DEEP PROCESS-DATA MINING FOR BUILDING OF ANALYTICAL MODELS: 2. INFLUENCE OF WINTER-SPRING TEMPERATURES AND PRECIPITATION ON SPRING FLOOD EXTREMES FOR MOUNTAIN RIVERS

Abstract

Using the standard methodology for deep process-data mining called system- analytical modeling, we have built a high-performance process-driven (analytical) model for description of winter-spring temperatures and precipitation influence on spring flood extremes for mountain rivers. In April, flood discharge peaks (with ice motion) cause emergency inun- dations and pose a constant threat to local population. The effect of the landscape structure of river basins, winter-spring temperatures and precipitation on spring flood discharge peaks and troughs (SFDP and SFDT) for April 1951–2020 was analyzed by the example of 34 medium and small rivers of the Altai-Sayan mountain country. We identified nontrivial SFDP/SFDT dependences on meteorological conditions, proposed their physical-hydrological substantia- tions and determined SFDP/SFDT sensitivities to meteorological factor variations as con- tributions to the observed SFDP/SFDT variances. The contributions of winter and spring precipitation, winter air temperatures and spring ones for SFDP/SFDT variances made up 34.7, 21.9, 7.8, 6.1% and 13.6, 18.8, 6.6, 1.5%, respectively.