• Published on

  • View

  • Download


  • ABSTRACT: Responses of the Wyoming Stream Integrity Index(WSII), a regionally calibrated multimetric index, were investigatedin relation to background elevational changes in water quality andhabitat conditions versus accelerated anthropogenic degradation atthe watershed scale. Assessments were conducted for three riversin southeast Wyoming: the Little Medicine Bow River, the MedicineBow River, and Rock Creek. Pearson correlation coefficients andregression models related core metrics and index scores to eleva-tional gradients of physicochemical variables. Velocity, substrate,and weighted habitat values were positively correlated to indexscores, while suspended solids was negatively correlated. The exclu-sive dependence of index scores on physical variables specifies thetype of environmental gradients the WSII is most robust in detect-ing. The individual core metrics Plecoptera taxa, Trichopterataxa, percent Trichoptera without Hydropsychidae, and percentnoninsects appeared most sensitive to physical changes and werethus driving associations between index scores and physical vari-ables. Despite strong correlations with physical variables, anoma-lies existed where habitat conditions were good, unknown stressorsexisted, or gradients were naturally occurring despite Poor indexscores (i.e., degraded stream conditions). Such findings illustratethe influence of regional variability on biotic indices and the impor-tance of identifying sufficient reference and impaired stream reach-es used to develop and calibrate multimetric indices relying onreference conditions.(KEY TERMS: biotic integrity; environmental indicators; nonpointsource pollution; benthic macroinvertebrates; sediment impacts;and semi-arid regions.)

    Miller, Scott W., Quentin D. Skinner, and Katta J. Reddy, 2004. Stream Assess-ments Using Biotic Indices: Responses to Physicochemical Variables. Journalof the American Water Resources Association (JAWRA) 40(5):1173-1188.


    Nonpoint source (NPS) pollution is the primarysource of water quality degradation in the UnitedStates, with siltation of aquatic environments identi-fied as the leading NPS pollutant (GPAC, 1992;USEPA, 2000). Miles of unassessed streams anddegraded stream conditions have prompted federal,state, and local monitoring of water quality to estab-lish baseline information from which to judge surfacewater quality and habitat impairments. The ubiqui-tous nature of NPS pollution has changed the focus ofthese monitoring efforts. Traditional emphasis onhuman health related pollutants has been expandedto include parameters for evaluating changes inecosystem health or biological integrity (Karr, 1993).Biological integrity is defined as the ability of a sys-tem to support and sustain a balanced, integrated,adaptive community of organisms, having a composi-tion and diversity comparable with natural habitatsof a given region (Karr et al., 1986; Gibson et al.,1996). The U.S. Environmental Protection Agency(USEPA) has responded to this new focus on biologi-cal integrity by promoting rapid bioassessment proce-dures that can be adapted to the individual needs ofstates and geographic areas. Specifically, bioassess-ments have been identified as the most effectivemeans to evaluate cumulative impacts from NPS pol-lution. Consequently, the USEPA has mandated allstates to implement biomonitoring programs by 2005(USEPA, 2002a).

    1Paper No. 03005 of the Journal of the American Water Resources Association (JAWRA) (Copyright 2004). Discussions are open untilApril 1, 2005.

    2Respectively, Graduate Research Assistant, Department of Fisheries and Wildlife, Oregon State University, Nash Hall, Room 104, Corval-lis, Oregon 97331; and Professor and Associate Professor, Department of Renewable Resources, University of Wyoming, P.O. Box 3354,Laramie, Wyoming 82072 (E-Mail/Miller:




    Scott W. Miller, Quentin D. Skinner, and Katta J. Reddy2

  • Macroinvertebrates have emerged as key biologicalindicators for use in rapid bioassessment proceduresbecause of their numerous advantages over more con-ventional monitoring tools (Lenat et al., 1981; Abel,1989; Barbour et al., 1999). A recent USEPA survey ofU.S. states, tribes, and territories revealed that 56 outof 57 in-place biomonitoring programs use aquaticmacroinvertebrates (USEPA, 2002b). Macroinverte-brates exhibit graded responses to a variety of pollu-tants, are differentially sensitive to pollutants, aresedentary, reflect localized impairments, and providefor a record of impacts over a longer period of timethan do chemical and physical analyses (Metcalfe,1989; Crunkilton and Duchrow, 1991).

    Geographically calibrated multimetric indices areone type of bioassessment tool endorsed by theUSEPA for their ability to provide objective andrepeatable methods of assessing community respons-es to environmental stressors (Karr et al., 1986; Gib-son et al., 1996; Barbour et al., 1999). Mulitmetricindices are comprised of individual metrics measuringassemblage structure, function, and pollution sensi-tivity. Individual metrics are selected based on theirability to discriminate between reference and stressedsites, as well as their ecological meaning (B.K. Jessupand J.B. Stribling, 2002, unpublished report, TetraTech., Inc.). Inherent within these indices is thatstress responses differ based on index composition(Yaun and Norton, 2003). Multimetric indices arecommonly developed for the purpose of assessingwater resources over large geographic areas, such asthe state or country level. Consequently, tests of indexefficacy, the validity of reference conditions, and useof stratification layers (e.g., ecoregions) have common-ly been conducted at large spatial scales (e.g., 105 m2).However, widespread acceptance of multimetricindices has expanded their use to the watershed scale,where conditions are assessed along elevational gradi-ents within and among stream systems.

    Here a case study is presented to evaluate a multi-metric index developed for the state of Wyoming, asemiarid region. Past studies of the WSII focusedindividually on datasets spanning the entire state ofWyoming or specific bioregions within the state (J.B.Stribling, B.K. Jessup, and J. Gerritsen, 2000, unpub-lished report, Tetra Tech., Inc.; B.K. Jessup and J.B.Stribling, 2002, unpublished report, Tetra Tech., Inc.).The goal of the study is to assess WSIIs ability to dif-ferentiate between background elevational changes inwater quality and habitat conditions versus accelerat-ed anthropogenic changes at the watershed scale.Specific objectives are to (1) identify index score andmetric responses to variability in chemical and physical parameters and (2) assess whether WSIIresponds predictably to environmental gradients andis thus robust in detecting changes in water quality or

    habitat conditions. A priori hypotheses predict exclu-sive WSII responses to physical parameters associat-ed with sediment and aquatic habitat, as opposed tovariables of water column chemistry because of dif-fuse land uses and the semiarid nature of the region.Therefore, responses of WSII index scores to variabili-ty in physical parameters were examined to identifyoutliers or site specific anomalies. Site anomalies rep-resent instances where the WSII potentially misclas-sified stream reaches as impaired or cases in which itfailed to detect changes in physical conditions. Bytesting these hypotheses one can assess WSIIs capac-ity to respond to environmental gradients and thusidentify impaired stream reaches or accelerated ratesof degradation. In addition, analysis identifiesresponses of individual core metrics, as well as thetypes of impacts and subsequent degradation WSII ismost capable of detecting. This source relationship iscritical to identifying accurate 303(d) listings requiredby the CWA as well as focusing restoration activitieson the proper causal factor of impairment.


    Study Area

    The Medicine Bow River watershed encompasses747,512 ha. Elevation varies from 1,920 to 3,600 mand is included in Omerniks (1987) Southern Rockiesand Wyoming Basins ecoregions. These correspond tothe Rockies and Wyoming Basins Bioregions used bythe Wyoming Department of Environmental QualityWater Quality Division (WDEQ-WQD). Rock Creek(RC), Little Medicine Bow River (LMB), and theMedicine Bow River (MB) drain the Medicine BowRiver watershed (Figure 1) and are classified as Class2 cold water game fisheries (WDEQ, 1998). RockCreek and the LMB River are tributaries of the MBRiver, which flows into the North Platte River. The cli-mate is cold continental and average annual meantemperature and precipitation at Medicine Bow,Wyoming, are 5.6C and 28.4 cm respectively. A snow-pack dominated hydrologic regime produces maxi-mum streamflows from April to June, which arerecorded on each stream by U.S. Geological Survey(USGS) stage recorder gauging stations. Geologic par-ent material in the LMB River drainage is composedof older Mesozoic granitic rock, and the MB River andRC drainage basins are composed of younger Protero-zoic sedimentary geologic material. Major land usesinclude sheep and cattle ranching, oil and gas devel-opment, forestry, and irrigated agriculture for grow-ing grass forage. Production of grass forage is



  • dependent upon flood irrigation, which may lowerstream peak flows during spring runoff events (Hill,1976; Brookes, 1992). Mineral exploration and urani-um mining occurred in the LMB drainage from 1955to the early 1980s (McCallum and Birch, 1980).

    Sampling Sites

    Nine sites were selected three per river toquantify responses in WSII index scores to chemical,physical, and biological parameters along each riveraxis (Figure 1). Criteria considered for site selectionincluded elevation, gradient, bioregions, geology, soils,land use, channel morphology, and ability to laterinterpret data in relationship to stream function(Stednick, 1991). Sites LMB-1, MB-1, and RC-1 werelocated in the Montane elevation zone; Sites LMB-2,MB-2, and RC-2 were located in the Foothill elevationzone; and Sites LMB-3, MB-3, and RC-3 were locatedin the Basin elevation zone following Schumms(1977) stratification of stream systems by elevation.Montane sites fall within the Rockies (Southern Rock-ies subregion) Bioregion, and Foothill and Basin sitesfall within the Wyoming Basin Bioregion. Streamchannel elevation, sinuosity, slope, and order were

    derived from USGS 1:24,000-scale topographic maps(Meador et al., 1993).


    Twenty-one chemical and physical water qualityparameters were measured from April to October of1999, 2000, and 2001. Samples were collected weeklyduring spring runoff (April through July) andbimonthly during baseflow conditions (Augustthrough October). Water quality parameters includedcalcium, magnesium, sodium, potassium, total dis-solved solids (TDS), total suspended solids (TSS), tur-bidity, carbonate, bicarbonate, fluoride, chloride,nitrate as N, orthophosphate, sulfate, total alkalinityas calcium carbonate, and hardness as calcium car-bonate. A DH-48 sampler was used to collect samplesusing the equal width increment method (Guy andNorman, 1970). Wyoming Analytical Services pro-cessed all samples according to the methods ofUSEPA (1995). Appropriate samples were fixed witheither nitric or sulfuric acid as specified by the stan-dard methods. Electrical conductivity (EC), pH, watertemperature, and dissolved oxygen were measured insitu using a YSI 610 DM multiparameter sonde (YSIInc., 1996). Discharge was calculated by measuringstream velocity with a Marsh McBirney flow meterand water depth and wetted width (Meador et al.,1993).

    Channel morphology and habitat attributes at eachsite were assessed using a 200 m section of the streamchannel following Plafkin et al. (1989). Values werecalculated for stream primary factors: channel bottomsubstrate, instream cover for fish, embeddedness, andvelocity/depth ratios. Channel shape, channel alter-ation, pool/riffle ratio, and width/depth ratio wererecorded as secondary variables. Bank vegetation pro-tection, lower bank stability, disruptive pressures,and width of riparian zone were recorded as tertiaryvariables. The USEPA procedure was modified toinclude a quantitative assessment of channel sub-strate using the Wolman (1954) pebble count method.One hundred particles were randomly selected withinthe bankfull width of the stream channel and mea-sured along the intermediate axis while walkingupstream at 45 degree angles from the respectivebank. Weighted habitat scores were calculated fromprimary, secondary, and tertiary habitat variables foreach sampling site and date using the relative impor-tance of each variable for supporting macroinverte-brates and/or fish assemblages.

    Macroinvertebrate samples were collected at eachsite during September 1999 and 2001 following theWDEQ-WQD Beneficial Use Reconnaissance Programprotocol (King, 1993). Sampling was confined to riffles



    Figure 1. Map of Study Area Illustrating LocationsWhere Chemical, Physical, and Biological

    Parameters Were Measured.

  • or runs representative of the study reach wheredepths were less than 0.61 meters and velocities wereless than 0.914 m/s. A composite of eight randomlyselected samples were collected at each site using a0.093 m2 surber sampler fitted with 500 m meshnetting and fixed with 70 percent ethanol. Sampleswere analyzed by Aquatic Biology Associates, Corval-lis, Oregon, using a minimum of 400 organisms percomposite sample and identified to genus or speciesfor better known taxa.

    Data Analysis

    Principle components analysis (PCA) reducedchemical and physical parameters from a set of inter-related variables to simplified sets of uncorrelatedcomponents (Afifi and Clark, 1997). Weighted leastsquares regression techniques were used to derive fac-tor scores for parameters within individual compo-nents. The PCA and regression analyses werefacilitated by use of the Statistical Package for theSocial Sciences (SPSS, 1998) and the Statistical Anal-ysis System (SAS Institute, 1999).

    Relationships among chemical, physical, and bio-logical data were analyzed using Pearson correlationcoefficients and simple linear regressions. Pearsoncorrelation analysis identified relationships betweenindividual parameters and biological index scores,while coefficients of determination and slope coeffi-cients quantified observed correlations. Limiteddegrees of freedom restricted the use of multipleregression techniques and multivariate analyses. Theresponse variable of metric scores was tested for inde-pen...


View more >