Haze Forecast Of Joint Ground-based GNSS And MODIS

In recent years,haze phenomenon has frequently occurred in major cities in China.It causes serious harm to people's health and social environment.The existing methods for monitoring forecasting fog and haze have certain limitations,such as low spatial and temporal resolution and high price.After many years of development and improvement,GNSS technology and Remote Sensing technology have the advantages of large global coverage,large data acquisition,high precision,and low cost.Therefore,monitoring and forecasting of haze weather by combining two technologies can make up for traditional monitoring methods.This paper uses Beijing as a research area,comprehensive use of GNSS technology and MODIS method for monitoring and forecasting of haze.The specific research content and achievements of this paper include:?1?Introduce the basic method of using the ZTD?Zenith Tropospheric Delay?to invert PWV?Precipitable Water Vapor?principle.Further analying the characteristics of AQI and ZTD,PM_2.5 and PWV in each season.It was found that the sequentially of AQI and ZTD,PM_2.5 and PWV were highly consistent during the process of haze generation and dissipation.Factors such as air pressure,temperature,relative humidity,and wind speed have different degrees of influence on haze weather.Especially during spring and winter in haze occurrence period,the correlation between air pressure,relative humidity and PM_2.5 is the most significant.?2?Introduce the principle of retrieving AOD?Aerosol Optical Depth?using MODIS images.According to the spatial distribution characteristics of AOD on March 14 and March16,2016,the AOD in Beijing gradually increased from the northwest to the southeast.Comparing the spatial distribution of AOD and PM_2.5 concentrations,it was found that PM_2.5concentrations were higher in regions with AOD in higher degree,whereas PM_2.5concentrations were lower in regions with AOD in lower degree,but individual regions were affected by water and other factors,so the effect does not show this rule.?3?The BP neural network was used to model the influence factors of six PM_2.5concentrations such as PWV,AOD and related meteorological elements.Spatial PM_2.5concentrations were used to obtain PM_2.5.5 concentration values,and the PM_2.5.5 predicted by the BP neural network was used.The spatial distribution of PM_2.5 concentration values and concentration values published by the monitoring station are compared,and the spatial distribution of the two PM_2.5 concentration values is generally consistent.