Merging and error analysis of regional hydrometeorologic anomaly forecasts conditioned on climate precursors

Abstract

Forecasts of hydroclimatic variables and incorporation of their error bounds are invaluable in water resources planning and operations under uncertainty. In this study, regional long-term operational hydrologic forecast models conditioned on climatic precursor are presented. The forecasts also include uncertainty intervals and confidence limits. The forecasts are based on the temporal and spatial variability of hydrometeorologic anomalies and their relationships with climatic interannual and intraseasonal El Nino-Southern Oscillation (ENSO). The forecast skills of the proposed model, which incorporates ENSO forecasts on tropical rainfall and streamflow, are compared with those that are unconditional and do not incorporate ENSO. Significantly improved skills are achieved by incorporating forecasted ENSO indices and their errors. The seasonal variability of the forecast model skills are also evaluated. These ENSO-based forecasts of regional and seasonal-to-interannual hydrometeorologic variables consistently merged with systematic error analysis can provide outputs for direct use in water resources planning and operation under uncertainty.