Close Window

Jackson State Univ.: "The surprise record meso-snowfall event of 1997 in central Mississippi - How surprising and how common?"

Final Report


Overall project objectives included answering the questions: What did it take for the record snowfall event? What were the forecast clues? What methods may have assisted forecasters? What would have made "the difference" to forecasters to predict a snowfall event? What is the frequency of snow events in Mississippi?

The project met its objectives of data collection and analysis for the record meso-snowfall event in Jackson (14 December 1997) and data collection for the preliminary analysis of snow events (1970-1999) in the state of Mississippi. The record event was analyzed and indicated that significant jet streaks through Mexico, and another moving northward from the Gulf of Mexico, were poorly handled by operational models (ETA and AVN) resulting in large errors in the predicted strength of the upper level system responsible for the snowfall.

The snowfall was largely the result of interactions between these jets, the deformation zone associated with an upper level feature previously in existence (and as clearly indicated by, and diagnosed with, water vapor imagery), a direct frontogenetical circulation in mid and upper levels (i.e., 500 mb, 300 mb, and 250 mb), and isentropic lift of a moisture pool from the east. Model data were used for MM5 simulations of the event based on the 1200 and 0000 UTC analyzed fields prior to the event. Diagnostics of the event confirmed the synoptically analyzed aspects and provided insight to the mesoscale processes involved.

In the case of historical snowfalls, results for the state indicate events as early as November and as late as April, apparent cycles of snow/no-snow years, and the obvious association of snowfall in the state with arctic outbreaks. All single day snowfalls in November were under one inch whereas only January-February-March had single day totals of over three inches. Only January had events (single day snowfall) in excess of eight inches. Most January and February events occurred in the early 1980s through the early 1990s. These data were organized into synoptic events so that classification of surface and upper air patterns may be made in the future. It is hoped thereafter that they will be analyzed using historical data to diagnose their dynamics.

Lessons learned from NWS side include:

Project participants and their roles included: Paul Croft and Alan Gerard, lead investigators on the diagnoses; Croft and Shundra Stewart (student) data collection and analysis; Croft, Gerard, Michelle Webb (student), and Shundra Stewart synoptic classification of snowfall events; Croft, Dr. Jan Hafner (new research staff at JSU`), and Gerard MM5 modeling and analyses.

Project accomplishments include preprints, presentations (by students as well as lead investigators), use of findings by the Jackson NWS office, presentation of results during joint workshop.

Major changes to the scope: given level of student abilities and the lack of certain data the project's progress was not as rapid as had been anticipated and thus were unable to complete diagnostics study of 1990s snowfall events.


Upgrade of NWS workstation on-site at JSU for use in map discussions and classes, hiring of undergraduate major for AMBER project work, exploring possibility of a cooperative proposal, an unsolicited NOAA proposal submitted for more formalized collaborative work. A co-sponsored workshop on medium range QPF was held at the JSU library for the NWS in Jackson and other offices. University PI attended a COMET workshop on numerical weather prediction during the summer. A joint workshop was held in April on modeling and visualization efforts of the JSU Meteorology Program for NWS and others in the professional community. Preprint publication, manuscript to be submitted, and considering an electronic paper submission of model runs for event.

From NWS: Interaction regarding the mesoscale modeling efforts, the April workshop, current ingesting of MM5 output into NWS operations (in GRIB format and via web site products), Shundra Stewart and Michelle Webb's work and interactions at the NWS office and time with staff.


Two abstracts were submitted "The Record Meso-Snowfall Event of 1997 in Jackson, Mississippi", by Paul J. Croft and Alan Gerard (to AMS and NWA); and "Preliminary Evaluation of Snowfall Events in Mississippi -How Rare?” by Paul J. Croft and Shundra Stewart (to NWA). Both were accepted for presentation at the NWA and AMS meetings.

Croft, P. J., and A. Gerard, 2000. The Surprise Record Meso-Snowfall Event of 1997 in Central Mississippi. In final preparation.

Croft, P. J., and A. E. Gerard, 2000. The record meso-snowfall event of 1997 in Jackson, Mississippi. Preprint for AMS Annual Meeting, Long Beach, California, 9-15 January 2000.

Webb, M., P. J. Croft, and A. E. Gerard, 2000. AMBER: Student experiences and products for operational use. Preprint for AMS Annual Meeting, Long Beach, California, 9-15 January 2000.

Croft, P. J., and S. Stewart, 2000. Snowfall Events in Mississippi. In preparation.


Benefits: The project work has given new and useful insight to snowfall events in Mississippi, which are applicable to other southern states, especially in cases of limited extent snowfalls. Undergraduate meteorology majors have shared in the findings and related these to their coursework. It also provides an excellent opportunity for the use of diagnostic software in classes. Our research indicates parameters and products useful in the forecasting of heavy snow. Particularly, this project has made clear the importance of examining frontogenesis fields, particularly at upper levels, when diagnosing the potential for heavy snow with upper lows. This is particularly true in cases where the synoptic pattern appears to be fairly innocuous, or where the models do not portray the possibility of a significant event. Additionally, our research is leading toward techniques using rawinsonde operations and water vapor to monitor the development of heavy snow in the short term (0 to 6 hours) and the potential use of locally generated MM5 output. In addition, more collaboration was initiated between the NWS office and other JSU Meteorology Program faculty in the building of a research community.

Problems: The project funding and budget set-up were not completed until March and, as such, the project did not get fully underway until the middle of March (rather than January as planned). Some of this time was made up over the summer, as the student worker was able to work at that time and by way of a no-cost extension. Unfortunately several other items also caused delays including: network outages, PT travel (and lack of release-time to devote to project), unanticipated travel of PI (as requested by the university for various other research projects), several departmental/building break-ins and stolen computers, PT in promotion year.

Benefits to the NWS office resulting thus far from the collaboration:

Only problem is getting MM5 data into format which could be easily used at our end. Joint work between JSU and NWS overcame this problem, and we are now ingesting MM5 data in Grib format.