Final Report

University Principal Investigator: Prof. Steven Koch
NWS Principal Investigator: Kermit Keeter

1. Project Objectives and Accomplishments

1.1 Introduction

COMET was established to increase opportunities for: mesoscale education and training, the conduct of collaborative research projects between universities and the National Weather Service, and the improvement of local forecasts by fostering such interactions between forecasters and researchers. COMET-funded joint activities between North Carolina State University (NCSU) and the Raleigh (RAH) WSFO have over the past three years reached a mature stage resulting in tangible and significant impacts on the local forecast process. The most critical forecast problems in the Southeastern U. S. have always been the principal focus of our collaborative studies and they continue to require an understanding of the effects of the region's diverse topography on the evolution of meteorological features.

Years of joint study experiences have led the principal investigators to develop a well focused perspective of NWS - University collaborations. "Lessons learned" from our experiences provide a context for interpreting the progress and value of our collaborative results. We have also adopted a "holistic" approach to our collaborations, meaning that the whole of all our efforts is greater than the sum of the various collaborative components. Accordingly, there has been a carryover of results from one project to another. Our views and approach were presented at the CSTAR NWS/University Science Workshop held on 25-27 September 1996 at NWS Headquarters:

"The ultimate goal of NWS RAH - NCSU collaborations is to integrate significant collaborative findings into the NWS forecast/warning process such that all, not just a few, forecasters are applying the finding in a correct and efficient manner. All other collaborative activities such as the research process, the development of forecast tools or conceptual models, presentations and publications, while essential, only represent intermittent successes along the way toward the goal of extensive operational proficiency."

Transferring research into extensive application in NWS operations is a complex process that proceeds at various rates depending on such factors as the complexity of the meteorological concepts, the availability of technological resources, and individual experiences and learning differences among forecasters. The transfer process is characterized by progressive levels of understanding where ideally forecasters advance from an introduction to a research finding to mastery of its application in forecast operations. Methods that we have utilized to facilitate the transfer process from research to operationally proficient use have included, most importantly, the practice, publication, testing, and operational implementation of forecast applications research applicable to the most important local/regional forecast problems. Our approach has also included many other elements:

• The development of conceptual models featuring the use of forecast macros that focus and streamline the forecast and warning processes
• The development of objective and subjective forecast/warning tools providing guidance for specific forecast concerns
• A locally run mesoscale numerical weather prediction model for studying various aspects of critical regional forecast problems
• The involvement of university faculty and graduate students in NWS operations and professional development, including joint severe and winter weather operations that provide researchers with an operational perspective and fosters collaborative relationships
  NWS forecasters auditing or taking graduate-level meteorology courses at NCSU and forecasters presenting topics as guest lectures in undergraduate classes
• Lectures given by NCSU faculty at NWS RAH seminars, colloquia, and workshops
• Joint development and use of software systems and research tools for analysis of NWS data
• Joint NWS-University case study team meetings to choose and facilitate the collaborative study of significant weather events
  The NWS RAH - NCSU Lecture Series and a recommended "Paper of the Month" featuring presentations on meteorological concepts and collaborative findings
• "Beyond AFOS Toward AWIPS" tutorials providing a succinct overview of less familiar meteorological concepts written with an operational perspective ("Why should this research matter to me?") employing a "hands on" component that brings the concept to life through modern data visualization (PC GRIDS, GEMPAK, GARP)
• "Operational Proficiency" guides serving as a comprehensive and ready reference for the periodic review of recommended procedures for seasonal forecast problems
• Sharing of the results of our COMET-sponsored work with other NWS field offices and the private sector, an extensive outreach program accomplished primarily through regional workshops and personal contacts

In developing these collaborative methods, there has been a general division of responsibility between the University and the Forecast Office. The RAH office generally depends upon NCSU researchers to guide and direct the collaborations toward the inclusion of progressive meteorological fields and innovative uses for the modernized technologies. The University serves as a resource for ensuring that the content of the tutorials and operational proficiency guides produced for the forecasters are correct. The University also provides technological expertise to the Forecast Office regarding data systems, in developing GEMPAK scripts, and in making the output from the locally-run mesoscale numerical model available in real time to forecasters. The University looks to the Forecast Office to provide an operational perspective, defining problems for joint investigation and putting collaborative findings into forecaster-friendly formats. The Forecast Office has the major responsibility for managing the integration of research findings into the operational forecast/warning process, which includes "outreach" activities where forecasters from other NWS offices or the media are familiarized with these findings. This division of responsibilities between NCSU and NWS RAH is detailed below.

The participants from NCSU during the 3-year project have consisted of several professors and a cadre of graduate students. Clearly, the limited COMET funding did not fully support all of this effort, but rather, it fostered an environment of mutual interest in the conduct of operational forecasting research activities. Many of these efforts were jointly supported by COMET, the Air Force's AFIT program (to which we and COMET are indebted), the Southeast Consortium on Severe Thunderstorms and Tornadoes, and the National Science Foundation. The faculty and students involved in the research listed below received at least part of their support from COMET, or their work was of direct relevance to COMET objectives (some of this work will be reported in full detail in other reports, such as the COMET Graduate Student Fellowship work of Devin Kramer and the COMET Partners Project work of Michael Black):

Dr. Steven Koch (Associate Professor) - Principal NCSU investigator for COMET Cooperative Project and grant administrator. He has primary decision-making responsibility for selection of research topics and cases, and advises graduate students in their research. Provides guest lectures at RAH seminars and workshops and offers expert review of tutorials. Developer of the realtime gravity wave detection system, and the primary person involved in project task T5. Has chief responsibility for assuring publication of the results of collaborative activities. Professor Koch guided these graduate students in this project:

Devin Kramer (Graduate Student/Task T4) - His chief responsibility has been to run the MASS model for study of the physics, dynamics, and numerics necessary to provide accurate forecasts of Cold-Air Damming events; primary developer of the MASS model homepage, NTRANS and GARP fields; works closely with Ron Humble, Gail Hartfield, Ed Delgado, and Kermit Keeter. His scientific contributions will be reported in detail in the final report on the COMET Graduate Student Fellowship.

Tony Krogh (Graduate Student/ Task T2) - His M. S. thesis study used vertical wind shear information obtained from the WSR-88D Velocity Azimuth Display Wind Product (VWP) to retrieve thermal advection and temperature fields. This research has established a foundation for future efforts to obtain knowledge of the unknown thermodynamic structure of nonclassical frontal systems in the Carolinas (in particular, our next COMET project will utilize this technique to study Cold Fronts Aloft).

Tony Ray (Graduate Student/Task T2) - His M. S. research showed that the new Raleigh WSR-88D could be used in conjunction with GOES and surface data to conduct mesoanalyses in North Carolina. He also showed the ability to identify various convergence boundary types and determine their importance in the convective initiation process. His research has had profound impact on operations, including recent attempts to do nowcasting and create convective outlooks in the Raleigh office.

Leanne Siedlarz (Graduate Student/Task T3) - Primary contribution was the completion of the study of gravity wave climatology using the Koch and O'Handley (1997) objective analysis methodology, in which high time resolution information from surface stations can be exploited to perform realtime mesoanalyses.

Dr. Al Riordan (Associate Professor) - Project participant in joint severe operations; documentation of the severe weather warning process; studies of severe weather events; conduct of a convective nowcast experiment in summer 1996. While not directly funded by this COMET program, the following projects involved interactions between NCSU and the NWSFO at Raleigh and addressed the types of research and operational problems that COMET supports (additional detail is provided here concerning these projects, since they were not included in the list of objectives from this COMET proposal):

Michael Black (Graduate Student) - An M. S. student who conducted an experiment during summer 1996 at the NWS RAH office to forecast local thunderstorm activity using the Raleigh WSR-88D radar and other observing systems, as a natural follow-on to the benchmark study of low-level convergence boundaries in North Carolina by Koch and Ray (1997). Black produced both a six-hour convective outlook and a 60-min nowcast for forecast sectors within 230 km of the radar. Forecast skill assessed using a Brier score showed dependency of skill on spatial resolution of the forecast domain and day of the week. The nowcast study supported the Ray study showing the feasibility of forecasting thunderstorm initiation near low-level boundary interaction regions, with a Critical Success Index achieved of 0.29, which is comparable to other recent attempts to perform nowcasting.

Debra Hoium (Graduate Student) - In 1992, the NWSFO at Raleigh began inviting selected faculty and students at North Carolina State University to participate in its warning activities. One purpose of the university involvement was to document and facilitate the process by which severe thunderstorm and tornado warnings are issued. An M. S. study by Debra Hoium analyzed and summarized the results using reflectivity and velocity products from the WSR-88D. The influence of these products was compared with information obtained in real time from communications with ground-truth sources. It appears that reflectivity products were most often favored as the product that triggered warnings, and that 80% of the warnings initiated or triggered by ground truth were issued for counties downstream of where the report originated. Ground truth was obtained in a timely manner (typically within 30 minutes after the warning was issued). Thus, information that severe weather had been observed could be used as an independent basis for future warnings. Detailed results of this study appear in a paper that has been accepted for publication (Hoium et al 1997).

Kimberly Kreis (Graduate Student) - This M. S. research of a damaging derecho along a stationary frontal boundary showed that, although the synoptic environment was very similar to that found for other warm-season progressive derechos, careful surface analysis provided the key in narrowing the focus for the storm tracks. On the afternoon of the day following the derecho, a severe hail storm developed in central North Carolina where hail stones as large as 7.5 cm diameter caused $3 million in damage. The interesting aspect of the event was that storms all developed very close to a mesoscale thermal boundary apparently formed by the previous day's derecho.

Robert Trayers (Graduate Student) - During the night of 6-7 January 1994, an intense squall line developed rapidly in the Carolinas and swept eastward to the coast producing extreme downburst winds and numerous weak tornadoes, all in an environment which initially appeared only marginally supportive of severe weather. An M. S. study of this event by Robert Trayers showed that the squall line formed as a Cold Front Aloft (CFA), approaching from the west, encountered unstable air aloft over a warm front. VAD wind profiles from the WSR-88D radar provided evidence of the arrival of the CFA. This study helped motivate a task under the new COMET project by Dr. Koch, whose objective is to employ mesoscale model forecasts and VAD displays from upstream stations as forecast guidance to better anticipate such events in the future.

Craig Souza (Graduate Student) - Completed his M. S. thesis early in the project, in which a methodology was developed for forecasting precipitation type that has found wide acceptance at NWS RAH and other regional NWS offices. This benchmark study is discussed in greater length throughout this report

Dr. Lian Xie (Associate Professor) - Project participant in development and realtime operation of the North Carolina Sound flooding model. This capability has been a key tool in the Raleigh forecast office, as is discussed later.

Participants from NWS RAH are numerous; in fact, by any measure (fraction of total staff involved, creative energies expended, etc.), the degree of involvement of RAH forecast staff in collaborative activities with the University is staggering. This kind of involvement is much more aptly described by the term Experimental Forecast Center than by merely "collaborators". Personnel who have been involved at some point during the 3-year project from NWS RAH are listed alphabetically:

Ruth Aiken (Lead Forecaster) - Familiarizes NCSU collaborators with the archive IV capabilities of Raleigh's Doppler radar and trains them to use the graphics tablet for retrieving various radar products. Serves as a member of the Joint NWS - NCSU Case Studies Team.

Phil Badgett (Graduate Intern) - Conducts case studies for verifying the performance of the project's winter storm forecast tools and for testing the collaborative conceptual model - Cold-Air Damming Spectrum.

Ken Cook (Forecaster) - Outreach participant for communicating the results of our collaborative efforts to the private sector.

Edward Delgado (Lead Forecaster) - Conducts ongoing NWS RAH - NCSU Lecture Series featuring presentations on collaborative findings and closely related issues and the "Paper of the Month" series which focuses on meteorological topics pertinent to collaborative projects. Coordinates MASS model simulation studies and serves as a member of the Joint NWS-University Case Studies Team. Principal instigator for the development of the Convective Outlook Product.

Greg Dial (Forecaster) - Worked closely with Tony Ray on the role of convergence boundaries in convective storm interactions prior to his departure from NWS RAH.

Rod Gonski (Lead Forecaster) - Coordinates the joint NWS - NCSU severe and winter weather operations. Provides critical reviews of collaborative tutorials, forecast macros, and operational proficiency reference guides. Co-authored the project's operational proficiency guide for severe pulse thunderstorms. Provided a forecasters perspective on Dr. Koch's recent publication "Operational Forecasting and Detection of Gravity Waves".

Steve Harned (MIC) - Assures that needed communications and computer hardware and software systems are acquired and operating in order to meet project objectives. Reviews contents of collaborative tutorials, operational proficiency guides, and forecast macros. Has provided guest lectures in NCSU meteorology classes.

Gail Hartfield (Forecaster) - The principal producer of forecast macros featuring the integration of collaborative findings and progressive meteorological fields into operations. Develops the "hands on" exercises that bring the meteorological concepts to life and provides experience in using modernized data formats. She is a principal author for the "Beyond AFOS" tutorial series. Provided an NWS operational perspective to Tony Krogh for investigating the use of the Doppler radar to retrieve horizontal thermal gradients and associated thermal advections. A principal developer of the Cold-Air Damming (CAD) Spectrum, she co-authored the Glossary Field Guide on CAD. Conducted case studies for testing the CAD Spectrum and for verifying the projects' winter storm forecast tools. Works closely with the COMET sponsored fellow (Devin Kramer) on MASS model
study of CAD.

Ron Humble (Lead Forecaster) - Works closely with NCSU to expand the utilization of RAH HP workstations in joint collaborations. Heads a team of NWS forecasters to develop GEMPAK and GARP scripts for integrating collaborative findings into NWS operations.

Jeff Jarvis (Intern) - Webmaster for the NWS RAH weather homepage that maintains a research component featuring COMET sponsored NWS RAH - NCSU collaborations and the collaborative NWS RAH - NCSU Lecture Series.

Kermit Keeter (Science Officer) - Principal NWS investigator for all COMET sponsored projects and programs. Supervises the collaborative activities of all NWS participants. Authored the Winter Storm Forecast Process Reference Guide featuring collaborative findings and techniques integrated into station operations. Co-authored the Severe Pulse Thunderstorm Reference Guide and the design of some of the stations' forecast macros integrating collaborative findings and progressive meteorological concepts into operations. He is a principal author of the "Beyond AFOS" tutorial series and a principal developer of the CAD Spectrum and Field Glossary Guide.

Paula Kurilla (Intern) - A participant on the HP development team writing GEMPAK-GARP scripts for integrating collaborative findings into NWS operations.

George Lemons (Warnings and Coordination Meteorologist) - A participant in the outreach activities helping to present to media meteorologists the collaborative contents of the winter weather forecast process reference guide. Assists in familiarizing NCSU collaborators in the use of NWS RAH's Doppler radar level IV archive data.

Michael Monepenny (Forecaster) - Works closely with collaborators providing an operational perspective on the use of modernized technologies to detect gravity waves and low-level convergence boundaries. Documents case studies on boundary-convective storms interactions and integrating the findings into the station's nowcasting program.

Rick Neuherz (Forecaster) - Serves as member of the collaborations HP development team. Documents case studies on collaborative findings featuring boundary-convective storms interactions. Works closely with Dr. Xie on applying the North Carolina Sounds Model into realtime NWS operations during Soundside flooding events.

Jan Price (Lead Forecaster) - Outreach participant for communicating the results of our collaborative efforts to the private sector.

Mike Vescio (Lead Forecaster) - Prior to his departure, played important role in identifying low-top, low-reflectivity severe storms and shallow frontal boundaries in North Carolina, both of which motivated COMET studies (on mesocyclone detection and on boundary detection and understanding). Currently at Storm Prediction Center.

Finally, participants from other NWS offices and organizations have involved people who have played crucial roles in this project:

Hugh Cobb (NWS ABQ and outreach participant)
Dr. Ken Crawford (consultant on project task T5)
John Elardo (NWS MHX and outreach participant)
Greg Fishel (WRAL-TV and outreach participant)
Doug Hoehler (NWS ILM and outreach participant)
Chris Holman (WTVD-TV and outreach participant)
Larry Lee (NWS GSP and outreach participant)
Ed Matthews (WFMY-TV and outreach participant)
Dr. Chris O'Handley (employee of SAIC, Inc., he played crucial role in project task T3)
Dr. Joseph Pelissier (former Deputy MIC at NWS RAH, now MIC at GSP)
Steve Swienckowski (WLFL-TV and outreach participant)
Dr. Ken Waight (employee of MESO, Inc.; under contract from NCSU, he was solely responsible for developing the realtime MASS model capability under project task T4)

1.2 Overview of Objectives and Accomplishments

As previously stated, our ongoing joint investigations have revealed that the forecaster's understanding of collaborative findings progresses through various levels before the finding can be implemented into NWS operations. The path to transferring collaborative findings into forecast operations is marked along the way by significant research/development and training accomplishments. These accomplishments also serve as the means and methods of reaching the ultimate goal of NWS-University collaborations-extensive operational proficiency. Our three-year proposal contained six objectives. Accomplishments achieved in pursuit of each of the six task objectives are now discussed. Other accomplishments that arose during the course of the project but were not specifically mentioned as an original project objective are also presented.

Task T1

Complete collaborative activities between NCSU and NWS RAH personnel initiated under an earlier COMET Outreach Project, either by publishing papers, testing and evaluating new forecast techniques, or refining and extending previously developed techniques and models. Also, continue to provide training opportunities in the proper utilization of these forecast tools to NWS personnel and members of the private sector through workshops and individual sessions.

A) A locally developed regression-nomogram technique for predicting winter precipitation type in North Carolina was improved and expanded to additional sounding sites using 40 years of sounding data and incorporated into a finished M.S. thesis (Souza 1994). This technique has been incorporated into forecast operations at the RAH office, shared with the private sector through workshops and other NWS field offices, and published (Keeter et al. 1995).

B) A method for delineating precipitation type patterns and associated synoptic winter storm types in the Carolinas was incorporated into forecast operations at RAH, and through workshops was shared with the private sector and other NWS field offices.

C) The Janowitz/Pietrafesa NCSU hydrodynamic model was applied with success in predicting the amount of flooding from the Pamlico Sounds associated with Hurricanes Emily and Bertha as they affected the Outer Banks of North Carolina in 1996. This model is run whenever hurricanes or strong winter storms are a threat to coastal areas bordering the Sounds.

D) Kermit Keeter published two papers during the early stages of this project that provide an operational perspective of East Coast cyclogenesis and an overview of the effects of topographyon the evolution of meteorological features in the region (Keeter et al. 1995; Gurka et al. 1995).Reprints of these articles are attached at the end of this report.

E) Presentations featuring these tools were made by project members at the first AMS-sponsored Training Symposium for Operational Forecasters in the Carolinas and Virginia cosponsored by NWS RAH and NCSU. This symposium held in November 1995 represents a major outreach undertaking by our group. This coming fall, we have scheduled presentations on the forecast process for winter storms to the meteorology staffs at the ABC, CBS, NBC, and FOX stations in Raleigh, Durham, Greensboro, and Winston-Salem. Copies of our Winter Storm Forecast Process Reference Guides have already been provided to these stations. We are also scheduled to return to the Airport Operations Conference this fall to again discuss our collaborative techniques for predicting precipitation types. At the Conference last October in Greensboro, we learned from a number of sources that our accurate forecasts of precipitation types for North Carolina were being well utilized by local airports. We have also helped other NWS offices to develop winter weather prediction tools and to conduct case studies using our cold-air damming classification system. The NWS office at Greer, South Carolina now has its own regression equations and nomograms for predicting precipitation types in the "Upstate" of South Carolina. The NWS Office at Wakefield, Virginia has been working with the University of Virginia to develop their own regional precipitation type forecast tools.

F) Assisted by technical review from University collaborators, NWS investigators have produced "Beyond AFOS Toward AWIPS" tutorials (see attachment) that include: an overview of the QG force-response perspective on meteorological processes, differential vorticity and thermal Laplacians, potential vorticity, and tropopause folds. The tutorials provide "hands on" case example exercises bringing the concepts to life and helping to further familiarize forecasters with the utility of GEMPAK and GARP scripts available on the HP workstations. These tutorials are paving the way for a better understanding of future collaborative investigations, including the influences of gravity waves, cold fronts aloft, and jet circulations on North Carolina weather.

G) Building upon a better understanding of meteorological concepts provided by the tutorials, our team has designed and developed PC GRID Forecast Macros that guide forecasters through meteorological data sets that are not available on AFOS. The macros assist the evaluation of the forecast problem(s) of the day by prompting the forecaster to recall the recommended utility of a given meteorological field for a given forecast problem. Given the limited time for forecast preparation in an operational setting, the macros are also an efficient means of streamlining the forecast process. Many forecast macros have been created and new ones are being designed and developed. Popular forecast macros well integrated into operations at Raleigh include: (i) CYCL which provides the dynamical and thermal contributions to the patterns of cyclogenesis seen in the Southeastern U.S.; (ii) DAMG which provide an overview of the various physical processes known to contribute to cold-air damming episodes in North Carolina; (iii) KRMT which provides the partial thickness fields used by the station as a tool for forecasting precipitation types; and (iv) JETS which features jet streaks, divergence aloft, and ageostrophic winds.

H) Once tried and tested strategies and tools are implemented into forecast operations, there is still a need to ensure extensive utilization and periodical review by the forecast staff. For those purposes, we have developed Operational Proficiency Reference Guides. A guide outlining the station's recommended forecast process for predicting winter storms in North Carolina has been developed and is included as an attachment to this report. The guide contains the results of years of NWS-NCSU COMET sponsored collaborations and represents the station's standard for accurately predicted winter storm events. Most Raleigh forecasters are now proficient with these techniques and conceptual models presented in the guide. Through the project's "outreach" activities, the guide has been sent to several other NWS forecast offices and to the NWS Training Center. This coming fall, the guide will be formally presented to the meteorological staffs at several North Carolina TV stations. A second comprehensive guide suitable for periodical review for the purpose of maintaining operational proficiency was completed on pulsating severe thunderstorms. It is a review of the environmental setting and radar signatures of the most frequent severe thunderstorm type that occurs in North Carolina. The contents of the guide were chosen, presented, and reviewed by COMET sponsored collaborators. This coming fall, the guide will be presented to forecasters at the Storm Prediction Center in Norman and to the meteorological staffs at several North Carolina TV stations.
Task T2

Develop and compare new techniques designed to detect boundary layer convergence boundaries in pre-convective situations from synthesizing WSR-88D radar, GOES-Next satellite, and ASOS data.

A) We accomplished our goal to learn to use the new technologies to locate and study low-level convergence boundaries in the convective boundary layer, and their role in thunderstorm initiation. This objective was the #2 research priority for the Operational Support Facility. The RAH staff helped the graduate student (Tony Ray) identify and assemble case studies using WSR-88D data for the purpose of developing a regional radar database. Tony finished his M. S. thesis work on this problem (Ray 1995), and the results have since been published (Koch and Ray 1997). A reprint of this article is attached at the end of this report. Much effort was expended towards disseminating our findings to the operational community, in addition to the Weather and Forecasting paper. This research was presented at AMS and NWA conferences and the Training Symposium for Operational Forecasters in the Carolinas and Virginia. Mr. Ray made a special trip to Morehead City to share his informative findings with forecasters assigned to the new NWS forecast office located there, and he also presented several local seminars at NWS RAH and NCSU. Of special note are the following:

This research has shown that it is possible to perform informative mesoanalyses of summertime convergence boundaries in central North Carolina by combining capabilities of the WSR-88D radar with high-resolution GOES visible imagery and conventional surface data. So-called "random" thunderstorm activity in central North Carolina was found to be either directly attributable to or strongly affected by such convergence zones. Boundary interactions frequently result in new thunderstorm cells. The transition zone between the Piedmont and the Coastal Plain is a preferred location for convergence boundaries. These "Piedmont Troughs" form under intense diurnal heating, and though they are capable of producing convection without any interaction with other features, they can easily be overlooked by the forecaster/analyst. Since they are such a prolific producer of thunderstorms, awareness of their existence should be of primary concern when nowcasting thunderstorms in North Carolina in the warm season.

We made suggestions that the threshold values on the Doppler velocity scale displayed on the WSR-88D PUP can and should be reset by the local NWS WSFO interested in improving the ability to detect boundaries. Since it is also quite difficult to switch the radar into clear air mode and keep it there, the default mode of the radar in the summer should be the clear air mode until thunderstorms threaten part of the warning area. These changes would provide the maximum amount of information about subtle features in the pre-convective environment.

This research has had profound impacts on local NWS operations, as is discussed later. Tony Ray has gone on to work at the Operational Support Facility (OSF) in Norman, where his experiences at NCSU and NWS RAH have proven to be very useful.

B) The M. S. degree research by Krogh (1996) concerned the development and use of a technique to retrieve temperature fields derived from the Velocity-Azimuth Display (VAD) product in Level IV WSR-88D data. The retrieved thermodynamic structure of shallow frontal systems in the mid-Atlantic region was compared to the NMC Eta and our real-time MASS model predictions. This research complements Ray's WSR-88D research on summertime air mass boundaries in North Carolina. In particular,

Vertical wind shears from multiple VAD sites in the mid-Atlantic region were used to estimate the time-height fields of temperature gradient from the geostrophic thermal wind relationship. By referencing "tie points" where a WSR-88D radar is collected with a rawinsonde and triangulating from other rawinsonde locations to the radars, the temperature measured by the rawinsonde was then used in an attempt to recover the temporally and spatially varying temperature and potential temperature fields from the retrieved temperature gradients.

We have compared these 3-hourly temperature retrievals to objective analyses of 12-hourly NWS rawinsonde data and to initial analyses and forecasts from our real-time mesoscale model. The results indicate that it is possible to retrieve meaningful and accurate horizontal temperature gradient information from the VAD, except where frictional and curvature effects ignored in the retrieval process are important. Time-height cross sections of temperature gradient and thermal advection from two cold air damming events display excellent temporal continuity and readily portray important differences in the strength and depth of the cold dome between the cases, as could only be implied in the surface analyses. Verification of the recovered thermal fields at the 12-hourly rawinsonde times is also encouraging. Although errors are significantly larger than for the temperature gradient fields, the temperature fields do display useful information about the time-varying depth and structure of the cold dome (Krogh and Koch 1996).

Task T3

Evaluate spectral and bandpass filter techniques for detecting mesoscale phenomena (such as gravity waves and thunderstorm pressure features) in a real-time setting with high temporal resolution ASOS and GOES-Next data. These data will be used to develop a database of such mesoscale phenomena for case study research and for climatological studies. In particular, a conceptual model of the environment favored by weather-producing gravity waves in the Southeast will be implemented for use as a forecasting tool.

A) Two goals were defined under this task: (1) to develop an objective analysis procedure that will utilize the high temporal information contained within ASOS data to make it possible to operationally detect gravity waves and other propagating mesoscale phenomena; and (2) to determine whether the Uccellini and Koch (1987, hereafter UK87) synoptic-scale conceptual model found for weather-producing gravity wave events, which consists of a jet streak propagating toward an upper-level ridge north of a surface frontal boundary and dynamically unbalanced flow aloft, is sufficiently general to allow for quantitative analysis of operationally available mesoscale model data to predict where and when waves should be expected to occur. Mesoscale gravity waves display periods of 1-4 h, have wavelengths of 50-500 km, and can have important effects upon the sensible weather.

Our COMET research has shown that real-time prediction, detection, and nowcasting of these mesoscale phenomena is feasible, due to recent major advances in operational observing and modeling systems. We completed the development of a methodology designed to operationally detect and forecast propagating mesoscale wave disturbances using ASOS and mesoscale model forecasts. The method has been presented in four operational forums: the NWS Training Symposium held in Raleigh in 1995, the NWA meeting held in Houston in 1995, the AMS Conference on Weather Analysis and Forecasting held in Norfolk in 1996, and repeatedly at the COMET COMAP course. In addition, the entire technique and its demonstrated utility have been published (Koch and O'Handley 1997). A flowchart that can be used by forecasters to both anticipate and nowcast mesoscale gravity waves was developed - upon finding an environment conducive to gravity wave occurrence from analysis of upper-air data and MesoEta model data, the forecaster activates an automated wave detection system that uses objectively analyzed bandpass-filtered ASOS fields. A reprint of our Weather and Forecasting paper is attached at the end of this report.

More specifically, our findings indicate that the ability to predict the likelihood of a gravity wave event rests upon recognizing the UK87 synoptic flow pattern in which such waves are consistently found to occur. The delineation of the most likely region for wave activity can be further refined by computing simple indicators of unbalanced flow and conducting a cursory search for a suitable wave "duct" with MesoEta model data.

Whenever and wherever a suitable gravity wave environment is found, the ASOS (Automated Surface Observing System) pressure data should be carefully monitored for evidence of gravity wave activity. Our automated gravity wave detection system consists of a time-to-space conversion (TSC) adaptation of the Barnes objective analysis scheme to permit waves with horizontal wavelengths of only 150 to be resolved. This has been demonstrated with both idealized waves (plane and arc-shaped) and with real data. The ASOS pressure and wind data first are subjected to an automated bandpass filtering scheme, then are objectively analyzed using TSC, and finally the resultant fields are animated as color perturbation maps for easy forecaster identification of the waves.

The TSC analysis scheme requires accurate knowledge of the wave propagation velocity. A method was developed and successfully tested for this purpose, which is based on an adaptation of wave ducting theory to the mesoscale model forecast fields. Our technique has the ability to extract the mesoscale wave signals from complex time series data. Smaller-scale gravity waves generated by convection are filtered out and are not analyzable.

The entire procedure was demonstrated by Koch and O'Handley (1997) with a gravity wave event that occurred during STORM-FEST. A solitary wave of depression formed as an upper-level jet streak approached an inflection axis in the diffluent height field downstream of the Rocky Mountains. This wave generation region was diagnosed from mesoscale model forecasts as being unbalanced. A wave duct was diagnosed north of a warm front in both the model forecasts and the STORM-FEST soundings over the region traversed by the observed waves. The analyzed pressure and wind perturbation fields successfully portray the evolution of the gravity wave into a wave train as strong thunderstorms developed with the wave. The mesoscale model produced a gravity wave similar in most respects to that analyzed prior to the development of convection. These results suggest that mesoscale gravity waves can be predicted and analyzed with operationally available data and numerical model guidance.

B) Leanne Siedlarz completed her M. S. thesis research, in which she examined the general applicability of the UK87 conceptual model for mesoscale gravity wave occurrence, and performed a climatology of waves as detected in STORM-FEST surface time series data (Siedlarz 1996). Coherent pressure pulse events with amplitudes _ 0.4 mb occurred 34% of the time during this six week period. The four strongest and most coherent wave events happened in an environment that qualitatively fit the UK87 conceptual model. Three detailed case studies were performed to demonstrate that the Koch and O'Handley (1997) wave tracking technique can be reliably used to detect and follow the evolution of mesoscale gravity waves with 5-min pressure data over several states. The case studies reveal interesting interactions between convection and the waves. Her study indicates that these methods hold promise for operational implementation once the full ASOS network is established and means are found to access the 5-min digital ASOS data in real time on the SAC computer. Our followup COMET project will address this need.

C) Although not originally mentioned in our COMET proposal, we have further extended the application of our gravity wave analysis technique to the Palm Sunday tornado outbreak that occurred on March 27, 1994 in the Southeastern U. S. Under combined support from COMET, the Southeastern Consortium on Severe Thunderstorms and Tornadoes, and the National Science Foundation, we completed an in-depth observational and numerical modeling study of the mesoscale dynamics operating in this event. The study has been accepted for publication in Monthly Weather Review (Koch et al. 1997). In particular, we were able to track coherent gravity waves in bandpass and TSC-analyzed barograph data and also the MASS model simulations using both 24 km and 12 km grid meshes. Interestingly, the waves in the model were generated by strong convection in Texas upstream of the outbreak. Our analysis shows these waves to be trapped Wave-CISK modes, and that they played a key role in the generation and modulation of frontal mesolows that were a primary factor in the development of supercell thunderstorms in Alabama and Georgia. Interactions between the waves and mesoscale jetlets produced by the Texas storms were also very revealing.

Task T4

Initiate plans for the implementation of an operational mesoscale modeling facility designed to operate on a low-cost workstation on an as-needed basis, with the capability to predict convective-scale phenomena at the WSFO, and with state-of-the art scientific visualization techniques designed to communicate the most information possible from the diagnostic fields predicted by the model. No hardware purchase is requested for this purpose.

A) Work progressed rapidly on the development of a real-time mesoscale modeling capability and its use in education, training, and experimental forecasting. Early accomplishments:

Under a subcontract, Dr. Ken Waight from MESO Inc. set up and configured the MASS (Mesoscale Atmospheric Simulation System) to run in real-time on a DEC Alpha workstation that was obtained from separate funding. The model setup allowed it to be run easily by a student.

Next the model postprocessing software was written to interpolate the model's sigma coordinate data to the more familiar pressure coordinates for use in GEMPAK, which was determined to be the system for viewing the model forecasts. GEMPAK Unix scripts were written to automatically produce a suite of model forecast products as determined by mutual consent between NWS RAH and NCSU.

An important event occurred when Dr. Waight permanently relocated from the MESO, Inc. headquarters in New York to our Raleigh office. This move in July 1995 assured that we would have a more durable and dependable model forecast facility.

B) The NWS office played much more than a passive, receptive role in the mesoscale modeling effort. Several important actions were taken by NWS RAH to prepare for the eventual use of the mesoscale model in operations. These accomplishments include the following:

Early in the project, discussions were held with Dr. Koch and MASS modelers Waight and Kramer to determine which model configuration and forecast products were desirable to best address operational forecasting problems. Accordingly, we agreed to run MASS on two nested grids using a high resolution planetary boundary scheme and simplified microphysical prediction equations. The innermost grid would have 15 km resolution and cover most of the mid-Atlantic region, while the outermost grid would be a 45 km mesh covering the area east of the Rockies. The list of MASS model diagnostic fields put together by Gail Harfield and Ed Delgado of NWS RAH was primarily concerned with the cold air damming and severe convection forecast problems. The product list grew and underwent substantial changes over time as the technology and needs changed, with the last one being at the very end of this project (April 1997), when more GARP functions were added for use with the MASS model. Delgado played an important role in training forecasters to manipulate MASS model fields using NTRANS/GARP.

Funding was obtained from the Office of Meteorology to obtain an HP workstation (similar to the SAC). This workstation was dedicated as the highest priority to the MASS model, primarily for viewing the model forecast fields. Problems occurred over many months with instability of this computer system, but by 1997 it was deemed sufficiently reliable for us to move the model post-processing to this platform. Once a router was obtained and installed by the electronics staff, the Raleigh office could get on the Internet and obtain MASS model digital data for display on the HP workstation. Thanks to the combined efforts of the RAH-NCSU team, the historical gap between the Raleigh forecast office and other NWS offices that have been more progressive with the development of the HP systems is closing.

Funding was obtained from UCAR/NWS for a fellowship in order to have a graduate student (Devin Kramer) dedicated to the development of model product fields and use of the model to better understand and predict frontal systems in the Carolinas. This development in August 1995 and the continued support for Dr. Waight were the keys to a successful mesoscale modeling facility in Raleigh by 1996, for which we are very appreciative to COMET for continuance of financial support.

C) Improvements were made to the modeling system throughout the course of this project to meet local forecast needs:

The model forecast fields were made available in a much shorter time beginning in 1997 when GEMPAK file conversion was shifted to an "on-the-fly" operation, instead of waiting until the entire model run was completed; the result was that MASS coarse mesh data became available by 3:00 a.m. (in time for the early morning briefing) instead of 7:30 a.m, and the fine mesh data became available by 7:00 a.m. instead of 11:30 a.m.

We increased the forecast period for the 45 km coarse grid from 24 h to 36 h to make the forecasts of greater use in operations.

Dr. Waight found the cause of and corrected the excessive PBL heat fluxes over the Gulf Stream. Detailed vegetation and sea surface temperature fields using AVHRR satellite data were added, because of the recognized importance that these fields could play in the boundary layer developments in convective and coastal frontogenesis situations.

A mesoscale humidity initialization using realtime GOES and radar data to improve on forecast heating rates and precipitation fields was added to MASS.

Dr. Waight assisted a programmer (Patricia Davis), supported under this COMET project, in the development of additional visualization capabilities for the MASS model output fields using IBM Data Explorer (DX), which can be run on the HP workstation and other Unix platforms. We now possess the capability to produce color four-dimensional displays of MASS fields to aid in understanding of three-dimensional airflow patterns and their relationship to the moisture and thermal fields. This capability was demonstrated at the CSTAR meeting.

D) Approximately mid-way through the 3-year project, additional strides were taken by the NWS RAH to progress from simply examining model fields of the same products that forecasters were already accustomed to viewing as output from the NCEP models to the next level for incorporating MASS into the forecast process:

First, Ed Delgado (NWS RAH), Dr. Koch and Devin Kramer (NCSU), and Dr. Waight (MESO, Inc.) worked together to produce a basic familiarization document for the NWS RAH staff titled "Prelude to the MASS". This document describes the model configuration and the process for incorporating MASS into the forecast arena. The end result was forecaster confidence in the model and an increased comfort level with this new source of information.

Second, the NWS and NCSU began using the MASS model as a forecaster training and teaching tool. Unconventional products (e.g., ageostrophic transverse circulations on isentropic cross sections) produced from the model began to be used to educate the NWS staff about mesoscale processes in conjunction with our lecture series, mini-workshops, training notes and tutorials (see Tasks Tl and T6). The MASS model's fine grid forecasts helped to train forecasters about the difference between precise model guidance versus detailed but incorrect model forecasts.

The third step of the learning process was begun toward the end of this 3-year project--namely, the application of the model to local forecast techniques, such as resolving the difficulty in identifying and differentiating among cold air damming events and their lookalikes, predicting winter precipitation type on the basis of predicted sounding structures and/or layer thicknesses, and the use of a mesoscale model in experimental severe weather outlooks. The MASS model is now having an impact on the forecast process, since it is being used to help in the construction of a mesoscale discussion that is added to the State Forecast Discussions, which are shared with other NWS forecast offices, as well as an internal Daily Experimental Convective Nowcast (see attachment showing an example internal forecast discussion). Impressions gathered thus far are that MASS produces detailed boundary layer wind, moist static energy, and precipitation fields from the fine grid that help in the short-range forecasting of where convection is most probable.

E) In summary, the NWS RAH staff is very excited about the introduction of the MASS model to the daily forecast process, since they are no longer restricted to a limited set of AFOS products produced from relatively coarse resolution models at infrequent intervals, and with no refinement of the model to directly deal with local forecast concerns. The MASS model has been running as an experimental mesoscale modeling system at NCSU with 90-95% reliability since January 1996. Other accomplishments concern the multitude of ways that the model fields may be viewed and manipulated by forecasters and researchers using the same tools:

A set of GEMPAK contour plots saved as color image files in GIF format are automatically sent to the MASS model home page. These computer displays replaced the reams of Postscript paper plots that we had been producing earlier for NWS examination. We quickly discovered that MASS output had little impact on the forecast process as long as we continued to produce paper copy. These Web products are viewable on a daily basis by forecasters or anyone else at the following URL: http://www2.ncsu.edu/ncsu/pams/meas/mass.

MASS model output in NTRANS format for forecaster viewing on N-AWIPS, which has been installed on the HP computer purchased by NWS RAH primarily for this purpose. One purpose of doing this was to allow forecasters at NWS RAH to examine animated hourly model output instead of the 6-hourly static fields being made available on the MASS home page. Another reason was to be able to readily intercompare MASS fields with other NCEP model guidance and observational data.

Tape archival of all MASS model coarse and fine grid runs, plus the datasets used to initialize the model for later retrieval. A very important advantage of running a mesoscale model locally is that beliefs (or hypotheses) expressed by forecasters about why certain events or kinds of phenomena were poorly forecast can be tested by re-running the model with different initial conditions or configurations in research studies. The locally run mesoscale model has opened up a gold mine of data for use in collaborative research and university education (see Task T6).

The initial suite of model products created using automated Unix scripts for the morning briefing has been continually updated with specialized model products. The choice of fields, display attributes, and timing for introduction of the new fields has always been determined by joint decisions between NCSU and the forecast staff. Examples of such displays include: vertical cross sections of transverse jet circulations, multiple-level ageostrophic wind and mixing ratio plots, quasi-geostrophic Q-vectors, partial layer thicknesses and probabilities of frozen precipitation (Keeter et al. 1995), and convergence of equivalent potential temperature (a convective predictor field). Raleigh forecasters are now regularly seeing tailor-made MASS model products for cold-air damming spectrum events including Froude numbers and a variety of meteograms. Examples of MASS model products used in winter storm forecasting are attached.

MASS model simulations and case studies are providing a better understanding about the capability of numerical models to predict cold-air damming spectrum events. Sharing this information with NWS forecasters promotes a better working knowledge of numerical models. For example, Devin Kramer's research is giving forecasters an awareness of the interconnections between a model's cloud physics, its handling of diabatic processes, and the resulting impact of the model's forecast of cold-air damming events. While these collaborative findings are indicating that there is still a great deal of room for models to improve in their handling of diabatic processes, forecasters are gaining a much better understanding of the whys of the model's limitations. This in turn places forecasters in a better position to use past forecast experiences, case studies, and subjective conceptual models as alternatives to assuming a perfect program interpretation for a given numerical model solution. Through our collaborative outreach efforts, the cold-air damming spectrum has gained regional acceptance as a useful conceptual model for classifying cold air damming and cool pooling events and for understanding the relative weight of the physical processes that drive spectrum events. There is evidence in forecast discussions from various NWS offices that the spectrum has helped ease case by case discrepancies that stemmed from the misclassification of events and the lack of a standard terminology for coordinating them.

Task T5

Ensure that a surface automated mesonetwork for the state of North Carolina designed after the fashion of the Oklahoma Mesonet will produce high quality, frequently sampled data that can be accessed and used in both academic and NWS environments for improving mesoscale analyses and forecasting. No hardware purchase is requested for this purpose.

A) We brought Dr. Ken Crawford to NCSU to present a seminar on the highly successful Oklahoma Mesonet and to discuss with NCSU and North Carolina state officials the possibility of funding the development of a similar mesonetwork in our state. Although the Oklahoma Mesonet is providing great economic payoffs to the citizens of Oklahoma (for education, agriculture, emergency management, etc.), it does not seem possible at the present time to find the required $2.5M state funding to develop such a network in North Carolina. Prof. Koch also became aware of the very significant involvement that would be required on his part to attract the funding and to manage such a system. Until such funding and management matters have some likelihood of being resolved, it does not appear that any additional action will be taken to develop a state mesonet here.

B) Nevertheless, Dr. Crawford's visit was very helpful for providing us with information needed to make rational decisions concerning the creation and maintenance of a state mesonetwork. In fact, recent discussions involving university officials has resulted in the initiation of a more active State Climate Office at NCSU headed by Dr. Sethu Raman. Dr. Koch serves on the Advisory Board for this office. Attempts are already being made by this office to bring together existing small-scale observing networks in North Carolina under one umbrella, and this office is seeking to find funding resources that would enable adding more surface stations, with the ultimate goal being one mesonet station per county. Dr. Raman has meetings set up with Dr. Crawford to discuss matters related to this need.

Task T6

Conduct lecture series on the dynamics of, and techniques for detecting, thermal moisture boundaries, frontogenetical circulations, jet streaks, density currents, bores, gravity waves, and other mesoscale phenomena capable of initiating deep convection. A new course on the principles of radar meteorology (including the use of laser disk computer-aided learning modules from COMET) will be offered. This will form a very useful complement to the NWS NEXRAD training program in which RAH will be engaged in the first year of this effort.

A) COMET-sponsored collaborators developed and maintain an ongoing seminars program that features topics from collaborative studies and closely related issues. Following is a compilation of the joint seminars that have been offered in this series:

18 September 1996
NCSU Graduate Student UCAR Fellow Devin Kramer
"Applying a Mesoscale Model (MASS) to Cold-Air Damming Events"

24 September 1996
NCSU Prof. Al Riordan
"Cold Fronts Aloft"

12 November 1996
NCSU Graduate Student Chris Vandersip
"Identifying Mesocyclones Using the WSR-88D"

22 November 1996
NWS RAH Lead Forecaster Ed Delgado
"(Re)Introduction to Isentropic Analysis and Cross Sections"

26 November 1996
NCSU Prof. Steven Koch
"A Short Course on Jet and Frontal Circulations (Part 1)"

5 December & 12 December 1996
NWS RAH Science Officer Kermit Keeter
"The Winter Storm Forecast Process"

24 January, 21 February, & 29 May 1997
NCSU Prof. Steven Koch
"A Short Course on Jet and Frontal Circulations (Part 2)"

I I March 1997
Meso, Inc. Dr. Ken Waight
"Evaluation of MASS Model Verification on Recent Cases"

26 March 1997
SUNY Prof. Lance Bosart
"Impact of the Appalachian Mountains on Cyclones and Fronts"

22 April 1997
NWS RAH Lead Forecaster Joel Cline
"A Discussion of GOES 8 Satellite Data: Interpretation and Resources"

B) Prof. Koch has developed a Web-based lecture series on Quasi-Geostrophic Jet Dynamics and Jet Transverse Circulations to supplement his lectures given at the NWS RAH office. The Web page contains MASS model simulations intended to exemplify principles given in the lectures, so that forecasters can apply their knowledge to real-world situations. This Web page is currently being modified to provide more of the fundamentals following Dr. Koch's attendance at the Unidata Workshop on Satellite Meteorology Instruction held in Boulder in June 1997.

C) Three new university courses developed by Prof. Koch that address important needs in mesoscal education and training (Radar Meteorology, Atmospheric Convection, and Mesoscale Precipitatio Systems) have all been influenced by the collaborative efforts with the NWS RAH. The COMET Computer-Based Learning System was acquired and used with great success in the radar course and a senior undergraduate course that Prof. Koch teaches on synoptic and mesoscale weather analysis and forecasting. All of these courses have benefited NWS staff, several of whom have sat in on lectures. In particular, Ed Delgado (NWS RAH) took both Radar Meteorology and Atmospheric Convection for credit.

D) Although not fully envisioned at the time that the original COMET proposal was written, the real-time MASS modeling project has enabled incorporation of this model as a new tool in university education, namely the Mesoscale Precipitation Systems class in Fall 1996. Students were required to apply concepts on mesoscale dynamics learned in lecture to the MASS model datasets. It is very difficult to obtain complete mesoscale datasets for such purposes in research, not to mention education. Whether the model forecasts were accurate is immaterial in this regard, since the MASS fields were treated as a dynamically consistent, four-dimensional datesets. The experience was incredibly valuable, because the 15 km resolution fields are particularly well suited for diagnostic analysis of such mesoscale processes as conditional symmetric instability, frontogenesis, and mesoscale gravity waves.

E) Steve Harned (MIC at NWS RAH) presented lectures and conducted several labs on map analysis in February 1996 in Dr. Koch's senior undergraduate class on Weather Analysis and Forecasting (MEA 444). Kermit Keeter (NWS RAH) gave invited lectures on "Regional Forecast Problems at the Raleigh Forecast Office" in April 1995 and on "The Winter Storm Forecast Process" in April 1997 in MEA 444.

F) The ideas and forecast tools generated from the COMET sponsored NWS RAH-NCSU collaborations have been shared with the private sector and with other NWS field offices. With assistance from and participation by NCSU faculty and graduate students, the RAH office has conducted regional workshops highlighting the effects of the area's diverse topography upon the evolution of meteorological features and their associated forecast problems. As our efforts produce more forecast tools, ideas, and conceptual models that are incorporated into the forecast process at the NWS RAH office, there is a need to ensure that others in nearby NWS forecast offices and in the private sector become familiar with the utility of these tools; and further, that all NWS RAH forecasters become and remain proficient in employing these tools operationally. This training is an absolute necessity for ensuring that collaborative research findings are correctly and extensively used in the operational arena. Informal training in the form of individual sessions continues to be conducted by project members for both NWS and private sector meteorologists. Formal training in the form of workshops and symposia have also occurred, including the Forecaster's Training Workshop conducted by the RAH office for forecasters at the Greenville-Spartanburg (GSP) office, and the Training Symposium for Operational Forecasters in the Carolinas and Virginia. The latter symposium hosted by the RAH NWS office was very successful and well attended. As a result, several NWS offices and a few meteorologists in the private sector are now understanding and employing COMET sponsored ideas and forecast tools, including: (a) a classification scheme for cold air damming and damming look-a-likes, (b) a classification scheme for low-level summertime boundary types detectable in WSR-88D imagery, and (c) objective schemes for predicting precipitation types in winter storms, and water levels on the Carolina Sounds in coastal storms.

2. Summary of University/NWS Exchanges

A) The move of both the Raleigh WSFO and several of the faculty (including Dr. Koch and his graduate students) to the Research III building on the Centennial Campus has fostered much stronger collaboration than was realizable prior to the colocation. This was particularly helpful in the Doppler radar studies that involved Tony Ray and Tony Krogh, because of their need to collect radar data in real time. Naturally, the colocation has also made it more convenient to have group meetings, present joint seminars, have forecasters attend university classes, and have Profs. Koch and Riordan and a cadre of selected students provide assistance to NWS RAH personnel during severe weather events and so become more aware of operational requirements and procedures. Students (especially Deb Hoium) could thereby document the warning process to determine the basis for the warning decisions made, and to allow forecasters to study the events afterwards.

B) As discussed earlier, the NWS at RAH and NCSU have provided valuable familiarization and training for the region's new forecasters on the area's most critical forecast problems. The results of joint climatological studies and case studies illustrating the region's variable weather conditions and the impact of the local topography on meteorological systems has also been made available to members of the private sector. These outreach exchanges are extremely valuable to the NWS as new field offices open and existing offices undergo immense changes during the modernization and transition period.

C) Our collaborations concerning mesoscale modeling, the forecast process, and regional forecast problems has cross-pollinated a number of closely related activities not directly funded under this project. Other projects dealing with the nowcasting and short-term forecasting of convective storms (e.g., Kreis 1995; Trayers 1996; Black 1997), the application of a forecast paradigm for predicting violent tornado outbreaks in the Southeastern U. S. (Hartfield and Gonski's PC GRIDS macros), and a better understanding of the severe storms warning process (Hoium et al. 1997) have all benefited from, been motivated by, and in turn affected this COMET-funded effort. Two examples of what we mean by this are provided. First, Black's nowcast experiment was undertaken largely because of the benchmark study of low-level convergence boundaries in North Carolina by Koch and Ray (1997); the results of this project will lay the foundation for future efforts to conduct thunderstorm nowcasting. Second, Trayers' study of cold fronts aloft in North Carolina has motivated us to undertake in our future COMET effort an intensive search for other such events using the WSR-88D VAD from Raleigh and surrounding sites and the MASS model.

D) A NWS Raleigh home page has been created, which features a research section that highlights COMET sponsored NWS-NCSU collaborations. The home page has stimulated a number of favorable comments and inquiries about university NWS joint applied studies. The URL of the home page is: http://www.nws.noaa.gov/er/rah/

These NWS-University collaborations are conducted in a setting where all participants are charged with a number of missions that limit their time for collaborative endeavors. Research faculty teach classes and conduct other sponsored research, students study for exams, forecasters issue forecasts and warnings. An appreciation for the workload of collaborative participants and empathy for the rigors of NWS rotating shift work is vital to the successful management of collaborative activities. From the NWS perspective, there is nothing in the job descriptions of NWS forecasters that, to a large degree, direct them toward collaborative activities. It must be demonstrated that collaborations will pay professional development dividends and result in better tools for the fundamental mission of issuing forecasts and warnings. This reality also necessitates a tolerance for intrusions of uncontrollable events (e.g. workload associated with Hurricane Fran) that infringe upon a planned schedule for carrying out collaborative activities. Realizing that the transfer of research into operations takes time, there must also be a commitment by investigators to continue a collaboration beyond the end of funding in order to realize the full potential of a collaborative finding. Finally, the educational value obtained by students working in a collaborative setting enhances that obtained in a traditional classroom setting and pays dividends to the meteorological community long after funding for the collaboration has ended.

3. Presentations and Publications

The following peer-reviewed publications, conference proceedings, theses, and presentations can be reported as being due either totally or in part to COMET sponsorship of this project:

Peer-reviewed publications (chronological order). Reprints are attached.

Gurka, J. J. E. P. Auciello, A. F. Gigi, J. S. Waldstreicher, K. K. Keeter, S. E. Businger, and L. G. Lee, 1995: Winter weather forecasting throughout the eastern United States. Part 11: An operational perspective of East Coast cyclogenesis. Weather and Forecasting, 10, 21-41.

Keeter, K. K., S. Businger, L. G. Lee, and J. S. Waldstreicher, 1995: Winter weather forecasting throughout the eastern United States. Part III: The effects of topography and the variability of winter weather in the Carolinas and Virginia. Weather and Forecasting, 10, 42-60.

Koch, S. E., and A. C. Ray, 1997: Mesoanalysis of summertime convergence zones in central and eastern North Carolina. Wea. and Forecasting, 12, 56-77.

Koch, S. E., and C. O'Handley, 1997: Operational forecasting and detection of mesoscale gravity waves. Wea.
and Forecasting, 12, 253-28 1.

Koch, S. E., D. Hamilton, D. Kramer, and A. Langmaid, 1997: Mesoscale dynamics in the Palm Sunday tornado outbreak. Mon. Wea. Rev. (cond. accept.).

Hoium, D. K., A. J. Riordan, J. Monahan, and K. K. Keeter, 1997: Severe thunderstorm and tornado warnings at Raleigh, North Carolina. Bull. Amer. Meteor. Soc. (cond. accept.).

M.S. Theses Completed from Partial COMET support to North Carolina State University (chronological order)

Souza, Craig, 1994: The operational forecasting/nowcasting of precipitation types in the Southeastern U. S.: Type of precipitation - descriptive and objective guidance (under the direction of Prof. A. J. Riordan).

Kreis, Kimberly W., 1995: Synoptic and mesoscale severe weather: A two day case study of a Derecho and local hailstorms under the direction of Prof. A. J. Riordan).

Ray, Charles A., 1995: Detection of summertime convergence zones in central and eastern North Carolina using the WSR-88D Doppler radar, 193 pp (under the direction of Prof. S E. Koch).

Krogh, Tony C., 1996: Determination of frontal structure in the mid-Atlantic region from WSR-88D Doppler Radar Velocity Azimuth Display, 68 pp (under the direction of Prof. S E. Koch).

Siedlarz, Leanne M., 1996: A climatology of mesoscale wave disturbances seen in mesonet data during STORM-FEST, 200 pp (under the direction of Prof. S E. Koch).

Trayers, Robert W., Jr., 1996: Winter severe weather: A case study of the intense squall line of 6-7 January 1995 (under the direction of Prof. A. J. Riordan).

Black, M. A., 1997: Experimental forecasts of precipitation in North Carolina (under the direction of Prof. A. J. Riordan).

Conference papers (chronological order)

Ray, C. A., S. E. Koch, and G. Dial, 1995: A study of summertime convergence boundaries in eastern North Carolina using the WSR-88D Doppler radar. 27th Conference on Radar Meteorology, Vail, CO, Amer. Meteor. Soc., 10-12.

Koch, S. E., Y.-L. Lin, M. L. Kaplan, M. Vescio, A. Langmaid, D. Hamilton, and D. Kramer, 1996: Mesolow dynamics in the Palm Sunday tornado outbreak. Preprints, 18th Conf. on Severe Local Storms, Amer. Meteor. Soc., San Francisco, CA, 47-51.

Delgado, E., K. Waight III, D. Kramer, and S. E. Koch, 1996: Employing the MASS model as a teaching and forecasting tool at the National Weather Service Forecast Office Raleigh, North Carolina. Preprints, 15th Conference on Weather Analysis and Forecasting, Norfolk, VA, Amer. Meteor. Soc., 176-179.

Koch, S. E., and C. O'Handley, 1996: Towards real-time detection and forecasting of mesoscale gravity waves. Preprints, 15th Conference on Weather Analysis and Forecasting, Norfolk, VA, Amer. Meteor. Soc., 253-256.

Kramer, D., S. E. Koch, and K. Waight III, 1996: Mesoscale model analysis of a cold air damming event in the Carolinas during December 18-19, 1995. Preprints, 15th Conference on Weather Analysis and Forecasting, Norfolk, VA, Amer. Meteor. Soc., 555-558.

Krogh, T., and S. E. Koch, 1996: Determination of frontal structure in the rnid-Atlantic region from WSR-88D Doppler Radar Velocity-Azimuth Displays. Preprints, 15th Conference on Weather Analysis and Forecasting, Norfolk, VA, Amer. Meteor. Soc., 419-421.

Siedlarz, L., and S. E. Koch, 1996: A climatology of mesoscale wave disturbances seen in surface mesonet data during STORM-FEST. Preprints, 15th Conference on Weather Analysis and Forecasting, Norfolk, VA, Amer. Meteor. Soc., 249-252.

Other presentations (chronological order)

Vescio, M., and K. Keeter, 1993: "Radar Signatures Associated with Severe Storms in the Carolinas", presented at National Weather Association annual meeting, Raleigh, NC.

Koch, S., and K. Keeter, "Storm Analysis and Training (START) Project, 1993: Future Plans for Collaboration between North Carolina State University and WSFO Raleigh", presented at National Weather Association annual meeting, Raleigh, NC.

Koch, S. E., 1994: Drylines, gravity waves, and downslope winds. Presented at the Unidata Workshop on Teaching Mesoscale Meteorology in the Age of the Modernized National Weather Service, Boulder, CO.

Marine forecasting for the coastlines of the Carolinas - a two day workshop held at NWS RAH in Nov 1994.

Variability of winter weather in the Carolinas and Virginia and the effects of topography on the evolution of meteorological systems - a two day workshop held at NWS RAH in Dec 1994.

Koch, S. E., 1995: Mesoscale structure in wintertime cyclones: The role of gravity waves. Presented at COMET Mesoscale Analysis and Prediction COMAP 95-1 Course, Boulder, CO.

Koch, S. E., 1995: Detection and forecasting of gravity waves in extratropical cyclones. Presented at COMET Mesoscale Analysis and Prediction COMAP 95-2 Course, Boulder, CO.

Badgett, P. 1995: The scourge of the Piedmont airmass: freezing rain and thunder. Presented at the 20th Annual Meeting of the National Weather Association, Houston, TX.

Hartfield, G., 1995: The spectrum of cold air damming and damming look-alikes. Presented at the 20th Annual Meeting of the National Weather Association, Houston, TX.

Keeter, K., R. Gonski, S. Harned, A. Riordan, and S. E. Koch, 1995: NWS-University joint severe weather operations. Presented at the 20th Annual Meeting of the National Weather Association, Houston, TX.

Koch, S. E., C. O'Handley, and K. Keeter, 1995: The potential for operational detection and forecasting of mesoscale gravity waves in the Modernized Weather Service. Presented at the 20th Annual Meeting of the National Weather Association, Houston, TX.

Koch , S. E., C. A. Ray, and K. Keeter, 1995: Use of the WSR-88D doppler radar, GOES imagery, and surface data in the mesoanalysis of summertime convergence boundaries. Presented at the 20th Annual Meeting of the National Weather Association, Houston, TX.

Hartfield, G., K. Keeter, D. Kramer, and S. E. Koch, 1996: Operational use of a mesoscale model within a diagnostic and forecast methodology for cold air damming and lookalike events. Presented at the 21st National Weather Association Annual Meeting, Cocoa Beach, FL.

The NWS Training Symposium for Operational Forecasters in the Carolinas and Virginia (Nov 1995).
The following talks and/or poster presentations were given by COMET participants at NCSU and NWS RAH:

"Prediction and detection of weather-producing gravity waves" (Koch)

"Employing the MASS model as a teaching and forecast tool at the NWS Forecast Office Raleigh" (Delgado,
Koch, Waight, and Kramer)

"WSR-88D radar detection of low level convergence boundaries in North Carolina" (Koch)

"Determination of frontal structure in the mid-Atlantic region from WSR-88D Velocity-Azimuth Displays"
(Krogh and Koch)

"Towards Operational Forecasting of Mesoscale Gravity Waves" (Koch and Siedlarz)

"The Spectrum of Cold Air Damming and Damming Look-Alikes" (Hartfield)

"The Influence of Topography in the Carolinas and Virginia on the Evolution of Meteorological Features and
their Associated Weather" (Cook)

"The North Carolina State Sound Basin Model for Predicting Water Levels on the Albermarle and Pamlico
Sounds" (Xie and Neuherz)

"Guidelines for Predicting the Transition Zone of Mixed Precipitation in North Carolina" (Keeter)

"Objective Forecast Scheme for Predicting Precipitation Types in the Southeastern U. S." (Keeter and Souza)

Forecaster's Training Workshop at the NWS GSP Office. The following talks were given by COMET
participants at NWS PAH:

"Understanding the Precipitation Type Forecast Process" (Keeter)

"The Spectrum of Cold Air Damming and Damming Look-Alikes" (Hartfield)

4. Summary of Benefits and Problems Encountered

4. 1. Benefits to North Carolina State University

A) over the course of the past three years, a close working relationship has developed between NCSU researchers and NWS RAH forecasters, not just because of the colocation, but primarily due to the strong mutual interest in project objectives and the realization that the collaboration has resulted in numerous tangible benefits to both groups. Since the beginning of this COMET Cooperative Project, NWS RAH has provided desk space in the forecast office to several students working on COMET research activities (Ray, Krogh, Kramer, and Vandersip) and has given them virtually unlimited access to their equipment for data gathering purposes. The RAH office has made it possible for us to obtain needed WSR-88D data on low-level convergence boundaries and VAD data from WSR-88D sites in the Carolinas and Virginia by setting up user functions on the PUP to dial up the needed data. This assistance has been crucial in making it possible to use the WSR-88D for studies of low-level boundaries. Moreover, we would probably have missed the majority of interesting events were it not for the interest expressed by the NWS personnel.

B) The assistance provided by NCSU personnel to NWS RAH during joint severe weather operations has actually been tremendously beneficial to the university. Students with no previous operational experience have been exposed to the operational environment and could see possible applications of their research studies. NWS RAH received needed assistance in mesoanalysis, seeking "ground truth" of severe weather, and keeping forecasters current with severe weather reports.

C) The close working relationship between NCSU students and faculty and NWS personnel has enabled both the Raleigh office and the university to actively seek to educate students and forecasters on mesoscale forecast problems and phenomena and the MASS model being run at NCSU; i.e., we have come to understand more fully that the educational process is a two-way process between research and forecasting. The MASS model forecasts have found their way into the requirements for completing a graduate-level mesoscale dynaymics course and an undergraduate-level synoptic weather forecasting course at NCSU. Forecasters feed back information to MASS modelers and researchers at NCSU about the performance of the model, interesting weather events that suggest the possible influence of various mesoscale phenomena (e.g., gravity waves and cold fronts aloft), and the forecast/nowcast process.

D) The NWS office has taken a highly proactive role in obtaining the manpower and computer and telecommunications systems needed to provide experimental mesoscale model forecasts to the forecaster. Special efforts by Kermit Keeter and Steve Harned in helping to gain UCAR/NWS support for the graduate student dedicated to the modeling effort (Devin Kramer), in obtaining the HP workstation dedicated to the MASS modeling project, and most recently, in gaining approval for access to real-time wideband WSR-88D data from the Raleigh radar are of special note.

E) Other benefits include a lecture in Dr. Koch's Weather Analysis and Forecasting class on mesoscale forecast problems from the operational perspective offered by Rod Gonski of NWS RAH, the sharing of operational perspectives on the nature of non-classical fronts in North Carolina by Gail Hartfield of RAH to Dr. Koch and Tony Krogh, the archiving of cases of cold air damming by RAH personnel for future research, and many other things too numerous to mention. Perhaps the biggest benefit of all is the atmosphere of true excitement, encouragement, and dedication exhibited by the office.

4.2. Benefits to the National Weather Service Raleigh Office

From the perspective of NWS RAH, much has been accomplished to benefit operational forecasting resulting from the collaborative activities sponsored by COMET.

A) This project has been an immense help to the Raleigh forecast office in its attempts to focus upon mesoscale processes and forecasting techniques. Collaborative interactions with NCSU has meant an increased awareness of and exposure to research; consequently, attitudes are being modified as NWS forecasters continue to evolve into operational scientists. This process has accelerated as NCSU opened even more doors to the information highway (e.g., Internet access), and provided new mesoscale data sets and tools (e.g., the MASS mesoscale model output).

B) The learning curve regarding the new operational technologies has accelerated through both formal and informal exchanges between the University and NWS. Forecaster knowledge of mesoscale models and the capabilities of the WSR-88D have both been enhanced by the collaborative activities involving research, formal education, and training. Knowledgeable graduate students have brought additional insights into the operational arena as they investigate thesis topics pertinent to operations (e.g., WSR-88D capability to detect shallow convergence boundaries). In addition, the NCSU efforts with MASS output has allowed NWS forecasters to view and work with non-traditional meteorological fields in the operational forecast process. With so much data available for viewing, NCSU and NWS collaborators have recently turned their attention to the need to streamline the forecast process. For example, meteorological scripts for recognizing and forecasting CAD events have been created with the limitations of forecast preparation time well in mind.

C) Despite being in an exceptionally difficult area for predicting precipitation types in winter storms, the Raleigh Forecast Office has one of the best winter storm watch and warning verification rates of any NWS office in the country. Spanning three COMET Cooperative Programs, the winter storm forecast process has been refined and standardized. During the past two years, NWS and University collaborators discovered the occurrence of "corridors of predominant precipitation type" that are associated with secondary cyclogenesis. Using the station's partial thickness techniques, these corridors have been successfully predicted over the past two winter weather seasons. Local newspapers, including the Raleigh News and Observer and the Greensboro Record, have featured articles on the accuracy of the forecasts and the techniques used. Last year alone, a number of Airport Authorities in North Carolina, including Greensboro and Laurinburg, indicated that the degree of specificity and accuracy in our precipitation type forecasts enabled them to save thousands of dollars by applying at the right time the right chemical treatments to their runways.

D) Through our collaborative outreach activities, media forecasters now have a better understanding of Raleigh's forecast techniques, which in turn has reduced the degree of variability and errors in winter weather forecasts made available to the public through and by the media. Last year we compiled an Operational Proficiency Reference Guide (see attachment). This guide features the collaborative components to the station's winter storm forecast process. The Winter Storm Forecast Process Reference Guide has been made available to other nearby forecast offices as well as the NWS Training Center. The guide is a reference source for seasonal and periodical review by forecasters and helps to ensure that the techniques are extensively used correctly and understood. Custom designed winter storm forecast products produced from the MASS model have been excellent tools for demonstrating the forecast utility of Raleigh's techniques to the media and other NWS forecast offices.

E) The research on low-level convergence boundaries (Koch and Ray 1997) has had considerable impact on operations. Recognition of the nature of shallow boundaries in North Carolina has improved the station's mesoscale surface analyses. The various categories of boundary types are being recognized, analyzed, and tracked from one forecaster to another with better continuity. The tendency to misidentify various boundary types as synoptic-scale fronts has decreased. These experiences led to the decision by the office to use Ray's work as the foundation for participating with Prof. Al Riordan in a short-term convective outlook and nowcasting experiment in summer 1996. Results of this experiment were reported in an M. S. thesis (Black 1997). Raleigh forecasters have experimented with a variety of meteorological fields pertinent to forecasting convective storms. These initial findings laid the foundation for current forecast macros and GEMPAK-GARP scripts used in producing convective outlooks by the Raleigh Office since May 1997. These outlooks provide short-term forecasts of the areal coverage, the timing for, and the type of convective storms expected in the station's county warning area. This product has become a reality, based in part, upon the use of the MASS model and the Koch and Ray (1997) and Black (1997) studies. The outlook product has received praise from storm spotter groups who use it to access the day's potential for activating the "Skywarn" spotters network.

F) Given NWS dependency on very limited AFOS data sets, prior to our collaborations the vast majority of Raleigh forecasters did not consider the possible influence of such mesoscale upper-level features as gravity waves, cold fronts aloft, jet circulations, potential vorticity, and tropopause folds on regional weather. Through COMET-sponsored activities including publications, the RAH-NCSU Lecture Series, training tutorials, forecast macros, and GEMPAK-GARP scripts, most Raleigh forecasters now have a basic understanding of mesoscale upper-level features and are aware that they can and do significantly affect the region's weather. Forecast macros and GEMPAK-GARP scripts were developed by collaborative participants and are being utilized to evaluate the roles of jet streak circulations, tropopause folds and potential vorticity in the region's weather events. Collaborative case studies are demonstrating the utility of the WSR-88D radar to show clues hinting at the presence of possible gravity waves and/or cold fronts aloft (this is a topic for intensive research in our follow-on COMET project). Better skill and more extensive utilization of these meteorological concepts will emerge as collaborations continue to promote these concepts to Raleigh forecasters.

G) Historically, there has been a tendency for NWS forecast offices to classify a variety of weather events as cold-air damming based solely upon pattern recognition associated with mean sea level pressure analysis. Many NWS forecasters were assuming cold-air damming whenever there was a ridge of high pressure extending along the eastern seaboard. Forecast discussions for the same event from different NWS offices often varied greatly as to the nature and expected evolution of potential cold-air damming events. During the past cool weather season, forecast discussions from Raleigh and nearby NWS offices indicated significant progress regarding the understanding of the various types of cool pooling and cold-air damming and the physical processes that drive these events. Our collaborative conceptual model providing a "spectrum of cold-air damming" events based upon the physical processes that drive them has reduced the occurrence of the more obvious errors of misclassifying events as cold air damming. The utilization of our CAD Glossary for Field Forecasters (see attachment) in forecast discussions has decreased weather coordination misunderstandings between NWS forecast offices. We have also shared with other NWS forecasters the preliminary findings concerning the performance of mesoscale models in portraying the evolution of CAD events as well as our CAD Forecast Macro which examines the physical processes that account for the evolution of CAD spectrum events. As a result, we have noted that Raleigh forecasters are now better able to more critically evaluate the MOS guidance, to better recognize the tendencies of the numerical weather models to erode the CAD dome too quickly, and to better anticipate the timing and occurrence of severe convection associated with the shallow boundary that marks the southern and eastern edge of the surface based cold air dome during "in situ" cold-air damming events.

H) Selected students and faculty, including COMET collaborators, continue to participate in joint severe weather operations that were begun under the sponsorship of the now defunct Southeast Consortium on Tornadoes and Severe Thunderstorms. These successful severe weather operations were expanded last year under the auspices of COMET to include winter storm events. Joint operations are a very popular means for student collaborators to gain an operational perspective of meteorology early in their career and are a valuable tool for fostering collaborative relationships.

I) Over the past several months, we began holding regular meetings between RAH-NCSU collaborators for the purpose of facilitating joint studies and the sharing of information on significant weather events affecting the area. We have found these meetings to be a very effective means of promoting case study events and providing feedback to the MASS modelers. Already, there has been extensive sharing of information on an unusual late night severe hailstorm event that has led to a joint study between a Raleigh forecaster and an NCSU undergraduate honors student, and will be featured at the upcoming NWA National Meeting.

4.3. Problems Encountered by North Carolina State University

A) Early in the project, the success of the real-time mesoscale model activity was severely limited by manpower restrictions and unreliability in the data ingest system. The lack of a student dedicated to running the model and producing output products and the fact that we did not have an in-house model expert created some real limitations in our ability to deliver useful products to forecasters and to use the real-time simulations for research and teaching purposes. These two manpower problems were resolved by having Dr. Ken Waight relocate to our office, and having the UCAR/NWS fellow (Devin Kramer) and the NWS person (Ed Delgado) dedicated to seeing that the model is run in real-time, that the required GEMPAK Unix scripts are written to produce model forecast products, and that proper training is conducted to transfer this new technology to forecasters. While we were only 30% successful in the first few months in producing daily model forecasts on demand, due in part to these problems and also because of an unreliable data ingest system, we are now at the 95% level of success. Profs. Koch and Riordan were awarded NSF support to acquire the needed computer equipment to update and replace our overtaxed computer server. However, lack of a permanent Local Data Manager (LDM) Administrator at the university has meant that graduate students are relied upon to maintain the LDM feed from Unidata that is the foundation for most of our efforts, including the MASS model. With the expected departure of Devin Kramer in September 1997, we are once again faced with this crisis.

B) The original goal was to use the second HP workstation purchased under COMET funding to convert the raw Mass model data to GEMPAK format used in N-AWIPS, create GEMPAK plots, and display the plots. Although we achieved the capability to display the plots using NTRANS on the HP SAC computer, that computer was dedicated to other NWS purposes for much of 1996. Thus, we were forced to use a less reliable workstation in the NCSU meteorology department to perform the GEMPAK functions; this caused severe problems and was a factor in our deciding to temporarily deactivate the MASS model home page in summer 1996. Although the problem seemed to have ameliorated earlier this year, recently the operational HP computer has begun to crash repeatedly. A hardware problem has been identified and the replacement parts are on order. Additionally, there has been a memory card problem on the HP research workstation. Thus, these problems have limited our use of the research HP computer for our collaborative efforts.

C) Other difficult problems concern the WSR-88D display issues, as this affects any ability to perform operational mesoanalysis using the radar to detect low-level convergence boundaries. For instance, much time and persistence are needed to keep the radar operating in clear air mode. Manual intervention is required to override the precipitation mode in which the radar usually is found on any convective day; Mike Black gave up on trying to keep the radar in the clear air mode in the convective nowcast experiment performed in 1996. Secondly, the vertical cross section capability of the WSR-88D was designed for study of deep systems (thunderstorms), not shallow convergence features, because the displays always extend to 60K ft; we have found that virtually no use can be made of these displays for nowcasting purposes. What is the solution? Clearly, a second radar cannot be purchased! Perhaps another PUP display can be built into the radar system in a future build of the system, but in the short term, it seems that the only solution is to try to display a precipitation mode reflectivity field that does not filter out the lower level (<20 dBZ) values whenever the radar meteorologist has the opportunity to provide this display without deleteriously affecting the warning process. Being able to simultaneously display unfiltered base reflectivity returns and filtered composite reflectivity returns is needed to maximize NWS capability to nowcast without interfering with the warning process.

D) A new staffing issue has arisen fairly recently with the change in policy regarding the Air Force Institute of Technology (AFIT) program. The Air Force is now sending most of its graduate students for education to its training facility at Wright-Patterson Air Force Base in Dayton, Ohio. Given the reality of very limited COMET funding (or the distinct possibility this year of zero funding for the Outreach Program!!!), we have relied in the past upon AFIT students to conduct much of the research proposed to COMET (our past project did not have a single penny in the budget for graduate student support!). In fact, every one of the M. S. theses listed in this report as "COMET-supported" except one (Michael Black) would not have been possible without AFIT support (and the COMET graduate student fellowship program that has supported Devin Kramer on the MASS model project is now dead). Not a single new AFIT student is coming to NCSU in Fall 1997; unless the situation changes, this can only have serious consequences for the Outreach Program at this university.

4.4. Problems Encountered by National Weather Service Raleigh Office

A) From 1994 to 1996, staff vacancies at RAH caused the principle NWS investigator to work a large number of operational shifts. Over several years, the collaborative activities between NWS RAH and NCSU have grown from a few studies with a limited number of participants to a far reaching program involving a large number of NWS and NCSU personnel. Throughout, the staffing levels at RAH have been equivalent to other WSFOs not engaged in pervasive collaborations with a university. It is unrealistic to expect RAH t o continue a demanding and pervasive collaborative relationship with NCSU with staffing levels held to those of any typical WSFO. This is especially true given the increased time that must be devoted to modernization requirements and the continued disproportionate forecast workload (relative to new offices) that the traditional WSFOs continue to carry. Essentially, the RAH office has been operating an experimental forecast unit for quite some time with no more than regular staffing levels.

B) Gains have been made in the continuing computer systems problems experienced by the NWS at Raleigh. Thanks to NCSU, the MASS model output has been available on its own home page through the Internet. Hardware and/or software problems were experienced when using the SAC computer for the MASS output display. These problems prevented the consistent display of MASS through N-AWIPS. Late last year, the Raleigh office took steps to correct the problem, and the MASS output was again available on the SAC HP in 1997; however, the recent hardware problems with the operational SAC computer mentioned above have once again prevented full use of the research SAC computer for this project.

C) Until recently, the RAH office experienced a chronic shortage of electronics technician (ET) help. This meant a reduction in system configuration time for the Electronics System Analyst (ESA) and a delay in using the two HP workstations to their full potential. The delay hampered some aspects of the NCSU-NWS collaborations, such as the full utilization of the MASS model forecast products in the operational arena. Over the past several months, the development and utilization of the HP workstations has proceeded through the appointment of a staff meteorologist (Ron Humble) to work with the Science Officer in the capacity of a SAC focal point. The change has helped to close the HP systems gap between Raleigh and some of the more progressive NWS offices.

D) The types of problems encountered from a NWS perspective have changed little over the course of several years of ongoing collaborations. Moreover, there continues to be a limitation of time and resources. The workload of the NWS collaborative participants is tremendous. NWS forecasters are asked to perform professional development activities while carrying out the rigors of rotating shift work. The past 18 months has been especially demanding for Raleigh forecasters who have worked five federally declared weather disastrous events. Our collaborative success is a testimonial to their professional integrity. Budget cutbacks affected the acquisition of needed hardware and software that would enhance our utilization of collaborative findings in operations. The lack of funding for participating in the sharing of scientific findings at conferences and workshops was most regrettable.