Temporal models of precipitation-productivity relationships currently depict the relationship of ecosystem productivity with precipitation as linear, yet growing evidence suggests that ecosystems may be differentially sensitive to dry versus wet years. In addition, underrepresented in temporal models is the influence that extreme years may have on the relationships between precipitation and productivity. This suggests that our ability to project the response of ecosystem productivity to predicted increases in precipitation variability is constrained by a lack of understanding about the influence of precipitation extremes, as well as the relative sensitivities of ecosystems to positive versus negative anomalies. In a two year experiment within a mesic grassland, we assessed the responses of three key carbon cycling processes – net photosynthesis, soil respiration and aboveground net primary productivity (ANPP) - to an experimental gradient of growing season precipitation amount and extremity. This gradient ranged from the lowest to highest growing season precipitation amounts in the 110-year record, employing 9 and 11 levels in 2015 and 2016, respectively.
Results/Conclusions
Soil respiration exhibited negative asymmetric responses to precipitation extremes, with clear evidence for saturating responses to the precipitation gradient. The response of ANPP to precipitation was more dynamic. We observed variable effects under extreme drought conditions, yet consistently large effects for wet extremes, with the largest divergence in ANPP from nominal conditions occurring under wet extremes. Furthermore, ANPP exhibited both saturating and linear responses to the precipitation gradient in 2015 and 2016 respectively, with the former associated with a higher intercept of ANPP. The unifying mechanism underlying the response dynamics of these processes was largely attributed to soil moisture impacts to the upper 20cm of the soil profile. We observed a saturating relationship of soil respiration with soil moisture, yet a consistently linear relationship with ANPP. These data highlight the differential sensitivities of production and respiration to changes in precipitation amount and extremity. In particular, our data suggests that inferences of ANPP-precipitation relationships may be contingent on how precipitation translates to impacts soil moisture availability. Such an understanding may be critical for the proper parameterization of models predicting future precipitation-production relationships.