Author(s): Chenlu Yu; Dong Wang
Linked Author(s):
Keywords: Agricultural drought; Information theory; Prediction; Soil moisture; Vine copula
Abstract: Accurate and reliable prediction of agricultural drought (AD) is essential for food security and water resources management. This study develops a probabilistic AD prediction framework integrating information theory with a vine-copula model. Partial Information Correlation (PIC) is employed for feature selection and pre-processing of hydro-meteorological variables. Monthly soil moisture (SM) is predicted using a C-vine copula quantile regression model, and probabilistic forecasts are generated via Monte Carlo simulations. Based on predicted SM, the Standardized Soil Moisture Index (SSMI) at multiple time scales (1, 3, 6, and 12 months) is calculated for indirect AD prediction, while SSMI is also directly predicted using PIC-selected predictors. A case study at four stations in Yunnan Province, China, shows that PIC effectively identifies key hydro-meteorological drivers, SM predictions remain robust despite inter-monthly variability, and AD predictability generally increases with time scale, with direct prediction outperforming indirect prediction. The framework demonstrates strong reliability and practical utility for agricultural irrigation planning and drought mitigation.
Year: 2026