The space-for-time substitution assumption is often used implicitly for studying ecological processes in static spatial data sets. Since ecological processes occur in time, this practice is problematic, especially in nonstationary environments. More processes might lead to the same spatial pattern, and instead of testing hypotheses on ecological processes by analyzing spatial variation in static data, it is more judicious to report the observed spatial patterns and only discuss which ecological processes are in concordance with the observed spatial pattern. Alternatively, it might be feasible to combine relatively sparse time-series data or experimental data with spatial variation data and analyze such data types in a common statistical framework.
