In 1972, the US Congress enacted the Clean Water Act (CWA) to protect US water resources. A critical section (305[b]) of the CWA calls for periodic accounting of the success or failure of efforts to protect and restore US waters. Over the years, several groups have reviewed the available data and concluded that they do not adequately describe the condition of US waters. The Wadeable Streams Assessment presents the first set of results from what will be a long-term partnership between the US Environmental Protection Agency, the individual states, tribal nations, and other federal agencies. The goal of this partnership is to fill critical information gaps that remain a deterrent to our ability to determine whether our policies and investments have resulted in improvement of US water resources.

The Wadeable Streams Assessment (WSA) provided the first statistically sound summary of the ecological condition of streams and small rivers in the US. Information provided in the assessment filled an important gap in meeting the requirements of the US Clean Water Act. The purpose of the WSA was to: 1) report on the ecological condition of all wadeable, perennial streams and rivers within the conterminous US, 2) describe the biological condition of these systems with direct measures of aquatic life, and 3) identify and rank the relative importance of chemical and physical stressors affecting stream and river condition. The assessment included perennial wadeable streams and rivers that accounted for 95% of the length of flowing waters in the US. The US Environmental Protection Agency, states, and tribes collected chemical, physical, and biological data at 1392 randomly selected sites. Nationally, 42% of the length of US streams was in poor condition compared to best available reference sites in their ecoregions, 25% was in fair condition, and 28% was in good condition. Results were reported for 3 major regions: Eastern Highlands, Plains and Lowlands, and West. In the West, 45% of the length of wadeable flowing waters was in good condition. In the Eastern Highlands, only 18% of the length of wadeable streams and rivers was in good condition and 52% was in poor condition. In the Plains and Lowlands, almost 30% of the length of wadeable streams and rivers was in good condition and 40% was in poor condition. The most widespread stressors observed nationally and in each of the 3 major regions were N, P, riparian disturbance, and streambed sediments. Excess nutrients and excess streambed sediments had the highest impact on biological condition; streams scoring poor for these stressors were at 2 to 3× higher risk of having poor biological condition than were streams that scored in the good range for the same stressors.

The US Environmental Protection Agency (EPA) conducted a Wadeable Stream Assessment (WSA) of all wadeable streams and rivers in the conterminous US between 1999 and 2005. The assessment was led by the EPA Office of Water, in cooperation with EPA regions, states, tribal nations, and the EPA Office of Research and Development (ORD). The WSA was implemented as 2 large-scale regional surveys of streams and rivers. Both studies used EPA's River Reach File (RF3) as the basis for the sample frame. The Environmental Monitoring and Assessment Program (EMAP) Western Pilot Study, conducted by ORD in cooperation with EPA Regions 8, 9, and 10 and 12 western states, assessed all streams and rivers in the 12 western states (EMAP-West). A stratified, unequal probability survey design (50 sites/state and additional sites in 5 intensive study areas) was used to select sites from all streams and river segments coded as perennial in RF3. The unequal selection depended on Strahler order, aggregated Omernik level III ecoregion, and special study region. The WSA study used the EMAP-West wadeable streams (WSA-West) and implemented a new design for the remaining 36 eastern conterminous states (WSA-East). The WSA-East design was an unequal probability survey design with unequal selection depending on Strahler order, Omernik Level II ecoregion, and EPA region. RF3 includes 5.29 million km of rivers and streams, of which 39% (2.07 million km) are coded as perennial. The WSA sample frame included 2.84 million km of streams (54% of the total length in RF3), of which 2.24 million km were in WSA-East and 0.60 million km were in WSA-West. Each selected site was classified on the basis of wadeability and the presence of flowing water. The estimated length of wadeable streams and rivers in the 48 conterminous states was 1.30 ± 0.025 (SE) million km (45.7 ± 1.1% of the stream length in the sample frame). Of this wadeable stream length, 78.6 ± 1.0% (1.02 million km) was estimated to be appropriate for sampling. Nationally, 11.5 ± 0.8% and 5.2 ± 0.6% of this length could not be sampled because of access denial or physical inaccessibility, respectively. The proportion of length affected by access denial was higher in Southern Plains, Northern Plains, and Xeric West aggregated ecoregions, whereas stream length affected by physical inaccessibility was greatest in the Western Mountains aggregated ecoregion. Improvements in the sample frame (RF3 and its successors National Hydrography Database [NHD] and NHD-Plus) would reduce field costs for national surveys.

The US Environmental Protection Agency (EPA) is revising its strategy to obtain the information needed to answer questions pertinent to water-quality management efficiently and rigorously at national scales. One tool of this revised strategy is use of statistically based surveys of aquatic resources such as lakes, wetlands, rivers, and streams. Implementing large-extent surveys involves decisions that reflect compromises between scientific rigor, consistent and practical implementation over large areas and many participants, and the realities of time and money. Primary constraints result from interactions among management objectives, time lines, funding, and institutional constraints of participants. Secondary constraints arise from the interaction between the survey design (geographic extent, sample size, use of existing data), logistics (sampling period, sample shipping, information management, crew expertise, field training), and the suite of ecological indicators selected (site-scale sampling design, field and laboratory protocols). We use our experience with the EPA's national Wadeable Stream Assessment and its Environmental Monitoring and Assessment Program Western Pilot Study to describe the key constraints and the resulting decisions made to implement those surveys successfully. The experiences from those surveys offer perspectives and information useful for effectively implementing future surveys of similar scope or spatial extent, including advanced planning, compatible survey designs, consistent methods, indicators, and cooperative research.

One of the biggest challenges when conducting a continental-scale assessment of streams is setting appropriate expectations for the assessed sites. The challenge occurs for 2 reasons: 1) tremendous natural environmental heterogeneity exists within a continental landscape and 2) reference sites vary in quality both across and within major regions of the continent. We describe the process used to set expectations for the multimetric index of biotic integrity (MIBI) and observed/expected (O/E) indices generated from predictive models used to assess stream condition for the US Wadeable Streams Assessment (WSA). The assessment was based on a reference-site approach, in which the least-disturbed sites in each region of the US were used to establish benchmarks for assessing the condition of macroinvertebrate assemblages at other sites. Reference sites were compiled by filtering WSA sample sites for disturbance using a series of abiotic variables. Additional reference sites were needed and were obtained from other state, university, and federal monitoring programs. This pool of potential reference sites was then assessed for uniformity in site quality and comparability of macroinvertebrate sample data. Ultimately, 1625 sites were used to set reference expectations for the WSA. Reference-site data were used to help define 9 large ecoregions that minimized the naturally occurring variation in macroinvertebrate assemblages associated with continental-wide differences in biogeography. These ecoregions were used as a basis for developing MIBI and O/E indices and for reporting results. A least-disturbed definition of reference condition was used nationally, but we suspect that the quality of the best extant sites in ecoregions, such as the Northern Plains and Temperate Plains, was lower than that of sites in other ecoregions. For the MIBI assessment, we used a simple modeling approach to adjust scores in ecoregions where gradients in reference-site quality could be demonstrated conclusively. The WSA provided an unparalleled opportunity to push the limits of our conceptual and technical understanding of how to best apply a reference-condition approach to a real-world need. Our hope is that we have learned enough from this exercise to improve the technical quality of the next round of national assessments.

Differences in sampling and laboratory protocols, differences in techniques used to evaluate metrics, and differing scales of calibration and application prohibit the use of many existing multimetric indices (MMIs) in large-scale bioassessments. We describe an approach to developing MMIs of ecological condition that is applicable to a variety of biological assemblage types and to spatially extensive (regional, national) aquatic resource surveys. The process involves testing the performance characteristics of candidate metrics in several categories that correspond to key dimensions of biotic condition. The performance characteristics include: information content (range), reproducibility, calibration for natural gradients, responsiveness to stressor gradients, and independence from other metrics. The best-performing metric from each category is included in the final MMI. The consistency of the process enables development of separate MMIs in different regions that can be combined in a national assessment and that are more comparable across regions and taxonomic groups than a set of independently developed MMIs would be. We provide an example of the process applied to macroinvertebrate data from the US Environmental Protection Agency's Wadeable Streams Assessment (WSA) from 3045 sites (of which 1390 were WSA probability sites). The MMIs developed for the WSA demonstrate the feasibility of conducting bioassessments at continental scales and provide a basis for interpreting existing MMIs from regional- and national-level perspectives.

We examined the effects of different regionalization schemes on the performance of River InVertebrate Prediction and Classification System (RIVPACS)-type predictive models in assessing the biological conditions of streams of the US for the National Wadeable Streams Assessment (WSA). Three regionalization schemes were considered: a single national predictive model (MOD1), separate predictive models for each of the 9 WSA aggregated Omernik level III ecoregions (MOD9), and 3 predictive models roughly corresponding to the western US, the Appalachian Mountains, and the Central and Coastal Plains (MOD3). The goal of the WSA was to assess stream condition at the national scale and at the scale of WSA aggregated ecoregions, so we compared the performance of the ratio of the observed number of taxa to the expected number of taxa (O/E) index estimated using different regionalization schemes at both of these spatial scales. We assessed model performance with a randomized resampling procedure, in which we set aside 10% of the reference sites, calibrated the model with the remaining sites, and applied the model to the set-aside sites. Performance statistics for the set-aside reference sites were accumulated over 10 iterations. When summarized at the national scale, mean model predictions of O/E for set-aside reference sites from the 3 different regionalization schemes were all reasonably close to 1. When summarized by the 9 aggregated ecoregions, MOD1 and MOD3 predictions of O/E differed systematically from 1 in certain aggregated ecoregions. Over all 9 ecoregions, the magnitude of these differences was significantly greater than observed with MOD9 predictions. Results from our analysis suggest that O/E values at test sites should be interpreted with respect to mean and SD of O/E of reference sites from the same region to minimize the effects of systematic biases in the predictions. RIVPACS-type predictive models also should be calibrated at a spatial scale similar to the scale at which summary statistics are reported.

Taxonomic identifications are central to biological assessment; thus, documenting and reporting uncertainty associated with identifications is critical. The presumption that comparable results would be obtained, regardless of which or how many taxonomists were used to identify samples, lies at the core of any assessment. As part of a national survey of streams, 741 benthic macroinvertebrate samples were collected throughout the eastern USA, subsampled in laboratories to ∼500 organisms/sample, and sent to taxonomists for identification and enumeration. Primary identifications were done by 25 taxonomists in 8 laboratories. For each laboratory, ∼10% of the samples were randomly selected for quality control (QC) reidentification and sent to an independent taxonomist in a separate laboratory (total n = 74), and the 2 sets of results were compared directly. The results of the sample-based comparisons were summarized as % taxonomic disagreement (PTD) and % difference in enumeration (PDE). Across the set of QC samples, mean values of PTD and PDE were ∼21 and 2.6%, respectively. The primary and QC taxonomists interacted via detailed reconciliation conference calls after initial results were obtained, and specific corrective actions were implemented (if needed) prior to a 2^nd round of comparisons. This process improved consistency (PTD = 14%). Corrective actions reduced the proportion of samples that failed the measurement quality objective for PTD from 71 to 27%. Detailed comparisons of results for individual taxa and interpretation of the potential causes for differences provided direction for addressing problematic taxa, differential expertise among multiple taxonomists, and data entry and recording errors. The taxa that proved most difficult (i.e., had high rates of errors) included many Baetidae, Odonata, Ceratopogonidae, selected groups of Chironomidae, and some Hydropsychidae. We emphasize the importance of experience and training and recommend approaches for improving taxonomic consistency, including documentation of standard procedures, taxonomic data quality standards, and routine and rigorous quality control evaluations.

A major goal of the national aquatic surveys being conducted by the Environmental Protection Agency in partnership with the states and tribes is to assess the relative importance, at a regional scale, of stressors that impact aquatic biota. The Wadeable Streams Assessment (WSA) was a prototype of these surveys, and it assessed 8 individual water-chemistry, physical-habitat, and landuse stressors in 2 ways. First, the WSA estimated the total length of streams in a region that was deemed to be in poor condition for each stressor considered separately. Estimates of stressor extent describe the prevalence and regional breadth of each stressor's potential effects. Second, the WSA estimated each stressor's relative risk, a measure widely used in human epidemiology. Relative risk estimates a stressor's association with biota in terms of the likelihood that poor stressor conditions and poor biota conditions co-occur in a region's streams. We describe how the population attributable risk, also borrowed from epidemiology, combines extent and relative risk into a single, overall measure of a stressor's regional effect. The attributable risk of a stressor is the % reduction in the regional extent of poor biological condition (measured here by the macroinvertebrate index of biotic integrity [MIBI]), that presumably would result from eliminating that stressor. Under the attributable risk assumptions, for example, the WSA data imply that the nationwide extent of poor MIBI conditions would be reduced by an estimated 26% if excess total P were eliminated as a stressor, but only by 3% if excess salinity were eliminated. We also illustrate the attributable risk for the combined effects of multiple, correlated stressors. Last, we discuss how best to interpret and apply attributable risk estimates.

We analyzed nutrient data from a probability survey of 1392 wadeable streams across the 48 conterminous states of the US and from intensified survey data in 921 streams in the Pacific Northwest (PNW) to examine different methods of setting nutrient criteria and to develop a nutrient stream typology. We calculated potential nutrient criteria for total P (TP) and total N (TN) by 3 methods (ecoregion population 25^th percentile of population, least-disturbed reference-site 75^th percentile, and disturbance modeling) and compared them with existing draft US Environmental Protection Agency (EPA) criteria within 14 national nutrient ecoregions. All criteria derived from the methods were highly correlated; however, absolute values within ecoregions differed greatly among approaches. Population 25^th percentiles of TP were almost always lower from statistically designed survey data than from found data. TN percentiles were more similar than were TP profiles, but they still tended to be lower from survey data than from found data. TP and TN population 25^th percentiles were lower (often by a factor of 2–6) than reference-site 75^th percentiles in all ecoregions. This result indicates that population 25^th percentiles cannot be used as surrogates for reference-site 75^th percentiles. Thirty-nine percent of the assessed national stream length exceeded TP criteria and 47% exceeded TN criteria when compared to nutrient criteria based on EPA Wadeable Stream Assessment reference-site 75^th percentiles. In the PNW data set, all disturbance regression model estimates of background nutrient concentrations were lower than reference-site 75^th percentiles. Regression tree analysis based on PNW reference sites used runoff, elevation, acid neutralizing capacity, forest composition, substrate size, and Omernik level III ecoregion as environmental class predictors to explain 46 to 48% of the total deviance in nutrient concentration. Reference-site nutrient concentrations varied widely among Omernik level III ecoregions in nutrient ecoregion II. Our analysis and the literature strongly suggest that 14 national nutrient ecoregions are too coarse to account for natural variation in stream nutrient concentrations. Setting appropriate national nutrient criteria will require finer-scale typology or classification of sites that better controls for natural variation.

Assessments of stream ecosystems often include an evaluation of riparian condition; a key stressor in riparian ecosystems is the presence of invasive plants. We analyzed the distribution of 12 invasive taxa (common burdock [Arctium minus], giant reed [Arundo donax], cheatgrass [Bromus tectorum], musk thistle [Carduus nutans], Canada thistle [Cirsium arvense], teasel [Dipsacus fullonum], Russian olive [Elaeagnus angustifolia], leafy spurge [Euphorbia esula], English ivy [Hedera helix], reed canarygrass [Phalaris arundinacea], Himalayan blackberry [Rubus armeniacus], and saltcedar [Tamarix spp.]) to characterize a portion of that stressor. Observations from 961 probability survey reaches and 355 additional reaches distributed across 12 western US states provided a statistically defensible foundation for trend monitoring, risk assessments, or economic evaluation of these 12 taxa over a large area. We estimate that ≥1 of these taxa are present in riparian areas on 47 ± 3.6% of the perennial stream length in the western US. One or more of these taxa were present in >⅓ of the reaches identified as least-disturbed (reaches that define reference condition and are used to quantify instream biotic integrity). Association between target invasive presence and instream biotic integrity varied, particularly as a function of ecoregion. Relationships were often statistically significant in the Mountain climatic region, sometimes significant in the Xeric climatic region, and never significant in the Plains climatic region. Regional variations in associations and confounding relationships between multiple factors suggest that multiple variables should be examined to explain or predict the presence of invasive species or their associations. Our survey illustrates strengths and limitations of collecting information on a limited number of invasive plants in riparian vegetation as part of general probability surveys of aquatic ecosystems. Our survey of only 12 somewhat arbitrarily selected invasive plants provided much information for a limited cost. We recommend including similar efforts in future surveys.

The increasing demand for tools that can score biological condition from aquatic community data has spurred the creation of many predictive models (e.g., observed/expected [O/E] indices) and multimetric indices (MMIs). The geographic and environmental scopes of these indices vary widely, and coverages often overlap. If indices developed for large environmentally heterogeneous regions provide results equivalent to those developed for smaller regions, then regulatory entities could adopt indices developed for larger regions rather than fund the development of multiple indices within a region. We evaluated this potential by comparing the performance (precision, bias, responsiveness, and sensitivity) of benthic macroinvertebrate O/E indices and MMIs developed for California (CA) with that of indices developed for 2 large-scale condition assessments of US streams: the Environmental Monitoring and Assessment Program Western Pilot Study (EMAP-West) and the western portion of the Wadeable Streams Assessment (WSA-West). WSA-West and EMAP-West O/E scores were weakly correlated with CA O/E scores, had lower precision than CA O/E scores, were influenced by 2 related natural gradients (% slope and % fast-water habitat) that did not influence CA O/E scores, and disagreed with 21 to 22% of impairment decisions derived from the CA O/E index. The WSA-West O/E index produced many fewer impairment decisions than did the CA O/E index. WSA-West and EMAP-West MMI scores were strongly correlated with the CA MMI scores. However, the WSA-West and EMAP-West MMIs produced many fewer determinations of impairment than did the CA MMI. EMAP-West and WSA-West MMIs were biased and differed in responsiveness compared with CA MMI. Thus, they might produce estimates of regional condition different from those from indices calibrated to local conditions. The lower precision of the EMAP-West and WSA-West indices compromises their use in site-specific assessments where both precision and accuracy are important. However, the magnitude of differences in impairment decisions was sensitive to the thresholds used to define impaired conditions, so it might be possible to adjust some of the systematic differences among the models to make the large-scale models more suitable for local application. Future work should identify the geographic and environmental scales that optimize index performance, determine the factors that most strongly influence index performance, and identify ways to specify accurate reference condition from geographically extensive reference-site data sets.

Inferences drawn from regional bioassessments could be strengthened by integrating data from different monitoring programs. We combined data from the US Geological Survey National Water-Quality Assessment (NAWQA) program and the US Environmental Protection Agency Wadeable Streams Assessment (WSA) to expand the scope of an existing River InVertebrate Prediction and Classification System (RIVPACS)–type predictive model and to assess the biological condition of streams across the western US in a variety of landuse classes. We used model-derived estimates of taxon-specific probabilities of capture and observed taxon occurrences to identify taxa that were absent from sites where they were predicted to occur (decreasers) and taxa that were present at sites where they were not predicted to occur (increasers). Integration of 87 NAWQA reference sites increased the scope of the existing WSA predictive model to include larger streams and later season sampling. Biological condition at 336 NAWQA test sites was significantly (p < 0.001) associated with basin land use and tended to be lower in basins with intensive landuse modification (e.g., mixed, urban, and agricultural basins) than in basins with relatively undisturbed land use (e.g., forested basins). Of the 437 taxa observed among reference and test sites, 180 (41%) were increasers or decreasers. In general, decreasers had a different set of ecological traits (functional traits or tolerance values) than did increasers. We could predict whether a taxon was a decreaser or an increaser based on just a few traits, e.g., desiccation resistance, timing of larval development, habit, and thermal preference, but we were unable to predict the type of basin land use from trait states present in invertebrate assemblages. Refined characterization of traits might be required before bioassessment data can be used routinely to aid in the diagnoses of the causes of biological impairment.

The species composition of benthic diatoms was related to environmental conditions in streams throughout the western US to develop diatom traits, indicators for assessment of biological condition and indicators for diagnosing stressors. We hypothesized that indicators based on species traits determined for subsets of streams with similar natural landscape features would be more precisely related to environmental conditions than would be indicators calculated based on species traits for all streams in the data set. The ranges of many environmental conditions were wide among western streams, and these conditions covaried greatly along a major environmental gradient characterized by positive correlations among % watershed disturbed by agricultural and urban land uses (% WD), conductivity, total N, total P, and % fine sediments. Species traits were calculated for 242 diatom taxa. Weighted average (WA) methods were used to define species environmental optima, and regression approaches were used to determine whether species were sensitive or tolerant to environmental conditions indicated by % WD, total P, total N, a nutrient multivariate index, pH, conductivity, % fine sediments, % embeddedness, and a watershed disturbance multivariate index. Indicators based on WA optima and sensitive/tolerant traits were highly correlated with these environmental conditions. Natural and anthropogenic conditions varied greatly among classes of streams grouped by climate regions, but indicators developed for the entire western US were consistently more accurate than were regional indicators. Indicators for individual stressors, such as total P, conductivity, and % embeddedness, were highly correlated with values of respective stressors, but covariation among all indicators and stressors indicated that only 1 environmental gradient was reliably reflected by the indicators. Thus, robust indicators of the biological condition of diatom assemblages were developed for streams of the western US, but development of stressor-specific indicators will require application of additional analytical approaches.