92nd American Meteorological Society Annual Meeting (January 22-26, 2012)

Tuesday, 24 January 2012: 1:30 PM
Utilizing Probabilistic Forecasts for Colorado River Reservoir Operations: Decision Making and Risk Management [INVITED]
Room 350/351 (New Orleans Convention Center )
Katrina Grantz, Bureau of Reclamation, Salt Lake City, UT; and H. Hermansen and S. Tighi

The Bureau of Reclamation manages reservoirs on the Colorado River from the headwaters down to the Mexican border. Probabilistic forecasts are utilized at various temporal scales, from daily operations and spring runoff routing to annual and multi-year planning. The Upper Basin reservoirs, from the headwaters to Lake Powell, are supply driven and operations depend heavily on forecasted inflows. Upper Basin risk management balances water supply, hydropower, flood control, endangered species, fish and wildlife, and recreation. Operations of Lower Basin reservoirs, from Lake Mead to the Mexican border, utilize Upper Basin probabilistic operational scenarios to manage risk by identifying a range of potential future Lake Mead elevations. This informs decision making for a diversity of stakeholders concerned with flood control, hydropower, recreation, environment, and water supply, particularly the probability of a shortage or surplus being declared by the Secretary of the Interior.

Reclamation uses several operations models to aid with decision making and risk management. The basin-wide Colorado River reservoir operations model, the “24-Month Study," provides a suite of projected future reservoir and river conditions. Reasonable minimum, reasonable maximum and most probable inflow forecasts are evaluated at key operational decision making times, but provide a limited range of projected conditions. A more stochastic-based operations model that utilizes multiple inflow scenarios is in the final stages of operational implementation. Individual reservoir routing models are also used to spatially and temporally fine tune projections of future conditions. Current challenges with probabilistic forecasts and their implementation in real-time reservoir management tools continue to provide ample research opportunities.

Supplementary URL: