Abstract

Predicting shifts of species geographical ranges is a fundamental challenge for conservation ecologists given the great complexity of factors involved in setting range limits. Distributional patterns are frequently modelled to "simplify" species responses to the environment, yet the central mechanisms that drive a particular pattern are rarely understood. We evaluated the distributions of two sandhopper species (Crustacea, Amphipoda, Talitridae), Talorchestia capensis and Africorchestia quadrispinosa along the Namibian and South African coasts, encompassing three biogeographic regions influenced by two different oceanographic systems, the Benguela and Agulhas currents. We aimed to test whether the Abundant Centre Hypothesis (ACH) can explain the distributions of these species' abundances, sizes and sex ratios and examined which environmental parameters influence/drive these distributions. Animals were collected during a once-off survey at 29 sites over c. 3500 km of coastline. The ACH was tested using a non-parametric constraint space analysis of the goodness of fit of five hypothetical models. Distance Based Linear Modelling (DistLM) was performed to evaluate which environmental traits influenced the distribution data. Abundance, size and sex ratio showed different patterns of distribution. A ramped model fitted the abundance (Ramped North) and size (Ramped South) distribution for A. quadrispinosa. The Inverse Quadratic model fitted the size distribution of T. capensis. Beach slope, salinity, sand temperature and percentage of detritus found on the shore at the time of collection played important roles in driving the abundance of A. quadrispinosa. T. capensis was mainly affected by salinity and the morphodynamic state of the beach. Our results provided only some support for the ACH predictions. The DistLM confirmed that the physical state of the beach is an important factor for sandy beach organisms. The effect of salinity and temperature suggest metabolic responses to local conditions and a role in small to mesoscale shifts in the range of these populations.