PNNL

Abstract

Emulation through pattern scaling is a way of rapidly producing climate fields (like temperature or precipitation) from existing Earth System Model (ESM) output that, while inaccurate, is often useful for a variety of downstream purposes. Conducting pattern scaling has historically been a laborious process, in large part due to the increasing volume of ESM output. Here we introduce Pangeo-Enabled ESM Pattern Scaling (PEEPS), a Jupyter notebook-based repository of pattern scaled output from CMIP6. Because all source data for PEEPS is cloud-based, users do not need to download and house the ESM output, meaning that emulated patterns are essentially on demand. Validating PEEPS on the CMIP6 archive for temperature, precipitation, and near-surface relative humidity, pattern scaling performs well over a variety of future scenarios except for regions in which there are strong, potentially nonlinear climate feedbacks. Although pattern scaling is normally conducted on annual mean ESM output, it works equally well on monthly mean ESM output. We identify several downstream applications for PEEPS, including impacts assessment and evaluating climate system uncertainties.