Remote Sensing Fine Estimation Model of PM<sub>2.5</sub> Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in China
The accurate prediction of PM<sub>2.5</sub> concentration across extensive temporal and spatial scales is essential for air pollution control and safeguarding public health. To address the challenges of the uneven coverage and limited number of traditional PM<sub>2.5</sub> gr...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-11-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/16/22/4306 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1846152525267337216 |
|---|---|
| author | Yiye Ji Yanjun Wang Cheng Wang Xuchao Tang Mengru Song |
| author_facet | Yiye Ji Yanjun Wang Cheng Wang Xuchao Tang Mengru Song |
| author_sort | Yiye Ji |
| collection | DOAJ |
| description | The accurate prediction of PM<sub>2.5</sub> concentration across extensive temporal and spatial scales is essential for air pollution control and safeguarding public health. To address the challenges of the uneven coverage and limited number of traditional PM<sub>2.5</sub> ground monitoring networks, the low inversion accuracy of PM<sub>2.5</sub> concentration, and the incomplete understanding of its spatiotemporal dynamics, this study proposes a refined PM<sub>2.5</sub> concentration estimation model, Bi-LSTM-SA, integrating multi-source remote sensing data. First, utilizing multi-source remote sensing data, such as MODIS aerosol optical depth (AOD) products, meteorological data, and PM<sub>2.5</sub> monitoring sites, AERONET AOD was used to validate the accuracy of the MODIS AOD data. Variables including temperature (TEMP), relative humidity (RH), surface pressure (SP), wind speed (WS), and total precipitation (PRE) were selected, followed by the application of the variance inflation factor (VIF) and Pearson’s correlation coefficient (R) for variable screening. Second, to effectively capture temporal dependencies and emphasize key features, an improved Long Short-Term Memory Network (LSTM) model, Bi-LSTM-SA, was constructed by combining a bidirectional LSTM (Bi-LSTM) model with a self-adaptive attention mechanism (SA). This model was evaluated through ablation and comparative experiments using three cross-validation methods: sample-based, temporal, and spatial. The effectiveness of this method was demonstrated on Beijing–Tianjin–Hebei urban agglomeration, achieving a coefficient of determination (R<sup>2</sup>) of 0.89, root mean squared error (RMSE) of 12.76 μg/m<sup>3</sup>, and mean absolute error (MAE) of 8.27 μg/m<sup>3</sup>. Finally, this model was applied to predict PM<sub>2.5</sub> concentration on Beijing–Tianjin–Hebei urban agglomeration in 2023, revealing the characteristics of its spatiotemporal evolution. Additionally, the results indicated that this model performs exceptionally well in hourly PM<sub>2.5</sub> concentration forecasting and can be used for PM<sub>2.5</sub> concentration hourly prediction tasks. This study provides technical support for the large-scale, accurate remote sensing inversion of PM<sub>2.5</sub> concentration and offers fundamental insights for regional atmospheric environmental protection. |
| format | Article |
| id | doaj-art-d201f68d3b204d67a092e33fde1e1500 |
| institution | Kabale University |
| issn | 2072-4292 |
| language | English |
| publishDate | 2024-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj-art-d201f68d3b204d67a092e33fde1e15002024-11-26T18:20:21ZengMDPI AGRemote Sensing2072-42922024-11-011622430610.3390/rs16224306Remote Sensing Fine Estimation Model of PM<sub>2.5</sub> Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in ChinaYiye Ji0Yanjun Wang1Cheng Wang2Xuchao Tang3Mengru Song4School of Earth Science and Space Information Engineering, Hunan University of Science and Technology, Xiangtan 411201, ChinaSchool of Earth Science and Space Information Engineering, Hunan University of Science and Technology, Xiangtan 411201, ChinaInstitute of Aerospace Information Innovation, Chinese Academy of Sciences, Beijing 100864, ChinaSchool of Earth Science and Space Information Engineering, Hunan University of Science and Technology, Xiangtan 411201, ChinaSchool of Earth Science and Space Information Engineering, Hunan University of Science and Technology, Xiangtan 411201, ChinaThe accurate prediction of PM<sub>2.5</sub> concentration across extensive temporal and spatial scales is essential for air pollution control and safeguarding public health. To address the challenges of the uneven coverage and limited number of traditional PM<sub>2.5</sub> ground monitoring networks, the low inversion accuracy of PM<sub>2.5</sub> concentration, and the incomplete understanding of its spatiotemporal dynamics, this study proposes a refined PM<sub>2.5</sub> concentration estimation model, Bi-LSTM-SA, integrating multi-source remote sensing data. First, utilizing multi-source remote sensing data, such as MODIS aerosol optical depth (AOD) products, meteorological data, and PM<sub>2.5</sub> monitoring sites, AERONET AOD was used to validate the accuracy of the MODIS AOD data. Variables including temperature (TEMP), relative humidity (RH), surface pressure (SP), wind speed (WS), and total precipitation (PRE) were selected, followed by the application of the variance inflation factor (VIF) and Pearson’s correlation coefficient (R) for variable screening. Second, to effectively capture temporal dependencies and emphasize key features, an improved Long Short-Term Memory Network (LSTM) model, Bi-LSTM-SA, was constructed by combining a bidirectional LSTM (Bi-LSTM) model with a self-adaptive attention mechanism (SA). This model was evaluated through ablation and comparative experiments using three cross-validation methods: sample-based, temporal, and spatial. The effectiveness of this method was demonstrated on Beijing–Tianjin–Hebei urban agglomeration, achieving a coefficient of determination (R<sup>2</sup>) of 0.89, root mean squared error (RMSE) of 12.76 μg/m<sup>3</sup>, and mean absolute error (MAE) of 8.27 μg/m<sup>3</sup>. Finally, this model was applied to predict PM<sub>2.5</sub> concentration on Beijing–Tianjin–Hebei urban agglomeration in 2023, revealing the characteristics of its spatiotemporal evolution. Additionally, the results indicated that this model performs exceptionally well in hourly PM<sub>2.5</sub> concentration forecasting and can be used for PM<sub>2.5</sub> concentration hourly prediction tasks. This study provides technical support for the large-scale, accurate remote sensing inversion of PM<sub>2.5</sub> concentration and offers fundamental insights for regional atmospheric environmental protection.https://www.mdpi.com/2072-4292/16/22/4306MODIS AODself-adaptive attention mechanismlong and short-term memory networkdeep learningPM<sub>2.5</sub> concentration prediction |
| spellingShingle | Yiye Ji Yanjun Wang Cheng Wang Xuchao Tang Mengru Song Remote Sensing Fine Estimation Model of PM<sub>2.5</sub> Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in China Remote Sensing MODIS AOD self-adaptive attention mechanism long and short-term memory network deep learning PM<sub>2.5</sub> concentration prediction |
| title | Remote Sensing Fine Estimation Model of PM<sub>2.5</sub> Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in China |
| title_full | Remote Sensing Fine Estimation Model of PM<sub>2.5</sub> Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in China |
| title_fullStr | Remote Sensing Fine Estimation Model of PM<sub>2.5</sub> Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in China |
| title_full_unstemmed | Remote Sensing Fine Estimation Model of PM<sub>2.5</sub> Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in China |
| title_short | Remote Sensing Fine Estimation Model of PM<sub>2.5</sub> Concentration Based on Improved Long Short-Term Memory Network: A Case Study on Beijing–Tianjin–Hebei Urban Agglomeration in China |
| title_sort | remote sensing fine estimation model of pm sub 2 5 sub concentration based on improved long short term memory network a case study on beijing tianjin hebei urban agglomeration in china |
| topic | MODIS AOD self-adaptive attention mechanism long and short-term memory network deep learning PM<sub>2.5</sub> concentration prediction |
| url | https://www.mdpi.com/2072-4292/16/22/4306 |
| work_keys_str_mv | AT yiyeji remotesensingfineestimationmodelofpmsub25subconcentrationbasedonimprovedlongshorttermmemorynetworkacasestudyonbeijingtianjinhebeiurbanagglomerationinchina AT yanjunwang remotesensingfineestimationmodelofpmsub25subconcentrationbasedonimprovedlongshorttermmemorynetworkacasestudyonbeijingtianjinhebeiurbanagglomerationinchina AT chengwang remotesensingfineestimationmodelofpmsub25subconcentrationbasedonimprovedlongshorttermmemorynetworkacasestudyonbeijingtianjinhebeiurbanagglomerationinchina AT xuchaotang remotesensingfineestimationmodelofpmsub25subconcentrationbasedonimprovedlongshorttermmemorynetworkacasestudyonbeijingtianjinhebeiurbanagglomerationinchina AT mengrusong remotesensingfineestimationmodelofpmsub25subconcentrationbasedonimprovedlongshorttermmemorynetworkacasestudyonbeijingtianjinhebeiurbanagglomerationinchina |