Population-level responses of stream macroinvertebrates to drying can be density-independent or density-dependent

The frequency and intensity of droughts and stream drying are likely to increase with climate change, but how freshwater macroinvertebrate populations respond to such disturbances is poorly understood. Theory predicts that stochastic disturbances should generate density-independent mortality, but density dependence is possible if mortality arises indirectly through changes in consumer-resource dynamics, or if mortality is mediated by disturbance refugia. Empirical evidence for density-dependent or density-independent mortality in freshwater invertebrates is scarce, but important to predict how population sizes will alter with changing climate. In a stream mesocosm experiment, we tested whether short-term drying events (5-6 days) resulted in population losses of macroinvertebrates, and whether any losses were density-dependent or density-independent. On each of 10 replicate drying events over a year and over a naturally occurring range of invertebrate densities, we measured pre- and post-disturbance densities. From these data, we calculated per capita mortality rates, tested statistically for associations between mortality rates and pre-disturbance densities, and assessed whether mortality differed seasonally. Invertebrate assemblages in the channels were derived from a chalk stream in south-west England. Analyses focussed on 12 core taxa that were common in the community, which included a taxonomically diverse group of worms, Gastropoda, Gammaridae, Diptera, Ephemeroptera and Trichoptera. Responses to drying were species-specific and three different response types were observed. Half of the taxa were resistant to drying and there was no apparent mortality, even in the near-total loss of free water (Potamopyrgus antipodarum, Bezzia sp., Chironomini, Radix balthica and two groups of Naididae). Four taxa showed density-independent mortality (Polycentropus flavomaculatus, Tanypodinae, Gammarus pulex, Baetis scambus), but differed in vulnerability with per capita loss rates ranging from 57 to 100%. Strong evidence of density-dependent mortality was observed in Orthocladiinae and suggested also for Tanytarsini. Mortality varied seasonally for only one taxon (Tanytarsini), with higher mortality in warm versus cool months. Particular response types were not necessarily associated with particular taxonomic groups or body shapes. This range of responses to drying by different, naturally co-occurring, populations suggests that community-level impacts of drying are underpinned by diverse, taxon-specific responses. The experiment provides some support for theoretical predictions of density-independent mortality, but also some of the first evidence that disturbance refugia can result in density-dependent mortality.