|Tornadic QLCS on 01 November 2004 over parts of Southwest Illinois -An interesting tornadic case.
Ron. W. Przybylinski - National Weather Service St. Louis. (January 2009)
During the mid afternoon of 1 November 2004 a non-supercell tornadic event occurred over parts of central through northeast Randolph county in southwest Illinois. The same convective system also produced wind damage over parts of central and northeast Washington county Illinois. Two tornadoes occurred with this case and damage were rated (F0) from parts of central through northeast Randolph county Illinois. The first tornado occurred at the Sparta Illinois country club 2 to 4 miles southwest of downtown Sparta, The second tornado touched down 1 mile southwest of Sparta and damaged several building and a Church in the downtown Sparta area. The first part of this event will disuss the pre and near storm environment utilizing upper air, sounding and surface analysis. Storm Prediction Center (SPC) mesoscale analysis output will reveal the degree of instability and shear over the Mid-Mississippi Valley Region. This will be followed by an indepth analysis of the storm reflectivity and Doppler velocity structures of the convective system. Warning Decision Support System (WDSS-II) from the National Severe Storms Laboratory was used to view the vertical cross-sections and plan views of reflectivity and Doppler velocity data from single Doppler analysis at KLSX (WFO St. Louis). Early studies on the relationship the reflectivity pattern and location of non-supercell tornadoes have been documented by McAvoy et al 2000. they showed that tornadoes occurred near the broken "S" reflectivity pattern. The sampling of their storms were within 100 km from their WSR-88D. Pence et al 1998 showed a tornadic bow echo case in early March 1996 over south-central Alabama. The tornado was located near the bottom of a "comma-shaped echo" behind the bow. They showed that tornado touchdown occurred when Rotational Velocity (Vr) values at 0.5° reached 24 m s-1 .Przybylinski and Schmocker (2002) studied the 11 Feb1999 HP to bow echo event which occurred within 100 km northwest to north of KLSX. After completing several time-heigh Vr cross-sections, one mesovortex showed the strongest Vr values (22 m s-1) below 2.5 km while the strongest Vr values of a second mesovortex (Delta V values 30 m s-1) was confined below 2.0 km. The non-supercell tornado occurred at the time of mesovortex identification leading to zero lead time (Weblink to the 11 Feb 1999 bow echo case).
As part of this study, we compared the convective mode of the 1 November 2004 non-supercell tornado case to the study of "Cool Season" convective modes completed by Burke and Schultz (2004). There has been little if any work completed on mesovortex evolutionary characteristics with "Cool Season" events. However, we will also show that the tornadic mesovortex on 01 November 2004 exhibited strong low-level rotation, had a lifetime exceeding 30 minutes, and revealed an overall depth of 4 km during the mature and later stages of the vortex. Other non-tornadic mesovorticies identified in this study were short-lived and showed weaker rotation, lasting 10 to 15 minutes and exhibited shallow depths. The second goal is to compare the mesovortex evolution of 1 November 2004 to "Warm Season" mesovortices. We will elaborate on the differences and similarities. Additional information on the overall reflectivity and mesovortex structures will be presented below. We will also briefly highlight current sampling limitation issues associated with this rapidly moving tornadic QLCSs.
2) Background Information
The definition of "Cool-Season" severe storm events vary widely amongst the research community. For example Wheatley and Trapp <2005> defines the "Cool Season" beginning in mid October and ending in March. Others like Britt and Glass (2006) define the "Cool Season" period from mid Novermber through mid March. Their defintion of a "Transition Period" would begin in October and end in mid November. In this case, we will use the later where the 01 November 2004 event occurred during the "Transistion Period" rather than the "Cool Season per se."
2) Large-scale environment for 1200 UTC - 01 November 2004
Mesoscale analysis showing the degree of instability at 1800 UTC is shown in figure 4. An area of surface-based CAPE extended from south-central Illinois through the Missouri bootheel wtih magnitudes exceeding 500 J kg-1. Surface CIN was negligible over this area. MLCAPE area was nearly coincident within SBCAPE however magnitudes were weaker with values of 250 J kg-1. A narrow ribbon of CIN less than -25 J kg-1 stretched along the Mississippi River from just south of St. Louis through the Boot heel and south into western Tennessee. This narrow ribbon coincident with the axis of surface dewpoints exceeding 68°F. MUCAPE field at 1800 UTC showed an area MUCAPE values exceeding 500 J Kg-from south-central Illinois to near the boot heel of Missouri. The stripped area shows the constant contours of lifted parcel level. Based on these derived RUC model analyses, one could suspect that greatest potential area of tornadogenesis may occur near or south of the warm fronal boundary over the northern part of the mid 60 degree surface dewpoint area, and within the area of MUCAPE. We should not exclude the 0-1 and 0-3 km SRH axis which extended from north-central Illinois through parts of southwest Illinois and into southeast Missouri.
Figure. 1 Upper air and sounding analysis for 1200 UTC 01 November 2004. (Far left) 500 mb; (Near left) 850 mb;
(Near right) Sounding from Little Rock Arkansas; (Far right) Sounding from Springfield MO. (Click on image for a larger image).
The 500 mb analysis showed a deep trough extending from eastern Montana through the four corners of the United States. Strong southwest flow was present over much of the Mid-Mississippi Valley region with winds of 45 to 60 kts. At 850 mb a warm frontal boundary stretched from central Indiana through northern Missouri and then southwest to a 850 mb low center over the Texas Panhandle region. Strong southwest flow of 35 to 40 kst was bringing warm and unstable air to the lower and middle Mississippi Valley region. 8H dewpoints reach 10° C from central Missouri through southwest Indiana. The Little Rock Arkansas sounding at 1200 UTC showed a moist layer extending from the surface to 680 mb, then a drier layer was present above 680 mb up to 350 mb. ML CAPE values at KLZK reached 1318 J Kg-1 while ML CIN was -85 J Kg-1. Deep layer shear (0 - 6 km) was 26 m s-1 while 0 - 3 km shear was 15 m s-1. Since the convective squall line passed east of KSGFMLCAPE values were quite low (37 J Kg-1) while ML CIN magnitudes exceeded 240 J Kg-1.
Figure. 2. Surface analysis for 1800 UTC 01 November 2004
(Click on image for a larger image).
The environment over the Mid-Mississippi Valley region at 1800 UTC was characterized as being unseasonably warm and humid with strong synoptic-scale forcing for early November. The 1800 UTC surface analysis (Fig. 2) revealed a number of interesting features. A strong thermal - dewpoint gradient extending from central Missouri eastward into eastern Kentucky revealed the location of the warm frontal boundary. A narrow ribbon of dewpoints exceeding 68°F along the Mississippi River from eastern Arkansas through southwest Illinois was moving northward by a southerly flow of 10 to 15 kts. The flat terrian stretching from eastern Arkansas through southwest Illinois appeared to play a role in advecting low-level moist air northward. From central Missouri a weak cold front extended west into east-central Kansas then southwest into central Oklahoma. A squall line was moving through south-central Missouri through east-central Arkansas at this time. A weak meso high was identified upshear of the line of convection while a wake low stretched from southwest Missouri into far northwest Arkansas.
b) Mesoscale Environment for 1800 UTC
Storm Prediction Center Mesoscale analysis and RUC model sounding data iwas used in this study to show the near storm environment
over parts of eastern Missouri and southwest Illinois.
Figure 3. Storm Prediction Center Mesoscale Analysis for 1800 UTC 20061101; (a) far left - 0 - 6 km shear vector; (b) near left 0-1 km shear (m/s); (c) 0-1 km storm-relative helicity (m2 s-2); (d) 0-3 km storm relative helicity (m2 s-2). (click on image for a larger image).
Storm Prediction Center Mesoscale analysis for a number of shear-based derived fields at 1800 UTC is shown in Fig 3a - 3d. Strong deep-layer shear (0 - 6 km) was present over much of Missouri and into central and southern Illinois.with magnitudes reaching 23 - 26 m s-1 (Fig 3a). Concurrently, shear within the lowest 1 km (Fig 2b.) was equally impressive with magnitudes reaching 15 m s-1 from southeast Missouri through central and southern Illinois. Shear within the 0 - 1 km appears to be a good predictor for mesovortex development. These values suggested an increased tornadic potential particularly from southeast Missouri through southern Illinois. The axis if highest 0-1 km storm-relative helicity (SRH) values extended from north-central Illinois through southwest Illinois and southwest into northeast Arkansas (Fig 2c). Highest values were found over north-central Illinois and northeast Arkansas. However the axis of SRH across parts of south-central and southwest Illiniois reached magnitudes of 200 m2 s-2 again suggesting the potential for possible tornadic development. A similar axis in the 0-3 km SRH layer was similar to the 0-1 km axis (Fig. 2d). Magnitudes over southwest Illinois varied from 200 - 250 m2 s-2.
Figure 5. RUC Sounding for Sparta Illinois (KSAR) at 1800 UTC.
RUC sounding for Sparta (SAR) Illinois at 1800 UTC was examined to showed the thermodynamic and vertical wind shear structures. The sounding analysis shows a very deep moist layer from the surface to 425 mb with a conditionally unstable lapse rate extending from the surface to 650 mb. This type of sounding is "typical" for cool-season tornado events where the moist layer is quite deep. MU CAPE magnitudes at SAR were 620 J kg-1 while CIN was very negligible. Bulk shear within the 0-6 km layer was 22 m s-1 (strong shear) while bulk shear values within the 0-3 km layer were 18 m s-1 suggesting the upper end of end of the moderate shear category.
3) Radar Analysis from KLSX
WSR-88D reflectivity and Doppler velocity data was used to show the changes in the reflectivity and velocity structures of the convective line that moved across parts of southwest Illinois during the early part of the afternoon. At 1934 UTC a narrow and linear high reflectivity convective line segment extended from northwest Randolph county Illinois into east-central Ste Genevieve county Missouri. A smaller secondary line segment stretched along and west of the larger segment over extreme west-central Randolph county through far northeast Ste Genevieve county MO. A third line segment extended mainly across western Perry county MO. Moderate precipitation covered parts of central Randolph county just east of the larger line segment. Stratiform precipitation dominated areas north of the larger convective line segment over St. Clair and eastern Monroe counties in southwest Illinois or southeast of the St. Louis metropolotian area. .
Figure. 6 Base reflectivity (left) and Storm-relative velocity (right) at 0.5° from KLSX at 1934 UTC (click on image for a larger image).
Storm-relative velocity data at this time showed two weak mesovortices along the southeast flank of the larger line segment and the third line segment over eastern Ste Genevieve county. A shear axis was noted to the north of the mesovortices along the eastern flank of the larger line segment shown. Early observations documented by Przybylinski (1988) showed that in the March 10, 1986 tornado outbreak over central Indiana showed that two tornadoes occurred between two convective line segments giving the overall appearance of a "comma-shaped" echo. Using WSR-88D reflectivity and Doppler velocity data, McAvoy et al 2000 revealed the occurrence of tornadoes near the northern and southern ends of two convective line segments also showing a "comma-like shaped" echo during the "cool season."
Figure. 7 Similar to figure except for 1939 UTC (Click in image for a larger image).
Five minutes later (1939 UTC) the convective line segments over west-central Randolph county continued to move northeast at approximately 20 m s-1 (Fig. 7) . Along with the larger convective line segments, the weak comma-shaped echo earlier identified at 1934 UTC was just west of the Mississippi River near the Ste-Genevieve - Perry county line. SRM velocity data at this time showed a well defined low-level convergent shear axis stretching from northwest Randolph county IL through far eastern Ste. Genevieve county MO. The strongest part of the shear axis was over west-central Randoph county or approximately 20 km southwest of Sparta (SAR) IL. The strongest part of the shear axis at lowest levels should be closely monitored for rapid mesovortex development. A weak mesovortex was still persistent on SRM velocity image continued to be embedded within the comma-shaped echo and was located over northern Perry county MO.
Figure 8. Same as figure 6 except for 1944 UTC (Click on image for a larger image).
At 1944 UTC, higher reflectivity values were obsevered along the nearly linear convective line segment as it approached central sections of Randolph county. A strong low-level reflectivity gradient along the eastern flank of the larger line segment over central Randolph county became more evident at this time. Concurrently, a much weaker reflectivity area east of the segment became more visible and suggested that this area was becoming the convective line's inflow region. Across north and northeast Randolph county, an area of higher reflectivity extended east-northeast to just north of Sparta Illinois. This reflectivity feature appeared to resemble a warm advection wing. The comma-shaped echo over the southern part of the line did not show much in the way of structural changes siince 1939 UTC as it remained weakly defined. Comparing storm modes of bow echo evolution from Burke and Schultz (2004) to this case revealed that the overall reflectviity pattern on 01 Noember resembled the Squall Line - Cell Merger mode (See Burke - Schultz 2004 paper). We observe areas of smaller cells merging along the downshear (eastern) flank of the larger convective line. The shear axis shown on 0.5° SRM velocity data across central Randolph county further intensified at this time and was coincident with the leading edge of the higher reflectivity line segment. The inital signs of developing mesovortex (MV -1) can be found at this time over west-central Randolph County Illinois (See figure 8 - SRM data) Enhanced low-level convergence was noted over the southern part of the segment 10 km northwest of Chester while a cyclonic convergent couplet was identified north of the enhanced low-level convergence or about 14 km west-southwest of SAR.
Figure 9. Same for figure 6 except for 1949 UTC. Additionally base reflectivity and storm-relative cross-sections bisected MV1 and the TVS. Cross-sections - viewing to the northwest - towards the KLSX WSR-88D (Click on image for a larger image).
The reflectivity field at 0.5° from KLSX at 1949 UTC continued to show a nearly linear convective structure extended north-south across the middle part of Randolph County Illinois. The best inflow region along the leading edge of the convective line became more evident with a weak reflectivity notch further evolving 5 to 9 km southwest of Sparta (SAR) IL. A poorly defined comma-shaped echo was still present west and southwest of Chester Illinois. Significant changes occurred in the SRM velocity data in which a Tornadic Vortex Signature (TVS) embedded within a larger mesovortex (MV 1) was identified 9 km west-southwest of SAR or 99 km southeast of KLSX. Delta-V magnitudes exceeded (45 m s-1). The strongest part of the mesovortex remained below 2 km (e.g. check Fig. 13b Vr trace). At the time when the TVS was present, no tornadic damage was found. Since using WDSS-II we were able to constructed vertical cross-sections of MV1 and the TVS. The right two figures show cross-sections of reflectivity and SRM Doppler analysis. The cross-sections are allowing us to view to the northwest towards the KLSX WSR-88D Doppler radar. The center of the TVS is 99 km away from KLSX WSR-88D. The reflectivity cross-section shows the maximum storm top height approximately 4.3 km while the height of the 50 dBz core extends to 2.0 km. The SRM velocity image reveals the TVS primary within the lowest 1.0 km with slightly weaker cyclonic rotation above the TVS. This type of vortex structure is similar to vorticies associated with tropical storm rainbands (G.V Rao - personnel communications). The strong inbound veloctiies above the low-level outbounds, left of the TVS, represents the storm's ascending branch to the rear while low-level outflow is noted below. At this time there was no tornadic damage.
Figure 10. Same as figure 6 except for 1954 UTC (Click on image for a larger image).
The TVS was not detected on the subsequent volume scan (1954 UTC) however a strong mesovortex was still identified at 1954 UTC with one radial separation between the maximum inbound and outbound isodops. Vr magnitudes at 0.5° slice was 18 m s-1 (36 kts) (using 8 bit velocity data). In the orginal plot of low-level mesovortex strength and tornado occurrence (lower part of main QLCS page) we found out that there is a high probability of tornado occurrence when low-level Vr magnitudes reached 18 m s-1 and greater. Vortex deepening also occurred between 1949 and 1954 UTC which is a second signal for increased tornado potential and the vortex column become vertically stretched. The reflectivity field at 1954 UTC continued show a nearly linear convective line, with reflectivity magnitudes exceeding 50 dBZ. This structure was similar to the reflectivity structure during the past three volume scans. The reflectivity and velocity cross-sections taken near the center of the mesovortex core is shown on the right side of Fig. 10. The depth of the convective line changed little since 1954 UTC. However, SRM velocity data showed an increase in the core diameter and the depth of the vortex. Scanning through the mesovortex core with the cross-section revealed that the vortex was tiltied down shear with height thus the highest part of the mesovortex was northeast of the current cross-section. Since there was very little if any bowing at this time, this observation goes against conventional thinking in which a mesoscale RIJ combined with local convective-scale downdrafts would aid in the initial development of a bowing segment and subsequent mesovortex evolution.. So why is this happening? - we will provide the answer later on this page. The weak reflectivity notch just southwest of SAR continued to represent the inflow region of MV1.
Between 1954 and 1959 UTC, the first of two tornadic damage tracks occurred 5 to 7 km southwest of SAR at the Sparta County Club. Several large trees across the country club golf course were snapped or uprooted by the tornado (See Fig. 13amesovortex (MV 1) track and damage mapping - first tornado damage track).
Figure 11. Same as figure 6 except for 1959 UTC (Click on image for a larger image).
Slight changes in the low-level overall reflectivity pattern were noted at 1959 UTC as MV1 was mainly embedded within the higher reflectivity region and began to redistribute the precipitation field. Slight bowing of the linear line was evident across parts of central and east-central Randolph county. The reflectivity inflow notch in the vicinity of SAR was become less discernable compared to earlier times as the mesovortex was mainly masked within heavier rainfall. SRM velocity data showed a moderate intensity mesocyclone approaching SAR with one radial separation between the max inbound (outbound) velocities. The strongest Vr shear values remained at the 0.5° slice (or 1.6 km AGL) while the depth of the vortex did not change between 1949 and 1959 UTC (See Fig. ). The first tornadic damage path ended at approximately 2000 UTC.approximately 5 km southwest of SAR.
Figure 12. Same as figure 6 except for 2004 and 2009 UTC (Click on image for a larger image).
A sequence of the base reflectivity and SRM velocity fields for 2004 and 2009 UTC are shown in Fig. 12. Base reflectivity and SRM velocitiy data at 2004 UTC showed that the developing bowing segment was moving over Sparta Illinois. MV1 was totally embedded the high reflectivity core region while a developing Rear Inflow Notch was present along the trailing flank of the bowing segment. A second tornado associated with MV1, touchdown started near the southwest side of Sparta (SAR) at approximately 2002 UTC and lasted for a few minutes. The tornado damaged a number buildings in town and took the top half of a church steeple. Window and parts of roof and some siding damage was documented along the tornadic damage path in town. The tornado lifted about 1/2 mile northeast of SAR.
Figure 13. (a) The left image shows the mesovortex track of MV 1 and the time of tornadoes and wind damage southwest through Sparta Illinois. (b) The middle image shows a time-height Vr trace of MV1. (c) The right image reveals the mesovortex core diameters for the lowest three elevation angles. (Click on image for a larger image).
The left image on Fig. 13 shows the mesovortex tracks of MV1 through MV4. MV 1 had the longest lifetime while MV 2 through 4 showed very short lifetmes during this period. The strength of rotation was much weaker with MV2 to MV4 compared to MV1. The "time-height Rotational Velocity" (Vr) trace with MV 1 revealed increasing mesovortex strength as Vr values at 0.5° slice increased from 15 to 18 m s-1 from 1944 to 1954 UTC. Within this period a TVS was identified at 1949 UTC at 0.5° slice. No damage occurred at the time of the TVS. MV1 also showed signs of deepening between between 1949 and 1954 UTC.from 3.2 km to 4.8 km. Additionally mesovortex core diameter (MV 1) decreased in size from 3.3 km to 2.2 km between 1944 and 1954 UTC. The first of two tornado touchdowns occurred 5 minutes later after all of these vortex characteristics were completed. The overall mesovortex pattern shown here is similar to "Warm Season" tornadic QLCS. Differences include much deeper vortex height in warm season events compared the "Cool Season counterparts. In some of the "Cool Season" events, low-level Vr magnitudes were similar to their "Warm Season" counterparts (Vr magnitudes ~ 20 m s-1). In the 1 November 2004 tornadic QLCS event low-level Vr magnitudes were less than 20 m s-1 (90 - 100 km range from the KLSX WSR-88D). The highest peak Vr values were 18 m s-1 below 2.5 km.
As of today, there are only a handful of studies on topic of "Cool Season" tornadic QLCS events. The 01 November 2004 tornadic QLCS occurred under a strong low-level and deep layer shear environment and weak instability (500 J Kg-1). Both tornadoes occurred south of the warm frontal boundary, over the northern part of greatest MUCAPE and within the axis of 0-1 km SRH. The convective line developed over parts of southeast Missouri and moved rapidly east-northeast into Randolph County over southwest Illinois. As the convective line moved across western and central Randolph County a nearly linear reflectivity pattern was observed. Weak convective cells merged along the eastern flank of the linear convective line similar to Mode F squall line - bow echo structure (Burke and Schultz 2004). Over western Randolph County, SRM velocity data showed a convergent shear axis prior to mesovortex development. Mesovortex (MV 1) formed at low-levels at 1944 UTC deepened and intensified at low-levels through 1954 UTC. Inintially the MV1 core diameter ranged around 3.5 km then dropped to 2.0 - 3.0 km just prior to and during the time of tornado occurrence. A TVS signature was noted at 1949 UTC where Delta-V values reached (45 m s-1 or 90 kts). At this time there was no reports of tornadic activity. The first or two tornadoes touched downs occurred at 1959 UTC or five minutes after low-level Vr magnitudes reacbed 18 m s-1 and the vortex reached its greatest height of 4.8 km. The second tornado touchdown occurred approximately 3 - 4 minutes later (2004 UTC) and caused damage to a number of building (roof and window damage) and destroyed a church steeple in Sparta Illinois. Both tornadoes were 90 to 100 km southeast of the KLSX WSR-88D.
One of the greater challenges for warning forecasters is viewing "Cool Season" tornadic or non-tornadic vortices at distances greater than 130 - 140 km from the RDA. Radar horizon issues may be the largest factor since the radar beam may "overshoot" the strong low-level mesovortex. Additional survey of environmental conditions may play a stronger role in these cases. After 2004 UTC, MV 1 weakened as it moved into southwest Washington County Illinois. MV 1 showed characteristics similar to "Warm Season" vortices (strong Vr values below 2.5 km and deepening vortex before tornadogeneis. Other vortices identified over southern Randolph County were short-lived and revealed weaker Vr values compared to MV1. As of today there are only a handful of studies on the topic of "Cool Season" tornadic QLCS events. The study of additional "Cool Season QLCS" events will allow us to further understand why some vorticies spawn tornadoes while other do not. Additionally, these studies will allow us the further refine today's mesovortex conceptual models.
Britt M. F. and F.H. Glass, 2006: Environmental and synoptic conditions associated with "Cool Season " strong and violent tornadoes in the North Central United States. Preprints 23rd Conf. on Severe Local Storms, Amer. Meteor. Soc., St. Louis MO.
Burke P.C., and D.M Schultz, 2004: A four year climatology of Cold-Season Bow echoes over the Continental United States. Wea Forecasting, 19, 1061 - 1074.
McAvoy, B.P.., W.A. Jones, and P.D. Moore, 2000: Invertigation of an unusal storm structure associated with weak to occasionally strong tornadoes over the Eastern United States. Preprints, 20th Conf. on Severe Local Storms, Amer. Meteor. Soc. Orlando Fl. 182 - 185.
Pence, K.J., J.T. Bradshaw, and M.W. Rose: 1998, Preprints, 19th Conf. on Severe Local Storms. Minneapolis MN, 147 - 150.
Przybylinski, R.W. and G.K. Schmocker, 2002:Characteristics of circulations associated with the 11 February 1999 over the Mid-Mississippi Valley Region. Preprints, 21st Conf. on Severe Local Storms. Amer. Meteor. Soc. San Antonio Texas.