Creating climate projections to support the 4th California Climate Assessment

We have generated daily, 1/16° spatial resolution (about 6 km, or 3.7 miles) climate projections over the state of California to support the 4th California Climate Assessment. We started with data from 32 coarse‐resolution global climate models (GCMs) from the Climate Model Intercomparison Project, version 5 (CMIP5) archive, which is the most recent archive of models developed for the United Nations Intergovernmental Panel on Climate Change reports. The data cover 1950‐2005 for the historical period, and include two future climate projections for the period 2006‐2100, one using medium (Representative Concentration Pathway [RCP] 4.5) and one using high (RCP 8.5) greenhouse gas and aerosol emissions scenarios. We removed systematic errors from the GCMs via bias correction and then downscaled the daily precipitation, minimum and maximum temperature data from the coarse‐resolution GCMs to the 1/16° latitude‐longitude grid using a statistical method called Localized Constructed Analogs (LOCA). The downscaled fields were applied to the VIC land surface/hydrological model to develop additional variables of interest to climate impact studies, including snow cover, soil moisture, runoff, water loss from plants, surface heat fluxes, etc. The final data set of 32 models is voluminous (~40 TB), so recommendations are included for selecting a subset of models if the full set cannot be accommodated. A subset of 10 models identified by the California Department of Water Resources as having a good simulation of California’s climate is described. For impact studies that cannot accommodate data from 10 models, a further reduced set of 4 models that substantially cover the results from the set of 10 is derived.