Long-range Forecasting of Seasonal Air Temperature Using the Time Analog Method

摘要

A method is proposed to forecast monthly mean anomalies of air temperature for three months of the moving season, on the basis of statistical analysis of a selected set of processes with temperature changes analogous to that in the initial season. The set of the analogous processes is used to reveal the most frequent values of temperature anomalies in their grouping by three equiprobable categories. The set of analyses for each station gives a field of predicted anomalies at the CIS area for every month of the season. The anomaly sign forecast verification is presented for 1981-2004.