热带气象学报2026，Vol.42Issue(4)：555-566,12.DOI:10.16032/j.issn.1004-4965.2026.052
恩平市空气负氧离子时空分布特征及日变化预测研究
Spatiotemporal Distribution Characteristics and Daily Variation Prediction of Air Negative Oxygen Ions in Enping
摘要
Abstract
Negative oxygen ion(NOI)concentration is one of the crucial indicators for air quality.Using the observational data from four NOI stations in Enping from 2022 to 2023,this study analyzed the spatiotemporal variations of NOIs.Based on multiple linear regression(MLR)and Light Gradient Boosting Machine(LightGBM)algorithm,the daily prediction models for NOI variations were established,incorporating meteorological and ecological factors.The results indicate that Enping has high NOI levels which are highly beneficial to human health.Overall,the annual mean NOI concentrations reaches 2 697 cm-3.Spatially,the ranking of NOI concentrations across stations are as follows:Qixingkeng＞Meteorological Bureau＞Aofeng Mountain＞Qingnanjiao.Diurnally,the NOI concentration exhibited a bimodal pattern,with two peaks at 05:00 BJT and 16:00 BJT,and in-between a trough at 10:00 BJT.The monthly averages of NOI concentrations were higher from May to September and lower in other months.Correlation analysis revealed that the daily average NOI concentrations were significantly positively correlated with relative humidity,precipitation,and temperature,but negatively correlated with average atmospheric pressure and PM2.5 concentrations.Both models tended to underestimate relatively larger observed values.However,compared to the MLR model,the LightGBM model demonstrated superior performance,yielding a lower root mean square error(RMSE),a lower absolute mean error(MAE)and a coefficient of determination closer to 1.These findings suggest that the LightGBM model provides a better fit to the observed data,thus a more accurate and robust tool for forecasting daily NOI variations in Enping.关键词
负氧离子/时空分布特征/模型预测/LightGBM机器学习Key words
negative oxygen ions(NOI)/spatiotemporal distribution characteristics/model-based prediction/LightGBM machine learning分类
天文与地球科学引用本文复制引用
何博翰,姚钰玲,朱怀卫,温文,肖悦,刘敏..恩平市空气负氧离子时空分布特征及日变化预测研究[J].热带气象学报,2026,42(4):555-566,12.基金项目
广东省气象局科学技术研究项目(GRMC2025M50) （GRMC2025M50）
广东省生态气象中心科学技术研究项目(STZX215040005) （STZX215040005）
广东省气象局科学技术研究项目(GRMC2023M38)共同资助 （GRMC2023M38）
