Stream Bioassessment Research Articles

Application of a benthic observed/expected-type model for assessing Central Appalachian streams influenced by regional stressors in West Virginia and Kentucky

Stream bioassessments rely on taxonomic composition at sites compared with natural, reference conditions. We developed and tested an observed/expected (O/E) predictive model of taxonomic completeness and an index of compositional dissimilarity (BC index) for Central Appalachian streams using combined macroinvertebrate datasets from riffle habitats in West Virginia (WV) and Kentucky (KY). A total of 102 reference sites were used to calibrate the O/E model, which was then applied to assess over 1,200 sites sampled over a 10-year period. Using an all subsets discriminant function analysis (DFA) procedure, we tested combinations of 14 predictor variables that produced DF and O/E models of varying performance. We selected the most precise model using a probability of capture at >0.5 (O/E₀.₅, SD = 0.159); this model was constructed with only three simple predictor variables--Julian day, latitude, and whether a site was in ecoregion 69a. We evaluated O/E and BC indices between reference and test sites and compared their response to regional stressors, including coal mining, residential development, and acid deposition. The Central Appalachian O/E and BC indices both showed excellent discriminatory power and were significantly correlated to a variety of regional stressors; in some instances, the BC index was slightly more sensitive and responsive than the O/E₀.₅ model. These indices can be used to supplement existing bioassessment tools crucial to detecting and diagnosing stream impacts in the Central Appalachian region of WV and KY.

Optimizing stream bioassessment: habitat, season, and the impacts of land use on benthic macroinvertebrates

Bioassessment of running waters should ideally be optimized to include sampling of the biota when and where they are most sensitive to anthropogenic disturbances, but direct comparisons of the responses of biota across habitats and seasons are lacking. We sampled benthic macroinvertebrates from nine boreal streams situated along an agricultural land use gradient in two seasons (spring and autumn) and two habitats (pools and riffles). Univariate (e.g., diversity) and multivariate (ordination scores) metrics, as well as biological traits, were used to assess changes in assemblage composition associated with agricultural land use. Abundances were generally higher in agricultural compared to forested streams, and in riffles compared to pools. Spring samples had lower mean abundances of several insect taxa (e.g., chironomid midges) compared to autumn samples, while abundances of non-insects (e.g., oligochaetes and Pisidium spp.) remained unchanged. Community turnover (correspondence analysis) had higher precision and sensitivity compared to diversity metrics, and samples from the spring and from riffles responded more to the land use gradient than those from autumn and pool habitats, respectively. The finding that catchment land use resulted in macrohabitat differences and, ultimately, differences in taxonomic composition between agricultural and forested streams and between pool and riffle habitats can be used to optimize future bioassessment based on macroinvertebrates.

Integration of invertebrate traits into predictive models for indirect assessment of stream functional integrity: A case study in Portugal

Biological traits are increasingly used for describing ecological functioning of stream benthic assemblages. Such approaches associate information on species distribution to their biological characteristics (e.g. life history, physiology, dispersal ability) providing a biological trait profile of communities. They may complement structural bioassessment measures using taxonomic composition by providing indirect information on stream ecological functioning, with the additional advantage of being less constrained by biogeographic differences. A multivariate predictive model, that provides a site-specific list of expected taxa at least disturbed conditions was recently developed for the bioassessment of Portuguese streams. Here, we tested if the inclusion of trait information in the model would also enable the detection of most common anthropogenic disturbances (i.e., organic contamination, hydrological disturbance) and provide diagnostic hints for causal relationships. We used existing information on 11 invertebrate biological traits and their 54 categories to convert the observed and expected taxonomic composition at several test sites into expected and observed trait compositions. The first three axes of a normalised PCA (Principal Components Analysis) performed on disturbance variables accounted for 42.7% of explained variability. The proportion of variance in disturbance explained by the three types of trait-based metrics (overall observed/expected trait composition, trait-category profile difference and traits profile dissimilarity) ranged between 9% and 32%. Our predictions made on the response of observed to expected trait categories for organic contamination were generally confirmed and demonstrated that disturbances resulted in a change in those traits conferring species resilience capacity and sensitivity to oxygen depletion, as well as a shift in the proportion of animals with filter-feeding behaviour. Variations in observed to expected trait-category differences showed that even a small increase in organic contamination led to a significant change in the biological trait profile, as expected. By contrast, only two out of 11 trait category predictions were confirmed for hydrological disturbance. Finally, we found that 4/11 and 9/11 observed to expected trait differences showed a significant deviation with organic contamination and hydrological disturbance, respectively, whereas all 11 observed to expected trait differences responded to overall disturbance. These changes in trait profiles reflect changes in the performance of invertebrate communities to cope with disturbance, which potentially can alter ecosystem functioning (e.g., energy flow or chemical cycling). In conclusion, the integration of biological trait information in an AUSRIVAS type predictive model allowed the detection of a general disturbance gradient and particularly organic contamination, which indicates their value in addition to taxonomic-based assessment.

Diatom communities as indicators of ecological status in Mediterranean temporary streams (Balearic Islands, Spain)

Diatoms are widely used for the bioassessment of permanent rivers and streams due to their broad distribution and their ability to integrate changes occurring in water composition and quality, but they have rarely been investigated in temporary streams. Water and epilithic diatom samples were collected from 60 sites in temporary streams distributed among the three largest Balearic Islands (Majorca, Minorca and Ibiza). The sampling design included only temporary streams that have water for at least 4–5 months during the annual period. The selected sites included minimally disturbed (reference sites) and other sites influenced by different degrees of human pressures. Each study site was sampled at least twice in different seasons to take into account the temporal variability in diatom assemblages ( n = 255). Non-metric multidimensional scaling (NMDS) and analyses of similarities (ANOSIM) were used to assess variation in the communities between a priori defined types and along to the degradation gradient, and to study the species tolerance to pollution. A diatom multimetric index (DIATMIB) was developed to assess the ecological status of temporary streams, as a combination of three metrics: abundance of sensitive taxa, abundance of tolerant taxa and a metric of chlorophyll a values. Individual diatom metrics and the multimetric DIATMIB were correlated in a significant way with phosphate, nitrate, ammonia, silicon dioxide, ash free dry mass and water electric conductivity. Cymbella vulgata, Diploneis oblongella and Encyonopsis microcephala appeared only in reference sites and in high and good status classes. These taxa were replaced by others such as Navicula veneta, Nitzschia inconspicua, N. frustulum and Planothidium frequentissimum that appeared in moderate, poor and bad status classes, which were characterized by high levels of organic/nutrient pollution. Our study validates the application of a diatom multimetric index as a good approach to classify the ecological status of temporary streams due to changes in the structure of diatom communities. We concluded that the multimetric index DIATMIB is a valuable tool for the assessment and future monitoring of temporary Mediterranean streams.

Region of influence method improves macroinvertebrate predictive models in Maryland

Stream biological assessment reflects not just conventional water quality, but an environmental quality that represents the integrity of the stream ecosystem. In Britain, Australia and the United States, macroinvertebrate predictive models were built and applied to stream assessment by employing multivariate analysis. There were variations in these models, where adaptations were made for different regions, but the philosophy underlying the models was similar: employ site classification to predict expected assemblage. Taxon assemblage is predicted from reference groups with similar stream features; these resulting models are RIVPACS-style models. Because every site has to belong to one group in the classification process, each reference group might include some dissimilar sites, and their dissimilarity in taxon assemblage impaired the results of taxon predictions from these models. To avoid this limitation, this study employed a Region-of-Influence-style (ROI-style) modeling method, selecting only similar reference sites and allowing each site to build its own reference group. Three different Region-of-Influence selection schemes were applied to improve the macroinvertebrate predictive model in Maryland: the Assessment by Nearest Neighbour Analysis (ANNA), the Burn's Region of Influence (BROI), and the New Datum Region of Influence (NROI) predictive schemes. The prediction results from ANNA, BROI, and NROI were compared, and the reference selections of each predictive scheme were examined. The comparison showed no preference for the total number of reference sites used by either predictive scheme. The number of reference sites did not correlate to the quality of reference sites used and thus does not control the predictability. The ROI-style model in Maryland had better prediction performance than the RIVPACS-style models, and could improve the bioassessment of streams.

Towards rapid bioassessment of wadeable streams in Brazil: Development of the Guapiaçu-Macau Multimetric Index (GMMI) based on benthic macroinvertebrates

We developed a multimetric index based on macroinvertebrates to assess ecological condition of wadeable streams in Rio de Janeiro State, southeast Brazil. To do so we used a set of 12 reference and 12 severely impaired sites sampled in the summer wet season as calibration sites. Nine metrics were selected based on their abilities to distinguish between reference and impaired sites and their independence from other metrics. Metrics retained in the final index were: family richness; Trichoptera family richness; Shannon family diversity; % Plecoptera individuals; % Ephemeroptera, Plecoptera and Trichoptera (EPT) individuals; % mollusk and Diptera individuals; % shredder individuals; proportion Chironomidae/Diptera individuals; and proportion Hydropsychidae/Trichoptera individuals. The final 9 metrics were standardized; then the Guapiaçu-Macau Multimetric Index (GMMI) was calculated, yielding a final index score ranging from 0 to 100. We validated the index through use of 19 sites subject to different levels of impairment, sampled both in summer and in winter. Our final multimetric index distinguished well between different levels of impairment and was strongly correlated with other disturbance measures. Temporal stability was also tested through use of 8 sites that were sampled in both seasons, and results showed no significant changes in the index final score among seasons. Our research represents one of the first initiatives in Brazil to develop a multimetric index for use in a spatially extensive bioassessment program extending across substantial natural and disturbance gradients. It can be a useful tool to help conserve streams in Atlantic forest conservation areas. We suggest that the methodology established here offers a foundation for similar index development and assessments in Brazil.

The role of live diatoms in bioassessment: a large-scale study of Western US streams

Diatom-based stream bioassessment is constantly being improved to meet the increasing demands of water quality management. This study examined whether percentage of live diatoms (PLD) in periphyton communities can be used as a metric of human disturbance in streams and rivers. The analyzed dataset (587 sites) was collected over the course of 3 years (2000–2002) from 12 Western US states (US EPA’s Environmental Monitoring and Assessment Program). The mean PLD in Western streams and rivers was low (34.50%) and highly variable (range 2.08–97.02%). It did not differ significantly between the Mountains (MT, 36.38%) and the Xeric (XE, 35.49%) ecoregions, but it was significantly (P < 0.05) greater than that in the Plains ecoregion (PL, 28.27%). PLD distinguished reference from impacted sites in the MT (P < 0.05) and somewhat in the PL (P = 0.05). However, PLD exhibited opposite patterns in the two ecoregions. It increased with human disturbance in the MT and decreased in the PL due to a potential subsidy-stress gradient of available resources. The different pattern may be largely interpreted by the quality of the reference conditions in each ecoregion. In the MT ecoregion, the selected reference sites may resemble very closely the natural state in this ecoregion. In contrast, human disturbance is much more pervasive in the low-land PL ecoregion and the “reference sites” may reflect the best attainable conditions in this ecoregion. PLD as a metric has potential for monitoring human disturbance of streams, if reference sites represent natural conditions and differing responses among regions are included in expected effects.

Which index is the best to assess stream health?

Biotic indices are widely used in monitoring the health status of various ecosystems. The choice of the best index is generally done qualitatively depending on a variety of aspects including cost and time. ROC (Receiver Operating Characteristic) methodology constitutes a valuable tool to compare objectively the diagnostic capabilities of different tests in addition to obtain decision thresholds. In this manuscript, ROC methodology is described and implemented for the first time in the context of stream bioassessment through benthic macroinvertebrates. Cut-off values that distinguish impaired from healthy sites are suggested. A new index called IBY-4 is also developed. IBY-4 accounts for the occurrence of Megaloptera, Plecoptera, Trichoptera and Elmidae in a target site and may achieve the best general performance in the study region concerning to Andean Tropical streams.

Choosing the best method for stream bioassessment using macrophyte communities: Indices and predictive models

The bioassessment and monitoring of the ecological status of rivers using macrophytes has gained new momentum since macrophytes were recognised as biological quality elements for the implementation of the European Water Framework Directive (WFD; EU/2000/60). Our objectives were to test the suitability of two predictive modelling approaches to macrophyte communities as a tool for water quality assessment, and to compare their performance with other more common approaches—the use of macrophytes as indicators of the trophic status of rivers and multimetric indices. We used floristic and environmental data that were collected in the spring of 2004 and 2005 from around 400 sites on rivers across mainland Portugal, western Iberia. We build two predictive models: MACPACS (MACrophyte Prediction And Classification System) and MAC (Macrophyte Assessment and Classification) based on RIVPACS and the BEAST methods, respectively. Whereas MACPACS is derived from taxa occurrence data, MAC uses a quantitative measure of taxa abundance. Both models showed good performance in predicting reference sites to the correct group and low rate of misclassification errors. However, they performed differently. MAC depicts a reliable response to the overall human-mediated degradation of fluvial systems, as does the multimetric index (RVI, Riparian Vegetation Index), but MACPACS presented only a poor correlation with the Global Human Disturbance Index and with the nutrients input. The incorporation of abundance data in vegetation predictive models appears to be particularly important to the detection of high levels of degradation. The values for correlations with physical–chemical pressure variables were lower than expected for MTR (Mean Trophic Rank) due to an insufficient number of scoring species found in Portuguese fluvial systems. Our results suggest that the most effective methods for bioassessment in Mediterranean-type rivers are either the RVI or the MAC predictive model.

Riparian Bird Communities as Indicators of Human Impacts Along Mediterranean Streams

Riparian areas link aquatic and terrestrial habitats, supporting species-rich bird communities, which integrate both terrestrial and aquatic processes. For this reason, inclusion of riparian birds in stream bioassessment could add to the information currently provided by existing programs that monitor aquatic organisms. To assess if bird community metrics could indicate stream conditions, we sampled breeding birds in the riparian zone of 37 reaches in 5 streams draining watersheds representing a gradient of agricultural intensity in central Italy. As a more direct indicator of water quality, stream macroinvertebrates were also sampled for computation of the Italian Extended Biotic Index (IBE). An anthropogenic index was calculated within 1 km of sampled reaches based on satellite-derived land-use classifications. Predictive models of macroinvertebrate integrity based on land-use and avian metrics were compared using an information-theoretic approach (AIC). We also determined if stream quality related to the detection of riverine species. Apparent bird species diversity and richness peaked at intermediate levels of land-use modification, but increased with IBE values. Water quality did not relate to the detection of riverine species as a guild, but two species, the dipper Cinclus cinclus and the grey wagtail Motacilla cinerea, were only observed in reaches with the highest IBE values. Small-bodied insectivorous birds and arboreal species were detected more often in reaches with better water quality and in less modified landscapes. In contrast, larger and granivorous species were more common in disturbed reaches. According to the information-theoretic approach, the best model for predicting water quality included the anthropogenic index, bird species diversity, and an index summarizing the trophic structure of the bird community. We conclude that, in combination with landscape-level information, the diversity and trophic structure of riparian bird communities could serve as a rapid indicator of stream-dwelling macroinvertebrates and, therefore, degradation of in-stream biotic integrity.

Selecting objectively defined reference sites for stream bioassessment programs

Our study develops and demonstrates an objective method for selecting reference sites for the assessment of ecological condition in freshwater ecosystems. The method uses widely available GIS data to group potential sites based on their natural environments. It then establishes the degree and types of human activities each site is exposed to prior to scoring the sites in each group by the relative amount of human activity present. Finally, the sites in each group with the least amount of human activity are categorized as reference sites, with the boundary between reference and test sites defined to maximize the distinctiveness of the two categories with respect to human activity. Application of this technique for the purpose of identifying headwater reference basins in rural areas of southwestern Ontario resulted in the classification of basins into six natural groups based on the dominant texture of the surface geology. Development of a human activity gradient indicated that basins varied according to the amount of exposure to agricultural activities with most basins having at least moderate exposure. Establishment of the reference test boundary indicated that the selected reference basins exhibited substantively lower extents of agricultural activity than test sites for most groups. Because this method uses only widely available GIS data, it enables rapid and cost-effective identification of candidate reference sites, even for large, remote, and understudied regions.

Should only live diatoms be used in the bioassessment of small mountain streams?

It is unclear whether differentiating live and dead diatoms would enhance the accuracy and precision of diatom-based stream bioassessment. We collected benthic diatom samples from 25 stream sites in the Northern Oregon Coast ecoregion. We counted live diatoms (cells with visible chloroplasts) and then compared the counts with those generated using the conventional method (clean counts). Non-metric multidimensional scaling (NMDS) showed that the diatom assemblages generated from the two counts were overall similar. The relationships between the two diatom assemblages (summarized as NMDS ordination axes) and the environmental variables were also similar. Both assemblages correlated well with in-stream physical habitat conditions (e.g., channel dimensions, substrate types, and canopy cover). The conventional diatom method provides taxonomic confidence while the live diatom count offers ecological reliability. Both methods can be used in bioassessment based on specific assessment objectives.

Assessing stream condition using macroinvertebrates and macrophytes: concordance of community responses to human impact

Freshwater bioassessment is typically based on well known taxonomic groups, such as fish and macroinvertebrates, while less studied groups are frequently excluded from bioassessment programs. We examined the community variation of stream macroinvertebrates and macrophytes (mainly aquatic bryophytes) in near-pristine reference sites and sites impacted by agriculture, forestry, or in-stream habitat alteration in western and central Finland. We specifically examined whether these taxonomic groups exhibited concordant variation in community structure and whether they could be used as surrogates for each other in stream bioassessment. The effects of human disturbance on taxonomic composition and community structure were investigated using canonical correspondence analysis and RIVPACS-type predictive models. The community concordance between macroinvertebrates and macrophytes was higher when test sites were included, suggesting that environmental degradation may increase community concordance between different biotic groups. However, despite increased concordance, macroinvertebrates and macrophytes responded to different stressors. The sensitivity of the two groups to human disturbance also differed markedly. The macroinvertebrate model determined 13 (with P t = 0+ as the threshold for species inclusion) and 12 (with P t = 0.4) out of 35 test sites as altered, whereas the macrophyte model detected impairment for only 7 sites (both probability levels), likely because of the low species richness of macrophytes in our study streams. Low richness of macrophytes may thus hinder their use in bioassessment based on taxonomic completeness, suggesting that measures that do not rely solely on species loss might perform better in at least high-gradient boreal streams.

A comparison between biotic indices and predictive models in stream water quality assessment based on benthic diatom communities

Diatoms are widely used in stream bioassessment due to their broad distribution, extraordinary variability and the ability to integrate changes in water quality. The indices Specific Polluosensitivity Index (SPI), standardized Biological Diatom Index (BDI), European Economic Community Index (CEC) and Generic Diatom Index (GDI), originally developed in France, are often applied in Portugal to evaluate stream ecological quality based on diatom communities. Alternatively, predictive models resulting from the comparison between the communities of the study site and those of a set of reference sites representing undisturbed or the best available conditions of a given region have been proposed as valuable methods for evaluating the ecological status of streams. In the present study, we applied the four above-mentioned widely used diatom-based indices (SPI, BDI, CEC and GDI) and a predictive model (MoDi) to 54 sites located in central Portugal to assess the sensitivity of the five methods to a range of anthropogenic disturbances cumulatively affecting streams and represented by 27 variables (e.g., organic enrichment, changes in morphology of the channels, integrity of the riparian corridor, land use in the catchment). The results were analyzed comparatively through Spearman correlations, Boxplots and Stepwise Discriminant Analysis. This study confirmed the sensitivity of diatoms to organic and nutrient contamination (showed by the MoDi, BDI, CEC and SPI) and revealed the importance of suspended solids (through the MoDi, GDI, and SPI). The relevance of modifications in land use to diatoms was shown by all methods applied, except for the GDI. The MoDI also revealed the importance of changes in the structure and morphology of the reach and the channel, like the construction of artificial walls or embankments and connectivity; the BDI also related its assessments with the riparian zone integrity; and the SPI was not useful in detecting morphological pressures. The GDI produced the most divergent assessments and was less effective in revealing the anthropogenic disturbances. The use of the predictive model (MoDi) is therefore a good method for the assessment of streams in central Portugal because it expresses a great diversity of quantitative and qualitative changes in freshwater systems reflected in the structure (species richness and abundance) of diatom communities.

The effect of excluding diatom taxa and reducing taxonomic resolution on multivariate analyses and stream bioassessment

The goal of the present study was to evaluate whether the Eastern Canadian Diatom Index (IDEC: Indice Diatomées de l’Est du Canada), which was built on fine-scale taxonomic resolution and included all observed diatom taxa, would perform similarly if a number of taxa were excluded from the analyses or if a reduced level of taxonomic resolution was used. The effects of excluding taxa and reducing the taxonomic resolution were evaluated by studying community structure variations in ordination analyses, and more specifically by comparing IDEC values calculated from the original diatom matrix with IDEC values obtained from simplified diatom matrices. The results showed that the exclusion of taxa based on the relative abundance criterion is the most appropriate, while the exclusion of taxa based on the frequency of occurrence criterion greatly affected the structure of the ordination. More specifically, taxa with a maximum relative abundance <2% can be excluded from the CA without markedly affecting the results. As a result, 125 taxa (40%) were excluded from the 311 taxa included in the original matrix without significantly affecting the performance of the IDEC. Excluding a greater number of taxa still allows for the distinction between impacted and reference sites, although subtle changes in the ecological status are lost. Ordinations based on presence/absence or genus-level identification resulted in a loss of information on subtle changes, but gross separation between impacted and reference sites was still possible.

Relationship of fish and macroinvertebrate communities in the mid-Atlantic uplands: Implications for integrated assessments

Indices developed for stream bioassessment are typically based on either fish or macroinvertebrate assemblages. These indices consist of metrics which subsume attributes of various species into aggregate measures reflecting community-level ecological responses to disturbance. However, little is known about the relationship between fish and macroinvertebrate metrics, or about how ecological health assessments are affected by assemblage-specific responses to disturbance. We used principal component analysis (PCA) and regression analysis of existing fish (n = 371) and macroinvertebrate (n = 442) stream bioassessment data from a multi-source dataset to determine broad scale, within-assemblage metric patterns, and to examine the intercorrelation of fish and macroinvertebrate metrics (n = 246) and their response to watershed area and land use/land cover gradients. Fish and macroinvertebrate metrics expressed as principal components (PCs) accounted for 72.4 and 85.4% of dataset variance, respectively, with PC-metric patterns reflecting aspects of stream impairment including water and habitat quality. Model components predicting fish metric response differed among fish PCs, with watershed area and macroinvertebrate metric response strongly correlated with the first fish PC, and remaining fish PC models consisting of watershed area, land use, and macroinvertebrate PCs. Correlation between fish and macroinvertebrate PCs, and models relating fish and macroinvertebrate PCs generally explained less variation (13–27%) than metric response models of fish (25–34%) and macroinvertebrates (8–38%) to watershed area and land use/land cover variables. Best-response models integrating fish and macroinvertebrate PCs, watershed area, and land use/land cover variables accounted for the greatest variation in fish PCs (32–50%) across sites. Because fish and macroinvertebrate metrics provide different information on ecological condition, integrated use of information from multiple groups may be appropriate when developing monitoring programs.

Bioassessment of streams with macroinvertebrates: effect of sampled habitat and taxonomic resolution

Two difficult decisions in the design of any bioassessment program based on stream macroinvertebrates are the number and types of habitats that should be sampled and the taxonomic level to which specimens should be identified. We used a large data set from biomonitoring of streams in the greater Sydney region, New South Wales, Australia, to compare bioassessment results obtained with an average-score-per-taxon type of biotic index among several habitats and between the taxonomic levels of family and genus. We evaluated the sensitivity of family- and genus-level indices calculated for 5 habitats (edges of pools, rocks in pools, riffles, aquatic macrophytes, and submerged wood) by considering relationships to chemical and microbial indicators of anthropogenic stress and capacity to distinguish sites affected by human activities from reference sites. Samples from rocks in pools did best on both criteria, and samples from edges of pools also did well. Genus-level index scores were slightly more strongly correlated with environmental variables than were family-level index scores, but taxonomic resolution had virtually no average effect on the degree to which samples from test sites differed from reference status, even though tolerance values for genera differed widely within some families. We attributed the weak effect of greater taxonomic resolution to the small number of identified genera in most families and the fact that many specimens could be identified to family but not to genus. The cost of discrete sampling from multiple habitats might sometimes be justified by its potential to detect habitat-specific impacts at particular sites. However, we conclude that it might be more cost-effective in broad-scale surveys to restrict sampling to the edges of pools, a habitat that occurs widely, or to consider assessment using composites of samples across multiple habitats. The small difference in sensitivity between the family- and genus-level indices suggests that, given its greater cost, bioassessment with fine-level taxonomy may be justified only in special circumstances, such as detection of subtle impacts. A tiered approach, in which only those families with wide intrafamilial variation in tolerance are identified to finer levels, is likely to be more cost-effective than identifying all taxa with fine-level resolution.

Comparing apples with apples: use of limiting environmental differences to match reference and stressor‐exposure sites for bioassessment of streams

AbstractHuman impacts on freshwater biota are often assessed by comparisons with regional reference sites judged to be free or nearly free from the influence of stressors of concern. These comparisons may involve predictive models that extrapolate from reference‐site data to estimate the biota that would occur at each assessment site in the absence of the stressors. This extrapolation often involves selection or weighting of data from particular reference sites, but it is seldom demonstrated whether this process results in a more accurate, precise and sensitive assessment than would be achieved from comparison with a simple unweighted combination of data from all reference sites, that is a null model. In addition, predictive models often rely on additive combinations of environmental predictor variables, but such combinations may poorly represent the natural control of spatial variation in biological communities by multiple limiting factors. In this paper, we describe a different type of reference‐site approach, based on the concept of limiting environmental differences (LEDs) as natural constraints on biological similarity among sites. In this method, an assessment site is compared with a sub‐set of environmentally matched reference sites selected by application of LED criteria to individual environmental variables. We illustrate this approach by its application to a potentially subtle impact: the effect of water abstraction on fish assemblages in unregulated streams in northeastern New South Wales, Australia. We compared (1) a LED‐based predictive model, (2) a null model and (3) a model based on cluster analysis and multiple discriminant analysis in the style of the widely followed River Invertebrate Prediction and Classification System (RIVPACS). The LED‐based model was about as accurate as the RIVPACS‐type model and the most precise and sensitive of the three, being best able to distinguish sites where fish assemblages departed from reference status. Copyright © 2007 John Wiley &amp; Sons, Ltd.

Integrating stream bioassessment and landscape ecology as a tool for land use planning

Summary1. Bioassessment has evolved significantly from a method of deciding whether an ecosystem exposed to stressors should ‘pass’ or ‘fail’ (or how badly it fails). Society wants some notion of what has caused any observed degradation of ecosystems, and what management strategies might improve degraded ecosystems. Managers also want to predict what negative or positive effects different land use strategies will have on the component ecosystems of a landscape, including lakes and streams.2. Here we illustrate an approach to providing these tools to managers with data from a bioassessment study of streams in the Fraser River Basin of British Columbia, Canada.3. Landscape scale descriptors of both the natural (e.g. catchment size, surficial geology) and stressor (e.g. hard rock mines, forest harvest) environment of each site were used to define the natural and stressor environments of each of the 242 stream sites.4. We classified 206 reference (relatively unexposed to human activity) sites using their benthic macroinvertebrate community composition, and then discriminated among the faunally defined groups with landscape scale descriptors of the natural environment of the sites.5. This discriminant function model allowed us to predict which group each of the test sites would be in if it were in reference condition, and then measure the relationship between the amount of human activity and the biota in each of these groups.6. These relationships were turned into projections of what will happen to a stream ecosystem's biota if the stressor environment is either improved or degraded. These projection models form the basis of evidence‐based land use planning that takes into account the health of freshwater ecosystems.

Development of a RIVPACS-type predictive model for bioassessment of wadeable streams in Wyoming

RIVPACS models produce a community-level measure of biological condition known as O/E, which is derived from a comparison of the observed (O) biota with those expected (E) to occur in the absence of anthropogenic stress. We used benthic macroinvertebrate and environmental data collected at 925 stream monitoring stations, from 1993 to 2001, to develop, validate, and apply a RIVPACS model to assess the biological condition of wadeable streams in Wyoming. From this dataset, 296 samples were identified as reference, 157 of which were used to calibrate the model, 46 to validate it, and 93 to examine temporal variability in reference site O/E-values. We used cluster analyses to group the model development reference sites into biologically similar classes of streams and multiple discriminant function analysis to determine which environmental variables best discriminated among reference groups. A suite of 14 categorical and continuous environmental variables best discriminated among 15 reference groups and explained a large proportion of the natural variability in biota within the reference dataset. Eleven of the predictor variables were derived from GIS. As expected, mean O/E-values for reference sites used in model development and validation were near unity and statistically similar. Temporal variability in O/E-values for reference sites was low. Test site values ranged from 0 to 1.45 (mean = 0.73). The model was accurate in both space and time and precise enough (S.D. of O/E-values for calibration data = 0.17) to detect modest alteration in biota associated with anthropogenic stressors. Our model was comparable in performance to other RIVPACS models developed in the United States and can produce effective assessments of biological condition over a broad, ecologically diverse region. We also provide convincing evidence that RIVPACS models can be developed primarily with GIS-based predictor variables. This framework not only simplifies the extraction of predictor variable information while potentially reducing expenditures of time and money in the collection of predictor variable information, but opens the door for development and/or application of RIVPACS models in regions where there is a paucity of local-scale, abiotic information.