Examining drivers of macroinvertebrate integrity in western US streams using structural equation models
On this page:
Macroinvertebrate integrity indices provide a comprehensive picture of ecosystem condition.. However, by integrating biological effects from multiple stressors such as nutrient enrichment and urbanization these metrics make it challenging to develop a mechanistic understanding of how various natural and anthropogenic drivers may affect macroinvertebrate integrity. Structural equation models (SEMs) are a powerful analytic approach to evaluate and quantify pathways in complex systems and offer promise to bridge the information gap between aquatic monitoring data and management needs. We applied an SEM framework to examine hypothesized watershed, riparian, and in-stream factors affecting macroinvertebrate integrity in western US streams using the US EPA National Rivers and Streams Assessment and StreamCat datasets. We compared model results from the Western Mountains (WMT) and Xeric (XER) ecoregions to examine differences in drivers between ecoregions. We found that our hypothesized models fit the datasets relatively well based on commonly used SEM model fit criteria (root mean square error of approximation < 0.08; comparative fit index >0.90). The main drivers of macroinvertebrate integrity were similar in the two ecoregions and included a morphometric measure of potential stream hydraulic energy (standardized total effect coefficient = 0.26), relative bed stability (0.29), and total nitrogen (-0.15). Urban land cover in the watershed was negatively associated with macroinvertebrate integrity, and these effects were all indirect and mediated through proximal drivers: relative bed stability and total nitrogen. Sulfate concentrations were negatively associated with macroinvertebrate integrity in the XER ecoregion but not in WMT. Identifying similarities and differences in drivers of macroinvertebrate integrity can guide management action/intervention that are appropriate to the ecoregional setting. Further, our findings illustrate that SEMs can provide a scientific framework to evaluate systems-level hypotheses and provide mechanistic insights by which natural and anthropogenic factors affect aquatic ecosystems.