COS 80-3
Effects of interannual precipitation variability on aboveground net primary production in global grasslands

Wednesday, August 13, 2014: 2:10 PM
Regency Blrm E, Hyatt Regency Hotel
Laureano A. Gherardi, School of Life Sciences, Arizona State University, Tempe, AZ
Osvaldo E. Sala, School of Life Sciences, Arizona State University, Tempe, AZ

Climate change will result in increased frequency of extreme events that lead to higher precipitation variability at multiple temporal scales. Consequences of increasing precipitation variability on the functioning of ecosystems have received considerable less attention than assessments of the effects of changes in precipitation amount. Expectations are that increased variability will negatively affect ecosystem functioning and specifically aboveground net primary production. In order to assess patterns of the relationship between different scales of precipitation variability and aboveground net primary production, we collected 88 existing long term datasets from four continents that had at least 10 consecutive years of productivity and precipitation data along a rainfall gradient from 150 to 1200 mm. We fitted various regression analyses that tested for the effects of precipitation amount, coefficient of variation and their interaction at various temporal scales and selected the best models using akaike information criterion.


Our results indicate that there is a positive overall effect of precipitation amount (b = 0.53, t (1254) = 16.42, P < 0.0001) and variability (b = 2.63, t (1254) = 4.22, P < 0.001) on ANPP. However, we found a significant interaction effect between precipitation coefficient of variation and site mean annual precipitation (b = -0.06, t (1254) = -4.86, P < 0.0001) with an inflection point at 390 mm mean annual precipitation. Therefore, increased interannual precipitation variability has a positive effect on ANPP in sites where mean annual precipitation is lower than 390 mm and a negative effect on sites where the mean annual precipitation is higher than 390 mm. These results are parallel to the inverse texture hypothesis that states that at low precipitation ecosystems benefit at coarser textures because water can penetrate deeper in the soil profile preventing evaporation losses while mesic environments benefit from thinner soil texture because it increases soil water retention capacity preventing deep percolation losses. Our findings highlight the importance of projected changes in precipitation variability and its effects on grassland ecosystems that occupy 40% of the terrestrial land area.