Between 0230 and 0300 UTC on September 21, 1996, a bow echo developed rapidly over extreme northeast Texas and moved into extreme northwest Louisiana, just north of Shreveport. This system produced severe wind gusts in excess of 55 kt and hail one inch in diameter near Shreveport, within a tornado watch area. Conditions were obviously recognized as favorable for the development of severe thunderstorms; however, prior to the development of the bow echo, the activity appeared rather benign on radar. It was mainly just a large area of rain and thunderstorms. Within a period of approximately 20 min, the convection underwent a rapid change in which it became a well-defined bow echo within a line echo wave pattern (LEWP). Almost as quickly as it developed, the bow echo then rapidly weakened again into a large area of rain and thunderstorms.
This event was well handled by the responsible NWS offices. The following analysis documents an excellent case of rapid bow echo development, and examines the synoptic environment that indicated favorable conditions for the development of bow echoes. Radar products documenting the rapid evolution of the convective system are discussed, along with implications for the warning process.
2. Synoptic Environment
The synoptic environment at 0000 UTC on September 21, 1996, is shown in the composite chart in Fig. 1. A large mid- and upper-level cyclone was centered over the northern Plains, with a well defined trough extending south into the southern Plains. A surface cold front associated with this system was pushing into the Arklatex region, providing low-level moisture convergence to help focus the development of organized deep convection.
The Shreveport upper-air observation at 0000 UTC on September 21 (Fig. 2) was quite unstable, with a lifted index of -6oC. In addition, a large area of dry, low theta-e air existed in the mid and upper levels above moist air in the lower levels, as shown by the steep lapse rate in the theta-e profile displayed in the upper left-hand corner of the sounding. Hence, strong convective instability was clearly in place. The wind profile was characterized by weak winds in the lowest levels of the atmosphere, but above 850 mb much stronger westerly winds existed with significant speed shear. This can also be seen in the hodograph (Fig. 3), which is straight-line above the lowest levels of the troposphere. Helicities were very low, and in fact were negative.
The pattern discussed above appears to be similar to the dynamic bow echo pattern outlined by Johns (1993). This is supported by the dry air available in the mid-levels for the enhancement of downdraft production; mid-level dry air is not normally associated with the dynamically weaker warm-type bow echo development. Although the low-level winds were fairly light, the strong mid- and upper-level winds over the region (due to the proximity of the trough moving through the southern Plains) and the straight-line hodograph showing a nearly parallel orientation of mid-and upper-level winds all seem to better support a dynamic bow echo pattern than that of the warm type bow echo development.
3. Radar Products
Figure 4 shows the KSHV 0.5o base reflectivity and base velocity products starting with 0228 UTC (Figs. 4a and 4c). At this time, rain and thunderstorms were over the Arklatex region, between Texarkana and Longview, Texas. The first feature of note was a short bowing line just south of Texarkana. Although this line later formed part of the LEWP associated with this event, the preliminary local storm report issued by Shreveport that evening did not indicate any severe weather with this particular feature. Even though the remainder of the activity did not look particularly impressive--either on reflectivity or Vertically Integrated Liquid (VIL) products (not shown)--base velocity for the same time (Fig. 4c) shows a small area of 36 to 49 kt inbound velocities over eastern Marion County, Texas, fairly close to the leading edge of the convection. In addition, a small area of lower reflectivities can be seen just to the west of this area of high inbound velocities (Fig. 4a).
By 0238 UTC, the area of inbound velocities had increased significantly (Fig. 4d), with a fairly pronounced area of 50 to 63 kt inbound velocity noted. During this same period, the area of lower reflectivity to the west of the high inbound velocity also became more apparent (Fig. 4b), and by 0248 UTC (Fig. 5a) it appears to have become a well defined rear inflow notch (RIN), a typical feature of bow echo storms (Przybylinski 1995).
In addition, the 0248 UTC reflectivity also shows that the convective line as it extended from Texarkana to northwest of Shreveport was starting to take on a LEWP configuration. A segment of the line definitely bowed out as it approached Caddo Parish and the Shreveport area, with the well defined RIN just behind the bowing segment. In addition, base velocity data (Fig. 5c) for the same time continued to show a large area of 36 to 49 kt inbound velocities, with a small area of 50 to 63 kt inbound velocity within it.
The complex still had a LEWP configuration in the 0258 UTC reflectivity (not shown), with a well defined bow echo signature having moved rapidly into northern Caddo Parish, Louisiana. The VIL (not shown) for the convective complex peaked at this time, although it only reached 41 kg m-2. At this time one-inch diameter hail was reported over northern Caddo Parish. It is possible the VIL was being underestimated due to the proximity of the convection to the radar, resulting in higher levels of the storm lying in the "cone of silence."
Reflectivity data (Fig. 5b) indicated the complex remained fairly well defined at 0303 UTC, with the RIN containing reflectivity values of less than 15 dBZ. Just to the north of the bowing segment over northern Caddo Parish, there was some hint of an inflow notch developing on the southeast flank of an area where a comma head echo may have been trying to form. However, storm-relative velocity data for the time (not shown) indicated little in the way of rotation.
Base velocity data (Fig. 5d) showed significant weakening over previous volume scans with only a small area of 36 to 49 kt inbound velocity still intact, and no 50- to 63-kt inbound velocities indicated; however, it was just shortly after this time (0308 UTC) that a report of a 57-kt gust was received from northern Caddo Parish. Hence, rather than an indication of an actual decrease in the wind, this diminishing trend was likely attributable to the fact that the area of strong winds was now nearly due north of the radar site, and the actual wind was more perpendicular to the beam, resulting in less of the total wind vector being measured by the radar.
Weakening of the overall complex was clearly evident, however, 10 min later at 0313 UTC (not shown). Although the RIN could still be discerned--and in fact a second RIN may have developed to the west of the original RIN--neither feature was particularly well defined, and the bowing segment and LEWP configuration had definitely become less impressive. This trend continued over the ensuing 15 min to 30 min, and no severe weather was reported in NWSO Shreveport's local storm reports from this particular complex between 0308 UTC and 0400 UTC.
Several implications for the warning process can be derived from this case. First, radar data from 0228 to 0248 UTC show just how rapidly storms can change from fairly routine in nature to a classic bow echo/LEWP-type structure. In a case such as this where the synoptic pattern was clearly favorable for bow echo development, the radar operator must be especially aware of the possibility of such rapid development.
The actual bow echo development in this case seems to have been preceded by the development of the RIN and a large area of 36- to 49-kt inbound velocities. Such precursors can be an early warning sign to the radar operator of impending bow echo development, but only if the operator has a good feel for the synoptic and mesoscale environment and the type of development that the environment favors. The critical role that knowledge of the environment plays in the warning process cannot be stressed enough.
Once the bow echo develops, the warning decision often becomes easier. However, it must be noted that in this event the strong inbound base velocity weakened even while the bow echo was becoming increasingly well defined, and a report of a severe wind gust was being received. This was likely due to the angle between the actual wind and the radar beam becoming more perpendicular. Radar operators must be aware of this process taking place as storms change their position relative to the RDA, and may want to rely more on reflectivity pattern than on velocity data when this phenomenon is observed.
Finally, VILs were quite unimpressive throughout this event. This may have been at least partially due to higher elevations of the storms lying in the "cone of silence," especially once the line moved into Caddo Parish, Louisiana. In addition to the possible role the "cone of silence" played, Johns (1993) has discussed that objective reflectivity-based severe weather criteria often fail during low instability dynamic bow echo events. Others (for example, Pfost and Gerard 1996) seem to indicate this can be the case even in dynamic bow echo events in which greater instability exists. In situations such as this, reflectivity patterns are usually much more important than the height of the high reflectivity or the VIL.
Johns, R.H., 1993: Meteorological conditions associated with bow echo development in convective storms. Wea. and Forecasting, 8, 294-299.
Pfost, R.L. and A.E. Gerard, 1996: "Bookend vortex" induced tornadoes along the Natchez Trace. Wea. and Forecasting (accepted for publication).
Przybylinski, R.W., 1995: The bow echo: Observations, numerical simulations, and severe weather detection methods. Wea. and Forecasting, 10, 203-218.
Fig. 1. Synoptic composite chart for 0000 UTC September 21, 1996. Surface cold front indicated, along with positions of upper-level low (L), mid-level trough (dashed), and location and strength of 500 mb wind maximum (arrow).
Fig. 2. 0000 UTC September 21, 1996 upper-air sounding for Shreveport. Trace in upper left corner is vertical profile of theta-e.
Fig. 3. 0000 UTC September 21, 1996 hodograph for Shreveport. Winds are in knots.
Fig. 4. 0.5o base reflectivity and base velocity from KSHV (Shreveport) WSR-88D for 0228 UTC (a, c) and 0238 UTC (b, d), September 21, 1996. Arrows indicate areas of low reflectivity (a and b) and high inbound velocities (c and d).
Fig. 5. Same as Fig. 4, except for 0248 UTC (a, c) and 0303 UTC (b, d). Arrows point to rear inflow notches (a and b), and areas of high inbound velocities (c and d).