RadNet 1.0: Exploring deep learning architectures for longwave radiative transfer [post]

Ying Liu, Rodrigo Caballero, Joy Merwin Monteiro
2020 unpublished
<p><strong>Abstract.</strong> Simulating global and regional climate at high resolution is essential to study the effects of climate change and capture extreme events affecting human populations. To achieve this goal, the scalability of climate models and the efficiency of individual model components are both important. Radiative transfer is among the most computationally expensive components in a typical climate model. Here we attempt to model this component using a neural
more » ... t using a neural network. We aim to study the feasibility of replacing an explicit, physics-based computation of longwave radiative transfer by a neural network emulator, and assessing the resultant performance gains. We compare multiple neural-network architectures, including a convolutional neural network and our results suggest that the performance loss from the use of convolutional networks is not offset by gains in accuracy. We train the networks with and without noise added to the input profiles and find that adding noise improves the ability of the networks to generalise beyond the training set. Prediction of radiative heating rates using our neural network models achieve up to 370x speedup on a GTX 1080 GPU setup and 11x speedup on a Xeon CPU setup compared to the a state of the art radiative transfer library running on the same Xeon CPU. Furthermore, our neural network models yield less than 0.1 Kelvin per day mean squared error across all pressure levels. Upon introducing this component into a single column model, we find that the time evolution of the temperature and humidity profiles are physically reasonable, though the model is conservative in its prediction of heating rates in regions where the optical depth changes quickly. Differences exist in the equilibrium climate simulated when using the neural networks, which are attributed to small systematic errors that accumulate over time. Thus, we find that the accuracy of the neural network in the "offline" mode does not reflect its performance when coupled with other components.</p>
doi:10.5194/gmd-2019-327 fatcat:5rvurgbg75fyfij3jc5xh57i4e