Statistical Properties of an Index of Biological Integrity Used to Evaluate Water Resources

Abstract

We determined the statistical properties of the index of biotic integrity (IBI) from electrofishing samples collected from Ohio streams. Although IBI is widely used to evaluate the condition of water resources by biologists and resource managers, expanding its role as a regulatory tool depends on statistical validation of its precision and power. We addressed these issues by constructing an additive variance model for IBI and testing the assumptions of that model directly with a bootstrap resampling algorithm and simulations using field data. Statistical properties of IBI supported the use of standard analysis techniques such as ANOVA for hypothesis testing. We determined with power analysis that IBI can distinguish between five and six nonoverlapping categories of biotic integrity based on a model that includes the effects of measurement error, variability of fish assemblages through time, and statistical interaction of location and time. We conclude that IBI is an effective monitoring tool that can be used to communicate qualitative assessments to the public and policy makers or to provide quantitative assessments for a legal or regulatory context based on confidence intervals or hypothesis testing procedures.