588WETLANDS, Vol. 22, No. 3, September 2002, pp. 588–615q2002, The Society of Wetland ScientistsHYDROLOGIC VARIABILITY AND THE APPLICATION OF INDEX OF BIOTICINTEGRITY METRICS TO WETLANDS: A GREAT LAKES EVALUATIONDouglas A. Wilcox1, James E. Meeker2, Patrick L. Hudson1, Brian J. Armitage3, M. Glen Black1, andDonald G. Uzarski4,51U. S. Geological SurveyGreat Lakes Science Center1451 Green RoadAnn Arbor, Michigan, USA 48105E-mail: [email protected] Resources ProgramNorthland CollegeAshland, Wisconsin, USA 548063Midwest Biodiversity Institute, Inc.P.O. Box 21561Columbus, Ohio, USA 43221-05614Department of ZoologyMichigan State UniversityEast Lansing, Michigan, USA 488245Present address:Annis Water Resources CenterLake Michigan Center, Grand Valley State University740 West Shoreline DriveMuskegon, Michigan, USA 49441Abstract: Interest by land-management and regulatory agencies in using biological indicators to detectwetland degradation, coupled with ongoing use of this approach to assess water quality in streams, led tothe desire to develop and evaluate an Index of Biotic Integrity (IBI) for wetlands that could be used tocategorize the level of degradation. We undertook this challenge with data from coastal wetlands of theGreat Lakes, which have been degraded by a variety of human disturbances. We studied six barrier beachwetlands in western Lake Superior, six drowned-river-mouth wetlands along the eastern shore of LakeMichigan, and six open shoreline wetlands in Saginaw Bay of Lake Huron. Plant, fish, and invertebratecommunities were sampled in each wetland. The resulting data were assessed in various forms againstgradients of human disturbance to identify potential metrics that could be used in IBI development. Ourresults suggested that the metrics proposed as potential components of an IBI for barrier beach wetlands ofLake Superior held promise. The metrics for Lake Michigan drowned-river-mouth wetlands were inconsistentin identifying gradients of disturbance; those for Lake Huron open embayment wetlands were yet moreinconsistent. Despite the potential displayed by the Lake Superior results within the year sampled, we con-cluded that an IBI for use in Great Lakes wetlands would not be valid unless separate scoring ranges werederived for each of several sequences of water-level histories. Variability in lake levels from year to yearcan produce variability in data and affect the reproducibility of data collected, primarily due to extremechanges in plant communities and the faunal habitat they provide. Substantially different results could beobtained in the same wetland in different years as a result of the response to lake-level change, with nochange in the level of human disturbance. Additional problems included limited numbers of comparablesites, potential lack of undisturbed reference sites, and variable effects of different disturbance types. Wealso evaluated our conclusions with respect to hydrologic variability and other major natural disturbancesaffecting wetlands in other regions. We concluded that after segregation of wetland types by geographic,geomorphic, and hydrologic features, a functional IBI may be possible for wetlands with relatively stablehydrology. However, an IBI for wetlands with unpredictable yet recurring influences of climate-induced,long-term high water periods, droughts, or drought-related fires or weather-related catastrophic floods or highWilcox et al., EVALUATION OF WETLAND IBI METRICS 589winds (hurricanes) would also require differing scales of measurement for years that differ in the length oftime since the last major natural disturbance. A site-specific, detailed ecological analysis of biological in-dicators may indeed be of value in determining the quality or status of wetlands, but we recommend thatIBI scores not be used unless the scoring ranges are calibrated for the specific hydrologic history pre-datingany sampling year.Key Words: biological indicators, fish, Great Lakes, human disturbance, hydrologic variability, Index ofBiotic Integrity (IBI), invertebrates, Lake Michigan, Lake Superior, plants, water-level fluctuations, wetlandsINTRODUCTIONMethods for assessing the quality of wetlands havereceived much attention in recent years. Developedspecifically for wetlands, the functional assessment ap-proach was popularized by the Wetland EvaluationTechnique (WET) (Adamus 1983, Adamus et al.1987). More recently, the Hydrogeomorphic approach(HGM) was introduced, which classifies wetlands bytype and incorporates physical and biological samplingof reference sites to assess their hydrologic, biogeo-chemical, plant community maintenance, and faunalcommunity habitat maintenance functions (Brinson1993, Brinson and Rheinhardt 1996). Alternatively, bi-ological assessment techniques in which biological or-ganisms are used as indicators of environmental healthor stress have been developed for a variety of aquaticenvironments (e.g., Karr 1981, Plafkin et al. 1989,Rankin 1989, Adamus and Brandt 1990, Rosenbergand Resh 1993, Kramer 1994, Lovett Doust et al.1994, Butterworth et al. 1995, Davis and Simon 1995),and some are presently being adapted for use in wet-lands. Adaptation has not been straightforward, how-ever, because wetland environments can differ fromother aquatic environments, both in the response tohydrologic changes and in the importance of plantcommunities as faunal habitat. Therefore, any such at-tempted adaptation must be tested successfully beforebeing implemented.The biological assessment approach receiving themost attention focuses on biological integrity, whichwas defined by Karr and Dudley (1981) as the abilityto support and maintain a balanced, integrated, adap-tive community of organisms having a species com-position, diversity, and functional organization com-parable to that of natural habitat of the region. Biolog-ical integrity is assessed using the Index of BiologicalIntegrity (IBI), a method developed by Karr (1981) inwhich fish communities were used to assess waterquality in streams of the midwestern United States(Karr 1981, 1991, Karr et al. 1986). The method hassince been extended to other aquatic ecosystems else-where (e.g., Steedman 1988, Minns et al. 1994, Dee-gan et al. 1997). A fish IBI is developed by samplingfish communities in reference streams and additionaldegraded streams that span a gradient of human dis-turbance, as determined by some external measure ofwater quality or human influence (e.g., extent of