Increasing global temperatures are projected to alter the frequency and intensity of precipitation events. Specifically, precipitation patterns are predicted to become dominated by extreme rainfall events punctuated with longer dry intervals. Changes in rainfall distribution are especially important in arid and semi-arid environments where precipitation driven spatial and temporal patterns of soil water content are a key control on ecosystem function. Throughout aridland systems discrete storms provide a pulse of soil water that drives biotic activity until available water is depleted. The "bucket model" predicts an increase in soil respiration and primary production in xeric ecosystems as extreme precipitation events alleviate drought stress through deep soil water recharge. To determine the effects of altered precipitation regimes on a Chihuahuan Desert grassland ecosystem, we experimentally altered precipitation frequency and intensity in a monsoon rainfall manipulation experiment at the Sevilleta LTER site in central New Mexico, USA. Treatments (n = 5) included ambient rainfall plus one 20 mm rain event each month (July-Sept) and ambient plus four 5 mm rain events each month. Soil processes were measured by in situ sensors that measured soil CO₂ and moisture at 30 minute intervals. Leaf-level photosynthesis and pre-dawn water potential of Bouteloua eriopoda were measured in association with experimental rainfall events. Aboveground net primary productivty (ANPP) was measured bi-annually.

Results/Conclusions

Our results support the "bucket model" which predicts that a small number of large events (more extreme events) will increase soil moisture in xeric ecosystems leading to an increase in soil respiration and primary production. When compared to ambient plots, large rainfall events significatly increased mean %SWC (soil water content) and this corresponded with a significant increase in soil CO₂ flux and ANPP (207.10 g m^-2 vs 97.60 g m^-2). Moreover, rainfall treatments significantly decreased pre-dawn leaf water potential and increased leaf-level photosynthesis by 32% (p < .05). Overall our results show that desert grasslands dominated by B. eriopoda are highly sensitive to changes in precipitation regime. Climate models consistently indicate a future with altered precipitation patterns and an increase in extreme precipitation events. Thus, understanding how the spatial and temporal patterns in precipitation will affect aridland ecosystems is important because water availability affects ecosystem carbon exchange through shifts in plant and microbial activity. Through this and similar rainfall manipulation experiments, we can begin to address the consequences of environmental change in aridland ecosystems.