10th International Aerosol Conference September 2 - September 7, 2018 America's Center Convention Complex St. Louis, Missouri, USA
Abstract View
Correlations between PM2.5 Concentration and Local Meteorological Conditions with Focus on Statistical Models to Retrieve Long Term PM 2.5 Proxy Data: A Case Study in Chengdu, China
LEI LUO, Xinying Tang, Pengping Wu, Ling Wang, Institute of Plateau Meteorology, CMA, Chengdu
Abstract Number: 1596 Working Group: Aerosols in Earth System
Abstract The metropolitan city of Chengdu in south-west china has been suffering from severe air pollution for many years due to its basin topography as well as unfavorable atmospheric conditions. During heavy pollution episodes in the city, PM2.5 is mostly the primary pollutants, particularly in winters. With governments implemented stringent emission reduction measures, the local anthropogenic pollutant source usually remains constant or even declined sometimes; therefore meteorological conditions must have been playing a major role in those pollution episodes. To get a clearer understanding on how PM2.5 concentration correlated with meteorological parameters, 3 year’s hourly meteorological and PM2.5 data from 2012 to 2014 were obtained from local observation stations. Linear regression and multi-variance analysis were applied to find the correlation between major meteorological variables and PM2.5 concentrations. The preliminary results showed that air temperature, pressure, relative humidity (RH) and wind speed may have caused significant different PM2.5 concentration levels. However, these meteorological factors demonstrate different effects to PM2.5 in different seasons. During the winter, RH has a significant positive correlation while temperature and pressure show a slightly negative correlation with PM2.5 concentrations, but these effects are not obvious in summer. Analysis also reveals that visibility has a strong positive linear correlation with PM2.5, this leads to the idea to derive a statistical model using visibility values to calculate long term proxy PM2.5 data for early history years. This proxy PM2.5 datasets may be very useful in assessing long term aerosol climate effects. Intensive stepwise recursive regression were conducted on dataset of 2013, resulted in a simple yet quite effective statistical model (R=0.67) with visibility variables contributes the most variance. The equation was used on 2012 dataset to calculate proxy PM2.5 concentration for each sample hour in 2012. The retrieved data agreed ideally with the observed PM2.5 concentrations with no significant statistic differences.