Advanced cyberinfrastructure for intercomparison and validation of climate models

Abstract

The current routine of comparison and validation in climate science is frequently static and of low efficiency, which hinders evidence-based decision making and scientific confidence. Due to the aggressively increasing resolution, complexity, and associated data volumes of climate models, objectively comparing multiple models and assessing their accuracy against observations is an ever-increasing challenge. We propose an integrated framework for harmonizing state-of-the-art cyberinfrastructure techniques with the user habits formed by long-term familiarity with existing community-oriented software. An open source prototype named COVALI is implemented and used to compare and validate the results of several widely-used climate models and datasets. Our results show that the proposed cyberinfrastructure-based strategy can significantly automate the comparison and validation processes in climate modeling. More importantly, the new strategy retains the existing user habits in the climate community while making it easier for scientists to adopt new technology in their research routine.