WSR-88D Detection of June 4, 1993 Derecho Event over Southeast Missouri, Southern Illinois and Western Kentucky


Bradley S. Small
National Weather Service Forecast Office
Des Moines, Iowa






On the morning of June 4, 1993, thunderstorms that produced severe wind damage along with F0 and F1 tornadoes moved across the Midwest from southern Missouri through southern Illinois and across much of Kentucky. This storm complex resembled characteristics similar to that of a Derecho (Johns and Hirt 1987). The WSR-88D at WSFO St. Louis, Missouri (KLSX) detected the storm structures very well giving forecasters excellent information to issue timely, accurate warnings and allowed them to pass information to adjacent NWS offices. This paper will focus upon a storm that evolved from a High Precipitation (HP) type supercell (Moller et al. 1990; Przybylinski et al. 1993) to a large bowing convective line segment. WSR-88D reflectivity and Doppler velocity data will be used in the analysis to show the detailed storm reflectivity and velocity structures.




A mechanism that enhanced the development of moist convection was an east-west oriented front that extended from extreme south central Missouri into northern Kentucky at 1300 UTC (all times here after UTC) (Figure 1). An associated axis of surface moisture flux convergence was noted along and south of the boundary from northern Arkansas to western Kentucky. The degree of instability was marked by an -8 to -10 surface Lifted Index axis that extended from central Oklahoma through western Kentucky and Tennessee (Figures 2a and 2b).


The Paducah, Kentucky (PAH) 1200 sounding further characterized the degree of instability with Convective Available Potential Energy (CAPE) of about 3500 J/kg. The 0-3 km vertical speed shear at PAH was just within the moderate category with a value of about 14 m/s (Figure 3). According to Weisman (1993), vertical shear is moderate when the shear through the lowest 2.5 km ranges from 12-19 m/s. Strong shear is defined when vertical shear over the layer exceeds 19 m/s. Additionally, little directional shear was observed with a nearly uniform west to southwest wind throughout the layer. Other factors that created a more favorable environment for convective development over the region included strong warm advection at 850 mb, an approaching 500 mb shortwave trough, and 300 mb diffluence (all not shown) across southern Missouri and southern Illinois at 1200.


Figure 1. Analyzed surface map (inches Hg) 1300 UTC 4 June 1993.

Figure 2a. ADAP fields from 1300 UTC 4 June 1993 showing surface moisture flux convergence {((g/kg)/hr)*10}.

Figure 2b. ADAP fields from 1300 UTC 4 June 1993 showing lifted index (°C).



Around 1330, the storm of interest (Storm A) initially showed reflectivity characteristics of a HP type supercell over southeast Missouri near southern Reynolds and northern Carter Counties (Figure 4). Wind damage began in Wayne County around 1400 and tracked east-southeastward across Scott and Mississippi Counties in southeast Missouri (5-30 miles south of Cape Girardeau (CGI)). Initial reports indicated weak tornado (F0 and F1) damage along with straight line winds at speeds exceeding 25 m/s (50 kts). As the storm matured into an organized bowing line segment (derecho stage), the magnitude and coverage of the straight line winds increased. Surface wind speeds by 1445 were estimated at between 40-45 m/s (80-95 kts) over parts of extreme southern Illinois and northern Kentucky (near PAH).




Scanning strategies employed by the WSR-88D during the event included operation in Volume Coverage Pattern (VCP) 21 as opposed to VCP 11. With the storm approaching the effective range of WSR-88D velocity measurement, this scan strategy was utilized to obtain the highest quality velocity data available. Increased velocity resolution resulted from a slower scanning rate and longer sample time. A combination of reflectivity (.54 and 1.1 nautical mile (nm) resolution), base velocity (.54 nm), and storm relative velocity (SRM) (.54 nm) products were used to diagnose the storm's reflectivity and velocity structures. Since the storm was located over 100 nautical miles (185 km) from the RDA, the two lowest elevation angles were used (0.5 and 1.5 degrees). At this distance, the beam center of the 0.5°, slice sampled the storm's mid-level structure at around 4.5 km (14,000 ft MSL). At 1.5°, the center beam is at 7.6 km (25,000 ft MSL). In this case, the storm's mid level features were sampled with little evaluation of lower-level characteristics. The forecaster should always consider the range of the storm from the RDA and where each particular elevation angle intersects the target.


Figure 3. Vertical wind field table and upper air sounding from 1200 UTC 4 June 1993 at Paducah, Kentucky (PAH).



Between 1330 and 1400, Storm A appeared to exhibit HP supercell reflectivity characteristics (Moller et al. 1990; Przybylinski et al. 1993) as it traversed across extreme northern Carter County into southern Wayne County Missouri. A broad mesoscale circulation (11-13 km diameter) observed on the 1333 Storm Relative Velocity (SRM) product (0.5° and 1.5° slices) was located within one of the storm's 40-50 dBZ core regions. Other small 50 dBZ reflectivity cores embedded within the HP storm suggested that several updraft centers may have been present along the storm's leading edge (Figures 4 & 5). The depth of the circulation exceeded 3 km while rotational velocities (Vr = (¦Vin¦ + ¦Vout¦ /2)) of 21 to 28 knots depicted the circulation's intensity.


Figure 4. WSR-88D base reflectivity field for 1333 UTC at 0.5° elevation slice.


During the next few volume scans, the magnitude of the mid-level region's circulation gradually intensified, however its diameter remained a broad feature (13 km dia.). Figure 6a reveals a time-height series plot of the vortex rotational velocities (Vr) located within what would eventually evolve into the large bowing segment's comma-head. The plot indicated that the circulation's early stages (1330-1345) were characterized as a broad and weak vortex having a diameter of 11-13 km and depth exceeding 3 km (10 kft). However, Vr values increased from 10.5 to 16.5 m/s (21-33 kts) from 1333 to 1350. With those increased Vr values, the circulation's diameter also decreased to 6.5 km (satisfying mesocyclone criteria) suggesting that the vortex was becoming more organized with time. Comparison of SRM to reflectivity data continued to show the circulation near the core of the maximum reflectivity that is consistent with the definition of an HP supercell. Magnitudes of Vr at these distances (105 to 110 nm) would qualify the vortex as a weak to moderate mesocyclone if using the National Severe Storms Lab's (NSSL) mesocyclone criteria threshold values (OSF 1995). At this point, F0 to F1 tornadoes and damaging winds were reported within and immediately south of the circulation's center as the storm moved across Wayne County.

Figure 5. WSR-88D storm relative velocity (SRM) field for 1333 UTC at 0.5° elevation slice, Mesoscale circulation denoted by dashed circle.


Figure 6a. Plots of Time/height cross-section of rotational velocities (Vr in kts) at 0.5° and 1.5° elevation slices.


Figure 6b. Plots of Time vs. 0.5° Circulation Diameter (km).


From 1350 to 1413, the storm began to evolve into a bowing structure similar to TYPE IV storm evolution (High Precipitation to Bow Echo, Przybylinski and DeCaire 1985) with storm movement from the west at 20 - 25 m/s. Reflectivity data at 0.5° from 1350 to 1425 showed the initial stages of evolution with a coherent region of 50 dBZ echo outlining the bowing structure (Figure 7). A Rear Inflow Notch (RIN) gradually developed along the storm's trailing edge (west-southwest of the circulation center) by 1413. The reflectivity core near the northern edge of the bow also further intensified and began resembling a spiral-like structure. The next several volume scans indicated the 50-55 dBZ reflectivity core expanded in size while a strong low-level reflectivity gradient and inflow notch were identified near the storm's forward flank. These reflectivity signatures appeared to indicate that the transport of cooler, drier mid-level air to the bowing line segment's leading edge was likely being enhanced, the storm's updraft center intensified and a vortex was still present. Intense damaging straight line winds would become an increased possibility over southern Bollinger, northern Stoddard and eventually into Scott County as lower theta-e air was presumably being channeled toward the storm's leading edge.


The evolution of the circulation's intensity is particularly interesting during the period 1350 through 1420 and seems to support this assumption. The highest Vr values were detected within the storm's mid-level region (6 - 8 km) during the first part of this period. However after 1356, the circulation's rotational velocities gradually increased within the vortex's lower regions (3 - 4 km) while Vr values at mid-levels significantly decreased. The stronger circulation over the storm's lower region likely strengthened the descending mesoscale Rear Inflow Jet and enhanced the production of widespread damaging straight line winds along the system's leading edge. SRM data further confirmed this observation with the presence of a strong circulation (Vr > 30 kts) within the 50 - 55 dBZ core region. These features all correlate well with the first occurrence of downburst wind damage and appear similar to numerical simulation results documented by Weisman (1993).


Figure 7. Four panel display of base reflectivity field for 1350, 1402, 1413, and 1425 UTC at 0.5° elevation slice.

The circulation from 1413 to 1431 slightly weakened as it progressed into northern Scott County (105-110 nm from RDA) with the vortex located near the northern reflectivity maximum. A plot of the circulation's intensity (Vr values) during this period showed this slight decrease in magnitude near 4 km as the storm evolved from a HP structure to a bowing line segment (Figure 6). The circulation continued to redistribute the precipitation and resulting in a spiral shaped reflectivity pattern at the northern end of the bow. By 1431, the Vr value was 26 knots at 0.5°, placing the circulation in the upper bounds of a weak vortex (using NSSL mesocyclone velocity threshold values) at 105 nm from the radar (105 nm). The broad circulation moved across the Cape Girardeau, MO (CGI) airport at 1430, where observed winds reached 28 m/s (56 kts) (Figure 8).


On the next volume scan (1437), base velocity data at 0.5° indicated a 25 m/s (50 kts) wind maxima within the vicinity of the RIN over northern Scott County (Figure 9a & 9b). The local wind maximum at 4.3 km is a result of the descending mesoscale Rear Inflow Jet (RIJ) current (Smull and Houze 1987; Przybylinski and Schmocker 1993). It is important to note that when viewing WSR-88D velocity data, one must remember that the radar did not capture the true magnitudes of the local velocity maximum since the beam was practically normal to the probable wind direction. Taking that into consideration, the presence of a 50-kt wind maximum is significant and further suggested there was a high probability that intense damaging winds were occurring at this time near the leading edge of the bow. In this case, the strongest winds likely occurred 7-15 km south of CGI.


Figure 8. Plot of damage for morning of June 4, 1993. "W" denotes severe wind damage (50 kts or greater) and "T" indicates tornado with F-Scale in parenthesis. Time is given in UTC.

Figure 9a. WSR-88D base reflectivity for 1437 at 0.5° elevation slice. Arrow points to location of Rear Inflow Notch (RIN) and corresponding Rear Inflow Jet.

Figure 9b. WSR-88D base velocity field for 1437 UTC at 0.5° elevation slice. Arrow denotes location of outbound velocity maximum (signifies descending mesoscale Rear Inflow Jet structure).

The reflectivity image at and before 1437 revealed a nearly solid 50 dBZ core along the leading edge of the broad bowing structure while a notch-like concavity was observed along the trailing flank. Low-level and elevated reflectivity scans showed nearly vertical convective towers along the leading edge of the bow. Reflectivity images of the convective system through 1437 were supportive of Weisman's second stage of numerical simulations, where a balance existed between the storm-induced cold pool circulation and the ambient shear.


Just south of the broad circulation after 1443, the leading edge of the nearly solid bowing structure further accelerated and fragmented near the apex of the bow (Figure 10). The fragmented bowing structure may have resulted from an accelerated descending mesoscale RIJ. The reflectivity pattern rapidly transformed into two 50 dBZ core regions near the ends of the bowing structure with weaker reflectivities near the apex. These reflectivity features are similar to the third stage of Weisman's numerical simulation, where the convective system's cold pool overwhelmed the ambient shear near the apex of the bow. Meanwhile, a balance between the system's cold pool and shear existed near the ends of the bow resulting in intense convective storms over these regions. This type of storm reflectivity evolution was similar to observations of a large bowing structure documented by Przybylinski and Schmocker (1993). Damage reports during this period (1440 - 1450) indicated that wind gust winds exceeded 40 m/s through the bow's apex with numerous swathes of extensive tree damage documented in this region.


Figure 10. WSR-88D base reflectivity for 1443 UTC at 0.5° elevation. Arrow points to fragmentation at apex of bowing structure.

The mid-level circulation, located within the northern 50 dBZ core region (Figure 11) during this period weakened, expanded in size and took on the characteristics of a `cyclonic bookend vortex' (Weisman 1993) with a diameter of 16 km. Weisman noted that the bookend vortex will often further strengthen and deepen the descending mesoscale RIJ current. The 25 m/s outbound velocity maximum first identified at 1437 South of the circulation had expanded over northern and central Scott County at 1448 (Figure 12) and continued to be aligned with the reflectivity RIN. Utilizing this knowledge and the fact that the radar beam remained fairly normal to the storm, these signatures further indicated that significant wind damage would continue to occur along and immediately south of the bookend vortex near the bowing segment's leading edge.


Figure 11. WSR-88D storm relative velocity field for 1448 at 0.5° elevation slice. Dashed circle denotes mesoscale circulation.

Figure 12. WSR-88D base velocity field for 1448 UTC at 0.5° elevation. Arrow denotes location of expanded outbound velocity maximum and descending mesoscale Rear Inflow Jet.

As the bowing structure further broadened after 1450, the storm appeared to imitate Weisman's fourth stage of numerical simulation. The cold pool further expanded and its associated circulation continued to overcome the ambient shear. Reflectivity and velocity images at 1500 displayed a broad but distinct bowing line segment and RIN outlined by the 35 dBZ echo near the trailing edge of the bow and south of the weakening circulation. After this time, velocity data was no longer available as the line of storms moved outside the 88D's 230 km (124 nm) range. However, valuable reflectivity information was still used as a component in the warning process. Although the radar was sampling well into the storm's mid-levels, the bowing structure along with the RIN feature was still evident as the storm progressed across southern Illinois and into central Kentucky. While reflectivity structures at this height will not always translate into damaging surface winds, in this instance significant wind damage similar to that received over southeast Missouri continued to be reported across the region.




During the morning of June 4, 1993, an organized cluster of thunderstorms developed over southern Missouri, moved into southern Illinois and across much of Kentucky. The convective system was characteristic of a Derecho with widespread significant wind damage and small-scale tornadoes along the system's leading edge throughout its history. Initial reflectivity structures suggested that the storm was a HP type supercell, but later evolved into a large bowing structure by the time it reached extreme southeast Missouri, southern Illinois and far western Kentucky.


The WSR-88D at St. Louis, Missouri (KLSX) captured the High Precipitation supercell to Bow Echo evolution that had similarities to the results of studies conducted by Moller et al. (1993) and Przybylinski and DeCaire (1985) (TYPE IV). Reflectivity, base velocity and storm relative velocity products from the radar detected the storm's transformation by capturing such features as mesoscale circulations, a large bowing line segment, Rear Inflow Notches (RIN) along the system's trailing edge and local wind maxima (Descending Mesoscale Rear Inflow Jet). The 88D representation was very similar to numerical simulations documented by Weisman (1993). As a result, forecasters could examine detailed storm structure and provide timely, accurate information to surrounding NWS offices.




The author would like to thank Ron W. Przybylinski (SOO - WSFO St. Louis, MO) and Karl A. Jungbluth (SOO - WSFO Des Moines, IA) for their guidance during the preparation of the paper and their comments and suggestions throughout the review process. The author would also like to extend his appreciation to Steven D. Thomas (MIC - WSFO St. Louis, MO), Ernest H. Goetsch (MIC - WFO Lincoln, IL) and J. John Feldt (MIC -WSFO Des Moines, IA), all past or present supervisors, for their encouragement to write this paper.




Johns, R.H, and W.B. Hirt, 1987: Derecho: Widespread Convectively Induced Windstorms. Wea. Forecasting, 2, 32-49.


Moller, A.R., C.A. Doswell III, and R.W. Przybylinski, 1990: High-precipitation supercells: A conceptual model and documentation. Preprints, 16th Conf. Severe Local Storms, AMS (Boston), 52-57.


Operational Support Facility (OSF) WSR-88D Operations Training Student Guide, 1995: WSFO Norman Modified NSSL Mesocyclone Criteria, 8-105.


Przybylinski, R.W., and D.M. DeCaire, 1985: Radar signatures associated with the derecho, a type of mesoscale convective system. Preprints, 14th Conf. Severe Local Storms, Indianapolis, AMS (Boston), 228-231.

____________, and G. K. Schmocker, 1993: The Evolution of a Widespread Convective Wind Storm Event Over Central and Eastern Missouri. Preprints, 13th Conf. on Weather Analysis and Forecasting, Vienna, Virginia, AMS (Boston), 461-465.


Smull B.F., and R.A. Houze, Jr. 1987: Rear Inflow in squall lines with trailing stratiform precipitation. Mon. Wea. Rev., 115, 2869-2889.


Weisman, M.L., 1993: The Genesis of Severe, Long-Lived Bow Echoes. J. Atmos. Sci., 50, 645-668.


USA.gov is the U.S. government's official web portal to all federal, state and local government web resources and services.