Knowledge-guided machine learning reveals pivotal drivers for gas-to-particle conversion of atmospheric nitrate

Aug 3, 2026

Particulate nitrate, a key component of fine particles, forms via the gas-to-particle conversion process governed by the nitrate conversion coefficient ε(NO). The ε(NO)–driver relationship is complex, nonlinear, and sensitive to ambient conditions, making conventional machine learning prone to results with limited physical meaning and even potential conflicts with established mechanisms. A supervised framework, multilevel nested random forest guided by theory, is presented to obtain interpretable driver identification. It robustly highlights NH, SO24, and temperature as pivotal drivers, while clarifying distinct daytime and nighttime contributions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Abstract

Particulate nitrate formation and ε(NO)

Limits of conventional machine learning

Theory-guided multilevel nested random forest

Identified key drivers and day/night differences

University of Birmingham Knowledge-guided machine learning reveals pivotal drivers for gas-to-particle conversion of atmospheric nitrate Xu, Bo; Yu, Haofei; Shi, Zongbo; Liu, Jinxing; Wei, Yuting; Zhang, Zhongcheng; Huangfu, Yanqi; Xu, Han; Li, Yue; Zhang, Linlin; Feng, Yinchang; Shi, Guoliang DOI: 10.1016/j.ese.2023.100333 License: Creative Commons: Attribution-NonCommercial-NoDerivs (CC BY-NC-ND) Document Version Publisher's PDF, also known as Version of record Citation for published version (Harvard): Xu, B, Yu, H, Shi, Z, Liu, J, Wei, Y, Zhang, Z, Huangfu, Y, Xu, H, Li, Y, Zhang, L, Feng, Y & Shi, G 2024,'Knowledge-guided machine learning reveals pivotal drivers for gas-to-particle conversion of atmospheric nitrate', Environmental Science and Ecotechnology, vol. 19, 100333. [https://doi.org/10.1016/j.ese.2023.100333](https://doi.org/10.1016/j.ese.2023.100333) Link to publication on Research at Birmingham portal General rights Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes permitted by law. •Users may freely distribute the URL that is used to identify this publication. •Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private study or non-commercial research. •User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?) •Users may not further distribute the material nor use it for the purposes of commercial gain. Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document. When citing, please reference the published version. Take down policy While the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has been uploaded in error or has been deemed to be commercially or otherwise sensitive. If you believe that this is the case for this document, [please contact UBIRA@lists.bham.ac.uk](please contact UBIRA@lists.bham.ac.uk) providing details and we will remove access to the work immediately and investigate. Download date: 02. Aug. 2026 Environmental Science and Ecotechnology 19 (2024) 100333 Contents lists available at ScienceDirect Environmental Science and Ecotechnology journal [homepage: www.journals. elsevier. com/environmental-science-and](homepage: www.journals. elsevier. com/environmental-science-and) ecotechnology/ | Original Research | | | | | --- | --- | --- | --- | | Knowledge-guided machine learning reveals pivotal drivers for gasto-particle conversion of atmospheric nitrate | | | | | Bo Xu a, b, Haofei Yu c, Zongbo Shi d, Jinxing Liu e, f, Yuting Wei a, b, Zhongcheng Zhang a, b, Yanqi Huangfu a, b, Han Xu a, b, Yue Li g, Linlin Zhang h, **, Yinchang Feng a, b,
Guoliang Shi a, b, *
a State Environmental Protection Key Laboratory of Urban Ambient Air Particulate Matter Pollution Prevention and Control, Tianjin Key Laboratory of Urban Transport Emission Research, College of Environmental Science and Engineering, Nankai University, Tianjin, 300350, China
b CMA-NKU Cooperative Laboratory for Atmospheric Environment-Health Research (CLAER), College of Environmental Science and Engineering, Nankai University, Tianjin, 300350, China
c Department of Civil, Environmental, and Construction Engineering, University of Central Florida, Orlando, FL, USA
d School of Geography Earth and Environment Sciences, University of Birmingham, Birmingham, B15 2TT, UK
e State Key Laboratory of Precision Measuring Technology and Instruments, Tianjin Key Laboratory of air Pollutants Monitoring Technology, School of Precision Instrument and Opto-electronics

FAQs

What controls the gas-to-particle conversion of atmospheric nitrate in this study?

The study focuses on the nitrate gas-to-particle conversion coefficient, ε(NO), which regulates how particulate nitrate forms from atmospheric gases.

Why are conventional machine learning approaches considered insufficient here?

Conventional methods may produce outputs with unclear physical interpretation and can even contradict known physical or chemical mechanisms due to ambient influences.

Which drivers are identified as pivotal for ε(NO), and how does the study treat daytime versus nighttime?

The approach highlights NH, SO24−, and temperature as key drivers, emphasizing NH contributions during both daytime (about 30%) and nighttime (about 40%), with reduced emphasis on less relevant factors compared with traditional random forest results.

Aug 3, 2026

Upload by Olivia Brown