Consequences of spatial autocorrelation for the analysis of metapopulation dynamics

The importance of spatial and temporal autocorrelation for the processes occurring at many different scales has been pointed out in the last few years. Although the role of spatial pattern in models on metapopulation dynamics has recently been recognized, there is a lack of empirical studies accounting for the consequences of considering autocorrelation in these processes. Here, we show that the conventional statistical methods used for determining the factors controlling species population dynamics (species density, sex ratio, and extinction events) can be strongly affected by spatial autocorrelation. Under these circumstances, the use of a spatially explicit model is highly recommended in order to demonstrate the real factor controlling species population dynamics. Moreover, presenting erroneous conclusions concerning the real factors governing species population dynamics could have far-reaching consequences in the management and conservation of many endangered species.