Insight into the Multi-decadal Effects of Floods on Aquatic Macroinvertebrate Community Structure in the Murray River Using Distributed Lag Nonlinear Models and Counterfactual Analysis

Abstract We describe the multi-decadal delayed effects of flood on macroinvertebrate community structure using 33 years of monitoring data on macroinvertebrates, water quality, and climate, and 51 years of hydrological data, spanning 2,300 km of the Murray River, Australia. We used distributed lag nonlinear models in a four-step analytical process, including 1) modelling macroinvertebrate community structure, represented as a set of principle coordinate axes, as a function of a lagged hydrologic index and other environmental variables using distance-based redundancy analysis 2) visualizing the patterns of delayed effects of flows on the PCO axes, 3) modelling the abundances of groups of taxa along individual PCO axes, and 4) combining the two sets of models in a counterfactual analysis to predict the community structure under flood and no-flood scenarios to describe the multi-decadal trajectory of the community following a flood. Our findings show an increase in abundance of most taxa of filtering-gathering collectors, scrapers, and shredders in the long term that implicates an influx of organic matter of all sizes, from particulate organic matter to coarse and large woody debris, that serves directly or indirectly as a food resource and/or habitat. Our approach enabled the isolation of a flood impact from the confounding effects of other flow events and environmental variables, overcoming a substantial challenge in ecohydrological studies.