A model of vegetation structure constraining productivity in response to changes in water availability

There has been an increasing need for forecasting aboveground net primary production (ANPP) -the main energy input for ecosystems- responses to climate variability. In arid ecosystems, ANPP increases linearly along spatial precipitation (PPT) gradients but inter-annual variation in production is loosely correlated with precipitation. Few studies have researched the mechanisms underlying the temporal dynamics of ANPP. We modeled ANPP response to water availability under the assumption that plant meristem density (MD) constrains ANPP by determining production potential. Plants were modeled as a meristem population that followed a density-dependant growth, with carrying capacity set by previous-year conditions. We parameterized the model for a shrub-grass community of the Patagonian steppe. We compared daily green biomass variation using a 19 year-long PPT record, and an artificial one that had similar mean, but that was highly variable in the occurrence of wet and dry years (+30% and -30% average PPT). Similarly to the ecosystem response, our modeled ANPP was linearly and positively related to annual PPT (R²=0.23). ANPP was also highly related to MD, but shrub MD explained ANPP better than grass MD did. ANPP showed an interesting response when simulated using a highly variable PPT series: average ANPP obtained from field observations was higher than productivity run with the artificial set. These results were mainly produced by hysteresis in the productivity response to high PPT variability. Changes in the periodicity of wet and dry years might alter the average productivity of a site, which may affect other ecosystems processes in the long-term.