Relics of beavers past: time and population density drive scale‐dependent patterns of ecosystem engineering

Abstract

Like many ecological processes, natural disturbances exhibit scale‐dependent dynamics that are largely a function of the magnitude, frequency and scale at which they are assessed. Ecosystem engineers create patch‐scale disturbances that affect ecological processes, yet we know little about how these effects scale across space or vary through time. Here, we investigate how patch disturbances by beavers Castor canadensis, ecosystem engineers renowned for their pond‐creation behavior, affect ecological processes across space and time. We evaluated how beaver population recovery influenced surface water dynamics in relation to population density over 70 years across multiple spatial scales (pond, watershed and regional) in northern Minnesota. Surface water area was positively related to population density at the watershed scale; however, despite variation in beaver densities (and therefore surface water area) at the watershed scale, regional‐scale surface water area was stable through time. This stability appears to have been driven by asynchronous beaver density fluctuations among watersheds, combined with the increasing importance of abandoned ponds. Beavers initially created and occupied larger ponds with greater surface water area, but through time shifted towards occupying smaller ponds. As ponds accumulated on the landscape proportionally more surface water was stored within abandoned ponds, which offset the smaller size of occupied ponds. Beaver engineering – driven by density‐dependent mechanisms and the legacy effects from abandoned ponds – not only follows general patterns of patch disturbance dynamics by creating a spatial mosaic of patches, but the organism‐created mosaic also appears to generate ecological stability at greater spatial scales. We suggest restoring beavers to landscapes is a viable method for increasing surface water storage and will ultimately help advance numerous conservation and rewilding objectives. Our study demonstrates that ecosystem engineering effects can be scale‐dependent, indicating researchers should evaluate the ecological impact of engineers across diverse spatiotemporal scales to fully understand their functional roles in ecosystems.