|
The objective of this study was to evaluate the biological integrity of Indiana coolwater resources based on "least impacted" reference conditions for establishing baseline conditions (Hughes et al. 1986). Least impacted reference sites are representative of the subbasin under study and reflect the better sites with minimum anthropogenic change. Least impacted is not synonymous with pristine. Rather, sites are selected for their representativeness of the area.
The following project goals were addressed during the coolwater biological criteria project:
- Develop a thermal model capable of identifying coolwater streams during sample index periods including May through October (Task A);
- Calibrate an index of biotic integrity using the rationale of Karr et al. (1986) (Task C);
- Validate the index using habitat, land use, and water quality measures (Task D).
- Review Existing IDEM Data (Task A)
Lyons et al. (1996) classified streams as coldwater, coolwater, and warmwater based on maximum daily temperatures. Coldwater streams never exceed 20 C, while coolwater stream maximum temperatures ranged between 22-26 C, and warmwater streams temperature profiles typically are >26 C and usually exceed 30 C. Temperature increase within streams may be caused by a variety of point and non-point sources. Non-point sources include changes in channel or water body size, sedimentation, reduction in streambank and canopy cover, irrigation return flows, irrigation withdrawals, stormwater runoff, low flow, hydromodifications, and unusually warm regional temperatures (Poole & Berman 2001).
The coolwater IBI is needed because many streams in Indiana were being classified as degraded warmwater streams and ending up on the 303(d) list. In actuality these streams were misclassified for designated use. The cost of a TMDL to find problems with these streams are much more expensive than the work it took to evaluate the relationships of these streams with coolwater. Further efforts are needed to calibrate coldwater and revised warmwater expectations.
The development of a coolwater thermal model was critical to the proper implementation of this project. The development of the coolwater temperature model was based on several assumptions. First that coolwater streams are represented by measuring daily maximum temperatures and that these temperatures must never exceed 26 C during the period of measurement. Second, daily temperatures were based on reliable information that would be persistent through time. The third assumption was that the model could be created by either repeat measures of a few streams or could be developed during a single year by measuring a variety of streams throughout the study area.
Since temperature was the critical selection factor and the basis for site selection, all sites included in the validation data were initially screened using the weekly temperature calibration. During the collection of each sample reach, a single series of physical and chemical measurements are obtained Included in this physical and chemical series is a single temperature reading from a single point in time.
Over one hundred eighty candidate metrics were considered for inclusion in the coolwater IBI (Appendix A). These contained all of the relevant metrics used in previous warmwater stream IBIs, plus several additional metrics (Simon & Lyons 1995; Hughes & Oberdorff 1999). Prior to the analyses the metrics were transformed to better approximate normality (a loge transformation for the number of individuals or biomass and an arcsine-square-root transformation for proportional metrics). Results of analyses were considered significant if a < 0.05.
First, the variation in metric values was examined in relation to two natural factors, drainage area and geographic location, that might influence fish assemblages. Appropriate metrics would have either little variation relative to these two factors or a strong, monotonic, biologically meaningful relation that could be easily taken into account in IBI calculations (Hughes et al. 1998; Lyons et al. 2001). This analysis was limited to the INBS reference condition samples from the development group to minimize the potential confounding effects of human impacts. Drainage area upstream of the sampling site (loge transformed) was used as a measure of stream size. Data from all size stream and river reaches and preliminary analyses of a subset of our large river reaches were included in the calibration. No substantial structural or compositional differences were observed in the data base, so we did not need to derive separate reference condition expectations for either the ecoregions nor the land uses in the final analyses. Regression analysis was used to evaluate patterns between each metric and drainage area, while an Analysis of Variance (ANOVA) for each metric was used to compare the potential differences for ecoregion or land use. No statistically significant relationship was observed for drainage area or ecoregion.
Next, metric performance relative to a gradient of human impact was evaluated using the validation samples. When examining the most- and least-degraded stream reaches, the assumption was that multiple-impact sites would have the most modified fish assemblages and least-impacted, the least, with the intermediate impact classes somewhere in between. Metrics that fit this pattern, that is showed least-impacted sites as having the best values (highest or lowest depending on the specific metric) and multiple-impact sites having the worst values, were considered appropriate for our IBI. For each potential metric, an analysis of variance (ANOVA) was used with a Duncan multiple-range, multiple-comparison test (DMC) to assess differences among impact classes. If the metric value at the least-impacted sites were related to stream size, drainage area (loge transformed) was included as a covariate in this analysis.
The final metrics chosen for inclusion in the coolwater IBI were based on their variation relative to natural factors, their relation to human impact, and whether they represented a unique aspect of the structure, composition, or functional organization of the fish assemblage (Hughes et al. 1998). Each final metric had an appropriate response pattern to both natural factors and human impacts. For those metrics that involved the same species and that were strongly correlated with each other (Pearson’s r > 0.6), a single representative metric was chosen for use in the index. The final metrics selected included at least one metric for each of the five attributes of fish assemblages that an IBI should include: species richness and composition, indicator species, trophic function, reproductive function, and individual abundance and condition (Simon & Lyons 1995).
Scoring criteria followed the classic 1, 3, and 5 scoring criteria established by Karr (1981) and Karr et al. (1986), which is consistent with previous adaptations of the IBI for other Indiana ecoregions (Simon 1991, 1994; Simon & Dufour 1998a, b) and large rivers (Simon 1992; Simon & Stahl 1998; Emery et al. 2003). A minimum possible score (0 points) was assigned when the metric value was below the level achieved by the development data set. For example, when a site did not possess a particular indicator species or guild, then the specific metric was assigned 0 points. We did not assign a zero score for percentage metrics. The overall IBI score was the sum of twelve metric scores and ranged between 0 and 60.
The IBI was validated with data from the test group by performing an ANOVA and a DMC on the land use test samples, with impact categories as the main effect and IBI score as the response variable. Index of biotic integrity scores were converted to a proportion from 0 to 1 and then arcsine-square-root transformed prior to analysis. The IBI was considered valid if there were significant differences among impact categories, with the least-impacted samples having the highest scores and the impacted samples the lowest. The index was validated using chemical, habitat, and land use categorical data. Every one of the twelve metrics were significant with multiple tiers of anthropogenic disturbance.
Back to Home Page
|
|
