Characterizing the responses of natural populations to habitat loss remains a primary conservation challenge. Nonlinear responses can produce threshold effects that are inherently difficult to predict. Simultaneously, short-term transient population dynamics can obscure the effects of environmental change by causing delays before the eventual consequences of habitat loss or restoration are realized. Agricultural intensification is implicated in widespread declines of wild bumblebees, which are an economically important pollinator group. Here, we study the effects of rapid habitat change on short- and long-term bumblebee population dynamics. Our objectives are to: 1) examine the conditions that lead to critical thresholds in loss of natural habitat, and 2) evaluate potential time-lags in population response to habitat loss or restoration. We link habitat-specific demographic models with an empirically-derived dispersal kernel for bumblebees to establish a classic source-sink population model. We then apply the tools of matrix population modeling to study the effects of rapid landscape change on short-term (i.e., transient) and long-term (i.e., asymptotic) dynamics using a series of landscape simulations.
Loss of natural habitat by conversion to agriculture had strongly negative and nonlinear effects on long-term population growth rate, λ, producing a clear threshold in the amount of natural habitat required to sustain bumblebee populations. Consistent with previous empirical studies, habitat fragmentation had weakly negative effects on λ. Rapid landscape change also had the potential to produce considerable transient dynamics. When natural habitat was rapidly degraded, transient dynamics produced positive population inertia and lagged population declines. Conversely, even in unrealistically optimistic scenarios where agricultural habitat could be immediately restored to natural habitat, transient dynamics produced negative population inertia and lagged population increases. Convergence time, the delay until λ was achieved, ranged from one year to over a decade and was generally longer when habitat was rapidly restored than when habitat was rapidly degraded. Additionally, up to eight years elapsed before short- and long-term population growth rates were in the same direction. Our results indicate that substantial delays can occur before population dynamics equilibrate following rapid landscape change, even for short-lived species such as bumblebees. Ultimately, our study emphasizes the importance of conserving wild pollinators, as transient dynamics may obscure critical habitat thresholds and hamper subsequent restoration efforts.