• Select Specific Areas for Study

By selecting for study a small number of basins for which hydrologic models have already been developed, communication and coordination between researchers will be enhanced and efforts will tend to be focused on the problems associated with forecasting rather than on building new hydrologic models for new locations or on other associated tasks.

• Generate a Perpetual Set of Ongoing Large-Scale Simulations

To facilitate research efforts, a set of regularly-updated large-scale GCM simulations should be available to meso-scale climate modelers and hydrologic forecasters. For example, a large-scale three-month temperature and precipitation forecast based on forecast sea surface temperatures could be made available on a web site and updated every two weeks.

• Increase Spatial Resolution

Hydrologic results have been shown to be very sensitive to spatial resolution of climate models and/or downscaling techniques. One way to avoid model bias due to these effects is to increase the spatial resolution of the climate models. Validation of models should also be at an appropriate spatial scale for the basin.

• Coordinate Data Sets

Data that does not change (historic climatological and hydrologic data, for example), should be from a common source so that validation or other results are based on the same data and are directly comparable. Having a single agreed upon source of data also saves time by eliminating repetitive data processing efforts by different teams of researchers. The data is prepared once and is made available to all.

• Develop and Standardize Measures of Performance for Forecasting Skill to Improve Assessment and Evaluation of Research

By creating standardized measures of skill for forecast efforts, improvements in skill or comparisons of different techniques will be possible in a well-defined and quantitative manner.

• Improve Understanding of El Nino/Southern Oscillation and Applications to Specific Systems

Correlations between ENSO effects and hydrology have been identified in the historic record for some river basins. Strong correlations between ENSO effects and precipitation/temperature have been identified for the Southern latitudes and forecasting methods have been developed and are being used, however the use of ENSO forecasts to providing streamflow forecasts in the Northern hemishere has not yet been brought to a very high level of development. A better understanding of ENSO effects might allow similar forecasting skill in the Northern latitudes.

• Improve Validation of Each Element of Modeling Packages

If explicit streamflow forecasts are to be made using linked climate/hydrologic/reservoir-operations models, then each component of these linked packages must be carefully validated, and as much model bias as possible removed at each step.

• Standardize Validation Data

Validation data should have a common source to allow direct comparisons of different techniques.

• Improve Coordination Between Modelers and End Users

By clearly defining the framework above, forecasting tools will tend to be developed for appropriate areas, at appropriate time scales, and in a such a way that the hydrologic forecasts can be used easily and directly by the water resources management community.

• Improve Coordination Between Modelers

The considerations above were identified as crucial to improving the practice of using hydrologic models to forecast streamflow.

• Develop Protocols and Agreements

By emphasizing and better defining the connections between forecasters and end-users, appropriate development and validation of forecasting tools will tend to follow. Existing services are not generally perceived as being valuable to the water resources management community, and efforts should be made to provide managers with more information about what can be done now. Potentially, institutional paths exist for transferring new streamflow forecasting technology to end users in an effective manner, but organization and coordination of such efforts would be required.