STATISTICAL MODELS FOR FORECASTING AVERAGE MONTHLY TEMPERATURE AND MONTHLY PRECIPITATION AMOUNT IN PERM

Abstract

The article proposes an approach to forecasting mean temperature and total precipitation for the upcoming month, based on the study of the regularities of the influence of statistical characteristics of temperature and precipitation of previous periods on them. Among the predictors, along with the basic statistical characteristics, we use the fractality index which is an indicator of the randomness/ determinism of the climate series. Within the framework of this approach, we have developed models of different levels to predict the temperature and total precipitation amount in the upcoming month. The main parameters of these models are described and the possibilities of their variation are indicated. Examples are given to illustrate the forecasting methodology using various types of models and include the results of quality control of the models, calculation of forecast accuracy and dependence of forecast accuracy of average temperature and precipitation on the month (climate season). When tested in 2020, models for forecasting temperature and precipitation for the upcoming month give good results: 9 correct forecasts of temperature anomalies out of 10 (90%) and 7 correct forecasts of precipitation anomalies out of 9 (77,7%).