Clements: “3357_c031” — 2007/11/9 — 12:43 — page 665 — #1 31 Descriptive Approaches for Assessing Ecosystem Responses to Contaminants 31.1 INTRODUCTION Now that we have an appreciation of the important processes that characterize ecosystems and the general approaches used to quantify these processes, we will turn our attention to the primary objective of this section.Aswithcommunity-levelassessments, ecotoxicologists interested in ecosys- tem responses to anthropogenic stressors employ descriptive, quasi-experimental, and experimental approaches. In the following two chapters, we will explore the use of these observational and experimental studies to link changes in primary and secondary production, nutrient cycling, and decomposition to contaminants and other anthropogenic stressors. In a separate chapter, we will con- sider effects of globally distributed and atmospheric stressors (e.g., acidification, NO x deposition, elevated CO 2 , and UV radiation) on these ecosystem processes. Investigations of ecosystem processes may be conducted across a range of spatial and temporal scales. Functional measures such as community metabolism or nutrient transport can be measured in isolated soil microbial systems or in whole forests or watersheds. However, as we move up the hier- archy of biological organization from individuals → populations → communities → ecosystems, we generally increase the spatial and temporal scales of our investigations. Because many experi- mental studies of ecosystem processes are often limited in spatial and temporal scale, descriptive approaches can provide very compellingand ecologically realistic results. As discussed in Chapter 23 for communities, the typical trade-off is that observational or correlative investigations only provide a catalog of potential causal explanations. A more powerful case for causation in descriptive stud- ies can be established by the application of strong inference (Platt 1964), other formal inferential methods such as stressor identification (Suter et al. 2002), or Bayesian inferential techniques. The initial definition of ecological integrity proposed by Karr (1991) included both structural and functional measures, and most ecologists would agree that assessing effects of anthropogenic stressors on ecosystems requires adequate characterization of both patterns and processes. The effic- acy of using functional measures to assess ecosystem responses to contaminants has received limited attention. As a consequence, development of functional criteria as indicators of ecological integrity has lagged behind more traditional approaches based on community structure (Bunn and Davies 2000, Gessner and Chauvet 2002, Hill et al. 1997). Kersting (1994) provides an excellent review of literature on the use of functional endpoints in freshwater field tests for hazard assessment of chemicals. Some assessments of ecological integrity measure patterns of community composition as a surrogate for ecosystem processes (Bunn and Davies 2000); however, patterns and processes are not necessarily related in some instances, especially in systems where disturbance is relatively weak. For example, Bunn and Davies (2000) measured stream metabolism at seven sites in southwestern Australia and relatedecosystemprocessestocommunitystructure.Althoughchangesingrossprimary production (GPP) and respiration were related to water quality, there was no relationship between water quality and macroinvertebrate community structure. 665 © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 666 — #2 666 Ecotoxicology: A Comprehensive Treatment The characterization of ecological integrity based exclusively on structural measures is incon- sistent with how most ecologists view ecosystems (Gessner and Chauvet 2002). We believe that restricting our analyses to mainly structural measures has provided a somewhat incomplete picture of how ecosystems respond to and recover from anthropogenic disturbances. Issues such as relative sensitivity, response variability, and functional redundancy have been considered when comparing the usefulness of structural and functional measures (Howarth 1991, Schindler 1987, 1988). Despite concern that changes in some ecosystem processes occur only after compositional changes and are therefore less useful, alterations in material cycling and energy flow are such fundamental properties of ecosystems that they should be included in ecological assessments. Leland and Carter (1985) argue that some functional processes in ecosystems are easier to quantify than relationships between abundance and environmental variables. These functional measures also integrate general charac- teristics of diverse communities, thus facilitating comparisons among different ecosystems. Given the recent interest in making comparisons across relatively broad geographic regions, functional measures may be more useful than structural measures because they are not dependent on specific taxa that are often restricted to a specific region. Finally, because causal mechanisms that control ecosystem processes are generally well understood, restoration strategies may be more obvious when based on functional measures (Bunn and Davies 2000). Although papers that report relative sensitivity of community metrics to contaminants are com- mon in the literature (Carlisle and Clements 1999, Kilgour et al. 2004), surprisingly few studies have compared responses across levels of biological organization (Adams et al. 2002, Bendell- Young et al. 2000, Cottingham and Carpenter 1998, Niemi et al. 1993, Sheehan 1984, Sheehan and Knight 1985). There is also the perception that quantifying ecosystem responses is logistically challenging compared to structural measures (Crossey and La Point 1988), an idea that has not been rigorously examined in the literature. Thus, we believe that it is premature to conclude that ecosystem processes are less sensitive or less reliable indicators of stress. In fact, some studies have reported that changes in ecosystem processes may occur in the absence of alterations in community structure (Bunn and Davies 2000). Niemi et al. (1993) reported that functional measures such as GPP were more sensitive indicators of recovery than structural measures. Similarly, Clements (2004) observed that community respiration was generally more sensitive to heavy metals than common structural measures such as abundance and species richness. Because alterations in community struc- ture are not necessarily related to ecosystem processes, we view these as complementary measures for assessing ecological integrity. More important, simultaneous assessment of pattern and process can provide insight into the mechanistic linkages between stressors and responses. Studies by Wal- lace and colleagues (Wallace et al. 1996) provide some of the best examples demonstrating how contaminant-induced alterations in structural characteristics (e.g., elimination of macroinvertebrate shredders) directly influence ecosystem processes (e.g., litter decomposition and export). There is also some evidence thatfunctional measures may be moredirectlyrelated to specific typesofstressors (Gessner and Chauvet 2002). Although many different processes could be used to assess ecosystem integrity, we will focus in this section on three functional measures: ecosystem metabolism (respiration, primary and second- ary production), litter decomposition, and nutrient cycling. As described in Chapter 30, a significant amount of background information characterizing these processes is available, although the level of development varies among ecosystem types. For example, lake ecologists have historically relied on functional measures, especially primary production, whereas lotic ecologists have tended to rely on structural measures (Gessner and Chauvet 2002). These differences have resulted in diver- gent approaches used to assess effects of contaminants in aquatic ecosystems. Similarly, studies of biogeochemical processes in terrestrialhabitats, especiallyinagricultural systems, focus primarily on factors that increase primary production, whereas aquatic ecologists have been more concerned with understanding factors that limit production as a way to control eutrophication (Grimm et al. 2003). The methodological approaches used to assess effects of contaminants on ecosystem metabolism are consequently different in aquatic and terrestrial ecosystems. © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 667 — #3 Descriptive Approaches for Assessing Ecosystem Responses to Contaminants 667 31.2 DESCRIPTIVE APPROACHES IN AQUATIC ECOSYSTEMS 31.2.1 E COSYSTEM METABOLISM AND PRIMARY PRODUCTION Energy flow and metabolism are fundamental properties of ecosystems that are also closely related to the transport of contaminants. Many of the same physical, chemical, and biological processes that influence the flow of energy between biotic and abiotic compartments also regulate the fate of chemicals. Primary production in aquatic ecosystems is particularly sensitive to many anthropogenic stressors. The effects of nutrient subsidies and input of organic materials on productivity have received considerable attention in streams, lakes, and marine ecosystems. In their description of subsidy-stress gradients, Odum et al. (1979) contrast the ecosystem-level effects of “utilizable” inputs such as nutrients and organic materials with toxic materials (Chapter 25). This analysis contrasts the role of nutrients such as N and P as both regulators of ecosystem production as well as stressors when threshold levels are exceeded. Input of nutrients associated with agricultural, domestic, industrial, and atmospheric sources are widely regarded as major stressors of aquatic ecosystems (National Research Council 1992). Whole ecosystem nutrient budgets calculated for several ecosystems reveal that inputs often exceed outputs, resulting in large amounts of nutrients being stored in a watershed (Bennett et al. 1999, Jowarski et al. 1992, Lowrance et al. 1985). Bennett et al. (1999) used amass-balance approach to estimate P storage based on inputs and outputs in the Lake Mendota (Wisconsin, USA) watershed. They reported that approximately 50% of the P entering the watershed was retained and could be readily mobilized by climatic, geologic, or hydrologic events. These increased nutrient levels in aquatic ecosystems are often associated with toxic algal blooms, increased plant growth, oxygen depletion, fish kills, and major shifts in community composition. Land-based inputs of nutrients also increase eutrophication and have negative effects on primary productionofmacrophytes in coastal areas. Using data compiled from an extensive literature survey, Valiela and Cole (2002) reported a strong inverse relationship between N loading and primary production of seagrass meadows in coastal marine areas. The percent of seagrass cover lost reached 100% as N loading approached 100 Kg N/ha/y. These effects resulted from reduced light supply associated with increased phytoplankton production. The damaging effects of N enrichment were significantly reduced in areas protected by salt marshes and mangroves. As described in Chapter 30, availability of N and P can directly regulate primary production and biomass accrual in aquatic ecosystems (Biggs 2000). However, the direct effects of nutrients on primary production are complex and may be mediated by other factors such as hydrologic char- acteristics and abundance of grazers. Riseng et al. (2004) used covariance structure analysis (CVA) to examine effects of hydrologic regime and nutrients in 97 midwestern U.S. streams. Increased nutrients in streams with high hydrologic variability resulted in greater algal abundance because grazers were reduced. In contrast, in more stable streams where grazers were abundant, algal pro- duction was limited and the net effect was an increase in herbivore production. Because hydrologic characteristics of a watershed are dependent on watershed physiography and climate, these factors may ultimately control responses to nutrient additions (Riseng et al. 2004). Alterations in primary productivity and respiration have been measured in response to chemical stressors other than nutrients in aquatic ecosystems. Crossey and La Point (1988) could not detect differences in GPP between metal-impacted and reference sites, but when data were normalized to algal biomass (as chlorophyll a), GPP was higher at the reference site. Hill et al. (1997) measured variance and sensitivity of several functional measures in the Eagle River, a Colorado (USA) stream impacted by metals. Results showed that measures of community metabolism (GPP, NPP, and res- piration) were lower at stations located downstream from heavy metal inputs (Figure 31.1). These functional measures were correlated with mortality of Ceriodaphnia dubia, and inhibition concen- trations (IC50 values) for respiration were comparable to LC50 values derived from these more traditional toxicological approaches. These results suggest that functional measures were about © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 668 — #4 668 Ecotoxicology: A Comprehensive Treatment Station ER-1r ER-3 ER-5 ER-12 ER-12a 0 2 4 6 8 10 12 14 Zn concentration (µg/L) 0 50 100 150 200 250 300 GPP Zn concentration GPP (g O 2 /m 2 /day) FIGURE 31.1 Effects of Zn on community respiration measured at different stations in the Eagle River, Colorado, United States. (Data from Tables 2 and 3 in Hill et al. (1997).) as sensitive as acute toxicity for quantifying effects of heavy metals. Clements (2004) compared structural (species richness and abundance of metal-sensitive mayflies) and functional (community respiration) responses of benthic macroinvertebrate communities to a mixture of Cd and Zn in stream microcosms (Figure 31.2). Both structuraland functional measures weresignificantly related to metal concentration, but effects on community respiration were generally greater than effects on species richness or abundance of metal-sensitive mayflies. Unlike studies conducted with diatoms and attached algae, research investigating effects of contaminants on emergent macrophytes has shown that primary production and photosynthesis of these groups are relatively insensitive. Bendell-Young et al. (2000) compared the response of several structural and functional endpoints (mutagenic responses, morphological deformities, mortality, community structure, and plant productivity) measured in wetlands receiving oil sands effluents. Photosynthetic rates of cattails (Typha latifolia L.), measured as CO 2 uptake, were actually greater in wetlands receiving processed water from oil sands, a response that contradicted expectations. These researchers concluded that structural changes in benthic communities and blood chemistry of fish were more sensitive indicators of stress than functional measures. Photosynthetic rate of salt marsh plants (Spartina alterniflora) was measured at reference and contaminated sites in the southeastern United States (Wall et al. 2001). Although significant negative effects on benthic detritivores were observed at a site heavily contaminated by mercury and PCBs, photosynthesis of Spartina was not affected. 31.2.2 SECONDARY PRODUCTION In addition to direct effects on primary producers, contaminants and other stressors may alter the amount and rate of energy flow to higher trophic levels. The utilization of available energy in an ecosystem is thus an important measure of ecological integrity. Perhaps the most common functional response related to energetics measured in aquaticecosystems is the abundance ofdifferent functional feeding groups (Rawer-Jost et al. 2000, Wallace et al. 1996). In part, the utility of functional feeding groups as a metric in ecological assessments is based on the assumption that specialist feeders such as scrapers and shredders are more sensitive to contaminants than generalist feeders such as collector- gatherers and filterers (Barbour et al. 1996). Although data on functional feeding groups are generally presented as abundance or density per unit area, and therefore not strictly a functional measure, © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 669 — #5 Descriptive Approaches for Assessing Ecosystem Responses to Contaminants 669 Total number of species 16 18 20 22 24 CCU 010 20304050 Community respiration 0.0 0.5 1.0 1.5 Total number of heptageniidae 200 250 300 FIGURE 31.2 Relationship between structural (total number of species; abundance of metal-sensitive hepta- geniid mayflies) and functional (community respiration) endpoints and heavy metal (Cd and Zn) concentration in stream microcosms. Heavy metal concentration was expressed as the cumulative criterion unit (CCU),defined as the ratio of the measured metal concentration to the hardness adjusted chronic criterion values for Cd and Zn. (Data from Clements (2004).) the assumption isthatcompositionofdifferent feeding groups reflects important ecosystemprocesses. For example, abundance of grazers, organisms that feed directly on periphyton and algae, is related to primary productivity in streams. Similarly, abundance of shredders, organisms that process leaf litter, regulates downstream transport of coarse particulate organic material (Wallace et al. 1982). Secondary production, which we have defined as the production of heterotrophic organisms, has been used to document effects of several stressors in aquatic ecosystems, including hydrologic modification (Raddum and Fjellheim 1993), pesticides (Whiles and Wallace 1995), urbanization (Shieh et al. 2002), and heavy metals (Carlisle and Clements 2003). Because secondary production integrates individual growth rates and population dynamics, it captures in a single measure several important aspects of energy flow through ecosystems.Although integration of these measures across levels of biological organization is a laudable goal in ecosystem ecotoxicology, measures of second- ary production are rarely included in biological assessments. Even measuring secondary production of individual species is highly labor intensive because it requires sampling populations with sufficient frequency to quantify individual growth rates, mortality, immigration, and emigration. The logistical © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 670 — #6 670 Ecotoxicology: A Comprehensive Treatment challenges associated with measuring secondary production will likely deter some researchers from using this endpoint in ecological assessments. Indeed, France (1996) argues that because secondary production is dependent on abundance, little additional information is gained by including these more labor-intensive approaches. However, because secondary production is a composite of indi- vidual mortality, growth rate, population abundance, and biomass, it represents a potential holistic indicator of ecosystem bioenergetics that is not reflected in these individual measures. Because most studies of secondaryproduction are based on detailed analysis of individual species or groups of related species, our understanding of energeticsfrom the perspective of entire ecosystems is somewhat limited (Shieh et al. 2002). To be useful as an ecosystem-level indicator, a measure of secondary production should include a significant number of dominant species in an ecosystem and should also be combined with data on trophic interactions. Sheehan and Knight (1985) compared patterns of community composition and secondary production at several sites along a gradient of metal contamination. Chironomids dominated benthic communities at metal-polluted sites, a finding commonly reported in the literature. However, despite large shifts in community composition among sites, relatively little difference in secondary production of chironomids was observed. Another challenge associated with using secondary production as an indicator of ecosystem integrity is that production may either increase or decrease, depending on the nature of the stressor. The theoretical basis for the difference in responses between subsidizing and toxic stressors was first described by Odum et al. (1979), but there have been relatively few empirical studies docu- menting this pattern. Shieh et al. (2002) estimated energy flow based on secondary production and trophic interactions at polluted and reference sites in a Colorado stream receiving urban discharges. Secondary production, which was primarily supported by detritus, increased by more than two times at the most impacted site due to the input of nutrients and organic materials (Figure 31.3a). In contrast to these patterns, Carlisle and Clements (2003) reported a decline in secondary production along a gradient of heavy metal contamination (Figure 31.3b). Differences in production among these streams were primarily a result of lower population abundance of metal-sensitive species, especially grazing mayflies. The large reduction in secondary production of herbivores likely had important cascading effects on trophic interactions and energy flow through this ecosystem. Results of these studies are consistent with predictions of the subsidy-stress hypothesis (Odum et al. 1979) and illustrate the contrasting effects of subsidizing materials and toxic chemicals on ecosystem processes. Reduced secondary production of zooplankton in lake ecosystems may result from increased mortality, lower growth rates, and/or shifts in size composition of dominant species. Hanazato (2001) reviewed effects of pesticides on zooplankton across levels of organization, from individuals to ecosystem-level responses. Ageneral trend observed in lakes receiving pesticides was a reduction in mean body size of zooplankton as a result of differential sensitivity among species. Hanazato (2001) speculated that one potential ecosystem-level consequence of altered size distributions was a reduction in the amount of energy transferred from primary producers to higher trophic levels. This reduced transfer efficiency was associated with a variety of anthropogenic stressors, including heavy metals, acidification, and nutrient enrichment. 31.2.3 DECOMPOSITION Litter decomposition is a fundamental ecological process that has been studied extensively, espe- cially in lotic ecosystems (Chapter 30). It integrates responses of a variety of biota, from bacteria and fungi to shredding macroinvertebrates (Niyogi et al. 2001, 2003). There is a large database available reporting decomposition rates of leaves and quantifying biotic and abiotic factors that influence litter decay under a variety of environmental conditions. Gessner and Chauvet (2002) provided an excellent and comprehensive review of numerous studies that used litter breakdown to quantify effects of physical and chemical stressors in streams. They make a compelling argument for the use of decomposition as an ecosystem indicator and provide specific criteria for assessing ecological integrity. Breakdown rate coefficients (k), measured by regressing remaining mass of © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 671 — #7 Descriptive Approaches for Assessing Ecosystem Responses to Contaminants 671 Reference Intermed N High N Production (g AFDM/m 2 /year) 0 10 20 30 40 50 60 70 0.0 0.5 1.0 1.5 2° Production Nitrogen Station Reference Low Zn Intermed Zn High Zn Production (g DW/m 2 /year) Zn (µg/L) NH 4 (mg/L) 0 1 2 3 4 5 6 0 50 100 150 200 250 300 2° Production Zinc (a) (b) FIGURE 31.3 Contrasting effects of nutrients (a) and heavy metals (b) on invertebrate secondary production. (Effects of nutrients from Tables 1 and 4 in Shieh et al. (2002). Effects of Zn from Tables 1 and 2 in Carlisle and Clements (2003).) litter against time, are generally reduced in disturbed ecosystems. Effects of contaminants on litter decay may result either from alterations in microbial processes or reduced abundance of macroin- vertebrate shredders (Figure 31.4). It is also necessary to distinguish effects of contaminants on biological processes, such as the elimination of shredders or reduced microbial activity, from effects due to physical characteristics of the system. Methodological approaches that exclude or include different groups of organisms can be used to separate the relative importance of these processes, thus allowing ecologists to isolate underlying mechanisms. Because of the diversity of approaches used to quantify litter decay and the large number of environmental factors that influence k, development of standardized techniques for assessing effects of contaminants is essential. The most comprehensive applications of leaf litter methodologies to investigate effects of con- taminants have been conducted in metal-polluted and acidified streams. Schulthesis and Hendricks (1999) and Schulthesis et al. (1999) measured macroinvertebrate community composition and leaf decomposition at sites upstream and downstream from an abandoned pyrite mine in southwest- ern Virginia (USA). Shredder abundance was greater and decomposition rates were 1.4–2.7 times faster at the reference site compared Cu-polluted sites. Remediation activities initiated during the © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 672 — #8 672 Ecotoxicology: A Comprehensive Treatment Physicochemical characteristics Microbial processes Contaminants and other stressors Shredder biomass and production Leaf decomposition FIGURE 31.4 Conceptual model showing the effects of contaminants and physicochemical characteristics on microbial processes, shredder biomass, and leaf litter decomposition. TABLE 31.1 The Influence of Aqueous Zn Concentration, Metal Oxide Deposition, and Nutrient Concentrations on Structural and Functional Endpoints Measured at 27 Stream Sites in the Rocky Mountains, Colorado (USA) Variable Independent Variables R 2 P-Value Leaf breakdown rate (k) [Zn], oxide deposition .72 .0001 Shredder biomass [Zn], oxide deposition .64 .0001 Microbial respiration Oxide deposition, nutrients .54 .0001 Data from Table 2 in Niyogi et al. (2001). study period allowed these researchers to compare recovery of structural and functional responses. Although community composition and abundance of shredders increased following improvements in water quality, the rate of leaf processing did not increase as expected, suggesting some resid- ual effects of Cu on microbial processes. In contrast to these studies, Nelson (2000) reported little effects of Zn contamination on decomposition rates of aspen (Populus tremuloides) in a Colorado Rocky Mountain stream, despite significant changes in community composition. These research- ers speculated that the lack of a response in their study resulted from the relative insensitivity of microbes, especially fungi, to the moderate levels of Zn contamination.Alternatively, because micro- bial activity is limited by cold temperatures in Rocky Mountain streams, leaf processing may be more dependent on invertebrate shredders (Niyogi et al. 2001), which were unaffected by Zn in this study (Nelson 2000). Comparative studies in streams across a gradient of heavy metal pollution provide the best oppor- tunity to quantify effects of stressors relative to other factors that regulate leaf decomposition. Niyogi et al. (2001) measured decomposition rates at 27 stream sites (8 reference and 19 metal-polluted) in the Rocky Mountains of Colorado, USA. In addition to its broad spatial scale, this study is unique because researchers quantified the relative influence of several stressors associated with mining pollution, including acidification, elevated Zn concentration, and metal oxide deposition. Litter decay coefficients (k) and shredder biomass decreased with increasing aqueous Zn concentration and deposition of metal oxides (Table 31.1). In contrast, microbial respiration was more influenced by metal oxide deposition and nutrients. Because decay coefficients were more closely related to © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 673 — #9 Descriptive Approaches for Assessing Ecosystem Responses to Contaminants 673 shredder biomass than microbial respiration, results of this study suggest that macroinvertebrates were more important than microbial processes in regulating leaf litter processing in streams (Niyogi et al. 2001). Carlisle and Clements (2005) related leaf decomposition to shredder secondary produc- tion and microbial respiration in reference and metal-polluted streams in Colorado, USA. Because leaf decomposition was measured as a function of shredder secondary production instead of shredder biomass, this study provided a unique opportunity to quantify effects of stressors on energy flow through an allochthonous food web. Results showed that shredders disproportionately contributed to leaf litter decay, and that species-specific differences in sensitivity to metals among shredders helped explain differences among streams. Stream acidification by atmospheric deposition or other sources can have direct effects on litter decomposition (Dangles et al. 2004, Griffith and Perry 1993, Tuchman 1993, Webster and Benfield 1986). Griffith and Perry (1993) attributed slower processing rate of litter in acidic streams to lower biomass of shredders. In contrast, differences in community composition of shredders were primarily responsible for differences in processing rates between neutral and more alkaline streams. Tuchman (1993) reported that declines in invertebrate shredders in acidified lakes were correlated with decreased litter breakdown rates. As with studies of metal-polluted streams, the most convin- cing evidence demonstrating a relationship between acidification and leaf decomposition has been obtained from spatially extensive surveys. Dangles et al. (2004) measured microbial respiration, litter decay, and shredder abundance and composition in 25 streams along a gradient of acidification in the Vosges Mountains, France. Breakdown rates varied 20-fold between acidified and neutral streams, with alkalinity and aluminum concentration explaining 88% of the variation. Reduced leaf decomposition in acidified streams was related to lower abundance and biomass of the amphipod, Gammarus fossarum, a functionally important and acid-sensitive species. The greater breakdown rate observed in coarse mesh bags (5.0 mm), which allowed shredder colonization, compared to fine mesh bags (0.3 mm), which excluded shredders, supported the hypothesis that microbial processes were relatively unimportant in this investigation (Dangles and Guerold 2001). Although pesticides and other organic contaminants are likely to have significant effects on litter decomposition by altering microbial processes and shredder communities, these stressors have received considerably less attention than heavy metals and acidification (Gessner and Chauvet 2002). Delorenzo et al. (2001) provided a comprehensive review of the effects of pesticides on microbial processes related to decomposition. The best examples showing the effects of organic chemicals on shredder biomass and subsequent alterations in leaf processing involve long-term experimental studies (Whiles and Wallace 1992, 1995) and stream mesocosm experiments (Stout and Cooper 1983), which will be described in Chapter 32. Swift et al. (1988) examined effects of dimilin, an insect growth regulator used for control of gypsy moths, on litter decomposition. Although laboratory bioassays with shredders showed significant mortality when shredders were fed dimilin- treated leaves, decomposition rates of treated leaves in the field were actually greater than controls. The faster processing rate of treated leaves was attributed to the potential carbon source that dimilin provided for bacteria. Most investigations of leaf litter decomposition report that decay coefficients are reduced in contaminated streams. However, stressors that subsidize an ecosystem (e.g., nutrients or organic materials) may have the opposite effect. Niyogi et al. (2003) measured breakdown of tussock grass (Chionocloa rigida) in 12 New Zealand streams along a gradient of agricultural develop- ment. Nutrients (N and P), the predominant stressors in this system, increased along this gradient and stimulated microbial respiration, invertebrate abundance, and the rate of litter decomposition. In contrast, the macroinvertebrate community index (MCI), a biotic index of organic pollution, showed increased stress along this same gradient. Similar findings were reported by Pascoal et al. (2001) for a stream in Portugal receiving elevated nutrients. Despite reductions in abundance of shredders at polluted sites, leaf breakdown rates were greater. These results serve to illustrate the importance of understanding mechanistic linkages among stressors, microbial processes, and macroinvertebrate community composition when using leaf decomposition to assess ecological integrity. © 2008 by Taylor & Francis Group, LLC Clements: “3357_c031” — 2007/11/9 — 12:43 — page 674 — #10 674 Ecotoxicology: A Comprehensive Treatment In addition to the direct effects of contaminants on the rate of litter decay, the concentrations of toxic chemicals may increase in decomposing of plant material, thus providing a direct link to detritus-based food chains. Windham et al. (2004) measured reduced decomposition of marsh grass (Spartina alterniflora) in a metal-contaminated marsh as compared to a reference site. The 10– 100 times increase in metal concentrations in decomposing litter was attributed to adsorption and microbial processes. 31.2.4 NUTRIENT CYCLING The majority of studies investigating effects of contaminants on nutrient cycling in aquatic ecosys- tems have focused on nitrification, denitrification, and other processes associated with N flux (Kemp and Dodds 2002, Kemp et al. 1990, Royer et al. 2004). Most of this research has been conducted within the context of understanding effects of nutrient enrichment, especially N and P, on freshwater and estuarine ecosystems. Eutrophication, caused by the release of excess nutrients, is regarded as the major threat to freshwater and coastal ecosystems in the United States (U.S. EPA1990). Greater than 50% of the impaired lake area and river reaches in the United States result from excess nutri- ents. Most of this impairment is associated with nonpoint source inputs from agricultural and urban activities (Carpenter et al. 1998), although atmospheric deposition is considered an important source of N to some areas. In particular, N inputs from the upper Midwest to the Gulf of Mexico have increased dramatically in the past several decades, and excess nutrients have had severe effects on water quality and community composition. A significant portion of the N from nonpoint sources is retained in aquatic ecosystems by biological processes such as microbial uptake as well as lat- eral exchange with the hyporheic zone. However, despite N retention in some aquatic ecosystems, a large amount of excess N is transported downstream. Royer et al. (2004) measured denitrification in headwater stream sediments located in agricultural areas. Because denitrification rates were low in these streams, there was relatively little influence on instream concentrations and therefore most of the NO 3 –N was transported downstream. These researchers concluded that previous estimates of denitrification rates may have overestimated N loss to the sediments. Because rates of nitrification and denitrification in aquatic ecosystems are dependent on concen- trations of ammonium (NH 4 ) and nitrate (NO 3 ), these processes are likely to increase in areas receiv- ing anthropogenic inputs. Kemp and Dodds (2002) measured effects of anthropogenic N on rates of nitrification and denitrification in pristine and agriculturally influenced watersheds. Whole stream nitrification and denitrification rates were greater at agriculturally influenced sites, most likely due to greater input of N. Despite greater denitrification, the large amount of anthropogenic N exceeded the natural retentive ability of the stream and a significant amount was transported downstream. As described in Chapter 30, retention of nutrients and organic materials is dependent on a number of physical, chemical, and biological characteristics. Headwater streams often represent the largest portion of the linear dimension of a watershed and are closely connected to surrounding riparian and terrestrial ecosystems. These systems are generally considered to be highly retentive of nutrients (Peterson et al. 2001). A similar situation exists in coastal marine ecosystems. Meta-analysis of data collected in coastal areas demonstrated that denitrification by fringing wetlands (e.g., salt marshes, mangroves) serves to intercept excess nutrients and protect seagrass meadows from anthropogenic N (Valiela and Coe 2002). Physical disturbances such as removal of vegetation and reduced habitat complexity may decouple denitrification and nitrification processes in aquatic ecosystems, thereby exacerbating the effects of nutrient enrichment (Kemp and Dodds 2002). 31.3 TERRESTRIAL ECOSYSTEMS 31.3.1 R ESPIRATION AND SOIL MICROBIAL PROCESSES Functional measures of ecosystem processes have also been used to characterize the impacts of contaminants in terrestrial ecosystems. In particular, effects of contaminants on microbial and soil © 2008 by Taylor & Francis Group, LLC [...]... Johnson, D and Hale, B., White birch (Betula papyrifera Marshall) foliar litter decomposition in relation to trace metal atmospheric inputs at metal-contaminated and uncontaminated sites near Sudbury, Ontario and Rouyn-Noranda, Quebec, Canada, Environ Pollut., 127, 65–72, 2004 Karr, J.R., Biological integrity: Along-neglected aspect of water resource management, Ecol Appl., 1, 66–84, 1991 Kauppi, P.E.,... physicochemical characteristics of soil such as pH, the amount of organic material, and soil composition Because these soil characteristics may also vary with metal contamination, field assessments of nitrification must account for natural differences at reference and contaminated sites For example, metal deposition from smelters is often associated with soil acidification, which would increase metal bioavailability... laboratory responses are necessary Furthermore, because soil pH and organic matter may affect nitrification rates and contaminant bioavailability, experimental approaches are necessary to determine if the relationship between nitrification and soil characteristics is a result of direct or indirect effects 31. 3.4 AN INTEGRATION OF TERRESTRIAL AND AQUATIC PROCESSES Biogeochemists generally recognize that a. .. Accumulation of litter in a forest near a lead-zinc-cadmium smelter was attributed to reduced decomposition associated with heavy metal contamination (Coughtrey et al 1979) McEnroe and Helmisaari (2001) measured decomposition of pine needles along a gradient of metal contamination in soil Decreased mass loss and reduced C:N ratios were attributed to metal contamination in soils and litter Metal concentrations... that adaptation of soil microbes to metals significantly reduced effects on nitrification Sauve et al (1999) questioned the utility of nitrification as an indicator of contamination because of its sensitivity to other factors such as pH and organic carbon Because of potential differences in effects of metals and other contaminants in field-collected and laboratory-spiked soils, studies that validate laboratory... decomposition can also affect nutrient cycling and growth of vegetation, thereby reducing soil organic content and increasing contaminant bioavailability (Derome and Nieminen 1998, Johnson and Hale 2004) Because contaminants may affect processing of detritus and accumulation of organic material in soils, the ability of an ecosystem to assimilate contaminants may be reduced because of lower organic content and subsequently... contrast, soils contaminated by sewage sludge may have elevated levels of organic material, which would likely reduce contaminant bioavailability There is also the potential for positive feedback relationships between the bioavailability of contaminants and soil properties For example, the amount of organic material in soils is in part regulated by nitrification and microbial decomposition (Figure 31. 7)... hydrologic extremes, Can J Fish Aquat Sci., 52, 2402–2422, 1995 Windham, L., Weis, J.S., and Weis, P., Metal dynamics of plant litter of Spartina alterniflora and Phragmites australis in metal-contaminated salt marshes Part 1: Patterns of decomposition and metal uptake, Environ Toxicol Chem., 23, 1520–1528, 2004 Zimakowska-Gnoinska, D., Bech, J., and Tobias, F.J., Assessment of the heavy metal pollution effects... of soil microalgae were associated with large (>90%) reductions in microbial activity (dehydrogenase, nitrate reductase) at field sites contaminated by pentachlorophenol (PCP) Megharaj et al (2000) measured changes in soil microbial function at sites contaminated by DDT and its metabolites Because sensitive bacteria were replaced by DDTtolerant microorganisms, total microbial biomass was found to be... clean and contaminated litter is measured at reference and impacted sites, is an effective approach for distinguishing the effects of contaminants in soils from those deposited on leaf litter Johnson and Hale (2004) measured metal accumulation and decomposition of white birch leaves at metal-polluted sites in Canada Decomposition rates were reduced at a metal-polluted site, but concentrations of metals . field-collected and laboratory-spiked soils, studies that validate laboratory responses are necessary. Furthermore, because soil pH and organic matter may affect nitrification rates and contaminant. andHale 2004, Kohleret al. 1995). Becausecon- taminants may affect microbial processesin the soil, litter palatability, and/or invertebrateabundance, additional studies that quantify the relative. streams was related to lower abundance and biomass of the amphipod, Gammarus fossarum, a functionally important and acid-sensitive species. The greater breakdown rate observed in coarse mesh bags