Citation:

Fergus, E., J. Renee Brooks, A. Herlihy, Ryan A Hill, Phil Kaufmann, R. Mitchell, AND P. Ringold. Improving valuation of freshwaters (II): Quantifying drivers of macroinvertebrate condition in western US streams with structural equation models. National Monitoring Conference, Virginia Beach, VA, April 24 - 28, 2023.

Impact/Purpose:

Key to developing effective environmental policy is assessing how candidate policies affect human well-being and valuation. NCEE in collaboration with OW and ORD seek to examine these relationships as mediated through aquatic ecosystems. Disentangling these complex relationships is challenging without appropriate analytic frameworks. We propose using a structural equation modeling (SEM) framework to evaluate and quantify the pathways in which natural and anthropogenic factors affect aquatic condition. We applied SEM to evaluate hypothesized watershed, riparian, and in-stream factors in affecting stream biotic condition using the US EPA National Rivers and Streams Assessment and StreamCat datasets. This information is a necessary component to bridge the gap linking management actions to ecosystem condition, and future work will expand to incorporate human valuation. The SEM framework is a promising analytic approach to leverage broad-scale data to develop mechanistic understanding of management effects on ecosystem condition.

Description:

Biotic indices from aquatic macroinvertebrate assemblages provide an integrated picture of ecosystem condition. As such, for assessment and management questions, there is value in developing a mechanistic understanding of how natural and anthropogenic drivers affect macroinvertebrate assemblage integrity. Structural equation models (SEMs) are a powerful analytic approach to quantify hypothesized causal pathways in complex systems and could bridge the information gap between aquatic monitoring data and assessment/management needs. We applied an SEM framework to examine how a hierarchy of interacting watershed, riparian, and in-stream factors may affect macroinvertebrate integrity in western US streams with data from the US EPA National Rivers and Streams Assessment (NRSA). Specifically, we examined the ratio of observed-to-expected taxonomic composition (O/E); a commonly used index in assessments. We compared model results from 595 streams sampled by NRSA in the Western Mountains (WMT) and Xeric (XER) ecoregions. Our hypothesized models fit the datasets relatively well based on commonly used SEM model fit criteria (RMSEA < 0.08; comparative fit index >0.90). The main drivers of macroinvertebrate integrity were similar in the two ecoregions, but their effect size relative to one another differed. Drivers included stream slope*depth [a morphometric measure of potential stream hydraulic energy (WMT βstd= 0.24; XER βstd= 0.43)], relative streambed stability (WMT βstd= 0.29; XER βstd= 0.16), and total nitrogen (TN) (WMT βstd= -0.15; XER βstd= -0.15). Urban land use in the watershed was negatively associated with O/E, affecting macroinvertebrate assemblage indirectly through proximal drivers (relative bed stability and TN). Riparian forest cover positively affected O/E, but indirectly through reductions in TN and improved bed stability. Identifying anthropogenic stressors and natural features affecting macroinvertebrate integrity can aid in the economic analyses of proposed management/intervention actions to support objectives of the Clean Water Act. Further, our findings illustrate that SEMs provide a scientific framework to evaluate systems-level hypotheses and provide mechanistic insights about natural and anthropogenic factors affecting aquatic ecosystems.     The views expressed in this presentation are those of the authors and do not necessarily represent the views or policies of the U.S. Environmental Protection Agency.

Record Details: