Estimation and application of indicator values for common macroinvertebrate genera and families of the United States

Ecological Indicators 7 (2007) 22–33 This article is also available online at: www.elsevier.com/locate/ecolind

Estimation and application of indicator values for common macroinvertebrate genera and families of the United States Daren M. Carlisle *, Michael R. Meador, Stephen R. Moulton II, Peter M. Ruhl National Water Quality Assessment Program, U.S. Geological Survey, 12201 Sunrise Valley Drive, MS 413, Reston, VA, USA Received 2 June 2005; received in revised form 6 September 2005; accepted 27 September 2005

Abstract Tolerance of macroinvertebrate taxa to chemical and physical stressors is widely used in the analysis and interpretation of bioassessment data, but many estimates lack empirical bases. Our main objective was to estimate genus- and family-level indicator values (IVs) from a data set of macroinvertebrate communities, chemical, and physical stressors collected in a consistent manner throughout the United States. We then demonstrated an application of these IVs to detect alterations in benthic macroinvertebrate assemblages along gradients of urbanization in New England and Alabama. Principal components analysis (PCA) was used to create synthetic gradients of chemical stressors, for which genus- and family-level weighted averages (WAs) were calculated. Based on results of PCA, WAs were calculated for three synthetic gradients (ionic concentration, nutrient concentration, and dissolved oxygen/water temperature) and two uncorrelated physical variables (suspended sediment concentration and percent fines). Indicator values for each stress gradient were subsequently created by transforming WAs into ten ordinal ranks based on percentiles of values across all taxa. Mean IVs of genera and families were highly correlated to road density in Alabama and New England, and supported the conclusions of independent assessments of the chemical and physical stressors acting in each geographic area. Family IVs were nearly as responsive to urbanization as genus IVs. The limitations of widespread use of these IVs are discussed. Published by Elsevier Ltd. Keywords: Macroinvertebrates; Indicator values; Tolerance values; Biological assessment; National scale

1. Introduction Most ecological assessments of streams include measures of benthic macroinvertebrate assemblages (USEPA, 2002; Hering et al., 2004). Biotic indices (sensu Johnson et al., 1993) are often used to describe * Corresponding author. 1470-160X/$ – see front matter. Published by Elsevier Ltd. doi:10.1016/j.ecolind.2005.09.005

the structure of biological assemblages based on known or hypothesized tolerances of taxa to pollution. Taxon tolerance estimates are often simplified to ordinal scales or ranks known as indicator values (hereafter IVs). Macroinvertebrate IVs for organic pollution and acidification have been developed and modified throughout Europe (Sandin and Hering, 2004; Sandin et al., 2004). In the U.S., macroinvertebrate IVs for

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

organic pollution were developed in the upper Midwest (Hilsenhoff, 1987) and widely applied elsewhere, but IVs for other pollutants (e.g., acidification, sedimentation) are limited. Macroinvertebrate IVs for a wide range of pollutants offer several potential benefits to biological assessments. First, they provide a quantitative basis for assigning taxa to general tolerance classes (e.g., intolerant, intermediate, tolerant), which is required for calculating some biological metrics (Barbour et al., 1999; Hering et al., 2004) and defining gradients of biological condition (sensu Gerritsen and Leppo, 2005). Second, if unique sets of taxa are sensitive to specific pollutants, IVs may facilitate diagnosing potential causes of biological impairment (e.g., Norton et al., 2000). Last, IVs could be used to

23

expand the geographic relevance of existing biotic indices (e.g., Hilsenhoff, 1987) and create new ones for additional pollutants. The widespread application of IVs is complicated by at least two methodological issues. First, few biological and pollutant data have been collected using the same methods across large spatial scales. As a consequence, IVs have been estimated for limited geographic areas and for ‘‘general’’ pollution (e.g., Barbour et al., 1999). Although some regional classifications have recently emerged (Yuan, 2004), IVs for specific pollutants that are potentially relevant at a variety of spatial scales are clearly needed. Second, the importance of taxonomic resolution in the application of IVs is unclear. Indicator values for coarsely resolved taxa (e.g., family and genus) may be

Fig. 1. Locations in the contiguous Unites States where the U.S. Geological Survey National Water-Quality Assessment Program collected macroinvertebrate and environmental data, 1993–2003, used in this report.

24

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

more widely applicable than those for species (Sandin and Hering, 2004), but may be less effective at detecting impairment. Because consistent, specieslevel taxonomy is rarely achieved in large-scale biological assessments, the potential benefits of IVs for higher taxonomic levels (e.g., family and genus) merit investigation. We address three objectives in this report. First, we estimate IVs for common macroinvertebrate genera and families to select chemical and physical pollutants. Second, we evaluate at two regional scales whether IVs detect changes in macroinvertebrate assemblages associated with urbanization. Third, we compare the usefulness of family- and genus-level IVs for addressing objective two.

2. Methods 2.1. Study area We used data collected by the U.S. Geological Survey National Water-Quality Assessment (NAWQA) Program from 1993–2003 (Fig. 1). The NAWQA program used a targeted sampling design within a

framework of 45 major drainage basins across the continental U.S. sampling locations were selected that drained basins representative of dominant land uses (e.g., agriculture, urban), including basins minimally influenced by anthropogenic disturbances (Gilliom et al., 1995). Macroinvertebrate communities and water chemistry were sampled (within 14-day interval/ instance) on 2100 instances at 1100 sites in the contiguous U.S. Sampled locations represented a wide range of environmental settings (Table 1). 2.2. Biological sampling Benthic macroinvertebrate samples were collected from a pre-defined sampling reach using the richesttargeted habitat (RTH) protocols described in Cuffney et al. (1993) and Moulton et al. (2002). In most cases, RTH samples were collected from either woody-snags or coarse-grained, riffle substrates. Discrete woodysnag collections were taken using a saw or lopping shears, and loss of motile organisms was minimized by holding a Slack Sampler (sampling area = 0.25 m2; 500 mm mesh) on the downstream side of the snag. At least two discrete snag collections were made at each of five locations throughout the sampling reach.

Table 1 Environmental conditions of streams sampled by the NAWQA program (1993–2003) and used in this report Environmental variable Basin variables Drainage area (km2) Forest land cover (%) Urban land cover (%) Agricultural land cover (%) Reach variables Elevation (m a.s.l.) Gradient (m/km) Mean fine (<2 mm) substrates (%) Water quality measures Ammonia (mg/L) Chloride (mg/L) Dissolved O2 (mg/L) Nitrite–nitrate (mg/L) pH Total phosphorus (mg/L) Specific conductance (mg/cm at 25 8C) Sulfate (mg/L) Suspended sediments (mg/L) Water temperature (8C)

Range 4–221498 0–100 0–93 0–99 1–3032 <1–178 0–100 0.002–2.880
Median

Mean

N

162 44 2 20

4350 46 9 32

1086 1086 1086 1086

216 2 21

421 5 36

1086 1086 1852

0.030 13 8.3 0.471 7.8 0.060 339 19 5 20

0.071 28 8.3 1.398 7.7 0.187 423 58 59 20

1880 1713 1996 1880 2041 1825 2112 1733 1547 2136

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

Discrete coarse-grained riffle samples were taken using a Slack Sampler and disturbance removal technique. Five discrete Slack samples were collected throughout representative riffle habitats in the sampling reach. All samples were elutriated in the field to remove inorganic debris, composited, then fixed using 10% buffered formalin. Samples were processed (sorted and identified) at the U.S. Geological Survey’s National Water Quality Laboratory (NWQL) Biological Group. Samples were processed using a 300 organism fixed-count method described in Moulton et al. (2000). Sorted benthic macroinvertebrates were identified, in most cases, to the genus or species levels and enumerated. Labeled voucher and reference specimens were preserved in vials of 70% ethanol or mounted in CMC-10 media on slides and deposited at the NWQL. Taxonomic accuracy and consistency were maintained by verification of referenced taxa and frequent, random verification of 10% of all identifications (Moulton et al., 2000). 2.3. Chemical sampling Chemical and physical water quality constituents were sampled using standardized methods (Fishman, 1993; Shelton, 1994). Several liters of water were composited from depth- or width-integrated samples. From this composited sample, 250–1000 mL splits were extracted and preserved in the field for laboratory analyses of unfiltered nutrients, ions, and suspended sediment concentrations. Data quality was maintained through adherence to quality assurance plans and quality control samples (e.g., replicates, field blanks, spikes). Dissolved O2 and temperature were measured directly from the stream using hand held probes. Specific conductance and pH were measured from split samples. Only chemistry data collected within 14 days of biological samples were included in this analysis. 2.4. Substrate size estimation The relative proportion of fine-grained streambed substrate (percent fines <2 mm) within each sampling reach was estimated using categorical instream substrate data collected using standard NAWQA methods (Meador et al., 1993; Fitzpatrick et al.,

25

1998). For each sampling reach a set of transects was established. At several points along each transect the dominant streambed substrate was visually evaluated and assigned to one of several substrate particle-size categories. This produced data on the frequency of occurrence of each substrate size category. Percent fines was derived by combining observations of silt/ clay and sand categories and then dividing by the total number of substrate observations taken within the reach. Only habitat data collected within 120 days of biological samples were included in this analysis. We justified this temporal lag because habitat and biological sampling were conducted at base flow conditions (although sometimes several weeks apart) and in the absence of antecedent high flows (Fitzpatrick et al., 1998). 2.5. Indicator value estimation We estimated IVs for families and genera by calculating abundance-weighted averages (WAs) of chemical and physical variables using the equation: WA j ¼ ðY1 X1 þ Y2 X2 þ    þ Yn Xn Þ=ðY1 þ Y2 þ    þ Yn Þ where WAj is the WA of taxon j, Y the taxon abundance in samples 1, . . ., n, and X is the value of the environmental variable in samples 1, . . ., n. All calculations were limited to taxa present in at least 100 samples. Preliminary analyses were used to determine the effect of censoring on WA estimates because nutrient data were censored by values below reporting limits. We first calculated WAs for ammonia, nitrite–nitrate, and total phosphorus. We then substituted left-censored values with half the reporting limit (Patton and Truitt, 1992; Fishman, 1993) and recalculated WAs. Although use of leftcensored values may bias estimates of central tendency, the effect diminishes as the number of observations increases (Newman, 1995). Weighted average estimates from censored and uncensored data were highly correlated for ammonia (n = 111, R2 = 0.99), nitrite–nitrate (n = 111, R2 = 1.00), and total phosphorus (n = 108, R2 = 0.99). We concluded that WA estimates were largely unaffected by

26

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

censoring and used left-censored data for all further analyses. Because many chemical and physical pollutants cooccur, we used principal components analysis (PCA) to generate independent ‘‘stressor gradients’’ prior to calculating WAs. Principal components analysis (PCA) computes synthetic variables that are linear combinations of correlated, original variables (Stenson and Wilkinson, 2004). Concentrations of ammonia, chloride, dissolved oxygen, nitrite–nitrate, total phosphorus, and sulfate, as well as water temperature, pH, and specific conductance were included in the PCA. Logtransformations were used to stabilize variances and a varimax rotation was used in the PCA (standardized variables). Correlations of original variables to each PCA factor were used to interpret the stress gradient potentially represented by each factor. Genus- and family-specific WAs were calculated for each PCA factor (with eigenvalue > 1). Suspended sediment concentrations and percent fines were uncorrelated with all other variables, so they were considered independent stressors for which taxa WAs were calculated. Although not used in this study, we also calculated genus and family WAs for all original chemical and physical variables, and these are available on the internet: http://water.usgs.gov/nawqa/ecology/. We assigned family and genus IVs for each synthetic stress gradient (PCA factor), suspended sediment, and percent fines, by transforming WAs into ordinal ranks. We used a 10-point ordinal classification (rank) scale because it coincided with existing tolerance classification schemes (e.g. Barbour et al., 1999). For each stressor, the tolerance class (1–10) of each taxon was determined based on the value of its WA and the percentiles of WA values across all taxa. Ranks were ordered such that tolerance increased from 1 to 10. 2.6. Bioassessment application We used independent data sets collected in New England and Alabama to evaluate whether IVs could be useful for bioassessment. We used data collected in 2000 as part of two studies to assess the effects of urbanization on stream ecosystems. Sites were selected that varied in the degree of watershed urbanization but were otherwise (e.g., stream size and physiography) as similar as possible (Coles et al.,

2004; Cuffney et al., 2005). For each site we calculated mean IVs (not abundance-weighted) for each stressor across all genera and families collected at that site. Indicator values were available for 87 of 117 genera (56 of 65 families) collected in New England and 86 of 108 genera (52 of 61 families) collected in Alabama. We compared the strength of association between a measure of urbanization (GISderived estimates of basin road density, Cuffney et al., 2005) and IVs using linear regression.

3. Results Principal components analysis revealed that many water quality variables were correlated (Table 2). Three PCA factors explained >70% of the total variation among nine chemical and physical variables. Factor one was composed of specific conductance, pH, and sulfate concentrations. The second factor was composed of nutrients and chloride. Water temperature and dissolved oxygen concentration were represented on the third factor. There were sufficient data to calculate IVs for 102 genera and 67 families (Appendix 1 and 2) for each synthetic stress gradient, suspended sediment concentration, and percent fines. Mean IVs of most stressors were highly correlated with road density (Table 3). In New England, road density was associated with 23–72% of the variation in genus IVs, and 56–75% of the variation in family IVs. Family IVs for dissolved oxygen/temperature were not related to road density in New England. In Alabama, road density was associated with 16–51% of the Table 2 Loadings of factors from principal components analysis (varimax rotation) on original nine water quality variables Variable

Factor 1

Factor 2

Factor 3

Specific conductance Sulfate pH Ammonia Chloride Nitrite–nitrate Total phosphorus Water temperature Dissolved oxygen

0.832 0.744 0.811 0.189 0.555 0.258 0.281 0.304 0.263

0.382 0.405 0.220 0.786 0.601 0.702 0.667 0.017 0.174

0.125 0.017 0.077 0.199 0.235 0.231 0.276 0.794 0.831

Variance explained (%)

28

26

17

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

27

Table 3 Regression statistics of mean indicator values regressed on road density in New England and Alabama, 2000 Indicator value

New England

Alabama

F

P

R2

F

P

R2

Genus Ions Nutrients DO/Temp Suspended sediment Percent fines

56.07 74.77 9.53 44.57 59.67

<0.00 <0.00 <0.00 <0.00 <0.00

0.66 0.72 0.23 0.61 0.68

8.07 30.15 0.23 6.47 15.54

0.01 <0.00 0.64 0.02 <0.00

0.21 0.51 0.00 0.16 0.34

Family Ions Nutrients DO/Temp Suspended sediment Percent fines

57.67 87.19 0.58 37.43 63.69

<0.00 <0.00 0.45 <0.00 <0.00

0.67 0.75 0.00 0.56 0.69

19.22 31.39 1.13 30.66 22.91

<0.00 <0.00 0.30 <0.00 <0.00

0.39 0.52 0.00 0.51 0.44

DO/Temp: dissolved oxygen/temperature. N = 30 for each geographic area.

variation in genus IVs, and 39–52% of the variation in family IVs. Neither genus nor family IVs for dissolved oxygen/temperature was related to road density in Alabama.

4. Discussion Pollutant tolerances of macroinvertebrate taxa are widely invoked in the analysis and interpretation of biological assessment data, but many tolerance estimates lack empirical foundation. To our knowledge this is the first publication of empirically derived IVs for the common macroinvertebrates of the U.S. based on consistent, nationwide sampling. Because our IVs were derived from data collected at a broad spatial scale, they may be sufficiently robust to apply at regional scales, as evidenced by tests with two independent data sets. Additional evaluations of their general applicability are required, but the IVs reported here should enhance tolerance classifications for many macroinvertebrate taxa. Application of national-scale IVs may enhance the detection and diagnosis of anthropogenic stressors at regional scales. Previous studies (Coles et al., 2004; Cuffney et al., 2005) in New England and Alabama revealed that mayfly (Ephemeroptera) and stonefly (Plecoptera) taxa richness declined in urbanized basins. Reanalysis of these data using IVs provided additional insights because variation in assemblage structure was

examined relative to specific stressor tolerances. Our results showed that stream macroinvertebrate assemblages in urbanized basins were dominated by taxa (e.g., Hydropsyche spp., Polypedilum spp., Caecidotea spp., and Sialis spp.) that are tolerant of many of the stressors we examined. Previous studies in New England (Coles et al., 2004) suggested that non-point runoff and wastewater effluent contributed nutrients and ions to urbanized streams, and that reductions in riparian shading and increased numbers of small impoundments were the dominant habitat alterations, which led to increases in stream water temperatures (Paul and Meyer, 2001). Observed losses of fluvialspecialist fish species in the same streams (Meador et al., 2005) also suggests that flow alteration was a cause of biological impairment. In Alabama, nutrients appeared to be more strongly associated with biological data than other potential stressors, which corroborates conclusions by Meador et al. (2005) that nutrients were likely the major cause of stream impairment in these streams. The apparently weak correlations between mean IVs and land use in Alabama streams was likely caused by drought conditions during the year in which sampling occurred (Cuffney et al., 2005). We acknowledge that other stressors (e.g., exotic species, and pesticides) that we did not address may have also influenced biological assemblages in both regions. Although we cannot definitively show which proximal chemical and physical stressors were responsible for biological degradation in New England and Alabama

28

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

streams, the application of national-scale IVs for a subset of possible stressors provided direct linkages between observed changes in assemblage structure and independent assessments of chemical and physical habitat. Taxonomic resolution is often a consideration in the analysis of macroinvertebrate data because identifications are dependent on specimen quality, maturity, and availability of appropriate keys. The appropriate taxonomic resolution may be related to the analysis method and spatial scale of study (Resh and McElravy, 1993). For example, family-level resolution may be equal to or better than genus-level resolution at detecting environmental gradients in multivariate ordinations of abundance data (Bowman and Bailey, 1997; Reynoldson et al., 2001), but less sensitive in other types of analyses (Hawkins et al., 2000; Schmidt-Kloiber and Nijboer, 2004; Waite et al., 2004). Few studies have examined the importance of taxonomic resolution in the application of IVs. Two European studies showed that familylevel IVs performed comparably to species-level IVs at regional spatial scales (Zamora-Mufioz and AlbaTercedor, 1996), but were less sensitive to degradation at smaller scales where species-level tolerance values were available (Sandin and Hering, 2004). Our results are in general agreement, and suggest that family-level IVs may be as effective as genus-level IVs in some bioassessment applications, but exclusive use of family-level IVs may be inadequate to detect some types of stressors. We did not compare the IVs in this report to published tolerance values because such comparisons would likely be ambiguous for at least two reasons. First, most published tolerance values (e.g., Armitage et al., 1983; Barbour et al., 1999; Bode et al., 2002; Yuan, 2004) were developed for specific geographic areas or for generalized pollutants so they are probably not comparable to our nationwide, stressor-specific IVs. Second, published IVs in the U.S. have been derived using a variety of methods (including best professional judgment), most of which are unpublished. We suspect that IV estimates vary for this reason alone, but the effects of methodological differences on IV estimates are largely unknown. We believe there is a great need for developing comparable IVs for macroinvertebrate taxa at a variety of spatial and taxonomic scales.

There are limitations to the widespread use of the IVs we report. First, our WA estimates are biased by the sampling design. Ideally, empirically based IVs would be estimated based on a random sampling design stratified by the known geographic range and expected habitats of each taxon. We suggest, however, that IVs based on ranks may be more robust than WAs to this potential bias. A second limitation is that our IVs may be ineffective for assessing stressors at local spatial scales because our estimates integrate tolerances among multiple genera and species across the country. High variation in pollution sensitivities within a family or genus will increase the uncertainty associated with IV estimates. Analyses of interspecific variation in tolerance are rare, but experimental (Clements, 2004) and observational (Carlisle and Clements, 2003) studies in Rocky Mountain streams documented substantive differences in metal sensitivity between two species of Drunella spp. and two species of Baetis spp. Research that assesses interspecific and inter-genus variability of pollution tolerance would strengthen the confidence with which specific genus- and family-level IVs are applied over large spatial scales. An additional weakness of our IVs is that, like all observational data, estimates probably include the confounding effects of unmeasured variables. Because multiple chemical, physical, and biological stressors frequently co-occur in nature, experimentation is required to reliably attribute causality to a single stressor. The need for experimentally derived IVs (sensu Clements, 2004) cannot be overstated, but such estimates are extremely rare. Lastly, our IV estimates are limited to a subset of all possible stressors that exist in streams. Indicator value estimates for a variety of additional chemical, physical, and biological stressors are clearly needed. Yuan’s (2004) tolerance values for habitat condition are a notable first step.

Acknowledgements We thank the many past and present biologists with the NAWQA Program who diligently collected the data used in this report, especially H. Zappia in Alabama and J. Coles in New England. Timely colleague reviews by R.E. Zuellig and M.D. Bilger improved the manuscript and are appreciated. Two anonymous reviewers

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

29

Appendix 1 Genus indicator values, alphabetized by order, family, and genus, for five stressors. See Table 2 and text for descriptions of stressors. SS: suspended sediment. Order

Family

Genus

Ions

Nutrients

DO/Temp

SS

Fines

Amphipoda

Gammaridae Hyalellidae Ancylidae Physidae Elmidae

Gammarus Hyalella Ferrissia Physa Ancyronyx Dubiraphia Macronychus Microcylloepus Optioservus Oulimnius Stenelmis Dineutus Psephenus Orconectes Atherix Ablabesmyia Brillia Cardiocladius Chironomus Cladotanytarsus Corynoneura Cricotopus Cryptochironomus Dicrotendipes Eukiefferiella Glyptotendipes Micropsectra Microtendipes Nanocladius Nilotanypus Pagastia Parachironomus Parakiefferiella Parametriocnemus Paratanytarsus Pentaneura Phaenopsectra Polypedilum Potthastia Pseudochironomus Rheocricotopus Rheotanytarsus Stempellinella Stenochironomus Sublettea Synorthocladius Tanytarsus Thienemanniella Tribelos Tvetenia Hemerodromia

8 7 9 10 1 10 2 9 9 3 8 1 5 9 5 2 2 8 10 8 1 10 8 3 6 9 2 7 9 7 5 2 5 7 3 9 4 6 4 8 1 3 3 1 4 4 4 7 1 4 7

10 6 4 8 10 5 10 7 5 6 7 9 2 5 2 8 4 5 9 6 9 9 9 10 4 8 9 8 9 5 3 10 10 6 10 6 9 9 2 5 10 8 1 10 2 4 7 6 7 5 7

1 10 9 2 3 6 7 9 2 4 7 10 6 6 5 10 2 2 2 5 6 4 6 8 1 9 1 4 8 9 1 9 1 3 6 9 4 7 3 10 9 3 3 10 5 2 7 5 10 1 7

10 7 9 5 9 9 9 7 4 1 7 10 3 4 6 5 3 2 2 6 7 7 6 8 2 10 2 7 10 6 4 9 8 4 7 5 3 7

6 9 5 8 10 8 10 5 6 2 4 9 1 7 4 9 7 5 5 10 10 8 9 10 5 10 5 6 10 7 4 9 10 6 8 8 7 8 4 7 9 9 3 10 3 3 9 8 10 3 8

Basommatophora Coleoptera

Decapoda Diptera

Gyrinidae Psephenidae Cambaridae Athericidae Chironomidae

Empididae

3 8 9 2 9 1 5 5 6 4 10

30

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

Appendix 1 (Continued ) Order

Ephemeroptera

Family

Genus

Ions

Nutrients

DO/Temp

SS

Fines

Simuliidae Tipulidae

Simulium Antocha Hexatoma Tipula Acentrella Baetis Fallceon Plauditus Caenis Drunella Serratella Epeorus Heptagenia Leucrocuta Rhithrogena Stenacron Stenonema Isonychia Tricorythodes Paraleptophlebia Rhagovelia Caecidotea Petrophila Corydalus Nigronia Sialis Elimia Argia Leuctra Acroneuria Perlesta Pteronarcys Brachycentrus Micrasema Glossosoma Helicopsyche Ceratopsyche Cheumatopsyche Hydropsyche Hydroptila Lepidostoma Nectopsyche Oecetis Chimarra Polycentropus Psychomyia Rhyacophila Neophylax Corbicula Pisidium Sphaerium

6 4 3 5 6 8 10 5 9 2 4 1 8 6 3 9 4 5 10 1 8 10 10 5 2 7 5 6 2 2 1 3 3 6 2 7 6 6 8 10 4 10 7 5 3 7 1 6 5 10 9

8 3 1 8 3 4 5 4 3 1 3 1 6 5 1 7 3 2 6 2 7 10 6 1 4 7 1 2 3 2 8 3 2 4 1 3 5 8 8 7 1 4 6 5 2 3 1 1 4 8 7

3 4 3 6 3 3 10 5 9 1 2 1 5 9 1 8 8 8 8 4 10 5 4 10 6 9 7 10 5 5 2 3 4 2 2 6 4 8 7 6 1 8 8 8 7 5 1 4 9 7 5

5 3 4 5 2 5 8 6 4 1 4 1 10 10 2 10 5 6 9 1 7 2 8 4 3 8 3 6 2 1 6 10 3

7 3 2 8 5 4 9 4 6 2 2 1 10 4 1 8 4 2 9 2 6 7 2 3 3 5 6 6 1 1 8 7 3 4 3 3 4 6 8 6 1 9 2 1 5 5 1 2 7 6 7

Baetidae

Caenidae Ephemerellidae Heptageniidae

Hemiptera Isopoda Lepidoptera Megaloptera

Mesogastropoda Odonata Plecoptera

Trichoptera

Isonychiidae Leptohyphidae Leptophlebiidae Veliidae Asellidae Pyralidae Corydalidae Sialidae Pleuroceridae Coenagrionidae Leuctridae Perlidae Pteronarcyidae Brachycentridae Glossosomatidae Helicopsychidae Hydropsychidae

Hydroptilidae Lepidostomatidae Leptoceridae

Veneroida

Philopotamidae Polycentropodidae Psychomyiidae Rhyacophilidae Uenoidae Corbiculidae Sphaeriidae

1 8 5 6 9 7 1 8 4 2 2 1 1 2 8 9

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

31

Appendix 2. Family indicator values, alphabetized by order, for five stressors. See Table 2 and text for descriptions of stressors. SS: suspended sediment. Order

Family

Ions

Nutrients

DO/Temp

Amphipoda

Gammaridae Hyalellidae Erpobdellidae Ancylidae Physidae Planorbidae Dytiscidae Elmidae Gyrinidae Hydrophilidae Psephenidae Cambaridae Athericidae Ceratopogonidae Chironominae Diamesinae Empididae Orthocladiinae Simuliidae Tabanidae Tanypodinae Tipulidae Enchytraeidae Baetidae Caenidae Ephemerellidae Heptageniidae Isonychiidae Leptohyphidae Leptophlebiidae Corixidae Veliidae Asellidae Pyralidae Lumbriculidae Corydalidae Sialidae Hydrobiidae Pleuroceridae Aeshnidae Calopterygidae Coenagrionidae Gomphidae Chloroperlidae Leuctridae Nemouridae Perlidae Perlodidae Pteronarcyidae Glossiphoniidae Brachycentridae Glossosomatidae

8 7 10 7 10 9 2 7 1 5 6 6 6 5 4 3 6 8 4 3 4 5 9 8 9 2 3 6 10 6 9 8 8 10 3 3 7 10 4 1 2 7 4 1 1 1 2 2 2 9 4 2

10 7 10 6 9 6 9 7 9 10 3 7 2 5 8 4 7 7 8 3 8 4 10 5 3 2 3 2 5 3 9 7 10 6 5 3 8 9 6 10 5 6 2 1 3 1 4 2 1 8 3 1

1 10 5 9 1 8 4 5 10 7 6 5 4 10 5 2 7 2 2 5 8 3 6 3 9 2 6 8 8 4 10 10 6 3 4 9 9 4 8 7 7 9 8 1 4 1 3 1 3 10 3 2

Arhynchobdellae Basommatophora

Coleoptera

Decapoda Diptera

Enchytraeida Ephemeroptera

Hemiptera Isopoda Lepidoptera Lumbriculida Megaloptera Mesogastropoda Odonata

Plecoptera

Rhynchobdellae Trichoptera

SS 9 8 5 7 3 9 6 10 10 3 7 5 7 8 4 10 5 4 2 6 5 5 7 4 1 4 5 10 6 10 6 5 7 1 4 9 7 3 2 10 7 3 2 1 3 2 4 3 1

Fines 3 7 9 5 10 5 8 5 10 9 1 6 3 7 9 6 9 8 8 7 8 4 10 5 5 2 4 2 8 3 6 5 9 2 4 2 4 3 9 10 10 7 2 1 1 3 3 1 5 6 4 3

32

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33

Appendix 2 (Continued ) Order

Tubificida Veneroida

Family Helicopsychidae Hydropsychidae Hydroptilidae Lepidostomatidae Leptoceridae Limnephilidae Philopotamidae Polycentropodidae Psychomyiidae Rhyacophilidae Uenoidae Naididae Tubificidae Corbiculidae Sphaeriidae

Ions 7 7 9 4 7 1 5 3 5 1 5 10 10 5 9

provided valuable advice for improving the clarity and relevance of this work.

References Armitage, P.D., Moss, D., Wright, J.F., Furse, M.T., 1983. The performance of a new water quality score system based on macroinvertebrates over a wide range of unpolluted runningwater sites. Water Res. 17, 333–347. Barbour, M.T., Gerritsen, J., Snyder, B.D., Stribling, J.B., 1999. Rapid Bioassessment Protocols for Use in Streams and Wadeable Rivers: Periphyton, Benthic Macroinvertebrates, and Fish, EPA 841-B-99-002. U.S. Environmental Protection Agency, Washington, DC. Bode, R.W., Novak, M.A., Abele, L.E., Heitzman, D.L., Smith, A.J., 2002. Quality Assurance Work Plan for Biological Stream Monitoring in New York State. Bureau of Water Assessment and Management, Division of Water, New York State Division of Environmental Conservation, Albany, NY. Bowman, M.F., Bailey, R.C., 1997. Does taxonomic resolution affect the multivariate description of the structure of benthic macroinvertebrate communities? Can. J. Fish. Aquat. Sci. 54, 1802–1807. Carlisle, D.M., Clements, W.H., 2003. Growth and secondary production of aquatic insects along a gradient of Zn contamination in Rocky Mountain streams. J. North Am. Benth. Soc. 22, 582–597. Clements, W.H., 2004. Small-scale experiments support causal relationships between metal contamination and macroinvertebrate community responses. Ecol. Appl. 14, 954–967. Coles, J.F., Cuffney, T.F., McMahon, G, Beaulieu, K.M., 2004. The effects of urbanization on the biological, physical, and chemical characteristics of coastal New England streams. U.S. Geological Survey Professional Paper 1695.

Nutrients 3 8 7 1 6 5 5 5 4 1 1 9 10 4 9

DO/Temp

SS

Fines

5 7 6 1 7 2 7 10 4 1 2 3 5 9 8

9 9 10 1 5 6 2 9 2 1 1 8 8 2 7

2 7 7 1 4 7 2 8 7 1 1 10 10 6 5

Cuffney, T.F., Gurtz, M.E., Meador, M.R., 1993. Methods for collecting benthic invertebrate samples as part of the National Water-Quality Assessment Program. U.S. Geological Survey Open-File Report 93-406. Available at http://water.usgs.gov/ nawqa/ecology/. Cuffney, T.F., Zappia, H., Giddings, E.M.P., Coles, J.F., 2005. Effects of urbanization on benthic macroinvertebrate assemblages in contrasting environmental settings: Boston, MA; Birmingham, AL; and Salt Lake City, UT. In: Brown, L.R., Gray, R.H., Hughes, R.M., Meador, M.R. (Eds.), Effects of Urbanization on Stream Ecosystems. American Fisheries Society, Bethesda, MD. Fishman, M.J., 1993. Methods of analysis by the U.S. Geological Survey National Water Quality Laboratory: Determination of inorganic and organic constituents in water and fluvial sediments. U.S. Geological Survey Open-File Report 93-125. Fitzpatrick, F.A., Waite, I.R., D’Arconte, P.J., Meador, M.R., Maupin, M.A., Gurtz, M.E., 1998. Revised methods for characterizing stream habitat in the National Water-Quality Assessment Program. U.S. Geological Survey, Water-Resource Investigations Report 98-4052. Gerritsen, J., Leppo, E.W., 2005. Biological condition gradient for tiered aquatic life use in New Jersey. Report to US Environmental Protection Agency. Tetra Tech, Inc., Owings Mill, MD. Gilliom, R.J., Alley, W.M., Gurtz, M.E., 1995. Design of the National Water-Quality Assessment Program: Occurrence and distribution of water-quality conditions. U.S. Geological Survey Circular 1112, Sacramento, CA. Hawkins, C.P., Norris, R.H., Hogue, J.N., Feminella, J.W., 2000. Development and evaluation of predictive models for measuring the biological integrity of streams. Ecol. Appl. 10, 1456– 1477. Hering, D., Moog, O., Sandin, L., Verdonschot, P.F.M., 2004. Overview and application of the AQEM assessment system. In: Hering, D., Verdonschot, P.F.M., Moog, O., Sandin, L. (Eds.), Integrated Assessment of Running Waters in Europe. Hydrobiology 516, 1–20.

D.M. Carlisle et al. / Ecological Indicators 7 (2007) 22–33 Hilsenhoff, W.L., 1987. An improved biotic index of organic stream pollution. Great Lakes Entom. 20, 31–39. Johnson, R.K., Wiederholm, T., Rosenberg, D.M., 1993. Freshwater biomonitoringusing individual organisms, populations, and species assemblages of benthic macroinvertebrates. In: Rosenberg, D.M., Resh, V.H. (Eds.), Freshwater Biomonitoring and Benthic Macroinvertebrates. Chapman & Hall, New York, pp. 40–158. Meador, M.R., Hupp, C.R., Cuffney, T.F., Gurtz, M.E., 1993. Methods for characterizing stream habitat as part of the National Water-Quality Assessment Program. US Geological Survey Open-File Report 93-408. Meador, M.R., Coles, J.F., Zappia, H., 2005. Fish assemblage responses to urban intensity gradients in contrasting metropolitan areas: Birmingham, AL and Boston, MA. In: Brown, L.R., Gray, R.H., Hughes, R.M., Meador, M.R. (Eds.), Effects of Urbanization on Stream Ecosystems. American Fisheries Society, Bethesda, MD. Moulton II, S.R., Carter, J.L., Grotheer, S.A., Cuffney, T.F., Short, T.M., 2000. Methods of analysis by the US Geological Survey National Water Quality Laboratory: processing, taxonomy, and quality control of benthic macroinvertebrate samples. US Geological Survey Open-File Report 00-212. Available at http:// water.usgs.gov/nawqa/ecology/. Moulton II, S.R., Kennen, J.G, Goldstein, R.M., Hambrook, J.A., 2002. Revised protocols for sampling algal, invertebrate, and fish communities as part of the National Water-Quality Assessment Program. US Geological Survey Open-File Report 02-150. Available at http://water.usgs.gov/nawqa/ecology/. Newman, M.C., 1995. Quantitative Methods in Aquatic Ecotoxicology. Lewis Publishers, Boca Raton, LA. Norton, S.B., Cormier, S.M., Smith, M., Jones, R.C., 2000. Can biological assessments discriminate among types of stress? A case study for the eastern cornbelt plains ecoregion. Environ. Toxicol. Chem. 19, 1113–1119. Patton, C.J., Truitt, E.P., 1992. Methods of analysis by the US Geological Survey National Water Quality Laboratory: determination of ammonium plus organic nitrogen by a Kjeldahl digestion method and an automated photometric finish that includes digest cleanup by gas diffusion: US Geological Survey Open-File Report 00-170. Paul, M.J., Meyer, J.L., 2001. Streams in the urban landscape. Annu. Rev. Ecol. Syst. 32, 333–365. Resh, V.H., McElravy, E.P., 1993. Contemporary quantitative approaches to biomonitoring using benthic macroinvertebrates. In: Rosenberg, D.M., Resh, V.H. (Eds.), Freshwater Biomoni-

33

toring and Benthic Macroinvertebrates. Chapman & Hall, New York, pp. 159–194. Reynoldson, T.B., Rosenberg, D.M., Resh, V.H., 2001. Comparison of models predicting invertebrate assemblages for biomonitoring in the Fraser River catchment, British Columbia. Can. J. Fish. Aquat. Sci. 58, 1395–1410. Sandin, L., Hering, D., 2004. Comparing macroinvertebrate indices to detect organic pollution across Europe: a contribution to the EC Water Framework Directive intercalibration. In: Hering, D., Verdonschot, P.F.M., Moog, O., Sandin, L. (Eds.), Integrated Assessment of Running Waters in Europe. Hydrobiology 516, 55–68. Sandin, L., Dahl, J., Johnson, R.K., 2004. Assessing acid stress in Swedish boreal and alpine streams using benthic macroinvertebrates. In: Hering, D., Verdonschot, P.F.M., Moog, O., Sandin, L. (Eds.), Integrated Assessment of Running Waters in Europe. Hydrobiology 516, 129–148. Schmidt-Kloiber, A., Nijboer, R.C., 2004. The effect of taxonomic resolution on the assessment of ecological water quality classes. In: Hering, D., Verdonschot, P.F.M., Moog, O., Sandin, L. (Eds.), Integrated Assessment of Running Waters in Europe. Hydrobiology 516, 269–283. Shelton, L.R., 1994. Field guide for collecting and processing stream-water samples for the national water-quality assessment program. U.S. Geological Survey Open File Report 94-455, Sacramento, CA. Stenson, H., Wilkinson, L., 2004. Factor Analysis. SYSTAT11. Statistics I. SYSTAT Software, Inc., Richmond, CA. US Environmental Protection Agency, 2002. Summary of Biological Assessment Programs and Biocriteria Development for States, Tribes, Territories, and Interstate Commissions: Streams and Wadeable Rivers, EPA-822-R-02-048. Environmental Protection Agency, Washington, DC. Waite, I.R., Herlihy, A.T., Larsen, D.P., Scott, N., Urquhart, S., Klemm, D.J., 2004. The effects of macroinvertebrate taxonomic resolution in large landscape bioassessments: an example from the Mid-Atlantic Highlands, USA. Freshwater Biol. 49, 474– 489. Yuan, L.L., 2004. Assigning macroinvertebrate tolerance classifications using generalized additive models. Freshwater Biol. 49, 662–677. Zamora-Mufioz, C., Alba-Tercedor, J., 1996. Bioassessment of organically polluted Spanish rivers, using a biotic index and multivariate methods. J. North Am. Benth. Soc. 15 (3), 332–352.