SSAS: Spatiotemporal Scale Adaptive Selection for Improving Bias Correction on Precipitation

Yiqun Liu (1), Junping Zhang (1), Lei Chen (2), Hai Chu (2), James Z. Wang (3), Leiming Ma (2)
(1) Fudan University, China
(2) Shanghai Central Meteorological Observatory, China
(3) The Pennsylvania State University, USA

By utilizing physical models of the atmosphere collected from current weather conditions, the Numerical Weather Prediction (NWP) model developed by the European Centre for Medium-range Weather Forecasts (ECMWF) can provide the indicators of severe weather such as heavy precipitation for an early-warning system. However, the performance of precipitation forecasts from ECMWF often suffers from considerable prediction biases owing to the high complexity and uncertainty for the formation of precipitation. The Bias Correcting on Precipitation (BCoP) was thus utilized for correcting these biases via forecasting variables including the historical observations and variables of precipitation, and these variables as predictors from ECMWF are highly relevant to precipitation. Existing BCoP methods, such as Model Output Statistics (MOS) and Ordinal Boosting Autoencoder (OBA), do not take advantage of both spatiotemporal dependencies of precipitation and scales of related predictors that can change with different precipitation. We propose an end-to-end deep-learning BCoP model, named the Spatiotemporal Scale Adaptive Selection (SSAS) model, to automatically select the spatiotemporal scales of the predictors via Spatiotemporal Scale-Selection Modules (S3M/TS2M) for acquiring the optimal high-level spatiotemporal representations. Qualitative and quantitative experiments carried out on two benchmark datasets indicate that SSAS can achieve state-of-theart performance, compared with eleven published BCoP methods, especially on heavy precipitation.

Full Paper
(high-resolution PDF, 11MB)

More information

Citation: Yiqun Liu, Junping Zhang, Lei Chen, Hai Chu, James Z. Wang and Leiming Ma, ``SSAS: Spatiotemporal Scale Adaptive Selection for Improving Bias Correction on Precipitation,'' IEEE Transactions on Cybernetics, vol. 52, no. , pp. -, 2021.

© 2021 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

Last Modified: June 8, 2021
© 2021