Disentangling the drivers of spatiotemporal dynamics in trophic metacommunities

Abstract

Understanding the relative influence of endogenous and exogenous factors on population dynamics has been a key focus of ecology since its inception. Endogenous processes such as dispersal and local interactions have historically been associated with maintaining "the balance of nature" in the form of equilibrium population dynamics, while exogenous processes such as environmental fluctuations have been deemed responsible for perturbing populations away from their natural equilibrium states. However, May (1974) showed that nonlinear species interactions could induce more complex dynamics such as limit cycles and chaos. In doing so, May helped put an end to the simplistic suggestion that the complexity of population dynamics could be used to infer their drivers, with endogenous processes being stabilizing and exogenous processes being destabilizing. Spatial synchrony, which measures the degree to which the abundances of disjunct populations are correlated over time, has been shown to be a much more potent tool for identifying the drivers of population dynamics and stability across scales. Synchrony and stability are expected to be inversely related, as increased synchrony across populations promotes the risk for stochastic extinction and thus disrupts the balance of nature (Gouhier et al. 2010a; Abbott 2011). This dissertation focuses on demonstrating how multiple drivers of synchrony interact to affect population dynamics, stability, and persistence. Chapter one explores how spatial and interspecific differences in recruitment affect the relationship between dispersal, synchrony, and stability in a trophic metacommunity. Using a keystone food web model, I show that the relationship between dispersal, synchrony, and stability can be complex. Specifically, intermediate levels of dispersal dampen population fluctuations and synchrony, no matter the degree of correlation in recruitment across species. However, high levels of dispersal generate large oscillations in population size, especially when recruitment is correlated across species, but buffers population abundances via a trophic decoupling effect when there are interspecific differences in recruitment. Thus, spatial and interspecific heterogeneity in recruitment can interact to produce complex relationships between dispersal, synchrony, and stability in trophic metacommunities. Chapter two removes the assumption present in many models that dispersal is described by a time-invariant statistical distribution. In reality, dispersal is temporally stochastic, and only mimics these static assumptions when it is averaged over many generations (Siegel et al. 2008), thus creating an implicit separation of time scales between local and regional dynamics. Using a trophic metacommunity with temporally stochastic and spatially aggregated dispersal that varies at the same time scale as local dynamics, I show that removing this separation of time scales disrupts the effect of dispersal-induced synchrony. Increasing dispersal shifts control of population dynamics from local interactions to regional processes, no matter the degree of spatial aggregation in dispersal. This results in an increase in the magnitude and the frequency of population fluctuations, which prevents spatial synchrony. Spatial aggregation in recruitment promotes boom-and-bust cycles and thus extinctions, which can be prevented by decreasing spatial aggregation or allowing species to disperse independently, which promotes stability. Overall, these results suggest that relaxing the implicit separation of time scales assumption in classic models is critical for understanding the relationship between dispersal, synchrony, and stability in nature. Finally, chapter three expands this work by not only focusing on the synchronizing effect of dispersal, but also the effect of spatial and temporal autocorrelation in the environment. Using a predator-prey model, I analyze the complex interplay between these synchronizing factors, and their effect on the dynamics, persistence, and stability of communities across scales. Low levels of dispersal, both in the absence of the environment and with weak environmental fluctuations, can induce non-stationary population dynamics. Temporal autocorrelation in the environment also disrupts the synchronizing ability of spatially autocorrelated environmental fluctuations, as well as the synchronizing effect of high levels of dispersal, even though reddened environments are expected to promote synchrony due to increased memory. Strong environmental fluctuations promote extinctions, especially under temporally autocorrelated environments, but dispersal can limit these extinctions, as long as the environment is spatially uncorrelated. These results suggest that the influence of autocorrelation in the environment on synchrony, stability, and persistence depends on the degree of environmental variability and dispersal. Taken together, the results presented in this dissertation suggest that dispersal-induced synchrony may be less common in systems characterized by heterogeneous dispersal or environments. These results are critical in a time of human-induced global change, as disruptions in processes such as dispersal and environmental fluctuations are likely. Specifically, recent work has shown that the environment is becoming more spatially and temporally autocorrelated, which removes the possibility for species to escape extreme events to spatial or temporal refugia (Di Cecco and Gouhier 2018). Thus, understanding the complex effects of and interactions between endogenous and exogenous processes is key in order to predict community responses to global change events.--Author's abstract