Assessing ecosystem sensitivity to alterations in precipitation regimes with three long-term experiments in a US tallgrass prairie

Background/Question/Methods

At large spatial and temporal scales, the relationship between precipitation inputs and ecosystem function (e.g. ANPP) in grasslands is robust, but at local scales there are many modifiers of this relationship. With forecast climate change, changing amounts, variability and timing of precipitation inputs have the potential to alter ANPP rapidly while changes in plant community structure may modify sensitivity (response in ANPP/unit change in precipitation) in the longer term.  We utilized ANPP and plant community composition data from three long-term (14+ years) experiments that manipulated water availability either directly or indirectly in a tallgrass prairie at the Konza Prairie Long-Term Ecological Research Station near Manhattan, KS. We assessed changes in sensitivity by regressing ANPP with annual precipitation amounts, and looked for concurrent shifts in plant community composition using multivariate analyses with species abundance data. We addressed the following questions: (1) How does altered water availability affect the sensitivity of ANPP to precipitation amount, and (2) over longer time scales, will changes in plant community composition alter sensitivity?

Results/Conclusions

We found that irrigation significantly decreased ANPP sensitivity to interannual variability in precipitation initially, but that as community composition changed, sensitivity increased substantially. In sites with chronic water stress (annually burned grassland on upland, shallow soils), ANPP was more sensitive to interannual variability in precipitation than in lowland deep soil sites. However, with community change to more xeric species in uplands, ANPP sensitivity to precipitation in uplands became similar to lowlands. Finally, in a long-term experiment in which precipitation variability was increased and accompanied by changes in genotypic composition of the dominant grasses (but not species composition changes), sensitivity was slightly decreased by this treatment that resulted in generally drier soils. These results suggest that ecosystem sensitivity to important climate drivers such as precipitation is not static and that shifts in community composition are important to consider when forecasting future ecosystem responses to climate change.