A Forecast-Refinement Neural Network Based on DyConvGRU and U-Net for Radar Echo Extrapolation

Precipitation nowcasting is very important for the sectors which critically depend on timely and accurate weather information. One of the challenges of precipitation nowcasting is radar echo extrapolation which predicts the radar echo images accurately. Nowadays, the methods of radar echo extrapolat...

Full description

Saved in:
Bibliographic Details
Main Authors: Jinliang Yao, Feifan Xu, Zheng Qian, Zhipeng Cai
Format: Article
Language:English
Published: IEEE 2023-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10138400/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850185594172342272
author Jinliang Yao
Feifan Xu
Zheng Qian
Zhipeng Cai
author_facet Jinliang Yao
Feifan Xu
Zheng Qian
Zhipeng Cai
author_sort Jinliang Yao
collection DOAJ
description Precipitation nowcasting is very important for the sectors which critically depend on timely and accurate weather information. One of the challenges of precipitation nowcasting is radar echo extrapolation which predicts the radar echo images accurately. Nowadays, the methods of radar echo extrapolation are mostly based on ConvRNNs. Unfortunately, as lead time increases, these methods unavoidably suffer from the problem that high reflectivity values are underestimated. Therefore, we propose a forecast-refinement neural network based on DyConvGRU and U-Net to improve the predicting ability for high reflectivity during radar echo extrapolation. To improve the model’s ability to predict high reflectivities, dynamic convolution, and the forecast-refinement architecture are applied. And to obtain more realistic results, the WGAN’s training strategy is adopted to train the forecast module and the refinement module. Through experiments on a radar dataset from Shanghai, China, the results show that our proposed method obtains higher Probability of Detection (POD), Critical Success Index (CSI), Heidke Skill Score (HSS), and lower False Alarm Rate(FAR).
format Article
id doaj-art-c673c99a7e4946c094d87db186f4a4a9
institution OA Journals
issn 2169-3536
language English
publishDate 2023-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-c673c99a7e4946c094d87db186f4a4a92025-08-20T02:16:40ZengIEEEIEEE Access2169-35362023-01-0111532495326110.1109/ACCESS.2023.328093210138400A Forecast-Refinement Neural Network Based on DyConvGRU and U-Net for Radar Echo ExtrapolationJinliang Yao0https://orcid.org/0000-0003-4689-3302Feifan Xu1https://orcid.org/0009-0002-6959-7303Zheng Qian2https://orcid.org/0000-0003-4904-2984Zhipeng Cai3https://orcid.org/0000-0003-1844-3646School of Computer Science and Technology, Hangzhou Dianzi University, Hangzhou, ChinaSchool of Computer Science and Technology, Hangzhou Dianzi University, Hangzhou, ChinaNingbo Meteorological Service Center, Ningbo, ChinaSchool of Computer Science and Technology, Hangzhou Dianzi University, Hangzhou, ChinaPrecipitation nowcasting is very important for the sectors which critically depend on timely and accurate weather information. One of the challenges of precipitation nowcasting is radar echo extrapolation which predicts the radar echo images accurately. Nowadays, the methods of radar echo extrapolation are mostly based on ConvRNNs. Unfortunately, as lead time increases, these methods unavoidably suffer from the problem that high reflectivity values are underestimated. Therefore, we propose a forecast-refinement neural network based on DyConvGRU and U-Net to improve the predicting ability for high reflectivity during radar echo extrapolation. To improve the model’s ability to predict high reflectivities, dynamic convolution, and the forecast-refinement architecture are applied. And to obtain more realistic results, the WGAN’s training strategy is adopted to train the forecast module and the refinement module. Through experiments on a radar dataset from Shanghai, China, the results show that our proposed method obtains higher Probability of Detection (POD), Critical Success Index (CSI), Heidke Skill Score (HSS), and lower False Alarm Rate(FAR).https://ieeexplore.ieee.org/document/10138400/Spatiotemporal sequence predictionradar echo extrapolationDyConvGRUU-Netdynamic convolutionWGAN
spellingShingle Jinliang Yao
Feifan Xu
Zheng Qian
Zhipeng Cai
A Forecast-Refinement Neural Network Based on DyConvGRU and U-Net for Radar Echo Extrapolation
IEEE Access
Spatiotemporal sequence prediction
radar echo extrapolation
DyConvGRU
U-Net
dynamic convolution
WGAN
title A Forecast-Refinement Neural Network Based on DyConvGRU and U-Net for Radar Echo Extrapolation
title_full A Forecast-Refinement Neural Network Based on DyConvGRU and U-Net for Radar Echo Extrapolation
title_fullStr A Forecast-Refinement Neural Network Based on DyConvGRU and U-Net for Radar Echo Extrapolation
title_full_unstemmed A Forecast-Refinement Neural Network Based on DyConvGRU and U-Net for Radar Echo Extrapolation
title_short A Forecast-Refinement Neural Network Based on DyConvGRU and U-Net for Radar Echo Extrapolation
title_sort forecast refinement neural network based on dyconvgru and u net for radar echo extrapolation
topic Spatiotemporal sequence prediction
radar echo extrapolation
DyConvGRU
U-Net
dynamic convolution
WGAN
url https://ieeexplore.ieee.org/document/10138400/
work_keys_str_mv AT jinliangyao aforecastrefinementneuralnetworkbasedondyconvgruandunetforradarechoextrapolation
AT feifanxu aforecastrefinementneuralnetworkbasedondyconvgruandunetforradarechoextrapolation
AT zhengqian aforecastrefinementneuralnetworkbasedondyconvgruandunetforradarechoextrapolation
AT zhipengcai aforecastrefinementneuralnetworkbasedondyconvgruandunetforradarechoextrapolation
AT jinliangyao forecastrefinementneuralnetworkbasedondyconvgruandunetforradarechoextrapolation
AT feifanxu forecastrefinementneuralnetworkbasedondyconvgruandunetforradarechoextrapolation
AT zhengqian forecastrefinementneuralnetworkbasedondyconvgruandunetforradarechoextrapolation
AT zhipengcai forecastrefinementneuralnetworkbasedondyconvgruandunetforradarechoextrapolation