Artificial intelligence and downscaling global climate model future projections

TL;DR

This paper critically reviews the application of AI and deep learning in downscaling global climate models, highlighting potential pitfalls and emphasizing the importance of traditional statistical methods and proper evaluation strategies.

Contribution

It provides a cautious assessment of AI/ML methods in climate downscaling, emphasizing the need for balanced evaluation and recognition of established statistical approaches.

Findings

AI/ML methods are often overestimated in effectiveness.

Traditional statistical methods remain valuable for climate downscaling.

Inappropriate evaluation strategies can mislead conclusions about AI/ML performance.

Abstract

A critical review of artificial intelligence and deep machine learning (AI/ML) applied to downscaling of global climate model simulations provides some words of caution, based on past experiences and well-established principles. Recent papers tend to ignore more subtle successes with statistics and mathematical based downscaling, and there are examples of inappropriate evaluation strategies and incomplete accounts of the scientific progress when it comes to climate downscaling. An incomplete description state-of-the-art and a dogmatic approach to evaluation may give a deceiving impression that AI/ML is superior to more statistics and mathematics based methods.