@inproceedings{klocek2021ms-nowcasting,
author = {Klocek, Sylwester and Dong, Haiyu and Dixon, Matthew and Kanengoni, Panashe and Kazmi, Najeeb and Luferenko, Pete and Lv, Zhongjian and Sharma, Shikhar and Weyn, Jonathan and Xiang, Siqi},
title = {MS-nowcasting: Operational Precipitation Nowcasting with Convolutional LSTMs at Microsoft Weather},
booktitle = {NeurIPS 2021 Workshop on Tackling Climate Change with Machine Learning},
year = {2021},
month = {December},
abstract = {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.},
url = {http://approjects.co.za/?big=en-us/research/publication/ms-nowcasting-operational-precipitation-nowcasting-with-convolutional-lstms-at-microsoft-weather/},
}