Abstract: The abstraction of surface water from rivers to meet human demands is one of the most common forms of water use and is predicted to increase in intensity in the future. Indeed, the combined impact of increasing abstraction and climate change is one of the greatest threats to river ecosystems. It is important, then, to understand how rivers respond to increasing magnitudes of abstraction to better manage and protect rivers. In addition, the examination of factors that might mitigate or exacerbate the effects of abstraction is also imperative in developing sound management practices. I conducted a meta-analysis on the impact of abstraction on river macroinvertebrates to examine 1) the shape of the response (linear vs nonlinear) to increasing magnitudes of abstraction, 2) the impact of two factors (land use and study design) on abstraction’s impact, and 3) the effect of a potential mechanism of impact (increased water temperature). I collated data from 49 studies and used mixed models to determine the shape of macroinvertebrate response to increasing magnitudes of abstraction and to examine the effects of land use, study design, and increased water temperature. Of six macroinvertebrate metrics examined, four were related to the magnitude of abstraction. Three metrics – benthic density, EPT density, and taxa richness – were best fit with a simple, linear model. The fourth metric, drift load, was best fit with a 2nd order polynomial. These results indicate that drift load shows a threshold response to increasing abstraction and that the other metrics show incremental changes with incremental changes in abstraction. However, the three metrics that showed linear responses also appeared to have thresholds. At low levels of abstraction these metrics tended to show zero or even positive responses and then at moderate levels of abstraction they showed increasingly negative responses. No effects of land use, study design or increased water temperature on macroinvertebrate responses to abstraction were found. While significant relationships were found for abstraction magnitude and macroinvertebrate response, a large amount of variability (in both slope and intercepts) existed in the response of specific river systems to increasing levels of abstraction, making it difficult to provide prescriptive measures of acceptable levels of abstraction. Further work should explore factors that influence the response to abstraction and whether rivers can be grouped (e.g., based on size, underlying geology, or ecoregion) in a fashion in which systems within a group display consistent responses to abstraction.
