Development of Wetland Bioassessment Methods in Wisconsin

Last Updated: March 2000

Contact Information

Dick Lillie
Wisconsin Department of Natural Resources
Bureau of Integrated Science Services
1350 Femrite Drive
Monona, WI 53716
Phone: (608) 221-6338
Email: LILLIR@dnr.state.wi.us

Purpose(s) of Project

• Develop a Biotic Index for Wisconsin's palustrine wetlands.
• Compare performance of one plant and two macroinvertebrate multimetric indices.
• Develop biological integrity rating system for classifying wetlands

Project History

Field studies for this project were conducted during the spring and summer of 1998, with laboratory analysis and data synthesis completed the following year. Funding was provided by a grant from the EPA's Region 5. A final report to EPA (Wetland Grant #CD985491-01-0) was prepared, and an article is being prepared for journal submission. The findings form the basis for continuing work, a second EPA-funded grant, in 2000 to refine and further evaluate the preliminary indices and expand communities covered to include amphibians, small mammals, diatoms, and zooplankton.

Study Design

During the early spring and summer of 1998, 104 palustrine depression wetlands distributed across the major ecoregions of Wisconsin were sampled. Study sites included a mixture of least-disturbed reference basins and degraded or restored wetlands. Study sites represented a range of vegetative cover types, water chemistries, and water duration.

Assemblages Monitored

• Macroinvertebrates
• Plants

Sampling Methods: Macroinvertebrates

We sampled macroinvertebrates early in the spring to minimize influences of immigration-emigration. Three different field collection procedures were evaluated and two laboratory approaches were used. On all 104 wetlands, we collected two sets of three standard D-frame net sweeps of approximately one meter length each. Sampling stations less than 60 cm deep were established at equally spaced points around the wetland perimeter that approximately trisected the basin (and assured coverage of the major plant communities present. The first set of net sweeps was concentrated (large coarse materials were rinsed, examined, and removed) into a one quart container. The second set of net sweeps was not field-concentrated, but rather was placed directly into a one-gallon plastic bag. Both sets were preserved in ethanol and returned to the laboratory for processing. The third set (used on a subset of 17 wetlands) of D-frame net sweep samples was placed on a coarse wire screen over a collection basin for a period of approximately 10 minutes. Organisms falling-or crawling-through the screen and entering the collection container were collected and preserved as above. This last set represented a "clean" sample that was much easier to sort and process than the standard samples.

Laboratory Methods: Macroinvertebrates

Macroinvertebrates were processed using a two-tiered approach. The first stage consisted of a fixed, 100-count sample (sensu Hilsenhoff Biotic Index procedures) using a grid-marked tray with 24 cells. Organisms were picked and sorted at a coarse taxonomic level, usually order or family only. Following completion of the 100-count sample, we processed the balance of the sample in its entirety, except for subsampling dominant taxa. The unconcentrated "bag" samples proved to be too large to process in an economical fashion, so only the complete set of 104 "field-concentrated" samples were processed. The 17 "screened" samples generally contained less than 100 total organisms and were processed completely.

Analytical Methods: Macroinvertebrates

We used SYSTAT (SPSS 1997) to perform all statistical and graphical analyses. Percentage data were transformed using the arc-sine, square-root transformation, and abundance data were either log-transformed or power-transformed as applicable. Metric development was based on a series of visual comparisons of community attribute responses to suspected measures of disturbance using box plots and jittered-dot density plots. Those attributes that exhibited evidence of separation between reference wetland conditions and wetlands suspected to be impacted by human disturbance were selected as potential metrics. Attributes that exhibited inconsistent or overlapping responses between impacted and reference systems were eliminated from further consideration.

Sampling Methods: Plants

We conducted simplified plant surveys during July 1998 using a combination of techniques. These included a subjective estimate of cover and an objective survey of percent cover and frequency of occurrence within six equidistantly spaced 20 by 50 cm rectangular quadrats positioned along each of three transects that trisected the wetland basin, for a total of 18 quadrats per wetland.

Laboratory Methods: Plants

No biomass or stem counts were performed. Voucher specimens were pressed and identified to species when possible, but most plant metrics were based on a coarser taxonomic level.

Analytical Methods: Plants

We developed the Wisconsin Wetland Plant Biotic Index (WWPBI) using the same procedures described for the two macroinvertebrate-based indices. Because we wanted to develop a practical tool for managers with limited botanical expertise, we lumped taxa at various taxonomic levels (e.g., family, genus) or structural groups (e.g., grass-like, emergents) for analysis. Importance values (average of percent cover and frequency of occurrence) were used as the attribute of concern for family-genus-species levels, and percent cover was used for emergent, submergent, floating-leafed, and open water attributes.

Other Data Collected

We also collected associated physical and chemical data on each wetland, including pH, alkalinity, conductivity, color, temperature, clarity, and depth. Riparian cover type within a 100-foot buffer area surrounding each wetland was subjectively estimated and recorded, as well as shade canopy cover.

Preliminary Findings

Three multimetric indices (two macroinvertebrate and one plant index) were developed. The Wisconsin Wetland Macroinvertebrate Index (WWMI) is a multimetric index based on 15 metrics derived from a total count of organisms in three composited net sweeps. A total of 69 community attributes were evaluated. The WWMI is composed of 12 abundance metrics, two richness metrics, and one percentage metric. Abundance metrics include mollusks, annelids, fairy shrimp, damselflies, pigmy backswimmers, water boatmen, limnephild caddisflies, total caddisflies, phantom midges, mosquitoes, soldier flies, and total invertebrates. Richness metrics are noninsects, and total taxa. The percent caddisflies is the only percentage metric. Apparent redundancies (e.g., caddisflies) in the metric may or may not be an issue; differences in taxonomic rate of development in wetlands due to thermal dynamics may require a certain amount of redundancy to ensure that important taxonomic groups are accounted for. The WWMI is used to rate, rank, or compare wetland biological condition.

The second macroinvertebrate index, termed the 100-Count Macroinvertebrate Biotic Index (100-Count MBI), is based on 10 metrics derived from a random pick of 100 organisms found in the three composited net sweeps. The 100-Count MBI is composed of nine percentage metrics and one richness metric (non-insect taxa). Percentage metrics include pigmy backswimmers, water boatmen, total "bugs," limnephilid caddisflies, total caddisflies, chironomids, soldier flies, and the sum of Ephemeroptera-Odonata-Trichoptera (EOT) taxa. Noninsects represent the only richness metric in the 100-Count MBI. The ninth percentage metric, mollusks, may be only useful in prairie-type wetlands. The 100-Count MBI is intended to be applied in the field by experienced staff as a means of rapid bioassessment.

The third index, the Wisconsin Wetland Plant Biotic Index (WWPBI), is based on eight (or nine) plant metrics derived from transect data (representing 18 quadrats) and is intended to serve as a supplementary index to the WWMBI to rate, rank, and compare wetland biological condition. Of 24 plant community attributes tested, only one richness (count) metric, one percent metric, and seven importance value-based metrics demonstrated consistent and predictable response. The single richness metric, total taxa, may require further modification after reaching some consistency regarding taxonomic resolution (currently mixed family-genus-species). Importance value-based metrics included Carex, reed canary grass, cattail, duckweed, bluejoint grass, and "good" plants (the sum of a group of plants including all Carex, Utricularia, Potamogeton, Leersia, Calamogrostis, Sagittaria, Polygonum, and Equisetum). An additional importance value-based metric, "pondweeds," would only be applied to wetlands with water duration exceeding seven months per year. The only percentage based metric, floating-leafed plants, would likewise only be applied to wetlands with water durations exceeding seven months.

Lessons Learned

Macroinvertebrates

• Water duration is an important factor shaping macroinvertebrate community composition and derived metrics that must be accounted for in metric scoring.
• A coarse level of taxonomic resolution (order and family) appears to be satisfactory in developing wetland macroinvertebrate metrics.
• Issues relating to redundancy among metrics, influences of water chemistry, differences among ecoregions, and seasonal variations need to be addressed in more detail. Undoubtedly, these factors need to be accounted for in establishing rating scores and/or in refining metrics for use in different areas or habitats.
• Basic differences exist in macroinvertebrate communities between wetlands representing wooded kettle depressions and open prairie-type depressions in Wisconsin.
• Our greatest difficulty was in selecting and assigning some measure of "human impact" to the study site wetlands. Further research will be required to quantify the degree of human impact in order to refine biological response metrics and indices.
• The WWMI does not appear to be stable across dates, but it is not designed to be. Consequently, its use is restricted to early spring.
• Each macroinvertebrate index has its own set of advantages and disadvantages; further refinement is required to enable their successful application in the field.

Plants

• Some biases were apparent in the WWPBI as reference kettles scored consistently higher than prairie wetlands.
• WWPBI scores in restored prairie wetlands were better than in many natural wetlands, suggesting that wetland restorations in Wisconsin may be adequate in terms of "restoring" the vegetative community (not true for macroinvertebrate response).
• The WWPBI shows some promise in its performance and, because of its taxonomic simplicity, it could be applied by nonbotanists.

Follow-Up and Continuing Work

We are in the process, with a second EPA wetland grant, of refining the macroinvertebrate and plant multimetric indices and expanding the community components to include zooplankton (in cooperation with Dr. Stanley Dodson, University of Wisconsin - Madison), diatoms (Paul Garrison, Wisconsin DNR), amphibians, and small mammals. Field work began in early April 2000, and is scheduled to continue through September 2000. Please contact Dick Lillie, Wisconsin DNR, at LILLIR@dnr.state.wi.us for additional details.