Climate for Data Scientists: Leveraging Artificial Intelligence and Machine Learning Expertise to Provide Decision Support Tools in a Changing Climate

Abstract:

The recent rise in Artificial Intelligence (AI) & Machine Learning (ML) in Earth Systems Science has enabled researchers to provide increasingly accurate, high-resolution predictions of future weather and climate. This means that data scientists – even those without traditional Earth Systems backgrounds – increasingly have a role to play in helping our society adapt to a changing climate. Here we provide a concise guide to extreme event attribution, regional climate modeling, and near-term climate prediction. We focus on: (1) how the models are developed in these literatures; (2) the sources of uncertainty in these models; (3) techniques for evaluating predictions from these models; and (4) tools to access and evaluate observational data & output for model runs and reanalyses. We focus on time scales of 1-10 years, as predictions over this time frame are crucial decision-support tools for leaders in business and policy. Our goal is to encourage collaboration between the AI / ML and ESS communities by lowering the barrier to entry for researchers with a computational background but no specific climate expertise.