Font Size: a A A

Aerosol Optical Thickness Estimating Using Spatiotemporal Mixed-effects Models And Elterman Models

Posted on:2024-09-06Degree:MasterType:Thesis
Country:ChinaCandidate:M S LiFull Text:PDF
GTID:2531307082981779Subject:Physical geography
Abstract/Summary:PDF Full Text Request
In the past 30 years,with the continuous advancement of China’s industrialization and urbanization process,people’s demand and consumption of energy resources have continued to increase,resulting in a huge load of air pollutant emissions,leading to increasingly serious air pollution problems in our country.Atmospheric aerosol is an important component of atmospheric pollutants,and aerosol optical depth(AOD)is an important indicator reflecting the aerosol load in the atmosphere.Therefore,accurately grasping the temporal and spatial variation characteristics of AOD can provide scientific theoretical basis and data support for regional air pollution prevention and control.At present,satellite remote sensing can provide AOD products with full coverage of time and space.However,due to factors such as cloud coverage and bright land surface,satellite products have serious shortages.How to effectively repair missing data has become a hot spot in the academic community.At present,the research on using satellite remote sensing and meteorological data to build a model to estimate AOD has been recognized by experts and scholars at home and abroad.However,most of the estimation models constructed in related studies are deterministic models,and it is difficult to accurately describe the spatiotemporal heterogeneity relationship between AOD and meteorological factors.Therefore,it is necessary to fully consider the spatiotemporal heterogeneity among different predictors,improve the AOD estimation model,realize higher-resolution AOD product estimation,and provide the possibility for accurate analysis of regional air pollution distribution characteristics and changing laws.In this paper,the Central-South Hebei Plain was taken as the research area,an AOD estimation model ST-EEM based on the spatio-temporal mixed effects model(STLME)and the Elterman model was constructed.First,MAIAC AOD,DEM,and meteorological element data were used as data sources,the Elterman model was used to estimate the aerosol elevation(ASH1);secondly,the meteorological factors such as water vapor pressure,relative humidity,temperature,atmospheric pressure,and wind speed and DEM were used to construct the spatio-temporal mixed effects model of ASH1;then,the ten-fold cross-validation method was used to verify the model,the root mean square error(RMSPE)and relative error(RPE)and other indicators were selected to evaluate the performance of the model;finally,the recalculated ASH1 parameter was brought into the formula of the Elterman model again to construct the ST-ERM model to obtain the final AOD estimation result.At the same time,the Nash coefficient(Ef)and the relative error(Er)were selected as indicators,and the estimation results of this model were compared with other models,and then the temporal and spatial distribution characteristics and changes of AOD in the central and southern Hebei Plain were analyzed from2016 to 2017,providing a theoretical basis for air pollution prevention and control.The main research conclusions are as follows:(1)Compared with the ERM,QRM and the ST-EEM model proposed in this paper has a greater improvement in AOD estimation accuracy,and has better applicability in data estimation in different years.The coefficient of determination R2 between the ST-AOD estimated by the ST-ERM model and the real MAIAC AOD is 0.70,which is increased by 0.53,0.47 and 0.309 than that of ERM,QRM and R-ERM models respectively;and RMSPE and RPE are both decreased.This shows that the ST-EEM model can not only better explain the role of the ASH1 parameter in AOD inversion,but also reduce prediction bias.At the same time,the data of 2017 is used to verify the model in this paper.Experimental results on two-year data show that the model has good time extension and can adapt to data inversion of different years.(2)The daily AOD value estimated by the ST-EEM model has a high temporal resolution.In 2016、2017,the model fitting R2 is 0.841,RMSPE is 0.382,and RPE is 56.789%.In 2017,the verification R2 is 0.823,RMSPE is 0.377,and RPE is 63.319%,indicating that the ST-EEM model has a high estimation accuracy.Among the 95 stations,there are 72 stations with Ef values between the predicted values of the ST-EEM model and the observed values in 2017greater than or equal to 0.50,and 92 stations with Er values less than 14%.The ST-EEM model is superior to the weather-based AOD prediction model proposed in previous studies.(3)Analyzing the spatiotemporal characteristics of the AOD inversion results of the ST-EEM model proposed in this paper,it can be known that the overall spatial characteristics of the AOD in the study area in 2016 and 2017 are high in the southwest,low in the northeast,high in the inland,and low on the coast.In addition to natural factors,socioeconomic factors are also responsible for the high AOD values in many urban areas in Southwest region.There are significant seasonal changes in the time characteristics,with the overall performance in summer being higher,followed by winter,and the overall performance in spring and autumn being relatively lower,which is related to the geographical location of the study area.Due to the special topography of the study area,pollutants tend to accumulate here,which leads to the increase of AOD concentration in this area.
Keywords/Search Tags:aerosol optical depth, spatiotemporal linear mixed-effects model, Elterman model, Analysis of spatiotemporal characteristics, Southern Central Hebei Plain
PDF Full Text Request
Related items