You are here:
Using Weighted-Averaging to Calibrate Stressor-Specific Genus Sensitivity Values for Assessing Causes of Stream Impairments
Citation:
Griffith, M. Using Weighted-Averaging to Calibrate Stressor-Specific Genus Sensitivity Values for Assessing Causes of Stream Impairments. 2022 SETAC North America Annual Meeting, Pittsburgh, PA, November 13 - 17, 2022.
Impact/Purpose:
This abstract describes research into the use of weighted averaging to create environmental stressor-specific tolerance values for macroinvertebrates. These tolerance values may be used to identify environmental stressors likely causing impairments of macroinvertebrate assemblages in streams. This research shows that use of the stressor-specific tolerance values in a simple metric, a sensitive genera ratio, is able to differentiate between different intervals along environmental gradients, including some metals, ions, nutrients, pH, and mean substrate diameter. This approach could be used by the Office of Water, regional and state biomonitoring programs as part of a toolbox for assessing the causes of biotic impairments in streams.
Description:
Effective water quality management is based on scientific evidence or demonstrated associations between at least two pieces of information: a stressor and a response. A large, combined data-set of USEPA and state macroinvertebrate and environmental data was split into a larger calibration data-set and a smaller test data-set. Environmental variables that are measures of stressors were selected for analysis that were uncorrelated and usually had several thousand site observations. Using weighted-averaging with the calibration set-set, stressor-specific genus sensitivity values were calibrated for the selected water chemistry [log10 (X+1) transformed, except for pH] or physical habitat variables. Rather than using these stressor-specific genus sensitivity values to directly infer stressor values, we used these values to classify genera into 4 sensitivity classes along each stressor gradient and calculate a metric, the Sensitive Genus Ratio = Number of Genera in the 2 Most Sensitive Classes / Total Number of Genera, using the test data-set. By using this ratio, the sites could be separated into intervals where the mean Sensitive Genera Ratio was statistically different between intervals for many of the tested environmental variables, including some metals, ions, nutrients, pH, and mean substrate diameter. However, there were weaknesses. Limited observations of an environmental variable in the data-sets limited the number of genera for which a tolerance value could be calculated. Also, when the tails of the distribution of an environmental variable in the data-set were small, the tails were often not statistically different compared to the center of the distribution.
