Do desert rodents form metacommunities?

Abstract

Numerous model systems have informed classical ideas of the deterministic structure of natural communities. Although a number of important insights have been gained from desert rodents, little is known regarding the relative role of local environmental and spatial processes that embody the emerging metacommunity paradigm in structuring these model systems. We compared patterns of community composition based on environmental characteristics and estimates of spatial isolation by distance to test predictions characterizing contemporary models of metacommunity structure. Analyses were based on rodent community structure at 31 sites within the Mojave Desert. Consistent with previous studies of rodent communities, local environmental characteristics accounted for a significant amount of variation in community structure. Nonetheless, spatial isolation by distance contributed significantly to structure as well, even after accounting for the shared effect between environmental and spatial descriptors. Indeed, regional collections of rodent communities represent metacommunities and a metacommunity perspective promises to complement much that has been learned from primarily local perspectives. Examination of data suggested that the mass effects model of metacommunity structure may best describe regional patterns of species composition, at least for this data set. Moreover, 2 further important insights emerge from analyses. First, important life-history differences among taxa translate into different spatial effects that are likely due to dispersal abilities. Second, despite the spatially uncorrelated nature of environmental variation across our sampling array much of the structure of this metacommunity could be attributed to the shared effects of spatial and environmental characteristics. Typically, empirical analyses of metacommunity structure have attributed this variation to spatial structure generated by environmental spatial autocorrelation, but our analyses suggested that this component of variation may to some degree represent a unique yet underappreciated spatial effect. Moreover, this substantive component of variation suggests that prior analyses of empirical metacommunity structure provide conservative estimates of regional spatial effects and that the role of dispersal in determining metacommunity structure may be greater than is commonly considered. © 2012 American Society of Mammalogists.