Conditioning Ensemble Streamflow Prediction Forecasts Using Climate Signals in the Midwestern U.S.

摘要

Assimilation of seasonal climate forecasts is widely recognized as a potential means of improving the accuracy and reliability of National Weather Service long lead-time (1- month to 1-year) Ensemble Streamflow Prediction (ESP) forecasts. Such forecasts are not widely used, however, for a number of reasons. In the Midwestern U.S., few strong and consistent teleconnections have been identified for forecasting, perhaps due to nonlinear interactions of both Atlantic and Pacific Ocean patterns. Further, there is no consensus that the resulting forecast skill is sufficient to support water resources decision making. In this paper, we evaluate a set of climate patterns that may serve as useful indicators for seasonal (1- to 3-month) streamflow forecasts in the Midwestern U.S. Streamflow persistence and predictability based on simulated soil moisture are used as benchmarks for comparison with climate-based forecasts. Various procedures are discussed for assimilating climate forecasts into ESP forecasts, and forecast skill is assessed using re-sampling procedures. Finally, the value of seasonal forecasts for water resources management applications is discussed.