Annual plant productivity as a function of species richness, diversity, and evenness over 17 years of rainfall variability in a Chilean semiarid shrubland

Over the past decade, studies in grasslands have assessed the relationship between species richness and biomass or productivity; however, these studies have produced contradicting conclusions. Often primary productivity and richness are strongly correlated, but weak or absent relationships have also been found. It could be possible that the strength of response that productivity and diversity have to common limiting resources differ, resulting in conflicting results. Similarly, few studies addressed the effects of diversity (H´) and evenness (J´) on productivity and these relationships could be relevant to explaining the functional role of diversity. Here, we tested the predictive effects of S, H´ and J´ on annual plant productivity in a Chilean semiarid shrubland, using a 17-year database (1989-2005) of annual plant cover (CA), collected in four 70m x 70 m plots. In this ecosystem, soil water and nutrient availabilities are modulated by strong interannual fluctuations in rainfall driven by El Niño Southern Oscillation (ENSO). Over the 17 years of data, rainfall varied between 11 (1998) and 356 mm (2002). In the same period, CA ranged from 28.6±6.4 (mean±1SE) to 421.8±20.1 m² and annual species richness ranged from 9 to 33 species. Plant productivity, estimated from cover, showed a positive relationship with H´, J´ and S. Similarly, these parameters increased with increasing rainfall. However, comparison of regression slopes showed that productivity increased 34% slower as a function of S than J´. We show that species richness alone was not the best predictor of ecosystem productivity. We argue that other components of diversity, species abundance and evenness, should be considered when evaluating the role of species on ecosystem function.