Impact of land use on macroinvertebrates from a multiscale perspective: enhancing structural equation models with inverse distance-weighted metrics

Abstract

Effective riverine ecosystem management and restoration are predicated upon comprehending the relationship between land use patterns, the aquatic environment, and macroinvertebrates. However, the processes by which land use affects riverine ecosystems are exceedingly complex, influenced by factors across varying special scales. To explore these multi-scale relationships, we centered our study on the Taizi River Basin (TRB) in northeastern China, encompassing data on land use, macroinvertebrates, physical habitats, and water quality. We employed a synergy of correlation analysis, multiple stepwise regression, and redundant analysis techniques to discern these relationships at diverse spatial extents. This approach led to the development of 24 structural equation models (SEMs) that delineate the pathways bridging land use patterns to various biological indices. Remarkably, the interpretive capacity of SEM-Benthic Index of Biotic Integrity (SEM-BIBI), SEM-Shannon Wiener index (SEM-H’), and SEM-Average Score Per Taxon (SEM-ASPT) models was 76%, 70%, and 66% respectively. Within the TRB, the primary pathway from land use pattern to biological index consisted of: land use (predominantly rural and agricultural) feeding into physical habitat (habitat complexity, combined characteristic of speed and depth), which in turn affected benthic indicators. A secondary pathway charts a course from land use, through water quality (electrical conductivity and total dissolved solids) and culminating in benthic variations. Furthermore, our findings indicated that inverse-distance-weighted (IDW) metrics, specifically the IDW distance based on flow length to the stream outlet and the nearest stream (iFLO_m and iFLS_m, respectively) could provide more accurate interpretations of the river’s ecological health status. Given the effects of rural areas and agriculture on macroinvertebrates, we recommend establishing 100-m wide riparian buffers. The optimal buffer range should be determined by evaluating various biological needs and considerations.