Close Window

North Carolina State Univ.: "Selective mesoscale analysis, research, and training (SMART) project"

Final Report

1. PROJECT OBJECTIVES AND ACCOMPLISHMENTS

1.1 Introduction

COMET-funded joint activities between North Carolina State University (NCSU) and the WFO Raleigh (RAH) have continued to produce tangible and significant impacts on the local forecast process over the past three years. The general goals of our COMET Outreach Project SMART (Selective Mesoscale Analysis, Research, and Training) Project are highly focused upon the most critical forecast problems in the southeastern United States. SMART builds upon our past accomplishments, exploits recent advances in data access and display systems, places a strong emphasis on mesoscale education and training programs, and aims to streamline the forecast process. This report discusses the accomplishments gained, benefits realized, and problems encountered in this project.

The five SMART task objectives as originally proposed were to:

The participants from NCSU and WFO RAH who have been involved in the research and forecast implementation activities in this project during this 3-year project are listed below:

Roles of NCSU Participants

Dr. Steven Koch (Principal NCSU Investigator):

Jeremy Moore:

Steve Saleeby:

Leanne Siedlarz:

Ruth Aiken:

Phil Badgett:

Rod Gonski:

Steve Harned:

Gail Hartfield:

Ron Humble:

Richard Jones:

George Lemons:

Mike Moneypenny:

Rick Neuherz:

Other persons not formally a part of the SMART project contributed to the accomplishment of the stated objectives:

  1. Dr. Ken Waight of MESO, Inc. freely contributed his time to improving the real-time MASS model during the first year of this project.

  2. Hugh Cobb of NWS-Wakefield worked with Dr. Koch on the development of a gravity wave on-line tutorial described below.

  3. Larry Lee of NWS-Greer served on the NCSU-NWS Workshop Program Committee.

1.2 Overview of Objectives and Accomplishments

Our experiences have shown that the forecaster's understanding of research findings from this collaboration progresses through various levels before a new technique can be implemented into NWS operations. The path to transferring research 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 for 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 presented in section 3.

Task T1

Complete publication of papers ensuing from the collaborative activities between NCSU and NWS RAH in our past (and current) COMET Cooperative Projects.

Mesoscale Gravity Wave Climatology

Our climatology study of mesoscale gravity waves was published (Koch and Siedlarz 1999). Pressure pulses displaying periods of 1-6 h were identified in mesonet surface pressure traces on 34% of the days throughout the 6-week data period. We discovered that gravity wave crests tend to be aligned with the cloud/rainbands. As convective entities mature into well-organized, mesoscale systems, the gravity waves strengthen. The events displaying the largest amplitudes occur in an environment in which a jet streak is approaching an inflection axis in the height field downstream from an upper-level trough, and primarily north of a surface warm front. These results suggest that this simple conceptual model for operational forecasting of significant gravity waves should prove useful, as suggested by Koch and O'Handley (1997).

Detection of Cold Fronts Aloft

Our structured methodology for detecting the presence of split fronts and cold fronts aloft (CFAs) in an operational forecast environment has been conditionally accepted for publication (Koch 2000). The advice and comments of Peter Hobbs, John Locatelli, and Mark Stoelinga of the University of Washington were solicited and accommodated in the text. This methodology was applied in the paper to a case in which a split front passed over a region of cold air damming in the southeastern U. S. A real-time mesoscale model and various products from the WSR-88D - including the Velocity-Azimuth Display Wind Profile (VWP) and hodograph products, plus a thermal advection retrieval scheme applied to the VWP data- were used to study this split front and an associated convective rainband.

Wet bulb temperature and vertical motion forecasts at 700 hPa from the model revealed the arc-shaped split front 300 - 500 km ahead of the surface cold front. As this mid-tropospheric front passed across the surface warm front and entered the cold air-damming region, model vertical cross section analyses showed that it created a deep elevated layer of potential instability. Furthermore, an ageostrophic transverse circulation associated with the split front provided the lifting mechanism for releasing this instability as deep convection. Analysis of the absolute geostrophic momentum field provided greater understanding of the structure of the split front and a deep tropospheric frontal system to its west that connected with the surface cold front.

An "S-inverted S" pattern in the zero isodop on WSR-88D radial velocity displays indicative of wind backing above wind veering indicated the presence of the split front in the observations (as did the hodographs). Detection of the passage of the split front could be discerned from temporal changes in the vertical profile of the winds, namely by the appearance of mid-level backing of the winds in VWP time-height displays. Because of the subtlety of this backing and the need to be more quantitative, a temperature advection retrieval scheme using VWP data was developed. The complex evolving structure of the split front was revealed with this technique. Results from this retrieval method were judged to be meteorologically meaningful, to exhibit excellent time-space continuity, and to compare reasonably well with the frontal structures evident in the mesoscale model forecasts. The thermal advection scheme can easily be made to function in operations, as long as there is real-time access to Level II radar data.

Task T2

Use mesoscale models to address critical forecast problems related to primarily topographically forced phenomena, such as shallow frontal boundaries and cold fronts aloft. Case studies will be carried out in conjunction with task T3. The goal is to develop flowcharts and other forecast methodologies to streamline the forecast process using the models and WSR-88D radar data as guidance.

Neither the RAH forecast office nor NCSU had reliable real-time access to the full resolution (32 km) MesoEta model data throughout this entire 3-year project, with the exception that during the last few months NCSU was able to acquire this data through the IDD from UNIDATA. However, the Raleigh office still cannot get this model data even on AWIPS! Hence, we continued to place strong reliance upon experimental mesoscale models run locally at NCSU. The MesoEta model has been used for some of the most recent studies and development efforts as described elsewhere in this report (namely, the 22 January 1999 CFA case and recent gravity wave events).

Early Use of the MASS Model

As described in our previous COMET Outreach project final report, we were one of the first universities to have a real-time mesoscale modeling capability. Through the able assistance provided by Dr. Ken Waight of MESO, Inc., we developed the ability to run the MASS model on a workstation in Dr. Koch's lab. We used MASS for some initial studies of cold air damming resulting in a completed M. S. thesis by Devin Kramer. During the first year of the SMART project, we attempted to maintain this model without any COMET funding specifically directed at that effort. During that time, we were able to make the MASS model forecast fields available in a much shorter time, because we shifted the GEMPAK file conversion to an "on-the-fly" operation, instead of waiting until the entire model run was completed. The result was that MASS fields became available for the early morning briefing as NTRANS files on the SAC computer. Researchers at NCSU constructed additional diagnostic fields from MASS model forecasts in response to the requests made by Gail Hartfield and Ed Delgado of WFO RAH. New fields added to the suite of products were primarily concerned with the cold air damming forecast problem (Q-vectors, partial layer thickness, probabilities of frozen precipitation, and Froude number meteograms). Delgado trained forecasters to use these tailor-made fields on NTRANS/GARP in a structured manner.

Development of Real-time MM5 Mesoscale Model Capability

As we gradually reduced our dependency upon the MASS model for this project, primarily because of lack of support to maintain the model, Dr. Koch sought out new opportunities to develop a real-time mesoscale modeling capability. Funding from Capitol Broadcasting Corporation and Cray/SGI, Inc. has allowed us to develop a real-time MM5 model on a dedicated SGI Origin 200 multi-processor workstation at the North Carolina Supercomputing Center (NCSC). The MM5 mesoscale model has been run in real-time at NCSC since early 1999. The Input/Output Applications Programming Interface (I/OAPI), developed at NCSC, was adapted for use in this project to enable model products to be produced as quickly as possible, minimizing the time the forecaster must wait for the numerical guidance to be available. I/OAPI converts the NetCDF formatted data output by the model into GEMPAK format as the model is being run. MM5 model hourly output is converted to GEMPAK format in a minute fraction of the time than has previously been possible. This capability should be of interest to other universities that are running MM5 models under COMET Outreach support.

The GEMPAK products are created for two purposes: 1) to produce images for our MM5 web site, and 2) to produce model grid files for viewing in GARP by students and researchers at NCSU. In conceiving the design of our MM5 web page, we first conducted an extensive survey of existing real-time mesoscale model web pages to evaluate which layouts and individual products would provide the greatest information and functionality to the forecaster. In addition to the traditional forecast products and table-based layouts (forecast time vs. pressure level/forecast product), we have been developing several "weather forecast paradigms", representing a real innovation in model-based weather web sites. These paradigms utilize a decision-tree approach to each forecast problem, as the forecaster is led through a series of posted questions to answer based upon the results of his/her input to earlier questions. We have nearly completed paradigms for Cold Air Damming and Precipitation Type to streamline the process of forecasting winter weather. Our MM5 model web page address is http://envpro.ncsc.org/projects/SECMEP.

Use of Mesoscale Models and Observations in Studies of Split Fronts and Cold Fronts Aloft

Our model-based studies of Cold Fronts Aloft (CFAs) indicate that convective destabilization caused by the advection of low equivalent potential temperature air at mid-levels behind the leading edge of the CFA and the vertical circulation transverse to the CFA are instrumental in generating severe weather over the southeastern U. S. Actually, the cases that we have examined in detail lend more credence to the notion of a Split Front (SF) than a CFA, since a surface cold front was present behind the mid-level front, rather than a "dry trough" indicated in CFA studies. We have been using the MASS, MM5, and MesoEta models, plus an abundance of observations including VWP wind profiles from the WSR-88D, wind profiler data, enhanced GOES satellite imagery, and NOWRAD mosaics to study split fronts that occurred on 19 December 1995 and 22 January 1999. In both cases, the split front triggered a convective rainband that maintained its integrity even as it propagated over a region of cold air damming; also, the mesoscale models predicted strong mid-level cold advection hundreds of kilometers ahead of the surface cold front. Differential thermal advection destabilized the atmosphere and helped to establish the conditions under which strong convection erupted. Vertical cross sections of absolute momentum gradient fields and equivalent potential temperature most clearly distinguished the split fronts in the model forecast fields. These fields and others (e.g., transverse ageostrophic circulation and relative humidity fields) are shown for the 19 December 1995 case on our web-based CFA/SF tutorial found at: http://www2.ncsu.edu/eos/info/mea/mea715_info/CFA_Tutorial.htm

The convection in the 22 January 1999 case was particularly notable, since it caused the largest tornado event ever recorded in the month of January in this part of the country. Jamie Mitchem has been studying this case for his M.S. degree research topic. The 19 December 1995 case was used in the NWS-NCSU Workshop on Cold Fronts Aloft, Gravity Waves, and Mesocyclones in the Southeastern U. S. to instruct forecasters about the dynamics and structure of CFAs and SFs and operational techniques for detecting them. A paper written on this case has been conditionally accepted for publication in Weather and Forecasting (Koch 2000). In summary, a highly structured process has been developed for detecting and predicting SFs and CFAs in mesoscale model fields and WSR-88D products:

1. The first step is to anticipate the general situation in which a CFA/SF is most likely to occur and to use the model forecasts as guidance to identify its signature in terms of the fields mentioned above. In particular, the forecaster should first search for evidence in the 600-700 mb fields of model-forecast vertical motion and wet bulb temperature of a forcing mechanism for a precipitation band observed well ahead of the surface front.

2. A series of vertical cross sections should then be constructed perpendicular to the suspected CFA/SF to examine isentropes of potential and equivalent potential temperature, transverse ageostrophic wind vectors, thermal advection, relative humidity, potential instability, and absolute momentum products. Dr. Koch created GARP macros for this purpose, made them available on the SOO contrib. directory, and used them in teaching COMAP99.

3. Finally, the forecaster should examine radial velocity and reflectivity fields from mid-elevation angle WSR-88D data, plus the derived VAD wind profiles and wind hodograph products for any signs that a CFA/SF has actually been detected by radar (see Task T3).

Use of Mesoscale Models in Studies of Gravity Waves

We have designed GARP macros to enable us to quickly determine whether the environment is conducive to generating significant, coherent gravity wave activity (Koch and O'Handley 1997). We also have developed a method to detect the waves themselves using bandpass-filtered forecast fields produced by the MesoEta model and our real-time MM5 model. Steve Saleeby is using this model as a part of his M.S. thesis. In general, these studies reveal that mesoscale models are not very reliable at predicting all the characteristics of specific observed waves, but they are quite helpful when it comes to predicting where and when waves will likely form.

Task T3

Continue development of the thermal retrieval technique from WSR-88D wind profiles with the goal of real-time application. Apply technique to studies of cold fronts aloft and shallow frontal boundaries of unknown origin, and synthesize the results with the mesoscale model forecast fields.

This task involves further testing and improvement of our technique for computing geostrophic temperature advection from the vertical wind shear obtained from the Velocity-Azimuth Display Wind Profile (VWP) product. This simple technique was applied earlier to the 19 December 1995 case and is currently being used in the 22 January 1999 case with modifications intended to make the product available in real-time to weather forecasters in Raleigh.

The 19 December 1995 Case

We have shown how WATADS and WSR-88D PUP displays can be profitably used to reveal the presence of split fronts and cold fronts aloft. In particular, an "S-backwards S" appearance in the zero isodop on a mid-level elevation scan indicates the presence of geostrophic warm advection at low levels and cold advection aloft. We also found the VAD wind hodograph product to be helpful. It is better to pick out a CFA or SF by observing backing winds in the mid-troposphere on plots of horizontal wind vectors in the VAD Wind Profile display, for then one can sense that a significant change has occurred (the hodograph and radial velocity fields produce only instantaneous displays). However, these signatures can be rather subtle, so it is for this purpose that we have developed a thermal retrieval scheme to highlight regions of strong geostrophic cold or warm advection with color-filled contours. The method utilizes the thermal wind approximation as applied to the VWP wind shears; thus, it must be used cautiously in regions of strong ageostrophy, such as in the friction layer.

We have shown that the thermal retrieval technique applied to the VWP data produces patterns of thermal advection that compare well with the mesoscale model forecasts (Koch 2000). Comparison between the VWP time-height cross-sections of wind vectors from the individual radar sites and the model winds in spatial cross sections was also very encouraging. However, we found that in the local vicinity of the split front, a significant ageostrophic wind field indicative of a transverse frontal circulation was present in the model-forecast fields. Although this represents a limitation to the thermal retrieval in a quantitative sense, the ability of the technique to distinguish the presence of the thermal advection fields is still of good qualitative value. Furthermore, the retrieved geostrophic thermal advection fields are much easier for a meteorologist to discern the presence of a CFA/SF than attempting to see backing of the winds with height, which is often quite subtle.

As mentioned above, this case formed the basis for one of the labs in the NWS-NCSU Workshop on Cold Fronts Aloft, Gravity Waves, and Mesocyclones in the Southeastern U. S. The CFA lab continues to act as a station drill for forecasters at the Blacksburg, Virginia office; the online tutorial on CFAs is a prerequisite. Mention of the existence of Cold Fronts Aloft has appeared in forecast discussions from several offices in the Mid-Atlantic and Southeastern regions.

Real-time Thermal Retrieval System and the 22 January 1999 Case

Level II radar data were acquired and have been partially processed for this second case. This case was also used as a test set to develop a real-time thermal retrieval analysis procedure. The RIDDS Level II Doppler radar data archival system was installed at NWSFO Raleigh in late 1997, and WSR-88D base data has been flowing to North Carolina State University in real-time since early 1998. We asked Dr. Tim Crum at the Operational Support Facility to persuade Michael Eilts and Kevin Thomas from the National Severe Storms Lab to develop a real-time display and analysis system to function with the Level II data from our RIDDS. We have been very fortunate to have a real-time version of this system (RT-WATADS), since without this capability, it would have been extremely difficult to produce such derived products as the VAD Wind Profile in real-time needed to conduct the thermal retrievals. We successfully tested this real-time CFA/SF detection system last month on a split front case. Output from the thermal retrieval scheme can now be made available to forecasters via e-mail notification with a link to a new web page dedicated to showing these products, as well as MesoEta model products directed to this problem.

While the above results demonstrating useful information about thermal advection are exciting, additional work is needed to improve on our method for estimating the temperature (or potential temperature) fields from the thermal gradients at the radar sites. A quantitative estimate of the influence of friction and other ageostrophic effects is needed, as is a way of constraining the analysis so that vertical consistency is built into the results.

Task T4

Work towards real-time application of the gravity wave prediction and detection scheme. Access to 5-min ASOS data and to digital mesoscale model forecast fields is required, as well as the porting of our existing software to the SAC.

Our earlier COMET research showed that real-time prediction, detection, and nowcasting of mesoscale gravity waves is feasible, due to recent major advances in operational observing and modeling systems. Koch and O'Handley (1997), hereafter KOH97, proposed a simple flowchart-based methodology designed to forecast and detect mesoscale gravity waves using Automated Surface Observation System (ASOS) and operationally available mesoscale models. The suggestions put forth in that paper have been successfully implemented by a graduate student (Steve Saleeby) as task T4 under the SMART effort.

A three-phase breakdown of the gravity wave mesoanalysis project was formulated and approved by the ASOS Project Office and the Eastern Division of the NWS. The three phases consist of the following:

1. During Phase I, we used only archived ASOS data from the National Climate Data Center (NCDC). The archived data were useful for learning how to reformat ASOS data so that it could be ingested into our software, and for development of the quality control, bandpass filtering, and objective analysis routines.

2. We began work on Phase II toward making the analysis system work in pseudo-realtime soon after the ASOS access authorization was granted. We performed many tests of the objective analysis system, and examined the results for maximum coherence and agreement with observed weather features. We also explored whether similar features were forecast by the MASS, MM5, and/or MesoEta models.

3. Our ultimate goal in Phase III was to superimpose the mesoanalyses directly onto digital radar and satellite imagery, and to enlist the aid of RAH forecasters in the use and assessment of this powerful new product. Products were produced as GIF images of pressure and wind perturbation fields on our home page, and Java animations were created. We have come a long ways toward meeting even this ambitious goal.

Gravity Wave Prediction

1. According to KOH97, gravity wave occurrences strongly favor a particular synoptic-scale flow pattern, consisting of a jet streak that is propagating away from the upper-level trough axis and towards the inflection axis in the height field. Gravity waves develop near the inflection axis and they propagate toward the downstream ridge axis and to the north of a low-level warm or stationary front. Though wave formation, shape, amplitude, and propagation direction vary between cases, this conceptual model remains a common link among case studies (Uccellini and Koch 1987).

2. Upon diagnosis of a favorable wave environment, it is then necessary to determine if unbalanced flow exists in relation to upper-level jet streak interactions. The two primary model-forecast indicators of unbalanced flow are the cross-stream Rossby number and the non-linear balance equation. Both indicators are easily examined from model forecast fields as system macros using GEMPAK and GARP.

3. Assuming that unbalanced flow is computed from the model fields, it is important to then determine if a wave duct is present. From a forecast sounding or cross section of potential temperature near the anticipated wave corridor, the forecaster looks for the presence of low level static stability and upper level conditional instability. As a further indicator of relative stability, KOH97 proposed the use of a wave-ducting factor that can be calculated from model layer data. If all the above criteria are met, then the real-time gravity wave analysis system is activated, as described below.

4. We have completed development of special procedures to make it easier for forecasters to detect the presence of gravity waves in complex mesoscale model forecast fields. In particular, a bandpass technique based upon a two-pass, dissimilar weight function version of the Barnes objective analysis scheme is applied to produce perturbation fields of pressure and the wave-normal wind component.

5. GARP macros for forecasting gravity waves were introduced to forecasters in COMAP99 and are also available by way of the SOO contributors directory. Unfortunately, the decision by the NWS to rely entirely in the future on AWIPS instead of N-AWIPS means that this script can no longer be used. New procedures must be programmed into AWIPS to accomplish what worked so well under GARP.

Implementation of the Automated Gravity Wave Detection System

For the purposes of studying mesoscale gravity waves and other mesoscale disturbances in near-real-time and in an operational forecast setting, North Carolina State University has developed the first automated gravity wave detection and analysis system. These tools have been constructed to demonstrate to the forecaster that real-time mesoanalysis is both possible and vital to nowcasting of convective weather events. Our methodology for gravity wave analysis presumes that 5-min ASOS data can be readily acquired from many sites in real time, analyzed, and displayed in an automated fashion. Several months were spent attempting to obtain permission from the NWS Systems Operation Center for real-time access. With the help of Steve Harned and NWS officials, authorization was finally granted in October 1997. The NWS RAH office then purchased serial modem dialup software (Procomm Plus with WinBatch) and a 33.6-Kb modem for use on a new PC in our lab (acquired from other funding) for the purpose of dialing into the ASOS datastreams.

1. The real-time system is a complex, multi-step process that makes use of the ASOS high-resolution 5-minute data retrievable by modem access. The data are then processed through a series of programs consisting of quality control, bandpass filtering, and objective analysis of the data. The final product consists of contoured fields of pressure and wave-normal wind perturbations, temperature, observed winds, and altimeter pressure that are available via the Internet as either single plots or animations. The results of the analysis system appear on our web page within a few hours after the data is downloaded at http://www4.ncsu.edu/unity/users/s/smsaleeb/www/gwave.html.

2. North Carolina State University has been granted limited real-time access to the ASOS database at certain times of the day in 3-hour segments. It takes approximately two minutes per station to attain and store the data via modem dial-up to each station. Due to these time constraints and the limited (12 hour) archive of 5-min data, a maximum of 9 - 10 hours of the most current data are requested, and no more than 75-80 stations are contacted per download cycle. Thus, it takes more than 2 hours to completely download one batch of stations to do an objective analysis. The area over which access has been granted to us extends from the Rocky Mountains, southward to the Gulf Coast, eastward to the Atlantic Ocean, and as far north as the southern Great Lakes. These time constraints demand that several overlapping sub-regions be created within the main mesoscale-network for data downloading. These sub-regions also provide the domains for the objective analyses.

3. The data analysis process is initiated upon completion of the completed download, as the station data files are automatically transferred to a Unix workstation for processing. The entire automated analysis process, including web page image creation, requires approximately 30 minutes of processing time before the results appear on the web site. The first step in the data processing is to extract information from each 5-min data string, and place the data in a GEMPAK formatted file for easier manipulation and management.

4. Next, the data are run through a series of data quality control scripts to attempt to detect and eliminate potentially erroneous data embedded within the time series data stream. The first check involves a search for isolated five-minute discontinuities in the data stream and replacement of these glitches with null values. The second check is to ensure that all data fall within a preset range of tolerable values; any data point outside of the established boundaries is given the null value. The third step involves a sequence of 3-point Laplacian functions to detect unrealistic slopes in the sequential time series data for the purpose of detecting extremely rapid changes in the data values. Another data check searches for 5 or 10-min spikes in the time series values that return back to the previous base state. Blips beyond ten minutes are assumed to be potentially real anomalies and not an error caused by hardware induced fluctuations. Finally, a Lagrangian cubic interpolation is performed to obtain the interpolated value of remaining missing points for time series gaps of up to 10 minutes (2 data points).

5. Prior to performing the objective analysis, the quality-controlled pressure and wave-normal wind data are subjected to bandpass filtering in order to obtain perturbation fields needed to clearly isolate the gravity waves. Waves with periods of 0.5 - 4.2 hours pass through the filter without appreciable loss of amplitude (the filter response curve is displayed on the web site). The filtered fields are useful for revealing the covariance between the wind and pressure perturbations, which is needed to confirm the existence of gravity waves.

6. The traditional Barnes scheme (as well as other schemes) cannot resolve waves of scales smaller than 2-4 Dn (the average station spacing). However, a time-to-space conversion (TSC) adaptation of this technique allows for analysis of these waves, as demonstrated by KOH97. The TSC technique requires the use of a representative advection vector to convert the off-time observations into spatial data (with an accordingly smaller weight than the on-time data). For this purpose, the mass-weighted mean wind vector in the conditionally unstable layer above the stably stratified duct layer is computed for each rawinsonde station within the region that will be objectively analyzed. According to the KOH97 methodology, this vector acts as an adequate surrogate for the winds at the critical level (where the gravity wave propagation vector matches the environmental winds). This vector also determines the surface wind component in the direction of wave propagation and the manner in which off-time observations are converted from time to spatial data in the objective analysis.

7. A series of tests were performed to determine the most appropriate values for the variables in the TSC-Barnes scheme. The results of these tests are displayed on the web page. Two of the tests concerned the values for the station spacing and the numerical convergence parameter, which together determine the amount of resolvable detail in the objective analysis. Another test examined the most appropriate value to use for the temporal weight parameter in the TSC scheme, which also controls the amount of spatial overlap of the off-time observations. The most important tests performed dealt with the issue of the accuracy and number of advection vectors needed to obtain the best TSC analysis. Using no vectors essentially digresses to the traditional Barnes scheme, resulting in mainly small, isolated circular perturbations in the analyses. Direct comparison between the TSC analysis and the traditional analysis gives visual evidence of the power of TSC to better resolve mesoscale features and maintain those as they propagate through the analysis region (see the web page). The waves produced from TSC are also better correlated to the associated satellite and radar observations. The mesoscale structure becomes appreciably more coherent with an increase in the number of available vectors. A single vector produces unrealistically linear perturbation features, whereas the use of multiple vectors offers the ability to reveal the arc-shaped nature of most waves and produces better wave coherence in space and time. It is fortunate that the noticeable improvement begins to decrease rapidly for more than four vectors, since there are typically no more than 4-6 available sounding stations per download region.

8. Whenever and wherever a suitable gravity wave environment is predicted to occur, the ASOS pseudo-realtime gravity wave analysis and prediction system is activated. We have set up an e-mail alias of recipients whom we believe are interested in this gravity wave analysis system, and we notify them when we plan to activate the system. We recently added an archive capability on the web site, so that historical cases that we have analyzed can be called up by anyone and reviewed.

9. We have periodically reacquainted NWS Raleigh staff with the gravity wave analysis and prediction methodology and shared the real-time analysis results with them in debriefings. This information has occasionally led to its inclusion in the State Forecast Discussion. For example, one such mention was made on 19 February 1999 with regards to our diagnosis of gravity wave activity in the MM5 and MASS models across the southeastern portion of North Carolina, and the effects that this would most likely have upon convection. The most recent briefing occurred for the 17-18 February 2000 wave event, which appears on the web page above. The most recent gravity wave analysis occurred on 11 March 2000, and successfully analyzed strong waves over the Ohio Valley, just as had been predicted to occur the day before by Dr. Koch by using the KOH97 conceptual model. In fact, in all the real-time cases analyzed, it has been possible to predict quite well the exact wave corridor a day or more in advance!

Our ultimate goal under this task is to be able to compare surface mesoanalyses of pressure perturbations to that forecast by mesoscale models, which produce gravity waves with uncertain reliability. In so doing, we seek to understand how generally important these phenomena are to the organization of mesoscale precipitation systems, and hence, to helping to address the Quantitative Precipitation Forecast problem, in the southeastern and Mid-Atlantic regions. This forms the topic of the M. S. thesis for Steve Saleeby.

Task T5

Develop progressive meteorological fields, seminars, and operational proficiency tutorials with "hand on components" to accelerate and fully incorporate the results of applied forecasting research into operations at the RAH office. Topics will emphasize the dynamics of, and techniques for detecting, shallow boundaries, gravity waves, cold fronts aloft and other topographically forced mesoscale phenomena of concern to regional forecasting.

We believe that educational materials designed to ensure that research findings are well understood and applied effectively in the operational arena should accompany the initial implementation of research results into the NWS forecast office. Accordingly, throughout the course of the SMART project we have developed notebooks and operational proficiency guides, created online tutorials, and held joint seminars.

NWS-NCSU Workshop on Cold Fronts Aloft, Gravity Waves, and Mesocyclones

The joint NWS-NCSU Workshop on Cold Fronts Aloft, Gravity Waves, and Mesocyclones in the Southeastern U. S. was held on the NCSU campus on 11-12 August 1998. Approximately 30 forecasters from the southeastern region plus other representatives of NWS Eastern Region, the Storm Prediction Center, and the Office of Meteorology attended. The Office of Meteorology financially supported this workshop. Liz Page at COMET and NCSU provided computer support for the lab exercises. The workshop was not just a series of lectures, which is a sure guarantee that what was learned would not be utilized on the forecast floor. Rather, this workshop included computer-based, hands-on exercises designed to allow the attendees to work through actual cases in a simulated operational setting using the methods for detection and prediction learned in lecture. In addition, we developed online tutorials and Training and Evaluation Modules for the Workshop, and made it easy for SOOs to download all of the lab materials used in the workshop for retraining purposes back at their offices after the end of the workshop.

Lectures were presented on these topics:

Detailed labs were created to allow forecasters to practice their newfound understanding of the concepts taught in the lecture. In addition, the lab instructions, datasets and scripts (all of which utilize GARP and WATADS) have been made available to the attendees. Dr. Koch and Jeremy Moore of NCSU worked with Gail Hartfield at WFO Raleigh on the macros, which are available from the SOO/SAC Home Page at COMET in the contrib/fdf/ directory of the SAC FTP server. The gravity wave lab datasets are also easily obtainable from COMET case study #012 (the Palm Sunday gravity wave event) available at the COMET source http://www.comet.ucar.edu/resources/cases/c12_27mar94/support.htm.

The cold fronts aloft lab material is available by ftp as tar files as shown:

ftp meaculpa.nrrc.ncsu.edu

login: nws

password: ftp4nws

cd /users/nws

Since that time, the Storm Prediction Center has repeatedly used the concepts and methodology for the analysis of Cold Fronts Aloft that were taught in the workshop. The Raleigh forecast office has utilized many of the recommendations made in the workshop that were based upon our research studies on mesocyclone detection using the WSR-88D. Gravity waves appearing in mesoscale model forecast fields have also been mentioned in State Forecast Discussions issued by the Raleigh forecast office, and in other discussions from the Wakefield forecast office.

Educational and training guides produced under this project

The tutorials include a "hands on" component involving scripts written for mesoscale model output and macros so as to help the forecaster better understand the application of progressive meteorological fields to the region's most critical local forecast problems. These education and training guides are saved for periodical review and future reference. The tutorials and other training materials developed under this project are listed below:

http://www4.ncsu.edu/unity/users/s/smsaleeb/www/gwave.html
http://www.nws.noaa.gov/er/akq/Gwave.htm

ftp meaculpa.nrrc.ncsu.edu

http://www2.ncsu.edu/eos/info/mea/mea715_info/CFA_Tutorial.htm

http:// www2.ncsu.edu/classes/mea444-sekoch/

2. SUMMARY OF UNIVERSITY / NWS EXCHANGES

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. These activities include a joint seminar series, a strong outreach program to the private sector and other NWS offices, training/education modules, and other things.

Seminars Program Seminars were primarily given in the first year of the SMART project, as the NCSU-NWS Workshop and COMAP99 sapped most of our education and training energies in 1998 and 1999, respectively. Following is a compilation of the joint seminars that have been offered by the collaborators in our joint seminar series:

Outreach to the Private Sector and other NWS Offices

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. We have provided valuable familiarization and training for local forecasters on the region's most critical forecast problems. The results of joint 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. We continue to share and refine the winter storm forecast process enhanced through COMET support. The techniques are well understood by the regional media and used in their forecast process, resulting in far better forecast coordination for the general public.

An important part of our outreach strategy is to hold formal training in the form of workshops and symposia. The joint NCSU-NWS Workshop described in detail above was the premier example of these workshops. In addition, a Forecaster's Training Workshop was conducted by the RAH office at the Greenville-Spartanburg (GSP) office. A Training Symposium for Operational Forecasters in the Carolinas and Virginia was also held in Raleigh. Kermit Keeter traveled to the Storm Prediction Center in October 1997 to provide training and familiarization on the winter storm forecast process for the Carolinas and Virginia. Most recently, the Raleigh forecast office sponsored a "Seasonal Winter Weather Forecast Workshop" (in October 1999) that was well attended by the regional weather media.

As a result of these outreach efforts, several NWS offices and a few meteorologists in the private sector are now understanding and employing COMET sponsored ideas and forecast tools. These include: (a) a classification scheme for cold air damming and damming look-alikes, (b) a classification scheme for low-level summertime boundary types detectable in WSR-88D imagery, (c) objective schemes for predicting the predominate winter precipitation types, and (d) frontal conceptual models invaluable for predicting the areal distribution and intensity of precipitation.

Winter Precipitation Forecasting in BUFKIT

A second example can be provided of activities not specifically funded by the COMET Outreach Program, but which includes interactions undertaken in the spirit of improving relations between the academic and operational forecasting communities. The SMART collaboration has made winter precipitation type forecast tools (not just information) available to other regional NWS offices, and several of those offices are now using these tools. Just recently, our winter weather forecast schemes have been added to BUFKIT, a sounding visualization software package that is widely used by NWS forecast offices. BUFKIT will make the collaboration's forecast tools far more accessible and efficiently used by the nation's weather services.

Dr. Koch: Co-Lead Instructor for COMAP99

The sabbatical by Dr. Koch as co-lead instructor of COMAP99 allowed him the opportunity to communicate through lectures and hands-on labs to Science Operations Officers the new findings, understandings, conceptual models, and tools for the diagnostic analysis of observational data and mesoscale model fields arising from our SMART efforts. Very favorable responses from the SOOs were received, particularly concerning the topics of frontal conceptual models (such as anafronts, split fronts and cold fronts aloft) and mesoscale gravity waves. Kermit Keeter and Gail Hartfield gave invited presentations to the class featuring the "cold-air damming spectrum", precipitation type forecasting, and the "5-step collaborative process".

Training and Evaluation Module

The Cold-Air Damming (CAD) Spectrum, a conceptual model developed under SMART, was published as a National Weather Service Eastern Region Training and Evaluation Module and distributed to all NWS Forecast Offices in the Eastern United States. The module was very well received. Its principal ideas have been frequently referred to in forecast discussions issued by NWS offices, improving the forecast coordination of these events.

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).

Koch, S. E., and C. O'Handley, 1997: Operational forecasting and detection of mesoscale gravity waves. Wea. and Forecasting, 12, 253-281.

Koch, S. E., and L. M. Siedlarz, 1999: Mesoscale gravity waves and their environment in the central U. S. during STORM-FEST. Mon. Wea. Rev., 127, 2854-2879.

Koch, S. E., 2000: Real-time detection of cold fronts aloft and split fronts using mesoscale models and WSR-88D radar products. Wea. and Forecasting (cond. accept.).

M.S. Theses completed from COMET support (chronological order)

Moore, J., 1997: Studies of Cold Fronts Aloft using the WSR-88D velocity profiles and a mesoscale model. Honors Research Project, North Carolina State University.

Vandersip, C., 1998: A single-Doppler radar study of kinematic and structural characteristics of mesocyclones in the Southeastern and Great Plains regions of the United States. M. S. Thesis, North Carolina State University, 168 pp.

Saleeby, S., 2000: Development and implementation of an automated system for mesoscale gravity wave analysis. M. S. Thesis, North Carolina State University (in preparation).

Conference papers (chronological order)

Koch, S. E., 1998: Mesoanalysis and modeling of the forcing for mesoconvective systems in the Palm Sunday tornado outbreak. Preprints, 12th Conference on Numerical Weather Prediction and 16th Conference on Weather Analysis and Forecasting, Phoenix, AZ, Amer. Meteor. Soc., J91-J94.

Kramer, D. P., and S. E. Koch, 1998: Realtime mesoscale model evaluation during Appalachian Cold Air Damming. Preprints, 16th Conference on Weather Analysis and Forecasting, Phoenix, AZ, Amer. Meteor. Soc., 341-343.

Vandersip, C., and S. E. Koch, 1998: Characteristics of mesocyclones in the Southeast and Southern Plains regions of the U. S.: A comparative study. Preprints, 16th Conference on Weather Analysis and Forecasting, Phoenix, AZ, Amer. Meteor. Soc., 207-209.

Other presentations (chronological order)

Aiken, R., and J. Moore, 1997: An unusual large hail event in eastern North Carolina 21 April 1997: Using the 88-D to investigate possible gravity waves and cold fronts aloft plus an evaluation of the hail detection algorithm's performance. 22nd Annual Meeting of the National Weather Association, Las Vegas, NV.

Hartfield, G., K. Keeter, and S. Koch, 1997: Designing tutorials as part of a professional development program. Annual Meeting of the National Weather Association, Las Vegas, NV.

Koch, S. E., and K. Keeter, 1997: Forecast applications research using the WSR-88D at North Carolina State University and NWS-Raleigh. NEXRAD Users Workshop, Asheville, NC.

Koch, S. E., 1997: The use of conceptual models in the forecast process for frontal precipitation events. Presentations at NWS-Raleigh and NWS-Greenville-Spartanburg WSFOs.

Koch, S. E., 1997: COMET Mesoscale Analysis and Prediction COMAP 97-2 Course, Boulder, CO. Koch, S. E., 1998: Cold fronts aloft. Joint NWS-NCSU Workshop on Mesoscale Meteorology, 11-12 August 1998, Raleigh, NC.

Koch, S. E., 1998: Forecasting and real-time mesoanalysis of gravity waves. Joint NWS-NCSU Workshop on Mesoscale Meteorology, 11-12 August 1998, Raleigh, NC.

Koch, S. E., and K. Keeter, 1998: Moving research findings into forecasting practice at NWS-Raleigh. National Weather Association Annual Meeting, Oklahoma City, OK.

Vandersip, C., 1998: Mesocyclones in the Southeastern U. S.: Findings and operational implications of a comparative WSR-88D study. Joint NWS-NCSU Workshop on Mesoscale Meteorology, 11-12 August 1998, Raleigh, NC. Hartfield, G., 1999: Cold Air Damming: An Introduction. Eastern Region Training and Evaluation Module No. 4, 16 pp.

Hartfield, G., 1999: The Spectrum of Cold Air Damming and Damming Look-Alikes. Forecaster's Training Workshop at the NWS GSP Office, Biloxi, MS.

Mitchem, J. D., and S. E. Koch, 1999: Mesoscale model analysis of a squall line associated with a cold front aloft. National Weather Association Annual Meeting, Biloxi, MS. Saleeby, S. and S. E. Koch, 1999: A pseudo real-time gravity wave analysis system. National Weather Association Annual Meeting, Biloxi, MS.

4. SUMMARY OF BENEFITS AND PROBLEMS ENCOUNTERED

4.1. Benefits to North Carolina State University

Benefits to NCSU faculty and students arising from this project have been numerous and enjoyable. This collaboration has provided assistance to the researchers needing data and in the development of analysis macros. University courses emphasizing mesoscale meteorology have directly benefited from the collaboration. The SMART project has enabled NCSU to attract additional funding for related research and development projects. Dr. Koch was the recipient of awards recognizing him for his achievements in operational forecasting research work. Finally, the university has enjoyed the popularity of several media events arising as the result of the accomplishments achieved in this program related to the detection and prediction of severe storms.

Providing Assistance in Obtaining Data and Developing Analysis Macros

Over the course of the past three years, a close working relationship has developed between NCSU researchers and NWS RAH forecasters. This relationship has not occurred just because of the collocation, though that has been essential, 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. The RAH office has made it possible for us to obtain needed WSR-88D data by setting up user functions on the PUP to dial up the needed data and assisting us in obtaining the RIDDS real-time radar ingest system. Steve Harned's acquisition of the software and the modem necessary to access ASOS data in realtime was a necessary step toward our goal of operational surface mesoanalysis. GARP scripts written by Ron Humble and Rick Neuherz helped enable the forecasters to use the MASS mesoscale model forecast fields on the SAC running NTRANS. Other macros were created jointly with NWS personnel in support of the NCSU-NWS Workshop. Training provided to Dr. Koch by Gail Hartfield and Rod Gonski on the PC GRIDS macros helped him to design the MM5 mesoscale model Web page to be an effective forecast tool.

Mesometeorological Education

We have actively sought to educate students as well as forecasters on mesoscale meteorology throughout the course of this project. We envisioned this education and training as a two-way process between researchers and forecasters. For example, the macros and mesoscale model data sets have become incorporated into the requirements for completing a graduate-level course on the Dynamics of Mesoscale Precipitation Systems taught by Dr. Koch (MEA 715), and are also utilized in his undergraduate-level synoptic weather forecasting course (MEA 444). In addition, forecasters feed back information to the researchers at NCSU about the performance of the experimental mesoscale models, interesting weather events that suggest the possible influence of various mesoscale phenomena, and the forecast/nowcast process. Furthermore, the time spent by Dr. Koch on helping to develop the various web pages for this collaboration has benefited his teaching of these concepts to university students.

Attraction of New Funding Sources

The North Carolina Supercomputing Center (NCSC) and Capitol Broadcasting, Inc. (WRAL-TV) approached Dr. Koch about development of a real-time MM5 modeling capability in early 1998. Both groups were impressed with our demonstrated ability to successfully develop the MASS model and MM5 web page and use them for NWS mesometeorological training purposes under COMET funding. The NCSC connection led to two separate, small grants - one from Cray/SGI, Inc. ($8K) and another from the U. S. Environmental Protection Agency ($10K), the latter for developing a real-time, coupled air chemistry-mesoscale model in conjunction with NCSC. A larger ($40K) one-year grant was awarded to NCSU and NCSC by Capitol Broadcasting, Inc. in 1999 to complete the development of the MM5 model and to design online forecast products specifically suited to their needs. We have just heard that an even larger grant from this company will be given to NCSU and NCSC to continue work on this model, perform a model evaluation, and produce on-air weather graphics.

Research Achievement Award

Our demonstrated success in conducting operational forecasting research and having those research results implemented in the operational arena has not gone unnoticed by the broader forecasting community. Dr. Koch was recognized for his achievements in this area by being the 1998 recipient of the Research Achievement Award from the National Weather Association.

Community attention to research on tornado detection

The NCSU News Services broadcast a major news release in 1998 about the collaborative research being conducted by NCSU researchers and NWS Raleigh forecasters on the development of new tools and training programs intended to aid in the identification of tornado-producing mesocyclones. Some of our other research activities were also mentioned in that news release, including the use of experimental mesoscale models and the Doppler radar thermal retrieval technique. Not only was this announced in the North Carolina State University Bulletin, but also numerous television and other news media picked up on this announcement.

4.2. Benefits to the National Weather Service Raleigh Office

From the perspective of NWS RAH, COMET sponsored collaborations have enhanced the profession of operational forecasters. Our forecasters now see their jobs as having a technique development and applied research component that goes beyond the practice of forecasting. They have come to feel comfortable with the partnership between themselves, faculty, and students recognizing their NWS operational perspective brings added value to the collaborations. Because of our extensive collaborations sponsored by COMET, NWS RAH attracts operational forecasters who are eager to participate in collaborations. This has also enhanced the forecast office's image with the media, the private sector, as well as the public. COMET has inspired nothing short of a cultural change within the NWS - a change where forecasters have expanded their role from just the practice of forecasting to include applied research focused on forecast technique improvements.

Unit Citation Award

The National Weather Service Forecast Office in Raleigh received a Unit Citation Award from NOAA in 1997 "in recognition of outstanding applied research activities leading to significant forecast improvements". This is the first time that NOAA has presented a field forecast office with an award for applied research, and in itself testifies to the intensity and success of the COMET-funded activities at NCSU and NWS RAH. Of particular mention are applications employing modernized technologies such as the Doppler radar and mesoscale numerical models, for which NCSU has played a key role in guiding the NWS staff toward understanding and utilizing new and progressive concepts and datasets.

Greater understanding of mesoscale models and WSR-88D capabilities

Through COMET-sponsored activities including the RAH-NCSU Lecture Series, the NCSU-RAH Workshop held in 1998, training tutorials, forecast macros, publications, and N-AWIPS 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. A few examples include the roles of: jet streak circulations, tropopause folds / potential vorticity, cold fronts aloft, gravity waves, and cold air damming in the region's weather events. Likewise, NCSU efforts with mesoscale models have allowed the forecasters to view and work with non-traditional meteorological fields in the operational arena. Collaborative case studies have also demonstrated the utility of the WSR-88D radar to show signatures of the presence of possible gravity waves and/or cold fronts aloft. Knowledgeable students and faculty have brought insights into the operational arena as they investigate research topics pertinent to operations using the radar.

Streamlining the forecast process

With so much data available for interrogation, NCSU and NWS RAH collaborators have emphasized the need to streamline the forecast process. For a specific forecast problem, a tutorial, operational proficiency guide, a guest lecture, a "paper-of-the-month", a PC GRIDS Macro (or GARP macro or AWIPS procedure) addressing the problem, and perhaps a case study, would all be coordinated to make for efficient use of the newly learned topic within the forecast process. In so doing, these components worked together to achieve the project goals.

Utilization of Conceptual Models

Prior to the SMART project, forecasters at the NWS office in Raleigh had little understanding of and exposure to conceptual models relating precipitation patterns to frontal structures, frontal circulations, and jet circulations. Dr. Koch has shown forecasters at the Raleigh NC, Greer SC, Morristown TN, and Columbia SC NWS offices how low-level frontal systems and upper-level jets interact to produce specific precipitation patterns and affect the onset, duration, and intensity of mesoscale precipitation events. Dr. Koch's case studies have provided an operational perspective to NWS forecasters. Some NWS forecasters at Raleigh have been trying to utilize these concepts in the short-term forecast and nowcast mission. Work continues on designing forecast macros that will guide the forecaster through the recommended meteorological fields important to ascertaining the potential effects of frontal structure, circulations, and interactions with upper jets on precipitation patterns.

Real-time mesoscale forecasting and warnings

Dr. Koch and his students in the NCSU Mesoscale Analysis and Forecasting Lab have recently begun issuing mesoscale gravity wave and split front alerts and making them available to WFO RAH in real-time through e-mail and direct contact with the lead forecaster. The e-mail contains a synopsis text message with a link to a web page, which shows graphical images of meteorological fields relevant to the evolution of the detected and/or predicted event. These alerts are assisting the WFO RAH in demonstrating the added value to the forecast process of accounting for these important meteorological features (see below the"5-Step Collaborative Process"). Currently, the forecast office is completing the development of AWIPS procedures for "split front" detection and for accounting for the synoptic environment conducive to the generation of significant gravity waves.

Winter Storm Forecast Process Reference Guide

Through our collaborative outreach activities, media forecasters now have a better understanding of Raleigh's forecast techniques. This has in turn reduced the degree of variability and errors in winter weather forecasts made available to the public through and by the media. The Winter Storm Forecast Process Reference Guide has been made available to other nearby forecast offices, the NWS Training Center, and the Storm Prediction Center as a reference source for periodical review by forecasters to help ensure that the techniques are understood and used correctly. Custom designed winter storm forecast products produced from the MASS model have been excellent tools for demonstrating the forecast utility of these techniques to the media and other NWS forecast offices.

Experimental Convective Outlooks

The availability of GARP scripts using the MASS model and the results of our earlier research on low-level convergence boundaries laid the foundation by the Raleigh forecast office for production of experimental convective outlooks regarding the areal coverage, timing, and type of convective storms expected in the station's county warning area. 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. Recognition of the nature of shallow boundaries in North Carolina has improved the station's mesoscale surface analyses. The various sub-categories of boundary types are being recognized, analyzed, and tracked from one forecaster to another with good continuity. The tendency to misidentify boundary types as synoptic-scale fronts has decreased.

Meteorological Tutorials

Dr. Koch served as a valuable resource for the development of the progressive meteorological field tutorials - "Beyond AFOS Towards AWIPS". His insight into the physical interpretation of potential vorticity was prominent in our latest tutorial - "Understanding the Tropopause and Potential Vorticity". Without Dr. Koch's participation in the review of and refinement to the tutorials, the RAH forecasters would be less prepared to use progressive meteorological fields in the forecast process. A forecast macro/script has been produced to encourage forecasters to evaluate potential vorticity fields and their possible effects upon the local weather patterns.

CSTAR Project Successfully Awarded to NCSU and NWS-Raleigh

Finally and quite importantly, the COMET Outreach program laid a foundation for NWS-NCSU collaborations allowing us to be in a position to obtain other research grants. The MM5 model funding by NCSC and Capitol Broadcasting, Inc. was mentioned earlier. Of even greater benefit not just to the WFO RAH office, but also to other offices in the region (Greer, Morehead City, Wilmington, and Wakefield) has been the awarding to NCSU of a large grant ($150K) from NOAA under the CSTAR program. This grant will fund the additional study of cold-air damming, shallow frontal boundaries (including the "coastal" and "wedge fronts), and split fronts. Unlike previous University-NWS collaborative endeavors, the CSTAR project directly involves multiple NWS forecast offices working jointly with NCSU. This project spins up in May 2000.

Clearly, the SMART project has been an immense help to the Raleigh forecast office in a multitude of ways. Additional specifics are provided below, bearing in mind the working philosophy of the NWS-NCSU collaborative process as each of these benefits is discussed, i.e., we hold to a 5-step process whereby new research findings ultimately are used operationally:

Periodical review for the staff is at least a seasonal necessity, while there is always a need for further refinement and improvement of the forecast/analysis methodology.

Mesocyclones in the Southeastern U.S.

Status: Level 5 (Periodical Review)

This study of mesocyclones by Chris Vandersip for his M.S. thesis research is a perfect example of findings from applied research being implemented into forecast operations in an effective manner. Characteristics of mesocyclones in the Southeast United States were found to be statistically different from their Plains counterparts (Vandersip 1998). These findings supported forecaster's subjective observations of the "mini-supercell" and low-topped convective nature of severe storms common to the Southeast. Moreover, mesocyclones in the Southeast are smaller in scale and spin up more rapidly than their cousins in the Plains states. The study also compared mesocyclone characteristics for tornadic and non-tornadic storms. Changes in mesocyclones indicative of tornadoes occurred in the lowest 20% of the storm's volume; tornadic storms showed significant azimuthal shear in both the low and mid levels, while non-tornadic storms indicated shear only in the mid levels. The study also found that the WSR-88D typically underestimates mesocyclone strengths by roughly 18, 23, and 33% at distances from the RDA of 100, 150, and 200 km. Finally, many mesocyclones as seen in the storm relative motion (SRM) products were found to be associated with asymmetrical velocity couplets characterized by a strong inbound (outbound) signal coupled with a weak outbound (inbound) signal.

The operational implications of these findings were presented to NWS-RAH forecasters by Vandersip and Koch in May 1998 and given to the station's SOO, WCM, and Warnings Team for follow-up action. These implications resulted in significant modifications to the radar surveillance strategies and warning decision practices employed by NWS-RAH. Specifically, the 2.0nm mesocyclone diameter nomogram is now typically used to better capture the strength of mesocyclones. Secondly, volume scan VCP 11 is now typically employed during severe weather instead of VCP 21. Thirdly, a "quick view" user function has been developed enabling the staff to view the low and mid level elevation slices as soon as possible. Fourth, maps showing the underestimation of mesocyclone characteristics at far range distances were made - at these distances, forecasters have been instructed to rely more on ground truth reports, storm history, and when in doubt about the potential of a possible tornado, to take the path of least regret.

Cold Fronts Aloft (CFA)

Status: Level 3 (Implementation)

Forecast scripts for detecting and predicting CFAs were given to all attending the NCSU-NWS Workshop in August 1998. Using instructional materials from the workshop and our on-line CFA tutorial, the attendees returned to their respective offices to instruct others on detecting and forecasting CFAs. To date, the Storm Prediction Center in Norman OK has successfully used the scripts to project storms in their Convective Outlook products. The NWS forecast office at Columbia, SC was able to identify in real-time a CFA case causing showers in their forecast area. The Blacksburg VA office recently identified an unusual cold front aloft occurring under northwesterly flow. In addition, workshop attendees from the NWS forecast office in Blacksburg VA have been able to repeatedly present the CFA lab portion of the workshop to their forecasters by downloading the lab data sets provided by NCSU.

Recommended procedures for the detection of CFAs include tailored diagnoses of numerical model fields using GARP scripts designed by the SMART team and the use of WSR-88D imagery (see Task T2). When viewed within the context of a CFA conceptual model presented in the SMART tutorials, these fields can be used to evaluate the numerical model's predictions of CFA evolution, and thereby assist in the proper interpretation of the mesoscale forcing for prefrontal bands of precipitation and severe weather. Currently, these macros are being rewritten at NWS RAH as procedures within AWIPS.

Mesoscale Gravity Waves (MGW)

Status: Level 2 Level 3 (Demonstrate Added Value Implementation)

These procedures show promise; however, forecast preparation time in operations is taxed due to the huge volume of data sets available for evaluation. The gravity wave prediction paradigm developed by NCSU researchers is critical to "alert" forecasters to a situation that is conducive to gravity wave development of significance to weather forecasting. Also critical for monitoring the waves once an "alert" has been made is the accessibility of real-time ASOS mesoanalyses if such a detection system is to be effectively utilized operationally (the current 4-5 hour delay is not extremely helpful for use in nowcasting).

The SMART team has developed GARP scripts that identify the synoptic-scale environment in which mesoscale gravity waves occur. Other scripts have been created to evaluate the presence of "unbalanced" flow and ducting conducive for gravity waves. Researchers at NCSU have demonstrated to NWS forecasters the ability to identify gravity waves using bandpass filtering techniques with ASOS data, as discussed above. It is now imperative that this capability be evaluated in terms of its added benefit to the forecast process. This is the third and final stage of our originally proposed real-time gravity wave prediction & analysis system.

Connections: Cold-air damming, split fronts, & mesoscale gravity waves

Status: Level 1 (Discover and Share)

Case studies conducted by Dr. Koch have pointed to the interaction of split fronts (SF) and mesoscale gravity waves (MGW) with cold-air damming (CAD) events. Moreover, rainbands generated by split fronts were shown to maintain their integrity while gliding across the cold air dome. Other case studies by Dr. Koch show the high static stability seen in CAD events to be a favorable ducting mechanism for the propagation of gravity waves. While the CAD Spectrum has done much to help forecasters understand and anticipate the evolution of damming, it does not account for the influence of split fronts and gravity waves on the evolution of CAD events or their impact on the associated sensible weather (e.g. enhanced rainfall, severe storms). The results of these and other similar case studies should allow for an expansion of the Spectrum model to include the interaction of CAD events with these other important meteorological features. Currently, NWS forecasters give little if any attention to the possible influence of these meteorological features during CAD episodes. The expansion of the Spectrum will help to correct this deficiency. We are happy to see that the CSTAR funding will permit this to happen.

4.3. Problems Encountered by North Carolina State University

The staffing issue resulting from the change in policy about 3 years ago regarding the Air Force Institute of Technology (AFIT) program continues to have negative consequences for our COMET Outreach projects. We used to rely in part upon AFIT students to conduct some of the research proposed to COMET. Although COMET was able to reinstate the funding for the UCAR/COMET Graduate Student Fellowship program two years ago, our proposal failed to be funded, in large part because of the intense competition for very limited funding. Thus, graduate student manpower is a continuing problem for this project.

The real-time gravity wave project will have only limited success for real-time nowcasting due to the communications limitations of the ASOS Project Office. Having access to less than 100 stations only 2-3 times daily, and using slow modem dialups that require 3 h to download this data means that we will always be faced with only pseudo-realtime mesoanalyses unless other methods can be devised for obtaining the data. Also, wavelet analysis holds promise as a more efficient method than bandpass filter analysis, since the latter method necessitates the loss of data at both ends of the time series.

4.4. Problems Encountered by NWS Raleigh Office

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 personnel, yet our staffing levels have been equivalent to other WFOs not engaged in extensive collaborations with a university. Making the mission of collaborations even more difficult was the disproportionate forecast workload (relative to new offices) that the established forecast offices carried until just recently. Now that the forecast workload is distributed evenly, all NWS offices are seeing an increasing amount of time that continues to be required for the implementation of the hardware and software supporting the NWS modernization. Moreover, during the course of this project, Raleigh forecasters have worked seven federally declared weather disastrous events, yet have been expected to participate in these collaborations. Our success is a testimonial to their professionalism and dedication to the project.

AWIPS was delivered to the field offices as an incomplete system in need of additional development. For over a year, all field offices have been engrossed in customizing AWIPS for their local areas and implementing key AWIPS components. The amount of time and work needed from field offices to prepare AWIPS for becoming the NWS operating system has been both surprising and disappointing. It has been difficult to understand the lack of documentation on AWIPS development. Many NWS endeavors, not just Raleigh's COMET collaborations, have been slowed and hindered by the work demands associated with AWIPS. While AWIPS provides many additional operational capabilities, it does not have some of the specific capabilities found with supplemental operating systems such as GARP scripts on LINUX based computers. This is especially true for the detection and prediction of split fronts. In general, forecasters have migrated away from the supplemental systems in favor of near total dedication to using AWIPS. This has slowed the implementation of split front forecasting techniques into the operational practices employed at the WFO RAH.