SECTION 1: PROJECT OBJECTIVES AND ACCOMPLISHMENTS
The primary goal of this project was to investigate the mechanisms responsible for the formation of snow bands associated with wintertime east coast cyclonic storms. The project then focused on determining how operational numerical weather prediction models could provide guidance that will improve forecasts of these events.
A preliminary object of our diagnostic work was the 30 December 1997 coastal storm that was responsible for heavy banded snowfall over Pennsylvania. We developed an analysis model, based on the semi-geostrophic approximation, that was able to explain the timing and location of the snow bands created by the release of conditional symmetric instability (CSI). Prominent features of the conceptual model (based on analysis of a 36-km resolution simulation using the PSU-NCAR MM5 numerical model) are:
The next focus was the Washington, DC snowstorm of 9 March 1999. Our motivation was to determine why there was a large discrepancy between operational Eta model forecasts for the event. In particular, a 24-hour forecast anticipated heavier snowfall, while a 12-hour forecast predicted very little snowfall. We wished to determine whether the problem arose because of poor model initializations or because a particular run did not capture a dynamical process. We examined a sequence of 15-km resolution NCAR-PSU MM5 and Eta model simulations initiated at 1800 UTC 8 March, 0000 UTC 9 March and 0600 UTC 9 March. The DC snowfall occurred between 1800 UTC and 2400 UTC 9 March.
We concluded that all model runs failed to capture the timing and structure of a dry slot that played a crucial role in initiating the heavy DC snowfall. In contrast to the December 1997 storm, CSI did not play a major role in this event. We discovered that the lifting of dry mid-tropospheric air ahead of the dry slot was able to destabilize the air by eliminating a capping inversion thus promoting convection and eventually heavy snow. The problem with the model runs was poor initialization and, in addition, inadequate horizontal resolution to capture the destabilization process. In addition, all runs produced a dry slot that moved much too slowly compared to observations. We suggest a method to estimate vertical motion and convective destabilization in dry slots using satellite data. This technique could aid operational forecasters anticipate the destabilization process.
1.2 Lessons Learned
In the proposal for this Outreach Project, our goal was to provide the forecaster with diagnostic tools based that he or she can use to assess the veracity and/or fine tune a particular model-produced forecast of banded precipitation. We had originally proposed investigating summertime, in addition to wintertime, events. However the complexity of the few cases that we investigated was not anticipated and we focused on wintertime cases.
We detailed two mechanisms that are associated with wintertime banded precipitation: the release of CSI and destabilization associated with dry slots. The tool that forecasters use to anticipate a CSI event that leads to banded snowfall is a region of negative moist potential vorticity (MPV) near mid-tropospheric frontogenesis. We demonstrated that the MPV definition used to diagnose Eta model output is incomplete and could miss regions of potential CSI. We propose providing forecasters with a more accurately defined MPV that will facilitate a more accurate assessment of the potential for CSI.
SECTION 2: SUMMARY OF UNIVERSITY / NWS EXCHANGES
The Academic partners have been regular participants in the annual fall Winter Weather Workshops held at the State College NWS office. In 1999, Clark presented a summary of progress made under the COMET project. Much of the computational work for this project (especially the high resolution simulations) was carried out at the local NWS office. A permanent benefit to the university and the NWS office should emanate from the cooperation initiated under this project. To facilitate data storage and analysis, a 34 GB hard drive was purchased with funds from this project and is now a permanent part of the NWS computer system.
Richard James (the graduate student supported by the project) has benefited from the opportunity to run MM5 using the resources available at the State College NWS office. Most importantly, the expertise of Richard Grumm was indispensable to Richard James’ work.
SECTION 3: PRESENTATIONS AND PUBLICATIONS
Clark, J. H. E., R. P. James and R. Grumm, 2002: A Reexamination of the Mechanisms Responsible for Banded Precipitation. Accepted by Monthly Weather Review.
Clark, J. H. E. and R. P. James, 2002: The Diagnosis of Vertical Motion within Dry Intrusions. Submitted to Weather and Forecasting.
James, Richard P., The Role of a Dry Intrusion in the 9 March 1999 Snowstorm, MS Thesis, August 2001
James. R. P., R. Grumm and J. H. E. Clark, July 2001:“The Role of a Mesoscale Dry Intrusion in the Washington, D.C. Snowstorm of March 9, 1999“ at the Ninth Conference on Mesoscale Processes in Ft. Lauderdale.
SECTION 4: SUMMARY OF BENEFITS AND PROBLEMS ENCOUNTERED
4.1 Academic Partner
Without the help of Richard Grumm and other NWS personnel this project would not have been possible. First of all, Grumm’s expertise is setting up the model runs was indispensable. Secondly, all of the runs were made on the NWS computer system.
An important benefit to the University from this project, as well as other COMET projects, is that undergraduate courses in weather forecasting are being taught jointly by NWS and University faculty. Thus the knowledge gained from these projects is being transferred to our students in a timely manner. The student is thus exposed to cutting edge forecasting techniques.
4.2 NWS Partner
The NWS in State College has benefited in several ways from this research. First, and foremost, we are able to continue a close interaction with the Penn State University. In recent meetings with Dr. Clark and Richard James, the impact of model resolution and data were discussed. This will contribute to our better understanding of banded precipitation events in general. An added benefit is the knowledge gained on how to configure and run local models and the impacts of the data sources on these local model runs.
In order to run the MM5 effectively, the NWS partner had to work with other PSU partners, some of whom are not directly involved with the project. This collaboration allowed the NWS partner to learn to configure the MM5 for case studies. Due to this effort, the NWS partner has gained valuable knowledge on implementing and testing the NCAR-PSU MM5 model.