Mi) (Equation 6-1)
i=1 i=1
and
N N
MW = Emma/Z11, (Equation 6-2)
i=1 i=1

where hi is the height and M; is the molecular mass of the sample HPSEC eluted at

volume i. The MW is commonly referenced as the average MW and this custom will be

maintained throughout the remainder of this paper. Polydisperity (p) is the ratio of the

weight-averaged MW and the number-averaged MW:

(Equation 6—3)

 

Low polydispersity indicates a DOC with a relatively narrow range of molecular weights.
The HPSEC system employed utilized a Gilson Model 303 pump (Middleton, WI), a
Waters Protein-Pak 125 modiﬁed silica column (Milford, MA) and UV detection at 254
nm on a Dionex Variable Wavelength Detector (Sunnyvale, CA) (Chin, Aiken et a1.

1994; Zhou, Cabaniss et a1. 2000). The mobile phase consisted of 0.1 m NaCl, 0.002 m

KHZPO4 and 0.002 m NazHPO4 solutions buffered to an approximate pH of 7 and

calibration was performed using random coil sodium polystyrene sulfonates

157

(Polysciences, Inc.) (1.8, 5.4, 8 and 18 kDa) and acetone (58 Da) (Zhou, Cabaniss et al.
2000)

DOC hydrophobicity was determined by the amount of DOC retained on a
hydrophobic solid phase extraction (SPE) media. Brieﬂy, surface water samples were
ﬁltered through a 0.45 pm ﬁlter, samples were passed through a styrene-divinyl benzene
SPE cartridge (Biotage, lsolute 101). The concentration of DOC was measured before
and after to obtain the fraction, or percentage, of DOC retained. The hydrophobicity of
DOC is related to its structure (Chapter 4) and details of the method used to isolate these
fractions can be found in an earlier chapter (Chapter 3).

Questionable data was not included in statistical analysis if it was deemed an
outlier by Dixon’s Q-test at the 95% conﬁdence level (Rorabacher 1991). SYSTAT
(version 12.02.00; San Jose, CA) was used for all statistical analysis. Unless otherwise
noted, and a—level of 0.05 was used to determine signiﬁcance. Due to the inﬂuence of
land use in determining the characteristics of DOC (Chapter 5), GLMS were used to
evaluate the inﬂuence of weather on DOC characteristics. GLMs are statistical models
used to investigate many environmental problems (El-Shaarawi and Piegorsch 2001), and
have been used successfully to identify environmental factors responsible for seasonal
variations in the production of organic acids by plants (Burkey, Neufeld et al. 2006).
GLMS utilize multiple regression coefﬁcients, each for different factors, which allows for
the inﬂuence of numerous factors to be assessed. The GLM for a single independent

variable can be described as:

Y = X1131 + X2152. .. Xan + e (Equation 6-4)

158

where Y is a vector of the dependent variable, X are vectors of the independent variables,
8 are vectors of regression coefﬁcients, and e is a vector of random errors. For each
combination of independent variables there is a linear regression coefﬁcient. A simpliﬁed
version of Equation 6-4 can be written excluding the regression coefﬁcient showing

simply the dependent and independent variables, for one independent variable:

A = B + C (Equation 6-5)

where A represents the dependent variable, B represents the independent variable and C
is a constant representing the random error. All models were evaluated to determine the
data was normally distributed, the errors had constant variance and the linear model
generally described all samples evenly. Once an appropriate GLM was established,
Tukey’s Honestly Signiﬁcant Difference (HSD) Test was used to evaluate differences
between factors responsible for DOC characteristics (Tukey 1949).

Two methods were used to determine the effectiveness of statistical models in
describing variations in DOC characteristics. The ﬁrst method used was to evaluate the r-
squared value of the resulting linear regression. The value of r-squared effectively
determines the accuracy of model predictions. GLMS unable to explain two-thirds of
observed variation were considered as highly suspect at best, and most likely to be an
indication that other factors were responsible for the observed variation. Because it is
possible to add and inﬁnite number of model parameters to effectively describe any

variation, regardless of the true meaning behind model parameters, another method was

159

required to evaluate the amount of bias and uncertainty present in suspected models. The
second method used to evaluate the effectiveness of statistical models was the Akaike
Information Criterion (AIC) (Akaike 1974). The AIC is based on the principle of
parsimony; the ideal model is one with the optimal combination of bias and variability.

Effectively, the AIC adds a penalty for the increased likelihood of bias and uncertainty. A

corrected AIC (AICC) was used in this study due to the limited sample size for some

comparisons (Hurvich and Tsai 1989).

Results and Discussion

For catchments dominated by vegetation (e.g. agricultural, golf course) a possible
negative correlation between NUVA and the mean weekly solar ﬂux density was
observed (Figure 6-1). Some of the sites dominated by vegetation that did not show a
decrease in NUVA with increased solar radiation were forested catchments. Using
NUVA as a surrogate for the aromaticity (Chin, Aiken et al. 1994), the observed decrease
in aromaticity with increased solar radiation suggests the breakdown of aromatic
moieties. Solar induced breakdown of DOC chromophores, sometimes call photo-
bleaching, has been reported for DOC in wetlands (Osbum, Morris et al. 2001; Waiser
and Robarts 2004) and marine waters (Del Vecchio and Blough 2004). A lack of photo-
degradation of DOC from forested catchments is likely due to the shielding from the
forest canopy (Frost, Larson et al. 2005).

After evaluating multiple other environmental parameters (e. g. pH, precipitation)
for correlations indicating factors which inﬂuence NUVA, the following simpliﬁed GLM

was found to be most effective in explaining NUVA variability (Figure 6-2a):

160

NUVA = LU + (LU x TEMPW) + (LU x SOLARW) + C (Equation 6-6)

where NUVA is the normalized UV absorbance at 280mm, LU is land use, TEMPW is the

water temperature (°C), SOLARw is the mean maximum solar ﬂux density (W m—z) for

the week prior to sampling, and C is a constant representing the random errors. As
described previously (Equation 6-4), for each combination of independent variables there
is a regression coefﬁcient determined by SYSTAT, these coefﬁcients are described in

Table 6-1. This GLM (Equation 6-6) accounts for 67% of the variation in NUVA

observed with an AICc of 93.8. While other parameters are likely to have major roles in

determining NUVA, evidenced by the poor model ﬁt, all variables were found to be
statistically signiﬁcant (p-values < 0.001). Based on Tukey’s HSD Test, statistically
signiﬁcant differences in NUVA were found for contrasts between four types of land
(Table 6-2). These differences were between catchments containing storm sewer
infrastructure (MSU, PL and URB) and catchments where the possibility of aquatic
processes were likely (FOR and GC).

While a large portion (33%) of the variation in NUVA remains unexplained, the
inﬂuence of water temperature and solar radiation was still found to be signﬁcant.
Microbial and photo-chemical processes may be responsible for changes in NUVA.
Heterotrophic microorganisms preferentially consume carbohydrates, proteins and other
nutrient rich sources leaving behind aromatic structures (Marschner and Kalbitz 2003).
As water temperature increases, microbial grth can be expected to increase. Therefore,
as water temperature increased, the inﬂuence of microbial processes would be expected

to increase aromaticity, relative to the concentration of DOC (i.e. NUVA). The other

161

process that could be responsible for changes in aromatic content is UV degradation of
aromatic moieties (Figure 6-1). Distinguishing between microbial and photo-chemical
processes within this set of samples is difﬁcult because water temperature is likely to rise
with mean weekly solar ﬂux. However, a decrease in the amount of photo-degradation in
forested sub-watersheds due to the forest canopy and increased microbial grth from
rising water temperatures may both contribute to a greater aromaticity for DOC ﬁ'om
forested areas.

The following simpliﬁed GLM accounted for 86% of the observed variation in

DOC MW (Figure 6-2b):

MW = LU + PRECIPW + (LU x SOLARm) + (LU x COND) + C (Equation 6-7)

where MW is the weight-averaged molecular weight (kDa), LU is land use, PRECIPW is
the total precipitation (mm) for the week prior to sampling, SOLAR,n is the mean
maximrun solar ﬂux density (W m-z) for the month prior to sampling, COND is the

. . -l . . .
conductrvrty (mS cm ) of the sample solution, and C rs a constant representing the

random errors. Again, regression coefﬁcients determined by SYSTAT are presented in
Table 6-1. In addition for this model accounting for nearly all the variability observed in
DOC MW, it was also found to have a relatively low AICc (-l7.2). With the exception of
the interaction between land use and solution conductivity (p-value = 0.08), all variable

were found to be statistically signiﬁcant (p-values < 0.009). Similar to the equation used

162

to describe the variability observed in MW, variations in polydispersity were explained

by the following equation (Figure 6-2c):
POLYD = LU + (LU x SOLARm) + (LU x COND) + C (Equation 6-8)

where POLYD is the polydispersity of DOC, LU is land use, PRECIPW is the total

precipitation (mm) for the month prior to sampling, SOLARm is the mean maximum solar

ﬂux density (W m'z) for the month prior to sampling, COND is the conductivity (mS cm-

1 . . . .
) of the sample solutron, and C rs a constant representing the random errors. Equation 6-

8 explains 79% of the variation observed in polydispersity with an AICc of 16.0 (Figure
6-2). Using the same model for polydispersity as the model used to explain MW resulted
in a similar r-squared value but a higher likelihood of bias and uncertainty (AICc = 45.8).
Variations observed in polydispersity were statistically signiﬁcant for all the factors
included in Equation 6-8: land use (p-value = 0.011), the interaction between land use
and the mean monthly solar ﬂux density (p-value = 0.002), and the interaction between
land use and conductivity (p-value < 0.001).

The inﬂuence of land use and solar ﬂux density on MW and polydispersity may
be attributed to plant grth by speciﬁc types of land use (i.e. golf course, forested)
(Lawlor 1995). The production of leaf soluble proteins has been shown to increase with
solar radiation (Liu, Xu et al. 2005). The production of proteins and other biological
compounds, such as terpenoids and ﬂavonoids (J affe, Rushdi et al. 2006), by terrestrial
plants can result in the production of high MW organic compounds that are likely

responsible for the increases in DOC MW. And unlike the GLM for NUVA, the GLMs

163

for MW and polydispersity were found to account for variability based on the monthly
mean solar ﬂux density rather than the weekly. This suggests the inﬂuence of solar
radiation on MW and polydispersity are more dependent on longer seasonal trends, such
as growing seasons.

There are multiple plausible explanations for the signiﬁcant inﬂuence of land use
dependent conductivity on DOC MW and polydispersity: (1) conductivity is simply a
seasonal indicator for urban landscapes where road salt is applied during the winter
months, (2) conductivity is indicative of where in the hydrograph samples were collected
(J ohnsen, Martinsen et al. 1987), (3) increased ionic strength, related to conductivity,
results in increased thermodynamic driving force for intermolecular assembly, and (4) the
ionic strength of surface water bodies inﬂuences microbial populations which produce or
transform DOC. Unfortunately, no clear trend between conductivity and MW is evident
(Figure 6-4), and we therefore cannot eliminate any of these possibilities.

Variations in hydrophobicity did not appear to be inﬂuenced by weather. Based
on the parameters investigated, the following GLM best described hydrophobicity

(Figure 6-2d):
%Ho = C + LU + ALK (Equation 6-9)

where %H0 is the percent DOC retained by hydrophobic mechanisms by SPE, LU is land

use, ALK is the sample alkalinity (mg ’L-1 as CaCO3), and C is a constant representing

the random errors. This model was only able to account for 64% of the variability in

hydrophobicity (Figure 6-2) and had a high AICc of 410.8, however, variations in land

164

use (p-value < 0.001) and alkalinity (p-value = 0.032) were found to be a signiﬁcant
source of variation.

Alkalinity may indicate the ﬂow path experienced by DOC during stormwater
runoff events. Rain water is low in alkalinity. Water which passes through soils and leaf
litter (e. g. interﬂow), where organic acids are likely to dissolve, are likely to have a low
alkalinity. Alternatively, ground water that has slowly seeped in to storm sewers and
surface water that has been trapped within storm sewers for extended periods of time will
contain a high amount of carbonates, silicates and other constituents which will increase
alkalinity. As these different water masses elute during the hydrograph, it is reasonable to
infer that the alkalinity will vary. Likewise, the production of DOC during stormwater
runoff events has been found to vary over the course of hydrographs based on catchment
hydrology (Schlesinger and Melack 1981; McGlynn and McDonnell 2003). DOC with
indirect ﬂow paths through leaf litter and soil may be stripped of hydrophobic DOC
(Kawahigashi, Kaiser et al. 2006). Furthermore, soils also demonstrate chromatographic
tendency to release predominantly hydrophilic DOC when ﬂushed (Kaiser and Zech
1998). As a result, alkalinity may correlate to DOC hydrophobicity because it serves as
an indicator of when in the hydrograph the samples were collected.

The hydrophobicity of DOC did exhibit some seasonal variability for forested
catchments (Figure 6-3). During the summer months, hydrophobic ﬁactions of DOC
were found to decrease. This is contrary to what would be expected when aromaticity and
MW increase, due to the grth of terrestrial plants. Additionally, as water temperatures
increase, so does microbial growth, resulting in the formation of more aromatic DOC,

which also increases hydrophobicity. In forested catchments the canopy prevents solar

165

radiation from reaching surface waters and causing photo-degradation. All of these
factors can be expected to result in DOC that is more hydrophobic during summer
months. It is suggested that the observed decrease in hydrophobicity may be due to the
preferential retention of hydrophobic fractions of DOC during transport to the sampling

location.

Conclusions

In addition to land use, variations in DOC aromaticity were found to correlate
with changes in solar radiation and water temperature. For sub-watersheds that contained
pooling water susceptible to high amounts of UV light exposure (i.e. golf course,
agricultural ﬁelds, and large rivers) a slight decrease in aromaticity suggests photo-
degradation did occur (Figure 6-1). For land uses with little direct solar radiation (i.e.
forested), increased microbial activity due to rising water temperatures may have been
more of a factor in determining the aromatic content of DOC, measured as NUVA (Table
6-2a). The interaction between solar radiation and land use was found to correlate with
the MW and polydispersity of DOC. GLMs using monthly mean solar ﬂux, conductivity
and land use were effective in explaining approximately 80% or more of the observed
variability (Figure 6-2). Environmental parameters were ineffective in reducing the
observed variability in DOC hydrophobicity (Figure 6-2d). The hydrophobicity of DOC
from forested sub-watersheds was found to be seasonally dependent (Figure 6-3).
Alkalinity was the most effective parameter found to reduce the variability observed in
DOC hydrophobicity. Results suggest hydrophobicity is related to the ﬂow path
experienced by DOC and watershed hydrologic characteristics. Overall, variations in

DOC characteristics from land uses with primarily vegetative cover were found to be

166

inﬂuenced by environmental factors that vary seasonally: water temperature, precipitation

and solar radiation.

167

Table 6-1. Regression coefﬁcients (13) for the GLMs describing DOC characteristics.

 

 

NUVA MW Polydi spersity Hydrophibicity

constant 2.185 1.198 1.445 33.20

B] 3.37E+00 5.76E-03 -2.43E-01 -1.49E+00

132 -5.35E-01 5.06E-01 -5.68E+00 -8.10E+00

B3 6.37E—01 1.83E+00 1.84E+00 -7.76E+00

B4 1.96E+00 -4.12E-02 1.21 E+00 -1.12E+01

135 -2.15E+00 -3 .28E+00 5.52E+00 1.67E+01

136 2.19E-01 7.78E-01 -3.62E+00 8.01 E+00

137 1.24E-01 7.16E-01 -3.34E-05 -2.14E—02

Bg -3.12E-02 -4.10E-04 7.44E-03

139 -7.44E-02 -1.25E-03 -2.71E-03

310 6.91 E-04 -3.19E-04 -1.46E-03

l3” 5.47E-02 3.28E-03 -5.31E-03

1312 -2.61 E-03 -8.62E-04 3.15E-03

1313 -6. 22E-03 -8 .45E-04 -2.5 8E-02

1314 2.07E—03 -4.1 1E-03 -2.24E+00

1315 1.01E-03 -9.80E—01 1.85E+00

016 -2. 09E-03 1.04E+00 2.15E—01

1317 8.25E-04 1.15E-01 -1.44E-01

1318 -4.38E-04 -6.28E-02 3.82E-01

1319 -6.26E-02

 

For NUVA the regression coefﬁcients represent the following variables: 01-6, land use; 07-12, land
use X water temperature (°C) ; and 1313.13, land use x mean weekly solar ﬂux density (W m'z).

For MW: the regression coefﬁcients represent the following variables: 131-6, land use; 137, total weekly

precipitation (mm), 133.; 3, land use x mean monthly solar ﬂux density (W m2); and 1314-19, land
use x conductivity (mS cm.1 ).

For polydisperity: the regression coefﬁcients represent the following variables: 01-6, land use; 137-12, land

use X mean monthly solar ﬂux density (W m2); and 1313-13, land use X conductivity (mS cm.1 ).

For hydrophobicity: the regression coefﬁcients represent the following variables: 0”,, land use; and B7,

alkalinity (mg L'1 as CaCO3).

168

Table 6-2. Differences determined by Tukey’s HSD Test (Tukey 1949) in NUVA, MW
and polydispersity resulting from the interaction between land use and parameters
describing weather and water chemistry.

 

 

DOC Characteristics
Land Use Contrasted Difference p-value
NUVA
FOR MSU 1.077 0.003
. FOR PL 0.982 0.023
FOR URB 1.196 0.001
GC URB 0.560 0.025
MW
GC MSU 0.659 0.001
GC PL 0.705 <0.001
GC URB 0.644 0.001
Polydispersity
AG GC -0.899 0.025
GC URB 0.753 0.034

 

169

 

 

 

 

1.0 . .
2.6 2.7 2.8 2.9 3.0 3.1
Log Weekly Mean Solar Flux Density

 

Figure 6-1. The inﬂuence of solar radiation, measured as loglo Week] mean solar ﬂux

density (W m-z), on the aromaticity, measured as NUVA (L mgC m - ) for samples from

catchments with the agricultural (El), forested (I), recreational golf course (0), and mixed
(A) land use.

170

NUVA MW

   

2.5 1
2.0
«a w: a
E at) 1.5 7'
a a 1‘
'8 8 1.0 T
0.5 .
0.0 111+, 1?: 1 .7.L_.l.;.11 1.._;.1._r..lx .'
0.0 2.0 4.0 0.0 0.5 1.0 1.5 2.0 2.5
Predicted Predicted
PolydiSPersity Hydrophobicity

   

“U '0
O Q) ~
2 a .._
8 3 d
a .9 [

0 1

1 0 .jil; . 1+;;i+411.1.1-1_1_1_.4l,1_L_LL.'14_;4_1.11111_1
0.0 2.0 4.0 0 10 20 3O 40 50 60 70
Predicted Predicted

Figure 6-2. Ability of generalized linear models (predicted) used to describe observed
DOC characteristics: (a) NUVA, (b) weight-averaged molecular weight, (c)
polydispersity, and (d) hydrophobicity.

171

 

50

 

 

 

 

40

Z?

"" 30 .

L; D

g 8

E20 A
O

10
0 .
2 4 6 8 10 12
Month

Figure 6-3. Seasonal inﬂuence on the hydrophobicity of DOC (%) for samples from
catchments with the agricultural (D), forested (I), recreational golf course (0), and mixed

(A) land use.

172

 

 

 

 

 

2.5

2 ...... ..D,.
Eb :
:a To U
'3 :
8 .
'6'
2 o a o

0.5

0 . . 1 .1 1 L

0 1 2 3 4 5
Conductivity

Figure 6-4. The inﬂuence of conductivity on the MW of DOC for sub-watersheds with
agricultural (El), forested (I), urban (9), MSU (O), recreational golf course (0), and
mixed (A) land use.

173

References

Akaike, H. (1974). "New Look at Statistical-Model Identiﬁcation." Ieee Transactions on
Automatic Control AC19(6): 716-723.

Amon, R. M. W. and R. Benner (1996). "Bacterial utilization of different size classes of
dissolved organic matter." Limnologv and Oceanography 41(1): 41-51.

Burkey, K. O., H. S. Neufeld, et al. (2006). "Seasonal proﬁles of leaf ascorbic acid
content and redox state in ozone-sensitive wildﬂowers." Environmental Pollution
143(3): 427-434.

Cabaniss, S. E., Q. Zhou, et al. (2000). "A Log-Normal Distribution Model for the
Molecular Weight of Aquatic Fulvic Acids." Environmental Science &
Technology 34(6): 1103-1109.

Canham, C. D., M. L. Pace, et al. (2004). "A spatially explicit watershed-scale analysis of
dissolved organic carbon in Adirondack lakes." Ecological Applicatitﬂ 14(3):
839-854.

Chin, Y. P., G. Aiken, et a1. (1994). "Molecular-Weight, Polydispersity, and
Spectroscopic Properties of Aquatic Humic Substances." Environmeﬁal Science
& Technology 28(11): 1853-1858.

Chin, Y. P., G. R. Aiken, et al. (1997). "Binding of pyrene to aquatic and commercial
humic substances: The role of molecular weight and aromaticity." Environmental
Science & Technology 31(6): 1630-1635.

Christian, B. W. and O. T. Lind (2007). "Multiple carbon substrate utilization by bacteria
at the sediment-water interface: seasonal patterns in a stratiﬁed eutrophic

reservoir." Iivdrobiologia 586: 43-56.

174

Clair, T. A. and B. G. Sayer (1997). "Environmental Variability in the Reactivity of
Freshwater Dissolved Organic Carbon to UV-B." Biogeochemistry 36(1): 89-97.

Cole, J. J ., W. H. McDowell, et al. (1984). "Sources and Molecular-Weight of Dissolved
Organic-Carbon in an Oligotrophic Lake." Q_il£0_s 42(1): 1-9.

Curtis, P. J. and D. W. Schindler (1997). "Hydrologic control of dissolved organic matter
in low-order Precambrian Shield Lakes." Biogeochemistg 36(1): 125-138.

Del Vecchio, R. and N. V. Blough (2004). "Spatial and seasonal distribution of
chromophoric dissolved organic matter and dissolved organic carbon in the
Middle Atlantic Bight." Marine Chemistry 89(1-4): 169-187.

El-Shaarawi, A. H. and W. W. Piegorsch, Eds. (2001). Encyclopedia of Environmetrics.
New York, Wiley and Sons.

Fischer, H., A. Sachse, et al. (2002). "Differential retention and utilization of dissolved
organic carbon by bacteria in river sediments." Limnology and Oceanography
47(6): 1702—1711.

Frost, P. C., J. H. Larson, et al. (2005). "Attenuation of ultraviolet radiation in streams of
northern Michigan." Journal of the North American Bentholggical Society 24(2):
246-255.

Ghabbour, E. and G. Davies, Eds. (2004). Humic Substances: Nature's most versatile
materials. New York, Taylor and Francis.

Hurvich, C. M. and C. L. Tsai (1989). "Regression and Time-Series Model Selection in

Small Samples." Biometrika 76(2): 297-307.

175

Imai, A., T. F ukushima, et al. (2001). "Fractionation and characterization of dissolved
organic matter in a shallow eutrophic lake, its inﬂowing rivers, and other organic
matter sources." Water Research 35(17): 4019-4028.

J affe, R., A. I. Rushdi, et a1. (2006). "Natural product biomarkers as indicators of sources
and transport of sedimentary organic matter in a subtropical river." Chemosphere
64(11): 1870-1884.

J ohnsen, S., K. Martinsen, et al. (1987). "Seasonal variation in composition and
properties of aquatic humic substances." Science of the Total Environment 62:
13-25.

Kaiser, K. and W. Zech (1998). "Rates of dissolved organic matter release and sorption in
forest soils." Soil Science 163(9): 714-725.

Kalbitz, K., S. Solinger, et a1. (2000). "Controls on the dynamics of dissolved organic
matter in soils: A review." Soil Science 165(4): 277-304.

Kawahigashi, M., K. Kaiser, et al. (2006). "Sorption of dissolved organic matter by
mineral soils of the Siberian forest tundra." Global Change Biology 12(10): 1868-
1877.

Kortelainen, P. (1993). "Content of Total Organic-Carbon in Finnish Lakes and Its
Relationship to Catchment Characteristics." gm Journal of Fi_s_heries an_d
Aquatic Sciences 50(7): 1477-1483.

Larson, J. H., P. C. Frost, et al. (2007). "Effects of upstream lakes on dissolved organic
matter in streams." Limnology and Oceanography 52(1): 60-69.

Lawlor, D. W. (1995). "Photosynthesis, Productivity and Environment." Journal of

Experimental Bota_ny 46: 1449-1461.

176

Lennon, J. T. and L. E. Pfaff (2005). "Source and supply of terrestrial organic matter
affects aquatic microbial metabolism." Aquatic Microbial Ecology 39(2): 107-
119.

Lindell, M. J ., H. Graneli, et al. (2000). "Seasonal photoreactivity of dissolved organic
matter from lakes with contrasting humic content." Canadian Journal of Fisheries
and Aquatic Sciences 57(5): 875-885.

Liu, L. X., S. M. Xu, et a1. (2005). "Solar UV-B radiation on growth, photosynthesis and
the xanthophyll cycle in tropical acacias and eucalyptus." Environmental and
Experimental Botzﬂy 54(2): 121-130.

Loh, A. N., J. E. Bauer, et al. (2006). "Dissolved and particulate organic matter source-
age characterization in the upper and lower Chesapeake Bay: A combined isotope
and biochemical approach." Limnology and Oceanography 51(3): 1421-1431.

Marschner, B. and K. Kalbitz (2003). "Controls of bioavailability and biodegradability of
dissolved organic matter in soils." Geoderma 113(3-4): 211-235.

McGlynn, B. L. and J. J. McDonnell (2003). "Role of discrete landscape units in

controlling catchment dissolved organic carbon dynamics." Water Resources

 

Research 39(4): SWC 3 1-18.

McKnight, D. M., E. D. Andrews, et a1. (1994). "Aquatic Fulvic-Acids in Algal-Rich
Antarctic Ponds." Limnology and Oceanogaphy 39(8): 1972-1979.

MDNR (2001). Michigan Land Cover/Use Classiﬁcation System - 2000: DRAFT.

Lansing, MI, Michigan Department of Natural Resources: 56.

177

Molot, L. A. and P. J. Dillon (1997). "Colour - mass balances and colour - dissolved
organic carbon relationships in lakes and streams in central Ontario." Canadian
Journal of Fisheries and Aquatic Sciences 54(12): 2789-2795.

Molot, L. A. and P. J. Dillon (1997). "Photolytic regulation of dissolved organic carbon
in northern lakes." Global Biogeochemical Cvcles 11(3): 357-365.

Osburn, C. L., D. P. Morris, et a1. (2001). "Chemical and Optical Changes in Freshwater
Dissolved Organic Matter Exposed to Solar Radiation." Biogeochemistry 54(3):
251-278.

Page, D. W., J. A. van Leeuwen, et al. (2001). "Tracing terrestrial compounds leaching
from two reservoir catchments as input to dissolved organic matter." Marine and
Freshwater Research 52(2): 223-233.

Porcal, P., J. Hejzlar, et al. (2004). "Seasonal and photochemical changes of DOM in an
acidiﬁed forest lake and its tributaries." Aﬁquﬂ Sciences 66(2): 211-222.
Qualls, R. G. and B. L. Haines (1992). "Biodegradability of Dissolved Organic-Matter in
Forest Throughfall, Soil Solution, and Stream Water." Soil Science Societv of

America Journal 56(2): 578-586.

Reche, I. and M. L. Pace (2002). "Linking dynamics of dissolved organic carbon in a
forested lake with environmental factors." Biogeochenﬂry 61(1): 21-36.

Rodriguez-zuﬁiga, U. F ., D. M. B. P. Milori, et al. (2008). "Changes in Optical Properties
Caused by UV-Irradiation of Aquatic Humic Substances ﬁom the Amazon River
Basin: Seasonal Variability Evaluation." Environmental Science & Technology

42(6): 1948-1953.

178

Rorabacher, D. B. (1991). "Statistical treatment for rejection of deviant values: critical
values of Dixon's "Q" parameter and related subrange ratios at the 95%
conﬁdence level." Anal. Chem. 63(2): 139-146.

Santschi, P. H., J. J. Lenhart, et al. (1997). "Heterogeneous processes affecting trace
contaminant distribution in estuaries: The role of natural organic matter." Marga
Chemisg 58(1-2): 99-125.

Schlesinger, W. H. and J. M. Melack (1981). "Transport of Organic-Carbon in the

Worlds Rivers." Tellus 33(2): 172-187.

 

Schumacher, M., I. Christl, et al. (2006). "Chemical composition of aquatic dissolved
organic matter in ﬁve boreal forest catchments sampled in spring and fall
seasons." Biggeochemistrgy 80(3): 263-275.

Sun, L., E. M. Perdue, et a1. (1997). "Use of elemental composition to predict
bioavailability of dissolved organic matter in a Georgia river." Limnology and
OceangLath 42(4): 714-721.

Tipping, E., J. Hilton, et al. (1988). "Dissolved Organic-Matter in Cumbrian Lakes and
Streams." Freshwater Biology 19(3): 371-378.

Traina, S. J ., J. Novak, et a1. (1990). "An Ultraviolet Absorbance Method of Estimating

the Percent Aromatic Carbon Content of Humic Acids." J Environ Oual 19(1):

 

1 51-153.

Tukey, J. W. (1949). "Comparing individual means in the analysis of variance."
Biometrics 5: 99-114.

Vannote, R. L., G. W. Minshall, et al. (1980). "River Continuum Concept." Canadian

Journal of Fisheries and Aggatic Sciences 37(1): 130-137.

179

Waiser, M. J. and R. D. Robarts (2004). "Photodegradation of DOC in a shallow prairie
wetland: evidence from seasonal changes in DOC optical properties and chemical

characteristics." Biggeochemistrv 69(2): 263-284.

 

Wickland, K. P., J. C. Neff, et al. (2007). "Dissolved organic carbon in Alaskan boreal
forest: Sources, chemical characteristics, and biodegradability." Ecosystems
10(8): 1323-1340.

Zhou, Q. H., S. E. Cabaniss, et al. (2000). "Considerations in the use of high-pressure
size exclusion chromatography (HPSEC) for determining molecular weights of

aquatic humic substances." Water Research 34(14): 3505-3514.

180

CHAPTER 7

CONCLUDING REMARKS

The work presented here demonstrates DOC characteristics vary depending on
land use and other environmental factors, and this variability likely inﬂuences Cu
complexation. In Chapter 3, variations in DOC characteristics were identiﬁed using a
new analytical technique that retains fractions of DOC based on speciﬁc bonding
interactions thought to be responsible for stabilizing humic substances. These variations
were found to be a function of the type of DOC evaluated. Additionally, Cu was found to
preferentially complex with DOC. The inﬂuence of DOC structure on Cu complexation
was explored ﬁ1rther in Chapter 4. Cu complexation was dependent upon both the
quantity and quality of DOC ligands present in solution. Aromatic structure and oxygen
ﬁmctional groups inﬂuenced Cu-DOC complexation. Overall, stronger complexation was
observed with aromatic moieties. Aromaticity played a signiﬁcant role in determining Cu
complexation with the high molecular weight fractions of DOC, while the oxygen content
inﬂuenced Cu complexation with low molecular weight fractions. DOC from terrestrial
and aquatic sources showed difference in the molecular structure of the low molecular
weight fraction of humic substances. The inﬂuence of land use on DOC characteristics
was demonstrated in Chapter 5. Catchments dominated by vegetation, such as
agricultural and forested land uses, were found to produce DOC higher in molecular
weight and aromatic structure than catchments with urban land uses. Results suggest the
preferential removal of hydrophobic fractions of DOC during transport from catchments

with indirect ﬂow paths. In Chapter 6, generalized linear models (GLMs) were used to

181

assess possible correlations between DOC characteristics and environmental factors in an
attempt to explain some of the variability observed in DOC characteristics. Results show
changes in solar radiation and water temperature, in addition to land use, correlate to
variations in DOC aromaticity. Solar radiation also correlated with DOC molecular
weight and polydispersity. By taking into account the inﬂuence of land use and
environmental factors (water temperature, precipitation and solar radiation), GLMs were
able to account for 80% or more of the variability observed in DOC characteristics.
Observed variations in DOC characteristics have a wide range of implications.
First, variations in DOC characteristics likely inﬂuence the microbial ecology present in
surface waters since DOC constitutes the base of aquatic food webs. As DOC
characteristics change, microbial communities will also likely change. While the dynamic
nature of these interactions are unknown, it is clear that not all DOC derived from
terrestrial sources is the same. This alone implies microbial communities, which
constitute the base of the aquatic food webs, likely differ in surface waters fed by
watersheds with dissimilar land uses. Second, variations in DOC characteristics also
inﬂuence the fate and transport of pollutants within surface water systems. As
demonstrated in Chapter 4, increases in DOC aromaticity will likely increase the amount
of Cu removed from solution. Cu is just one of the many trace-metals known to complex
with DOC. Furthermore, the solubility and reactivity of many organic constituents are
also inﬂuenced by DOC. With differences in characteristics observed for DOC derived
from catchments composed of various land uses, it can be expected that the transport of
pollutants within these systems will differ. Third, the connection between land use and

DOC characteristics offers another factor to consider when assessing the impact of land

182

use alterations. For example, based on the results discussed in chapter 5, if forested
landscapes are developed into residential properties there will likely be a change in DOC
characteristics which will impact surface water ecology and the transport of pollutants.
Given that human activity will continue to alter watershed surfaces, appropriate
precautions and treatment systems are required to ensure surface water quality. The
results of this work suggest stormwater treatment systems that utilize indirect ﬂow paths
and vegetative cover (i.e. rain gardens) will likely alter DOC characteristics and may
produce DOC similar to that produced from natural landscapes.

Future research is required to determine the impact variations in DOC
characteristics have on microbial ecology and the transport of trace metals, other than Cu,
and organic constituents. More extensive experiments are required to determine if the
correlations observed between molecular structure and Cu complexation hold for a wider
range of DOC, not just isolated humic substance. Additionally, the environmental factors
identiﬁed through the use of GLMs can serve as a guide for further research aimed to
describe the mechanisms responsible for DOC formation and transformation. For
example, the proposed removal of hydrophobic fractions during transport through
indirect ﬂow paths should be speciﬁcally tested. Research is required to design
stormwater treatment systems that not only remove pollutants of interest but also
produces DOC with characteristics similar to those found in natural environments.

Ultimately, a better understanding of the chemical and biological roles of DOC in
aquatic systems is required to accurately describe the dynamic processes within surface

water systems.

183

APPENDICES

184

APPENDIX A

METHODS

185

Alkalinity Analysis

The Gran titration method was used to determine sample alkalinity (Gran, 1950;

Gran, 1952). Details of this method can be found elsewhere (Stumm and Morgan, 1981).

Protocol for determining alkalinity

1.

Weigh 50mL disposable beaker (Fisher Scientiﬁc, part number 01-291-10).
Record weight (in grams) of empty beaker (W1).

Pour ~15mL into dry pre-weighted disposable beaker for alkalinity titration.
Re-weigh beaker to determine exact volume of sample titrated. Record weight
(in grams) of beaker with sample (W2). The volume (in milliliters) of sample
(V s) titrated is the difference between W2 and W1.

V5 = W2 — W1 (Equation A-l)
Record volume of sample titrated.

Add clean micro-stir bar (Fisher Scientiﬁc, part number 14-511-97) to beaker.
Place beaker on stir plate.

Insert calibrated pH probe so that probe tip is above stir bar.

Turn on stir plate brieﬂy (~5 sec) to ensure solution is completely mixed.
With stir plate turned off, measure and record solution pH and temperature
(°C).

Turn on stir plate and add 100uL of normalized 0.02N H2804 to sample. Aﬁer
solution is fully mixed (~5 sec) turn off stir plate and measure pH. Record

volume of acid added and resulting pH.

186

10. Continue to add normalized 0.02N H2SO4 to sample solution until pH drops

11.

below 3.5, using the stir plate as before to ensure solution is ﬁ111y mixed.
Record volume of acid added and resulting pH for each addition.

It should be noted that the volume of acid added depends on the solution. Four
or more data points between pH 4.5 and 8.3 and below 4.5 (total of 8 data
points) are necessary to obtain a good estimate of alkalinity. If the solution has
relatively high alkalinity (>250mg L'1 as C3CO3), the a few lmL additions of
acid may be appropriate. If the solution has relatively low alkalinity (<50mg
L'I as CaCO3), the total amount of acid added during titration may only be a
few milliliters. As the solution approaches the theoretical equivalence points
for carbonate dissociation (pH = 4.5 and 8.3) reduce the volume of acid added
to accurately capture large changes in pH.

Titration is complete when 4 data points (pH and volume of acid added) have
been recorded for solution with pH below 4.5. Titrated solution can be

disposed of and beaker and stir bar can be washed for reuse.

12. Calculate the total volume of acid (mL) added for each pH measurement.

13. Enter the data into online alkalinity calculator (http://or.water.usgs.gov/alk/):

a. Enter sample temperature (°C) — temperature at beginning of titration

b. Enter speciﬁc conductance (uS cm'l) - use conductivity from ﬁeld
sampling multi-probe

c. Select other from dropdown menu for acid concentration and specify
other acid concentration as 0.02N.

(1. Enter sample volume (mL) - Vs

187

e. Select yes for ﬁltered sample.

f. Select buret titration for titration type.

g. Past delimited data in titration data box and determine the order of
data points (pH in ﬁrst column, total volume of acid used for during
titration in second column or total volume of acid used for during
titration in ﬁrst column, pH in second column).

h. Select Gran function plot method for analysis method

i. Select advanced speciation method for speciation method

j. Click Calculate!

Table A-1. Example data for Gran Titration.

Vol. Acid Added (uL) Total Acid (mL) pH
0 0.000 7.76
500 0.500 7.43
1000 1.500 6.84
1000 2.500 6.58
1000 3.500 6.37
1000 4.500 6.15
1000 5.500 5.78
1000 6.500 5.09
500 7.000 4.01
100 7.100 3.82
100 7.200 3.69
100 7.300 3.59
100 7.400 3.51
100 7.500 3.44

14. Inspect Gran plot to make sure there were no problems with titration or data
entry.
15. From the Alkalinity Calculator output, record the Gran F1 bicarbonate

alkalinity (mg L'1 as CaCO3).

188

References

Gran, G., 1950. Determination of the equivalence point in potentiometric titrations. Acta
Chemica Scandinavica, 4: 559-577.

Gran, G., 1952. Determination of the equivalence point in potentiometric titrations - Part
II. The Analyst, 77: 661-671.

Stumm, W. and Morgan, M., 1981. Aquatic Chemistry. John Wiley & Sons, New York,

780 pp.

189

Cartridge Preparation

Preparation of anion and cation extraction cartridges

1. Place Extract-CleanTM 20pm polyethylene ﬁit (Alltech, part number 211408) into
bottom of 8.0ml Ultra-CleanTM treated polypropylene reservoir (Alltech, part
number 70218) using large end (end pipette bulb would be placed) of clean class
pipette.

2. Sequentially label cartridge with preﬁx to identify cation exchange (C-), anion
exchange without a molecular weight cutoff (A-) and anion exchange with a 1
kDa molecular weight cutoff (X-) (e. g. A-005 means ﬁfth anion exchange
cartridge prepared).

3. Cap cartridges (Alltech, part number 220600 and 220710).

4. Weigh out appropriate mass of resin material for each cartridge into a clean
100mL weigh boat.

a. Anion exchange: 1.0000 — 1.0100g of AG-MPl or AG-l X8 resin.

b. Cation exchange: 2.0000 — 2.0100g of Chelex 100 resin

5. Using approximately 4m1 of ultra-clean water (>18MQ), rinse resin from weigh
boat into appropriate cartridge.

a. If more water is necessary to rinse all of the resin from weigh boat,
carefully draw down water from cartridge using vacuum suction ﬁrst, then
rinse remaining resin into cartridge.

b. There should be approximately 2cm between top of tube and water in

cartridge when done.

190

6. Cap both ends of cartridge and swirl tube to remove large air bubbles.

7. Open top cap and rinse remaining resin from sidewalls into slurry at bottom.

a. If more water is necessary to rinse resin down, carefully draw down water
from cartridge using vacuum suction ﬁrst, then rinse remaining resin to
bottom.

b. The cartridge should be approximately half-full.

8. Place 10 new cartridges (capped) into an acid washed 400mL glass beaker and
cover beaker with Paraﬁlm.

9. Fill bottom of glass beaker with ultra-clean water to prevent beaker from ﬂoating
in sonicator bath (water level should be just above slurry depth in tubes).

10. Place beaker with new cartridges in sonicator bath for 20min to remove small
bubbles in bottom of cartridge.

11. After sonicating, remove caps from top of cartridge and place Extract-Clean
20pm polyethylene frit on top of resin. Compress resin slightly by the using large
end (end pipette bulb would be placed) of a clean class pipette with ﬁit.

12. Recap cartridge to prevent drying; always keep water above resin.

13. Store upright in vial rack with vial rack sealed in side plastic bag.

Counter-ion conversion for ion-exchange cartridge
Converting the counter-ion on ion-exchange cartridges requires
polytetraﬂuoroethylene (PTFE) tubing (Supelco, part number 57276) and solid phase

extraction (SPE) tube adapter (Supelco, part number 57020-U) be preassembled and

191

ultra-cleaned before use. To assemble, pull tubing through adapter so that ﬂange in

tubing creates a seal on the narrow end of the adapter, as shown below.

IllI
\ﬂ—J

tube adapter

PTFE tubing ﬂange

Figure A-l. Attachment of adapter to PTFE tubing.
Procedure to convert chloride to ﬂuoride counter-ion (for AG MP-l resin)

1. Prepare 1L of 1N NaOH by dissolving 17.01 g of NaOH in IL beaker and
ﬁlling to 1L with ultra-pure (>18MQ) water. Make sure NaOH completely
dissolves before using; this typically requires the use of a stir bar and stir
plate.

2. Prepare 1L of 1N NaF by dissolving 19.00g of NaF in beaker and ﬁll to IL
with nano-pure water. Make sure NaF completely dissolves before using; this
typically requires the use of a stir bar and stir plate.

3. Record cartridge identiﬁcation (ID) in lab notebook that will be converted and
write “F” on SPE tubes with sharpie to denote ﬂuoride counter ion.

4. Place cartridges on vacuum manifold.

5. Remove tops and draw water within cartridge down so that meniscus touches
top frit.

6. Rinse ultra-clean preassembled PTFE tubing and SPE tube adapter with ultra-
pure water.

7. Fill tubing with ultra-pure water (to ensure siphon) and block end of tubing at

SPE adapter with ﬁnger. Place open end of tubing in IN NaOH solution.

192

10.

11.

12.

13.

14.

15.

16.

Quickly, to avoid loosing siphon, attach tubing to top of SPE cartridge with
SPE tube adapter.

Repeat steps 6-7 until all cartridges have tubing with siphon to 1N NaOH
solution. It may be necessary to tape tubing together to prevent it from
popping out of beaker.

Turn on vacuum manifold and pass 40mL of 1N NaOH through cartridges.
Keep valve on glass block open so that ﬂow rate is approximately SmL min'l.
Discard rinse eluent collected in 40mL glass vials. Return vials to glass block.
Do not disconnect tubing to SPE cartridges.

Rinse cartridges with another 40mL of 1N NaOH (step 9). Do not allow
cartridges to run dry!

Discard rinse eluent collected in 40mL glass vials. Return vials to glass block.
Do not disconnect tubing to SPE cartridges.

Replace beaker containing 1N NaOH solution with beaker containing 1N NaF
solution. Remove open end of tubing from IN NaOH solution and place in 1N
NaF solution.

Rinse cartridges with 20mL of 1N NaF.

Aﬁer rinsing with NaF, remove tubing from solution and cartridges. Draw
down solution within cartridges until the bottom of meniscus touches upper
frit.

Cartridges are now converted. Cap and store up right in tray with bag taped

shut. Write “converted” and the date on bag.

193

Procedure to convert chloride to iodide counter-ion (for AG MP-l resin)

1.

10.

Prepare 1L of 1N NaI by dissolving 126.9g of Nal in beaker and ﬁll to 1L
with nano-pure water. Make sure Nal completely dissolves before using; this
typically requires the use of a stir bar and stir plate.

Record cartridge IDs in lab notebook that will be converted and write “I” on
SPE tubes with sharpie to denote iodide counter ion.

Place cartridges on vacuum manifold.

Remove tops and draw water within cartridge down so that meniscus touches
top ﬁit.

Rinse ultra-clean preassembled PTFE tubing and SPE tube adapter with ultra-
pure water.

Fill tubing with ultra-pure water (to ensure siphon) and block end of tubing at
SPE adapter with ﬁnger. Place open end of tubing in IN NaI solution.
Quickly, to avoid loosing siphon, attach tubing to top of SPE cartridge with
SPE tube adapter.

Repeat steps 5-6 until all cartridges have tubing with siphon to 1N Nal
solution. It may be necessary to tape tubing together to prevent it from
popping out of beaker.

Turn on vacuum manifold and pass 40mL of 1N Nal through cartridges. Keep
valve on glass block open - so that ﬂow rate is approximately SmL min".
Discard rinse eluent collected in 40mL glass vials. Return vials to glass block.
Cartridges are now converted. Cap and store up right in tray with bag taped

shut. Write “converted” and the date on bag.

194

Procedure to convert hydroxide to ﬂuoride counter-ion (for AG-l X8 resin)

1.

Prepare 1L of 1N NaF by dissolving 19.00g of NaF in beaker and ﬁll to IL
with nano-pure water. Make sure NaF completely dissolves before using; this
typically requires the use of a stir bar and stir plate.

Record cartridge IDs in lab notebook that will be converted and write “F” on
SPE tubes with sharpie to denote ﬂuoride counter ion.

Place cartridges on vacuum manifold.

Remove tops and draw water within cartridge down so that meniscus touches

top frit.

. Rinse ultra-clean preassembled PTFE tubing and SPE tube adapter with ultra-

pure water.

Fill tubing with ultra-pure water (to ensure siphon) and block end of tubing at
SPE adapter with ﬁnger. Place open end of tubing in IN NaF solution.
Quickly, to avoid loosing siphon, attach tubing to top of SPE cartridge with
SPE tube adapter.

Repeat steps 5-6 until all cartridges have tubing with siphon to 1N NaF
solution. It may be necessary to tape tubing together to prevent it from
popping out of beaker.

Turn on vacuum manifold and pass 20mL of 1N NaF through cartridges.
After rinsing with NaF, remove tubing from solution and cartridges. Draw

down solution within cartridges until the bottom of meniscus touches upper

frit.

195

10. Cartridges are now converted. Cap and store up right in tray with bag taped

shut. Write “converted” and the date on bag.

Counter-ion conversion notes:

0 Conversion procedures are for 2 batches of 10 cartridges (20 cartridges total)

0 It is possible to reuse tubing after conversions, only if conversions follow the
order OH, F, then I.

0 Sodium iodide is expensive and cannot be stored for more than 1 week. If

only 10 cartridges will be made, reduce amount of conversion solution to

500mL.

Procedure to condition anion exchange cartridges

1.

2.

Place cartridges on vacuum manifold.

Remove tops and draw water within cartridge down so that meniscus touches top
frit.

Rinse ultra-clean preassembled PTFE tubing and SPE tube adapter with ultra-pure
water.

Fill tubing with ultra-pure water (to ensure siphon) and block end of tubing at
SPE adapter with ﬁnger. Place open end of tubing in ultra-pure water solution.
Quickly, to avoid loosing siphon, attach tubing to top of SPE cartridge with SPE
tube adapter.

Repeat steps 3-4 until all cartridges have tubing with siphon to ultra-pure water
solution. It may be necessary to tape tubing together to prevent it from popping

out of beaker.

196

Draw down solution in cartridge so that bottom of meniscus touches the top ﬁit.
Do not allow cartridge to run dry!

Add 1mL of dilute (0.02%) hydroﬂuoric acid (HF).

Connect ultra-pure water tubing.

Draw through cartridge ~10mL of ultra-pure water into plastic falcon tubes

(V WR, product number 21008-931).

10. Discard HF solution in hazardous waste container.

11. Remove plastic falcon tubes and replace with 40mL silica TOC vials.

12. Rinse cartridges with 40mL of ultra-pure water twice (80mL total).

13. Draw water level down to approximately 1cm from top ﬁit.

14. Cartridge is now conditioned. Place red check mark on cartridge to denote

conditioned.

15. Cap and store up right in tray with bag taped shut. Write “conditioned” and the

date on bag.

Procedure to prepare and condition non-ionic cartridges

1.

2.

Mark cartridges with date and sequential number (ex: 020701, 020702, etc.)

Place cartridges on vacuum manifold.

. Add 1mL methanol.

Open manifold valve to allow methanol to soak entire media.
Once soaked, draw down methanol in cartridge so that bottom of meniscus
touches the top ﬂit. Do not allow cartridge to run dry!

Add 1mL of dilute (0.02%) HF.

197

10.

11.

12.

13.

14.

15.

16.

17.

Remove tops and draw water within cartridge down so that meniscus touches top
ﬁit.

Rinse ultra-clean preassembled PTFE tubing and SPE tube adapter with ultra-pure
water.

Fill tubing with ultra-pure water (to ensure siphon) and block end of tubing at
SPE adapter with ﬁnger. Place open end of tubing in ultra-pure water solution.
Quickly, to avoid loosing siphon, attach tubing to top of SPE cartridge with SPE
tube adapter.

Repeat steps 8-9 until all cartridges have tubing with siphon to ultra-pure water
solution. It may be necessary to tape tubing together to prevent it from popping
out of beaker.

Draw ~10mL of ultra-pure water through cartridge into plastic falcon tubes.
Discard methanol-HF solution in hazardous waste container.

Remove plastic falcon tubes and replace with 40mL silica TOC vials.

Rinse cartridges with 40mL of ultra-pure water twice (80mL total).

Draw water level down to approximately 1cm from top ﬁit.

Cartridge is now conditioned. Place red check mark on cartridge to denote
conditioned.

Cap and store up right in tray with bag taped shut. Write “conditioned” and the

date on bag.

Procedure to condition cation exchange cartridges

1.

Place cartridges on vacuum manifold.

198

. Remove tops and draw water within cartridge down so that meniscus touches top
ﬁit.

. Rinse ultra-clean preassembled PTFE tubing and SPE tube adapter with ultra-pure
water.

. Fill tubing with ultra-pure water (to ensure siphon) and block end of tubing at
SPE adapter with ﬁnger. Place open end of tubing in ultra-pure water solution.
Quickly, to avoid loosing siphon, attach tubing to top of SPE cartridge with SPE
tube adapter.

. Repeat steps 3-4 until all cartridges have tubing with siphon to ultra-pure water
solution. It may be necessary to tape tubing together to prevent it ﬁ'om popping
out of beaker.

. Rinse cartridges with 40mL of ultra-pure water twice (80mL total).

. Draw water level down to approximately 1cm from top ﬁit.

. Cartridge is now conditioned. Place red check mark on cartridge to denote
conditioned.

. Cap and store up right in tray with bag taped shut. Write “conditioned” and the

date on bag.

199

DOC-Trace Metal Fractionation

Dissolved organic carbon (DOC) and trace metal fractions were collected by
passing samples through solid-phase extraction (SPE) cartridges (in parallel) using trace
element clean 60 mL syringes arranged on a syringe pump (Harvard Apparatus) to
control the sample ﬂow rate. Syringes were connected to SPE tubes with the use of
cartridge adapters (Alltech). With the bottom of the SPE tube capped, syringes were
ﬁlled with 60 mL of sample from the top. Once ﬁlled, the SPE tube cap was removed, the
syringe plunger was inserted and ~10 mL of sample was forced through the SPE cartridge
and discarded (without the plunger contacting the aqueous sample). Flushing 10 mL of .
sample effectively replaced the ~2 mL of ultra-pure water remaining from cartridge
conditioning and allowed the SPE media to equilibrate with the sample solution. Eluent
from cartridges was collected in 40 mL borosilicate glass vials for DOC analysis. Aﬁer
the DOC samples were collected, SPE cartridges were temporarily removed from the
syringes to be reﬁlled. At all times the SPE media remained fully saturated with excess
sample. Syringes were then reloaded with sample solution to 40 mL. Again plungers
were inserted and placed back into the syringe pump. Eluent was collected in ultra-clean
30 mL polypropylene bottles for trace metals analysis. Used cartridges were then frozen
for the possible future extraction of retained compounds.

DOC-Trace Metal Fractionation Procedure
1. Place syringe pump (Harvard Apparatus) so that syringes and cartridges are

vertical and adequate space is available below to collect eluent (setup shown

below).

200

 

syringe pump

60mL syringes ——>
adapters —>

SPE cartridges

—>
collection vials or bottles —>|:| I] [I [I I] [l I]

Figure A-2. Arrangement of syringes and cartridges on for sample collection.

2.

3.

Set ﬂow rate to 1 mL min'l.

Place ultra-clean cartridge adapter on end of ultra-clean 60ml syringe.

. Remove plunger from syringe tube.

. Remove top cap from SPE cartridge and connect to syringe tube via cartridge

adapter. The Oasis HLB cartridges requires the use of Teﬂon tape (DuPont, part

number T-27730A) to ensure a good seal.

. Fill open syringe tube from top with 60 mL of sample.
. Remove bottom cap from cartridge.

. Insert syringe plunger and push 10 mL through SPE cartridge into waste beaker

(can be disposed of down drain). Be sure that rubber cap on plunger does not

contact sample!

. Place syringe-SPE assembly on syringe pump.

. Repeat steps 3-8 until each type of SPE cartridge required for fractionation are

loaded with sample solution.

201

11.

Record cartridge identiﬁcation numbers (IDs) and label clean 40 mL borosilicate
glass vials for DOC analysis with sample II). Add one of the following sufﬁxes to
end of base sample ID to indicate the fraction eluent collected (e.g. the eluent
collected from anion-F cartridge for a sample collected at location AG on August

8, 2007 would be labeled AG080807-F).

Table A-2. Cartridge ID key.

12.

13.

14.

15.

16.

17.

18.

Sufﬁx Cartridge

C cation
F anion-F
H H-bonding

Hi extended hydrophobic
Ho hydrophobic
I anion-I

X anion-lkDa
Arrange vials under appropriate SPE cartridges and turn on syringe pump.
After vials are ﬁlled (~40 min), turn off syringe pump, replace bottom cap on SPE
cartridge and cap DOC vials. Fraction samples should form an inverted meniscus
above top of the vial so that no air is trapped when vials are capped.
Immediately after collecting DOC samples, place vials in dark ﬁidge (4 °C).
While keeping syringe-SPE cartridge assemblies’ vertical remove from syringe
pump and place in vial tray.
Separate syringe from cartridge adapter (leaving adapter attached to SPE tube).
Dispense excess sample present in syringe (~10 mL) back into bulk solution of
ﬁltered sample (1 L glass beaker).

Remove plunger from syringe tube. Make sure plunger (with rubber cap) is not

contaminated by setting it down on unclean surface or touching anything.

202

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Reconnect syringe tube to SPE cartridge via adapter.

Fill open syringe tube from top with 40 mL of sample.

Remove bottom cap from cartridge.

Insert syringe plunger and push 10 mL through SPE cartridge into waste beaker
(can be disposed of down drain). Be sure that rubber cap on plunger does not
contact sample!

Place syringe-SPE assembly back on syringe pump.

Repeat steps 15-23 until each type of SPE cartridge required for fractionation are
re-loaded with sample solution.

Label 30mL ultra-clean HDPE bottles for trace metal analysis with sample ID (as
was done for DOC vials in step 11).

Arrange bottles under appropriate SPE cartridges and turn on syringe pump.
After vials are ﬁlled (~30 min, during this step cartridges can be run dry), turn off
syringe pump, replace bottom cap on SPE cartridge and cap DOC vials. Samples
for trace metal analysis should be ﬁlled to the base of the bottle neck, not
completely full like DOC samples.

Add 180 11L of Optima nitric acid to each trace metal fraction, cap bottles and
store in ﬁidge (4 °C) until analysis.

Remove syringe-SPE cartridge assembly from syringe pump.

If sample remains in cartridge or syringe, push solution through cartridge until
SPE media is dry.

Disconnect syringe-SPE cartridge assembly.

Place adapter in bin to be washed.

203

33. Discard used syringe in sharps box for disposal.

34. Cap both ends of SPE cartridge and place in Ziplock bag. Place all cartridges used
to fraction a sample in the same bag. On outside of bag write base sample name,
the date and your initials. Place bag in freezer.

35. Repeat steps 29-34 for all cartridges used.

204

Identifying Unique Land Use Micro-Watersheds

Catchments for sampling were initially identiﬁed using geographic information
systems (GIS) by overlaying hydrologic data on land use classiﬁcations to determine
rough boundaries of micro-watersheds that contained unique land use characteristics in
ArcMAP 9.1 (ESRI, 2003). Land use was identiﬁed based on the Michigan Land
Cover/Use Classiﬁcation System (MDNR, 2001b). Micro-watersheds with unique land
use characteristics were identiﬁed by conceptually moving sampling locations from the
farthest point upstream (i.e. water source or headwater) of the smallest hydrologic units
(intermittent streams) down stream until just before the catchment would contain more
than one land use. For example, the traditional method for determining watershed
boundaries, based solely on topography, often produces catchments with multiple land

uses, sample location A.

 

 

 

 

forested , . . . . . ....... ,
olooooo°°......”. ‘ do

I
I

 

 

/ l.
’ agricultural |

a - l

’ O
J

 

 

 

 

 

     

suburban

Stream Flow

 

 

Figure A-3. Differences in land use for traditional and new watershed boundaries.

205

By starting at the farthest point upstream and moving downstream until just before the
catchment would contain multiple land uses (i.e. moving the sampling location farther
upstream) a sampling location is identiﬁed for a micro-watershed that contains only one
type of land use, sample location B. After identifying potential micro-watersheds were
then inspected to determine if sampling locations existed where overland ﬂow could be
collected. Micro-watersheds were manually delineated based on topography. For urban
systems and agricultural ﬁelds where storm sewers and drain tiles ﬁelds alter natural
watershed boundaries, maps containing drainage networks were used to accurately
describe catchments boundaries. Land use data was based on the IFMAP/GAP Lower
Peninsula Land Cover raster data set (MDNR, 2001a). Land use within the Ramey
Chandler drain in Ingham County was updated to include the type of land use present at
the time of sampling. Topography was determined by county digital elevation maps

(MDEQ, 2005).

Procedure used to identify unique land use micro-watersheds within the Remey-
Chandler Drainage
1. The following datasets were downloaded from the Michigan Geographic Data
Library (www.michigan.gov/cgi_/):
a. Lower Peninsula Land Cover 2001 (IFMAP/GAP Lower Peninsula Land
Cover)
b. Clinton Digital Elevation Model
0. Ingham Digital Elevation Model

d. I4-Grand Watersheds

206

e. Clinton MI Geographic Framework Hydrography
f. Ingham MI Geographic Framework Hydrograph y

2. A Personal Geodatabase (RemyChandler.mdb) was set up for the Remey Chandler
drain using the following coordinate system:

AD_1 983_Hotine_Ob1ique_Mercator_Azimuth_Natural_Ori gin
Hotine_Oblique_Mercator_Azimuth_Natural_Ori gin
False_Easting: 2546731 .496000

F alse_Northing: ~43 54009.816000

Scale_Factor: 0.999600

Azimuth: 337.255560

Longitude_Of_Center: -86.000000

Latitude_Of_Center: 45.309167

3. All ﬁles previously downloaded were then imported into the RemyChandler
personal geodatabase and projections were adjusted to align all layers.

4. AutoCAD ﬁles containing drainage maps for storm sewer networks and tile drain
ﬁelds for Clinton and Ingham counties were obtained ﬁ'om the Clinton County
Drain Commissioner’s Ofﬁce and Michigan State University. These ﬁles were
then imported into the RemyChandler personal geodatabase.

5. Blueprints containing the storm sewer networks (Section 18) from the City of East
Lansing were scanned and converted to digital images by Capital Imaging
(Lansing, MI). Images were rectiﬁed to GCS_North_American_1983 coordinate
system.

Table A-3. Master conversion ﬁle used to rectify data and the resulting bounding
coordinates.

17.313583 13.054817 626894.263280 246147.222737

8.171042 14.276063 622325.134476 246668.000578

14.957873 3.434961 625801.770873 241316.418030
9.832593 3.389521 623199.756292 241255.200115

207

Bounding coordinates

Horizontal

In decimal degrees
West: -84.495625

East: -84.481747

North: 42.750289
South: 42.742617

In projected or local coordinates
Left: 623002.406532
Right: 624123.153837
Top: 245521.961176
Bottom: 244690.416466

6. From the Grand River watershed shape ﬁle (14-Grand Watersheds), the Remey
Chandler Drain sub-watershed was selected (EPA 14-digit Hydrologic Unit Code:
4050004060120) and exported to its own shape ﬁle. This shape ﬁle was called
Watershed_Boundary and temporarily contained only one feature.

7. The raster dataset used for land cover/use (IF MAP/GAP Lower Peninsula Land
Cover) was cut and converted to a shape ﬁle for ﬁuther editing.

a. The IFMAP/GAP Lower Peninsula Land Cover raster datasets was
clipped by Remey Chandler drainage watershed boundary using Hawth’s
Analysis tools for ArcGIS
(www.spatialecologv.com/htools/cliprasterbypolysphp).

b. The newly clipped raster data ﬁle was converted to a shape ﬁle using the

Export to Shapeﬁle feature of ET Geowizards 9.3 (www.ian-ko.com). The

 

new shapeﬁle was named Old_landuse.

208

8.

10.

c. Using the SampleShedCleanup script written in python, similar polygons
with the same land uses were grouped together.

The Old_landuse shapeﬁle was manually updated using current land cover/use
obtained visually from a MrSID raster database of the study area obtained from
AirPhotoUSA (www.airphotousa.com) and renamed Updated_Landuse.
By overlaying multiple layers containing the hydrology, digital elevation models
(DEMs), and relevant drainage networks (i.e. tile drains and storm sewer
networks), potential sampling locations were identiﬁed. As explained previously,
potential sampling locations with unique land use characteristics were identiﬁed
by conceptually moving the farthest point upstream (i.e. water source or
headwater) of the smallest hydrologic units (intermittent streams) down stream
until just before a catchment would contain more than one land use. From a
potential sampling location, the DEM was used to manually delineate a micro-
watershed. Using the Watershed_Boundary shape ﬁle as a template, rrricro-
watersheds were created by sketching a new polygon and using the Cut Polygon

Features option in ArcEditor. The new shape ﬁle was called sample_sheds.

After multiple sampling locations were identiﬁed and potential micro-watersheds
were created in the sample_sheds shape ﬁle, potential sampling locations were
ﬁeld investigated to ensure the digital information reﬂect real world
characteristics. The following characteristics were required to conﬁrm the validity

of all sampling locations:

a. current land cover/use did not deviate from that described by the

Updated_Landuse shape ﬁle,

209

it was possible to access the sampling location (e. g. there were no fences

restricting access),

Surface water runoff would be free ﬂowing stormwater runoff event (i.e.

no backwater effects would create mixed water samples),

the topography did not visually deviate from that described in the DEM,

and, no additional drainage networks (i.e. storm sewers or tile drain ﬁelds)

were present in catchments that were not described in GIS databases.

210

Molecular Weight Characterization

The molecular weight (MW) and polydispersity of DOC was determined by size
exclusion chromatography (HPSEC). The HPSEC system employed utilized a Gilson
Model 303 prunp (Middleton, WI) with a 20uL loop, a Waters Protein-Pak (125
angstrom, 7.8 x 300mm, part number WAT084601) modiﬁed silica column (Milford,
MA) and UV detection at 254nm on a Dionex Variable Wavelength Detector (Sunnyvale,
CA) (Chin et al., 1994; Zhou et al., 2000). The mobile phase consisted of 0.1M NaCl,
0.002M KHzPO4 and 0.002M Nazi-IP04 solution buffered to an approximate pH of 6.8
(Meier et al., 2004) and the ﬂow rate was 0.5 mL min'l. Calibration was performed using
1.8, 5.4, 8, 18 and 35kDa average MW random coil sodium polystyrene sulfonate
standards (Polysciences, Inc.) and acetone (58Da) (Zhou et al., 2000). To prevent
biofouling, the column was rinsed with and stored in 0.05% sodium azide solution before
and after running samples.

The number-averaged MW (Mn) and weight-averaged MW (M.,) were calculated

by the following equations:

N N
M1, = Zn, Emmi) (Equation A-2)
i=1 i=1
and
N N
Mw = Zh,(M,)/2h, (Equation A-3)
i=1 '=1

where h, is the height and M,- is the molecular mass of the sample HPSEC eluted at
volume i. Polydisperity (p) is the ratio of the weight-averaged MW and the number-

averaged MW:

211

Mw
Mn

 

p = (Equation A-4)

The extent of symmetrical (A) and asymmetrical (sk) band-broadening were calculated
for the 5.4kDa standard in every calibration to ensure the primary mechanism of
separation was size exclusion. The extent of symmetrical band-broadening was calculated

using the following equation:

 

 

M M
A =[ n .. reported + w __ observed J (Equation A-5)

M n _ observed M w _ reported

where Mn_,eponed and Mw_reponed are the known average molecular number and average
molecular weight for the standard evaluated, and Mn_obscmd and Mw_obsemd are the
average molecular number and average molecular weight for the standard evaluated using
HPSEC. The extent of asymmetrical band-broadening was calculated using the following

equation:

Q—l .
sk = — E uatron A-7
(0+1) ( q )

where

 

 

M M
Q = [ n _ reported x w _ observed ] (Equation A- 8)

M n _ observed M w _ reported
No additional corrections were deemed necessary in order to calculate MW
characteristics when A< 1.05 and sk < 0.05 (Yau et al., 1979).
Sample preparation
1. Immediately after sampling, ﬁlter water sample through an acid washed (18%

HCl and deionized-distilled water for 24hrs) 0.45pm glass ﬁber ﬁlter into acid

212

washed (18% HCl and deionized-distilled water for 24hrs) 30mL polypropylene
bottle.

2. Once ﬁlled, bottle is capped and stored in the dark at 4°C.

3. Open new 5mL syringe and Millex-GV 0.22pm ﬁlter (Millipore; Billerica, MA).
Without touching bottom of syringe or removing ﬁlter from packaging, remove
plunger from syringe and attached syringe to ﬁlter unit. Fill syringe with 4mL of
sample. Taking care not to touch rubber syringe cap to sample solution, insert
plunger and discard 1mL of ﬁlter eluent.

4. Fill Alcott 8x35mm autosampler vial (Grace; Deerﬁeld, IL) with 900uL of
ﬁltered sample and insert cap.

Standard preparation

1. The following solutions were prepared in acid washed (18% HCl and deionized-
distilled water for 24hrs) glass bottles with Teﬂon caps using ultra-pure water
(>1 8mQ).

Table A-4. MW standards for calibration.

MW Standard Concentration Standard Average MW (Da)
1 3 mL L" acetone 58
2 275 mg L" polystyrene sulfonate 1,800
3 225 mg L" polystyrene sulfonate 4,600
4 210 mg L'1 polystyrene sulfonate 8,000
5 200 mg L" polystyrene sulfonate 18,000
6 140 mg L" polystyrene sulfonate 35,000

2. For each set of calibration standards, a blank standard was prepared using ultra-
pure water (>18mﬂ).
3. For each sample run, a Suwannee River fulvic acid check standard (~5mg L")

was also prepared.

213

Instrument start-up

1.

Connect mobile phase solution (0.1M NaCl, 0.002M KHzPO4 and 0.002M
NazHPO4 buffered to an approximate pH of 6.8) and purge with helium (~2min).
After purging mobile phase, close valve and pressurize system.

Turn on UV detector. Set wavelength to 254nm and output range to 0.01au
(sampling rate should be once every second).

Turn on pump at a ﬂow rate of 0.2mL min". Every 20 minutes, increase ﬂow rate
0.1mL min". Backpressure for clean system should remain below 300psi. If
backpressure is higher the ﬁit most likely needs to be changed.

After 1hr warm up, the ﬂow rate should be 0.5 mL min'1 and the instrument is

ready for use.

Calibration and sample analysis

1.

Each run should start with the following sequence described in Table A-5.

where n is the number of vials run during analysis.

Run up to 20 samples between calibration curves with the Suwannee River FA
sample placed somewhere in the middle.

After standards have been run, plot log MW of known standards versus the
retention time of peak maximums and perform linear regression. Linear
regression can be used to calculate the molecular mass M; for unknown samples.
Before calculating M for samples, select one standard or sample (typically the
DDI water blank) with similar chemistry without DOC for baseline correction.
Subtract this baseline from each sample to correct for water dip and baseline shift

observed during the course of the run.

214

Table A-5. Typical arrangement of calibration standards.

Vial # Description
1 mobile phase
2 DD] water
3 MW standard 1
4 MW standard 2
5 MW standard 3
6 MW standard 4
7 MW standard 5
8 MW standard 6

n-6 DDI water
n-5 MW standard 1
n-4 MW standard 2
n-3 MW standard 3
n-2 MW standard 4
n-1 MW standard 5
n MW standard 6

5. Using a minimum peak height of 3 times the standard deviation of the signal

produced by the blank baseline, integrate and evaluate unknown DOC samples

using equations 1-3.

6. After the run is complete, use equations 4-6 to calculate the amount of
symmetrical (A) and asymmetrical (sk) band-broadening for standard #3 (5.4
kDa). If A > 1.05 or sk > 0.05, samples should be rerun.

Instrument shutdown

1. After last sample, ﬂush column with mobile phase turning ﬂow rate down 0.1mL
min'l every 20 minutes.

2. After 1hr of ﬂushing system with mobile phase (ﬂow rate should be 0.2mL min’

1), switch the mobile phase to 0.5mM sodium azide solution (pH ~ 6.2). At a ﬂow

rate of 0.2mL min") the system should be ﬂushed for 2hrs. After 2hrs, the system

215

can be turned off. The column should always be stored in 0.5mM sodium azide
solution.
Additional notes

1. Never allow backpressure to exceed 3,000psi. Once backpressure exceeds
1,000psi, complete run and change ﬁit. Backpressure should be <300psi for
system with new frit, guard-column and main column.

2. Never allow column to run dry. If column runs dry, ﬂow paths will collapse
and the column is ruined.

3. If samples are from relatively pristine waters, samples should be adjusted to
approximate pH and ionic strength of mobile phase before ﬁltering into

autosampler vials.

216

References
Chin, Y.P., Aiken, G. and Oloughlin, E., 1994. Molecular-Weight, Polydispersity, and
Spectroscopic Properties of Aquatic Humic Substances. Environmental Science &

Technology, 28(11): 1853-1858.

Meier, M., Chin, Y.P. and Maurice, P., 2004. Variations in the composition and
adsorption behavior of dissolved organic matter at a small, forested watershed.

Biogeochemistry, 67(1): 39-56.

Yau, W.W., Kirkland, 1.]. and Bly, DD, 1979. Modern Size-Exclusion Liquid
Chromatography: Practic of Gel Permeation and Gel Filtration Chromatography.

Wiley-Interscience, New York, 315-326 pp.

Zhou, Q.H., Cabaniss, SE. and Maurice, RA, 2000. Considerations in the use of high-
pressure size exclusion chromatography (HPSEC) for determining molecular

weights of aquatic humic substances. Water Research, 34( 14): 3505-3514.

217

Sample Collection Procedure

Sample preparation

Before leaving the laboratory to collect ﬁeld samples, a ﬁeld kit is required for

each sample to be collected. Field kits consist of the following items:

0 1 — pair of powder-free nitrile gloves (Kimberly-Clark, part number 50603)

0 1 — 3” x 1” strip of Paraﬁlm (V WR, part number 52858-000)

o 1 — 1L ultra-clean wedge shaped polypropylene bottle (ISCO, part number 68-
3700-001). Sample bottles should be labeled with unique identiﬁcation (ID)
numbers using permanent ink.

In addition to ﬁeld kits, a multi-parameter water quality meter (Horiba U-10 or YSI 556)
is used to measure dissolved oxygen (DO), conductivity, pH, temperature, salinity
(calculated by instrument based on conductivity) and turbidity at each sampling location.
The multi-parameter water quality meters should be calibrated according to the
instrument instruction manual before each sampling trip. A cooler with ice is also helpful
to store samples when collecting more than a few (4) samples during hot weather. Using
a cooler prevents samples from being exposed to UV light and helps to maintain natural
sample temperature. Samples should not be allowed to exceed 25°C.

Once samples are collected from the ﬁeld, they are ﬁrrther processed into

additional sample bottles. Laboratory analysis requires the following items to determine
alkalinity, total suspended solids (TSS), anions, major cations, molecular weight (MW),

total nitrogen and total phosphorus (TN/TP), trace metal and dissolved organic carbon

218

(DOC) analysis (including 7 fractions typically run for DOC and trace metal analysis) for
each sample:

0 8 — 30mL ultra-clean high-density polyethylene (HDPE) bottles for cation
analysis. Bottles should be wrapped twice with %” orange laboratory tape
(TimeMed, part number T-3460-6) and marked “Cations”. Note: One bottle is
used to collect an unfractioned sample for measurement of the total amount of
major cations and trace metals present in solution; the other seven samples are
for fractions. If more fractions are collected, additional bottles are necessary.

0 1 — 30mL ultra-clean HDPE bottle for anion analysis. Bottle should be
wrapped twice with %” blue laboratory tape (TimeMed, part number T-3460-
7) and 250uL of ACS grade 37% formaldehyde (Mallinckrodt, part number
5016-04) should be added. Bottle should be labeled for “Anions” analysis and
“+250uL of formaldehyde” to denote the added preservative.

o 1 — 30mL acid washed HDPE bottle (Fisher Scientiﬁc, part number 01-288-
33) for MW analysis. Bottle should be soaked in 18% HCl bath for ~24hrs,
rinsed with 3 times with ultra-pure (>18 M0) water, soaked in ultra-pure
water for ~24hrs and dried. Once dry, bottles should be wrapped twice with
3A” yellow laboratory tape (TimeMed, part number T-3460-2) and labeled
“MW”.

0 1 —— 30mL acid washed HDPE bottle (Fisher Scientiﬁc, part number 01-288-
33) for TN/TP analysis. Bottle should be soaked in 18% HCl bath for ~24hrs,

rinsed 3 times with ultra-pure water, soaked in ultra-pure water for ~24hrs and

219

dried. Once dry, bottles should be wrapped twice with 3/4” green laboratory
tape (TimeMed, part number T-3460-3) and labeled “TN/T P”.

o 8 — 40mL clean amber colored, borosilicate glass vials (I-Chem, part number
8346-0040) capped with polytetraﬂuorethylene (PTFE)/silicone septa
(0.010/0.050in; Alltech, part number 95322) and open top screw caps (20-
400; Alltech, part number 95321). Vials should be acid washed in 18% HCl
bath for ~24hrs, rinsed 3 times with ultra-pure water, soaked in ultra-pure
water for ~24hrs and dried in oven at 550°C for 1hr. Septa should be soaked
in ultra-pure water for ~24hrs and dried on clean counter paper (V WR, part
number 52857-120). Clean and capped vial should be wrapped with 3/4” white
laboratory tape (TimeMed, part number T-3460-1) and labeled “TOC”. Note:
One bottle is used to collect an unfractioned sample for the total amount of
major cations and trace metals present in solution; the other seven are for
fractions. If more fractions are collected, additional bottles are necessary.

0 1-6 (depending on sample turbidity) - acid washed 47mm type A/E glass ﬁber
ﬁlters (Pall Corporation, part number 61631). Filters should be acid washed in.
18% HCl bath for ~24hrs, rinsed 3 times with ultra-pure water, soaked in
ultra-pure water for ~24hrs, and dried in clean hood.

0 1 — ultra-clean ﬁlter assembly: ﬁlter ﬂask, fritted glass support and Buchner
funnel (Kontes, part number 953 845-0000). Aultra-puretionally, a clamp for
ﬁlter assembly, vacuum tubing and a vacuum pump are required to ﬁlter
samples.

0 Optima Nitric Acid (Fisher Scientiﬁc, part number A467-250)

220

Sample collection

Collecting samples required two people, one person who handles only ultra-clean

equipment (clean hands) and one who handles other sampling equipment (dirty hands).

Below describes the procedure for each person by identifying them as clean hands and

dirty hands.

1.

Both clean hands and dirty hands begin sampling by putting on a new pair of
powder-free nitrile gloves (Kimberly-Clark, part number 50603). A new pair
is required for each sample collected.

Dirty hands — open ﬁeld kit bag and hold for the clean hands

Clean hands — remove second pair of nitrile gloves from inside ﬁeld kit and
places on top of ﬁrst pair (double gloving).

Dirty hands — record bottle ID, sampling location, time, and weather
conditions on sampling sheet.

Clean hands — remove 1L sample bottle from ﬁeld kit. At sampling location
clean hands should remove cap from IL sample bottle and rinse both 3 times
with sample water. '

Dirty hands —- while clean hands is collecting sample, collect any additional
data or samples (i.e. multi-parameter water quality meter measurements).
Clean hands -— collect sample and cap bottle. If possible, submerge the sample
bottle completely and cap while underwater. Sample should be collected to

minimize the amount of air trapped in sample bottle (the goal is not to have an

air bubble).

221

8. Clean hands — remove Paraﬁlrn from ﬁeld kit and wrap around sample cap to
ensure seal.

9. Clean hands — place full sample bottle back in ﬁeld kit bag.

10. Dirty hands — close ﬁeld kit bag and seal with tape or by tying knot.

11. If collecting more than one sample, place used ﬁeld kits upright (to minimize

the possibility of sample leakage) in a cooler.

Sample processing
After collecting samples, ﬁeld kits are taken back to the laboratory to ﬁlter and
further processes aliquots for speciﬁc analyses. Directions below are for isolating and
preserving samples intended for the measurement of alkalinity, TSS, anions, major
cations, MW, TN/T P, trace metal and DOC (including 7 fractions typically run for DOC
and trace metal analysis). Analyst should wear a new pair of nitrile gloves for each
sample.
1. Remove sample bottles from ﬁeld kit bags and allow them to reach room
temperature (usually take 30-60 minutes).
2. Collect TSS and ﬁlter sample
a. Remove ultra-clean type A/E ﬁlter from pre-weighed, pre-numbered
zip-lock bag and place in a new ultra-clean ﬁlter assembly (Kontes,
part number 953845-0000).
b. Clamp ﬁlter paper between the ﬁitted glass support and funnel.
c. Attach vacuum tubing to vacuum pump.

d. Shake sample vigorously to ensure fully mixed.

222

. Turn on vacuum pump and ﬁlter 100mL of sample at a time, collecting

all eluent in the 1L ultra-clean ﬂask.
If ﬁlter becomes clogged (takes longer than 60 sec to pass 100mL
through ﬁlter):

i. Completely ﬁlter the last 100mL aliquot of the sample and
allow the sample to dry by passing air through the ﬁlter by
vacuum suction.

ii. Turn off vacuum pump
iii. Remove ﬁlter paper and place in pre-weighed ultra-clean zip-
lock bag.
iv. Place a new ultra-clean type A/E ﬁlter in the ﬁlter assembly
currently being used and restart at step b. Repeat as necessary
. If ﬁlter does not clog, continue adding 100mL aliquots until the full 1L
of sample has been ﬁltered.
. Turn off vacuum pump
Without breaking the seal on the glass ﬁber ﬁlter, remove ﬁitted glass
support and Buchner funnel (with clamp) and place on a dirty 1L ﬁlter
ﬂask.
Cover original ﬂask containing ﬁltered sample with Paraﬁlm.
. Turn on vacuum pump.
Rinse ﬁlter and Buchner ﬁrnnel with ultra-pure water so that all

particles are removed from glassware onto ﬁlter surface.

223

m. Continue to pull air through ﬁlter paper using vacuum suction until
ﬁlter paper is dry.

n. Turn off vacuum pump

0. Remove ﬁlter paper and place in pre-weighed ultra-clean zip-lock bag.

p. All ﬁlter papers and their numbered zip-lock bags used for each
sample should be sealed in a larger zip-lock bag. The larger zip-lock
bag should be labeled with the sample ID, the date and time of
ﬁltration and the analyst’s initials.

q. Immediately place TSS samples in freezer for storage.

r. The ﬁitted glass support and Buchner funnel can be placed in the dirty
glassware bin for washing. The spare ﬁlter ﬂask containing ultra-pure
water rinse can also be emptied and placed in the dirty glassware bin
for washing; however it can still be used for ﬁltering additional
samples if necessary.

8. Filtered samples in glass ﬂasks can be stored cover with Paraﬁlrn in
dark at 4°C for up to a week. If samples are refrigerated, they should
be allowed to reach room temperature before ﬁrrther processing.

3. Divide ﬁltered sample into appropriate containers for further analysis. From
1L glass ﬂask:

a. Pour ~15mL into pre-weighted disposable beaker for alkalinity

titration.

224

i.

Fill 40mL clean amber colored, borosilicate glass vials and cap with
PTFE/silicone septa for TOC analysis (seal sample without an creating
an air bubble). Label vial with sample ID.

Fill 30mL acid washed HDPE bottle for MW analysis. Label bottle
with sample ID.

Fill 30mL acid washed HDPE bottle for TN/TP analysis. Label bottle
with sample ID.

Fill 30mL ultra-clean HDPE bottle (containing 250uL of 37%
formaldehyde) for anion analysis. Label bottle with sample ID.

Fill 30mL ultra-clean HDPE bottles for major cation and trace metal
analysis. Add 180uL of Optima nitric acid to prevent precipitation.
Label bottle with sample ID.

Perform DOC ﬁactionation.

Place vials for TOC analysis and bottles for molecular weight, anion
and cation analysis in 4°C ﬁidge for storage.

Place bottles for TN/TP analysis in freezer for storage.

4. Complete necessary analyses.

a.

Alkalinity should be run immediately (same day the sample was
collected).

Anion, TN/T P and DOC samples should be run within 1 week of
collecting the sample.

Maj or cations and trace metal analysis should be run within 6 months

of collecting the sample.

225

5. After all samples and fractions have been collected the glass ﬂasks can be

washed according the ultra-clean procedure.

Additional notes
0 If additional sample remains after all samples have been collected, duplicate
samples can be collected. A duplicate sample for unfractioned DOC is extremely
helpful for identifying any possible dilutions that may be necessary for running

fractioned and unfractioned DOC samples.

226

Sample Watershed Cleanup Script

##
## sshedcleanpy

## Created on: Monday December 05, 2005 10:30AM

## Created by: Shawn P. McElmurry and Saradhi Balla

## Usage: sshedclean <Input_Shapeﬁle> <Diss_Field>

## Description: This program consolidates and recalculates landuse areas within micro-
watersheds.

## Landuse area is calculated ﬁrst and then summed based on the dissolve ﬁeld.

## Individual landuse codes will create individual features. This will

## delete all ﬁelds except landuse levels and labels, and shape area.

## MAKE BACKUP BEFORE RUNNING SCRIPT!

HH
I] n

 

# Import system modules
import sys, string, 08, win32com.client
# Create the Geoprocessor object
gp = win3200m.client.Dispatch("esriGeoprocessing.GpDispatch.1")
# Load required toolboxes...
gp.AddToolbox("CzlProgram Files/ArcGIS/ArcToolbox/Toolboxes/Spatia1 Statistics
Tools.tbx")
gp.AddToolbox("Cz/Program F iles/ArcGIS/ArcToolbox/Toolboxes/Data Management
Tools.tbx")
gp.AddToolbox("C:/Docurnents and Settings/Shawn/Application
Data/ESRI/ArcToolbox/My Toolboxes/ Shawn's Tools.tbx")
#Set the input shapeﬁle to clean
Input__Shapeﬁle = sys.argv[ 1]
Dis_Field = sys.argv[2]
Output_shed = Input_Shapeﬁle
# Process: Calculate Areas...
try:
# Output of calculate area creates temporary shapeﬁle that will later be deleted...
temp_shed = os.path.dimame(Input_Shapeﬁle) + "\\temp_" +
os.path.basename(lnput_Shapeﬁle)
gp.CalculateAreas_stats(Input_Shapeﬁ1e, temp_shed)
except:
gp.addMessage("ERROR-Failed to calculate areas")
del gp
sys.exit
# Process: Dissolve Landuses...
try:
gp.delete(lnput_Shapeﬁle)
except:
gp.addMessage("ERROR-Failed to delete input shapeﬁle.")
del gp
sys.exit

227

try:

# Dis_Field = "CODE"

Stat_Field = "LABEL F IRST;LEVEL1 FIRST;LABELI FIRST;LEVEL2
FIRST;LABEL2 FIRST;LEVEL3 FIRST;LABEL3 FIRST;F__AREA SUM"

gp.Dissolve_management(temp_shed, Output_shed, Dis_Field, Stat_Field)
except:

gp.addMessage("ERROR-Failed to dissolve landuses.")

del gp
sys.exit
try:
gp.delete(temp_shed)
except:
gp.addMessage("ERROR-Failed to delete temp ﬁle.")
del gp
sys.exit
gp.addMessage(" ----- Calculate and dissolve successfulll")
## RENAME FIELD

 

 

rename_ﬁle = Output_shed
# Need to cycle through ﬁelds that have been created by dissolve to rename...
try:
ﬁelds = gp.ListFields(rename_ﬁle)
ﬁelds.reset()
ﬁeld = ﬁelds.next()
while ﬁeld:
if ﬁeld.Name == "FIRST_LABE":
old_ﬁeld = ﬁeld.Name
new_ﬁeld = "LABEL"
gp.Rename_Field(rename_ﬁle, old_ﬁeld, new_ﬁeld)
elif ﬁeld.Name = "FIRST_LEVE":
old_ﬁeld = ﬁeld.Name
new_ﬁeld = "LEVELl "
gp.Rename_Field(rename_ﬁle, old_ﬁeld, new_ﬁeld)
elif ﬁeld.Name = "FIRST_LA_l":
old_ﬁeld = ﬁeld.Name
new_ﬁeld = "LABEL1"
gp.Rename_Field(rename_ﬁle, old_ﬁeld, new_ﬁeld)
elif ﬁeld.Name = "FIRST_LE_1":
old_ﬁeld = ﬁeld.Name
new_ﬁeld = "LEVEL2"
gp.Rename_Field(rename_ﬁle, old_ﬁeld, new_ﬁeld)
elif ﬁeld.Name == "FIRST_LA_2":
old_ﬁeld = ﬁeld.Name
new_ﬁeld = "LABEL2"
gp.Rename_Field(rename_ﬁle, old_ﬁeld, new_ﬁeld)
elif ﬁeld.Name = "FIRST_LE_2":

228

old_ﬁeld = ﬁeld.Name
new_ﬁeld = "LEVEL3"
gp.Rename_Field(rename_ﬁle, old_ﬁeld, new_ﬁeld)
elif ﬁeld.Name = "FIRST_LA_3":
old_ﬁeld = ﬁeld.Name
new_ﬁeld = "LABEL3"
gp.Rename_Field(rename_ﬁle, old_ﬁeld, new_ﬁeld)
elif ﬁeld.Name = "SUM_F_AREA":
old_ﬁeld = ﬁeld.Name
new_ﬁeld = "SHED_AREA"
gp.Rename_Field(rename_ﬁle, old_ﬁeld, new_ﬁeld)
break
ﬁeld = ﬁelds.next()
gp.addMessage(" ----- Rename ﬁeld successfulll")
except:
gp.addMessage("ERROR-Did Not Rename Fieldl")
del gp
sys.exit

 

gp.addMessage("

 

ﬂ)
gp.addMessage("CONGRATULATIONS," + rename_ﬁle + " has been cleaned!")
gp.addMessage("

 

 

 

II)

229

Total Organic Carbon Analysis

Dissolved organic carbon (DOC) was determined by total organic carbon analysis
(TOC) using an 01 Analytical Model 1010 Wet Oxidation Total Organic Carbon
Analyzer (College Station, TX). The automated analysis utilizes heated-persulfate
oxidation to measure the total amount of organic carbon present in aqueous samples
(1998). Generally, samples were ﬁltered through a 0.45pm ﬁlter into acid washed
borosilicate glass vials that were capped with TEE/silicone liners and refrigerated until
analysis, typically within 1-2 days. Clean gloves and counter paper should be used at all

times when performing procedure described below.

Sample preparation

1. Soak 40mL amber colored, borosilicate glass vials (I-Chem, part number S346-
0040) in 18% HCl acid bath for 24hrs.

2. Soak polytetraﬂuoroethylene (PTFE)/silicone septa (0.010/0.050in; Alltech, part
number 95322) and open top screw caps (20-400; Alltech, part number 95321) in
ultra-pure water for 24hrs.

3. Remove vials from acid bath after 24hrs, rinse in ultra-pure water and bake in
oven 550°C for 1hr.

4. Remove septa from ultra-pure water after 24hrs and dry on counter paper in hood
(usually ~12hrs).

5. After baking for 1hr, remove vials from oven, cover with aluminum foil and allow

vials to come back to room temperature (usually ~2hrs).

230

Sample collection and preservation

1.

Fill sample vials until meniscus becomes inverted above the vial top. When
capping with septa, attempt to minimize the amount of air trapped. If sample vial
is full, it should be possible to seal vial without creating an air bubble.
Immediately after sampling, place sample vials in dark ﬁidge (4°C) for up to one
week. Although no limit on storage time is given by Stande Methods (1998),

run all samples with in one week of sample collection.

Operating the 01 Analytical 1010 TOC Analyzer

1.

Turn on ultra-pure nitrogen (>99.99% or better) gas supply to a line pressure of
60 psi.

Turn on TOC analyzer and auto sample tray. Allow analyzer to fully boot.

Check to make sure rinse bottle is empty. Empty any rinse water down drain with
copious amounts of tap water.

Fill water reservoir with ultra-pure water (>18MO). If this jug is needs to be
ﬁlled, carefully remove the cap (do not twist lines) and make sure the submerged
tubing remains clean when ﬁlling bottle. This can often be accomplished by
placing the lines on clean counter paper. After rinsing the tubing with clean water,
place lines back into reservoir and make sure nitrogen is bubbling.

Make sure acid and oxidant reagents for TOC machine are full (>250mL for 59

vial run) and fresh (< 3 months old).

231

a. To make acid, ﬁll 1L bottle with approximately 500mL of ultra-pure water
(>18MQ), add 59mL of ACS grade (85%) phosphoric acid and ﬁll to with
ultra-pure water.

b. To make oxidant, add 100gm of low organic carbon sodium persulfate (01
Analytical, part number 178848) to acid washed 1L volumetric ﬂask, ﬁll
to IL with ultra-pure water(>18MQ), add acid washed stir bar and stir
until fully dissolved.

6. Make sure there is a sufﬁcient amount of unused desiccant. For a 24 hr run, 3A of
the tube should be unused (blue). Every time the desiccant tube is replaced a leak
check should be performed.

7. Turn on computer and click the winTOC 1010 icon.

8. Select System I.

9. Make sure there is an IR signal (red number in lower left of window). If there is
no signal an error warning will be displayed in message window. Allow system to
warm up until IR signal stabilizes around 6,000-10,000.

10. Remove the cover to observe the pumps and injection loops.

11. Flush system with reagents:

a. Click instrument on the toolbar; click diagnostics; type in 99 for oxidant
pump and 99 for acid pump. Click pump.

b. This should make both pumps perform 99 strokes.

0. Watch for leak.

d. Notice glass reaction chamber ﬁlling with solution

232

6.

Without closing Diagnostics window; in the miscellaneous block, click
drain. After all the liquid has drained from reaction vessel, click drain

again to turn it off (keep the diagnostics window open).

12. Check the alignment of autosampler.

a.

g.

With a vial in sample tray position #1, place the sample tray on metal
platform making sure to line up center hole and guide pine.

Once the sample tray is positioned securely— click go to ﬁrst vial in the
autosampler block. This will cause the tray to rotate the ﬁrst sample vial
directly under the needle.

Again in the Diagnostics window, click calibrate in the Autosarnpler
block.

Manually align the center of the vial (while the autosampler tray is still
positioned securely on metal platform) with the sample needle with the
center of vial by moving metal platform in any direction. Click done.
While still in the diagnostics window and without adjusting the tray or vial
after the previous calibration, click needle down. The auto-sarnpling
needle should go into the center of the septa of vial #1. When ﬁnished,
click needle up.

Click home and the tray should adjust so that the tray is positioned with
the needle slightly ahead vial #1.

Close diagnostics window.

13. Setup output ﬁles.

a.

Click setup on the toolbar and then Win T 0C Output on pop-down menu.

233

b. Enter subdirectory (typically your name, up to 6 characters long)
c. Enter log ﬁle name (typically the date in MMDDYY format)
d. Set the output ﬁle preﬁx (typically the month and day in MMDD format).

The preﬁx and counter number are used to create result ﬁles.

.0

Reset output ﬁle counter to desired number.
Sample analysis

1. Under the instrument menu, select sequence.

2. Load standards and samples in the following order:

Table A-6. Sequence of TOC standards.

Position # Sample Name Run Type

1 DDI Water Sample
2 TOC STD #1 Std 1
3 TOC STD #2 Std 2
4 TOC STD #3 Std 3
5 TOC STD #4 Std 4
6 TOC STD #5 Std 5
7 Sample 1 Sample
etc.

Before the run is started and any calibration is performed, reagent blanks should
be performed. For an instrument that has not been turned off since the last time
samples were run and no changes have been made to reagents, 3 reagent blanks
should be run. For an instrument that has run samples recently (7 days) and no
changes have been made to reagent, or reagents have been changed since the last
calibration, 6 reagent blanks should be run. Any time the instrument has been idle

for longer than 7 days, 9 reagent blanks have been run.

234

3. Every 6-20 samples, check standards should be run. For these standards, the run

type should be changed to the appropriate check standard in the drop down list.

235

References

Standard Methods, 1998. Method 5310 C-Total Organic Carbon, Heated-Persulfate
Oxidation Method. Standard Methods for the Examination of Water and
Wastewater. American Public Health Association, American Water Works

Association, and Water Environment Federation, Washington, DC, 5-22 pp.

236

Ultra-Clean Procedure

Procedures described below should be performed in an EPA class 100 clean room. The
procedure employs strong acids and requires the use of safety goggles, acid apron, closed
toed shoes, long sleeves, long pants and appropriate gloves. While thick rubber gloves are
typically used when handling strong acids, the rubber gloves initially used were found to
leach metals. Due to the potential for contamination, two pairs of powder-free nitrile
gloves (Kimberly-Clark, part number 50603) were used instead. Because nitrile gloves
are not designed to withstand the strong acid solutions used they should be changed
frequently, especially when the gloves begin to change color (from purple to blue). Other
gloves may be more appropriate (6. g. Kimberly-Clark Kim Tech Pure G3 Cleanroom
Acid Gloves, part number 40222), however the extent these gloves leach trace metals has
not been investigated and should be veriﬁed before routine use.

For all cleaning procedures described below, the following acid baths should be
prepared in appropriate size containers (heavy duty plastic containers designed to
withstand strong acids).

0 18% HCl solution prepared in ultra-pure (>18MQ) water using trace metal grade
acid (Fisher Scientiﬁc, part number A508-212)
0 35% HNO3 solution prepared in ultra-pure water using trace metal grade acid
(Fisher Scientiﬁc, part number A509-212)
Cleaning procedure for 30mL bottles and 60mL syringes
1. The following acid baths should be prepared in appropriate size containers (heavy

duty plastic containers designed to withstand strong acids).

237

a. 18% HCl solution prepared in ultra-pure water using trace metal grade
acid (Fisher Scientiﬁc, part number A508-212)

b. 35% HNO3 solution prepared in ultra-pure water using trace metal grade
acid (Fisher Scientiﬁc, part number A509-212)

. While wearing powder-ﬂee nitrile gloves, open 60mL syringes (Becton

Dickinson, part number 309620) and place rubber caps in ultra-clean volumetric

ﬂask.

. Place new 30mL hi gh-density polyethylene (HDPE) bottles and caps (narrow

mouth bottles are available from Fisher Scientiﬁc, part number 01-288-33), and

60mL syringes (plungers and tubes) in 18% HCl bath. Batches of 21 syringes and

~60 bottles works well. Soak for ~ 24hrs.

. Rinse rubber caps 3 times with ultra-pure water. Fill volumetric ﬂask containing

rubber caps with ultra-pure water, cover with Paraﬁlrn (VWR, part number

52858-000) and soak for greater than 24hrs. Do not allow Paraﬁlm to contact

solution.

. Carefully transfer bottles and syringes (plungers and tubes) from 18% HCl bath to

35% HNO3 bath. Soak for ~24hrs at 20 °C.

. Remove rubber caps from ultra-pure water solution, shake/tap caps to help

remove water and place in tubs lined with counter paper (VWR, part number

52857-120), paper side up. Allow to dry, typically 2-3 days.

. Fill bottles with 35% HNO; and cap. Place capped bottles in square polycarbonate

bin suitable for hot acid bath. Fill bin with ultra-pure water to submerge bottles.

Transfer syringes (plungers and tubes) in square polycarbonate bin suitable for

238

acid bath. Fill bin with syringes with 35% HNO3. Place sealed bins in hot water
bath at 45 °C for ~24hrs.

8. Remove bins from water bath. Pour HNO3 from square polycarbonate bin ﬁlled
with syringes back into HNO3 bath. Rinse syringes 3 times with ultra-pure water
and soak for ~24hrs. For the square polycarbonate bin ﬁlled with capped bottles,
pour ultra-pure water down drain. With plenty of water from tap (neutralize if
necessary), uncap acid washed bottles and dispose of used acid. Rinse bottles and
caps 3 times with ultra-pure water and soak for ~24hrs. Be careful not to allow tap
water to contaminate acid washed labware.

9. Cut 3 new 3’ x 3’ pieces of counter paper (VWR, part number 52857—120). Place
one, paper side up, on counter and line a drying tub with another. The third piece
of counter paper will be used to cover tub during drying. Remove syringes and
bottles from ultra-pure water soak, tap labware on counter paper to remove as
much water as possible and place in tub lined with counter paper. Allow to dry,
typically 2-3 days.

10. Once dry, cap sample bottles, replace syringe cap on plunger head and insert
plunger into syringe barrel. Place bottles and syringes in plastic bags and seal
bags with laboratory tape (TimeMed, part munber T-1260-l). On tap mark the
number of bottles or syringes in bag. Double bag each set of ultra-cleaned items
and seal with lab tape. Label outside of bag with: “ULTRA-CLEAN”, quantity,

your name and date.

239

Cleaning procedure for 1L sample bottles and glassware

The following procedure should be used for 1L wedge shaped polypropylene bottles

(ISCO, part number 68-3 700-001), or similar size sample bottles, and any glassware

that comes in contact with water samples intended for trace metal analysis. For

example, the removal of total suspended solids (TSS) required a ﬁltration assembly to

be ultra-clean for each sample: ﬁlter ﬂask, fritted glass support and Buchner funnel

(Kontes, part number 953845-0000).

1.

Wash bottles and glassware using liquid soap solution of diluted Liqui-Nox
detergent (VWR, part number 21837-027) and rinse 3 times with ultra-pure water.
Place in 18% HCl bath. Soak for ~ 24hrs.

Carefully transfer from 18% HCl bath to 35% HNO3 bath. Soak for ~24hrs at 20
°C.

Remove from acid bath, rinse 3 times with ultra-pure water and soak for ~24hrs.
Cut 3 new 3’ x 3’ pieces of counter paper (VWR, part number 52857-120). Place
one, paper side up, on counter and line a drying tub with another. The third piece
of counter paper will be used to cover tub during drying. Remove bottles and
glassware from ultra-pure water soak, tap on counter paper to remove as much
water as possible and place in tub lined with counter paper. Allow to dry,
typically 2-3 days.

Once dry, place bottles in plastic bags and seal with laboratory tape (TimeMed,
part number T-1260-l). On tap mark the number of bottles in bag. Double bag
each set of ultra-cleaned items and seal with lab tape. Label outside of bag with:

“ULTRA-CLEAN”, quantity, your name and date. Glassware should be covered

240

with Paraﬁlm (VWR, part number 5285 8-000), when appropriate, and stored
separately from other glassware.
Additional notes

0 When working with labware at multiple stages in the ultra-clean process, begin by
moving labware at the clean end of the process forward (i.e. from ultra-pure water
rinse to drying tubs) and work backwards to the dirty end, the start of the acid
washing procedure (i.e. placing labware in 18% HCl acid bath).

0 To decrease drying time, tubs can be left uncovered in hood. Additionally,
periodically tapping labware on counter paper to remove water droplets will

greatly decrease drying time. Drying time can be reduced to as little as 1-2 days.

241

APPENDIX B

SAMPLE LOCATIONS

242

Table B-1. Sample location key.

 

 

Location ID Latitude Logitude
8 42.793732 -84.499154
9 42.791795 -84.496150
14 42.782244 -84.483383
17 42.773184 -84.488345
18 42.782102 -84.472273
28 42.782299 -84.453063 -
37 42.813222 -84.538336
64 42.789039 -84.515634
84 42.819833 -84.549150
87 42.765698 -84.512780
88 42.791685 -84.476050
95 42.753488 -84.505666
96 42.757112 -84.503660
98 42.763575 -84.486242
99 42.746776 -84.502062
100 42.758727 -84.493629
101 42.798424 -84.495828
A 42.685137 -84.483903
AG 42.685137 -84.483903
W 42.718248 -84.476216
F 42.718248 -84.476216
BW 42.718248 -84.476216
CR 42.942920 -85.804550
CS 42.773208 -84.488388
CSO 42.930600 -85.743680
DC 43.010620 -85.936450
DLM 42.976680 -85.876080
EE 42.790740 -84.485250
ERC 42.716124 -84.469627
FL 42.727458 -84.477729
G 42.710925 -84.488876
. GC 42.710925 -84.488876
GRER 42.534722 -84.623056
GRFU 42.964230 -85.676600
GRGR 42.968137 -85.676360
GRIO 42.971944 -85.069167
GRJA 42.281476 -84.409214
GRLA 42.750321 -84.555266

243

Table B-1 (cont’d)

 

 

Location ID Latitude Longitude
H131 42.948830 -85.7061 70
HHS 42.800029 -84.464392
LGEA 42.828056 -84.759444
OF22 42.729692 -84.483753
OF37 42.729554 -84.494557
OF38 42.730193 -84.491759
OF52 42.728088 -84.471678
OF53 42.728277 -84.462343
PL 42.712558 -84.482343
RCMSU 42.728569 -84.481242
RCR 42.728951 -84.482149
RCWI 42.683056 -84.219167
RRRO 43.072203 -85.597486
SW 42.728534 -84.463888
U28 42.920780 -85.764750

 

244

Table B-2. Samples collected.

 

 

Sample ID Date Sampled Sample Source Landuse ID
37-0425 April 25, 2007 Creek 11
14-0726 July 26, 2006 Creek 22
99-0816 August 16, 2006 Pond 1 12
100-0710 July 10, 2006 Creek 1 13
100-0711 July 11, 2006 Creek 113
100-080707 August 7, 2007 Creek 1 13
84-0425 April 25, 2007 Creek 1 13
64-0623 June 23, 2006 Pond 115
87-0425 April 25, 2007 Pond 122
87-0623 June 23, 2006 Pond 122
87-0711 July 11, 2006 Pond 122
96-0425 April 25, 2007 Creek 124
96-0622 June 22, 2006 Creek 124
SW-0829 August 29, 2006 Pond 411
BW-0622 June 22, 2005 Pond 431
BW-0729-T July 29, 2005 Pond 431
BW-103 1 October 31, 2006 Pond 431
BW-112905 November 29, 2005 Pond 431
BW-52506 May 25, 2006 Pond 431
BW-040407 April 4, 2007 Pond 431
BW-052507 May 25, 2007 Pond 431
BW-0829 August 29, 2006 Pond 431
BW-112905 November 29, 2005 Pond 431
BW-52506 May 25, 2006 Pond 431
17-0726 July 26, 2006 Creek 1121
17-080707 August 7, 2007 Creek 1121
95-0816 August 16, 2006 Pond l 121
98-0623 June 23, 2006 Pond 1121
CS-080707 August 7, 2007 Outfall 1121
HHS-080707 August 7, 2007 Pond 1
HR-0425 April 25, 2007 Pond 1121
HR-0623 June 23, 2006 Pond 1121
HR-0710 July 10, 2006 Pond 1 121
HR-080707 August 7, 2007 Pond 1121
ERC-071907 July 19, 2007 Runoff 1449
PL-031407 March 14, 2007 Pond 1449
PL-040407 April 4, 2007 Pond 1449
PL-052507 May 25, 2007 Pond 1449

 

245

Table B-2 (cont’d)

 

 

Sample ID Date Sampled Sample Source Landuse ID
PL-071907 July 19, 2007 Pond 1449
PL-52506 May 25, 2006 Pond 1449
PL—0307 March 7, 2006 Pond 1449
PL-0622 June 22, 2005 Pond 1449
PL-0628 June 28, 2006 Pond 1449
PL-0710 July 10, 2006 Pond 1449
PL—0711 July 11, 2006 Pond 1449
PL—0729-T July 29, 2005 Pond 1449
PL—1031 October 31, 2006 Pond 1449
PL-112905 November 29, 2005 Pond 1449
AG-031407 March 14, 2007 Pooling 211 1
AG—040407 April 4, 2007 Pooling 21 1 1
AG-0829 August 29, 2006 Pooling 2111
AG-52506 May 25, 2006 Pooling 2111
AG-0307 March 7, 2006 Pooling 2111
AG-0622 June 22, 2005 Pooling 2111
AG-112905 November 29, 2005 Pooling 2111
08-0622 June 22, 2006 Creek 2113
09-0622 June 22, 2006 Creek 2113
101-0711 July 11, 2006 Creek 2113
18-0726 July 26, 2006 Creek 2231
18-080707 August 7, 2007 Creek 2231
28-0726 July 26, 2006 Creek 2231
28-080707 August 7, 2007 Creek 2231
OF38-1107-1 October 30, 2006 Outfall 12644
OF38-1107-2 November 7, 2006 Outfall 12644
OF38-1107-3 November 7, 2006 Outfall 12644
OF38-1107-4 November 7, 2006 Outfall 12644
OF38-1107-5 November 7, 2006 Outfall 12644
OF 3 8-1 107-6 November 7, 2006 Outfall 12644
OF38-1107-7 November 7, 2006 Outfall 12644
OF38-1107-8 November 7, 2006 Outfall 12644
OF38-1107-9 November 7, 2006 Outfall 12644
OF38-0425-1 April 25, 2007 Outfall 12644
OF38-0425-2 April 25, 2007 Outfall 12644
OF38-0628 June 28, 2006 Outfall 12644
OF38-0724-12 July 24, 2006 Outfall 12644
OF 38-0725 July 25, 2006 Outfall 12644
OF 38-0725-5 July 25, 2006 Outfall 12644

 

246

Table B-2 (cont’cy

 

 

Sample ID Date Sampled Sample Source Landuse ID
OF 3 8-0725-A1 July 25, 2006 Outfall 12644
OF -22-0726 July 26, 2006 Outfall 12644
OF22-0803-01 August 3, 2006 Outfall 12644
OF 37-0726 July 26, 2006 Outfall 12644
OF38-0802- 14 August 2, 2006 Outfall 12644
OF38-0803-01 August 3, 2006 Outfall 12644
OF3 8-0803-A1 August 3, 2006 Outfall 12644
OF3 8-0803-A2 August 3, 2006 Outfall 12644
OF3 8-1 130-1 November 30, 2006 Outfall 12644
OF 3 8-1 1 30-10 November 30, 2006 Outfall 12644
OF38-1130-11 November 30, 2006 Outfall 12644
OF38-1130-12 November 30, 2006 Outfall 12644
OF 3 8-1 1 30-13 November 30, 2006 Outfall 12644
OF 3 8-1 130-13 November 30, 2006 Outfall 12644
OF 3 8-1 130-15 November 30, 2006 Outfall 12644
OF 3 8-1 130-16 November 30, 2006 Outfall 12644
OF38-1130-17 November 30, 2006 Outfall 12644
OF38-1130-2 November 30, 2006 Outfall 12644
OF 3 8-1 1 30-4 November 30, 2006 Outfall 12644
OF3 8-1 130-5 November 30, 2006 Outfall 12644
OF3 8-1 130-9 November 30, 2006 Outfall 12644
OF 3 8-1 1 30-A10 November 30, 2006 Outfall 12644
OF38-1 130-A1 1 November 30, 2006 Outfall 12644
OF 3 8-1 1 30-A9 November 30, 2006 Outfall 12644
OF 3 8- 1 201-12-D December 1, 2006 Outfall 12644
OF 3 8-1201 -14 December 1, 2006 Outfall 12644
OF 3 8-1201- 1 6 December 1, 2006 Outfall 12644
OF38-1201-17 December 1, 2006 Outfall 12644
OF38-1201-19 December 1, 2006 Outfall 12644
OF38-1201-21 December 1, 2006 Outfall 12644
OF38-1201-22 December 1, 2006 Outfall 12644
OF38-1201-23 December 1, 2006 Outfall 12644
OF38-1201-24 December 1, 2006 Outfall 12644
OF38- 1206 December 6, 2006 Outfall 12644
OF38-1208-2 December 8, 2006 Outfall 12644
OF 3 8-1208-3 December 8, 2006 Outfall 12644
OF38-1212-1 December 12, 2006 Outfall 12644
OF 3 8-12 1 2-2 December 12, 2006 Outfall 12644
OF38-619-0 June 20, 2007 Outfall 12644

 

247

Table B-2 (cont’d)

 

 

Sample ID Date Sampled Sample Source Landuse ID
OF38-619-1 June 19, 2007 Outfall 12644
OF38-619-13 June 20, 2007 Outfall 12644
OF38-619-14 June 20, 2007 Outfall 12644
OF38-619-15 June 19, 2007 Outfall 12644
OF38-619-16 June 19, 2007 Outfall 12644
OF38-619-17 June 19, 2007 Outfall 12644
OF38-619-18 June 19, 2007 Outfall 12644
OF38-619-19 June 19, 2007 Outfall 12644
OF 38-619-2 June 19, 2007 Outfall 12644
OF 38-619-20 June 19, 2007 Outfall 12644
OF38-619-25 June 19, 2007 Outfall 12644
OF38-619-5 June 19, 2007 Outfall ' 12644
OF38-619-6 June 19, 2007 Outfall 12644
OF38-6l9-8 June 20, 2007 Outfall 12644
OF38-619-9 June 19, 2007 Outfall 12644
OF52-0726 July 26, 2006 Outfall 12644
OF53-0726 July 26, 2006 Outfall 12644
88-080707 August 7, 2007 Pond 19331
EE-080707 August 7, 2007 Pond 19331
EE-0816 August 16, 2006 Pond 19331
GC-031407 March 14, 2007 Pond 19331
GC-040407 April 4, 2007 Pond 19331
GC-052507 May 25, 2007 Pond 19331
GC-071907 July 19, 2007 Pond 19331
GC-52506 May 25, 2006 Pond 19331
GC-0307 March 7, 2006 Pond 19331
GC-0622 June 22, 2005 Pond 19331
GC-0628 June 28, 2006 Pond 19331
GC-0710 July 10, 2006 Pond 19331
GC-0729-T July 29, 2005 Pond 19331
GC-0816 August 16, 2006 Pond 19331
GC-1031 October 31, 2006 Pond 19331
GC-112905 November 29, 2005 Pond 19331
CR-0615 June 15, 2006 River mixed
CSO-0615 June 15, 2006 River mixed
DC-O615 June 15, 2006 River mixed
DLM-0615 June 15, 2006 River mixed
FL-0307 March 7, 2006 River mixed
FL-0628 June 28, 2006 River mixed

 

248

Table B-2 (cont’d)

 

 

Sample ID Date Sampled Sample Source Landuse ID
GR-0615 June 15, 2006 River mixed
GRER-0711 July 11, 2006 River mixed
GRFU-0615 June 15, 2006 River mixed
GRGR-0808 August 8, 2006 River mixed
GRIO-0808 August 8, 2006 River mixed
GRJA-0711 July 11, 2006 River mixed
GRLA-0808 August 8, 2006 River mixed
GRPO-0808 August 8, 2006 River mixed
H131-0615 June 15, 2006 River mixed
LGEA-0808 August 8, 2006 River mixed
RCMSU-0808 August 8, 2006 River mixed
RCR 0404 April 4, 2006 River mixed
RCR-0226 February 26, 2007 River mixed
RCR-1005 October 5, 2006 River mixed
RCWI-0808 August 8, 2006 River mixed
RRRO-0808 August 8, 2006 River mixed
U28-0615 June 15, 2006 River mixed

 

249

APPENDIX C

DATA

250

Table C-l. Molecular characterization data.

 

 

Average Average
Sample ID quA -1 Molecular Molecular Polydispersity
(L gC cm ) Number Weight (Da)
(Da)
08-0622 2.06 1349.2 1571.9 1.17
09-0622 1.86 1212.8 1525.3 1.26
100-0710 1.99 921.0 1027.0 1.12
100-0711 2.11 684.5 1151.7 1.68
100-080707 2.22
101-0711 1.72 1171.9 1412.6 1.21
14-0726
17-0726
17-080707 1.96
18-0726
18-080707 1.63 1241.8 1319.1 1.06
28-0726
28-080707 1.90 1266.5 1336.0 1.05
37-0425 2.67
64-0623 1.44
84-0425 2.65
87-0425 2.24
87-0623 1.76 1171.3 1266.9 1.08
87—07 11 2.25 981.9 1292.4 1.32
88-080707 0.92
95-0816 1.19 974.1 1094.7 1.12
96-0425 2.15
96-0622 2.05 1129.8 1509.1 1.34
98-0623 1.36 1150.2 1334.1 1.16
99-0816 0.81
AG-0307
AG-031407
AG-040407 1279.2 1330.3 1.04
AG-O622
AG-0829 2.90 754.3 1927.4 2.56
AG-l 12905 3.40
AG-52506 2.25 681.2 1106.3 1.62
BW-040407 3.21 1562.5 2168.1 1.39
BW-052507 3.18 665.6 1952.6 2.93

 

251

Table C-l. (cont’d)

 

 

Average Average
Sample ID NU.1VA -1 Molecular Molecular Polydispersity
(L gC cm ) Number Weight (Da)
(Da)
BW-0622
BW-O729-T 4.49
BW-0829 3.17 852.5 2151.9 2.52
BW-1031 2.39
BW-112905
BW-l 12905 2.47
BW-52506 2.36 1160.4 1673.1 1.44
BW-52506 3.43 1243.1 2375.4 1.91
CR-0615 1.874 1201.0 1528.1 1.27
CS-080707 2.28 974.8 1119.4 1.15
CSO—061 5 1.622 607.6 1549.3 2.55
DC-0615 1.717 1400.5 1552.9 1.11
DLM-0615 11.410 1241.3 1557.7 1.25
EE-080707 1.58 1241.6 1348.8 1.09
EE-0816 1.77 1105.6 1323.0 1.20
ERC-071907 1.588 324.0 776.0 2.40
FL-0307
F L-O628 1.905 1026.8 1262.4 1.23
GC-0307
GC-031407
GC-040407
GC-052507 2.60 560.2 1425.4 2.54
GC-0622
GC-0628 2.38 846.9 1465.5 1.73
GC-0710 2.35 1097.9 1442.5 1.31
GC-07 1907 2.36
GC-0729-T 3.43
GC-0816 2.56 1075.5 1548.0 1.44
GC-1031 2.60
GC-l 12905 2.92
GC-52506 1.97 1065.8 1494.0 1.40
GR-0615
GRER-0711 2.057 1314.3 1628.6 1.24
GRFU-0615 1.882 1486.3 1604.7 1.08

 

252

Table C-l. (cont’d)

 

 

Average Average
Sample ID NU.1VA -1 Molecular Molecular Polydispersity
(L gC cm ) Number Weight (Ba)
4113)
GRGR-0808 1 171.4 1266.5 1.08
GRID-0808 1077.9 1567.9 1.45
GRJA-0711 1.886 1245.0 1428.7 1.15
GRLA-0808 1.927 809.2 1139.7 1.41
GRPO-0808 836.1 1188.1 1.42
H131-0615 1.850 1490.8 1603.4 1.08
HHS-080707 1.46 1176.6 1290.0 1.10
HR-O425 2.51
HR-0623 1.69
HR-0710 2.27 767.7 1143.0 1.49
HR-080707 2.32
LGEA-0808 2.375 1000.8 1616.5 1.62
OF-22-0726 2.08 651.7 940.7 1.44
OF 22-0803-01 585.3 852.6 1.46
OF37-0726 2.43
OF38-0425-1 2.25
OF38-0425-2 1.36
OF38-0628 1.07 942.3 1000.5 1.06
OF38-0724-12 1.30 863.8 1027.8 1.19
OF38-0725 1.54 607.3 914.9 1.51
OF38-O725-5
OF38-O725-A1 1.65 545.7 804.0 1.47
OF38-0802-14 2.15
OF38-0803-01 1.81
OF38-0803-A1
OF38-0803-A2 1.83 1023.3 1209.2 1.18
OF38-1107-1
OF38-1107-2
OF38-1107-3
OF38-1107-4
OF 38-1 107-5
OF38-1107-6
OF 38-1 107-7
OF38-1107-8

 

253

Table C-l. (cont’d)

 

 

Average Average
Sample ID NU.1VA -1 Molecular Molecular Polydispersity
(L gC cm ) Number Weight (Da)
(Da)
OF38-1107-9
OF38-1130-1 0.96 893.5 1267.7 1.42
OF38-1130-10 1.42
OF38-1130-11 1.39 793.7 1290.6 1.63
OF38-1130-12
OF38-1130-13 0.89 713.2 1216.7 1.71
OF38-1130-13 1.57
OF38-1130-15 1.30 806.2 1956.2 2.43
OF38-1130-16 1.70 1209.7 1716.6 1.42
OF38-1130-17 1.29
OF38-1130-2 1.05 902.8 1268.7 1.41
OF38-1130-4 1.54 1001.6 1546.0 1.54
OF38-1130-5 1.33 683.4 1129.0 1.65
OF38-1130-9 1.27 770.0 1282.2 1.67
OF38-1130-A10 1.51 981.8 1543.8 1.57
OF38-1130-A11 1.57 1082.3 1604.8 1.48
OF38-1130-A9
OF38-1201-12-D
OF38-1201-14
OF38-1201-16 1315.5 1740.0 1.32
OF38-1201-17 1330.2 2112.8 1.59
OF38-1201-19 1126.5 1652.8 1.47
OF38-1201-21 1111.5 1672.8 1.50
OF38-1201-22 1016.0 1631.8 1.61
OF38-1201-23
OF38-1201-24 1159.1 1736.8 1.50
OF38-1206
OF38-1208-2
OF38-1208-3
OF38-1212-1
OF38-1212-2
OF 3 8-6 1 9-0 1.04
OF38-619-1 1.40 666.8 860.5 1.29
OF38-619-13 1.62 495.6 763.5 1.54

 

254

Table C-l. (cont’d)

 

Average

 

Avera e
Sample ID NU.1VA -1 Molecular Molecular Polydispersity
(L gC cm ) Number Weight (Da)
0):!)
OF38-619-14 1.50 466.4 663 .3 1.42
OF38-619-15 1.38 432.6 682.4 1.58
OF38-619-16 1.45 441.7 622.0 1.41
OF38-619-17 1.44 421.2 687.2 1.63
OF38-619-18 1.43 591.2 769.3 1.30
OF38-619—19 1.42
OF38-619-2 1.39 563.7 996.4 1.77
OF38-619—20 1.35 306.1 860.3 2.81
OF38-619-25 1.41 381.4 749.8 1.97
OF 38-61 9-5 1.76 665.6 986.0 1.48
OF38-619-6 2.16 598.4 848.3 1.42
OF38-619-8 2.24 958.6 963.8 1.01
OF38-619-9 2.07
OF52-0726 2.55 688.7 1145.3 1.66
OF53-O726 2.21 902.3 1244.1 1.38
PL-0307
PL-031407
PL-040407
PL-052507 2.069 346.4 849.8 2.45
PL-O622
PL-0628 6.918 1090.9 1104.0 1.01
PL-0710 1.941 905.5 1009.3 1.11
PL-0711 1.789 1023.2 1223.9 1.20
PL-071907 1.808 400.2 846.4 2.1 1
PL-0729-T 1.636
PL-1031 2.104
PL-l 12905 3.087
PL-52506 2.117 838.8 1036.8 1.24
RCMSU-0808 890.0 1416.6 1.59
RCR 0404
RCR-0226
RCR-1005
RCWI-0808 2.337 913.3 1315.2 1.44

 

255

Table C-l. (cont’d)

 

Average

 

Avera e
Sample ID NU.1VA -1 Molecular Molecular Polydispersity
(L gC cm ) Number Weight (Da)
(Da)
RRRO-0808 1348.5 1365.5 1.01
SW-0829 2.48 938.8 1680.8 1.79
U28-0615 1.896 1431.8 1589.2 1.11

 

256

Table C-2. DOC fraction data: total carbon in sample (CT), anion-F eluent (i.e. carbon
not retained by cartridge) (CF), anion-I eluent (C1), extended hydrophobic e1uent(CHi),
hydrophobic eleuent (CH0), H-bonding eluent (CH), and anion-lkDa eluent (Cx) (ND -

non-detect).

 

 

Sample ID CT CF C] Cm CH0 CH Cx
08-0622 12.21 10.23 9.03 5.09 3.36 0.34 9.26
09-0622 14.22 12.52 11.29 2.81 3.28 0.24 11.42
100-0710 29.50 24.81 27.04 13.05 11.50 1.87 24.49
100-0711 19.36 15.09 14.17 9.03 7.87 2.15 14.97
100-080707 12.43 11.10 10.81 6.76 6.04 2.40 10.68
101-0711 11.65 10.91 9.60 3.97 2.64 0.54 10.24
14-0726 13.76 12.97 11.70 3.80 2.94 0.84 12.49
17-0726 7.66 7.02 6.29 2.53 2.26 0.77 6.32
17-080707 12.52 11.36 10.79 6.57 5.77 1.64 10.79
18-0726 7.75 6.57 5.90 2.71 2.51 0.94 6.00
18-080707 9.27 8.52 7.68 2.74 2.21 0.62 8.44
28-0726 11.08 10.37 9.51 2.90 2.89 0.88 9.74
28-080707 9.45 8.33 7.47 3.04 2.72 1.14 7.77
37-0425 2.34 1.81 1.86 1.50 1.30 1.19 1.33
64-0623 12.91 10.31 9.60 3.02 2.96 0.67 9.08
84-0425 6.05
87-0425 3.12 1.56 1.41 1.59 1.40 0.50 1.14
87-0623 7.09 3.78 6.97 3.10 2.34 0.20 3.41
87-0711 7.48 5.98 5.11 3.71 3.47 1.31 5.39
88-080707 8.80 7.53 6.89 1.88 1.46 ND 6.85
95-0816 7.87 6.95 6.26 2.23 1.74 0.36 6.02
96-0425 12.49
96-0622 12.13
98-0623 15.79 11.59 11.65 4.40 3.88 0.54 10.97
99-0816 5.59 4.51 4.01 1.48 1.18 ND 3.78
AG-0307 9.68 5.92 9.29 3.47 2.84
AG-031407 9.84 5.47 8.46 ND 2.41 ND 5.81
AG-040407
AG-0622 19.26 15.97 2.27 1.28
AG-0829 24.87 23.86 23.54 13.71 7.89 1.83 22.89
AG-112905 7.14 2.05 1.27
AG-52506 10.31 3.29 4.28 2.76 9.64 6.68 7.66
BW-040407 32.57 29.29 29.57 11.61 9.93 15.97 17.14
BW-052507 43.64 37.75 9.71 6.38 8.43 8.39

 

257

Table C-2. (cont’d)

 

sample ID CT CF C] Cm CH0 CH Cx
BW-0622 19.83 16.84 3.09 2.37
BW—0729-T 25.65 22.10 18.85 3.74 2.66
BW-0829 22.33 20.97 20.43 5.83 4.49 6.63 11.57
BW-1031 66.61 59.18 47.84 28.77 20.52 23.64 40.20
BW-112905
BW-112905 26.53 17.05 12.58
BW-52506 16.12 14.14 13.86 4.09 2.98 0.02 14.39
BW-52506 28.91 26.25 25.78 6.40 4.49 4.94 13.15
CR-0615 9.50 7.76 2.26 7.05 ND 7.09
CS-080707 13.83 12.22 11.62 7.29 6.37 2.79 11.78
CSO-0615 10.79 8.25 3.26 2.81 ND 7.07
DC-0615 9.84 7.76 6.77 2.44 1.91 ND 6.79
DLM-061 5 1 .56 ND ND ND ND ND ND
EE-080707 l 1.42 10.40 9.29 2.37 1.91 0.00 9.94
EE-0816 12.93 12.19 11.23 3.22 2.44 0.19 11.56
ERC-071907 23.74 18.94 17.88 11.04 9.52 0.75 18.43
FL-0307 9.34 7.04 7.92 1.67 1.82
FL-0628 11.39 9.83 ND 2.53 8.20 ND 9.08
GC-0307 15.72 13.28 13.83 3.60 2.68
GC-031407 12.87 12.22 11.21 4.01 3.32 1.70 11.58
GC-040407
GC-052507 29.12 27.93 24.21 5.43 5.00 0.26 27.24
GC-0622 19.83 17.69 4.10 2.43
GC-0628 23.61 18.86 20.02 6.30 5.68 9.39 20.27
GC-0710 19.80 18.60 17.30 4.94 4.10 2.15 17.94
GC-071907 30.21 29.16 26.57 6.44 6.08 1.20 28.30
GC-0729-T 22.97 20.22 22.16 5.41 3.38
GC-0816 27.40 26.25 24.59 7.15 5.52 1.04 25.10
GC-103l 29.35 28.35 26.84 6.57 7.77 1.60 27.76
GC-112905 36.58 7.57 3.50
GC-52506 23.25 21.59 20.43 4.68 3.90 0.53 21.34
GR-O615
GRER-0711 13.42 12.87 11.88 3.07 2.77 0.79 12.20
GRFU-0615 9.67 9.10 7.05 2.24 2.27 ND 7.10

 

258

Table C-2. (cont’d)

 

 

Sample ID CT CF C1 Cm CH0 CH Cx
GRGR-0808
GRIO-0808
GRJA-0711 10.81 10.12 9.34 2.51 1.75 0.24 9.61
GRLA-0808 9.34 8.67 7.99 2.48 2.22 0.46 8.11
GRPO-0808
H131-0615 9.89 8.17 7.24 2.41 2.14 0.03 7.01
HHS-080707 11.04 9.93 9.17 ND 2.08 0.48 9.42
HR-0425 4.30 3.44 3.77 2.31 2.42 1.28 3.33
HR-0623 13.92 11.51 9.81 6.10 4.76 1.15 10.59
HR-0710 33.75 30.24 29.13 16.15 14.19 4.55 29.96
HR-080707 10.93 10.12 9.96 5.96 5.25 2.90 8.60
LGEA-0808 20.92 20.01 18.57 4.23 0.62 19.59
OF-22-0726 6.98 5.65 5.41 ND 2.89 0.90 5.13
OF22-0803-01
OF37-0726 7.16 5.76 5.20 3.21 2.66 0.75 5.17
OF 38-0425-1 5.33
OF3 8-0425-2 7.92
OF38-0628 5.62 3.80 5.55 1.83 1.76 0.15 3.31
OF38-0724-12 4.75 4.16 4.40 1.55 1.29 ND 3.57
OF38-0725 9.02 7.85 8.29 3.93 3.68 0.42 7.33
OF38-0725-5 0.08 0.02 0.03 0.01 ND ND ND
OF38-0725-A1 10.08 8.58 9.00 4.62 4.24 0.06 7.95
OF 38-0802-14
OF38-0803-01
OF38-0803-A1
OF 38-0803-A2
OF38-1107-1 5.31 4.72 4.73 1.49 1.33 0.12 3.98
OF38-1 107-2 5.42 4.59 5.00 1.40 1.42 ND 3.81
OF38-1107-3 13.45 9.43 7.77 3.62 2.99 ND 9.10
OF38-1107-4 16.03 11.71 10.43 6.18 5.80 1.52 10.92
OF38-1107-5 12.24 9.23 5.21 4.49 3.84 0.45 8.32
OF38-1107-6 11.06 8.15 5.89 5.24 4.70 1.70 7.87
OF38-1107-7 12.90 8.70 8.15 6.77 5.61 0.95 8.13
OF38-1107-8 12.18 8.06 7.10 6.04 5.52 1.40 7.81

 

259

Table C-2. (cont’d)

 

SampleID CT CF C1 Cm CH0 CH Cx

 

OF38-l 107-9 5.89 4.79 4.53 3.53 2.97 ND 3.91
OF38-l 130-1 5.72 4.82 5.06 2.11 1.77 ND 4.46
OF38-1130-10 4.09 3.45 3.40 2.00 1.77 0.26 2.85
OF38-1130-11 6.89 5.48 5.18 3.62 3.31 1.03 4.86
OF38-1130-12 0.08

OF38-1130-13 6.04 4.82 4.72 3.13 2.82 1.00 4.27
OF38-1 l30-13

OF38-l l30-15 6.82 5.28 5.01 3.54 3.23 0.76 4.60
OF38-l l30-16 4.42 3.73 3.70 1.99 1.79 0.90 3.21
OF38-1130-l7 6.91 5.58 5.15 3.53 3.20 0.77 4.86

OF38-1130-2 5.22 4.39 4.64 1.61 1.22 ND 3.82
OF38-11304 4.92 4.27 4.22 2.47 2.40 0.32 3.72
OF38-1130-5 6.93 5.56 5.28 3.59 3.30 0.86 4.89
OF38-1130-9 5.28 4.38 4.35 2.82 2.59 0.40 3.88

OF38-ll30-A10 5.81 4.95 4.86 2.85 2.58 0.82 4.32
OF38-1130-All 5.49 4.80 4.67 2.73 2.42 0.81 4.25
OF38-l l30-A9 5.72

OF38-1201-l2-D

OF38-1201-l4

OF38-1201-16

OF38-1201-17

OF38-1201-19

OF38-1201-21

OF38-1201-22

OF38-1201-23

OF38-l201-24

OF38-1206 5.81

OF38-1208-2 7.22

OF38-1208-3 4.44 3.82 3.82 0.02 0.68 ND 3.17
OF38-1212-1 8.16

OF38-1212-2 3.81

OF38-619-0 4.50 3.94 3.76 1.28 1.20 0.20 3.37
OF38-6l9-l 8.06 6.90 7.02 2.79 2.50 0.34 6.47
OF38-619-l3 8.98 7.78 5.11 5.02 0.96

 

260

Table C-2. (cont’d)

 

 

Sample ID CT CF C1 Cm CH0 CH Cx
OF38-619-14 10.66 9.28 9.06 6.77 6.24 1.37 8.58
OF38-619-15 12.77 11.15 10.89 8.05 7.08 1.59 10.61
OF38-619-16 12.57 11.07 11.08 8.10 7.53 1.63 10.73
OF38-619-17 12.70 11.33 11.11 8.03 7.44 1.58 10.86
OF38-619-18 11.58 10.22 9.99 7.41 6.56 1.52 9.91
OF38-619-19 12.99 11.59 11.27 8.31 7.43 1.55 11.31
OF38-619-2 17.49 11.38 10.49 8.30 8.05 3.57 10.93
OF38-619-20 13.30 11.66 11.59 8.39 7.73 1.74 11.18
OF38-619-25 12.80 11.41 11.30 8.33 7.52 1.57 11.01
OF38-619-5 9.03 6.47 6.17 3.83 3.48 0.93 5.96
OF38-619-6 6.65 4.71 4.69 3.17 2.95 1.10 4.38
OF38-619-8 5.90 4.29 3.34 3.10
OF38-619-9 4.88 3.58 3.58 2.64 2.26 1.04 3.04
OF 52-0726 10.27 8.65 8.17 4.45 4.07 1.43 7.90
OF53-0726 8.24 7.49 6.96 2.15 1.81 0.22 6.81
PL-0307 5.45 3.05 5.43 2.19 2.12
PL-031407 4.00 3.36 3.19 2.03 1.63 0.85 2.81
PL-040407
PL-052507 21.65 17.09 15.52 10.12 8.76 0.64 16.55
PL-0622 11.54 9.55 4.98 4.66
PL-0628 3.83 ND ND 2.08 ND ND ND
PL-0710 31.22 26.54 25.35 15.53 14.02 3.85 25.80
PL-0711 7.49 5.32 5.36 3.86 2.98 1.46 4.97
PL-071907 13.50 12.36 11.75 5.23 4.74 0.62 11.72
PL-0729-T 9.72 6.30 6.28 2.96 2.01
PL-103l 6.37 5.24 5.09 2.34 2.02 0.38 4.74
PL-112905 5.36 2.85
PL-52506 8.79 6.67 6.54 3.44 3.16 ND 6.39
RCMSU-0808
RCR 0404 11.59 10.89 9.86 3.09 2.67 10.09
RCR-0226
RCR-1005 9.19 8.67 7.78 2.47 2.11 0.11 8.06
RCWI-0808 9.20 8.67 7.63 2.09 1.96 0.20 7.84

 

261

Table C-2. (cont’d)

 

SampleID CT CF C1 Cm CH0 CH Cx

 

RRRO-0808
SW-0829 13.88 13.20 12.02 4.11 3.38 1.28 12.53
U28-0615 9.60 8.16 7.06 0.36 1.56 ND 7.12

 

262

   

u11111111111111111tht