Using WSR-88D Shear Products During Severe Storm Events
George R. Wilken
NWSFO Little Rock, Arkansas
Although the WSR-88D provides excellent support for warnings through many different products, severe weather events push the equipment and personnel to the maximum. Good radar observing techniques demand that more than one product or technique be used to verify what is suspected. Trends are very important and should be watched closely from volume scan to volume scan for all products used. This paper examines shear products.
Shear. In using WSR-88D shear products (calculated VR and the shear value in the alphanumeric [A/N] meso [M] product), priorities can be set during the observation of the storm signatures. These priorities allow warning personnel to concentrate on the locations of mesocyclones that would likely produce tornadoes. Other than observing on the A/N M product what the base of the mesocyclone is, it is difficult to determine if a tornado touchdown has occurred without spotter confirmation. In addition to identifying significant storms, good lead times and a low false alarm ratio (FAR) are obtained by using shear products.
Combined Shear. Most tornadoes in the U.S. are ranked at the lower end of the Fujita intensity scale (F0 - F2). Often these tornadoes are not detectable until one or two volume scans before touchdown. In most cases, personnel delay issuing a warning because they note only weak velocity and shear signatures. For that reason, such events may go unwarned or have little warning lead time. On the other hand, the Combined Shear (CS) product has the potential to detect these weak events (within limits) and provide some warning lead time in these very difficult situations.
This paper examines the various WSR-88D shear products and suggests some operational applications for use during both strong and relatively weak events. Since the intent is to focus on WSR-88D presentations and the interpretation of products, atmospheric conditions preceding the events will be treated only superficially.
2. Rotational Velocity (VR) and Shear Values
The WSR-88D software allows the PUP operator to calculate a VR/Shear value. This product is available by positioning the cursor and clicking on the opposite values of highest velocity, and utilizing the shortest distance between these two values. The correct way to select points for calculation of the values is shown in Fig. 1 (NWS 1997). Once calculated, the values appear in the lower right-hand corner of the velocity product that was selected.
The A/N meso product also lists shear values in the far right-hand column of the product, although the values listed may be somewhat different from those calculated for the on-screen VR product. The difference is related to how the operator selects points on-screen.
At NWSFO Little Rock, it has been empirically determined that VR values approaching the 40-45 kt range, and shear values approaching .020 sec-1 (shown as E-3/s on the A/N meso screen), are good base values to define the formation of a tornado. The NEXRAD Operational Support Facility (OSF) has determined a comparable set of values as noted below.
Shear values are obtained by adding the absolute values of the maximum inbound and outbound values, dividing by the distance between them, and then converting from units of shear per hour to units of shear per second. Obviously, as the distance between the maximum inbound and maximum outbound velocities decreases, the shear value will increase.
The current mesocyclone strength nomogram assumes a circulation diameter of 3.5 nm. In a March 17, 1997 memorandum to all WSR-88D sites (Burgess 1997), mesocyclone nomograms portray the strength down to a 1 nm diameter, so the diameter must be taken into account when assessing the strength of any mesocyclone. The OSF and NSSL have developed the following threshold shear values for guidelines:
Less than 70 nm More than 70 nm
For minimum mesocyclones: .005/sec .003/sec
For strong mesocyclones: .020/sec .010/sec
These values were valid for the March 1, 1997 tornadic outbreak in Arkansas, as seen in Tables 1 and 2. For reference, the WSR-88D algorithms default to the following low shear thresholds (the mesocyclone algorithm also uses momentum):
Mesocyclone Shear: 7.2/hr = .002/sec
TVS Shear: 72/hr = .02/sec
To calculate rotational velocity (VR), take the absolute value of the difference between the highest inbound and highest outbound velocities in the circulation and divide by 2. The rotational velocity remains the same regardless of how far apart the two values are.
3. The Combined Shear (CS) Product
This product is useful in weakly sheared environments. Six cases have been examined at NWSFO Little Rock. Three of those cases are shown later in this paper. Most WSR-88D sites have observed mesocyclone alerts being generated along approaching squall lines, or have had F0 or F1 events that were difficult, if not impossible, to detect using velocity or other observed products or parameters. Therefore, when weak shears are expected, such as with an approaching cold front, put the CS product on the Routine Product Set (RPS) list and used it in conjunction with other items such as Spectrum Width (SW).
Using the WSR-88D Build 9.0 software, the CS product is user-selectable at one elevation cut; and the 1.5 deg elevation cut is recommended. This product is useful generally within 75 nm of the radar. Storm configuration is better defined by the CS product and therefore is easier to interpret. In addition, a threshold value of 80 units of shear, especially in a cluster or group, or showing up as part of a trend upward in the suspect area, is indicative of possible intensification.
The CS product is very "noisy," and some experience is needed before using it. The product requires somewhat intensive WSR-88D central processing unit (cpu) usage due to the calculations involved, so, some loadshedding must be anticipated when the product is produced. The Little Rock WSR-88D has been set for Volume Coverage Pattern (VCP) 11 and a 50 product RPS list on a few occasions and has shown some loadshedding, but none that could not be tolerated. A 40 product RPS list is recommended, with a transition to a 50 product RPS list (at Unit Control Position (UCP) sites) as experience dictates.
4. Utilizing WSR-88D Shear and Meso Products (M) During the March 1, 1997, Severe Weather Outbreak in Arkansas
Although it appeared March 1 would see an outbreak of severe weather in Arkansas, the magnitude of the event was not expected from the severe weather parameters shown on the 1200 UTC Little Rock sounding (Fig. 2). Instability was significant at -5C for both the 300 mb and 500 mb lifted indices, with the Showalter Index at -4C. Especially noteworthy was the Energy/Helicity Index (EHI) which showed a value of 1.99. At Little Rock, EHI values in excess of 1.0 might cause one to expect severe weather, and the higher the index, the more significant the event. Davies (1993) suggests that values below 2.0 indicate significant mesocyclone-induced tornadoes are unlikely to occur, and values from 2.0 to 2.4 indicate mesocyclone-induced tornadoes are possible, but unlikely to be strong or long-lived. Even when the 1200 UTC sounding was modified, this guideline by Davies did not appear valid for this day.
Storm motion from the SHARPII program was 250 deg at 34 kt. When the sounding was modified for the forecast maximum temperature and other parameters, the air was more stable and the EHI lower, remaining somewhat above a value of 1.0. Storm motion changed to 225 deg at 55 kt, which proved to be a good forecast as the severe weather developed and moved northeastward. The "VIL of the Day" was set at 52 with a 500 mb temperature of -12.6C. The 1200 UTC Little Rock hodograph showed a straight-line configuration which indicated that neither right or left-moving thunderstorms were favored.
The midshift Lead Forecaster, seeing that some favorable parameters existed and noting the Storm Prediction Center (SPC) forecast of moderate risk of severe for Arkansas, called in an extra forecaster to work the PUP at about 1300 UTC. Convection began to strengthen around 1800 UTC, and therefore, a second forecaster was called in around 1930 UTC to help with the event. The convective activity quickly increased in both coverage and intensity shortly after the second forecaster arrived. The radar was immediately put in VCP11 to better monitor the storms.
Convective Development and Warning Technique
Since the activity quickly became widespread, the PUP A/N meso product was used to identify storms of note and watch mesocyclones in the developmental stage. A threshold shear value of 20 x 10-3 sec-1 was used as previously mentioned. Whether the meso product identified the circulation as uncorrelated shear, 3-D correlated shear, or a Meso, the circulation was watched closely from volume scan to volume scan for deepening of the circulation and lowering of the base of the mesocyclone. Close attention was paid to the increase in shear and decreasing diameter.
Note in Figs. 3a,b the change in the storm identified as "D0" from 2022 to 2027 UTC. In five minutes the storm went from uncorrelated shear with a shear value of 7 units, to a mesocyclone with a shear of 9 units. When a storm such as "D0" appeared to be developing into a tornado, the location was quickly passed along to the PUP operator, who then looked closely at the velocity values of the Storm Relative Motion (SRM) product, looked at Spectrum Width (SW) and also calculated a VR/Shear set of values. This method of "spotting" from the PUP A/N terminal allowed the PUP operator to focus on the more important storms and not be distracted by any one storm. This procedure also helped to reduce the false alarm ratio (FAR) by identifying those storms for which warnings were not immediately required.
It is important to note that the A/N meso product identified a storm close-in to the radar which had not been seen using only the PUP graphic products. This storm produced a Tornadic Vortex Signature (TVS) within two volume scans after identification and issuance of a tornado warning (Figs. 4a,b). The storm was primarily identified through a "first-look," that is, a routine investigation of all storms and checking all elevation angles via a User Function (UF) when first sitting down at the PUP. Without this action, the storm could have easily become tornadic with no warning.
TVS signatures were produced in the Arkadelphia to Little Rock tornado only when the activity was within about 30 nm of the radar. Values of Threshold Pattern Vector (TPV) and TVS Threshold Shear (TTS) are set at the default values at Little Rock, but will likely be changed when several storms, including the March 1 outbreak, are examined later using WATADS (the WSR-88D Algorithm Testing and Display System).
Values of VR, shear, distance from the radar, diameter of the meso and azimuth and range coordinates are seen as Tables 1 and 2. Although range from the radar affects the net shear value, Table 1 shows how the shear values varied as the storm moved toward Little Rock. It is interesting to note that as the storm moved into southern Saline County, it had a shear value of 0.39 sec-1 and a diameter of 0.5 nm. On the very next volume scan, the shear had dropped to .009 sec-1 and the diameter had increased to 2.4 nm. The storm survey showed a break in the damage path at this time, from northern Hot Spring County to southern Saline County. On the next five volume scans, the shear steadily increased and the diameter gradually decreased as the tornado strengthened to an F4 in northern Saline County, with a shear value of .102 sec-1 (Fig. 5).
As seen in Table 2, for the storm that developed near the NWS office in North Little Rock, shear values diminished two volume scans after this tornado moved into White County. A damage survey showed that a break in the path also occurred at this time. The tornado then strengthened at 2107 UTC and the shear increased dramatically. The damage path resumed and continued for
62 mi. Data on the storms for March 1, 1997, may be observed on the NWSFO Little Rock homepage at http://www.srh.nws/ftproot/lzk/html/.
One of the most difficult aspects of observing tornadic storms with the WSR-88D is determining whether or not a tornado is on the ground. The observations for March 1 provided a basis for how shear might be used in this determination. For the time being, it is recommended that no assumptions be made when observing a decrease in shear, unless verified by other products.
Another product used to verify the more active areas was the Base Spectrum Width (SW). The SW product showed "clusters" of higher values, with a high value core, wherever the tornadic storms were occurring (Fig. 6a,b). This provided a cross-check for the other products used. At the National Weather Association Annual Meeting in December, 1996, Les Lemon recommended use of the SW product for detection of vortices and shear within the beam.
General Comments on the March 1 Event
Warnings for the March 1 event were issued using the WISE-II applications software (Jackson 1996). The latest available data should be used with this program to issue the warning, and in fast-moving situations, care should be taken that all locations and times in the warning have not already been observed. County coordinators after the March 1 event specifically noted the usefulness of the location arrival times in the warnings that were issued.
At NWSFO Little Rock dissemination of the warning on the NOAA Weather Radio is speeded by having the NWA operator stand close to the warning desk so data can be entered quickly on a form. Counties are also called in advance to provide a "heads-up" when the storm is moving their way, or to provide the recently issued warning on a more timely basis. When the tornado approached the Little Rock National Airport (Adams Field), tower personnel were alerted by telephone of its approach.
5. The Use of Combined Shear
The Killer Tornado of November 11, 1995
a. General Conditions
The approach of a prefrontal squall line during the night of November 11, 1995, provided a good opportunity, in retrospect, to look at the use of the Combined Shear (CS) product. As the squall line moved through central Arkansas, distractions were provided to the PUP operator by several shear alerts (uncorrelated shear) along the leading edge of the line. In addition, at least two TVS alerts to the south of the eventual tornado touchdown drew attention away from the developing tornado. Forecasters on duty that night issued a severe thunderstorm warning for several counties, including Prairie County where the tornado eventually touched down.
As the line approached the northwest portion of Prairie County (Fig. 7a, upper left), it appeared that a bow echo was developing. The CS product showed the gradual bowing and subsequent "circulation wrapup" two volume scans before the tornado touched down (Fig. 7a). This might have provided at least five minutes of lead time for this F2 tornado. The severe thunderstorm warning provided about ten minutes of lead time, and since the area was under a tornado watch the warning included the statement "...tornadoes can and do frequently occur..." for the warned area.
The tornado touched down (Fig. 7b, upper left) north of the community of Des Arc in northern Prairie County. It destroyed a home and killed two adults. An infant was tossed several hundred yards into an open field, but escaped serious injury. The tornado was on the ground for no more than ten minutes, moving to the northeast and lifting before moving into an adjacent county. The CS product showed the entire sequence from beginning to end (Figs. 7a,b).
b. Utilizing the CS Product
Two basic observations may be made in using the CS product. Although as mentioned previously it is rather noisy, the CS product isolates the combined azimuthal and radial shear and provides the observer with an image free of conflicting echo patterns, so it may divulge suspicious configuration as it did on November 11. In this particular situation, it was empirically determined that 100x10-4 sec-1 (or 100 units of) shear was the danger point; that is, a warning might have been in effect before that point was reached. This value of shear, along with configuration, produced the severe weather.
As the tornado touched down the peak shear was seen in the area of rotation. The lower left frame of Fig. 7a shows the storm about five minutes before tornado touchdown. At 0509 UTC (Fig. 7b) the circulation intensified and showed a core of 100 units of shear. An observation of the trend of intensification of shear, along with the development of the rotation, appear to have been signals that severe weather was imminent.
c. Spectrum Width
A glance at the SW product during this episode showed that values increased one volume scan before tornado touchdown, so although this product indicated a location that should be examined, it did not produce a longer lead time than the CS product.
The Bow Echo Storm of January 18, 1996
a. General Conditions
On January 18, 1996, an approaching cold front produced a line of thunderstorms and almost continuous meso alerts, primarily for uncorrelated shear. This type of situation is very distracting, with the operator generally not knowing if any of the alerts are to be believed. The line of thunderstorms developed into a bow echo which caused significant damage in Conway (south central Faulkner County). Some minor injuries were reported, along with damage to mobile homes. At least one aircraft was destroyed at the airport and other planes were severely damaged.
b. Utilizing the CS Product
Values of 90 units (90 10-4sec-1) of shear were located just to the north of the bowing line and the CS product provided a good view of the line's configuration and evolution. Figure 8 shows a four-panel view of the line during the most active part of its lifetime, as it approached Conway and moved across the area. The arrows in Fig. 8 identify where the strongest shear of 90 units was located just north of the wave crest. After the line moved across Conway, it generally decreased in strength. Although the line was still bowing, the values of shear dropped to less than 90 units.
This case suggests that with strong configuration, a value of 80 units of shear might be considered a threshold for warning action. The bowing of the line became evident in the imagery before the line entered Faulkner County (location of Conway). This might have provided the PUP operator with ample lead time to issue a severe thunderstorm warning, or at least flag this area as one to watch for a possible warning.
The Bow Echo Storm of March 25, 1996
a. General Conditions
A line of thunderstorms moved toward the Little Rock metropolitan area, and Meso alerts (primarily uncorrelated shears) were produced which taxed the PUP operator with much information to be evaluated. As the line moved into the Little Rock area winds of 65 to 78 mph were reported in Faulkner and Pulaski Counties, with the highest wind reported at the Little Rock Air Force Base. Trees and lines were down and some roof damage was reported. A hangar door was pushed off its track at Adams Field in Little Rock.
b. Utilizing Combined Shear
Both shear and configuration produced some concern as the line moved toward Little Rock. Approximately four pixels of shear of 100 units appeared in the 0124 UTC image (Fig. 9, upper left). Just five minutes later, the shear had increased in areal coverage and was located to the north of the bow. This area of shear was coincident with the damage that was reported over the central and northern portion of Pulaski County. In utilizing the CS product for this particular case, the PUP operator would observe the bowing configuration of the line and the trend upward in shear values. Finally, the increase to 100 units of shear on the 0124 UTC image showed that the storm had become severe. Note the double arrows in the upper right and lower left quadrants of Fig. 9 which identify high shear values clustered and aligned along the bowing line. As the line continued eastward the area of shear diminished and the line did not subsequently produce damage in the counties to the east of Little Rock.
Through the use of WSR-88D products (some routinely provided) PUP operators may identify particular radar echo patterns which may develop into severe storms. In particular, correlation of developing mesocyclones with shear values, both on the alphanumeric Meso product and by calculating VR, can provide longer lead times. When numerous echoes are present, such as the March 1, 1997, outbreak in Arkansas, this technique allows the PUP operator to focus on the more dangerous storms. Threshold shear values developed by the OSF should be used as a guideline when utilizing shear as an indicator.
Combined Shear, which can be placed on the Routine Product Set (RPS) list, can be useful in weakly sheared events to show both configuration of the echo pattern and strength of the shear. Recommended uses that have been described in this paper are with the various types of shear that accompany an approaching cold front, bow echoes, and for single or multiple-cell thunderstorms that look suspicious.
Recognition and use of the various types of shear may perhaps lead to a rising Probability of Detection (POD) and a lowering False Alarm Ratio (FAR).
Burgess, D.W., 1997: Tornado Warning Guidance. Contained in a memo to all WSR-88D sites, OSF/OTB, Norman, OK, 28 pp.
Davies, J.M., 1993: Hourly Helicity, Instability and EHI in Forecasting Supercell Tornadoes. Preprints, 17th Conf. on Severe Local Storms, St. Louis, MO, Amer. Meteor. Soc., 107-111.
Jackson, G. E., 1996: WISE Warning Locator v4.50. Computer applications program, NWSO San Angelo, TX.
National Weather Service, 1997: Some Guidelines for Interpreting the VR/Shear Function, NEXRAD Operational Support Facility, Norman, OK, 3 pp.