Author(s): Jeongwoo Lee; Minwoo Park; Gihyun Park; Hyun-Han Kwon
Linked Author(s): Hyun-Han Kwon
Keywords: GloSea6 Drought Prediction Spatio-temporal Downscaling
Abstract: Accurate hydro-meteorological predictions are essential for effective water resource management and drought mitigation. This study refines GloSea6 climate data using a multivariate Non-stationary Hidden Markov Model (MNHMM) to enhance its spatial and temporal resolution for hydrological modeling. GloSea6 Hindcast data were downscaled with observations from the Automated Surface Observing System (ASOS) in South Korea, and the derived MNHMM parameters were applied to GloSea6 Forecast data to generate daily rainfall and temperature sequences across multiple weather stations. The MNHMM effectively captures temporal and regional climate variations through dynamic transition probabilities, overcoming the limitations of stationary models. This approach improves rainfall projections, streamflow simulations, and hydrological drought indices, enabling more accurate drought predictions on a daily scale over a three-month forecast horizon
Year: 2025