MBFE-UNet: A Multi-Branch Feature Extraction UNet with Temporal Cross Attention for Radar Echo Extrapolation

Radar echo extrapolation is a critical technique for short-term weather forecasting. Timely warnings of severe convective weather events can be provided according to the extrapolated images. However, traditional echo extrapolation methods fail to fully utilize historical radar echo data, resulting i...

Full description

Saved in:
Bibliographic Details
Main Authors: Huantong Geng, Han Zhao, Zhanpeng Shi, Fangli Wu, Liangchao Geng, Kefei Ma
Format: Article
Language:English
Published: MDPI AG 2024-10-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/16/21/3956
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850062381374242816
author Huantong Geng
Han Zhao
Zhanpeng Shi
Fangli Wu
Liangchao Geng
Kefei Ma
author_facet Huantong Geng
Han Zhao
Zhanpeng Shi
Fangli Wu
Liangchao Geng
Kefei Ma
author_sort Huantong Geng
collection DOAJ
description Radar echo extrapolation is a critical technique for short-term weather forecasting. Timely warnings of severe convective weather events can be provided according to the extrapolated images. However, traditional echo extrapolation methods fail to fully utilize historical radar echo data, resulting in limited accuracy for future radar echo prediction. Existing deep learning echo extrapolation methods often face issues such as high-threshold echo attenuation and blurring distortion. In this paper, we propose a UNet-based multi-branch feature extraction model named MBFE-UNet for radar echo extrapolation to mitigate these issues. We design a Multi-Branch Feature Extraction Block, which extracts spatiotemporal features of radar echo data from various perspectives. Additionally, we introduce a Temporal Cross Attention Fusion Unit to model the temporal correlation between features from different network layers, which helps the model to better capture the temporal evolution patterns of radar echoes. Experimental results indicate that, compared to the Transformer-based Rainformer, the MBFE-UNet achieves an average increase of 4.8% in the critical success index (CSI), 5.5% in the probability of detection (POD), and 3.8% in the Heidke skill score (HSS).
format Article
id doaj-art-76fcceabf8ad429883a2fcce173154ba
institution DOAJ
issn 2072-4292
language English
publishDate 2024-10-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-76fcceabf8ad429883a2fcce173154ba2025-08-20T02:49:56ZengMDPI AGRemote Sensing2072-42922024-10-011621395610.3390/rs16213956MBFE-UNet: A Multi-Branch Feature Extraction UNet with Temporal Cross Attention for Radar Echo ExtrapolationHuantong Geng0Han Zhao1Zhanpeng Shi2Fangli Wu3Liangchao Geng4Kefei Ma5School of Computer Science, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Computer Science, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Computer Science, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Software, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Atmospheric Science, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaSchool of Software, Nanjing University of Information Science and Technology, Nanjing 210044, ChinaRadar echo extrapolation is a critical technique for short-term weather forecasting. Timely warnings of severe convective weather events can be provided according to the extrapolated images. However, traditional echo extrapolation methods fail to fully utilize historical radar echo data, resulting in limited accuracy for future radar echo prediction. Existing deep learning echo extrapolation methods often face issues such as high-threshold echo attenuation and blurring distortion. In this paper, we propose a UNet-based multi-branch feature extraction model named MBFE-UNet for radar echo extrapolation to mitigate these issues. We design a Multi-Branch Feature Extraction Block, which extracts spatiotemporal features of radar echo data from various perspectives. Additionally, we introduce a Temporal Cross Attention Fusion Unit to model the temporal correlation between features from different network layers, which helps the model to better capture the temporal evolution patterns of radar echoes. Experimental results indicate that, compared to the Transformer-based Rainformer, the MBFE-UNet achieves an average increase of 4.8% in the critical success index (CSI), 5.5% in the probability of detection (POD), and 3.8% in the Heidke skill score (HSS).https://www.mdpi.com/2072-4292/16/21/3956radar echo extrapolationdeep learningUNetmulti-branch
spellingShingle Huantong Geng
Han Zhao
Zhanpeng Shi
Fangli Wu
Liangchao Geng
Kefei Ma
MBFE-UNet: A Multi-Branch Feature Extraction UNet with Temporal Cross Attention for Radar Echo Extrapolation
Remote Sensing
radar echo extrapolation
deep learning
UNet
multi-branch
title MBFE-UNet: A Multi-Branch Feature Extraction UNet with Temporal Cross Attention for Radar Echo Extrapolation
title_full MBFE-UNet: A Multi-Branch Feature Extraction UNet with Temporal Cross Attention for Radar Echo Extrapolation
title_fullStr MBFE-UNet: A Multi-Branch Feature Extraction UNet with Temporal Cross Attention for Radar Echo Extrapolation
title_full_unstemmed MBFE-UNet: A Multi-Branch Feature Extraction UNet with Temporal Cross Attention for Radar Echo Extrapolation
title_short MBFE-UNet: A Multi-Branch Feature Extraction UNet with Temporal Cross Attention for Radar Echo Extrapolation
title_sort mbfe unet a multi branch feature extraction unet with temporal cross attention for radar echo extrapolation
topic radar echo extrapolation
deep learning
UNet
multi-branch
url https://www.mdpi.com/2072-4292/16/21/3956
work_keys_str_mv AT huantonggeng mbfeunetamultibranchfeatureextractionunetwithtemporalcrossattentionforradarechoextrapolation
AT hanzhao mbfeunetamultibranchfeatureextractionunetwithtemporalcrossattentionforradarechoextrapolation
AT zhanpengshi mbfeunetamultibranchfeatureextractionunetwithtemporalcrossattentionforradarechoextrapolation
AT fangliwu mbfeunetamultibranchfeatureextractionunetwithtemporalcrossattentionforradarechoextrapolation
AT liangchaogeng mbfeunetamultibranchfeatureextractionunetwithtemporalcrossattentionforradarechoextrapolation
AT kefeima mbfeunetamultibranchfeatureextractionunetwithtemporalcrossattentionforradarechoextrapolation