Predictive meteorological model based on fixed-length graphical figures

Abstract

The solution of the problem of increasing the success of forecasting meteorological quantities with a limited amount of initial information is presented. As a tool for predicting weather conditions, it is proposed to use predictive meteorological models based on graphical analysis of a time series of meteorological variables. The construction of these models is carried out using the methods of technical analysis, probability theory, the theory of large numbers, logical calculus and combinatorics. As the main tool, graphical figures are used, which are determined based on the analysis of chronological series of atmospheric parameter values and reflect the patterns of their variability. In the process of practical use of the proposed model, conditional probabilities of possible ways of development of atmospheric processes are estimated, and the most probable outcome in the current situation is selected. The obtained meteorological model has been successfully tested in predicting the height of the lower cloud cover and visibility range in the Arctic zone of the Russian Federation, a specific feature of which is a sparse network of meteorological stations. The success characteristics of the forecasts developed on its basis significantly exceed similar indicators of existing methods.