Impacts of climate change on California's rangeland production: sensitivity and future projection

Abstract

Rangelands support many important ecosystem services and are highly sensitive to climate change. Understanding temporal dynamics in rangeland gross primary production (GPP) and how it may change under projected climate change, including more frequent and severe droughts, is critical for ranching communities to cope with future changes. Covering ~10% of Californias climatologically and topographically diverse landscapes, annual rangelands express varying sensitivity to precipitation fluctuation and warming. Herein, we examined how climate regulates temporal dynamics of annual GPP in Californias annual rangeland across scales, based on 20 years of satellite record derived GPP at 500-meter resolution since 2001. We built gradient boosted regression tree models for 23 ecoregion subsections in our study area, relating annual GPP with 30 climatic variables and disentangling the partial dependence of GPP on each climate variable. Our analysis showed that GPP was most sensitive to growing season precipitation amount; GPP decrease as much as 200 g C/m2/yr when growing season precipitation decreased from 400 to 100 mm/yr in one of the driest subsections. We also found that years with more evenly distributed growing season precipitation had higher GPP. Warmer winter minimum air temperature enhanced GPP in approximately two-thirds of the subsections. In contrast, average growing season mean and maximum air temperatures showed a negative relationship with annual GPP. When forced by downscaled future climate projections, changes in future rangeland productivity at the ecoregion subsection scale were more remarkable than at the state level; this suggests rangeland productivity responses to climate change will be highly variable at the local level. Further, we found large uncertainty in precipitation projections among the four climate models used in this study. Specifically, drier models predicted a larger degree of reduction in GPP, especially in drier subsections. Our machine learning-based analysis highlights key regional differences in GPP vulnerability to climate and provides insights into the intertwining and potentially counteracting effects of seasonal temperature and precipitation regimes. This work demonstrates the potential of using remote sensing to enhance field-based rangeland monitoring and, combined with machine learning, to inform adaptive management and conservation within the context of weather extremes and climate change.