Interpretable spatio-temporal modeling for soil temperature prediction
Soil temperature (ST) is a crucial parameter in Earth system science. Accurate ST predictions provide invaluable insights; however, the “black box” nature of many deep learning approaches limits their interpretability. In this study, we present the Encoder-Decoder Model with Interpretable Spatio-Tem...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-12-01
|
| Series: | Frontiers in Forests and Global Change |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/ffgc.2023.1295731/full |
| _version_ | 1797384312966873088 |
|---|---|
| author | Xiaoning Li Xiaoning Li Yuheng Zhu Qingliang Li Hongwei Zhao Jinlong Zhu Cheng Zhang |
| author_facet | Xiaoning Li Xiaoning Li Yuheng Zhu Qingliang Li Hongwei Zhao Jinlong Zhu Cheng Zhang |
| author_sort | Xiaoning Li |
| collection | DOAJ |
| description | Soil temperature (ST) is a crucial parameter in Earth system science. Accurate ST predictions provide invaluable insights; however, the “black box” nature of many deep learning approaches limits their interpretability. In this study, we present the Encoder-Decoder Model with Interpretable Spatio-Temporal Component (ISDNM) to enhance both ST prediction accuracy and its spatio-temporal interpretability. The ISDNM combines a CNN-encoder-decoder and an LSTM-encoder-decoder to improve spatio-temporal feature representation. It further uses linear regression and Uniform Manifold Approximation and Projection (UMAP) techniques for clearer spatio-temporal visualization of ST. The results show that the ISDNM model had the highest R2 ranging from 0.886 to 0.963 and the lowest RMSE ranging from 6.086 m3/m3 to 12.533 m3/m3 for different climate regions, and demonstrated superior performance than all the other DL models like CNN, LSTM, ConvLSTM models. The predictable component highlighted the remarkable similarity between Medium fine and Very fine soils in China. Additional, May and November emerged as crucial months, acting as inflection points in the annual ST cycle, shaping ISDNM model’s prediction capabilities. |
| first_indexed | 2024-03-08T21:33:40Z |
| format | Article |
| id | doaj.art-e39d4526b88f4ef09db693216f305db2 |
| institution | Directory Open Access Journal |
| issn | 2624-893X |
| language | English |
| last_indexed | 2024-03-08T21:33:40Z |
| publishDate | 2023-12-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Forests and Global Change |
| spelling | doaj.art-e39d4526b88f4ef09db693216f305db22023-12-21T04:35:07ZengFrontiers Media S.A.Frontiers in Forests and Global Change2624-893X2023-12-01610.3389/ffgc.2023.12957311295731Interpretable spatio-temporal modeling for soil temperature predictionXiaoning Li0Xiaoning Li1Yuheng Zhu2Qingliang Li3Hongwei Zhao4Jinlong Zhu5Cheng Zhang6College of Computer Science and Technology, Jilin University, Changchun, ChinaCollege of Computer Science and Technology, Changchun Normal University, Changchun, ChinaCollege of Computer Science and Technology, Jilin University, Changchun, ChinaCollege of Computer Science and Technology, Changchun Normal University, Changchun, ChinaCollege of Computer Science and Technology, Jilin University, Changchun, ChinaCollege of Computer Science and Technology, Changchun Normal University, Changchun, ChinaCollege of Computer Science and Technology, Changchun Normal University, Changchun, ChinaSoil temperature (ST) is a crucial parameter in Earth system science. Accurate ST predictions provide invaluable insights; however, the “black box” nature of many deep learning approaches limits their interpretability. In this study, we present the Encoder-Decoder Model with Interpretable Spatio-Temporal Component (ISDNM) to enhance both ST prediction accuracy and its spatio-temporal interpretability. The ISDNM combines a CNN-encoder-decoder and an LSTM-encoder-decoder to improve spatio-temporal feature representation. It further uses linear regression and Uniform Manifold Approximation and Projection (UMAP) techniques for clearer spatio-temporal visualization of ST. The results show that the ISDNM model had the highest R2 ranging from 0.886 to 0.963 and the lowest RMSE ranging from 6.086 m3/m3 to 12.533 m3/m3 for different climate regions, and demonstrated superior performance than all the other DL models like CNN, LSTM, ConvLSTM models. The predictable component highlighted the remarkable similarity between Medium fine and Very fine soils in China. Additional, May and November emerged as crucial months, acting as inflection points in the annual ST cycle, shaping ISDNM model’s prediction capabilities.https://www.frontiersin.org/articles/10.3389/ffgc.2023.1295731/fullsoil temperaturedeep learninginterpretable modelmachine learning (ML)LSTM (long short term memory networks) |
| spellingShingle | Xiaoning Li Xiaoning Li Yuheng Zhu Qingliang Li Hongwei Zhao Jinlong Zhu Cheng Zhang Interpretable spatio-temporal modeling for soil temperature prediction Frontiers in Forests and Global Change soil temperature deep learning interpretable model machine learning (ML) LSTM (long short term memory networks) |
| title | Interpretable spatio-temporal modeling for soil temperature prediction |
| title_full | Interpretable spatio-temporal modeling for soil temperature prediction |
| title_fullStr | Interpretable spatio-temporal modeling for soil temperature prediction |
| title_full_unstemmed | Interpretable spatio-temporal modeling for soil temperature prediction |
| title_short | Interpretable spatio-temporal modeling for soil temperature prediction |
| title_sort | interpretable spatio temporal modeling for soil temperature prediction |
| topic | soil temperature deep learning interpretable model machine learning (ML) LSTM (long short term memory networks) |
| url | https://www.frontiersin.org/articles/10.3389/ffgc.2023.1295731/full |
| work_keys_str_mv | AT xiaoningli interpretablespatiotemporalmodelingforsoiltemperatureprediction AT xiaoningli interpretablespatiotemporalmodelingforsoiltemperatureprediction AT yuhengzhu interpretablespatiotemporalmodelingforsoiltemperatureprediction AT qingliangli interpretablespatiotemporalmodelingforsoiltemperatureprediction AT hongweizhao interpretablespatiotemporalmodelingforsoiltemperatureprediction AT jinlongzhu interpretablespatiotemporalmodelingforsoiltemperatureprediction AT chengzhang interpretablespatiotemporalmodelingforsoiltemperatureprediction |