MS-nowcasting: Operational Precipitation Nowcasting with Convolutional LSTMs at Microsoft Weather
- Resource Type
- Working Paper
- Authors
- Klocek, Sylwester; Dong, Haiyu; Dixon, Matthew; Kanengoni, Panashe; Kazmi, Najeeb; Luferenko, Pete; Lv, Zhongjian; Sharma, Shikhar; Weyn, Jonathan; Xiang, Siqi
- Source
- NeurIPS 2021 Workshop on Tackling Climate Change with Machine Learning, 2021. https://www.climatechange.ai/papers/neurips2021/19
- Subject
- Computer Science - Machine Learning
Physics - Atmospheric and Oceanic Physics
- Language
We present the encoder-forecaster convolutional long short-term memory (LSTM) deep-learning model that powers Microsoft Weather's operational precipitation nowcasting product. This model takes as input a sequence of weather radar mosaics and deterministically predicts future radar reflectivity at lead times up to 6 hours. By stacking a large input receptive field along the feature dimension and conditioning the model's forecaster with predictions from the physics-based High Resolution Rapid Refresh (HRRR) model, we are able to outperform optical flow and HRRR baselines by 20-25% on multiple metrics averaged over all lead times.
Comment: Minor updates to reflect final submission to NeurIPS workshop