Geographic Graph Network for Robust Inversion of Particulate Matters

Although remote sensors have been increasingly providing dense data and deriving reanalysis data for inversion of particulate matters, the use of these data is considerably limited by the ground monitoring samples and conventional machine learning models. As regional criteria air pollutants, particu...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autor principal:	Lianfa Li
Formato:	article
Lenguaje:	EN
Publicado:	MDPI AG 2021
Materias:	geographic graph hybrid network graph convolution neighborhood feature PM<sub>2.5</sub> PM<sub>10</sub> spatiotemporal modeling Science Q
Acceso en línea:	https://doaj.org/article/5c9ce40044584b0c87f31633bf973a43
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:5c9ce40044584b0c87f31633bf973a43
record_format	dspace
spelling	oai:doaj.org-article:5c9ce40044584b0c87f31633bf973a432021-11-11T18:54:12ZGeographic Graph Network for Robust Inversion of Particulate Matters10.3390/rs132143412072-4292https://doaj.org/article/5c9ce40044584b0c87f31633bf973a432021-10-01T00:00:00Zhttps://www.mdpi.com/2072-4292/13/21/4341https://doaj.org/toc/2072-4292Although remote sensors have been increasingly providing dense data and deriving reanalysis data for inversion of particulate matters, the use of these data is considerably limited by the ground monitoring samples and conventional machine learning models. As regional criteria air pollutants, particulate matters present a strong spatial correlation of long range. Conventional machine learning cannot or can only model such spatial pattern in a limited way. Here, we propose a method of a geographic graph hybrid network to encode a spatial neighborhood feature to make robust estimation of coarse and fine particulate matters (PM<sub>10</sub> and PM<sub>2.5</sub>). Based on Tobler’s First Law of Geography and graph convolutions, we constructed the architecture of a geographic graph hybrid network, in which full residual deep layers were connected with graph convolutions to reduce over-smoothing, subject to the PM<sub>10</sub>–PM<sub>2.5</sub> relationship constraint. In the site-based independent test in mainland China (2015–2018), our method achieved much better generalization than typical state-of-the-art methods (improvement in R<sup>2</sup>: 8–78%, decrease in RMSE: 14–48%). This study shows that the proposed method can encode the neighborhood information and can make an important contribution to improvement in generalization and extrapolation of geo-features with strong spatial correlation, such as PM<sub>2.5</sub> and PM<sub>10</sub>.Lianfa LiMDPI AGarticlegeographic graph hybrid networkgraph convolutionneighborhood featurePM<sub>2.5</sub>PM<sub>10</sub>spatiotemporal modelingScienceQENRemote Sensing, Vol 13, Iss 4341, p 4341 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	geographic graph hybrid network graph convolution neighborhood feature PM<sub>2.5</sub> PM<sub>10</sub> spatiotemporal modeling Science Q
spellingShingle	geographic graph hybrid network graph convolution neighborhood feature PM<sub>2.5</sub> PM<sub>10</sub> spatiotemporal modeling Science Q Lianfa Li Geographic Graph Network for Robust Inversion of Particulate Matters
description	Although remote sensors have been increasingly providing dense data and deriving reanalysis data for inversion of particulate matters, the use of these data is considerably limited by the ground monitoring samples and conventional machine learning models. As regional criteria air pollutants, particulate matters present a strong spatial correlation of long range. Conventional machine learning cannot or can only model such spatial pattern in a limited way. Here, we propose a method of a geographic graph hybrid network to encode a spatial neighborhood feature to make robust estimation of coarse and fine particulate matters (PM<sub>10</sub> and PM<sub>2.5</sub>). Based on Tobler’s First Law of Geography and graph convolutions, we constructed the architecture of a geographic graph hybrid network, in which full residual deep layers were connected with graph convolutions to reduce over-smoothing, subject to the PM<sub>10</sub>–PM<sub>2.5</sub> relationship constraint. In the site-based independent test in mainland China (2015–2018), our method achieved much better generalization than typical state-of-the-art methods (improvement in R<sup>2</sup>: 8–78%, decrease in RMSE: 14–48%). This study shows that the proposed method can encode the neighborhood information and can make an important contribution to improvement in generalization and extrapolation of geo-features with strong spatial correlation, such as PM<sub>2.5</sub> and PM<sub>10</sub>.
format	article
author	Lianfa Li
author_facet	Lianfa Li
author_sort	Lianfa Li
title	Geographic Graph Network for Robust Inversion of Particulate Matters
title_short	Geographic Graph Network for Robust Inversion of Particulate Matters
title_full	Geographic Graph Network for Robust Inversion of Particulate Matters
title_fullStr	Geographic Graph Network for Robust Inversion of Particulate Matters
title_full_unstemmed	Geographic Graph Network for Robust Inversion of Particulate Matters
title_sort	geographic graph network for robust inversion of particulate matters
publisher	MDPI AG
publishDate	2021
url	https://doaj.org/article/5c9ce40044584b0c87f31633bf973a43
work_keys_str_mv	AT lianfali geographicgraphnetworkforrobustinversionofparticulatematters
_version_	1718431647306612736

Geographic Graph Network for Robust Inversion of Particulate Matters

Ejemplares similares