Spatiotemporal Weather Data Predictions with Shortcut Recurrent-Convolutional Networks: A Solution for the Weather4cast challenge
This is an incremental improvement for weather forecasting researchers, applying a hybrid method to a specific competition dataset.
The paper tackled predicting satellite-based weather data evolution for the Weather4cast 2021 Challenge, using a neural network with GRU, residual blocks, and U-Net-like shortcuts, which retained sharp features initially but blurred later predictions due to uncertainty.
This paper presents the neural network model that was used by the author in the Weather4cast 2021 Challenge Stage 1, where the objective was to predict the time evolution of satellite-based weather data images. The network is based on an encoder-forecaster architecture making use of gated recurrent units (GRU), residual blocks and a contracting/expanding architecture with shortcuts similar to U-Net. A GRU variant utilizing residual blocks in place of convolutions is also introduced. Example predictions and evaluation metrics for the model are presented. These demonstrate that the model can retain sharp features of the input for the first predictions, while the later predictions become more blurred to reflect the increasing uncertainty.