| With the continuous progress of Surveying and Mapping Science and technology,the application field of GNSS technology has been expanding and gradually applied to many fields such as environmental monitoring,earthquake prediction,meteorology and so on.The application of GNSS in meteorology is of great significance.In recent years,fog and haze frequently occur in all parts of the country,which has a great impact on people’s production and survival.PM2.5 is the main component of haze.Spatial and temporal evolution of PM2.5 concentration is helpful to understand the development and status of air pollution.PM2.5 concentration monitoring in China started late,and is mainly obtained by physical methods.At present,the methods to obtain PM2.5 concentration are mainly based on ground monitoring and satellite remote sensing inversion.The former is expensive and complex,while the latter has low precision,which is not conducive to the analysis of temporal and spatial evolution characteristics of PM2.5 concentration.Therefore,scholars at home and abroad have proposed a method to establish PM2.5concentration model,which is to be selected.The PM2.5 concentration model was established with appropriate independent variables and models.Water vapor plays an important role in the formation of PM2.5.When water vapor concentration is saturated,it combines with fine dust in the air to form particles and participates in the formation of PM2.5.At the same time,as a reactant,water vapor reacts with air pollutants to form fine particles,which belong to PM2.5.At present,the main methods to obtain water vapor data are radio sounding,water vapor radiometer,satellite remote sensing inversion,etc.Among the above methods,radio sounding and water vapor radiometer are expensive and workload,and the precision of satellite remote sensing inversion is low,which causes many interference to obtain water vapor data.However,the retrieval of water vapor by GNSS is not only unaffected by weather conditions,but also has the advantages of accuracy,rapidity and high spatial and temporal resolution,and has been widely used.In view of the advantages of GNSS water vapor and its role in the formation of PM2.5,GNSS water vapor is used as an independent variable to construct PM2.5concentration model.Because the mechanism of PM2.5 concentration change is complex and affected by the external meteorological conditions and the observation of internal air pollution,the appropriate model is selected to establish PM2.5concentration model and analyze the temporal and spatial evolution characteristics ofPM2.5 concentration.The main contents and conclusions of this paper are as follows:1.Study the relationship between GNSS vapor and PM2.5 concentration.In this paper,by studying the correlation between GNSS water vapor and PM2.5concentration in time series,the change process of GNSS water vapor corresponds to the corresponding change of PM2.5 concentration.Due to the obvious seasonal variation of GNSS water vapor,the correlation of GNSS water vapor after wavelet decomposition and reconstruction is compared.After decomposition and reconstruction,the correlation results are higher,and GNSS water vapor can be used as an independent variable to participate in the construction of PM2.52.Establishment of PM2.5 concentration model by BP neural network.Taking the winter and spring seasons in Hebei Province as an example,suitable independent variables are selected for the internal and external factors affecting PM2.5concentration.PM10,SO2,CO,NO2 and O3 with high correlation are selected for air pollution observation.GNSS water vapor and wind speed are selected for external meteorological conditions.The fitting effect of the model is better when the pollution concentration level is good,light pollution,moderate pollution and heavy pollution.3.Establish a multivariate PM2.5 concentration model.The independent variables involved in the model construction were further selected.The PM10 with the highest correlation was selected as the internal factor,and the GNSS water vapor and wind speed as the external meteorological conditions.Compared with the PM2.5concentration model with PM10 as the independent variable,the multivariate concentration model had better effect,and the deterministic model had better fitting effect in pollution grade,mild pollution and moderate pollution. |
