Unmasking Compensating Errors in Numerical Weather Models: A Process-Additive Neural Network Approach

Yuiko Ichikawa, Edward Groot, Jack Atkinson

This work introduces a Process Additive Neural Network (PANN) framework to disentangle compensating biases in atmospheric models, where errors in different physical processes can o…