Landscape scale and contaminant effects on aquatic assemblage structure.

Abstract

Biological surveys are routinely used throughout the United States to identify localized impairments in aquatic ecosystems. This approach however, has had limited acceptance for in situ assessment situations in determining specific sources or causes of observed impairments as required under Section 303(d) of the Clean Water Act. While the best approach for determining the effect of urban impacts on streams is to directly compare biological data before and after urban impacts this approach is rarely used because of the lack of historical or pre-disturbance data. Traditionally, the source-cause investigation focused on using aquatic life chemical criteria as benchmarks, resulting in a "violation perspective" methodology that emphasizes specific water quality criteria being exceeded. Relying solely on this approach can be problematic since not all environmental stressors will have established criteria (e.g., sediment criteria are lacking) and those having criteria may not be sufficiently protective of portions of the aquatic resource (e.g., ammonia). This violation perspective assumes that intermittent chemical sampling and analysis will eventually discover the variables (contaminants) causing the impairment and emphasizes a select few water quality criteria exemplifying the “pollutant” focused approach as opposed to a broader and more comprehensive pollution focused approach. Furthermore, chemical water quality criteria are further removed from the designated use, which is more directly measured by the biota and minimizes type I and II assessment errors that would otherwise be more frequent. Evaluating aquatic systems using the violation perspective becomes increasingly more problematic due to increasing water samples collection costs, increased analysis costs for possible chemical stressors, and determining the identity among the thousands of possible stressors. Imperative to this discussion is that slightly elevated contaminant concentrations, synergistic effects, or sporadic spikes could adversely affect fish assemblage structure. As a result, these factors can potentially result in a biological impairment without the occurrence of specific chemical criteria violations. Nationally, the perception of causality for biologically impaired systems has shifted from point-source influences to more diffuse non-point source influences. Difficulty in tracking these pervasive non point-source impacts, combined with the lack of pre-determined signatory relationships with biological assemblage patterns creates a more complex problem. One way of increasing our knowledge of signatory relationships is through multivariate analysis utilizing the definable relationships between aquatic assemblage structure and quantifiable environmental stressors. The purpose of this research was multifaceted. We investigated the relationship between stressor response models associated with an urban landscape, multiple assemblage response, and fish assemblage nutrient response. Essentially the study area for this research encompassed data collected from across the State of Indiana. The nature of the analysis performed resulted in this volume of data being compartmentalized into discreet spatially driven subsets that were analyzed independently. To determine the responsiveness of fish assemblages to stressors associated with an urban landscape we targeted the Salt Creek Watershed. Salt Creek is a Lake Michigan tributary in Northwest Indiana, USA, which drains a watershed experiencing rapid urbanization as part of the expansion of the Greater Chicago metropolitan area. The watershed supports a managed coldwater fishery comprised principally of the introduced Skamania strain of the steelhead (Oncorhynchus mykiss). The sustainability of this watershed depends on the proper management of warm water tributaries and salmonid water in the Salt Creek mainstem. Twenty-three fish species were collected in the Salt Creek watershed and were numerically dominated by creek chub (Semotilus atromaculatus) and green sunfish (Lepomis cyanellus) both of which are tolerant to a wide range of environmental conditions. Habitat quality, measured using the Qualitative Habitat Evaluation Index (QHEI), showed that the watershed was generally degraded and scores ranged from 12-69. Fourteen parameters were significantly correlated with reach scale ecological health and biological integrity. Factor analysis found three factors explained 69% of the contributed variance in the watershed fish assemblage. The first factor included habitat measures comprised of the QHEI score and three of its metrics (i.e., channel, riparian and instream cover scores) and explained 36 percent of data variability. The second factor was comprised of two contaminants (i.e., TDS and Chloride) and one local-scale land-use variable (Agriculture) that explained an additional 20 percent of the variability. The third factor was comprised of two local scale land-use variables (i.e., riparian zone and water) explaining 13percent of the variability. To evaluate the responsiveness of multiple aquatic assembles to watershed stress we target the Big Oaks National Wildlife Refuge. The Big Oaks National Wildlife Refuge encompasses the northern 51,000 acres of the former Jefferson Proving Ground (JPG) which was used from 1940-1995 as a munitions testing facility. Since 2000 the U.S. Fish and Wildlife Service has utilized the northern 51,000 acres of JPG for ecosystem-based management in conjunction with continued use by the U.S. Department of Army and Indiana Air National Guard for air-to-ground training. An investigation of factors that explained the variance in fish, crayfish, and macroinvertebrate assemblage structure and function was based on catchment and reach-scale land use, habitat, contaminants, and water quality. Habitat quality, measured using the Qualitative Habitat Evaluation Index (QHEI), showed that scores ranged from 25 to 85 (average 61.36 + 10.08). The substrate score, instream cover, riffle-run score, and channel score were the primary factors contributing to declining QHEI scores. Factor analysis found four factors explained 69 percent of the contributed variance in the fish assemblage, two factors accounted for 56 percent of the total variance in macroinvertebrate assemblages, and two factors explained 49 percent of the cumulative variance in crayfish assemblages. Overall drivers of assemblage structure were associated with broad scale issues of wastewater treatment, ground water, and land-use. Our results show that fish, macroinvertebrate, and crayfish assemblages respond to similar broad scale stimulus; however, the specific physical/chemical constituent responsible for the stress may vary, and the realized magnitude of the overall stress on the system may be expressed by each organismal group differently. Our data suggest that varying organismal groups can respond independently and stress reflected in one assemblage may not necessarily be observed in another.Finally, we evaluated nutrient response in fish assemblages targeting a large data set collected from the Indiana portion of the Corn Belt Plain Ecoregion. Due to the complex interactions between the various forms of Nitrogen and Phosphorus within respective cycles, Total Nitrogen (TN) and Total Phosphorus (TP) cycling interactions can no longer be accepted as sole limiting factors in either marine or freshwaters. This study is conducted as part of the U.S. Environmental Protection Agency (USEPA) desire to development regional nutrient thresholds. The first objective of this study is to develop a biotic model capable of determining the contributions of various nutrients, including Nitrogen components and TP, in streams using fish assemblages. The second objective is to establish an approach for designating defensible nutrient biotic index (NBI) score thresholds and corresponding nutrient concentrations, above which fish assemblages show alterations due to increasing nutrient concentrations. Sampling within Indiana’s portion of the Corn Belt and Northern Great Plain Nutrient Ecoregion occurred from 1996-2007 at 1274 sites. Nutrient data were reviewed for outliers and then sorted into three groups relative to drainage class. Each group was arranged into 15 ranges or “bins” using the Jenks optimization method in Arc GIS 9.3. Next, sites were assigned to each bin relative to observed concentrations. These bin assignments were used to populate the species occurrence model for nutrient optima calculation. Nutrient optima were calculated by dividing the sum of the weighted proportion of times a species occurred in each bin by the un-weighted proportion of times a species occurred in each bin. The derived nutrient optima were divided into eleven equal ranges, by nutrient, and tolerance scores (0-10) assigned with respect to each species derived optima. Nutrient tolerance scores were used to calculate Nutrient Biotic Index (NBI) scores for each sampling site by summing the number of individuals of a given species at the site and multiplying times that species tolerance value then dividing by the total number of individuals at the site. A single break point was observed for unionized ammonia, which showed an NBIUnionized Ammonia score shift between 0.003 and 0.03 (mg/L). The mean NBIUnionized Ammonia scores were 3.09 and 3.29, respectively. Nutrient Biotic IndexUnionized Ammonia scores were significantly correlated with IBI score and IBI integrity class. Three break points were observed for Nitrogen, Nitrate+Nitrite, demonstrating a significant NBINitrate+Nitrite score shift at mean concentrations of 0.13 mg/L, 1.09 mg/L, 3.15 mg/L and 6.87 mg/L respectively. The mean NBINitrate+Nitrite scores were 5.58, 5.37, 5.82 and 6.25, respectively. The observed relationship produced a convex curve suggesting an enrichment signature. Nutrient Biotic IndexNitrate+Nitrite scores were significantly correlated with IBI score and IBI integrity class. Two break points were observed for Total Kjeldahl Nitrogen (TKN), which were significant. The mean concentrations of TKN were 0.4 mg/L, 0.68 mg/L, and 1.27 mg/L, respectively. The mean NBITKN scores were 2.73, 3.10, and 3.37, respectively. Nutrient Biotic IndexTKN scores were significantly related to IBI score and IBI integrity class. Two break points observed for TN were significant at concentrations of 0.56 mg/L and 3.30 mg/L. The mean NBITN scores were 4.60 and 4.85, respectively. Nutrient Biotic IndexTN scores were not significantly related to IBI score or IBI integrity class. Two significant break points were observed for TP. The mean concentrations of TP were 0.07 mg/L and 0.32 mg/L, respectively and mean NBITP scores were 3.43 and 3.58, respectively. Nutrient Biotic IndexTP scores were significantly related to IBI score and IBI integrity class. Two break points were observed for Chlorophyll a (periphyton), which were significant. Mean concentrations were 10.15 mg/m2 and 134.14 mg/m2, respectively. Mean NBIPeriphyton scores were 3.75 and 4.20, respectively. Nutrient Biotic IndexPeriphyton scores were not significantly related to IBI score, but were significantly related to IBI integrity class. Four break points were observed for Chlorophyll a (phytoplankton), which occurred at Chlorophyll a (phytoplankton) concentrations of 2.33 μg/L, 10.98 μg/L and 49.13 μg/L, respectively. The mean NBIPhytoplankton scores were 3.43, 3.85 and 5.02, respectively. Nutrient Biotic IndexPhytoplankton scores were significantly related to IBI score and IBI integrity class. Nutrient criteria concentration was interpreted for NBI and IBI integrity class relationships to establish protective nutrient concentration benchmarks. Proposed mean protection values are 3.0 μg/L for Unionized Ammonia, 130 μg/L for Nitrogen, Nitrate+Nitrite, 40 μg/L for TKN, 70 μg/L for TP, and 2.33 μg/L for Chlorophyll a (phytoplankton). Criteria established at or below these benchmarks should protect for both biological integrity of fish assemblages in Indiana as well as nutrient loadings into the Gulf of Mexico.