This area currently has one active project:

Fine resolution integrated model of the Rio Grande Basin
Winter, Springer, Costigan (LANL)

This project provides a scientific computational foundation for water resource decisions by coupling detailed physical models of the atmosphere, land surface hydrology (including plant communities) and groundwater hydrology of the Rio Grande Basin. Policy makers and water resource managers must increasingly decide among complex alternative allocation and activities. Their decisions routinely affect multiple competing users of water (including human beings, endangered species, etc.) whose interactions are difficult to anticipate without detailed physical models of highly resolved, basin-scale water budgets. Furthermore, their decisions sometimes have unintended consequences because of non-linear interactions among system components.

This project supports science-based decision making in three ways:

1. Increased physical understanding through experimentation. Physics-based models allow classical "hypothesize and test" science to be conducted on realistic virtual watersheds.
2. Scenario evaluation. Coupled physical models of the atmosphere, land surface hydrology and groundwater hydrology allow decision makers to evaluate the effects of alternatives on the entire Rio Grande system and their interactions.
3. Prediction. High performance computing supports Monte Carlo simulation of detailed basin-sized realizations. Development of deterministic equations for the moments of hydrologic variables (pressure head, runoff, etc.) yield efficient estimators and confidence intervals for system state variables.

Development of the Virtual Watershed Laboratory, a general computing environment for developing hydrologic models on parallel processors, has been postponed. The effort was too ambitious given the other tasks we are currently trying to accomplish.

During the period of October 2001 through July 2002) the following was accomplished:

• A physics-based approach to coupling hydrologic regimes was developed for the stream aquifer interface. The scaling approach has application to other interfaces such as land surface and atmosphere.
• We completed our work on stochastic models of heterogeneous groundwater systems and published them in Water Resources Research.
• Parallel Application Workspace (PAWS) software was applied to the coupling problem. PAWS allows individual components to transmit data when required rather than being collected by a single processor and distributed through files.

The findings, results, and implications are:

• We produced soil moisture maps at 100 m for the Rio Grande Basin above Cochiti Reservoir for the early 1992/93 water year. Once these have been validated, they will provide an unprecedented level of detail about the distribution of a critical variable.
• We developed a method for closing moment differential equations that is robust against the large variances found in heterogeneous groundwater systems. This promises to significantly extend the range of applications of stochastic models because most groundwater systems are highly heterogeneous.
• We developed boundary conditions for stream-aquifer interactions that conserve mass and momentum on average and are faithful to the small-scale physics found near the interface. We characterized conditions under which boundary conditions reduce to a simple difference between pressures in the stream and aquifer, which is the ad hoc model assumed in most of hydrology. We give conditions in terms of Reynolds and Froude numbers under which the ad hoc model is acceptable.

This work has been supported by a $500K/year match from Los Alamos National Laboratory as part of its commitment to SAHRA.

Plans for the Next Reporting Period

During the next reporting period, we intend to accomplish the following:

1. Complete land surface-atmosphere coupling
   a) Examine alternatives to current statistical downscaling of atmosphere-land surface link
   
b) Develop up-scaling for land surface-atmosphere link
2. Couple land surface-groundwater models
3. Couple stream-groundwater-land surface models
   
4. Computer Science
   a) Integrate 1-3 under PAWS
   b) Visualization
   
5. Develop requirements for additional modules
   a) Farm
   b) Riparian areas
   c) Socio-economic
   d) Water resources infrastructure (reservoirs and irrigation systems)
   

The project will provide basic commodity inputs to UNM economic models. We will also collaborate with Eric Small to develop method for up-scaling experimental results to 100 m scale. We will validate the models using the remote sensing model (SEBAL) implemented by Jan Hendrickx. We will also provide water balance inputs to water quality models (salinity).