PM_2.5 Estimation In South America Andean Region Based On Multi-Source Data

Air pollution is one of the main causes of mortality in the world,PM2.5 Is an important pollutant especially in big cities,recently PM2.5 emissions have increased its negative effects on human health.According to previous studies,in the majority of regions of the world,annual median concentrations of air pollution are higher than the WHO guideline values.Remote sensing and GIS are powerful tools for the monitoring of air pollution in areas with lack of ground monitoring such as some countries in South America.The present study used multi-source data including MODIS aerosol data,atmospheric,vegetation,and population data from different Models and sensors.Integration data and processing by GWR function was applied to estimate values of PM 2.5 in South-America Andean region and Rio de Janeiro city,for the 2015 year.The main research contents and conclusions are as follows.1.In South-America Andean region there is a strong correlation between PM2.5and AOD,each month got correlation coefficient between 70-90%,Among them,the correlation coefficient of September,October,November,and December are more than 97%.Therefore,the GWR model can be used to estimate PM2.5.2.The present study used a GWR model,for this,we obtained explanatory variable(AOD,atmospheric temperature,wind speed,PBLH,vegetation,and population),then integration data and common least squares(OLS)estimations were applied before GWR function calculations,the PM2.5 was estimated for two scales,South American Andean region,and Rio de Janeiro City.GWR model validation showed good precision,average absolute deviation(MAD)of 3.5,Mean square error(MSE)of 20,root mean square error(RMSE)of 4.5,Average absolute percent error(MAPE)of 25 and squared correlation coefficient(R2): 0.83.The inversion of PM2.5results showed that:(1)The lowest concentration of PM 2.5 in the South American Andean region was 0.2-22?g/m3,mainly distributed in the north of the region(Colombia and Ecuador),the highest PM 2.5 was 40-78?g/m3,mainly distributed in central Chile during May,June,and July.In August,PM 2.5 in central Chile got 47?g/m3,while Bolivia,Peru,and other countries PM 2.5 changes very little in the year,remained below 40?g/m3(35-40?g/m3).In November most of the Andean region got PM 2.5 values under 22?g/m3(17-22?g/m3),In December PM2.5 was under 28?g/m3(17-28?g/m3),in this month only some locations in Colombia raised near to 35?g/m3.(2)In The local case in Rio de Janeiro city,the highest values of PM2.5 were during May,June,and July,with 18?g/m3,25?g/m3 and 18?g/m3 respectively,mainly distributed in the central,southern and eastern parts of the region.The highest value in August was of 26?g/m3,mainly distributed in the middle,north,and west of the city.In November the highest value was 48?g/m3.In December PM2.5 fell to 18?g/m3 and it was concentrated in the south.The lowest values were between 0.2-10?g/m3.These results suggest that this approach is useful for estimating of PM2.5distributions in a large-scale,especially for regions without PMs monitoring sites,but,in order to get better results in future studies it is necessary to improve the model including more ground monitoring data and higher spatial resolution remote sensing data.