Operational Seasonal Streamflow Forecasting Using Climate Information

Accuracy of seasonal forecasts is improving due to recognition of large-scale climatic patterns. We discuss the need to identify the scope of the forecast and choose an appropriate presentation method for the forecast in order to successfully implement seasonal forecasts in operation. Probabilistic forecasts best represent the type of information contained in climate-based forecasts, and degrading them to deterministic forecasts can limit their effectiveness. Differences in climatic patterns have an impact on the accuracy of forecasts for different geographic regions that is demonstrated by the more accurate preliminary results for a river in the Pacific Northwest than one in the Colorado headwaters. We use non-parametric statistical methods to choose predictors from standard climate indices and sea-surface temperatures and forecasting streamflow.