Drought Definitions and Forecasts for Water Resources Management

Interest in the potential impacts of climate variability on existing and planned water resource projects is commonplace. Climate impact investigations have focused primarily on evaluating extreme events with large potential impacts , like floods and droughts. Droughts, in particular, are difficult to analyze because of their slow onset, spatial variability, heterogeneous impacts, and uncertain management actions. R ecent studies have attempted to define drought characteristics as a function of precipitation, streamflow, soil moisture, and storage level (for developed systems). This paper extends these other studies by focusing on drought in the Geum River basin of Korea with three primary goals: 1) generating an operational definition of drought, 2 ) creating mid-term forecasts of climate variability, and 3) developing a decision support system for drought management that incorporates these other two features. The drought definition is derived from a series of computer simulations that use past precipitation, past and current streamflow, and reservoir storage to anticipate the onset and continuation of a drought event. Next, ensemble climate forecasts (developed from NCEP forecasts) are used to generate predictions of streamflows for the basin. These forecasts are mid -range in length (up to six month) and are derived from global spectral models (GSMs). These monthly meteorological forecasts are then interpolated to the finer hydrologic model scale . Finally, a decision support system is created to integrate the current degree of drought intensity and streamflow forecasts into a management plan. The decision support system indicate s the likelihood of inadequate supplies and proposes management actions.