A Novel Reference-Based and Gradient-Guided Deep Learning Model for Daily Precipitation Downscaling
The spatial resolution of precipitation predicted by general circulation models is too coarse to meet current research and operational needs. Downscaling is one way to provide finer resolution data at local scales. The single-image super-resolution method in the computer vision field has made great...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-03-01
|
| Series: | Atmosphere |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4433/13/4/511 |
| _version_ | 1797436951962320896 |
|---|---|
| author | Li Xiang Jie Xiang Jiping Guan Fuhan Zhang Yanling Zhao Lifeng Zhang |
| author_facet | Li Xiang Jie Xiang Jiping Guan Fuhan Zhang Yanling Zhao Lifeng Zhang |
| author_sort | Li Xiang |
| collection | DOAJ |
| description | The spatial resolution of precipitation predicted by general circulation models is too coarse to meet current research and operational needs. Downscaling is one way to provide finer resolution data at local scales. The single-image super-resolution method in the computer vision field has made great strides lately and has been applied in various fields. In this article, we propose a novel reference-based and gradient-guided deep learning model (RBGGM) to downscale daily precipitation considering the discontinuity of precipitation and ill-posed nature of downscaling. Global Precipitation Measurement Mission (GPM) precipitation data, variables in ERA5 re-analysis data, and topographic data are selected to perform the downscaling, and a residual dense attention block is constructed to extract features of them. By exploring the discontinuous feature of precipitation, we introduce gradient feature to reconstruct precipitation distribution. We also extract the feature of high-resolution monthly precipitation as a reference feature to resolve the ill-posed nature of downscaling. Extensive experimental results on benchmark data sets demonstrate that our proposed model performs better than other baseline methods. Furthermore, we construct a daily precipitation downscaling data set based on GPM precipitation data, ERA5 re-analysis data and topographic data. |
| first_indexed | 2024-03-09T11:10:59Z |
| format | Article |
| id | doaj.art-5c3eca45235c4ac5b5af07d8faaaa612 |
| institution | Directory Open Access Journal |
| issn | 2073-4433 |
| language | English |
| last_indexed | 2024-03-09T11:10:59Z |
| publishDate | 2022-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Atmosphere |
| spelling | doaj.art-5c3eca45235c4ac5b5af07d8faaaa6122023-12-01T00:45:51ZengMDPI AGAtmosphere2073-44332022-03-0113451110.3390/atmos13040511A Novel Reference-Based and Gradient-Guided Deep Learning Model for Daily Precipitation DownscalingLi Xiang0Jie Xiang1Jiping Guan2Fuhan Zhang3Yanling Zhao4Lifeng Zhang5College of Meteorology and Oceanology, National University of Defense Technology, Nanjing 211101, ChinaCollege of Meteorology and Oceanology, National University of Defense Technology, Nanjing 211101, ChinaCollege of Meteorology and Oceanology, National University of Defense Technology, Nanjing 211101, ChinaSchool of Computer, National University of Defense Technology, Changsha 410000, ChinaThe PLA 31010 Units, Beijing 100081, ChinaCollege of Meteorology and Oceanology, National University of Defense Technology, Nanjing 211101, ChinaThe spatial resolution of precipitation predicted by general circulation models is too coarse to meet current research and operational needs. Downscaling is one way to provide finer resolution data at local scales. The single-image super-resolution method in the computer vision field has made great strides lately and has been applied in various fields. In this article, we propose a novel reference-based and gradient-guided deep learning model (RBGGM) to downscale daily precipitation considering the discontinuity of precipitation and ill-posed nature of downscaling. Global Precipitation Measurement Mission (GPM) precipitation data, variables in ERA5 re-analysis data, and topographic data are selected to perform the downscaling, and a residual dense attention block is constructed to extract features of them. By exploring the discontinuous feature of precipitation, we introduce gradient feature to reconstruct precipitation distribution. We also extract the feature of high-resolution monthly precipitation as a reference feature to resolve the ill-posed nature of downscaling. Extensive experimental results on benchmark data sets demonstrate that our proposed model performs better than other baseline methods. Furthermore, we construct a daily precipitation downscaling data set based on GPM precipitation data, ERA5 re-analysis data and topographic data.https://www.mdpi.com/2073-4433/13/4/511precipitation downscalingdeep learningsuper-resolution |
| spellingShingle | Li Xiang Jie Xiang Jiping Guan Fuhan Zhang Yanling Zhao Lifeng Zhang A Novel Reference-Based and Gradient-Guided Deep Learning Model for Daily Precipitation Downscaling Atmosphere precipitation downscaling deep learning super-resolution |
| title | A Novel Reference-Based and Gradient-Guided Deep Learning Model for Daily Precipitation Downscaling |
| title_full | A Novel Reference-Based and Gradient-Guided Deep Learning Model for Daily Precipitation Downscaling |
| title_fullStr | A Novel Reference-Based and Gradient-Guided Deep Learning Model for Daily Precipitation Downscaling |
| title_full_unstemmed | A Novel Reference-Based and Gradient-Guided Deep Learning Model for Daily Precipitation Downscaling |
| title_short | A Novel Reference-Based and Gradient-Guided Deep Learning Model for Daily Precipitation Downscaling |
| title_sort | novel reference based and gradient guided deep learning model for daily precipitation downscaling |
| topic | precipitation downscaling deep learning super-resolution |
| url | https://www.mdpi.com/2073-4433/13/4/511 |
| work_keys_str_mv | AT lixiang anovelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT jiexiang anovelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT jipingguan anovelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT fuhanzhang anovelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT yanlingzhao anovelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT lifengzhang anovelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT lixiang novelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT jiexiang novelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT jipingguan novelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT fuhanzhang novelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT yanlingzhao novelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling AT lifengzhang novelreferencebasedandgradientguideddeeplearningmodelfordailyprecipitationdownscaling |