Modelling lake and stream macroinvertebrate communities to improve estimates of biological conditions for the valuation of freshwater ecosystems

The U.S. Environmental Protection Agency (EPA) is conducting national surveys to capture the existence value of inland freshwaters. Recent work has found that people’s willingness to pay for improvements in water quality depends upon the biological conditions near them. Therefore, EPA economists need spatial estimates of biological conditions near survey respondents to accurately measure the value of freshwater. The US EPA’s National Aquatic Resource Surveys (NARS) provide representative samples of water bodies in the conterminous United States (CONUS) that can be used to predict biological conditions at unsampled water bodies, thereby providing estimates of conditions. We used machine learning models to produce spatial interpolations of macroinvertebrate richness, a common measure of biological condition, in lakes and streams across the CONUS. For streams and lakes, multivariate random forest (MVRF) models performed better than stacked single-taxon random forest models. The MVRF models explained about 40% of the spatial variation in macroinvertebrate richness in both streams and lakes. Further, this performance was achieved with a single national model, which alleviates complications associated with regionalization. Based on the MVRF models, many macroinvertebrate genera were strongly influenced by temperature, land cover, and runoff, independent of ecosystem type. However, the extent of forest cover and nitrogen deposition was more important for macroinvertebrates in lakes than in streams. The differences in anthropogenic drivers of biodiversity between streams and lakes suggest prioritizing different management strategies such as buffers for agricultural streams and air pollution mitigation for lakes. This study is part of a larger effort within EPA to improve valuation of aquatic resources that includes scenario modeling to understand shifts in habitat and water quality resulting from regulatory options in management.

Impact/Purpose

It is not well understood how human activity in watersheds differs in their influence on stream and lake biological conditions. Understanding the relative impacts of watershed activities on these different ecosystem types could improve the way we prioritize management of these systems at the national scale. Further, this understanding will help support current work by EPA economists to improve the way EPA captures the existence value of freshwater ecosystems and our understanding of the factors influencing this value. In this talk, we describe work to compare the relative roles of different watershed factors on stream and lake macroinvertebrate communities. We found that different human-related activities are associated with communities in these ecosystems, implying that management of each ecosystem could benefit from consideration of these different watershed stressors. Specifically, stream communities were more influenced by the amount of nearby row crops, whereas lake communities were more strongly associated with extent of forest cover and nitrogen deposition. ovel multivariate machine learning models. The results of this work will support efforts by NCEE economists to improve national valuation of freshwater ecosystems when conducting analysis of proposed Clean Water Act regulations and contributes to StRAP subproduct SSWR.401.3.2.2 – Empirical models to interpolate benthic macroinvertebrate observed/expected ratios, or other biological indicator(s) of aquatic ecosystem health, from NARS stream and lake condition to HUC12 or HUC8 units over CONUS under Product SSWR.401.2.2 – Interpolation and stressor-response analyses that extend the use of NARS data to support regulatory program needs.

Citation

Jansen, L., R. Hill, D. Kopp, S. Rumschlag, AND L. Yuan. Modelling lake and stream macroinvertebrate communities to improve estimates of biological conditions for the valuation of freshwater ecosystems. American Fisheries Society Annual Meeting, Honolulu, HI, September 15 - 19, 2024.