File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Conference Paper: A Local Spatial Kriging Applied to the PM2.5 Concentration Estimation

TitleA Local Spatial Kriging Applied to the PM2.5 Concentration Estimation
Authors
KeywordsAnisotropy
Fine particulate matter
Kriging
Point pattern
Issue Date2021
Citation
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2021, v. 12753 LNCS, p. 205-221 How to Cite?
AbstractAir pollution in China has aroused special concerns from the general public in recent years. Accurately estimating air pollutant concentrations, especially fine particulate matter (i.e., PM2.5), is of importance to understand the distribution patterns of pollutants and then facilitate the control of their emissions and the reduction of public exposure. While estimating ambient PM2.5 concentrations based on monitoring station measurements is inherently complex, geostatistical spatial interpolation could provide a potential solution. However, standard Kriging may fail in yielding an accurate estimate as it assumes a universally homogeneous semi-variogram and lacks in considering the non-stationary spatial process that arises from anisotropy and non-randomized point pattern. This study proposes an algorithm named Point Pattern Local Anisotropy Kriging (PPLAKriging) that can derive a locally adaptive semi-variogram. Specifically, this algorithm determines an optimum neighboring search range in terms of the local distribution pattern of surrounding points followed by a local anisotropy detection. A geographical coordinates transformation is then performed for ease of calculation. To validate the proposed model in interpolating PM2.5 concentration, a regional experiment is conducted. The results show that there are substantial benefits in considering local features under different neighborhood environmental conditions. In the experiment, PPLAKriging proves to be superior to the standard ordinary Kriging, root-mean-squared-error decrease by 33.9%.
Persistent Identifierhttp://hdl.handle.net/10722/329961
ISSN
2023 SCImago Journal Rankings: 0.606

 

DC FieldValueLanguage
dc.contributor.authorYao, Shiqi-
dc.contributor.authorHuang, Bo-
dc.date.accessioned2023-08-09T03:36:46Z-
dc.date.available2023-08-09T03:36:46Z-
dc.date.issued2021-
dc.identifier.citationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2021, v. 12753 LNCS, p. 205-221-
dc.identifier.issn0302-9743-
dc.identifier.urihttp://hdl.handle.net/10722/329961-
dc.description.abstractAir pollution in China has aroused special concerns from the general public in recent years. Accurately estimating air pollutant concentrations, especially fine particulate matter (i.e., PM2.5), is of importance to understand the distribution patterns of pollutants and then facilitate the control of their emissions and the reduction of public exposure. While estimating ambient PM2.5 concentrations based on monitoring station measurements is inherently complex, geostatistical spatial interpolation could provide a potential solution. However, standard Kriging may fail in yielding an accurate estimate as it assumes a universally homogeneous semi-variogram and lacks in considering the non-stationary spatial process that arises from anisotropy and non-randomized point pattern. This study proposes an algorithm named Point Pattern Local Anisotropy Kriging (PPLAKriging) that can derive a locally adaptive semi-variogram. Specifically, this algorithm determines an optimum neighboring search range in terms of the local distribution pattern of surrounding points followed by a local anisotropy detection. A geographical coordinates transformation is then performed for ease of calculation. To validate the proposed model in interpolating PM2.5 concentration, a regional experiment is conducted. The results show that there are substantial benefits in considering local features under different neighborhood environmental conditions. In the experiment, PPLAKriging proves to be superior to the standard ordinary Kriging, root-mean-squared-error decrease by 33.9%.-
dc.languageeng-
dc.relation.ispartofLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)-
dc.subjectAnisotropy-
dc.subjectFine particulate matter-
dc.subjectKriging-
dc.subjectPoint pattern-
dc.titleA Local Spatial Kriging Applied to the PM2.5 Concentration Estimation-
dc.typeConference_Paper-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1007/978-3-030-85462-1_19-
dc.identifier.scopuseid_2-s2.0-85115165676-
dc.identifier.volume12753 LNCS-
dc.identifier.spage205-
dc.identifier.epage221-
dc.identifier.eissn1611-3349-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats