Deep Learning-Based Univariate Prediction of Daily Rainfall: Application to a Flood-Prone, Data-Deficient Country
There are several attempts to model rainfall time series which have been explored by members of the hydrological research communities. Rainfall, being one of the defining factors for a flooding event, is rarely modeled singularly in deep learning, as it is usually performed in multivariate analysis....
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | Atmosphere |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4433/14/4/632 |
| _version_ | 1797606463226511360 |
|---|---|
| author | Imee V. Necesito Donghyun Kim Young Hye Bae Kyunghun Kim Soojun Kim Hung Soo Kim |
| author_facet | Imee V. Necesito Donghyun Kim Young Hye Bae Kyunghun Kim Soojun Kim Hung Soo Kim |
| author_sort | Imee V. Necesito |
| collection | DOAJ |
| description | There are several attempts to model rainfall time series which have been explored by members of the hydrological research communities. Rainfall, being one of the defining factors for a flooding event, is rarely modeled singularly in deep learning, as it is usually performed in multivariate analysis. This study will attempt to explore a time series modeling method in four subcatchments located in Samar, Philippines. In this study, the rainfall time series was treated as a signal and was reconstructed into a combination of a ‘smoothened’ or ‘denoised’ signal, and a ‘detailed’ or noise signal. The discrete wavelet transform (DWT) method was used as a reconstruction technique, in combination with the univariate long short-term memory (LSTM) network method. The combination of the two methods showed consistently high values of performance indicators, such as Nash–Sutcliffe efficiency (NSE), correlation coefficient (CC), Kling–Gupta efficiency (KGE), index of agreement (IA), and Legates–McCabe index (LMI), with mean average percentage error (MAPE) values at almost zero, and consistently low values for both residual mean square error (RMSE) and RMSE-observations standard deviation ratio (RSR). The authors believe that the proposed method can give efficient, time-bound results to flood-prone countries such as the Philippines, where hydrological data are deficient. |
| first_indexed | 2024-03-11T05:15:32Z |
| format | Article |
| id | doaj.art-f511af3294c3486caf4a33de6e818cc4 |
| institution | Directory Open Access Journal |
| issn | 2073-4433 |
| language | English |
| last_indexed | 2024-03-11T05:15:32Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Atmosphere |
| spelling | doaj.art-f511af3294c3486caf4a33de6e818cc42023-11-17T18:16:37ZengMDPI AGAtmosphere2073-44332023-03-0114463210.3390/atmos14040632Deep Learning-Based Univariate Prediction of Daily Rainfall: Application to a Flood-Prone, Data-Deficient CountryImee V. Necesito0Donghyun Kim1Young Hye Bae2Kyunghun Kim3Soojun Kim4Hung Soo Kim5Department of Civil Engineering, Inha University, Incheon 22212, Republic of KoreaInstitute of Water Resources System, Inha University, Incheon 22212, Republic of KoreaInstitute of Water Resources System, Inha University, Incheon 22212, Republic of KoreaDepartment of Civil Engineering, Inha University, Incheon 22212, Republic of KoreaDepartment of Civil Engineering, Inha University, Incheon 22212, Republic of KoreaDepartment of Civil Engineering, Inha University, Incheon 22212, Republic of KoreaThere are several attempts to model rainfall time series which have been explored by members of the hydrological research communities. Rainfall, being one of the defining factors for a flooding event, is rarely modeled singularly in deep learning, as it is usually performed in multivariate analysis. This study will attempt to explore a time series modeling method in four subcatchments located in Samar, Philippines. In this study, the rainfall time series was treated as a signal and was reconstructed into a combination of a ‘smoothened’ or ‘denoised’ signal, and a ‘detailed’ or noise signal. The discrete wavelet transform (DWT) method was used as a reconstruction technique, in combination with the univariate long short-term memory (LSTM) network method. The combination of the two methods showed consistently high values of performance indicators, such as Nash–Sutcliffe efficiency (NSE), correlation coefficient (CC), Kling–Gupta efficiency (KGE), index of agreement (IA), and Legates–McCabe index (LMI), with mean average percentage error (MAPE) values at almost zero, and consistently low values for both residual mean square error (RMSE) and RMSE-observations standard deviation ratio (RSR). The authors believe that the proposed method can give efficient, time-bound results to flood-prone countries such as the Philippines, where hydrological data are deficient.https://www.mdpi.com/2073-4433/14/4/632discrete wavelet transformlong short-term memory networkrainfall |
| spellingShingle | Imee V. Necesito Donghyun Kim Young Hye Bae Kyunghun Kim Soojun Kim Hung Soo Kim Deep Learning-Based Univariate Prediction of Daily Rainfall: Application to a Flood-Prone, Data-Deficient Country Atmosphere discrete wavelet transform long short-term memory network rainfall |
| title | Deep Learning-Based Univariate Prediction of Daily Rainfall: Application to a Flood-Prone, Data-Deficient Country |
| title_full | Deep Learning-Based Univariate Prediction of Daily Rainfall: Application to a Flood-Prone, Data-Deficient Country |
| title_fullStr | Deep Learning-Based Univariate Prediction of Daily Rainfall: Application to a Flood-Prone, Data-Deficient Country |
| title_full_unstemmed | Deep Learning-Based Univariate Prediction of Daily Rainfall: Application to a Flood-Prone, Data-Deficient Country |
| title_short | Deep Learning-Based Univariate Prediction of Daily Rainfall: Application to a Flood-Prone, Data-Deficient Country |
| title_sort | deep learning based univariate prediction of daily rainfall application to a flood prone data deficient country |
| topic | discrete wavelet transform long short-term memory network rainfall |
| url | https://www.mdpi.com/2073-4433/14/4/632 |
| work_keys_str_mv | AT imeevnecesito deeplearningbasedunivariatepredictionofdailyrainfallapplicationtoafloodpronedatadeficientcountry AT donghyunkim deeplearningbasedunivariatepredictionofdailyrainfallapplicationtoafloodpronedatadeficientcountry AT younghyebae deeplearningbasedunivariatepredictionofdailyrainfallapplicationtoafloodpronedatadeficientcountry AT kyunghunkim deeplearningbasedunivariatepredictionofdailyrainfallapplicationtoafloodpronedatadeficientcountry AT soojunkim deeplearningbasedunivariatepredictionofdailyrainfallapplicationtoafloodpronedatadeficientcountry AT hungsookim deeplearningbasedunivariatepredictionofdailyrainfallapplicationtoafloodpronedatadeficientcountry |