Attention-Based Dual-Source Spatiotemporal Neural Network for Lightning Forecast
Accurate lightning forecast is significant for disaster prevention and reduction. However, the mainstream lightning forecast methods, which mainly rely on numerical simulations and parameterizations, can hardly cope with the spatiotemporal deviations. Meanwhile, the rapid and complex evolution of li...
Main Authors: | , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2019-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/8886440/ |
id |
doaj-b91e5063a2b948448ac11cb241fabc9c |
---|---|
record_format |
Article |
spelling |
doaj-b91e5063a2b948448ac11cb241fabc9c2021-03-30T00:21:13ZengIEEEIEEE Access2169-35362019-01-01715829615830710.1109/ACCESS.2019.29503288886440Attention-Based Dual-Source Spatiotemporal Neural Network for Lightning ForecastTianyang Lin0https://orcid.org/0000-0002-1228-8385Qingyong Li1Yangli-Ao Geng2Lei Jiang3Liangtao Xu4Dong Zheng5Wen Yao6Weitao Lyu7Yijun Zhang8https://orcid.org/0000-0002-7573-424XBeijing Key Laboratory of Traffic Data Analysis and Mining, Beijing Jiaotong University, Beijing, ChinaBeijing Key Laboratory of Traffic Data Analysis and Mining, Beijing Jiaotong University, Beijing, ChinaBeijing Key Laboratory of Traffic Data Analysis and Mining, Beijing Jiaotong University, Beijing, ChinaBeijing Key Laboratory of Traffic Data Analysis and Mining, Beijing Jiaotong University, Beijing, ChinaState Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing, ChinaState Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing, ChinaState Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing, ChinaState Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing, ChinaDepartment of Atmospheric and Oceanic Sciences, Fudan University, Shanghai, ChinaAccurate lightning forecast is significant for disaster prevention and reduction. However, the mainstream lightning forecast methods, which mainly rely on numerical simulations and parameterizations, can hardly cope with the spatiotemporal deviations. Meanwhile, the rapid and complex evolution of lightning regions go beyond the traditional extrapolation-based forecast methods. In this work, we propose a data-driven neural network model for hourly lightning forecast, which exploits both the numerical simulations and the recent historical lightning observations. The two kinds of data complement each other and play different roles at different stages of the forecast. The use of dual-source data greatly increases the amount of information available to improve the forecasting performance. To handle the variability of deviation patterns in numerical simulations, we introduce a channel-wise attention mechanism, which adaptively adjusts the proportion of each simulated parameter to maximize the useful information. The attention mechanism also enables the model to reveal the contribution of each simulated parameter for the forecast. Experimental results on a real-world dataset show that the proposed method outperforms several baseline methods. Ablation studies further demonstrate the effectiveness of our data fusion approach and attention module.https://ieeexplore.ieee.org/document/8886440/Deep learninglightning forecastspatiotemporal data miningconvolutional neural networkchannel-wise attention |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Tianyang Lin Qingyong Li Yangli-Ao Geng Lei Jiang Liangtao Xu Dong Zheng Wen Yao Weitao Lyu Yijun Zhang |
spellingShingle |
Tianyang Lin Qingyong Li Yangli-Ao Geng Lei Jiang Liangtao Xu Dong Zheng Wen Yao Weitao Lyu Yijun Zhang Attention-Based Dual-Source Spatiotemporal Neural Network for Lightning Forecast IEEE Access Deep learning lightning forecast spatiotemporal data mining convolutional neural network channel-wise attention |
author_facet |
Tianyang Lin Qingyong Li Yangli-Ao Geng Lei Jiang Liangtao Xu Dong Zheng Wen Yao Weitao Lyu Yijun Zhang |
author_sort |
Tianyang Lin |
title |
Attention-Based Dual-Source Spatiotemporal Neural Network for Lightning Forecast |
title_short |
Attention-Based Dual-Source Spatiotemporal Neural Network for Lightning Forecast |
title_full |
Attention-Based Dual-Source Spatiotemporal Neural Network for Lightning Forecast |
title_fullStr |
Attention-Based Dual-Source Spatiotemporal Neural Network for Lightning Forecast |
title_full_unstemmed |
Attention-Based Dual-Source Spatiotemporal Neural Network for Lightning Forecast |
title_sort |
attention-based dual-source spatiotemporal neural network for lightning forecast |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2019-01-01 |
description |
Accurate lightning forecast is significant for disaster prevention and reduction. However, the mainstream lightning forecast methods, which mainly rely on numerical simulations and parameterizations, can hardly cope with the spatiotemporal deviations. Meanwhile, the rapid and complex evolution of lightning regions go beyond the traditional extrapolation-based forecast methods. In this work, we propose a data-driven neural network model for hourly lightning forecast, which exploits both the numerical simulations and the recent historical lightning observations. The two kinds of data complement each other and play different roles at different stages of the forecast. The use of dual-source data greatly increases the amount of information available to improve the forecasting performance. To handle the variability of deviation patterns in numerical simulations, we introduce a channel-wise attention mechanism, which adaptively adjusts the proportion of each simulated parameter to maximize the useful information. The attention mechanism also enables the model to reveal the contribution of each simulated parameter for the forecast. Experimental results on a real-world dataset show that the proposed method outperforms several baseline methods. Ablation studies further demonstrate the effectiveness of our data fusion approach and attention module. |
topic |
Deep learning lightning forecast spatiotemporal data mining convolutional neural network channel-wise attention |
url |
https://ieeexplore.ieee.org/document/8886440/ |
work_keys_str_mv |
AT tianyanglin attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast AT qingyongli attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast AT yangliaogeng attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast AT leijiang attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast AT liangtaoxu attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast AT dongzheng attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast AT wenyao attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast AT weitaolyu attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast AT yijunzhang attentionbaseddualsourcespatiotemporalneuralnetworkforlightningforecast |
_version_ |
1724188474977812480 |