| With the rapid development of the world economy and the increasing needs of people for a better life,it is the common responsibility of mankind to strengthen the protection of the ecological environment and improve the quality of the ecological environment.In the past two decades,governments of all countries have continuously strengthened the detection of atmospheric parameters and collected massive atmospheric parameter data.However,due to the different storage structures and parameter types of different data sources,scholars have many problems in atmospheric data information mining.Therefore,it is of great significance to achieve environmental governance,carbon peaking,carbon neutralization,etc.by organically integrating different atmospheric data sources,mining,analyzing and forecasting them.This paper focuses on data fusion and information mining of multi-source atmospheric parameter data.The research work is as follows:(1)On the basis of summarizing the research on the spatial and temporal distribution of atmospheric parameters at home and abroad,using the long-term accumulated radiosonde data and satellite observation data,based on the inverse distance difference and Kriging interpolation method,the paper constructs a nonlinear surface interpolation algorithm with higher longitude and latitude spatial resolution at different levels of multidimensional atmospheric parameters for data fusion,and constructs a unified data storage structure for the fused data,which is convenient for later data analysis and research.(2)After processing multi-source heterogeneous data and building a unified data storage system,this paper selects two atmospheric parameters,atmospheric CO concentration and cloud top pressure,to conduct a study on the characteristics of temporal and spatial distribution,and finds new phenomena of long-term changes in atmospheric parameters,so as to reflect and study the global and regional environmental quality,and provide valuable reference materials for green development and air pollution prevention in the world.In the analysis of the two atmospheric parameters,the main research ideas are spatial changes of longitude and latitude,annual changes,seasonal changes and monthly changes.First,the MK trend test and Sen slope estimation methods are used to mine the spatio-temporal information of CO concentration and cloud top pressure.Then,the Pearson correlation test and EOF mode decomposition algorithms are used to explore the important factors that affect their changes,and the impact of atmospheric parameter changes on the environment and the energy budget of the earth atmosphere system is excavated.Finally,based on massive long time series atmospheric data,this paper uses VMD to improve LSTM,which has achieved excellent results in time series prediction.Through a large number of experimental comparisons,it is found that the error rate of the improved time series prediction model is reduced by 1.7% compared with the original model,and the prediction error can be controlled at about 0.5%.(3)The data fusion model constructed,the data analysis means used and the improved time series prediction algorithm are applied to practice.A multi-source atmospheric data mining and prediction system is developed,which realizes the functions of data collection,data storage and management,data processing and visual analysis.It can help users quickly analyze the changes of atmospheric parameters and propose specific management measures in time.Figure [44] Table [9] Reference [79]... |
PDF Full Text Request