TRU-NET: A Deep Learning Approach to High Resolution Prediction of Rainfall

Rilwan Adewoyin, Peter Dueben, Peter Watson, Yulan He, Ritabrata Dutta

Machine Learning · 20 August 2020 · ECMWF

In one sentence. TRU-NET, a convolutional-recurrent U-Net with cross attention, predicts high-resolution rainfall from climate model fields with lower RMSE than a dynamical model's rainfall output.

Abstract

Climate models (CM) are used to evaluate the impact of climate change on the risk of floods and strong precipitation events. However, these numerical simulators have difficulties representing precipitation events accurately, mainly due to limited spatial resolution when simulating multi-scale dynamics in the atmosphere. To improve the prediction of high resolution precipitation we apply a Deep Learning (DL) approach using an input of CM simulations of the model fields (weather variables) that are more predictable than local precipitation. To this end, we present TRU-NET (Temporal Recurrent U-Net), an encoder-decoder model featuring a novel 2D cross attention mechanism between contiguous convolutional-recurrent layers to effectively model multi-scale spatio-temporal weather processes. We use a conditional-continuous loss function to capture the zero-skewed %extreme event patterns of rainfall. Experiments show that our model consistently attains lower RMSE and MAE scores than a DL model prevalent in short term precipitation prediction and improves upon the rainfall predictions of a state-of-the-art dynamical weather model. Moreover, by evaluating the performance of our model under various, training and testing, data formulation strategies, we show that there is enough data for our deep learning approach to output robust, high-quality results across seasons and varying regions.

Labels

Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.