Improving Climate Bias and Variability via CNN-Based State-Dependent Model-Error Corrections

We develop an approach to correct biases in the atmospheric component of the Community Earth System Model using convolutional neural networks (CNNs) to create a corrective model parameterization for online bias reduction. By predicting systematic nudging increments derived from nudging toward the ERA5-reanalysis, our method dynamically adjusts the model state, outperforming traditional corrections based on climatological increments alone. Our results show significant root mean squared error improvements across all state variables, with land precipitation biases reduced by 25%–35%, seasonally dependent. Notably, we observe an improvement to the Madden-Julian Oscillation (MJO), where the CNN-corrected model successfully propagates the MJO across the maritime continent, a challenge for many current climate models. This advancement underscores the potential of using CNNs for real-time model correction, providing a robust framework for improving climate simulations. This advancement highlights the potential of CNNs for real-time model correction, improving climate simulations and bridging observed and simulated dynamics.