1. Invertebrate drift is a fundamental process in streams and rivers. Studies from laboratory experiments and small streams have identified numerous extrinsic (e.g. discharge, light intensity, water quality) and intrinsic factors (invertebrate life stage, benthic density, behaviour) that govern invertebrate drift concentrations (# m−3), but the factors that govern invertebrate drift in larger rivers remain poorly understood. For example, while large increases or decreases in discharge can lead to large increases in invertebrate drift, the role of smaller, incremental changes in discharge is poorly described. In addition, while we might expect invertebrate drift concentrations to be proportional to benthic densities (# m−2), the benthic–drift relation has not been rigorously evaluated.
2. Here, we develop a framework for modelling invertebrate drift that is derived from sediment transport studies. We use this framework to guide the analysis of high-resolution data sets of benthic density and drift concentration for four important invertebrate taxa from the Colorado River downstream of Glen Canyon Dam (mean daily discharge 325 m3 s−1) that were collected over 18 months and include multiple observations within days. Ramping of regulated flows on this river segment provides an experimental treatment that is repeated daily and allowed us to describe the functional relations between invertebrate drift and two primary controls, discharge and benthic densities.
3. Twofold daily variation in discharge resulted in a >10-fold increase in drift concentrations of benthic invertebrates associated with pools and detritus (i.e. Gammarus lacustris and Potamopyrgus antipodarum). In contrast, drift concentrations of sessile blackfly larvae (Simuliium arcticum), which are associated with high-velocity cobble microhabitats, decreased by over 80% as discharge doubled. Drift concentrations of Chironomidae increased proportional to discharge.
4. Drift of all four taxa was positively related to benthic density. Drift concentrations of Gammarus, Potamopyrgus and Chironomidae were proportional to benthic density. Drift concentrations of Simulium were positively related to benthic density, but the benthic–drift relation was less than proportional (i.e. a doubling of benthic density only led to a 40% increase in drift concentrations).
5. Our study demonstrates that invertebrate drift concentrations in the Colorado River are jointly controlled by discharge and benthic densities, but these controls operate at different timescales. Twofold daily variation in discharge associated with hydropeaking was the primary control on within-day variation in invertebrate drift concentrations. In contrast, benthic density, which varied 10- to 1000-fold among sampling dates, depending on the taxa, was the primary control on invertebrate drift concentrations over longer timescales (weeks to months).