Functional syndromes as indicators of ecosystem change in temperate grasslands

The lack of an organizing conceptual framework to address ecosystem changes reduces our capacity to distinguish biophysical from direct human impacts on grassland dynamics. This is particularly important for subhumid temperate grasslands, one of the world's most threatened biomes. We identified and mapped 4 functional syndromes of grassland change at the wettest end of its worldwide distribution, the Campos in Uruguay. Syndromes were defined by differences in precipitation use efficiency (PUE, ANPP/precipitation), and in precipitation marginal response (PMR, slope of the linear regression between ANPP and precipitation) between two periods (1981-1995 and 2001-2011). Temporal trends in aboveground net primary production (ANPP, obtained by splicing two sources of NDVI, LTDR and MOD13Q1) were also characterized. To rule out the effect of precipitation we analyzed temporal trends of the residuals from the relationship between ANPP and annual precipitation (RESTRENDS). Functional syndromes associated with increases in seasonality or in the abundance of annual vegetation (Delta PMR > 0, Delta PUE < 0, 14,000 km(2)) and vegetation cover loss (Delta PUE < 0 and Delta PMR < 0, > 5000 km(2)) were the most abundant. ANPP trends were significantly negative in 3.7% of the area (2475 km(2)) and only positive in 0.3%. However, RESTRENDS were significant in 11% of the area (> 7700 km(2)), and mostly negative (in -7200 km(2)). Most of these negative trends and residual trends were associated to seasonality increase and vegetation loss syndromes. These patterns were consistent with observed changes in the region. We highlight that this conceptual framework is suitable for describing patterns of change and potential causes. Moreover, it provides policymakers with a novel tool to guide management and conservation policies, pointing to sites where intervention (i.e. conservation, restoration) is needed.