There is currently one active project in this sub-area:

Regional scale simulation and prediction
J. Roads, J. Chen (UCSD)

In the past, limited computer resources and lack of spatially distributed meteorological data at sufficient spatial and temporal scales have made it impossible to simulate coupled hydrologic systems at scales fine enough to be confident of the accuracy of underlying physical theories. Moreover, data about hydrologic systems have usually been limited to only a few samples, even though Earth systems exhibit considerable heterogeneity in both space and time. We have therefore developed a global to regional climate modeling system, the Global Spectral Model/Regional Spectral Model/Variable Infiltration Capacity macro-scale model (GSM/RSM/VIC) to simulate and predict short-term (synoptic) and long-term (climatological) characteristics of the Southwest. This modeling system not only is capable of potentially developing long-term simulations, but is also used to study coupled land-atmosphere processes.

Dr. Roads and Dr. Cui (2001) initially developed an off-line hydrologic simulation system using either observations or model variables to force a macro-scale hydrologic model focused on the U.S. Southwest. We have also developed a web site showing near-real-time daily to monthly forecasts for the U.S. Southwest, with a focus on hydrology components (see http://ecpc.ucsd.edu/projects/uastc/). Chen and Roads updated this hydrologic simulation and prediction system and augmented the web site with a description of the hydrologic models being developed for this project as well as links to sites describing the atmospheric models. We then began to validate the VIC streamflow in comparison to observed streamflow data from the Rio Grande.

During the reporting period, Chen and Roads compared observed streamflow data from the Rio Grande with the VIC output. Because of the heavy regulation of the Rio Grande, streamflow observations did not adequately assess the performance of the macro-scale hydrologic model. Moreover, little data was available to evaluate other model processes. We are therefore now beginning to use LDAS products as a benchmark to evaluate the RSM/VIC for this region. We also intend to use developing SAHRA snow products to evaluate the VIC snow water equivalent, since this variable provides some indication of the model's verisimilitude.

We have now demonstrated real time simulation and prediction capability for the U.S. Southwest hydrology in general and the Rio Grande in particular. This real time system includes simulation and prediction of various land surface properties, such as soil moisture, snow, streamflow, evaporation, and other energy fluxes. This system has now run continuously since Sept. 27, 1997, and covers the entire period of the SAHRA project. All of this output data is archived and is being continuously re-analyzed as we better understand various model deficiencies for this region, and also how to improve these features.

We have developed a seasonal hydrologic simulation and forecast system using RSM/VIC over the Southwest and have begun to study and analyze the skill and limitations of the macro-scale hydrologic model simulations and forecasts. Our research will help to develop a regional-scale hydrologic simulation and prediction system that can be easily connected to current regional atmospheric models and eventually to higher resolution hydrologic and groundwater models.

We have not yet been able to fully validate this developing system, in part due to lack of adequate data, which is beginning to become available for this region. It should also be mentioned that streamflow observations in heavily managed rivers and streamflow from physical models cannot be directly compared. We are therefore looking for other sources of data, including US LDAS products to help evaluate our simulation and prediction system for this region.

This work is being done in conjunction with the Scripps Experimental Climate Prediction Center, which is interested in first developing a hydrologic prediction system for the U.S. Southwest, in collaboration with SAHRA researchers, and then transferring the developed model to the rest of the U.S. and eventually to other global regions. The partial postdoctoral support from SAHRA is augmented with ECPC funds. The PI (Roads) devotes his time at no charge to the SAHRA project.

Plans

We are studying model-simulated runoff over the Rio-Grande at basin and regional scales and will eventually begin to study the Colorado River Basin. After validating and improving the VIC macro-scale land surface model, we are then going to integrate the model with the RSM, and study the hydrology forecast skill at higher resolution and for longer temporal horizons. Demonstrating a reliable hydrologic simulation and prediction system on a variety of scales is essential for decision makers. We therefore believe our proposed research is a key part of SAHRA.

Our proposed research could integrate well with other modeling activities that study local, micro-, and medium-scale hydrologic processes over the Rio Grande. In particular, our research provides an interface to a regional scale atmospheric model on a variety of spatial scales and a variety of temporal scales ranging from hours to decades. Also, field observations from TA1 can be used to evaluate our model performance and thereby improve the parameterizations of hydrologic processes in the VIC model. As confidence is gained in our model system, it could be used to extend the characteristics of scattered field measurements to the entire SAHRA region.