Advancing spatio-temporal analysis of ecological data: Examples in R

The article reviews main principles of running geo-computations in ecology, as illustrated with case studies from the EcoGRID and FlySafe projects, and emphasizes the advantages of using R computing environment as the most attractive programming/scripting environment. Three case studies (including R code) of interest to ecological applications are described: (a) analysis of GPS trajectory data for two gull-birds species; (b) species distribution mapping in space and time for a bird species (sedge warbler; EcoGRID project); and (c) change detection using time-series of maps. The case studies demonstrate that R, together with its numerous packages for spatial and geostatistical analysis, is a well-suited tool to produce quality outputs (maps, statistical models) of interest in Geo-Ecology. Moreover, due to the recent implementation of the maptools and sp packages, such outputs can be easily exported to popular geographical browsers such as Google Earth and similar. The key computational challenges for Computational Geo-Ecology recognized were: (1) solving the problem of input data quality (filtering techniques), (2) solving the problem of computing with large data sets, (3) improving the over-simplistic statistical models, and (4) producing outputs of increasingly higher level of detail