Comparison of simulating visibility using XGBoost and IMPROVE method: a case study in East China

The prediction accuracy of atmospheric visibility significantly impacts daily life. However, there is a relative scarcity of research on post-processing methods for visibility obtained from the WRF-Chem atmospheric chemistry model results. In order to explore a more accurate method for visibility ca...

Full description

Saved in:
Bibliographic Details
Main Authors: Xin Zhang, Yue Wang, Zibo Zhuang, Yuxi Liu, Chengduo Yuan, Lei Su, Jingyuan Shao, Pak-Wai Chan
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-01-01
Series:Frontiers in Environmental Science
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fenvs.2024.1534113/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1832593933304070144
author Xin Zhang
Yue Wang
Zibo Zhuang
Yuxi Liu
Chengduo Yuan
Lei Su
Jingyuan Shao
Pak-Wai Chan
author_facet Xin Zhang
Yue Wang
Zibo Zhuang
Yuxi Liu
Chengduo Yuan
Lei Su
Jingyuan Shao
Pak-Wai Chan
author_sort Xin Zhang
collection DOAJ
description The prediction accuracy of atmospheric visibility significantly impacts daily life. However, there is a relative scarcity of research on post-processing methods for visibility obtained from the WRF-Chem atmospheric chemistry model results. In order to explore a more accurate method for visibility calculation, we conducted a study on the meteorological conditions in the East China region during a heavy pollution period from October 1 to 23 in the year of 2022. The meteorological data were processed using both the XGBoost (XGB) model and the IMPROVE to calculate visibility. The results indicate that XGB outperforms the IMPROVE in various aspects. The visibility improved from a correlation of 0.56–0.71 with the use of XGB. And in comparison with the IMPROVE equation, XGB demonstrated a statistically significant reduction in RMSE by 1.96 km. Even in regions where the IMPROVE performs poorly, XGB demonstrates superior performance. In regions where the correlation simulated by the IMPROVE equation is less than 0.2 (Anqing and Nanyang), XGB still performs well, achieving correlations of 0.69 (Anqing) and 0.68 (Nanyang). Throughout the entire study period, the average visibility results obtained by XGB deviate by only 0.07 km from the observed values. These findings underscore the importance of incorporating the XGBoost model into WRF-Chem visibility simulations, as it significantly improves the accuracy of visibility predictions.
format Article
id doaj-art-6ed29e9523e9441191699c031993099f
institution Kabale University
issn 2296-665X
language English
publishDate 2025-01-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Environmental Science
spelling doaj-art-6ed29e9523e9441191699c031993099f2025-01-20T07:19:56ZengFrontiers Media S.A.Frontiers in Environmental Science2296-665X2025-01-011210.3389/fenvs.2024.15341131534113Comparison of simulating visibility using XGBoost and IMPROVE method: a case study in East ChinaXin Zhang0Yue Wang1Zibo Zhuang2Yuxi Liu3Chengduo Yuan4Lei Su5Jingyuan Shao6Pak-Wai Chan7Flight Academy of Civil Aviation University of China, Tianjin, ChinaDepartment of Safety Science and Engineering, Civil Aviation University of China, Tianjin, ChinaAviation Meteorological Research Institute, Civil Aviation University of China, Tianjin, ChinaState Environmental Protection Key Laboratory of Environmental Pollution and Greenhouse Gases Co-control, Chinese Academy of Environmental Planning, Beijing, ChinaDepartment of Flight Area Management, Harbin International Airport Co., Ltd., Harbin, ChinaDepartment of Safety Science and Engineering, Civil Aviation University of China, Tianjin, ChinaFlight Academy of Civil Aviation University of China, Tianjin, ChinaHong Kong Observatory, Hong Kong, Hong Kong SAR, ChinaThe prediction accuracy of atmospheric visibility significantly impacts daily life. However, there is a relative scarcity of research on post-processing methods for visibility obtained from the WRF-Chem atmospheric chemistry model results. In order to explore a more accurate method for visibility calculation, we conducted a study on the meteorological conditions in the East China region during a heavy pollution period from October 1 to 23 in the year of 2022. The meteorological data were processed using both the XGBoost (XGB) model and the IMPROVE to calculate visibility. The results indicate that XGB outperforms the IMPROVE in various aspects. The visibility improved from a correlation of 0.56–0.71 with the use of XGB. And in comparison with the IMPROVE equation, XGB demonstrated a statistically significant reduction in RMSE by 1.96 km. Even in regions where the IMPROVE performs poorly, XGB demonstrates superior performance. In regions where the correlation simulated by the IMPROVE equation is less than 0.2 (Anqing and Nanyang), XGB still performs well, achieving correlations of 0.69 (Anqing) and 0.68 (Nanyang). Throughout the entire study period, the average visibility results obtained by XGB deviate by only 0.07 km from the observed values. These findings underscore the importance of incorporating the XGBoost model into WRF-Chem visibility simulations, as it significantly improves the accuracy of visibility predictions.https://www.frontiersin.org/articles/10.3389/fenvs.2024.1534113/fullvisibility simulationXGBoostIMPROVE equationWRF-Chem modelmachine learningatmospheric chemistry
spellingShingle Xin Zhang
Yue Wang
Zibo Zhuang
Yuxi Liu
Chengduo Yuan
Lei Su
Jingyuan Shao
Pak-Wai Chan
Comparison of simulating visibility using XGBoost and IMPROVE method: a case study in East China
Frontiers in Environmental Science
visibility simulation
XGBoost
IMPROVE equation
WRF-Chem model
machine learning
atmospheric chemistry
title Comparison of simulating visibility using XGBoost and IMPROVE method: a case study in East China
title_full Comparison of simulating visibility using XGBoost and IMPROVE method: a case study in East China
title_fullStr Comparison of simulating visibility using XGBoost and IMPROVE method: a case study in East China
title_full_unstemmed Comparison of simulating visibility using XGBoost and IMPROVE method: a case study in East China
title_short Comparison of simulating visibility using XGBoost and IMPROVE method: a case study in East China
title_sort comparison of simulating visibility using xgboost and improve method a case study in east china
topic visibility simulation
XGBoost
IMPROVE equation
WRF-Chem model
machine learning
atmospheric chemistry
url https://www.frontiersin.org/articles/10.3389/fenvs.2024.1534113/full
work_keys_str_mv AT xinzhang comparisonofsimulatingvisibilityusingxgboostandimprovemethodacasestudyineastchina
AT yuewang comparisonofsimulatingvisibilityusingxgboostandimprovemethodacasestudyineastchina
AT zibozhuang comparisonofsimulatingvisibilityusingxgboostandimprovemethodacasestudyineastchina
AT yuxiliu comparisonofsimulatingvisibilityusingxgboostandimprovemethodacasestudyineastchina
AT chengduoyuan comparisonofsimulatingvisibilityusingxgboostandimprovemethodacasestudyineastchina
AT leisu comparisonofsimulatingvisibilityusingxgboostandimprovemethodacasestudyineastchina
AT jingyuanshao comparisonofsimulatingvisibilityusingxgboostandimprovemethodacasestudyineastchina
AT pakwaichan comparisonofsimulatingvisibilityusingxgboostandimprovemethodacasestudyineastchina