Basin‐scale monthly rainfall forecasts with a data‐driven model using lagged global climate indices and future predicted rainfall of an adjacent basin

Future long-term rainfall forecasts are valuable for operating water supply facilities and managing unusual droughts. This study proposes a novel approach to forecast basin-scale monthly from lagged global climate indices, antecedent historical data of targeted basin, forecasted nearby basin using data-driven model. The is applied the Han River Geum South Korea, May June, prone drought occurrence. An artificial neural network (ANN), widely used model, was employed build forecasting models basins. Two types ANN were constructed: one uses predictors indices that have been typically in previous studies, other further an adjacent first attempted this by considering strong concurrent relationship between optimal architectures determined through Monte Carlo cross-validation (MCCV) process which repeated subsampling training datasets carried out reduce output variance obtain ensemble forecasts. results show proposed model with input variables past target provides better predictive performance than without basin's rainfall. categorical skill based on good: hit rates Heidke scores ranged 50.9 66.0% 0.29 0.49, respectively. confirm as variable can enhance model's ability predict future