Using machine learning method on calculation of boundary layer height

Abstract

Traditional observation methods of the boundary layer are susceptible to external factors. In particular, detection data are relatively scarce in special areas such as plateaus and oceans. However, occultation observation data have the unique advantage of global coverage, and the boundary layer information of the occultation profile is rich. Based on the machine learning algorithm, this paper combines GPS occultation technology to improve the algorithm and constructs the boundary layer height simulation model based on the actual situation and analyzes its functional structure one by one. Moreover, in order to verify the validity of the simulation model, a simulation analysis is performed in combination with actual data. In addition, the performance of the model is verified from multiple aspects, and the results of the research are calculated by mathematical statistics, and the corresponding statistical diagram is drawn. Through experimental analysis, it can be known that the model constructed in this paper can effectively simulate the boundary height and has certain practical effects.