| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a U.S. Geological Survey, 6000 J Street, Sacramento, CA 95819
b International Water Management Institute, P.O. Box 2075, Colombo, Sri Lanka
* Corresponding author (jafleck{at}usgs.gov).
Received for publication June 10, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: DBP, disinfection by-product • DOC, dissolved organic carbon • HAA, haloacetic acid • NMR, nuclear magnetic resonance • STHMFP, specific trihalomethane formation potential • SUVA, specific ultraviolet absorption • THM, trihalomethane • TTHMFP, total trihalomethane formation potential
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Land subsidence is a major issue of concern in the delta because previous management of the peat islands has caused some of their land-surface elevations to subside as much as 6 m below sea level, resulting in serious threats to levee stability (Weir, 1950; Rojstaczer et al., 1991). Current-day land subsidence in the delta (1–3 cm yr–1) primarily is attributed to microbial oxidation and decomposition of the peat soils (Weir, 1950; Prokopovich, 1985; Deverel and Rojstaczer, 1996; Deverel et al., 1998). One strategy of subsidence mitigation focuses on restoring wetlands on previously farmed peat soils, thereby causing a fundamental change in the biogeochemical cycling of carbon on delta islands. Year-round shallow flooding of peat soils to develop freshwater wetlands has resulted in a net carbon gain in the system, mitigating and possibly reversing the effects of subsidence (Deverel et al., 1998).
Land-use change from agriculture to freshwater wetlands on delta islands may have large effects on DOC quantity and quality in drainage water. Deverel and Rojstaczer (1996) indicated that persistent flooding of shallow oxidized peat soils resulted in high concentrations of DOC (48–110 mg L–1) in drainage water. These results suggest that flooding of shallow, peat soils to form wetlands may contribute to increased DOC concentrations and loads in drainage water. Natural organic matter in the drinking source water reacts with chlorine, added as a disinfectant, to form DBPs, such as THMs and HAAs. Some of these DBPs are known carcinogens and are regulated in drinking water by the USEPA (2003). This issue is of particular concern in the delta because water diverted from the delta supplies drinking water to more than 22 million people.
Differences in carbon sources, decomposition rates and pathways, and carbon availability potentially have profound effects on the carbon forms that reach the delta channel waters and are diverted for drinking water. Agricultural drains already are considered a major source of THM precursors to the delta (Amy et al., 1990). Understanding and quantifying the effects that wetland restoration has on the biogeochemical and hydrologic processes that control DOC and DBP precursor production should provide water-resource managers the valuable knowledge to make informed decisions on how to best manage delta island land use.
The objective of this article is to discuss management and environmental factors controlling the biogeochemical processes responsible for DOC and DBP precursor release in peat soils. This discussion is founded upon results from the study's investigation as to whether the conversion of a heavily managed agricultural field to a wetland in 1997 has affected the DOC and DBP precursor release from the field and, potentially, into delta waterways. By comparing the potential release of DOC and DBP precursors from this restored wetland to nearby agricultural fields, the overriding biogeochemical factors that lead to DBPs in drinking water from the delta were determined in a natural field setting.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Soils were sampled from four monitoring sites within each of the two agricultural fields. An intact core sampler that allows discrete sectioning of soil cores was used to obtain peat soil profile samples for description of vertical zonation of soil properties. The wetland sites were sampled simultaneously at three monitoring sites within each habitat. Four replicate samples were collected at each monitoring site of both of the agricultural fields and in the wetland. Cores were kept inside acrylic liners, capped, and placed on ice during transport to the lab where they were kept at 4°C until extrusion. Sampling was done in August 2000, October 2000, January 2001, March–April 2001, and August 2001 to capture major seasonal variability throughout the year.
Laboratory Procedures
Individual cores from the agricultural fields were extruded and sectioned into four depths (top, 0–5 cm; transition, 5–15 cm; bottom, 15–30 cm; and subsoil, 30–60 cm). Rather than the strict depth increment used in the agricultural fields, wetland cores were separated into three horizons on the basis of observable features: a surface sample of loosely consolidated sediment, a transitional horizon that was more consolidated and extended to approximately 15 cm, and a bottom horizon that extended down to 30 cm. Soil samples of each depth increment at each monitoring site were sectioned and replicate samples were composited and stored as a stock sample (i.e., Ag1:2, 0–5 cm) in the lab at <4°C.
Subsamples were collected from the stock samples, weighed, and dried in an oven at 60°C. The dried samples were weighed to determine moisture content and homogenized using a mini ball mill (Model 8000; SPEX, Edison, NJ) in preparation for total soil carbon and nitrogen analysis. Elemental C and N analyses were performed using a CHN analyzer (Model 2400 Series II; PerkinElmer, Wellesley, MA).
Dissolved organic C quantity and quality were measured by extracting 20 g (dry weight) of stock sample soil within 7 d of collection. The extraction included adding 100 mL of organic-free water to the sample, shaking vigorously for 15 min, centrifuging the mixture at 10000 rpm (14500 RCF) for 8 min, and gravity-filtering the supernatant through a 2.7-, 1.6-, and 0.2-µm filter stack. The extraction was performed twice on each sample and composited to obtain 200 mL of filtrate sample. The filtrate was analyzed for total DOC using a Shimadzu (Kyoto, Japan) TOC-5000A total organic carbon analyzer (Fram et al., 2002). Ultraviolet (UV) adsorption at 254 nm (UV254) was analyzed on the filtrates using a spectrophotometer (Model UV/Vis Lambda 3B; PerkinElmer). From these data, carbon quality was assessed using specific ultraviolet absorbance (SUVA, L mg–1 C cm–1), calculated by normalizing UV254 to DOC (UV254/DOC). Higher SUVA values reflect a higher aromatic content of the DOC (Fujii et al., 1998). Trihalomethane formation potential (THMFP) was measured using a modified version of USEPA Method 502.2 (Fram et al., 2002). Specific trihalomethane formation potential (STHMFP, mmol mol–1) was calculated by normalizing THMFP to DOC on a molar basis (Fram et al., 2002). A subset of samples also was analyzed for haloacetic acid formation potential (HAAFP) using USEPA method 552.2. Specific haloacetic acid formation potential (mmol mol–1) was calculated by normalizing HAAFP to DOC on a molar basis.
Microbially available or labile C, another indicator of C quality, was quantified by analyzing CO2 production from soils under ambient air with a 2:1 ratio of water to dry soil (on a mass basis). The soils were placed in airtight glass jars and incubated at 20°C for 45 d, and the vessels were purged, sampled, and analyzed every 2 to 3 d. Initial incubations proceeded for 110 d to establish time periods for changes in C pools. It was determined that the most labile C was expelled in the first 10 d and a secondary labile pool was released between 10 and 40 d. Therefore, available C was defined as the quantity released during the entire 40 d. The jars were sampled in intervals of 2 to 3 d to reduce excessive buildup of gas in the headspace. Carbon dioxide production was measured by analyzing samples of headspace gas on a gas chromatograph (Model HP5890; Hewlett-Packard, Palo Alto, CA.).
Three soil samples that released DOC with considerably different STHMFP were analyzed using solid-state cross polarization magic angle spin nuclear magnetic resonance (CPMAS NMR) using methods similar to Fujii et al. (1998). Using CPMAS NMR analysis, we compare the relative quantities of carbon functional groups in the organic matter of the soils to assist in characterizing carbon compositions in the soil and to provide some insight to the origins of the THM precursors. The carbon functional groups are defined similarly by Fujii et al. (1998) as follows: (i) Aliphatic I (0–60 ppm), primarily sp3 hybridized carbons bonded to other carbons, carbons bonded to nitrogen and sulfur also can occur in this region, methoxyl carbons occur in the region; (ii) Aliphatic II (60–90 ppm), hetero-aliphatic carbons, primarily sp3 hybridized carbons bonded to oxygens, including ether, alcohol, and carbohydrate carbons; (iii) Anomeric (90–110 ppm), acetal and ketal carbons and anomeric carbons of carbohydrates; (iv) Aromatic (110–160 ppm), primarily aromatic and olefinic carbons; (v) Carboxyl (160–190 ppm), carboxylic acid carbons; ester, amide, and lactone carbons also can overlap with carboxyl; and (vi) Ketone (190–230 ppm), primarily ketones and aldehydes.
Data Analysis and Statistical Methods
Due to the complexity of the data collected, multiple scales of analysis were required. Data first were grouped and analyzed by site to assess large-scale trends and interactions caused by management effects. Within-site analyses (depth and season) were performed to identify environmental effects not associated with management practices. Field sampling variability was assessed using replicate samples collected during field sampling.
Data analysis by depth was based on preliminary results. Preliminary analysis indicated that comparing the plow and below-plow layers was most informative for investigating differences in THM formation potential in the agricultural sites. The Ag1 site showed no significant differences between depths for any variable, while the Ag2 site showed significant differences between all depths for some variables (p < 0.05, data not shown). Overall, differences between the two upper samples (0–5 and 5–15 cm) and the two lower samples (15–30 and 30–60 cm) were minor, while differences between the upper and lower soils were more pronounced. Therefore, for the agricultural sites, Ag1 and Ag2, data from the 0- to 5- and 5- to 15-cm soil depths were combined as "plow layer" and 15- to 30- and 30- to 60-cm depths were combined as "below-plow layer." The plow layer represents the zone with consistently greater aeration for the longest period of the year at both sites, while the below-plow layer has less consistent and more restricted aeration at various times. At the wetland sites, the important contrasts were observed between the loose, unconsolidated sediment layer and the underlying soils. Therefore, the "top" layer described in the methods was called "sediment" and the "transition" and "bottom" samples from the methods section were combined into a single wetland "soil" term. Aeration conditions in the wetland soil are considered restricted throughout the year, so differences are attributed to carbon inputs.
The basic design of the study is a split-plot, or nested, experiment investigating differences between sites, depths, and seasons. Statistical analyses were done using nonparametric methods. Differences between populations were determined using the Mann–Whitney rank sum test and Kruskal–Wallis analysis of variance on ranks; pair-wise comparisons were performed using Dunn's pair-wise comparison test. Regressions were performed using the Spearman rank order method (Devore and Peck, 1997).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Strong, positive correlations between SUVA and STHMFP existed within sites, except for the Ag1 site, which had a relatively limited range of values (Fig. 3 , Table 1). The correlations' slopes differed greatly between agricultural and wetland sites (Fig. 3, Table 1). A significant (p = 0.003) positive correlation also was found between SUVA and the HAA precursor content of the DOC (specific haloacetic acid formation potential, SHAAFP) for the Ag2 site, but there were no significant (p < 0.05) correlations for SUVA and SHAAFP for the other sites (Table 1). Specific trihalomethane formation potential was a good correlate for SHAAFP in the agricultural soil extracts; however, no correlation was observed between the two DBPs for the DOC released by the wetland soils (Fig. 4) . The SHAAFP was highly variable but still significantly higher than the STHMFP in the Wo site (Fig. 4). The Wv site also was highly variable but not significantly different than Ag1 or Ag2 (p < 0.05).
|
|
|
Whereas seasonal differences were minor compared with differences by depth and site, investigation of specific depths within sites revealed two important trends (Fig. 6) . First, for the Ag1 surface soil, the STHMFP of the extracted DOC reached its maximum in the winter and spring under flooded conditions and in the presence of crop residue, and its minimum in late summer when no residue remained and conditions were driest (Fig. 6a). The wetland sediment layer also showed a seasonal trend (Fig. 6b), with STHMFP greater in the summer and fall (warmer temperatures) than in winter (p < 0.05). The trend in the wetland sediment layer is important because, like the Ag1 site (Fig. 6a), the seasonal differences occur in concert with environmental factors. The timing of the seasonal differences in the wetland sediment indicate an interaction with seasonal carbon dynamics, particularly decomposition due to microbial activity and observed algal blooms, although no differences were seen in DOC concentrations between seasons for any of the sites.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The peat soils' decomposition states, driven by historical management, also dominate differences observed in DOC reactivity at different depths at the agricultural sites. Because the Ag1 and Ag2 soils were formed under similar conditions with similar carbon inputs, the differences in DOC release between the soils and their respective layers is probably a result of their varying extent of decomposition due to differences in drainage and oxidation. The relatively high STHMFP values for the Ag2 below-plow-layer soils (mean of 9.0 compared with 6.9 for plow layer) are probably a result of the relatively less decomposed state of the deeper horizon, which remains below the water table throughout the year and has maintained much of its original soil organic matter characteristics derived from the wetland conditions in which it was formed. In contrast, the Ag2 plow layer soil and the soils at the Ag1 site have undergone extensive oxidation due to artificial drainage, which has significantly altered the peat soil, creating a soil nearly unrecognizable to its original condition.
Identifying the dependence of DOC and DBP precursor release on historical management of the delta peat soils is important, but the question of what portion of the DOC constitutes the precursor pool remains. One tool available to address this question directly is NMR analysis. Numerous studies have reported changes in the carbon functional group composition during decomposition of drained peat soils using NMR (Hatcher et al., 1983; Hammond et al., 1985; Preston et al., 1987; Krosshavn et al., 1992; Zech et al., 1992; Baldock et al., 1997). In general, the NMR analyses reported in this study (Fig. 7) also support the compositional differences found elsewhere between well-decomposed and poorly decomposed peat soils. But, there is one overriding difference between the NMR results from this study and those in the literature; the loss of alkyl groups in the more oxidized soils. A defining feature of NMR spectra during decomposition involves the accumulation of alkyl groups, either through production of alkyl C from microbial decomposition or through "selective preservation" of the refractory alkyl sources such as lipids, waxes, and resins (Hatcher et al., 1983; Hammond et al., 1985; Preston et al., 1987; Krosshavn et al., 1992). The data presented here show lower amounts of alkyl groups (Aliphatic I, 0–62) in the more decomposed soils (Fig. 7), rather than an accumulation. This lower content of alkyl groups coincides with significantly lower STHMFP for the DOC from the more decomposed soils (Fig. 5e and 7), suggesting that alkyl functional groups may be related to production of DOC with propensity to form THMs, or precursor content. Although the NMR data from this study are very limited, the hypothesis that the content of Aliphatic I groups in the soil organic matter may be related to the content of THM precursors in pore water may help to explain the poor relation between DOC aromaticity (SUVA) and STHMFP in delta waters (Fujii et al., 1998; Fram et al., 1999). Whereas the authors concede that the data are limited, it merely is one piece of the puzzle that supports the hypothesis that a decomposable portion of the organic matter is a source of DBP precursors.
Continuing with our hypothesis and inferring from the NMR data, we can state that if aerobic decomposition of soil organic matter tends to degrade the forms of carbon that produce THMs in the agricultural plow layer soils, then DOC derived from less decomposed anaerobic wetland soils probably would have a high precursor content. Although historical management dominates total DOC and THM precursor release from the study's soils, the recent shift toward wetland conditions at the wetland sites already seems to show the effects wetland conditions have on the precursor content of the DOC (STHMFP) extracted from these soils. The STHMFP of the DOC extracted from the soils from the wetland sites already is significantly higher than the DOC extracted from either agricultural site (10.7 and 8.2 mmol mol–1 vs. 7.6 and 7.8 mmol mol–1, respectively; p < 0.05) after only 3 yr of inundation. The establishment of wetland conditions has shifted the primary soil organic matter decomposition pathway from aerobic toward anaerobic, similar to the original conditions under which the peat soil formed. This change in soil redox conditions has led to a shift in the microorganism population that is responsible for organic matter decomposition (D.A. Bossio, unpublished data, 2003), which has led to release of DOC with higher precursor content (STHMFP). The higher SUVA and STHMFP values for the wetland sites support the hypothesis put forth in a recent study that biogeochemical processes in constructed wetlands create and contribute a substantial amount of DOC that is aromatic and produces high amounts of THMs (Rostad et al., 2000). These results solidify the hypothesis that the organic matter decompositional environment strongly influences DOC quality and THM precursor release.
Comparing wetland habitats, the hypothesis that organic matter input quality is a determinant of DOC reactivity is demonstrated. This is illustrated in examining the slope of the correlations between SUVA and STHMFP for the wetland samples (Table 1). The steepness of the slope is driven strongly by the sediment layer at both sites, suggesting that source of the organic matter may be the dominant variable in STHMFP of the DOC extracted from soils within the wetland habitat. The high precursor content of the Wo sediments indicates that open-water wetland vegetation, such as submergent plants (Myriophyllum spp.), duckweed (Lemna spp.), and algae are implicated as sources of organic carbon that lead to production of high levels of THM precursors. In an investigation of the microbial communities at these same sites, a unique community was identified in the sediment layer of the Wo site, which had an unusually high organic matter turnover rate in the saturated environment (D.A. Bossio, unpublished data, 2003). The highly labile forms of carbon supplied by decomposition of organisms such as algae may account for the high turnover rates. These data suggest that the quality of the organic matter supply to the sediments under open-water wetland conditions significantly influence the quality of DOC (STHMFP) and may be as important as the decomposition conditions in the soil. This hypothesis is in agreement with another study where water from an upstream lake contributed DOC with an inordinate amount of THM precursors (STHMFP) to a downstream water treatment plant (Pomes et al., 2000). In that study, Pomes et al. determined that bacterial and algal contributions from the lake supplied only 33% of the DOC to the treated water but accounted for more than 70% of the THMs produced upon chlorination. The authors suggested that further research is needed to identify nonhumic (i.e., bacterial and algal) sources of THM precursors.
Seasonal trends in the data support the importance of organic matter source and decomposition environment in the production of THM precursors. With the removal of large-scale differences due to management, seasonal trends are apparent (Fig. 6). The wetland sediment layer shows a spike in STHMFP during the time of observed algal bloom in the pond while the agricultural fields' surface soils show an increase in STHMFP during decomposition of corn residue under wet conditions (Fig. 6). The importance of new organic matter and decomposition environment to THM precursor production may suggest that the labile carbon pool is the origin of the precursors; however, while available (labile) carbon measurements showed a strong seasonal trend, they were not related to the THM content of the DOC. The contribution from the seasonal production of THM precursors to the study exemplifies the importance of short-term influences on THM precursor content of DOC, despite the dominance of long-term factors in total THM precursor production. However, while seasonal trends were apparent within their specific environments, they were overshadowed by the larger-scale differences between managements.
Thus far we have discussed only THMs; however, other DBPs, such as HAAs, are of equal concern and are regulated as well. Whereas there is no basis for the belief that different DBPs would possess the same precursors and, thus, be derived from the same environmental conditions, it would be valuable for managers and policymakers to identify a relation between DBP formation potentials. A promising result moving toward this goal showed that the formation of THMs and HAAs is well correlated in water extracted from agricultural soils, indicating a similar precursor environment (Fig. 4). However, HAA precursors proved to be far more prevalent than THM precursors in wetland soils, particularly in the open water areas of the wetland (Fig. 4). The conditions in the open wetland seem to favor HAA production more than THM production. This implies that using THMFP measurements as a proxy for other DBPs may lead to erroneous conclusions. The variable nature of formation of different DBPs from differing carbon sources is well known (Reckhow and Singer, 1985; Larson and Weber, 1994). However, their behavior in the natural system remains unknown. In this study, THMs provide a minimum baseline for HAA production under different managements. It seems that a shift from agricultural to wetland conditions, especially open water environments, will lead to greater DBP formation in drinking water diverted from the delta. Haloacetic acids probably will be more elevated by a shift to wetlands than THMs.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  J. C. Chomycia, P. J. Hernes, T. Harter, and B. A. Bergamaschi Land Management Impacts on Dairy-Derived Dissolved Organic Carbon in Ground Water J. Environ. Qual., March 1, 2008; 37(2): 333 - 343. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. T. Chow, F. Guo, S. Gao, and R. S. Breuer Trihalomethane Reactivity of Water- and Sodium Hydroxide-Extractable Organic Carbon Fractions from Peat Soils J. Environ. Qual., January 3, 2006; 35(1): 114 - 121. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. T. Chow, F. Guo, S. Gao, and R. S. Breuer Trihalomethane Formation Potential of Filter Isolates of Electrolyte-Extractable Soil Organic Carbon J. Environ. Qual., October 12, 2005; 34(6): 1992 - 1997. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Vadose Zone Journal | Journal of Plant Registrations | ||||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal |
