Analysis of Drought Dynamics Using SPI and SARIMA Models: A Case Study of the Rostov Region, Russia

Abstract

Based on precipitation data from six weather stations covering the period 1960– 2024, this study presents a retrospective analysis of drought dynamics in the Rostov Region, Russia, and evaluates the potential of the SARIMA model for forecasting moisture regime fluctuations. The Standardized Precipitation Index (SPI) was employed as the primary drought indicator. Two key phases of crop development were analyzed: the vegetation initiation period (March–May), assessed using the three-month SPI of May (SPI-3), and the full active growing season (April– September), assessed using the six-month SPI of September (SPI-6). The Mann-Kendall test revealed a non-significant positive trend in SPI-3 across all stations, while SPI-6 trends were non-significant and varied in direction. The highest frequency of drought events, based on both SPI-3 and SPI-6, occurred during 1960– 1969, with a general decline in subsequent decades. The lowest drought frequency was observed during 2010–2019. Notably, the frequency of extreme droughts has shown an increasing trend, posing significant risks to agricultural productivity. Although SARIMA modeling proved useful for short-term forecasting, its application was limited by unrealistic long-term projections and deviations from climatic norms. Consequently, drought forecasts were restricted to a two-year horizon. Nonetheless, the SARIMA approach remains a valuable supplementary tool for anticipating precipitation dynamics and drought events. © 2025 Federation of Eurasian Soil Science Societies.