Enhanced Generalized Regression Neural Network for Soil Moisture Estimation Over the Qinghai-Tibet Plateau
Soil moisture (SM) is a critical parameter in maintaining the balance of water cycle and energy budgets between climate system and the Earth's environment. Generalized regression neural network (GRNN) has been substantially verified as a powerful model for SM estimation due to the ability of ca...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9763319/ |
| _version_ | 1818208756461207552 |
|---|---|
| author | Ling Zhang Zhaohui Xue Yujuan Zhang Jiayi Ma Hao Li |
| author_facet | Ling Zhang Zhaohui Xue Yujuan Zhang Jiayi Ma Hao Li |
| author_sort | Ling Zhang |
| collection | DOAJ |
| description | Soil moisture (SM) is a critical parameter in maintaining the balance of water cycle and energy budgets between climate system and the Earth's environment. Generalized regression neural network (GRNN) has been substantially verified as a powerful model for SM estimation due to the ability of capturing complex, non-linear relationships between predictors and responses. However, GRNN builds a full adjacency matrix using Gaussian kernel, which is computationally expensive and may ignore the local structure. In addition, it is laborious to optimize the “spread” parameter. To overcome the above issues, we propose an enhanced generalized regression neural network (EGRNN) for SM estimation, where two main adaptations are made. On the one hand, the City block distance instead of the Euclidean distance is used for building Gaussian kernel. On the other hand, <italic>k</italic>-nearest neighbors (<italic>k</italic>-NN) is adopted to yield an empirically sparse adjacency matrix. As the key advantage, the proposed EGRNN weakens the sensitivity to outliers since large differences are weighted more heavily by using Euclidean distance than City block distance. Another advantage is that EGRNN models more local and discriminant information in the pattern layer since only the data points within neighbors are connected by using <italic>k</italic>-NN. Experiments conducted in the Qinghai-Tibet Plateau (QTP) demonstrate that; 1) EGRNN outperforms the other four neural network models, with R = 0.9485 and RMSE = 0.0325 cm<sup>3</sup>/cm<sup>3</sup>; 2) It can well capture spatial-temporal dynamics and has higher consistent with the in-situ measurements; 3) It adapts well to different in-situ networks and has better generalization performance. |
| first_indexed | 2024-12-12T04:49:52Z |
| format | Article |
| id | doaj.art-6f62d04fa3784b2991ec97b8570372ea |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-12-12T04:49:52Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-6f62d04fa3784b2991ec97b8570372ea2022-12-22T00:37:31ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352022-01-01153815382910.1109/JSTARS.2022.31669789763319Enhanced Generalized Regression Neural Network for Soil Moisture Estimation Over the Qinghai-Tibet PlateauLing Zhang0Zhaohui Xue1https://orcid.org/0000-0002-1672-317XYujuan Zhang2Jiayi Ma3https://orcid.org/0000-0002-0723-653XHao Li4https://orcid.org/0000-0002-5678-7282School of Earth Sciences and Engineering, Hohai University, Nanjing, ChinaSchool of Earth Sciences and Engineering, Hohai University, Nanjing, ChinaSchool of Earth Sciences and Engineering, Hohai University, Nanjing, ChinaSchool of Earth Sciences and Engineering, Hohai University, Nanjing, ChinaSchool of Earth Sciences and Engineering, Hohai University, Nanjing, ChinaSoil moisture (SM) is a critical parameter in maintaining the balance of water cycle and energy budgets between climate system and the Earth's environment. Generalized regression neural network (GRNN) has been substantially verified as a powerful model for SM estimation due to the ability of capturing complex, non-linear relationships between predictors and responses. However, GRNN builds a full adjacency matrix using Gaussian kernel, which is computationally expensive and may ignore the local structure. In addition, it is laborious to optimize the “spread” parameter. To overcome the above issues, we propose an enhanced generalized regression neural network (EGRNN) for SM estimation, where two main adaptations are made. On the one hand, the City block distance instead of the Euclidean distance is used for building Gaussian kernel. On the other hand, <italic>k</italic>-nearest neighbors (<italic>k</italic>-NN) is adopted to yield an empirically sparse adjacency matrix. As the key advantage, the proposed EGRNN weakens the sensitivity to outliers since large differences are weighted more heavily by using Euclidean distance than City block distance. Another advantage is that EGRNN models more local and discriminant information in the pattern layer since only the data points within neighbors are connected by using <italic>k</italic>-NN. Experiments conducted in the Qinghai-Tibet Plateau (QTP) demonstrate that; 1) EGRNN outperforms the other four neural network models, with R = 0.9485 and RMSE = 0.0325 cm<sup>3</sup>/cm<sup>3</sup>; 2) It can well capture spatial-temporal dynamics and has higher consistent with the in-situ measurements; 3) It adapts well to different in-situ networks and has better generalization performance.https://ieeexplore.ieee.org/document/9763319/City block distancegeneralized regression neural network (GRNN)Qinghai-Tibet plateau (QTP)soil moisture (SM) estimationsparse adjacency matrix |
| spellingShingle | Ling Zhang Zhaohui Xue Yujuan Zhang Jiayi Ma Hao Li Enhanced Generalized Regression Neural Network for Soil Moisture Estimation Over the Qinghai-Tibet Plateau IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing City block distance generalized regression neural network (GRNN) Qinghai-Tibet plateau (QTP) soil moisture (SM) estimation sparse adjacency matrix |
| title | Enhanced Generalized Regression Neural Network for Soil Moisture Estimation Over the Qinghai-Tibet Plateau |
| title_full | Enhanced Generalized Regression Neural Network for Soil Moisture Estimation Over the Qinghai-Tibet Plateau |
| title_fullStr | Enhanced Generalized Regression Neural Network for Soil Moisture Estimation Over the Qinghai-Tibet Plateau |
| title_full_unstemmed | Enhanced Generalized Regression Neural Network for Soil Moisture Estimation Over the Qinghai-Tibet Plateau |
| title_short | Enhanced Generalized Regression Neural Network for Soil Moisture Estimation Over the Qinghai-Tibet Plateau |
| title_sort | enhanced generalized regression neural network for soil moisture estimation over the qinghai tibet plateau |
| topic | City block distance generalized regression neural network (GRNN) Qinghai-Tibet plateau (QTP) soil moisture (SM) estimation sparse adjacency matrix |
| url | https://ieeexplore.ieee.org/document/9763319/ |
| work_keys_str_mv | AT lingzhang enhancedgeneralizedregressionneuralnetworkforsoilmoistureestimationovertheqinghaitibetplateau AT zhaohuixue enhancedgeneralizedregressionneuralnetworkforsoilmoistureestimationovertheqinghaitibetplateau AT yujuanzhang enhancedgeneralizedregressionneuralnetworkforsoilmoistureestimationovertheqinghaitibetplateau AT jiayima enhancedgeneralizedregressionneuralnetworkforsoilmoistureestimationovertheqinghaitibetplateau AT haoli enhancedgeneralizedregressionneuralnetworkforsoilmoistureestimationovertheqinghaitibetplateau |