Identifying Key Stressors Driving Biological Impairment in Freshwater Streams in the Chesapeake Bay Watershed, USA

Biological communities in freshwater streams are often impaired by multiple stressors (e.g., flow or water quality) originating from anthropogenic activities such as urbanization, agriculture, or energy extraction. Restoration efforts in the Chesapeake Bay watershed, USA seek to improve biological conditions in 10% of freshwater tributaries and to protect the biological integrity of existing healthy watersheds. To achieve these goals, resource managers need to better understand which stressors are most likely driving biological impairment. Our study addressed this knowledge gap through two approaches: 1) reviewing and synthesizing published multi-stressor studies, and 2) examining 303(d) listed impairments linked to biological impairment as identified by jurisdiction regulatory agencies (the states within the watershed and the District of Columbia). Results identified geomorphology (i.e., physical habitat), salinity, and toxic contaminants as important for explaining variability in benthic community metrics in the literature review. Geomorphology (i.e., physical habitat and sediment), salinity, and nutrients were the most reported stressors in the jurisdictional impairment analysis. Salinity is likely a major stressor in urban and mining settings, whereas geomorphology was commonly reported in agricultural settings. Toxic contaminants, such as pesticides, were rarely measured; more research is needed to quantify the extent of their effects in the region. Flow alteration was also highlighted as an important urban stressor in the literature review but was rarely measured in the literature or reported by jurisdictions as a cause of impairment. These results can be used to prioritize stressor monitoring by managers, and to improve stressor identification methods for identifying causes of biological impairment.

Spatial correlation of macroinvertebrate assemblages in streams and the implications for bioassessment programs

Stream bioassessment using benthic macroinvertebrate assemblages is widely implemented by regulatory agencies, yet a critical question regarding spatial autocorrelation and sample independence remains: How much of a stream network does a point sample represent? Macroinvertebrate assemblages vary along a longitudinal gradient, likely due to a combination of natural and anthropogenic factors that alter water physiochemistry and habitat. A better understanding of how these gradients affect macroinvertebrate assemblage variance could prevent spatial over- and under-sampling within bioassessment efforts. This project investigated longitudinal patterns (10 s of km) of macroinvertebrate assemblages in 14 Wisconsin streams. Spatial autocorrelation was assessed using Moran's I and other multivariate methods with an emphasis on estimating the distance at which assemblages no longer display spatial correlation. Within most streams, there was a linear, direct relationship between assemblage dissimilarity and longitudinal distance, with distance to independence (DTI) ranging from 1.7-13.5 km. DTI was most strongly affected by conductivity, which is often a surrogate for a suite of anthropogenic effects. With increasing conductivity, DTI increases, suggesting more homogenous assemblages in disturbed streams. Natural factors like watershed size, channel gradient, and riparian slope also affected DTI. Considering spatial correlation in monitoring designs could improve the efficiency and application of regulatory bioassessment programs.

Spatial and Temporal Patterns of Macroinvertebrate Assemblages in the River Po Catchment (Northern Italy)

In the last decade, large scale biomonitoring programs have been implemented to obtain a robust understanding of freshwater in the name of helping to inform and develop effective restoration and management plans. A comprehensive biomonitoring dataset on the macroinvertebrate assemblages inhabiting the rivers of the Po Valley (northern Italy), comprised a total of 6762 sampling events (period 2007–2018), was analyzed in this study in order to examine coarse spatial and temporal trends displayed by biotic communities. Our results showed that macroinvertebrate compositions and derived structural and functional metrics were controlled by multiple environmental drivers, including altitude and climate (large scale), as well as habitat characteristics (local scale). Altitude proved to be the primary geographic driver, likely due to its association with thermal and precipitation regimes, thus explaining its overriding influence on macroinvertebrate assemblages. Significant temporal variations were observed across the study period, but notably in 2017, the overall taxonomic richness and diversity increased at the expense of Ephemeroptera, Plectoptera and Trichoptera taxa during an unprecedented heatwave that occurred across southern Europe. The detail of this study dataset allowed for important environmental attributes (e.g., altitude, habitat characteristics) shaping biotic communities to be identified, along with ecologically vulnerable regions and time periods (e.g., extreme climatic events). Such research is required globally to help inform large-scale management and restoration efforts that are sustainable over long-term periods.

The effects of pastoral intensification on the feeding interactions of generalist predators in streams

Land‐use change can alter trophic interactions with wide‐ranging functional consequences, yet the consequences for aquatic food webs have been little studied. In part, this may reflect the challenges of resolving the diets of aquatic organisms using classical gut contents analysis, especially for soft‐bodied prey. We used next‐generation sequencing to resolve prey use in nearly 400 individuals of two predatory invertebrates (the Caddisfly, Rhyacophila dorsalis, and the Stonefly Dinocras cephalotes) in streams draining land with increasingly intensive livestock farming. Rhyacophila dorsalis occurred in all streams, whereas D. cephalotes was restricted to low intensities, allowing us to test whether: (i) apparent sensitivity to agriculture in the latter species reflects a more specialized diet and (ii) diet in R. dorsalis varied between sites with and without D. cephalotes. DNA was extracted from dissected gut contents, amplified without blocking probes and sequenced using Ion Torrent technology. Both predators were generalists, consuming 30 prey taxa with a preference for taxa that were abundant in all streams or that increased with intensification. Where both predators were present, their diets were nearly identical, and R. dorsalis's diet was virtually unchanged in the absence of D. cephalotes. The loss of D. cephalotes from more intensive sites was probably due to physicochemical stressors, such as sedimentation, rather than to dietary specialization, although wider biotic factors (e.g., competition with other predatory taxa) could not be excluded. This study provides a uniquely detailed description of predator diets along a land‐use intensity gradient, offering new insights into how anthropogenic stressors affect stream communities.

Modeled hydrologic metrics show links between hydrology and the functional composition of stream assemblages

Flow alteration is widespread in streams, but current understanding of the effects of differences in flow characteristics on stream biological communities is incomplete. We tested hypotheses about the effect of variation in hydrology on stream communities by using generalized additive models to relate watershed information to the values of different flow metrics at gauged sites. Flow models accounted for 54-80% of the spatial variation in flow metric values among gauged sites. We then used these models to predict flow metrics in 842 ungauged stream sites in the mid-Atlantic United States that were sampled for fish, macroinvertebrates, and environmental covariates. Fish and macroinvertebrate assemblages were characterized in terms of a suite of metrics that quantified aspects of community composition, diversity, and functional traits that were expected to be associated with differences in flow characteristics. We related modeled flow metrics to biological metrics in a series of stressor-response models. Our analyses identified both drying and base flow instability as explaining 30-50% of the observed variability in fish and invertebrate community composition. Variations in community composition were related to variations in the prevalence of dispersal traits in invertebrates and trophic guilds in fish. The results demonstrate that we can use statistical models to predict hydrologic conditions at bioassessment sites, which, in turn, we can use to estimate relationships between flow conditions and biological characteristics. This analysis provides an approach to quantify the effects of spatial variation in flow metrics using readily available biomonitoring data.

The impact of feral camels (Camelus dromedarius) on remote waterholes in central Australia

The Katiti and Petermann Aboriginal Land Trusts (KPALT) in central Australia contain significant biological and cultural assets, including the World Heritage-listed Uluṟu-Kata Tjuṯa National Park. Until relatively recently, waterbodies in this remote region were not well studied, even though most have deep cultural and ecological significance to local Aboriginal people. The region also contains some of the highest densities of feral dromedary camels (Camelus dromedarius) in the nation, and was a focus area for the recently completed Australian Feral Camel Management Project. Within the project, the specific impacts of feral camels on waterholes were assessed throughout the KPALT. We found that aquatic macroinvertebrate biodiversity was significantly lower at camel-accessible sites, and fewer aquatic taxa considered ‘sensitive’ to habitat degradation were found at sites when or after camels were present. Water quality at camel-accessible sites was also significantly poorer (e.g. more turbid) than at sites inaccessible to camels. These results, in combination with emerging research and anecdotal evidence, suggest that large feral herbivores, such as feral camels and feral horses, are the main immediate threat to many waterbodies in central Australia. Management of large feral herbivores will be a key component in efforts to maintain and improve the health of waterbodies in central Australia, especially those not afforded protection within the national park system.

Drivers of macroinvertebrate community structure in unmodified streams

Often simple metrics are used to summarise complex patterns in stream benthic ecology, thus it is important to understand how well these metrics can explain the finer-scale underlying environmental variation often hidden by coarser-scale influences. I sampled 47 relatively pristine streams in the central North Island of New Zealand in 2007 and (1) evaluated the local-scale drivers of macroinvertebrate community structure as well as both diversity and biomonitoring metrics in this unmodified landscape, and (2) assessed whether these drivers were similar for commonly used univariate metrics and multivariate structure. The drivers of community metrics and multivariate structure were largely similar, with % canopy cover and resource supply metrics the most commonly identified environmental drivers in these pristine streams. For an area with little to no anthropogenic influence, substantial variation was explained in the macroinvertebrate community (up to 70% on the first two components of a partial least squares regression), with both uni- and multivariate approaches. This research highlights two important points: (1) the importance of considering natural underlying environmental variation when assessing the response to coarse environmental gradients, and (2) the importance of considering canopy cover presence when assessing the impact of stressors on stream macroinvertebrate communities.

How many mountains can we mine? Assessing the regional degradation of Central Appalachian rivers by surface coal mining

Surface coal mining is the dominant form of land cover change in Central Appalachia, yet the extent to which surface coal mine runoff is polluting regional rivers is currently unknown. We mapped surface mining from 1976 to 2005 for a 19,581 km(2) area of southern West Virginia and linked these maps with water quality and biological data for 223 streams. The extent of surface mining within catchments is highly correlated with the ionic strength and sulfate concentrations of receiving streams. Generalized additive models were used to estimate the amount of watershed mining, stream ionic strength, or sulfate concentrations beyond which biological impairment (based on state biocriteria) is likely. We find this threshold is reached once surface coal mines occupy >5.4% of their contributing watershed area, ionic strength exceeds 308 μS cm(-1), or sulfate concentrations exceed 50 mg L(-1). Significant losses of many intolerant macroinvertebrate taxa occur when as little as 2.2% of contributing catchments are mined. As of 2005, 5% of the land area of southern WV was converted to surface mines, 6% of regional streams were buried in valley fills, and 22% of the regional stream network length drained watersheds with >5.4% of their surface area converted to mines.

Experimental and environmental factors affect spurious detection of ecological thresholds

Threshold detection methods are increasingly popular for assessing nonlinear responses to environmental change, but their statistical performance remains poorly understood. We simulated linear change in stream benthic macroinvertebrate communities and evaluated the performance of commonly used threshold detection methods based on model fitting (piecewise quantile regression [PQR]), data partitioning (nonparametric change point analysis [NCPA]), and a hybrid approach (significant zero crossings [SiZer]). We demonstrated that false detection of ecological thresholds (type I errors) and inferences on threshold locations are influenced by sample size, rate of linear change, and frequency of observations across the environmental gradient (i.e., sample-environment distribution, SED). However, the relative importance of these factors varied among statistical methods and between inference types. False detection rates were influenced primarily by user-selected parameters for PQR (tau) and SiZer (bandwidth) and secondarily by sample size (for PQR) and SED (for SiZer). In contrast, the location of reported thresholds was influenced primarily by SED. Bootstrapped confidence intervals for NCPA threshold locations revealed strong correspondence to SED. We conclude that the choice of statistical methods for threshold detection should be matched to experimental and environmental constraints to minimize false detection rates and avoid spurious inferences regarding threshold location.

An approach for determining bioassessment performance and comparability

Many organizations in the USA collect aquatic bioassessment data using different sampling and analysis methods, most of which have unknown performance in terms of data quality produced. Thus, the comparability of bioassessments produced by different organizations is often unknown, ultimately affecting our ability to make comprehensive assessments on large spatial scales. We evaluated a pilot approach for determining bioassessment performance using macroinvertebrate data obtained from several states in the Southeastern USA. Performance measures evaluated included precision, sensitivity, and responsiveness to a human disturbance gradient, defined in terms of a land disturbance index value for each site, combined with a value for specific conductance, and instream habitat quality. A key finding of this study is the need to harmonize ecoregional reference conditions among states so as to yield more comparable and consistent bioassessment results. Our approach was also capable of identifying potential areas for refinement such as reevaluation of less precise, sensitive, or responsive metrics that may result in suboptimal index performance. Higher performing bioassessments can yield information beyond "impaired" versus "unimpaired" condition. Acknowledging the limitations of this pilot study, we would recommend that performance evaluations use at least 50 sites, 10 of which are ecoregional reference sites. Efforts should be made to obtain data from the entire human disturbance gradient in an ecoregion to improve statistical confidence in performance measures. Having too few sites will result in an under-representation of certain parts of the disturbance gradient (e.g., too few poor quality sites), which may bias sensitivity and responsiveness estimates.

Ecological integrity of streams related to human cancer mortality rates

Assessments of ecological integrity have become commonplace for biological conservation, but their role for public health analysis remains largely unexplored. We tested the prediction that the ecological integrity of streams would provide an indicator of human cancer mortality rates in West Virginia, USA. We characterized ecological integrity using an index of benthic macroinvertebrate community structure (West Virginia Stream Condition Index, SCI) and quantified human cancer mortality rates using county-level data from the Centers for Disease Control and Prevention. Regression and spatial analyses revealed significant associations between ecological integrity and public health. SCI was negatively related to age-adjusted total cancer mortality per 100,000 people. Respiratory, digestive, urinary, and breast cancer rates increased with ecological disintegrity, but genital and oral cancer rates did not. Smoking, poverty, and urbanization were significantly related to total cancer mortality, but did not explain the observed relationships between ecological integrity and cancer. Coal mining was significantly associated with ecological disintegrity and higher cancer mortality. Spatial analyses also revealed cancer clusters that corresponded to areas of high coal mining intensity. Our results demonstrated significant relationships between ecological integrity and human cancer mortality in West Virginia, and suggested important effects of coal mining on ecological communities and public health. Assessments of ecological integrity therefore may contribute not only to monitoring goals for aquatic life, but also may provide valuable insights for human health and safety.

Estimating the effects of excess nutrients on stream invertebrates from observational data

Increased nutrient concentrations in streams and rivers have altered biological structure and function. Manipulative studies have provided insights into different mechanisms by which changes in nutrient concentrations influence aquatic biota, but these studies are limited in spatial scope and in their quantification of nutrient effects on aggregate measures of the invertebrate assemblage. Observational data provide a complementary source of information to manipulative studies, but these data must be analyzed such that the potential effects of spurious correlations are minimized. Propensity scores, a technique developed to analyze human health observational data, are applied here to estimate the effects of increased nutrients on the total taxon richness of stream invertebrates in a large observational data set collected from the western United States. The analysis indicates that increases in nutrient concentration are strongly associated with and cause decreases in invertebrate richness in large, but wadeable, open-canopied streams. These decreases in invertebrate richness were not mediated by periphyton biomass, a commonly proposed mechanism by which nutrients influence invertebrates. In smaller, closed-canopied streams, increases in nutrients were associated with small increases in total richness that were not statistically significant. Using propensity scores can greatly improve the accuracy of insights drawn from observational data by minimizing the potential that factors other than the factor of interest may confound the results.

Linking multimetric and multivariate approaches to assess the ecological condition of streams

Few attempts have been made to combine multimetric and multivariate analyses for bioassessment despite recognition that an integrated method could yield powerful tools for bioassessment. An approach is described that integrates eight macroinvertebrate community metrics into a Principal Components Analysis to develop a Multivariate Condition Score (MCS) from a calibration dataset of 511 samples. The MCS is compared to an Index of Biotic Integrity (IBI) derived using the same metrics based on the ratio to the reference site mean. Both approaches were highly correlated although the MCS appeared to offer greater potential for discriminating a wider range of impaired conditions. Both the MCS and IBI displayed low temporal variability within reference sites, and were able to distinguish between reference conditions and low levels of catchment modification and local habitat degradation, although neither discriminated among three levels of low impact. Pseudosamples developed to test the response of the metric aggregation approaches to organic enrichment, urban, mining, pastoral and logging stressor scenarios ranked pressures in the same order, but the MCS provided a lower score for the urban scenario and a higher score for the pastoral scenario. The MCS was calculated for an independent test dataset of urban and reference sites, and yielded similar results to the IBI. Although both methods performed comparably, the MCS approach may have some advantages because it removes the subjectivity of assigning thresholds for scoring biological condition, and it appears to discriminate a wider range of degraded conditions.

River-stream connectivity affects fish bioassessment performance

Stream fish bioassessment methods assume that fish assemblages observed in sample sites reflect responses to local stressors, but fish assemblages are influenced by local factors as well as regional dispersal to and from connected streams. We hypothesized that fish movement to and from refugia and source populations in connected rivers (i.e., riverine dispersal) would weaken or decouple relations between fish community metrics and local environmental conditions. We compared fish-environment relations between streams that flow into large rivers (mainstem tributaries) and streams that lack riverine confluences (headwater tributaries) at multiple spatial grains using data from the USEPA's Environmental Monitoring and Assessment Program in the mid-Atlantic highlands, USA (n = 157 sites). Headwater and mainstem tributaries were not different in local environmental conditions, but showed important differences in fish metric responses to environmental quality gradients. Stream sites flowing into mainstem channels within 10 fluvial km showed consistently weaker relations to local environmental conditions than stream sites that lacked such mainstem connections. Moreover, these patterns diminished at longer distances from riverine confluences, consistent with the hypothesis of riverine dispersal. Our results suggest that (1) the precision of fish bioassessment metrics may be improved by calibrating scoring criteria based on the spatial position of sites within stream networks and (2) the spatial grain of fish bioassessment studies may be manipulated to suit objectives by including or excluding fishes exhibiting riverine dispersal.

Coupling GIS and multivariate approaches to reference site selection for wadeable stream monitoring

Geographic Information System (GIS) was used to identify potential reference sites for wadeable stream monitoring, and multivariate analyses were applied to test whether invertebrate communities reflected a priori spatial and stream type classifications. We identified potential reference sites in segments with unmodified vegetation cover adjacent to the stream and in >85% of the upstream catchment. We then used various landcover, amenity and environmental impact databases to eliminate sites that had potential anthropogenic influences upstream and that fell into a range of access classes. Each site identified by this process was coded by four dominant stream classes and seven zones, and 119 candidate sites were randomly selected for follow-up assessment. This process yielded 16 sites conforming to reference site criteria using a conditional-probabilistic design, and these were augmented by an additional 14 existing or special interest reference sites. Non-metric multidimensional scaling (NMS) analysis of percent abundance invertebrate data indicated significant differences in community composition among some of the zones and stream classes identified a priori providing qualified support for this framework in reference site selection. NMS analysis of a range standardised condition and diversity metrics derived from the invertebrate data indicated a core set of 26 closely related sites, and four outliers that were considered atypical of reference site conditions and subsequently dropped from the network. Use of GIS linked to stream typology, available spatial databases and aerial photography greatly enhanced the objectivity and efficiency of reference site selection. The multi-metric ordination approach reduced variability among stream types and bias associated with non-random site selection, and provided an effective way to identify representative reference sites.

Stream communities along a catchment land-use gradient: subsidy-stress responses to pastoral development

When native grassland catchments are converted to pasture, the main effects on stream physicochemistry are usually related to increased nutrient concentrations and fine-sediment input. We predicted that increasing nutrient concentrations would produce a subsidy-stress response (where several ecological metrics first increase and then decrease at higher concentrations) and that increasing sediment cover of the streambed would produce a linear decline in stream health. We predicted that the net effect of agricultural development, estimated as percentage pastoral land cover, would have a nonlinear subsidy-stress or threshold pattern. In our suite of 21 New Zealand streams, epilithic algal biomass and invertebrate density and biomass were higher in catchments with a higher proportion of pastoral land cover, responding mainly to increased nutrient concentration. Invertebrate species richness had a linear, negative relationship with fine-sediment cover but was unrelated to nutrients or pastoral land cover. In accord with our predictions, several invertebrate stream health metrics (Ephemeroptera-Plecoptera-Trichoptera density and richness, New Zealand Macroinvertebrate Community Index, and percent abundance of noninsect taxa) had nonlinear relationships with pastoral land cover and nutrients. Most invertebrate health metrics usually had linear negative relationships with fine-sediment cover. In this region, stream health, as indicated by macroinvertebrates, primarily followed a subsidy-stress pattern with increasing pastoral development; management of these streams should focus on limiting development beyond the point where negative effects are seen.

The biological condition gradient: a descriptive model for interpreting change in aquatic ecosystems

The United States Clean Water Act (CWA; 1972, and as amended, U.S. Code title 33, sections 1251-1387) provides the long-term, national objective to "restore and maintain the ... biological integrity of the Nation's waters" (section 1251). However, the Act does not define the ecological components, or attributes, that constitute biological integrity nor does it recommend scientific methods to measure the condition of aquatic biota. One way to define biological integrity was described over 25 years ago as a balanced, integrated, adaptive system. Since then a variety of different methods and indices have been designed and applied by each state to quantify the biological condition of their waters. Because states in the United States use different methods to determine biological condition, it is currently difficult to determine if conditions vary across states or to combine state assessments to develop regional or national assessments. A nationally applicable model that allows biological condition to be interpreted independently of assessment methods will greatly assist the efforts of environmental practitioners in the United States to (1) assess aquatic resources more uniformly and directly and (2) communicate more clearly to the public both the current status of aquatic resources and their potential for restoration. To address this need, we propose a descriptive model, the Biological Condition Gradient (BCG) that describes how 10 ecological attributes change in response to increasing levels of stressors. We divide this gradient of biological condition into six tiers useful to water quality scientists and managers. The model was tested by determining how consistently a regionally diverse group of biologists assigned samples of macroinvertebrates or fish to the six tiers. Thirty-three macroinvertebrate biologists concurred in 81% of their 54 assignments. Eleven fish biologists concurred in 74% of their 58 assignments. These results support our contention that the BCG represents aspects of biological condition common to existing assessment methods. We believe the model is consistent with ecological theory and will provide a means to make more consistent, ecologically relevant interpretations of the response of aquatic biota to stressors and to better communicate this information to the public.

Predictive models attribute effects on fish assemblages to toxicity and habitat alteration

Biological assessments should both estimate the condition of a biological resource (magnitude of alteration) and provide environmental managers with a diagnosis of the potential causes of impairment. Although methods of quantifying condition are well developed, identifying and proportionately attributing impairment to probable causes remain problematic. Furthermore, analyses of both condition and cause have often been difficult to communicate. We developed an approach that (1) links fish, habitat, and chemistry data collected from hundreds of sites in Ohio (USA) streams, (2) assesses the biological condition at each site, (3) attributes impairment to multiple probable causes, and (4) provides the results of the analyses in simple-to-interpret pie charts. The data set was managed using a geographic information system. Biological condition was assessed using a RIVPACS (river invertebrate prediction and classification system)-like predictive model. The model provided probabilities of capture for 117 fish species based on the geographic location of sites and local habitat descriptors. Impaired biological condition was defined as the proportion of those native species predicted to occur at a site that were observed. The potential toxic effects of exposure to mixtures of contaminants were estimated using species sensitivity distributions and mixture toxicity principles. Generalized linear regression models described species abundance as a function of habitat characteristics. Statistically linking biological condition, habitat characteristics including mixture risks, and species abundance allowed us to evaluate the losses of species with environmental conditions. Results were mapped as simple effect and probable-cause pie charts (EPC pie diagrams), with pie sizes corresponding to magnitude of local impairment, and slice sizes to the relative probable contributions of different stressors. The types of models we used have been successfully applied in ecology and ecotoxicology, but they have not previously been used in concert to quantify impairment and its likely causes. Although data limitations constrained our ability to examine complex interactions between stressors and species, the direct relationships we detected likely represent conservative estimates of stressor contributions to local impairment. Future refinements of the general approach and specific methods described here should yield even more promising results.