Combining Machine Learning and Numerical Simulation for High-Resolution PM2.5 Concentration Forecast

Forecasting ambient PM2.5 concentrations with spatiotemporal coverage is key to alerting decision makers of pollution episodes and preventing detrimental public exposure, especially in regions with limited ground air monitoring stations. The existing methods rely on either chemical transport models...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Environmental science & technology Ročník 56; číslo 3; s. 1544
Hlavní autori: Bi, Jianzhao, Knowland, K Emma, Keller, Christoph A, Liu, Yang
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: 01.02.2022
ISSN:1520-5851, 1520-5851
On-line prístup:Zistit podrobnosti o prístupe
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract Forecasting ambient PM2.5 concentrations with spatiotemporal coverage is key to alerting decision makers of pollution episodes and preventing detrimental public exposure, especially in regions with limited ground air monitoring stations. The existing methods rely on either chemical transport models (CTMs) to forecast spatial distribution of PM2.5 with nontrivial uncertainty or statistical algorithms to forecast PM2.5 concentration time series at air monitoring locations without continuous spatial coverage. In this study, we developed a PM2.5 forecast framework by combining the robust Random Forest algorithm with a publicly accessible global CTM forecast product, NASA's Goddard Earth Observing System "Composition Forecasting" (GEOS-CF), providing spatiotemporally continuous PM2.5 concentration forecasts for the next 5 days at a 1 km spatial resolution. Our forecast experiment was conducted for a region in Central China including the populous and polluted Fenwei Plain. The forecast for the next 2 days had an overall validation R2 of 0.76 and 0.64, respectively; the R2 was around 0.5 for the following 3 forecast days. Spatial cross-validation showed similar validation metrics. Our forecast model, with a validation normalized mean bias close to 0, substantially reduced the large biases in GEOS-CF. The proposed framework requires minimal computational resources compared to running CTMs at urban scales, enabling near-real-time PM2.5 forecast in resource-restricted environments.Forecasting ambient PM2.5 concentrations with spatiotemporal coverage is key to alerting decision makers of pollution episodes and preventing detrimental public exposure, especially in regions with limited ground air monitoring stations. The existing methods rely on either chemical transport models (CTMs) to forecast spatial distribution of PM2.5 with nontrivial uncertainty or statistical algorithms to forecast PM2.5 concentration time series at air monitoring locations without continuous spatial coverage. In this study, we developed a PM2.5 forecast framework by combining the robust Random Forest algorithm with a publicly accessible global CTM forecast product, NASA's Goddard Earth Observing System "Composition Forecasting" (GEOS-CF), providing spatiotemporally continuous PM2.5 concentration forecasts for the next 5 days at a 1 km spatial resolution. Our forecast experiment was conducted for a region in Central China including the populous and polluted Fenwei Plain. The forecast for the next 2 days had an overall validation R2 of 0.76 and 0.64, respectively; the R2 was around 0.5 for the following 3 forecast days. Spatial cross-validation showed similar validation metrics. Our forecast model, with a validation normalized mean bias close to 0, substantially reduced the large biases in GEOS-CF. The proposed framework requires minimal computational resources compared to running CTMs at urban scales, enabling near-real-time PM2.5 forecast in resource-restricted environments.
AbstractList Forecasting ambient PM2.5 concentrations with spatiotemporal coverage is key to alerting decision makers of pollution episodes and preventing detrimental public exposure, especially in regions with limited ground air monitoring stations. The existing methods rely on either chemical transport models (CTMs) to forecast spatial distribution of PM2.5 with nontrivial uncertainty or statistical algorithms to forecast PM2.5 concentration time series at air monitoring locations without continuous spatial coverage. In this study, we developed a PM2.5 forecast framework by combining the robust Random Forest algorithm with a publicly accessible global CTM forecast product, NASA's Goddard Earth Observing System "Composition Forecasting" (GEOS-CF), providing spatiotemporally continuous PM2.5 concentration forecasts for the next 5 days at a 1 km spatial resolution. Our forecast experiment was conducted for a region in Central China including the populous and polluted Fenwei Plain. The forecast for the next 2 days had an overall validation R2 of 0.76 and 0.64, respectively; the R2 was around 0.5 for the following 3 forecast days. Spatial cross-validation showed similar validation metrics. Our forecast model, with a validation normalized mean bias close to 0, substantially reduced the large biases in GEOS-CF. The proposed framework requires minimal computational resources compared to running CTMs at urban scales, enabling near-real-time PM2.5 forecast in resource-restricted environments.Forecasting ambient PM2.5 concentrations with spatiotemporal coverage is key to alerting decision makers of pollution episodes and preventing detrimental public exposure, especially in regions with limited ground air monitoring stations. The existing methods rely on either chemical transport models (CTMs) to forecast spatial distribution of PM2.5 with nontrivial uncertainty or statistical algorithms to forecast PM2.5 concentration time series at air monitoring locations without continuous spatial coverage. In this study, we developed a PM2.5 forecast framework by combining the robust Random Forest algorithm with a publicly accessible global CTM forecast product, NASA's Goddard Earth Observing System "Composition Forecasting" (GEOS-CF), providing spatiotemporally continuous PM2.5 concentration forecasts for the next 5 days at a 1 km spatial resolution. Our forecast experiment was conducted for a region in Central China including the populous and polluted Fenwei Plain. The forecast for the next 2 days had an overall validation R2 of 0.76 and 0.64, respectively; the R2 was around 0.5 for the following 3 forecast days. Spatial cross-validation showed similar validation metrics. Our forecast model, with a validation normalized mean bias close to 0, substantially reduced the large biases in GEOS-CF. The proposed framework requires minimal computational resources compared to running CTMs at urban scales, enabling near-real-time PM2.5 forecast in resource-restricted environments.
Author Bi, Jianzhao
Liu, Yang
Knowland, K Emma
Keller, Christoph A
Author_xml – sequence: 1
  givenname: Jianzhao
  surname: Bi
  fullname: Bi, Jianzhao
– sequence: 2
  givenname: K Emma
  surname: Knowland
  fullname: Knowland, K Emma
– sequence: 3
  givenname: Christoph A
  surname: Keller
  fullname: Keller, Christoph A
– sequence: 4
  givenname: Yang
  surname: Liu
  fullname: Liu, Yang
BookMark eNpNkD1PwzAYhC1UJNrCzOqRJcEfsWNGFFGKlALiY67e2G9ao8SGOPn_VC0D050enU66W5BZiAEJueYs50zwW7ApxzTm3DKlSnNG5lwJlimj-OyfvyCLlL4YY0IyMyeuin3jgw87ugG79wFpjTAcAQRHn6ceB2-ho---nzoYfQy0jQNd-90-e8MUu-nIXjciV7SKwWIYh1NuFQe0kMZLct5Cl_DqT5fkc_XwUa2z-uXxqbqvM5BMj5njViNKKYRqyrYpTFEajto0II0sNSqGTljn0GqnDzNNoV2pwHLQ4LDRYkluTr3fQ_yZDmdse58sdh0EjFPaCs3vBC-ZkuIXKHxdYg
ContentType Journal Article
DBID 7X8
DOI 10.1021/acs.est.1c05578
DatabaseName MEDLINE - Academic
DatabaseTitle MEDLINE - Academic
DatabaseTitleList MEDLINE - Academic
Database_xml – sequence: 1
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod no_fulltext_linktorsrc
Discipline Engineering
Environmental Sciences
EISSN 1520-5851
GroupedDBID ---
-DZ
-~X
..I
.DC
.K2
3R3
4.4
4R4
53G
55A
5GY
5VS
6TJ
7X8
7~N
85S
AABXI
AAHBH
ABBLG
ABJNI
ABLBI
ABMVS
ABOGM
ABPPZ
ABQRX
ABUCX
ACGFS
ACGOD
ACIWK
ACJ
ACPRK
ACS
ADHLV
ADUKH
AEESW
AENEX
AFEFF
AFRAH
AGXLV
AHGAQ
ALMA_UNASSIGNED_HOLDINGS
AQSVZ
BAANH
BKOMP
CS3
CUPRZ
EBS
ED~
F5P
GGK
GNL
IH9
JG~
LG6
MS~
MW2
PQQKQ
ROL
RXW
TN5
TWZ
U5U
UHB
UI2
UKR
UPT
VF5
VG9
W1F
WH7
XSW
XZL
YZZ
ZCA
ID FETCH-LOGICAL-a306t-d1c6ee33225b7fb484781e68ba38376e50ed2cddec6d61c0846d75ac1a6adeb62
IEDL.DBID 7X8
ISICitedReferencesCount 49
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000743731400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1520-5851
IngestDate Thu Oct 02 10:28:33 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 3
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-a306t-d1c6ee33225b7fb484781e68ba38376e50ed2cddec6d61c0846d75ac1a6adeb62
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
OpenAccessLink https://ntrs.nasa.gov/citations/20220002174
PQID 2619217053
PQPubID 23479
ParticipantIDs proquest_miscellaneous_2619217053
PublicationCentury 2000
PublicationDate 2022-02-01
PublicationDateYYYYMMDD 2022-02-01
PublicationDate_xml – month: 02
  year: 2022
  text: 2022-02-01
  day: 01
PublicationDecade 2020
PublicationTitle Environmental science & technology
PublicationYear 2022
SSID ssj0002308
Score 2.5797904
Snippet Forecasting ambient PM2.5 concentrations with spatiotemporal coverage is key to alerting decision makers of pollution episodes and preventing detrimental...
SourceID proquest
SourceType Aggregation Database
StartPage 1544
Title Combining Machine Learning and Numerical Simulation for High-Resolution PM2.5 Concentration Forecast
URI https://www.proquest.com/docview/2619217053
Volume 56
WOSCitedRecordID wos000743731400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LTwIxEJ6oeNCDD5T4Tk28FvbhtrsnYwjEgxAOariRvjAcXNAFf78zpYiJFxPv3WzTnZn9pjPfNwA3hc6k04Xi2kYRRwsxXKdJzLOC9M4wgRae9f7yKPv9fDgsBuHCrQptlauY6AO1nRq6I295pE_aL-nd7J3T1CiqroYRGptQSxHKkFXL4VotHOG1p8JlmCJR-etb2iduKVM16RWxIRGq_Fck9r-X7v5_N3YAewFYsvulJRzChivrsPtDbrAOjc6a1YZLg1tXR2AxKmg_KYL1fHOlY0F39ZWp0rL-YlnXwWcmb2HeF0O0y6hLhFMFYGm_bNBLmhlrExWyDHq8jIZ_GlXNj-G523lqP_AwfoEr_ERzbmMjnEvJ47Uc69ucWKlO5FpRVitcFjmbGAyPRliB54hIxspMmVgJZZ0WSQO2ymnpToBJzOoSBJqGaLGpigqV6jy3Mi6UdGNXnML16mxHaN5Us1Clmy6q0fp0z_6w5hx2EqIn-K7qC6iN0YXdJWybz_mk-rjy1vEFSunFSA
linkProvider ProQuest
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Combining+Machine+Learning+and+Numerical+Simulation+for+High-Resolution+PM2.5+Concentration+Forecast&rft.jtitle=Environmental+science+%26+technology&rft.au=Bi%2C+Jianzhao&rft.au=Knowland%2C+K+Emma&rft.au=Keller%2C+Christoph+A&rft.au=Liu%2C+Yang&rft.date=2022-02-01&rft.issn=1520-5851&rft.eissn=1520-5851&rft.volume=56&rft.issue=3&rft.spage=1544&rft_id=info:doi/10.1021%2Facs.est.1c05578&rft.externalDBID=NO_FULL_TEXT
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1520-5851&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1520-5851&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1520-5851&client=summon