High-Resolution Estimation of Daily PM<sub>2.5</sub> Levels in the Contiguous US Using Bi-LSTM with Attention
Estimating surface-level PM<sub>2.5</sub> concentrations at any given location is crucial for public health monitoring and cohort studies. Existing models and datasets for this purpose have limited precision, especially on high-concentration days. Additionally, due to the lack of open-so...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/17/1/126 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841549017868140544 |
|---|---|
| author | Zhongying Wang James L. Crooks Elizabeth Anne Regan Morteza Karimzadeh |
| author_facet | Zhongying Wang James L. Crooks Elizabeth Anne Regan Morteza Karimzadeh |
| author_sort | Zhongying Wang |
| collection | DOAJ |
| description | Estimating surface-level PM<sub>2.5</sub> concentrations at any given location is crucial for public health monitoring and cohort studies. Existing models and datasets for this purpose have limited precision, especially on high-concentration days. Additionally, due to the lack of open-source code, generating estimates for other areas and time periods remains cumbersome. We developed a novel deep learning-based model that improves the surface-level PM<sub>2.5</sub> concentration estimates by capitalizing on the temporal dynamics of air quality. Specifically, we improve the estimation precision by developing a Long Short-Term Memory (LSTM) network with Attention and integrating multiple data sources, including in situ measurements, remotely sensed data, and wildfire smoke density observations, which improve the model’s ability to capture high-concentration events. We rigorously evaluate the model against existing products, demonstrating a 2.2% improvement in overall RMSE, and a 9.8% reduction in RMSE on high-concentration days, highlighting the superior performance of our approach, particularly on high-concentration days. Using the model, we have produced a comprehensive dataset of PM<sub>2.5</sub> estimates from 2005 to 2021 for the contiguous United States and are releasing an open-source framework to ensure reproducibility and facilitate further adaptation in air quality studies. |
| format | Article |
| id | doaj-art-bbcc50dc62934c5a94d647e689d0eba5 |
| institution | Kabale University |
| issn | 2072-4292 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj-art-bbcc50dc62934c5a94d647e689d0eba52025-01-10T13:20:18ZengMDPI AGRemote Sensing2072-42922025-01-0117112610.3390/rs17010126High-Resolution Estimation of Daily PM<sub>2.5</sub> Levels in the Contiguous US Using Bi-LSTM with AttentionZhongying Wang0James L. Crooks1Elizabeth Anne Regan2Morteza Karimzadeh3Department of Geography, University of Colorado Boulder, Boulder, CO 80302, USANational Jewish Health, Denver, CO 80206, USANational Jewish Health, Denver, CO 80206, USADepartment of Geography, University of Colorado Boulder, Boulder, CO 80302, USAEstimating surface-level PM<sub>2.5</sub> concentrations at any given location is crucial for public health monitoring and cohort studies. Existing models and datasets for this purpose have limited precision, especially on high-concentration days. Additionally, due to the lack of open-source code, generating estimates for other areas and time periods remains cumbersome. We developed a novel deep learning-based model that improves the surface-level PM<sub>2.5</sub> concentration estimates by capitalizing on the temporal dynamics of air quality. Specifically, we improve the estimation precision by developing a Long Short-Term Memory (LSTM) network with Attention and integrating multiple data sources, including in situ measurements, remotely sensed data, and wildfire smoke density observations, which improve the model’s ability to capture high-concentration events. We rigorously evaluate the model against existing products, demonstrating a 2.2% improvement in overall RMSE, and a 9.8% reduction in RMSE on high-concentration days, highlighting the superior performance of our approach, particularly on high-concentration days. Using the model, we have produced a comprehensive dataset of PM<sub>2.5</sub> estimates from 2005 to 2021 for the contiguous United States and are releasing an open-source framework to ensure reproducibility and facilitate further adaptation in air quality studies.https://www.mdpi.com/2072-4292/17/1/126air pollutionPM<sub>2.5</sub>deep learningspatiotemporal modelingpublic dataset |
| spellingShingle | Zhongying Wang James L. Crooks Elizabeth Anne Regan Morteza Karimzadeh High-Resolution Estimation of Daily PM<sub>2.5</sub> Levels in the Contiguous US Using Bi-LSTM with Attention Remote Sensing air pollution PM<sub>2.5</sub> deep learning spatiotemporal modeling public dataset |
| title | High-Resolution Estimation of Daily PM<sub>2.5</sub> Levels in the Contiguous US Using Bi-LSTM with Attention |
| title_full | High-Resolution Estimation of Daily PM<sub>2.5</sub> Levels in the Contiguous US Using Bi-LSTM with Attention |
| title_fullStr | High-Resolution Estimation of Daily PM<sub>2.5</sub> Levels in the Contiguous US Using Bi-LSTM with Attention |
| title_full_unstemmed | High-Resolution Estimation of Daily PM<sub>2.5</sub> Levels in the Contiguous US Using Bi-LSTM with Attention |
| title_short | High-Resolution Estimation of Daily PM<sub>2.5</sub> Levels in the Contiguous US Using Bi-LSTM with Attention |
| title_sort | high resolution estimation of daily pm sub 2 5 sub levels in the contiguous us using bi lstm with attention |
| topic | air pollution PM<sub>2.5</sub> deep learning spatiotemporal modeling public dataset |
| url | https://www.mdpi.com/2072-4292/17/1/126 |
| work_keys_str_mv | AT zhongyingwang highresolutionestimationofdailypmsub25sublevelsinthecontiguousususingbilstmwithattention AT jameslcrooks highresolutionestimationofdailypmsub25sublevelsinthecontiguousususingbilstmwithattention AT elizabethanneregan highresolutionestimationofdailypmsub25sublevelsinthecontiguousususingbilstmwithattention AT mortezakarimzadeh highresolutionestimationofdailypmsub25sublevelsinthecontiguousususingbilstmwithattention |