Currently, global and regional identifications of Ecosystem Functional Types make only use of remote sensing variables related to the amount and timing of carbon gains, i.e., surrogates of primary production, seasonality and phenology (Ivits et al., 2013, Paruelo et al., 2001, Alcaraz-Segura et al., 2006, 2013b). However, despite that primary production is the most integrative descriptor of ecosystem functioning (Virginia and Wall, 2001), there are other components of the matter and energy fluxes between the land-surface and the atmosphere that are not explicitly considered. We aim to develop an enhanced EFT classification scheme, with a more comprehensive and hierarchical identification that incorporates carbon, radiation, heat, and water exchanges.

Based on the idea that Ecosystem Functional Types encapsulate  fundamental biophysical properties we developed a methodology to determinate essential variables to describe and understand functioning of terrestrial ecosystems worldwide. We aim to define which metrics should be used to synthesize temporal dynamics of ecosystem functioning including production level, seasonality and phenology. Using Google Earth Engine we perform stratified random sampling of MODIS data along the fourteen terrestrial biomes in a global approach (Olson et al., 2001). For each biome we developed Principal Components Analysis (PCA) of the MODIS time-series and we studied the degree of correlation between the principal axes and a set of metrics that describe ecosystem function dynamics. This approach allows us represent ecosystem functioning with a reduced set of variables with biological meaning and at the same time capture most of the variation in terrestrial ecosystems.