Process-data mining of regional climate changes in the Altai-Sayan mountain country and the West Siberian Plain
摘要
The proposed methodology for deep process-data mining of climate changes was applied to two large regions of Eurasia: the Altai-Sayan mountain country (2,000,000 km2) and the West Siberian Plain (3,000,000 km2). We used the observational data on the average monthly air temperatures and monthly precipitation from 23 reference weather stations for 1951–2023. These climatic characteristics were converted into dimensionless units by normalizing to their long-term average monthly “in situ” values and averaging over weather stations. As a result, they become almost the same for any site of the regions and thus turn into spatial ones. The reverse transition from dimensionless characteristics to those measured in °C and mm is easily performed through their multiplying by the long-term average monthly “in situ” values. The error of the created statistical-analytical models in describing the global-regional climatic changes does not exceed 5.4% and 3.7% of variances for the interannual variations in temperatures and precipitation at any sites of the mountains and the plain. The rate of increase in annual average temperatures and annual precipitation has exceeded 1 °C and 80 mm per 100 years. Basing on the simple criterion for assessing climate sensitivity to environmental factor variations, the impact of global-regional and local environmental factors on climate has been characterized. Both in the mountains and on the plain, global-regional factors form about 60% and 35% of interannual variances of the average monthly air temperatures and monthly precipitation, respectively. In turn, numerous local factors are responsible for approximately 40% and 65% of these variances.