(Excerpted from the 25 July 1997 final report. Mr. Westrick continued the research begun under the fellowship.)
The original goals of this research project were to use the Penn State/ NCAR MM5 model to experiment with various approaches for improving the model initialization in an effort to improve model simulations. Originally the plan was to include the LAPS analysis into the model. Due to problems with the LAPS analysis in the Pacific Northwest, the thesis has changed slightly, although it will still have a focus on the importance of accurate model initialization. Mr. Westrick will be coupling a distributed hydrology model to the MM5 for use in riverflow research. The hydrology model that is being employed is the University of Washington Distributed Hydrology-Soil-Vegetation model (DHSVM). This model has been designed to be used in complex orography at the watershed basin scale. Although most of the previous applications of this model have concentrated on timescales larger than single flooding events, Mr. Westrick will be concentrating on the application at high temporal (one hour timestep) and spatial (4 - 12 km) scale. A recent flooding event on the Snoqualmie River Basin has been chosen as a case study.
At present Mr. Westrick is still in the middle of his case study. A considerable amount of time was spent writing software to allow the MM5 to interface with the hydrology model. Much effort was also required to become familiar with the hydrology model and a considerable amount of time was spent interfacing with the hydrologists in the Civil Engineering Department. The required software has been completed and Mr. Westrick is now in the model sensitivity/verification stage. It is too early to provide definitive results from the case study as he is still assessing the relative impact of each models performance on the simulated riverflow. Mr. Westrick is spending a considerable amount of time writing software so all sources can be included for a comprehensive multi-model verification. When finished with the research work, the software that has been written for the MM5/DHSVM interface will be available for future research. Mr. Westrick also intends to have the system set up to run in near real-time through this next winter season to assess the long term performance of the coupled model. Archival of this data for climatological research will be performed. He hopes to complete his thesis and submit a paper to an AMS journal by next Spring.
Mr. Westrick has also spent a considerable amount of time designing, coding and implementing a comprehensive verification program for the real-time MM5 run at the university. This program is now providing statistics on a daily basis. This summer another student and Mr. Westrick have been working on including the Eta model output into the verification program. They hope to have this implemented before next winter. The forecasters at the Seattle office have expressed an interest in how the two models compare.