| In recent decades,China has experienced a rapid urbanization and socio-economic development,but the accompanied atmospheric pollution has been a serious environmental burden.PM2.5 is one of principal pollutants of atmospheric environment,which has obvious negative effects on public health.PM2.5.5 contentrations is high and increase rapidly in China,combined with its numerous and dense population,the public health resks related to PM2.5 pollution is particularly serious.At present,ground monitoring is the main souce of PM2.5 data,and the data basis of researches on the health effects of PM2.5 pollution.The monitoring sites were still sparse with obvious regional and urban-rural imbalance,which has been a data bottleneck for accurate risk assessment of PM2.5 pollution exposure and severely limited researches on the health effects of PM2.5 pollution.While,estimation or simulation for PM2.5 concentrations with a high temporal-spatial resolution was considered as a potential method for data collection and supplement data missing to perform accurate health risk assessment.The current study selected the Yangtze River Delta as the study region.On the basis of exploring the spatiotemporal variation characteristics of PM2.5 concentrations,quantifing the effects of land use and meteorological factors on the spatial difference and spatiotemporal variation of PM2.5 concentrations,the study estimated daily PM2.5concentrations with a spatial resolution of 3 km using a hierarchical spatial-temporal model with land use as spatial covariates and meteorological factors as spatial-temporal covariates.The study aimed at obtaining high-precision and spatiotemporal full-covered PM2.5 concentrations data to compensate for the lack of spatiotemporal data of PM2.5 concentrations and provide accurate data basis and support for studies on the health effects of PM2.5 pollution.The study summarized the main conclusions as follows:?1?The temporal variations of PM2.5 concentrations mainly showed short time-scale fluctuations and a seasonal trend of decrease followed by increaseIn 2016,the minimum,average,and maximum of PM2.5 concentrations was respectively 8.17?g/m3,46.44?g/m3,and 150.23?g/m3 in the study region.There were a large number of high PM2.5 concentrations during the period of from January to March and from November to December,simultaneously,the difference in concentration values among sites was also large.PM2.5 concentrations presented a seasonal trend of decrease and then increase,the fluctuation range of PM2.5 concentrations and the difference in in concentration values among sites presented a variation trend of decreasing first and then increasing.PM2.5 concentrations presented potential periodic fluctuations of different time scales,while the drastical fluctuation of short time scale and the seasonal change trend of decreasing first and then increasing were the main characteristics.?2?The spatial distribution of PM2.5 concentrations generally presented a pattern of lower in the south and higher in the northThe spatial distribution of monitoring PM2.5 concentrations showed that PM2.5concentrations was high at central and north sites,but was low at south sites,presented a pattern of high in northwest inland and low in southeast coastal.On the city scale,it was higher in the cities along the line from Suzhou to Nanjing and Taizhou,Yangzhou in the north-central region,and lower in Ningbo,Taizhou,and Zhoushan in the southeast.Spatially full coverage PM2.5 concentrations showed that the areas in the north of the line from Jiaxing to Nanjing was a concentrated and contiguous region of high PM2.5 concentrations,it was generally low in the south of the line and some high concentrations were just dispersedly distributed in main urban areas.?3?Land use and meteorological factors had important impacts on the spatial difference and spatiotemporal variation of PM2.5 concentrations respectivelyIn the long term,PM2.5.5 concentrations presented negative correlation with precipitation,temperature,and wind speed?p<0.001?,but presented negative correlation with air pressure?p<0.001?.In different time periods,PM2.5.5 concentrations and meteorological factors had common periods of different scales and phase relations.The negative coherent relationship between PM2.5 concentrations and wind speed was most significant,PM2.5 concentrations was greatly affected by wind speed.Land use was the dominant factor influcing the spatial distribution of PM2.5concentrations,NOx also had important impacts.The interactions between each pairs of factors including land use,NOx,road density,population density,and night light index were all bivariate enhancement.The central and northern part of the Yangtze River Delta was the high-polluted area,because it was affected by the interaction of multiple factors.The influence of meteorological factors on PM2.5 concentration has obvious spatiotemporal non-stationarity.On one hand,the effects of meteorological factors on PM2.5 concentrations has obvious seasonal variations,such as the negative effects of precipitation on PM2.5 concentrations showed a weakening trend before August,showed a increase trend after August.On the other hand,there were also significant spatial differences in the effects of meteorological factors on PM2.5 concentrations.Such as,during the 1th-5th week,precipitation had a strong negative effects in cities in southern Jiangsu,and the effects of air pressure turned from positive to negative from north to south.During the 32th-36th week,the effects of precipitation on PM2.5concentrations was negative only in several cities in the north and Taizhou in the south,air pressure showed positive effects in Jiaxing,but showed negative effects in most other cities,and so on.?4?The spatial-temporal modeling theory and effective covariates determine the accuracy of PM2.5 concentrations estimatingBased on the monitoring data with missing values,the study realized the estimation of PM2.5 concentrations with a spatial and temporal resolution of 3 km×day in the Yangtze River Delta region.The cross-validated results showed that the total RMSE is 10.28?g/m3,the R2 is 0.88,and the 95%confidence level covered 94.27%of all the observations,indicating the estimated results had high accuracy.In rural areas outside the city,the estimated results also showed good accuracy?R2=0.82?.The spatial distribution of estimated PM2.5 concentrations showed that there was a contiguous high-polluted region in the north of the study region during a period of heavy haze pollution?19 in December 2016?,indicating that the estimated results capture the heavy haze pollution process well.The hierarchical spatial-temporal model simulated the spatial-temporal process of PM2.5 concentrations as time variation trends with spatial disparities,determined the spatial-temporal variabilities can be adequately captured.The land use and meteorological covariates was confirmed to effective estimating and predicting indicators of PM2.5 concentrations.The cross-validated R2 was 0.88,which reached a higher accuracy of PM2.5 concentrations estimation and simulation.The study finally obtained the PM2.5 concentrations data with high precision and full coverage in space and time,which can provide accurate data support for the researches on the health effects of PM2.5 pollution. |
PDF Full Text Request