An Application of Deep Learning Technique to Improve Subseasonal to Seasonal Rainfall Forecast over Java Island, Indonesia
Subseasonal to seasonal (S2S) rainfall forecast can benefit several sectors, such as water resources, hazard management, and agriculture. However, the forecast remains challenging due to its lack of skill. This study applies Convolutional AutoEncoders (ConvAE), a deep learning technique, to improve...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Bogor Agricultural University
2022-11-01
|
| Series: | Journal of Natural Resources and Environmental Management |
| Subjects: | |
| Online Access: | https://journal.ipb.ac.id/index.php/jpsl/article/view/39478 |
| _version_ | 1827834112439746560 |
|---|---|
| author | Adyaksa Budi Raharja Akhmad Faqih Amsari Mudzakir Setiawan |
| author_facet | Adyaksa Budi Raharja Akhmad Faqih Amsari Mudzakir Setiawan |
| author_sort | Adyaksa Budi Raharja |
| collection | DOAJ |
| description |
Subseasonal to seasonal (S2S) rainfall forecast can benefit several sectors, such as water resources, hazard management, and agriculture. However, the forecast remains challenging due to its lack of skill. This study applies Convolutional AutoEncoders (ConvAE), a deep learning technique, to improve the quality of the S2S rainfall forecast. Seven S2S model output incorporated with Subseasonal Experiments Projects (SubX), including CCSM4, CFSv2, FIMr1p1, GEFS, GEOS_v2p1, GEPS6, and NESM, are corrected using the ConvAE approach. We combine 407 ground observations and the CHIRPS dataset using regression kriging methods producing gridded daily precipitation data with 0.05° spatial resolution. We utilize this dataset as a label to train ConvAE models and to perform bias corrections to all members of the SubX forecasts data. The results show that ConvAE is able to increase the quality of weekly S2S rainfall forecasts over Java, Indonesia. The Correlation Coefficient for 1 – 4 weeks lead time are improved from: 0.76, 0.715, 0.692 and 0.722 towards 0.809, 0.751, 0.719 and 0.74, respectively. Furthermore, the average CRPSS improves between 20 – 30% for all lead times.
|
| first_indexed | 2024-03-12T05:45:15Z |
| format | Article |
| id | doaj.art-3ee2eda13e0f47a194be2438e8a658c5 |
| institution | Directory Open Access Journal |
| issn | 2086-4639 2460-5824 |
| language | English |
| last_indexed | 2024-03-12T05:45:15Z |
| publishDate | 2022-11-01 |
| publisher | Bogor Agricultural University |
| record_format | Article |
| series | Journal of Natural Resources and Environmental Management |
| spelling | doaj.art-3ee2eda13e0f47a194be2438e8a658c52023-09-03T05:40:06ZengBogor Agricultural UniversityJournal of Natural Resources and Environmental Management2086-46392460-58242022-11-0112410.29244/jpsl.12.4.587-598An Application of Deep Learning Technique to Improve Subseasonal to Seasonal Rainfall Forecast over Java Island, IndonesiaAdyaksa Budi RaharjaAkhmad Faqih0Amsari Mudzakir Setiawan1Department of Geophysics and Meteorology, Faculty of Mathematics and Natural Sciences, IPBCenter for Climate Change Information, Meteorological Climatological and Geophysical Agency (BMKG) Subseasonal to seasonal (S2S) rainfall forecast can benefit several sectors, such as water resources, hazard management, and agriculture. However, the forecast remains challenging due to its lack of skill. This study applies Convolutional AutoEncoders (ConvAE), a deep learning technique, to improve the quality of the S2S rainfall forecast. Seven S2S model output incorporated with Subseasonal Experiments Projects (SubX), including CCSM4, CFSv2, FIMr1p1, GEFS, GEOS_v2p1, GEPS6, and NESM, are corrected using the ConvAE approach. We combine 407 ground observations and the CHIRPS dataset using regression kriging methods producing gridded daily precipitation data with 0.05° spatial resolution. We utilize this dataset as a label to train ConvAE models and to perform bias corrections to all members of the SubX forecasts data. The results show that ConvAE is able to increase the quality of weekly S2S rainfall forecasts over Java, Indonesia. The Correlation Coefficient for 1 – 4 weeks lead time are improved from: 0.76, 0.715, 0.692 and 0.722 towards 0.809, 0.751, 0.719 and 0.74, respectively. Furthermore, the average CRPSS improves between 20 – 30% for all lead times. https://journal.ipb.ac.id/index.php/jpsl/article/view/39478bias correction; convolutional autoencoders; rainfall forecasts; sub-seasonal to seasonal |
| spellingShingle | Adyaksa Budi Raharja Akhmad Faqih Amsari Mudzakir Setiawan An Application of Deep Learning Technique to Improve Subseasonal to Seasonal Rainfall Forecast over Java Island, Indonesia Journal of Natural Resources and Environmental Management bias correction; convolutional autoencoders; rainfall forecasts; sub-seasonal to seasonal |
| title | An Application of Deep Learning Technique to Improve Subseasonal to Seasonal Rainfall Forecast over Java Island, Indonesia |
| title_full | An Application of Deep Learning Technique to Improve Subseasonal to Seasonal Rainfall Forecast over Java Island, Indonesia |
| title_fullStr | An Application of Deep Learning Technique to Improve Subseasonal to Seasonal Rainfall Forecast over Java Island, Indonesia |
| title_full_unstemmed | An Application of Deep Learning Technique to Improve Subseasonal to Seasonal Rainfall Forecast over Java Island, Indonesia |
| title_short | An Application of Deep Learning Technique to Improve Subseasonal to Seasonal Rainfall Forecast over Java Island, Indonesia |
| title_sort | application of deep learning technique to improve subseasonal to seasonal rainfall forecast over java island indonesia |
| topic | bias correction; convolutional autoencoders; rainfall forecasts; sub-seasonal to seasonal |
| url | https://journal.ipb.ac.id/index.php/jpsl/article/view/39478 |
| work_keys_str_mv | AT adyaksabudiraharja anapplicationofdeeplearningtechniquetoimprovesubseasonaltoseasonalrainfallforecastoverjavaislandindonesia AT akhmadfaqih anapplicationofdeeplearningtechniquetoimprovesubseasonaltoseasonalrainfallforecastoverjavaislandindonesia AT amsarimudzakirsetiawan anapplicationofdeeplearningtechniquetoimprovesubseasonaltoseasonalrainfallforecastoverjavaislandindonesia AT adyaksabudiraharja applicationofdeeplearningtechniquetoimprovesubseasonaltoseasonalrainfallforecastoverjavaislandindonesia AT akhmadfaqih applicationofdeeplearningtechniquetoimprovesubseasonaltoseasonalrainfallforecastoverjavaislandindonesia AT amsarimudzakirsetiawan applicationofdeeplearningtechniquetoimprovesubseasonaltoseasonalrainfallforecastoverjavaislandindonesia |