Timescape: a space-time interpolation algorithm for complex ecological systems

Forest ecological systems are complex environments. Often they show both space and time patterns of variability of the values of the measured variables. The state-of-the-art algorithms currently employed in forestry research are focused on spatial variability (geostatistics) or on time-related variability (time series analysis), but not both.We propose a relatively simple, innovative algorithm which handles space- and time-variability on the same grounds, specially suited for the typical needs of forest ecology research: sparse information, inhomogeneous data collection and far from optimal space distribution of the samples. The Timescape algorithm has already been implemented in Java language, it is distributed under an open license, in two flavours: a global version for regional to world scaled projects, using geographical coordinates, and a local version for detailed scale projects, using projected coordinates. The software packages are designed in order to run on most now adays standard hardware configurations and operating systems. The output of Timescape modelling is a three-dimensional (time space) bulk voxel model from which both time series and geographical raster layers can be extracted. The model is georeferenced so it can be integrated into an already established GIS workflow.