Close Window

Saint Louis Univ.: "Mesoscale convective storms and quantitative precipitation forecasting across the middle Mississippi and Ohio River valleys"

Final Report

1.1 Introduction: Summary of overall project objectives and a description of project participants.

This project involves cooperative research among four NWS forecast offices, St. Louis, MO; Louisville, KY; Paducah, KY; and Slidell, LA and Saint Louis University. There are two separate research objectives of the project. One team of researchers dealt with the quantitative precipitation forecast (QPF) and Quantitative Precipitation Estimation (QPE) problems within the mid-Mississippi  and Ohio River Valleys. The main objective of this team was to determine those parameters and techniques that help to accurately forecast the areal extent and magnitude of significant (>2 inches/24 hours) precipitation events on a regional scale, evaluate numerical models' QPF, as well as existing QPF methods, and develop a regional climatology of warm and cool season heavy precipitation events for both the Mississippi and Ohio River Valleys. The QPF team includes the following people:

Saint Louis University:

Dr. James Moore
Mr. Patrick Market (finished Ph.D., July 1999)
Mr. Scott Watson (finished M.S., February 2000)
Mr. Stephen Klaus (finished M.S., February 2000)
Mr. John Gagan (began research May 1999)
Ms. Heather Sumner (began research September 1999)

Louisville, KY WSFO:

Mr. Ted Funk
Mr. Robert Cox
Mr. Chad Swain
Mr. Stephen Marien

St. Louis, MO WSFO:

Mr. Fred Glass

Paducah, KY NEXRAD WSO: 

Mr. Michael York
Mr. James Keysor

Slidell, LA WSFO:

Mr. Alan Johnson

The university participants' role was to archive datasets of significant precipitation events (>2 inches/24 hours) in the mid-Mississippi and Ohio River valleys, including observed data (surface, upper air,  profiler, raingage data, satellite imagery, and gridded model data), to create a regional climatology of  heavy precipitation events, distribute our SLUBREW diagnostic software to each NWS office, write FORTRAN code to assist in parameter development, investigate numerous case studies of warm and cold season precipitation events, and research the precipitation efficiency problem. They also worked on compositing the different synoptic types contributing to heavy precipitation. The role of the WSFO St. Louis participants was to create a heavy rain and heavy snow climatology for Missouri, while collaborating on individual case studies. The Paducah team  focused on the evaluation of model performance during heavy precipitation events. The Slidell team has developed an automated approach to estimating precipitation amounts using gridded model data. They helped the university participants to adapt their techniques to both the mid-Mississippi and Ohio river valleys. The Louisville team compiled a climatology of cold season precipitation events (snow > 4 inches 24 hours-1) in the Ohio River Valley while also diagnosing individual events in detail. Although these were the core tasks assigned to each group, there was considerable overlap in tasks among the project participants. Also, the groups have collaborated to write presentations and preprints for conferences, and write papers submitted to journals.

Our other research team studied mesoscale convective systems (MCSs) that produce damaging straight-line surface winds tornadoes across the mid-Mississippi and Ohio River valleys. The overall objectives of this research were to investigate the structure, evolution, and sensible effects of rear inflow jets (RIJs), "bookend vortices", tornadic and non-tornadic mesocyclones, mid-altitude radial convergence (MARC) signatures, and multicellular versus supercellular structures associated with linear and bow-shaped MCSs. Members of this research team include the following people:

Saint Louis University:

Dr. Yeong-jer Lin
Mr. Jason Martinelli (began research May 1999)
Mr. James O'Sullivan

St. Louis, MO WSFO:

Mr. Ron Przybylinski
Mr. Gary Schmocker
Mr. Tom Spriggs (new member)

Louisville, KY WSFO:

Mr. Ted Funk
Mr. Van DeWald


Mr. Pat Spoden
Mr. Jeff Hovis

Each of the MCS teams investigated case studies pertaining to their areas of interest. The university researchers worked on case studies of severe MCSs in Missouri and Illinois. Ron Przybylinski and Dr. Lin have worked with each of the graduate students to diagnose the pre-storm environment as well as study WSR-88D imagery for key signatures pertaining to strong straight-line wind or tornado development. The Louisville and Paducah teams studied severe MCSs that went through Kentucky. The NWS offices' job was to collect WSR-88D in archive level II format while the university team archived surface, upper air, satellite and profiler data for each interesting case. The SLUBREW software was also used at the Louisville and St. Louis offices to assist in diagnosing the pre-storm environment. WATADS software from NSSL was used by university and NWS researchers to analyze WSR-88D archive level II data for each case. Collaboration was extensive on preprint articles and papers submitted to journals.

1.2  Description of research/development activities and accomplishments carried out to date.

QPF Team

The research team at Saint Louis University concentrated on completing the significant rainfall climatology for Kentucky and southern Indiana and evaluating the Eta-32 model quantitative precipitation forecasts (QPFs) for significant rainfall events over the Mid-Mississippi and Ohio River Valleys. The former work was predominately done by Mr. Stephen Klaus, while Mr. Scott Watson worked on the latter research. They both completed the work and their M.S. degrees during the course of the COMET research. They presented their research at the Missouri Academy of Science annual meeting, the National Weather Association's annual meeting, and at the COMET Regional Workshop (see section 3).

As a result of Steve Klaus' and Scott Watson's research we now have compiled many cases from 1992 to present on which to do research. Scott assembled over 40 cases since February 1998 in which two inches or more of liquid precipitation fell in 24 h, as documented from the River Forecast Center's (RFCs) network of cooperative rain gages. We also have captured from the web the 12-36 hour Eta-32 forecasts for these cases (based upon 00 UTC runs of the model) for comparison. His work has helped us to better understand the deficiencies in model QPF so that WFOs may better use the Eta as guidance. We have sent copies of Scott's thesis to the St. Louis WSFO and Steve's thesis to the Louisville WSFO for their information. As noted in section 3, their powerpoint presentations have been place on our COMET web page for other WFOs to download.

Fred Glass and Jim Moore have stepped up their efforts to complete the compositing for the 21 elevated thunderstorm cases in order to begin writing a journal article for Weather and Forecasting. A rough draft of this paper should be completed by 30 June 2000 and submission is scheduled for late July 2000.

We have also prepared an abstract on the 3 January 2000 heavy rain event in Kentucky for the AMS Severe Local Storms Conference to be held in Orlando, FL in September 2000 and the Missouri Academy of Science annual meeting in Columbia, MO which was held on 15 April 2000 (see section 3). These projects represent "overlap" projects between the end of the COMET-sponsored research and the beginning of the NOAA-sponsored research associated with the Cooperative Institute for Precipitation Systems (CIPS) which began this May.

The SLU QPF group provided upper air data for the Louisville QPF group to use in creating maps for their heavy snowfall climatology for Kentucky. They are used the SLUBREW software to create maps and composites for each case. We have helped them in compositing these cases, much as we did for the elevated thunderstorm cases.

We have also provided many maps from SLUBREW for the Paducah, KY QPF team to analyze for the three case studies that they have been examining. Unfortunately, they cannot run the SLUBREW software and cannot generate the maps themselves (see section 4.1 for details). SLU graduate students have generated the maps that we have sent to Paducah.

Jim Moore has spent a great deal of time re-reviewing the paper "Flash Flood Forecasting and Detection Methods", which will be a chapter in the AMS monograph on severe local storms.

From WSFO Louisville:

Louisville project members (Ted Funk, Rob Cox, Chad Swain, and Steve Marien) have focused their attention on developing statistics and a climatology of heavy snow events (4 inches or more in 24 h) across Kentucky and southern Indiana. Approximately 40 heavy snow events have been collected during the period 1982-1996 for study. We have also printed out snow plot maps for all events on the SAC workstation. These maps currently are being contoured and analyzed. The goal is to determine various snow statistics, including snowfall occurrence versus state climatological regions, annual and monthly distributions, normalized statistics, etc.

We have downloaded numerous data files from SLU's ftp server in support of our COMET heavy snow climatology work. Using SLUBREW software, many maps were composed and printed for a number of heavy snow events for the period 1982-1996 across southern Indiana and Kentucky. From these maps, we made numerous environmental composites of several cases, a task that is ongoing. Composites include surface, upper air, frontogenesis, isentropic, θe, and stability. We have noticed two major environmental patterns conducive to heavy snowfall in the Ohio Valley.

The first pattern consists of broad southwest mid-level flow with an embedded southern stream shortwave, and relatively weak surface low development. In this pattern, isentropic lift, low-to-mid-level frontogenesis, and jet streak dynamics play vital roles. A northern and southern stream jet often are present. The southern jet helps focus low-level northward moisture transport and warm advection into the Ohio Valley (including southerly 850 mb winds and an overall east-west thermal gradient). The right entrance region of the northern jet appears to be instrumental in helping keep cold air in place in low-levels to prevent or delay a changeover to rain, despite southerly warm advection. In addition, frontogenetical forcing and isentropic lift are enhanced within the pronounced right entrance region, thereby contributing to strong lift and heavy snowfall production. The 17 January 1994 snowstorm, which produced 1-2 feet of snow across north-central Kentucky in less than 12 hours, was a classic example of the Type I pattern.

The second type features deep surface and upper air low development where heavy snow usually falls within a comma head structure on satellite imagery, a pattern most frequent across eastern Kentucky. In this pattern, frontogenetical forcing and isentropic lift occur within the comma head producing a zone of heavy snowfall. Low-level winds typically are directed more from an east-to-west direction with a north-to-south oriented thermal gradient. The 13-14 March 1993 "Storm of the Century" was a prime example of Type II events, which produced 1-2 feet of snow in eastern Kentucky. We plan to continue our investigation of important processes and their interrelationship in generating heavy snowfall in the Ohio Valley.

From our composite map events, we selected four significant snow cases for oral presentation at the November 1999 COMET Workshop at SLU, two events per major heavy snow type. Rob Cox and Chad Swain made the presentation, which also included Ohio Valley heavy snow statistical information. Results of these events have been posted on the Science and Technology section of our Web site. This spring, we plan to submit these results as an abstract to the NWA Annual Meeting in October 2000 in Silver Spring, Maryland.

In November/December 1999, Ted Funk developed numerous winter weather procedures in AWIPS to help forecasters 1) interpret model data, 2) assess various atmospheric processes important to the production of heavy cool season precipitation, and 3) determine precipitation type. These processes are important to heavy snow production in the Ohio Valley, a topic of research in our COMET work. In March 2000, Ted began developing AWIPS procedures for severe weather and heavy convective rainfall.

From WSFO St. Louis:

Fred Glass continued to work on his synoptic climatology of widespread heavy rainfall events (rainfall exceeding 4 inches/24 hours) for Missouri. The most recent work focused on cold season events (Nov-Mar) and involved case selection and segregating the cases based on similar conditions. Fred also gave a presentation on the climatology portion of his heavy rainfall study at a heavy rain forecasting workshop at the Springfield WFO. In preparation for the talk, he spent time enhancing and updating his talk in Corel Presentations. Over the next several months, additional work will be completed comparing the 50 year distribution of heavy rainfall events to El Nino/La Nina episodes. Fred also has finished up the work of Mr. Dave Metze who compiled a heavy snow (>4 inches/24 hours) climatology for Missouri before leaving WSFO St. Louis. The El Nino/La Nina statistics will also be done for the 50 year heavy snow climatology. These presentations will then be added to the SLU and WFO St. Louis internet home pages. Future plans are to create a detailed talk on Flash Flood Forecasting in both Corel Presentations and Microsoft Power Point.

Recently work has also begun studying storm scale and mesoscale aspects of flash flood events. Complete data sets are available for a handful of cases from recent years.

Fred also spent quite a bit of time reviewing several papers and a research proposal. The first paper was an on-going review of the "Flash Flood Forecasting and Detection Methods" chapter for the AMS monograph on severe local storms. The other was a paper involving forecasting nocturnal thunderstorms with heavy rain

From NWSO Paducah:

During the past six months Jim Keysor became actively involved  in the heavy precipitation portion of the COMET study. Jim and Mike York selected two convective heavy rainfall events that directly affected the Paducah County Warning Area (CWA). Jim selected a cold season event that occurred on 21 January 1999. Mike selected a warm season event that occurred on 4-5 June 1998. Using the methodology suggested by the graduate meteorology program at Saint Louis University, Mike and Jim examined several parameters associated with moisture, lift, and instability. They then compared the observed parameters with Eta model forecasts. A summary of their findings are appended to this report (see appendices A and B).

One of the key findings of the warm season evaluation was that mesoscale outflow boundaries became the driving force during later stages of the heavy rainfall event. The Eta model handled the initial stages of the event well, primarily because synoptic scale triggers  were dominant. As the event progressed, mesoscale outflow boundaries became the primary forcing mechanism, and Eta quantitative precipitation forecasts became less accurate.

The findings in our reports have yet to be shared with other  meteorologists. One of the main objectives over the next few months is to finalize the written draft of the report and verbally present our findings.

In addition, WFO Paducah will be taking these studies and applying the information to another local heavy rain event that occurred on 3 January  2000. This is planned to be a brief local study to see if the parameters studied  in the previous events could be used to locate the area of heavy rainfall on the 3 January event. We will also develop local heavy rainfall procedures on AWIPS for the staff to use based upon these studies.

Severe MCS Team

From WSFO St. Louis:

Over the past year, the MCS group at WSFO St. Louis continued work on a number of objectives including: 1) MARC climatology, 2) vortex climatology, 3) MCS - damaging winds climatology, 4) pre-convective environment climatology and 5) further investigate the spectrum of MCS evolution. The fifth objective keyed upon the study of isolated cell - convective line mergers and pre-existing outflow boundary - convective line interactions. Nine of the ten MCSs which occurred during the 1998 convective season have been examined. Ron Przybylinski and Gary Schmocker have completed two of the ten cases this spring. This work included the completion of the pre-convective environment and Doppler analyses of each case. Gary has worked on a third case and continues to work on the MARC climatology. The climatology will include the examination of 22 MCS cases. Since a number of detailed damage surveys were completed in 1998, we are looking closely at the relationship of MARC magnitudes to the degree of damaging winds at the surface. Nearly a third of the 1998 cases under investigation occurred either during the late night or early morning hours. These nocturnal convective cases have proven to be challenging as well as interesting with the MARC study. Tom Spriggs has studied a fourth MCS damaging winds case. Ron Przybylinski is assisting Tom with the Doppler analyses of this case.

The 29 June MCS has been studied by two groups. This convective system formed over northern Nebraska and moved east-southeastward through south-central Iowa, parts of far northeast and east-central Missouri and much of Illinois. Jay Martinelli (Grad Student at SLU) is working with Ron Przybylinski and Dr. Frank Lin on this case when the MCS moved through parts of Iowa and northeast Missouri. Mark Britt has worked with Ron and two forecasters at NWSFO Lincoln (ILX) covering a number of aspects when the convective system moved through much of Illinois. Jay has focused his work during the period of high-precipitation supercell - bow echo evolution while Mark and forecasters from Lincoln have keyed upon bow echo - circulation evolution over Illinois. Undergraduate students from St. Louis University have also worked on other MCS cases which occurred during the spring of 1998. One other MCS damaging winds case has been given to the Paducah Kentucky group for study. This particular case covered both parts of the St. Louis and Paducah County Warning Areas (CWA).

With the new data sets from 1998, we continued to test and revise the MARC and vortex evolution conceptual models. The new data sets are shedding new insights about the characteristics of MARC and vortex evolution under a number of pre-convective environment settings.    

We worked on a COMET WEB site at NWS St. Louis. The web site has links to the Saint Louis University COMET and CIPS home pages.

From WSFO Louisville:

The manuscript entitled, "Storm Reflectivity and Mesocyclone Evolution Associated with the 15 April 1994 Squall Line over Kentucky and Southern Indiana" was published as a journal article in the December 1999 issue of  Weather and Forecasting (Vol. 14, No. 6, pages 976-993). Reprints were ordered and sent to all co-authors and to CRH/SSD.

Ted Funk and Van DeWald spent much time developing a comprehensive hands-on workshop in Fall 1999 that we presented at the November 1999 COMET Workshop at SLU. The 14 May  1995 squall line and bow echo event featured very damaging winds, tornadoes, hail, mid-altitude convergence signatures, and flash flooding. Attendees were given data (including radar loops) to simulate real-time, from which they were required to identify and explain crucial storm signatures to make intelligent and timely warning decisions. Attendees felt the 2-hour workshop session provided excellent training. We posted the case in the Science and Technology section of NWS Louisville's Web site ( and submitted it to the SOO Resource Center so that other NWS offices could benefit from the training as well.

We continued to research the 20 April 1996 squall line and bow echo event. We also wrote an abstract based on this case entitled, "WSR-88D Reflectivity and Velocity Trends of a Damaging Bow Echo Event on 20 April 1996 over South-Central Indiana and Central Kentucky."  The abstract was submitted in March 2000 to the  20th Conference on Severe Local Storms to be held on 11-15 September 2000 in Orlando, Florida. Damaging winds, several F0-F2 tornadoes, embedded HP supercell structure, and squall line-outflow boundary interactions were noted in this case.

From NWSO Paducah:

Pat Spoden and Jeffrey Hovis studied the Derecho event from 17 May1999. They presented this at the COMET workshop at Saint Louis University in November 1999. The presentation utilized loops of satellite and radar data to reinforce the concept of looking at all of  the data, not just WSR‑88D data. This presentation is available on the internet at:

The title of the presentation is "Observations of the Damaging Wind Event of 17 May 1999 in the Lower Ohio Valley."

Pat Spoden and Mary Lamm presented a final presentation on the 5 May 1996 derecho at the COMET Workshop in November, 1999. It was titled "Observations of Flow Structure and Mesoscale Circulations Associated  with the 5 May 1999 Asymmetric Derecho in the Lower Ohio Valley". They presented initial findings at the first COMET Workshop in 1996. A  paper was also published (with the same title) in the Preprints of the 19th Conference on Severe Local Storms (AMS 1998). The presentation is available on the internet at:

Both of these presentations were offered to the staff at WFO Paducah before the COMET Workshop in November.    

COMET Regional Workshop

The COMET Research Workshop was held from 16-18 November 1999 in the Argentum Room of the Busch Memorial Center on the campus of Saint Louis University.   A copy of the agenda for the workshop is appended to this report. Approximately 75 people attended the workshop. The first day primarily focused on the QPF problem and was lead off by Mike Eckert's invited presentation entitled, "Quantitative Precipitation Forecasting for the Next Decade". This day also highlighted work by Fred Glass, Jim Moore, Rob Cox, Chad Swain, and graduate students Steve Klaus and Scott Watson. A presentation from Greg Gust of the NWSFO at Grand Forks, SD was also included on this day. Simultaneous QPF workshops were conducted by Jim Moore and Fred Glass. Jim's workshop focused on a cold season heavy rain event in Missouri while Fred's workshop focused on a fall convective rain/severe weather event in Kansas and Missouri.

Day two was highlighted by an invited talk by Dr. Mike Biggerstaff entitled, "Automatic Detection and Prediction of Damaging Straight-line Winds in Bow Echoes". Members of the severe MCS team including Gary Schmocker, Ron Przybylinski, Jeff Hovis, Pat Spoden Christopher Jones, James Keysor and Mary Lamm presented talks based upon their WSR-88D research on bow echoes. Invited NWS presentations were also given by Brian Klimowski, Bill Togstad, Jeff Evans (SPC), Gene Brusky, Ray Wolf, Alan Gerard and Karl Jungbluth. It was a full day of MCSs and WSR-88D imagery!

On the last day of the workshop two papers were presented by Jim O'Sullivan and Greg Gust on model forecasts of severe convection. Two simultaneous workshops were presented by Ted Funk and Ron Przybylinski on the operational forecasting of severe weather attending bow echoes. We finally concluded the conference at about 12:30 p.m. on 18 November.

A complete set of notes was given to all participants in a looseleaf binder. Selected talks were also put on our anonymous ftp site where they could be downloaded by participants to show to their fellow forecasters. In addition, additional "clean" copies of each workshop were made available for participants to use back home at their office. Since many of the attendees were SOOs, having the notes and workshop materials fits in with the "train the trainer" concept developed at COMET. Most of the presentations were place on our COMET web site, allowing people to download them.


A) University - NWS Exchanges / Benefits

Jim Moore and Pat Market visited the NWSFO in Tulsa, OK from 4-6 January 1999 at the request of Mr. Steve Amburn, the SOO at Tulsa.. Jim presented a seminar on the use of isentropic analysis techniques to forecast winter weather to the forecasters. Jim and Pat also spent a day working on isentropic macros for AWIPS for the staff to use in their daily forecast regime. They also presented a talk on the 10 April 1997 Surprise Snowstorm to the local AMS chapter in Tulsa.

Jim Moore presented an invited 1½ hour seminar on Quantitative Precipitation Forecasting at the Des Moines, IA National Weather Association 3rd annual Doppler Radar and Severe Local Storms Conference on 26 March 1999.            

Jim Moore presented an invited luncheon presentation titled, "A Short  History of Isentropic Analysis" to the 2nd Isentropic Conference held in Lancaster, Pennsylvania at Millersville University on 10 April 1999. Louis Uccellini, Ralph Petersen and several NWS forecasters from nearby forecast offices were in attendance at this University-sponsored conference.

Jim Moore presented a 1 ½  hour talk on "Summer Thunderstorm Safety" to the citizens of Glen Carbon, IL on 2 June 1999 at the request of the Chief of Police in Glen Carbon. This was Jim's first power point presentation.

Last December, Dr. Lin was invited by the National Science Council in Taiwan to attend the Second Workshop on Doppler Radar at the National Central University in Chung-Li, Taiwan. Professor Robert Houze from the University of Washington and Dr. Wen-Chau Lee from NCAR were also invited to participate in the workshop. Dr. Lin gave the invited talk entitled, "Structural Features of Downburst-producing Storms Determined from WSR-88D Data", at the workshop. The talk focused on the applicability and application of using the MCS's mid-altitude radial convergence (MARC) signature to "nowcasts" of a downburst-producing Midwestern storm. He utilized six case studies, conducted under the support of the current COMET cooperative project, to demonstrate the MARC signature for its potential as an indicator of the onset of damaging surface winds. Specifically, we found that the MARC velocity signature frequently extended from the edge of the multicell complex's right-front flank to the left flank. Intense reflectivity gradients along the storm's leading edge often coincided with the zone of strong radial convergence at middle levels of the MCS. The ribbon of outbound Doppler velocities signified the storm's updraft center. Upwind from the broad updraft center was a region of moderate to strong intensity inbound velocities. These case studies have shown a good correlation between values of MARC of at least 25-30 m s-1 to subsequent damaging surface wind gusts.

Dr. Lin's talk was well received by most of the 50 participants at the workshop. Professor Houze was also very interested in his presentation. He was particularly impressed by our approach which employs both theory and observed Doppler radar data to investigate the linkage between the MARC signature and damaging surface winds gusts. Professor Houze suggested at the workshop that such studies must continue in the future.

While in Taiwan, Dr. Lin also gave another invited talk at the Department of Atmospheric Sciences, National Taiwan University in Taipei, Taiwan. The title of his talk was "Preliminary Observational and Numerical Results of a midwestern MCS determined from Observed Data and a Mesoscale Numerical Model."  This study employed both the WSR-88D Doppler data and results from the MM5 model to investigate the structure and evolution of a tornado-producing storm during VORTEX-95. The study is somewhat related to the ongoing COMET project. We intend to pursue this research further in the future.

B) NWS Exchanges - Benefits 

From WSFO St. Louis:

Ron Przybylinski completed second reviews of the following two manuscripts:

Moller A. R., 1999: Severe Convective Storm Forecasting. (Meteorological Monograph).  

Gerard, A. and C. Mead, 1999: Evolution of a Mini-Bow Echo Occurring Behind a Prefrontal Squall Line. (Paper to be published in NWA Digest).

Saint Louis University student - WSFO St. Louis Visits / Exchanges

A  number of visits occurred during the past year between students at Saint Louis University and NWS staff members Ron Przybylinski and Gary Schmocker. The following is a summary of these visits:

1) Ron Przybylinski worked with SLU doctoral candidate Jim O'Sullivan on a number of COMET-related activities.  Ron assisted Jim with the following tasks:

a) Evaluation of the MM5 forecast output to observed and kinematic fields associated with the 11 February 1999 bow echo event. Also gave Jim an overview of the storm-reflectivity and Doppler velocity structures associated with this bow echo case.

b) Preparation of a MM5 presentation for the SOO Sub-regional meetings held in Kansas City Missouri in February 1999. Attendees at the meeting included: SOOs from 7 WFOs, Preston Leftwich (Regional SOO Coordinator - Central Region), FSL representative, Don Burgess and other OSF personnel and other NOAA attendees. Much of this work was completed at the university.

c) Writing portions of his manuscript for his doctoral prospectus (Ph.D. candidacy). Worked with Jim with the pre-convective environment and radar analysis sections of his perspectives.

d) Assisted Jim in the re-writing of two abstracts for the National Weather Association Annual Meetings.

2) Jay Martinelli (Grad Student - SLU)

a) Provided Jay a comprehensive listing of reference papers for background reading prior to working on his case.

b) Began work on Jay's outline for his case study.   Jay will focus on the 29 June 1998 HP storm - bow echo evolution across Iowa and parts of northeast Missouri.

3) MM5 Evaluation:

Ron Przybylinski has provided evaluations of MM5 output to Dr. Frank Lin, Dr. Chuck Graves, and Jim O'Sullivan from two severe weather cases this past spring. In these evaluations, a number of derived and base fields were surveyed and compared to current observations.

4) Student Volunteer Program (Saint Louis Univ - WSFO St. Louis)

Since the Fall of 1998, a Student Volunteer Program was re-established between WSFO St. Louis and the Meteorology Department at Saint Louis University. During the period from fall 1998 through spring 1999, three senior-level undergraduate students participated in the program, one during the fall term (Dan Valle; now a graduate student) and two during the spring term (Corey Chaskelson, and Marc Singer; both now graduate students). Since a number of MCS cases from the spring of 1998 still needed attention, each student was asked to complete the pre-convective environment analysis for their assigned cases. During the term, they would visit with Ron Przybylinski either at the NWS office or at Saint Louis University on a weekly or bi-weekly basis. Ben Able (Professor), and Jim Moore would also participate in this program providing guidance during the term. This program was quite successful, giving senior students an opportunity to work on COMET-related MCS cases with Ron and Gary Schmocker. Ron Przybylinski, also introduced the WSR-88D Doppler radar data to the cases they were completing. The introduction of Doppler data gave them an additional perspective about the case. In addition, Ron  provided reading materials and good discussions about MCS evolution/conceptual storm models  to the students. In all, the students greatly benefited greatly from this program.

Mark Singer (senior at Saint Louis Univ. in 1999) joined Ron Przybylinski on a damage survey over parts of southwest Illinois on 13 February 1999. The damage was associated with the 11 February 1999 tornado event. Mark was quite impressed with the damage caused by a non-supercell tornado and gained invaluable experience in understanding the characteristics of tornadic damage patterns. WSFO St. Louis continues to offer the opportunity to students at Saint Louis University to participate in such surveys.

On 28 April 1999; Ron Przybylinski gave a two hour presentation to three graduate students from Dr. Lin's Graduate Level Radar course. The presentation covered: a) characteristics of the WSR-88D; b) two severe weather case studies (27 February 1999 hail event and 8 April1999 tornado event). The students gained insights and a greater degree of understanding on supercell reflectivity/tornadic circulation evolution with the later case.

From WSFO Louisville:

QPF team:

Ted Funk held COMET project meetings at WSFO Louisville with members of the heavy precipitation group in October and November 1999 and in February 2000. In these meetings, we discussed current work progress, heavy snow parameter compositing issues, needed work over the next few months, necessary preparatory work for the November 1999 COMET Workshop in St. Louis (October and November meetings), and development of a conference paper based on the results to date (February meeting).

NWS Louisville conducted its annual NWS/Media Winter Weather Seminar on December 8, 1999 at Louisville. A total of 27 people attended, including 8 Louisville TV meteorologists (at least 1 from all 4 local stations), 5 (3) weather specialists from the Fort Campbell (Fort Knox) military reservation, and 11 WSFO Louisville staff members. As part of the seminar, Ted Funk discussed meteorological processes important to heavy precipitation production and their interrelationship. This discussion was based partly on findings from our COMET heavy snow climatological research across Kentucky and southern Indiana. Rob Cox presented preliminary COMET research results from this snow climatology, including snow statistical information and heavy snow case composites. Rob's talk provided enhanced pattern recognition skills to WSFO Louisville forecasters.

WSFO Louisville also conducted a Media Winter Weather Seminar on December 9, 1999 at WTVQ-TV Channel 36 in Lexington. A total of 8 people attended, including 6 Lexington TV meteorologists and 2 staff members from NWS Jackson, KY. Ted Funk gave a similar presentation as that on December 8 in Louisville. In addition, Chad Swain presented the same talk that Rob Cox had given in Louisville.

Ted Funk gave 3 identical 2-hour Staff Winter Weather Seminars to the WSFO Louisville staff on December 15, 17, and 22, 1999. An in-depth analysis and evaluation of various processes responsible for producing heavy precipitation was presented, including isentropic lift, jet streak dynamics, ageostrophic flow, frontogenesis, Q vectors, elevated instability, and their interrelationships. Results from our COMET snow research also were presented.

Severe MCS team

Ted Funk gave a severe weather training workshop at WSFO Paducah, KY on March 20, 2000. The 2.5-hour workshop included a hands-on exercise on the May 14, 1995 bow echo and tornado event in central Kentucky (the same workshop as presented at the COMET Workshop at SLU in November 1999). WSR-88D imagery from the January 3, 2000 severe squall line event, including the Owensboro, KY bow echo tornado, was shown and discussed as well. The seminar proved very useful to the Paducah staff in their efforts to prepare for the Spring 2000 severe weather season.

WSFO Louisville held its annual NWS/Media Severe Weather Seminar on March 29, 2000 at NWS Louisville. As part of the seminar, Ted Funk discussed convective storm structure and trends using animated WSR-88D radar imagery, and results of a heavy rainfall climatology across Kentucky and southern Indiana from 1982-1996. The storm structure presentation was based partly on our COMET severe weather research on squall lines and bow echoes. The heavy rainfall climatology was based on the Master's degree thesis work of Stephen Klaus, graduate student at Saint Louis University, which achieved a key COMET research goal.

On March 30, 2000, Ted Funk traveled to Lexington, KY to present a Media Severe Weather Seminar at WKYT-TV Channel 27 for the benefit of the Lexington media. Ted presented the same information as that discussed at the March 29 seminar in Louisville.

From NWSO Paducah:

WFO Paducah's increase in knowledge due to participation in the  COMET project has been enormous. One of the main areas where COMET has proved successful is in the training of tornadoes in bowing line segments. This training was applied to the Warrick County tornado of 19 April 1996. This helped us to identify (after the fact) the type of storm which produced the F2 tornado. We then applied the research into training for the staff and for emergency managers/media during the spring of 1997. We continue to see this "type" of tornado in our area during the cool season. This is now one of the main training topics each year.

The biggest payoff from the COMET Project was for the long lead time of the Owensboro, KY tornado of 3 January 2000. It is believed that this tornado (F3) was a product of a bowing line segment. The SOO was able to identify storm type and morphology during the warning  process due to the COMET information gleaned from research. Several members of the WFO Paducah team plan to look at this case in more detail. AWIPS has given us the ability to closely monitor storms close to our CWA. There have been at least two times within the past six months where we were able to identify bowing line segments (outside of our CWA) which subsequently produced tornadoes. One of these cases will be presented to the Paducah Staff this fall. It is believed that bowing line segment tornadoes are common in the Lower Ohio Valley.

It is expected that the participation of WFO Paducah in the QPF side of the project will soon produce many benefits for the staff. We  have already learned more about heavy rainfall due to the studies done by Fred Glass (WSFO STL) and Jim Moore (SLU). That information is used by the staff during the forecasting of heavy rain events. The information gained from the research done at WFO Paducah will assist us  in writing heavy rain procedures in AWIPS. We also plan in doing a brief study on the heavy rainfall from 3 January 2000.


M.S. Theses completed under the supervision of Dr. James Moore:

Stephen J. Klaus, 2000: The Creation of a Synoptic Climatology for Significant Rainfall Over Kentucky and Southern Indiana.

Scott A. Watson, 2000: Evaluation of the Eta-32 During Significant Rainfall Events.

M.S. Thesis completed under the supervision of Dr. Frank Lin:

Bill Hope, 1999: A Doppler Radar Comparison Study of Mesocyclones over Central Ohio on 26 July 1997.  

Ph.D dissertation completed under the supervision of Dr. James Moore:

Patrick Market, 1999: Jet Streak Modification via Diabatic Heating During Periods of Intense Cool Season Precipitation.

Presentations given at the annual meeting of the Missouri Academy of Science in Cape Girardeau, MO on 17 April 1999:

Klaus, S.J. and J.T. Moore, 1999: The Creation of a Synoptic Climatology of Heavy Rainfall over Kentucky and southern Indiana.

Watson, S.A. and  J.T. Moore, 1999: A Comparison of Quantitative Precipitation Forecasting Schemes.         

Presentations given at the annual National Weather Association Meeting  in Biloxi, MS on 18-22 October 1999:

Moore, J.T., P.S. Market, and S.M. Rochette, 1999: Comparing Adiabatic and Diabatic Vertical Motions in Isentropic Coordinates.

Market, P.S. and J.T. Moore, 1999: Jet Streak Backbuilding and the Influence of Latent Heating.

Klaus, S.J., J.T. Moore, and C.E. Graves, 1999: The Creation of a Synoptic Climatology for Significant Rainfall Over Kentucky and Southern Indiana.

Watson, S. A., J.T. Moore, and C.E. Graves, 1999: Validation of Eta-32 Model Quantitative Precipitation Forecasts for Rainfall Events Exceeding Two Inches.

Schmocker, G.K., and R.W. Przybylinski, 1999: A Detailed WSR-88D Radar and Damage Survey and a Severe Bow Echo Event on 14 June 1998 over the Mid-Mississippi Valley Region.

Przybylinski, R.W., G.K. Schmocker, J.M O'Sullivan and Y.J. Lin, 1999: Detailed WSR-88D Observations of a Cool-Season Tornadic Bow Echo Event on 11 February 1999 over the Mid-Mississippi Valley Region: A Unique Tornado Event.

O'Sullivan, J.M., R.W. Przybylinski, and Y.J. Lin, 1999: A Mesoscale Model Simulation of a Midwestern Linear Mesoscale Convective System

O'Sullivan, J.M. and R.W. Przybylinski,1999: A Comparison of Operational Models and a Local Mesoscale Model During a Squall Line Event from a Forecasting Perspective.

Presentations given at the annual meeting of the Missouri Academy of Science in Columbia, MO on 15 April 2000:

Nolan, S.M. and J.T. Moore, 2000: Convective Nature of Mesoscale Snow Events of Two Midwestern Storms.

Gagan, J.P. and J.T. Moore, 2000: Examination of a Training Cold Season Heavy Rain Event Over the Ohio River Valley.

Sumner, H., C.E. Graves, and J.T. Moore, 2000: A Climatology of Elevated Convection in the National Weather Service Central Region.

Abstracts submitted to the AMS 20th Severe Local Storms Conference for September 2000:

Moore, J.T. and J.P Gagan: Examination of a Training Cold Season Heavy Rain Event Over the Ohio River Valley.

Funk, T.W. and V.L. DeWald: WSR-88D Reflectivity and Velocity Trends of a Damaging Bow Echo Event on 20 April 1996 over South-Central Indiana and Central Kentucky.

Journal article published in Weather and Forecasting:

Funk, T.W., K .E. Darmofal, J.D. Kirkpatrick, V.L. DeWald, R.W. Przybylinski, G. K. Schmocker, and Y.-J. Lin:  Storm Reflectivity and Mesocyclone Evolution Associated with the 15 April 1994 Squall Line over Kentucky and Southern Indiana. Weather and Forecasting, 14, 976-993.

Journal articles in review:

Scott M. Rochette, James T. Moore, and Patrick S. Market, 2000:  The Importance of Parcel Choice in Elevated CAPE Computations. Accepted for publication in the National Weather Digest.

Patrick S. Market, James T. Moore, and Scott M. Rochette, 2000:  On Calculating Vertical Motions in Isentropic Coordinates. Submitted to the National Weather Digest for review (12/99).

Jim Moore presented a two-hour lecture and a two-hour laboratory exercise at the COMET COMAP course in Boulder, CO. This course is a seven week course for Science and Operations Officers within the NWS. This work took place from 4-6 August 1999.

The COMET Regional Workshop was held from 16-18 November 1999. The agenda and a complete set of notes was given to each workshop participant. In addition Microsoft "Powerpoint" and Corel "Presentations" talks were made available on the Department of Earth and Atmospheric Sciences' anonymous ftp site.

Greg Gust's two presentations were made available on the web at and

Ted Funk has made his workshop material available on the web at:

The Louisville QPF team has also posted their results of the Kentucky/southern Indiana snow climatology at the following web address:

The results of our Eta-32 evaluation and heavy rain climatology for Kentucky/southern Indiana have been placed in the aforementioned anonymous ftp site as powerpoint presentations. They are also available at our web site at the following address:

Gary Schmocker's presentation on MARC is on the WFO St. Louis (LSX) WEBsite. Ron's presentation on the mini-climo of MCS evolution across the Mid-Mississippi Valley has been enhanced and placed on the LSX web site as well. Plans are being made to place Fred Glass's presentation on the heavy rainfall climo on the LSX WEBsite in the near future. Both the QPF and MCS workshops developed by Fred and Ron have also been placed on the WEB. Information relating to the topics of the MARC signature and MCS / vortex evolution with convective lines across the Mid-Mississippi Valley region can be seen  on the following WEB address from WFO St. Louis:

A copy of the agenda from the COMET Regional Workshop is appended to this report.


4.1 University's Perspective

Dr. Lin's graduate "Radar Meteorology" course students went to the NWSFO in St. Louis on 28 April 1999. During a 2 hour presentation Ron Przybylinski showed several cases on the WSR-88D PUP of severe linear and bow echo squall lines. He also answered questions from the group. We hope to do more of these kinds of exercises in the future as well as "play" with the new AWIPS system at the forecast office. During time of "severe clear" Ron has told us that we can come out with some students to explore the AWIPS environment on one of their "spare" machines. We fully intend to take advantage of his offer.

On 2 June 1999 Jim Moore presented  "Summer Thunderstorm Safety" to the citizens of  Glen Carbon, IL at the Civic Center (1½ hrs). COMET-related research material was presented in this talk.

We have provided the SLUBREW software to the Paducah, KY WSFO but to date they have not been able to compile it as they do not have a FORTRAN compiler. They also require a dedicated laser printer for the SAC workstation in order to create maps for case studies. We are hopeful that these two problems can be overcome in the coming year as we anticipate working with the Paducah QPF team during our CIPS research.

Probably this biggest problem that we have had over the past year is in securing WSR-88D from NCDC. We typically tell Ron Przybylinski which time periods and stations' data we need on 8 mm tape for WATADS. He e-mails our request into NCDC. We were hoping for about a one month turn around, but typically turn around times have exceeded 4 months. This is simply unacceptable. It has hampered our ability to interrogate WSR-88D data for specific heavy rain cases.

4.2  NWS Perspective

From WSFO Louisville:

Although the Spring 1999 severe weather season was relatively quiet compared to previous years, WSFO Louisville forecasters used their knowledge of bow echo and squall line structure derived from in-house training and COMET research results to properly assess storms that did occur and issue the appropriate warnings.

On 28 May 1999, Ted Funk received an e-mail from Marty Trexler, Lead Forecaster at the WSFO Tallahassee, FL office but formerly a Forecaster and COMET project member at WSFO Louisville. In Tallahassee, Marty was on the radar when the southern portion of a line of non-severe thunderstorms suddenly began to accelerate ahead of the rest of the line. The developing bow structure also began to exhibit a comma-head (mesolow) pattern indicative of possible tornado formation within the bow.   In other words, the storm quickly transformed from non-severe to a tornadic severe cell. Knowing what was occurring, Marty quickly issued a tornado warning for the southern portion of the line. Ten to fifteen minutes later, the warning verified as tornado damage was reported from the warned storm. Enhanced storm structure knowledge and quick, accurate radar interpretation was essential in this case or else a tornado warning may not have been issued. In his e-mail, Marty credited Ted Funk for training him "superbly" and also stated that "the bottom line is the info you (Ted) taught me and the research from the COMET project was the difference."  This e-mail apparently was forwarded to Central and Southern Region Headquarters, the Office of Meteorology, and COMET as a "success" story.

During the Winter 1999-2000 season, a few significant storm systems brought the Ohio Valley rain and snow. In studying model data for these events, several forecasters at WSFO Louisville commented on how the information presented at our Winter Weather Seminars, i.e., our COMET snow climatology and the physical processes responsible for heavy precipitation production, was very helpful in assessing data and preparing subsequent forecasts. In one case in particular, the model forecasted jet dynamics and frontogenetical forcing suggested heavy rainfall potential across central Kentucky, although the model QPF did not. Forecasters decided to include heavy rainfall in the forecast, which subsequently verified.

During the 3 January 2000 severe squall line event in Kentucky, an organized bowing segment within the line produced wind damage and at least one major tornado along its path. The tornado did major damage in the city of Owensboro, Kentucky in Daviess County. WSFO Paducah did an excellent job in issuing timely tornado warnings for this storm. As the storm approached Owensboro, Pat Spoden, SOO at Paducah, contacted Ted Funk at Louisville to discuss storm trends. This coordination effort confirmed the serious threat to Daviess County and to Owensboro, which allowed ample tornado warning lead time. NWS Director Jack Kelly commended WSFO Paducah for their timely warnings as well as the effective communication between the Paducah and Louisville NWS offices.

Central Region Headquarters has determined that WSFO Louisville ranks among the top 6 Central Region offices in issuing timely and accurate flash flood warnings. This means that Louisville has a false alarm rate (FAR) of less than 0.5 and an average lead time of more than 25 minutes. Heavy precipitation training and research efforts were instrumental in placing WSFO Louisville near the top of Central Region offices in this crucial warning category.

From WSFO St. Louis:

(Further information will be sent under separate cover.)


A. Preliminary Study of the 4-5 June 1998 Heavy Rain Event in the Lower Ohio Valley
Mike York (member, QPF team in Paducah)

1. Background

Beginning in the early morning hours of 4 June and continuing  into the very early morning hours of 5 June 1998, a series of mesoscale convective systems (MCSs) produced heavy rainfall over Kentucky. Southwest Kentucky was in the primary path of the MCSs , which produced a total of 2.58 inches of rain at the Paducah National Weather Service office. The total at Paducah was representative of other rainfall reports from co-operative observers across the area. In general, the highest amounts were along the Tennessee border, with gradually decreasing amounts to the north. Totals ranged from a maximum of 3.17 inches at Mayfield to 1.32 inches at Henderson, KY.

A review of the surface observations from Paducah indicates most of the rain fell during three distinct convective events, each lasting an average of  four hours. The event with the highest total, about 1.65 inches, occurred during the evening of 4 June. The other two events each produced around one-half inch. Hourly rainfall rates were greatest during the evening event of 4 June, when up to .43 inches fell in one hour.

2. Synoptic Pattern

An analysis of the 12 UTC 4 June upper air data revealed a moderately strong west-southwesterly flow pattern at 500 mb. An embedded weak shortwave was apparent over western Arkansas. Downstream from this shortwave, upward vertical motion was observed at 700 mb over the Middle Mississippi Valley, from eastern Missouri into western Tennessee. The nose of a 35-knot low level (850 mb) jet was located over eastern Missouri. At the surface, a warm frontal boundary extended eastward from a weak low pressure center near Tulsa, Oklahoma. The location of the warm front was just north of a line from Little Rock, Arkansas to Memphis, Tennessee then northeast to near Clarksville, Tennessee. The wavy appearance of the warm front was likely due to the influence of mesoscale outflow boundaries.

By 00 UTC 5 June, some amplification of the mid- and upper-level flow had occurred. A longwave trough was observed over Texas and Oklahoma, with broad ridging downstream across the Mississippi Valley. Despite the weak upper- level ridge, an area of 700 mb upward vertical motion  persisted over the Mid-Mississippi Valley. The upward vertical motion was likely enhanced by the presence of low- level convergence in the vicinity of a surface warm front and weak low-level jet. The warm front surged north across parts of southern Missouri and western Kentucky during the day, and extended from near West Plains, Missouri through the southern tip of Illinois to south-central Kentucky near Bowling Green.    

Moisture fields in the region of concern were favorable for the formation of deep convection with heavy rain. At 12 UTC 4 June, precipitable water values of 1.5 inches or greater extended as far north as the Missouri Bootheel and far southwestern Kentucky. There was little change in these values during the day, except for a slight increase over southern Missouri. Mean relative humidity in the surface to 500 mb layer was between 50 and 70 percent at 12 UTC. By 00 UTC, it had risen to around 70 percent over the Lower Ohio Valley and southeast Missouri.   

 Lower levels were moistened by the proximity of the low level jet. 850 mb moisture convergence was noted over much of southern Missouri, far western Kentucky, and southern Illinois at 12 UTC. By 00 UTC, the best 850 mb moisture convergence had shifted southward into western Tennessee, but there was still some convergence over the region of concern. An 850 mb θe ridge was nearly stationary from eastern Oklahoma to northern Alabama during the day. The depth of the moisture was sufficient for widespread convection, as evidenced by high K indices. At 12 UTC K-indices were around 40 from the Ozarks to the Kentucky Lake region. A slow drop through the 30s was noted during the day.

Thermodynamics were favorable for elevated convection. Surface-based lifted indices were near zero over southeast Missouri and western Kentucky through the day. Surface CAPE values just over 1000 J kg-1 at 12 UTC decreased below 1000 over Kentucky and Illinois by 00 UTC. Surface-based instability was much greater over Arkansas, where CAPE values rapidly increased to 3000 J kg-1 and lifted indices were in the -4 to -8 range.           

3.  Diagnosis of the Eta Model

The ETA model output from the 4 June 00 UTC run was compared against observed conditions at the 12 and 24-hour valid times of the model. Several parameters of moisture, lift, and instability were evaluated.

The ETA 12 and 24-hour output verified well on the synoptic scale. Mesoscale outflow boundaries were not detected by the model, and these played a major role in the formation of convective systems. For this reason, the model's quantitative precipitation forecast (QPF) was only marginally accurate.

The formative stages of this event closely resembled the "frontal type" of flash flood documented by Maddox et al. (1979). At 12 UTC 5 June, a warm front was oriented from west to east in the vicinity of the Kentucky/Tennessee border. A persistent moist, southerly low-level flow was overriding the boundary. The upper-level flow was nearly parallel to the boundary, with a weak upper ridge axis over the threat area.

As the event unfolded during the afternoon of 5 June, the situation became more similar to the "Mesohigh type" of flash flood documented by Maddox et al. (1979). One of the key surface features at 00 UTC 5 June was an outflow boundary that extended from central Arkansas east through the Memphis, Tennessee vicinity and then northeast to about 40 miles northwest of Nashville, Tennessee. The low-level flow just above this boundary was from the south around 20 knots.

In many respects, this event fit the typical profile of a warm-season heavy rain event in the Lower Ohio Valley. Maddox et al. (1979) demonstrated that 59 percent of the flash flood cases he studied occurred with a weak upper trough approaching an upper level ridge axis. During the 24-hour period of concern in this case, the upper-level ridge axis was nearly quasi-stationary over the threat area. The main driving factor for this convection was a moist, southerly flow overriding a surface boundary. Maddox found that the seasonal peak of "mesohigh type" flash flood events occurs in June.   

4. Conclusion

The initial stages of this event were closely tied to synoptic scale triggers well-predicted by the Eta model. As the event unfolded, mesoscale influences became increasingly prominent, and the accuracy of the Eta model QPF decreased. According to Eckert (2000), this is a familiar sequence of events. The model's bullseye of heavy precipitation rates will be near the location of the strongest synoptic scale forcing and moisture inflow. As mesoscale outflow boundaries propagate away from the original convective clusters, the strongest forcing and moisture convergence will shift away from the model's heaviest rain area. This occurs mainly during low-shear situations, when the synoptic scale flow is least likely to disrupt outflow boundaries.

Since the current suite of short-term models is not designed to predict mesoscale outflow boundaries, the forecaster needs to watch for their formation and anticipate their movement. These features play a major role in the propagation of MCSs. In the case of 4-5 June 1998, outflow boundaries were the primary factor in the persistence of heavy convection over the Lower Ohio Valley as they refocused convective activity in the same area over a long time period.

B. Preliminary Study of the 21-22 January 1999 Heavy Rain Event in the Lower Ohio Valley
James Keysor (member, QPF team in Paducah)

1. Background

Beginning in the late evening hours on 21 January and continuing until the early evening hours of 22 January, a series of convective complexes produced heavy rainfall across the mid-Mississippi and lower Ohio valleys. A general two to four inch rainfall occurred across most of the area, with the National Weather Service Office at Paducah recording a 24-h total of 3.06 inches. This total appeared to be representative of most of the observations across the region. 

The surface observations from both Evansville, Indiana and Paducah, Kentucky reveal that a majority of the rain occurred between 06 UTC and 15 UTC on 22 January. Hourly rainfall rates were  greatest during this time, with several hours having rainfall rates exceeding 0.50 inches per hour.    

2. Synoptic Pattern

Analysis from 12 UTC 21 January revealed a deepening 500 mb trough over the southern plains with several embedded short wave troughs coming through the base of said feature. The strongest of these waves was just west of Amarillo, Texas at 12 UTC. The nose of the 850 mb low-level jet of 20 m s-1 (40 knots) was oriented from eastern Texas northeast into northern Mississippi. At 500 mb, a 50 m s-1 (100 knots) jet streak was noted along the Gulf Coast states extending from central Mississippi into Georgia. At the surface, an area of low pressure was located over central Oklahoma with a warm front snaking its way eastward across southern Missouri into southern Illinois. Precipitable water values in excess of one inch were evident across much of the mid-Mississippi valley, with 850 mb moisture convergence noted along and just west of the Mississippi river over southeast Missouri. South of the warm front, over much of southeast Missouri and western Kentucky, moderate instability was present with lifted indices (LI) less than 0°C and CAPE values around 1000 J kg-1.

By 00 UTC 22 January, the 500 mb trough over the southern plains was still deepening with several embedded short waves troughs still evident from Oklahoma northward into Missouri. The winds at 850 mb had backed in response to the approaching upper trough, with a 20 m s-1 (40 knots) jet now oriented from Louisiana northward into southeast Missouri. The strongest 500 mb winds of 50 m s-1 (100 knots) remained over the lower Mississippi valley. Because of the stronger low-level southerly flow, a greater depth of moisture was evident at 850 mb. This was indicated by the K-indices showing a 34° C bulls-eye over the eastern part of Missouri. At the surface, the area of low pressure which has been over central Oklahoma 12 h previously, had moved into southwest Missouri. 

This motion had helped push the warm front northward toward central Illinois and Indiana. A second area of low pressure was developing just south of the Texas panhandle. Instability parameters had increased, with LIs of -3 to -5 °C nosing into the Missouri boot heel.

By 12 UTC 22 January, the 500 mb flow had cutoff over the Texas panhandle, while the main upper trough slowly edged eastward. Several weaker short wave troughs were still evident from Texas into Missouri. The strongest of the these waves was just rounding the base of the upper trough and ejecting northeastward toward Oklahoma. The 850 mb jet had strengthened to 27.5 m s-1 (55 knots) and propagated a bit to the east, but was still oriented from the Gulf of Mexico northward into western Kentucky. The 500 mb jet remained at 50 m s-1 (100 knots), but was now oriented along the eastern edge of the upper trough from eastern Texas northeastward into south central Missouri. This orientation of the 500 mb jet put most of southeast Missouri and far western Kentucky in the favorable right entrance region, where divergence is favored. Moisture convergence at 850 mb was focused over western Kentucky. An 850 mb theta-e ridge axis was also seen extending from southern Illinois through western Kentucky into western Tennessee. CAPE values had also risen to nearly 1500 J kg-1 across much of western Kentucky.

3. Diagnosis of the Eta Model

The 00 UTC 21 January Eta run was analyzed through 48 hours, and was used primarily for this study. The 12-h forecast for 12 UTC 21 January had a deepening 500 mb trough over the southern plains with an embedded short wave trough near Amarillo, Texas. This was in close agreement with the actual data, although an analysis of the 558 dm height line revealed that the 12 h forecast was not quite deep enough with the upper trough. At the surface, the 12-h forecast had a surface low just west of Oklahoma City, Oklahoma with an associated warm front from central Missouri into central Illinois. This frontal placement was 25 to 50 miles further north than the actual frontal position. Because of this, the model forecast placed the 850 mb to 700 mb Q-vector  convergence zone too far to the north, closer to the front. The forecast 850 mb wind maximum of 15 m s-1 (30 knots) was 5 m s-1 (10 knots) weaker than observed. The forecast was also much too fast in backing the 850 mb flow to the south. This was most likely a direct response to the model not being deep enough with the upper trough and therefore being more progressive with the upper features.   The instability and moisture parameters all lined up fairly well, with only slight variations seen from the observed data.

The 24-h forecast for 00 UTC 22 January revealed a cutoff low at 500 mb over the Oklahoma panhandle. The forecast cutoff was nearly 12 h too fast compared with the observed data. At the surface, the 24-h forecast had a surface low over southwest Missouri with a warm front extending through central Illinois and Indiana. This frontal placement was 50 miles too far north as compared to the actual position. The 24-h forecast also failed to recognize the second area of low pressure which had developed just south of the Texas panhandle. The 850 mb wind maxima of 17.5 m s-1 (35 knots) was still under forecast with respect to speed, but the orientation from Louisiana northward toward the Missouri bootheel was fairly good. K-indices from 30 to 34°C were under represented on the 24-h forecast, with the model forecasting the maxima further to the south than actually occurred. The forecast placement of the 850 mb theta-e ridge axis from eastern Arkansas through eastern Missouri lined up well with the observed position. The general position of the instability parameters was also quite good, although the 24-h forecast under represented the large region of LIs of -5°C that had already nosed into the Missouri bootheel by 00 UTC 22 January.

The 36-h forecast for 12 UTC 22 January continued to show a cutoff low at 500 mb. It was now positioned along the Red River Valley between Oklahoma and Texas. This was once again too far south and east as compared with the observed upper air data.Several short wave troughs were seen embedded within the upper flow. The forecast had a shortwave over central Texas, but failed to resolve a  potent short wave trough over southwest Arkansas. Winds at both 850 mb and 500 mb aligned well with the observed data with respect to speed. The forecast position was too far east, once again in response to the model being to0 progressive with the upper pattern. The forecast was too far east with the positioning of the 850 mb theta-e ridge axis. It also failed to forecast the 850 mb moisture convergence bulls-eye over western Kentucky. The instability parameters were fairly well represented by the 36-h forecast, although CAPE values of 1500 J kg-1 were underdone.

4. Conclusions            

The heavy rain event of 21-22 January 1999 was quite similar to the heavy precipitation events noted by Market et al. (1995). Similarities between their 2 cool season events and this event include:

(1) Low-level jets of at least 25 m s-1 from the south-southwest which aided in providing ample moisture and lift for the event

(2) High precipitable water values in excess of 1 inch (over 150% of normal for the cool season)

(3) Weak surface cyclones

(4) Height and wind fields at 500 mb which developed and evolved slowly

Much like the 16-17 January 1994 snow case studied by Market et al. (1995), upper-level frontogenesis, associated with a strong 500 mb trough, forced a backbuilding upper level jet. This was the ingredient necessary to provide an extended episode of upper-level divergence across the heavy rain area.

The Eta model performed well overall, although it was too progressive with the upper-level features. However, because of its progressive nature, the model tended to under forecast such parameters as moisture depth and moisture convergence.