Neural general circulation models optimized to predict satellite-based precipitation observations

TL;DR

This paper introduces a neural network-based climate model trained on satellite precipitation data, significantly improving the simulation of precipitation patterns, extremes, and diurnal cycles over traditional models.

Contribution

It presents a hybrid neural GCM trained directly on satellite observations, enhancing precipitation simulation accuracy and outperforming existing climate models.

Findings

Reduced biases in precipitation simulation

More realistic distribution of precipitation extremes

Improved diurnal cycle representation

Abstract

Climate models struggle to accurately simulate precipitation, particularly extremes and the diurnal cycle. Here, we present a hybrid model that is trained directly on satellite-based precipitation observations. Our model runs at 2.8$^\circ$ resolution and is built on the differentiable NeuralGCM framework. The model demonstrates significant improvements over existing general circulation models, the ERA5 reanalysis, and a global cloud-resolving model in simulating precipitation. Our approach yields reduced biases, a more realistic precipitation distribution, improved representation of extremes, and a more accurate diurnal cycle. Furthermore, it outperforms the mid-range precipitation forecast of the ECMWF ensemble. This advance paves the way for more reliable simulations of current climate and demonstrates how training on observations can be used to directly improve GCMs.