Calibration of the biological condition gradient in Minnesota streams: a quantitative expert-based decision system

Abstract

The Biological Condition Gradient (BCG) is a conceptual model that describes changes in aquatic communities with increasing levels of anthropogenic stress. The gradient represented by the BCG has been divided into 6 levels of condition that biologists consider readily discernible in most areas of North America. We developed quantitative BCG models for 7 warm-water stream types in Minnesota for both fish and macroinvertebrates. Panels of aquatic biologists calibrated the general BCG model to Minnesota streams by assigning test samples (271 macroinvertebrate and 288 fish samples) to BCG Levels 1 to 6. From the panelists’ descriptions of their criteria for assigning sites to levels, a set of quantitative operational rules was developed for performing the same task. We developed a decision model based on fuzzy-set theory to account for discontinuities and to identify when BCG assignments might be intermediate between adjacent levels. This model captures the consensus professional judgment of the panel and uses panel-derived rules. Decisions based on the quantitative model for macroinvertebrates exactly matched 77% of the panel decisions, 89% within ½ BCG level, and 100% within 1 BCG level. Decisions based on the quantitative fish model exactly matched 70% of the panel decisions, 86% within ½ BCG level, and 99% within 1 BCG level. The BCG provides a tool to interpret aquatic biological condition along a gradient of naturalness and is consistent across stream types and political boundaries. It includes documentation of baselines to prevent inadvertent shifting, and the BCG logic rules are transparent, a desirable property for communicating condition, management goals, and water-quality criteria.