Estimating Surface Downward Longwave Radiation Using Machine Learning Methods
The downward longwave radiation (<i>L<sub>d</sub></i>, 4–100 μm) is a major component of research for the surface radiation energy budget and balance. In this study, we applied five machine learning methods, namely artificial neural network (ANN), support vector regression (S...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-10-01
|
| Series: | Atmosphere |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4433/11/11/1147 |
| _version_ | 1827703726080524288 |
|---|---|
| author | Chunjie Feng Xiaotong Zhang Yu Wei Weiyu Zhang Ning Hou Jiawen Xu Kun Jia Yunjun Yao Xianhong Xie Bo Jiang Jie Cheng Xiang Zhao |
| author_facet | Chunjie Feng Xiaotong Zhang Yu Wei Weiyu Zhang Ning Hou Jiawen Xu Kun Jia Yunjun Yao Xianhong Xie Bo Jiang Jie Cheng Xiang Zhao |
| author_sort | Chunjie Feng |
| collection | DOAJ |
| description | The downward longwave radiation (<i>L<sub>d</sub></i>, 4–100 μm) is a major component of research for the surface radiation energy budget and balance. In this study, we applied five machine learning methods, namely artificial neural network (ANN), support vector regression (SVR), gradient boosting regression tree (GBRT), random forest (RF), and multivariate adaptive regression spline (MARS), to estimate <i>L<sub>d</sub></i> using ground measurements collected from 27 Baseline Surface Radiation Network (BSRN) stations. <i>L<sub>d</sub></i> measurements in situ were used to validate the accuracy of <i>L<sub>d</sub></i> estimation models on daily and monthly time scales. A comparison of the results demonstrated that the estimates on the basis of the GBRT method had the highest accuracy, with an overall root-mean-square error (RMSE) of 17.50 W m<sup>−2</sup> and an R value of 0.96 for the test dataset on a daily time scale. These values were 11.19 W m<sup>−2</sup> and 0.98, respectively, on a monthly time scale. The effects of land cover and elevation were further studied to comprehensively evaluate the performance of each machine learning method. All machine learning methods achieved better results over the grass land cover type but relatively worse results over the tundra. GBRT, RF, and MARS methods were found to show good performance at both the high- and low-altitude sites. |
| first_indexed | 2024-03-10T15:24:57Z |
| format | Article |
| id | doaj.art-c16bbea363a34c6b9c7fabf93e6350e8 |
| institution | Directory Open Access Journal |
| issn | 2073-4433 |
| language | English |
| last_indexed | 2024-03-10T15:24:57Z |
| publishDate | 2020-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Atmosphere |
| spelling | doaj.art-c16bbea363a34c6b9c7fabf93e6350e82023-11-20T18:10:12ZengMDPI AGAtmosphere2073-44332020-10-011111114710.3390/atmos11111147Estimating Surface Downward Longwave Radiation Using Machine Learning MethodsChunjie Feng0Xiaotong Zhang1Yu Wei2Weiyu Zhang3Ning Hou4Jiawen Xu5Kun Jia6Yunjun Yao7Xianhong Xie8Bo Jiang9Jie Cheng10Xiang Zhao11State Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaState Key Laboratory of Remote Sensing Science, Jointly Sponsored by Beijing Normal University and Institute of Remote Sensing and Digital Earth of Chinese Academy of Sciences, Beijing 100875, ChinaThe downward longwave radiation (<i>L<sub>d</sub></i>, 4–100 μm) is a major component of research for the surface radiation energy budget and balance. In this study, we applied five machine learning methods, namely artificial neural network (ANN), support vector regression (SVR), gradient boosting regression tree (GBRT), random forest (RF), and multivariate adaptive regression spline (MARS), to estimate <i>L<sub>d</sub></i> using ground measurements collected from 27 Baseline Surface Radiation Network (BSRN) stations. <i>L<sub>d</sub></i> measurements in situ were used to validate the accuracy of <i>L<sub>d</sub></i> estimation models on daily and monthly time scales. A comparison of the results demonstrated that the estimates on the basis of the GBRT method had the highest accuracy, with an overall root-mean-square error (RMSE) of 17.50 W m<sup>−2</sup> and an R value of 0.96 for the test dataset on a daily time scale. These values were 11.19 W m<sup>−2</sup> and 0.98, respectively, on a monthly time scale. The effects of land cover and elevation were further studied to comprehensively evaluate the performance of each machine learning method. All machine learning methods achieved better results over the grass land cover type but relatively worse results over the tundra. GBRT, RF, and MARS methods were found to show good performance at both the high- and low-altitude sites.https://www.mdpi.com/2073-4433/11/11/1147downward longwave radiationmachine learningGBRTenergy budgetrandom forest |
| spellingShingle | Chunjie Feng Xiaotong Zhang Yu Wei Weiyu Zhang Ning Hou Jiawen Xu Kun Jia Yunjun Yao Xianhong Xie Bo Jiang Jie Cheng Xiang Zhao Estimating Surface Downward Longwave Radiation Using Machine Learning Methods Atmosphere downward longwave radiation machine learning GBRT energy budget random forest |
| title | Estimating Surface Downward Longwave Radiation Using Machine Learning Methods |
| title_full | Estimating Surface Downward Longwave Radiation Using Machine Learning Methods |
| title_fullStr | Estimating Surface Downward Longwave Radiation Using Machine Learning Methods |
| title_full_unstemmed | Estimating Surface Downward Longwave Radiation Using Machine Learning Methods |
| title_short | Estimating Surface Downward Longwave Radiation Using Machine Learning Methods |
| title_sort | estimating surface downward longwave radiation using machine learning methods |
| topic | downward longwave radiation machine learning GBRT energy budget random forest |
| url | https://www.mdpi.com/2073-4433/11/11/1147 |
| work_keys_str_mv | AT chunjiefeng estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT xiaotongzhang estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT yuwei estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT weiyuzhang estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT ninghou estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT jiawenxu estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT kunjia estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT yunjunyao estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT xianhongxie estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT bojiang estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT jiecheng estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods AT xiangzhao estimatingsurfacedownwardlongwaveradiationusingmachinelearningmethods |