成都市PM2.5浓度变化的影响因素交互作用研究

为探究大气环境中污染物与气象要素交互作用对PM_2.5浓度变化的影响特征，利用成都市2014~2020年逐日大气污染物资料以及同期的气象资料，采用广义相加模型（GAMs）分析不同影响因素对当地PM_2.5浓度变化的影响效应.结果表明，单影响因素GAMs模型中，无论全年还是冬季，PM_2.5浓度与平均气温（T）、相对湿度（RH）、平均风速（Wind）、降水量（Prec）、O₃、NO₂、SO₂和CO间均呈非线性关系，其中CO、NO₂、SO₂、T和Wind对PM_2.5浓度影响较大，与全年不同的是，冬季T和O₃对PM_2.5浓度变化的影响效应较全年明显减弱.多影响因素的GAMs模型中，T、Wind、RH、CO、NO₂、SO₂和O₃这7个解释变量对PM_2.5浓度变化的影响均较显著，构建的全年多影响因素GAMs模型调整后的R²=0.759，方差解释率为76.42%，冬季R²=0.708，方差解释率为72.2%，无论是全年还是冬季，CO都是PM_2.5浓度变化的主导影响因素.GAMs交互效应模型发现，全年弱低温（7℃左右）+高相对湿度+高浓度CO+高浓度NO₂+高浓度SO₂协同作用条件下有利于PM_2.5浓度的生成；冬季低Wind+高RH+高浓度CO+高浓度NO₂+高浓度SO₂共存条件下有利于PM_2.5的生成，即该条件对PM_2.5浓度的生成有协同放大效应.运用GAMs模型能够对PM_2.5污染的主导影响因素进行识别，并定量化分析影响因素单效应及其交互作用对PM_2.5浓度变化的影响特征，对PM_2.5浓度污染防控研究具有重要指示意义.

Interactive effects of the influcening factors on the changes of PM2.5 concentration

To explore the influence characteristics of the interaction effects between meteorological elements and ambient air pollutants on particulate matter with an aerodynamic less than 2.5 (PM_2.5), daily air pollutants data and meteorological data during the same period from 2014 to 2020 in Chengdu were collected. Generalized Additive Models (GAMs) were adopted to explore the effects of different factors on PM_2.5 concentration of Chengdu.The results of single-factor GAMs showed that the relationship between PM_2.5 concentration and daily average temperature (T), relative humidity (RH), wind speed (Wind), precipitation (Prec), ozone (O₃), nitrogen dioxide (NO₂), sulfur dioxide (SO₂), and carbon monoxide (CO) all were nonlinear no matter in the whole year or in winter. It was also found that CO, NO₂, SO₂, T, and Wind all had greater impact on PM_2.5 concentration. Furthermore, effects of T and O₃ on PM_2.5 concentration in winter were weaker than that inwhole year. In the multi-factor GAMs, the combined effects of T, RH, SO₂, NO₂, O₃ and CO had significant impacts on the change of PM_2.5 concentration.For whole year, the adjusted judgment coefficient (R²) of the multi-factor gams model was 0.759 and the variance explanation rate was 76.42%. For winter, the adjusted R² of gams model was 0.708 and the variance explanation rate was 72.2%. CO was the most important influencing factor no matter in whole year or in winter. In the interaction GAMs, for the whole year,it was found that the synergetic effect of moderate low T (around 7℃) + high RH + high concentration of CO + high concentration of NO₂+ high concentration of SO₂ were beneficial to the formation of PM_2.5 in Chengdu, which means this condition had a synergistic amplification effect on the formation of PM_2.5 concentration. For winter, the coexistence of low wind + high RH + high CO + high NO₂ + high SO₂ were beneficial to the formation of PM_2.5, which condition had a synergistic amplification effect on the formation of PM_2.5 concentration. It was found that GAMs model could not only be used to identify the dominant influencing factors of PM_2.5 pollution, but also quantitatively analyze the influence of single effect and interaction of influencing factors on the change of PM_2.5 concentration, which was great significance for local to prevent and control PM_2.5 pollution.