In the context of GCIP one of the eventual aims of the modeling effort is to generate inputs for operational hydrological and water resources management models over a range of time scales up to interannual. The specific GCIP objective for this area is to:

Improve the utility of hydrologic predictions for water resources management up to seasonal and interannual time scales.

The area of water resource applications is one of growing importance for GCIP because of both strong interest within NOAA and the priorities of the GEWEX Hydrometeorology Panel (GHP). GCIP is already carrying out research related to this topic. The University of Arizona has prepared summaries of relationships between GCIP and the water resources sector. Relevant studies have been carried out to determine the effects of the spatial scale of precipitation inputs to hydrologic models for streamflow forecasts. Studies have also been done to characterize the scaling properties of precipitation in order to develop a wavelet scheme for downscaling precipitation for input into hydrological models. Work on distributed hydrologic models will facilitate the coupling of hydrologic and atmospheric models for further studies involving the prediction of water budget components. The results of some of this research have already been applied in a water resource assessment project being carried out in the Columbia River Basin.

The goal of the GCIP Hydrologic and Water Resources Modeling (HWRM) research activities is to provide the physical understanding, and modeling expertise, to allow the GCIP's objective with respect to water resources stated above to be met. The geographic focus of the HWRM research activities will be within the GCIP region (the Mississippi River basin) and within the GCIP project period. With respect to the latter, the current implementation plan for GCIP extends through 2000, but follow-on work (possibly collaboratively with the Pan American Climate Study, PACS, and possibly beyond the current GCIP study area) would continue through at least 2005.

3.1 Approach

The approach will be to link with GCIP coupled modeling and data collection activities, to produce more accurate streamflow forecasts, and in turn, to develop methods of utilizing those forecasts for water management purposes. The lead times to be emphasized in GCIP HWRM activities will be longer than the currently accepted upper limit of weather forecasts, which is currently about one week, and up to interannual.

3.1.1 Hydrological Modeling

Present operational hydrologic forecast models in use by opearational agencies, such as the National Weather Service, and water management agencies in the GCIP region do not incorporate explicit representations of the effects of vegetation on surface hydrology that have been developed by the land surface community, nor do they model the surface energy budget. Further, operational forecast models generally make limited use of available soils, land use and remote sensing information of the kind that have been assembled by GCIP and NASA's Mission to Planet Earth. In the past, the land surface models utilized, for instance, in numerical weather prediction and climate models have had much more sophisticated representations of the surface energy balance than of runoff production. However, this situation is changing, and many land surface schemes now include hydrologic components that account for infiltration, surface runoff, and subsurface runoff and water storage. As GCIP progresses, subsurface storage, and other hydrologic processes such as snow accumulation and ablation, and soil freeze-thaw characteristics, will need to be represented better.

Nonetheless, macroscale hydrologic models appear to have potential for improving hydrologic prediction for several reasons. First, they can easily be made consistent with the spatial scale of numerical weather prediction and climate models, insofar as they are, for the most part, grid-based. Second, the more physically based representations offer the potential for greatly reducing the necessity of site-specific calibration both, because of their use of directly observed vegetation, and climate, data, and, because the large scale implementations are amenable to parameter regionalization methods. Finally, the models are consistent with evolving methods for initializing hydrologic variables (such as soil moisture and snow cover) needed by weather prediction and climate forecast models, hence the possibility exists for eventual integration of climate and long-range hydrologic forecasts. The latter is particularly important as ensemble precipitation forecast methods evolve. For instance, NOAA's Climate Prediction Center is moving to a system that will use ensemble forecasts from global and regional numerical prediction models to simulate possible future precipitation outcomes over periods up to several months. The methods to be used will include a range of statistical approaches to post-process model output information, for simulating fine scale space-time characteristics of precipitation not represented in model output, and to accounting for short-term forecast uncertainty that may not be included in NWP ensemble products.

Related Science Issues: Can physically based macroscale hydrologic models be used to produce streamflow forecasts that are more accurate, and/or require less initial and ongoing logistical support, than traditional hydrologic forecasting methods? What is the role of calibration in the implementation of physically based hydrologic models?

3.1.2 Water Resources Modeling

Previous GCIP Investigators and LSA planning meetings (such as the LSA-E meeting held in Huntsville in November, 1996) have concluded that improvements in short and long-range weather forecasting represent the strongest potential tie between the GCIP and water resources applications communities. Therefore, the current concept is that GCIP will play a central role in developing an experimental water resources forecast capability, as follows:

Science Questions: How can uncertain information from seasonal-to-interannual climate forecasts be used to improve water resources management?

3.1.3 Linkages to Coupled Modeling

There are important linkages between the HWRM and Coupled Modeling research activities. For instance, the stated function of the Coupled Modeling Research is to foster development and application of coupled models within GCIP to improve weather and climate prediction up to S/I time scales, and to improve the hydrological interpretation of meteorological predictions. The Coupled Modeling activities have identified three key science questions, the second of which (To what extent can meteorological predictions be given hydrological interpretation?) is central to the objectives of the Water Resources Activities. The hydrological modeling activities of the HWRM therefore, form the bridge between coupled modeling products, and GCIP is water resources objectives.

Science Questions: What is the predictability of surface hydrologic variables using global model ensemble forecasts? How can global model ensemble forecasts be downscaled to hydrologically relevant space-time scales in a computationally efficient manner?

3.2 Needs from other GCIP Research Areas:

The needs of the HWRM PRA from other PRAs are:

From DACOM: Access to ensemble climate forecasts of surface variables (downward solar and longwave radiation, precipitation, temperature, humidity, pressure, wind), downscaled to the regional level (nominally about 50 km resolution) over part or all of the GCIP region, for forecast lead times of 1-6 months. Initially, the HWRM research questions can be addressed using retrospective forecasts (real time is not necessary). Access to the desired forecast products will require protocols with the major modeling centers, e.g., the Climate Prediction Center of NCEP, ECMWF, and the International Research Institute.

From the Coupled Modeling: Downscaling techniques to produce surface fields of ensemble forecasts at regional scales (roughly 50 km, see above). Consideration should be given to computationally efficient methods that might produce downscaled fields with statistical characteristics similar to those that would be achieved via nested model applications, such as statistical disaggregation methods or resampling from archived nested model simulations.

From the Diagnostic Studies: Evaluation of ensemble forecast surface fields to offer insight into model strengths and weaknesses at climate forecast time scales.

3.3 Near-Term Priorities:

The near-term priorities of the HWRM research activities are:

Both of these priorities will require a shift in emphasis of GCIP to include a stronger focus on climate, as opposed to weather forecasts (e.g., forecast lead times of one to six months, and eventually longer, rather than hours to days). Although weather forecast model products, e.g., analysis fields, can be used for some aspects of testing of hydrologic and water resources models, stronger interactions with the climate forecast community will be essential for the HWRM research activities to achieve the research goal and GCIP objective.

3.4 Long-term Priorities

In the longer term (e.g., beyond 2000) it is expected that the HWRM research activities will focus on water resources in the western U.S., which are currently outside GCIP's geographic area. The hydrologic processes of concern in the West (such as, e.g., snow accumulation and ablation in mountainous regions) are, in some respects, more amenable to improved hydrologic forecasting than are the water resource systems of the Mississippi River basin. Also, linkages between seasonal-to-interannual climate variations and tropical ocean processes (which currently appear to offer the best hope for accurate seasonal-to-interannual climate forecasts) are generally stronger in the West than in the current GCIP region, so the West arguably offers a better water resources testbed for GCIP models than does its current region. In any event, a high priority for GCIP in this time frame is development of a demonstration application of seasonal forecast tools in at least one of the major water resource systems of the West.