Development of user-centric products requires the linking of identified user needs with research on forecast tools. This presentation will describe the process of integrating two lines of research that began in parallel: assessments of the needs of reservoir managers at the U.S. Bureau of Reclamation (USBR), and research to improve medium-range weather forecasting (also known as intraseasonal forecasts). Based on interviews, decision process analysis, and participant-observation, user studies of USBR Lower Colorado River management identified substantial needs for information on longer-term weather forecasts as well as seasonal climate. These needs relate to management objectives in the lower Colorado River basin including goals to maximize storage in lower Colorado River Reservoirs; balancing flood control and storage criteria; improving planning for releases for irrigation given the influence of temperature anomalies on demand; and to carry out other USBR goals such as environmental releases. The user assessment found that intra-seasonal forecast information may be useful throughout the water year, as reservoir planning and management adjusts to both observed and forecasted conditions of winter snow accumulation, spring runoff, warm-season irrigation, and municipal and industrial uses all year. At each stage in planning, anomalous intra-seasonal temperature or precipitation conditions may significantly influence storage and releases planned, often affecting planning for many months in the future.

Recently, forecast techniques have been developed that dramatically improve the skill of probabilistic weather forecasts during week 1 (Hamill et al. 2005) as well as improvements in week 2. These products are based on a 25-year reforecast database developed at NOAA/ESRL and used to demonstrate how a current numerical intra-seasonal forecast could be statistically calibrated using the reforecasts (Hamill et al. 2004). The resulting products include forecasts of temperature, precipitation, and other variables that are presented as maps of tercile probability and analog probability forecasts of precipitation downscaled to 32 km resolution. The ESRL technique now is synthesized into the CPC operational product and the tercile probability forecasts run operationally at the NOAA National Weather Service.

However, these forecasts are not yet in formats appropriate to many potential users. We are working with the USBR Lower Colorado Office (LCO) as pilot case for developing of these products for a broader community of water managers in the U.S. West through its operations. Because USBR/LCO reservoir management challenges and goals are similar in many ways to other large-scale reservoir management across the West, the results on usability of the products are likely to be transferable to other areas. The USBR also works closely with a large number of its own stakeholders in reservoir management, so developing products that support these interactions will provide insight into the needs of a broader community of water managers. This presentation will describe the efforts to work with the USBR to improve the useability of these products for their Lower Colorado operations, by reducing complexity and improving the accessibility of the forecasts.