Metapopulation models predict the temporal response of two macrophytes to drought in a subtropical water reservoir

Abstract

Water level fluctuations are among the main drivers of the temporal dynamics of aquatic species, and droughts represent a major disturbance that can affect aquatic populations. We applied a metapopulation perspective to model the temporal variation in the patch occupancy of local populations of two species of submerged macrophytes (Egeria densa and E. najas) that colonise a large reservoir (Itaipu, Paraguay/Brazil border) and suffered an intense drought disturbance. The severe water drawdown caused a decrease in the numbers of patches that were colonised by both macrophyte species. This disturbance decreased the fractions of patches occupied by E. densa and E. najas from 13% and 38% to 0.9% and 8.5%, respectively. The recovery of the fraction of patches colonised by populations of these two species had similar responses to the disturbance: a time lag phase, followed by an exponential increase and finally stabilization of the number of occupied patches. Recovery of the fractions of patches that were colonised by E. densa and by E. najas took 80 and 38 months, respectively. A time lag in the temporal metapopulation, followed by an exponential increase in patch occupancy, strongly indicates that both populations recovered by passive dispersal of fragments between patches. Indeed, a metapopulation model that accounts for the fraction occupied and the water level (as an indication of disturbance) explained the temporal dynamics of both species (coefficients of determination = 78% for E. densa and 43% for E. najas) relatively well. Our results indicate that aquatic species that follow a metapopulation dynamic and depend on dispersal among sites to recover may be severely affected in a future climate change scenario, in which extreme events (including droughts) will tend to be more frequent.