Spatiotemporal imputation of MAIAC AOD using deep learning with downscaling
Aerosols have adverse health effects and play a significant role in the climate as well. The Multiangle Implementation of Atmospheric Correction (MAIAC) provides Aerosol Optical Depth (AOD) at high temporal (daily) and spatial (1 km) resolution, making it particularly useful to infer and characteriz...
Main Authors: | , , , , , , , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | unknown |
Published: |
eScholarship, University of California
2020
|
Subjects: | |
Online Access: | https://escholarship.org/uc/item/1rm291mz |
id |
ftcdlib:oai:escholarship.org:ark:/13030/qt1rm291mz |
---|---|
record_format |
openpolar |
spelling |
ftcdlib:oai:escholarship.org:ark:/13030/qt1rm291mz 2023-10-01T03:49:49+02:00 Spatiotemporal imputation of MAIAC AOD using deep learning with downscaling Li, Lianfa Franklin, Meredith Girguis, Mariam Lurmann, Frederick Wu, Jun Pavlovic, Nathan Breton, Carrie Gilliland, Frank Habre, Rima 2020-02-01 application/pdf https://escholarship.org/uc/item/1rm291mz unknown eScholarship, University of California qt1rm291mz https://escholarship.org/uc/item/1rm291mz public Bioengineering Clinical Research Aerosol Optical Depth MAIAC MERRA-2 GMI Replay Simulation Deep learning Downscaling Missingness imputation Air quality Physical Geography and Environmental Geoscience Geomatic Engineering Geological & Geomatics Engineering article 2020 ftcdlib 2023-09-04T18:04:00Z Aerosols have adverse health effects and play a significant role in the climate as well. The Multiangle Implementation of Atmospheric Correction (MAIAC) provides Aerosol Optical Depth (AOD) at high temporal (daily) and spatial (1 km) resolution, making it particularly useful to infer and characterize spatiotemporal variability of aerosols at a fine spatial scale for exposure assessment and health studies. However, clouds and conditions of high surface reflectance result in a significant proportion of missing MAIAC AOD. To fill these gaps, we present an imputation approach using deep learning with downscaling. Using a baseline autoencoder, we leverage residual connections in deep neural networks to boost learning and parameter sharing to reduce overfitting, and conduct bagging to reduce error variance in the imputations. Downscaled through a similar auto-encoder based deep residual network, Modern-Era Retrospective analysis for Research and Applications Version 2 (MERRA-2) GMI Replay Simulation (M2GMI) data were introduced to the network as an important gap-filling feature that varies in space to be used for missingness imputations. Imputing weekly MAIAC AOD from 2000 to 2016 over California, a state with considerable geographic heterogeneity, our full (non-full) residual network achieved mean R2 = 0.94 (0.86) [RMSE = 0.007 (0.01)] in an independent test, showing considerably better performance than a regular neural network or non-linear generalized additive model (mean R2 = 0.78-0.81; mean RMSE = 0.013-0.015). The adjusted imputed as well as combined imputed and observed MAIAC AOD showed strong correlation with Aerosol Robotic Network (AERONET) AOD (R = 0.83; R2 = 0.69, RMSE = 0.04). Our results show that we can generate reliable imputations of missing AOD through a deep learning approach, having important downstream air quality modeling applications. Article in Journal/Newspaper Aerosol Robotic Network University of California: eScholarship Merra ENVELOPE(12.615,12.615,65.816,65.816) |
institution |
Open Polar |
collection |
University of California: eScholarship |
op_collection_id |
ftcdlib |
language |
unknown |
topic |
Bioengineering Clinical Research Aerosol Optical Depth MAIAC MERRA-2 GMI Replay Simulation Deep learning Downscaling Missingness imputation Air quality Physical Geography and Environmental Geoscience Geomatic Engineering Geological & Geomatics Engineering |
spellingShingle |
Bioengineering Clinical Research Aerosol Optical Depth MAIAC MERRA-2 GMI Replay Simulation Deep learning Downscaling Missingness imputation Air quality Physical Geography and Environmental Geoscience Geomatic Engineering Geological & Geomatics Engineering Li, Lianfa Franklin, Meredith Girguis, Mariam Lurmann, Frederick Wu, Jun Pavlovic, Nathan Breton, Carrie Gilliland, Frank Habre, Rima Spatiotemporal imputation of MAIAC AOD using deep learning with downscaling |
topic_facet |
Bioengineering Clinical Research Aerosol Optical Depth MAIAC MERRA-2 GMI Replay Simulation Deep learning Downscaling Missingness imputation Air quality Physical Geography and Environmental Geoscience Geomatic Engineering Geological & Geomatics Engineering |
description |
Aerosols have adverse health effects and play a significant role in the climate as well. The Multiangle Implementation of Atmospheric Correction (MAIAC) provides Aerosol Optical Depth (AOD) at high temporal (daily) and spatial (1 km) resolution, making it particularly useful to infer and characterize spatiotemporal variability of aerosols at a fine spatial scale for exposure assessment and health studies. However, clouds and conditions of high surface reflectance result in a significant proportion of missing MAIAC AOD. To fill these gaps, we present an imputation approach using deep learning with downscaling. Using a baseline autoencoder, we leverage residual connections in deep neural networks to boost learning and parameter sharing to reduce overfitting, and conduct bagging to reduce error variance in the imputations. Downscaled through a similar auto-encoder based deep residual network, Modern-Era Retrospective analysis for Research and Applications Version 2 (MERRA-2) GMI Replay Simulation (M2GMI) data were introduced to the network as an important gap-filling feature that varies in space to be used for missingness imputations. Imputing weekly MAIAC AOD from 2000 to 2016 over California, a state with considerable geographic heterogeneity, our full (non-full) residual network achieved mean R2 = 0.94 (0.86) [RMSE = 0.007 (0.01)] in an independent test, showing considerably better performance than a regular neural network or non-linear generalized additive model (mean R2 = 0.78-0.81; mean RMSE = 0.013-0.015). The adjusted imputed as well as combined imputed and observed MAIAC AOD showed strong correlation with Aerosol Robotic Network (AERONET) AOD (R = 0.83; R2 = 0.69, RMSE = 0.04). Our results show that we can generate reliable imputations of missing AOD through a deep learning approach, having important downstream air quality modeling applications. |
format |
Article in Journal/Newspaper |
author |
Li, Lianfa Franklin, Meredith Girguis, Mariam Lurmann, Frederick Wu, Jun Pavlovic, Nathan Breton, Carrie Gilliland, Frank Habre, Rima |
author_facet |
Li, Lianfa Franklin, Meredith Girguis, Mariam Lurmann, Frederick Wu, Jun Pavlovic, Nathan Breton, Carrie Gilliland, Frank Habre, Rima |
author_sort |
Li, Lianfa |
title |
Spatiotemporal imputation of MAIAC AOD using deep learning with downscaling |
title_short |
Spatiotemporal imputation of MAIAC AOD using deep learning with downscaling |
title_full |
Spatiotemporal imputation of MAIAC AOD using deep learning with downscaling |
title_fullStr |
Spatiotemporal imputation of MAIAC AOD using deep learning with downscaling |
title_full_unstemmed |
Spatiotemporal imputation of MAIAC AOD using deep learning with downscaling |
title_sort |
spatiotemporal imputation of maiac aod using deep learning with downscaling |
publisher |
eScholarship, University of California |
publishDate |
2020 |
url |
https://escholarship.org/uc/item/1rm291mz |
long_lat |
ENVELOPE(12.615,12.615,65.816,65.816) |
geographic |
Merra |
geographic_facet |
Merra |
genre |
Aerosol Robotic Network |
genre_facet |
Aerosol Robotic Network |
op_relation |
qt1rm291mz https://escholarship.org/uc/item/1rm291mz |
op_rights |
public |
_version_ |
1778534134492168192 |