Principles of long-term forecasting of natural processes: case study of the Asian territory of Russia

An important class of natural processes is formed by those whose development is due to causal relationships with some observed factors that are ahead of them in time, which makes it possible to use such relationships for long-term forecasting. The principles of such forecasting largely determine the choice of methodology used in the development of forecasts, as well as their justifiability. Therefore, their improvement is a promising direction for the development of technologies for the prevention of dangerous consequences of the processes under consideration, as well as an urgent safety problem in emergency situations. In the longterm forecasting of the studied processes, the principle “the more accurate the modeling, the more accurate the forecasting” is traditionally applied, and it is also assumed that their development occurs according to an inertial scenario. To solve the problem under consideration, the principle of “to each his own” can also be applied, suggesting that for each section of the studied territory, forecasting is carried out taking into account the significant factors of the process under consideration, corresponding specifically to this section. When identifying such factors, it is assumed that the scenario for the development of the process under study will not change in the future. This allows us to use as a predictor of the predictive model, the factor of the process under study, the connection with which has been strengthened in the past. A method is proposed for identifying areas for which the considered factor of the process under study has the specified property. It has been established that there are few areas of the Asian Territory of Russia where the scenario of the development of any hydrometeorological processes is inertial. As a result, a priori estimates of the justifiability of long-term forecasts of such processes here, as a rule, differ significantly from its a posteriori estimates. Following the principle of “to each his own” leads to the need to develop a specialized predictive model for each studied area of the territory under consideration, the predictors of which are factors that have previously been strengthened. The more sites under consideration, the higher the complexity of developing the required number of specialized models. Therefore, it is important to test hypotheses about the abundance of sites where a priori estimates of the feasibility of the forecast are underestimated. The validity of the hypotheses is confirmed by the example of such a process as interannual variations in average monthly surface air temperatures in the Asian territory of Russia for the spring months and their factors such as changes in the number of landscape fires that occurred in the previous year in Yakutia.