MESOs are essentially supercells (rotating thunderstorms)
TVS (tornadic vorticity signature) are supercells which exhibit on radar a tornadic signature (hook echo, shear parameters detected as sufficient enough by doppler radar)
Below from
http://www.wdtb.noaa.gov/resources/PAPE ... 99sec1.htm
Discriminating between thunderstorms likely to produce a tornado and those not likely to produce a tornado is very difficult. Operational meteorologists tasked with issuing tornado warnings are faced with many challenges. Many inputs, including radar signatures, spotter reports, and the mesoscale environment, play a part in the decision-making process. The most often used WSR-88D signatures in this process are a three-dimensional analysis of the magnitude of mesocyclone rotation in a storm, the strength of gate-to-gate shears, or presence of a Bounded Weak Echo Region (BWER) or hook echo. Many times, radar data has helped generate tornado warnings with significant lead time. However, occasionally tornadoes have occurred without warnings, or warnings based upon rotation have resulted in false alarms. The primary reasons for this are that the radar suffers from sampling limitations, and the tornado formation process is not yet well understood.
The information in this document is meant to provide meteorologists with some of the latest research findings related to the tornado warning decision-making process. This information is based on preliminary research results from the Verification of the Origins of Rotation in Tornadoes EXperiment (VORTEX) (Rasmussen et al. 1994), as well as the statistical analysis of a significant number of WSR-88D storm cases from a variety of sites across the U.S. as part of the mesocyclone and tornado detection algorithm research being performed at NSSL.
Research continues at the National Severe Storms Laboratory (NSSL) and elsewhere to improve the ability to discriminate between tornadic and non-tornadic storms. One of the primary goals of VORTEX (which was conducted in Central and Southern Great Plains) was to increase our understanding of why some storms produce tornadoes while others do not. While detailed analysis of the data is ongoing, some observations can be made and considered important guidance for the warning forecaster. A list of tornado warning considerations, including some gathered from analysis of the VORTEX data sets, will be presented in Section 2.
The WSR-88D Build 10 includes the new Tornado Detection Algorithm (TDA) (Mitchell et al. 1998) which has replaced the pre-Build 10 TVS Algorithm (hereafter "old TVS Algorithm"). This software upgrade represents a significant improvement in the way the original algorithm detects tornadic vortex signatures. For example, less stringent strength thresholds are used in the TDA than in the old TVS Algorithm, and a more robust vertical association scheme is used. In addition, the TDA will operate independently of the WSR-88D Mesocyclone Algorithm (88D-MA) and will not require the presence of a mesocyclone detection. Section 3 of this document includes information on the Build 10 TDA to supplement the Build 10 training guides.
NSSL has expended considerable effort to improve the performance of WSR-88D algorithms that are used in the tornado warning decision-making process. The NSSL Mesocyclone Detection Algorithm (MDA) (Stumpf et al. 1998) includes the ability to detect a broader spectrum of storm-scale vortices such as low-topped or mini supercells, and displays the algorithm output with a variety of color-coded tables, ranking attributes, and trend charts. In addition, the NSSL MDA has the capability to provide automated support to forecaster decision-making through the use of neural network probabilities. The NSSL MDA Algorithm will be a candidate for the first software upgrade after implementation of the WSR-88D Open Systems RPG (ORPG). The NSSL MDA is available right now via the WATADS software for testing on previously-collected WSR-88D data sets. NSSL also continues the experimental development of the TDA. This slightly more robust version of the Build 10 TDA, known as the NSSL TDA, is also available via WATADS. NSSL has also developed a BWER Algorithm which has been undergoing testing since 1998.
Section 4 of this document contains statistics from the NSSL MDA, NSSL TDA, and the NSSL BWER Algorithm on a large data set of tornado-producing storms as well as some non-tornadic supercell events. The location and storm type for the 43 days which comprised the statistical radar guidance data base are given in Table 1. This 43-day data set is larger, by 14 cases, than the data set used to develop the statistics in the 1st Edition of the Tornado Warning Guidance document (Spring 1997). This expanded 43-case data set is more geographically (and synoptically) diverse, and contains eight new "null" cases (no tornadoes reported).
Section 5 summarizes this document. With the delivery of ORPG to the field in two to three years, radar assistance in the detection and warning of tornadoes should be improved. For the near term and even beyond any known improvements, meteorologists are encouraged to continue using ALL possible information at their disposal in order to issue successful tornado warnings. This includes spotter reports, storm history, mesoscale environment, and storm/vortex evolution and trends (including reflectivity trends).
a) Supercells tend to produce significant tornadoes in regions with enhanced near-ground storm-relative helicity. In many situations, enhanced low-altitude helicity will be associated with backed and strengthened surface winds. All available low-altitude wind data should be monitored, including routine surface observations, mesonetwork data, and lowest-tilt radial velocity. Mesoanalysis is very important. The near-storm environment can vary dramatically over fairly short distances and is subject to rapid change (this was observed on many VORTEX storm days; see Markowski et al. 1998)
b) Because of baroclinic effects along shallow boundaries, the immediate cool side of the boundary is often an area of strongly enhanced horizontal vorticity. Even lacking wind and temperature data, the mere presence of a boundary should lead to heightened awareness, and storms crossing or interacting with boundaries merit special scrutiny for rapid increases in rotation in their lower altitudes. This implies that forecasters need to remain aware of the locations of radar fine lines, satellite-indicated cloud lines, and mesonet-detected surface temperature gradients and wind-shift lines. Storms interacting with boundaries should be closely monitored because tornadogenesis can occur rapidly.
c) The chances for significant tornadoes on the cool side of boundaries decrease as low-altitude cold air becomes increasingly deep and CAPE approaches zero. These are difficult to assess in real time, but the key fact is that if the air still contains some CAPE, despite being relatively cool, the potential for significant tornadoes exists; the potential is greatest where the cool air is shallowest and the enhanced helicity resides near the ground (the immediate cool side of a boundary, where the temperature difference between the two air masses is still rather small).
d) A high quality spotter network is vital. Whenever possible, one should not rely solely on radar data for making warning decisions, as even storms having strong low-altitude mesocyclones may not produce a tornado (this was observed several times during VORTEX). Warning forecasters should make use of all available information including reliable spotter reports (especially when storms are at far ranges), storm history, other remote sensing tools (e.g., satellite), as well as having a good understanding of the mesoscale environment of the storm.
Be aware that some storms may produce tornadoes rapidly with little advance warning in the way of algorithm detections or rotation at the 0.5 volume scan. Often, the only low-altitude precursor from radar in these situations is an area of strong, low-altitude (0-2 km AGL) convergence (Burgess and Magsig 1998) below the base of the organizing mesocyclone (remember that 0-2 km AGL information is only observable out to about 65 nm). Also, second and succeeding mesocyclone cores (cyclic mesocyclogenesis) typically have very short organizing stages as they quickly form over a large depth and strengthen rapidly. Therefore, explosive development can take place during the period of a single volume scan. The opposite (rapid dissipation) was also observed during VORTEX.
f) Not every TVS forms at mid altitudes and builds downward over time with the embryonic tornado. Trapp et al. (1998) observed that some TVSs develop rapidly near the surface or simultaneously at low and mid altitudes, especially in squall-lines (but also in some supercells). Be aware of both types of TVS development, and anticipate low-altitude development with squall lines.
g) Storm motion and tornado motion (direction and speed) may be significantly different. For example, on two VORTEX days (6/2/95 and 6/8/95), there were several instances where the parent thunderstorm was moving toward the northeast while the tornado was moving north. In addition, for another case, the tornado's forward movement was measured at 60 mph only to become nearly stationary before it dissipated. (Learning more about the reasons for changes in tornado motion will be a topic of future VORTEX research). Be careful about issuing tornado warning locations based on the storm cell centroid motions; use the motion of the radar vortex signature, whenever available.
h) In many instances, the radar-observed vortex signature can, depending on range, appear to dissipate prior to the actual dissipation of the tornado, as the shrunken tornado vortex (or tornado cyclone) becomes increasingly difficult to observe given WSR-88D sampling limitations. This period without a radar-observable vortex signature may include the most intense and damaging phase of the tornado. It is a good rule of thumb to continue tornado warnings for a few volume scans following the dissipation of the radar-observed vortex signature, especially in the absence of reliable spotter information and/or during nighttime hours.
i) Data collected during VORTEX using the Doppler On Wheels (DOW), and data from a variety of WATADS-analyzed WSR-88D cases, verify that a variety of vortex scales occur within storms, ranging from the scale of the actual tornado (and even its sub-vortices), up to the scale of the rotating updraft/downdraft of the supercell storm (mesocyclone), with vortices intermediate to these scales also occurring (sometimes referred to as the tornado cyclone). Some data suggest that these vortices may be embedded within each other, or that some vortices may taper or widen in diameter at different heights. Radar users should be aware that the WSR-88D, with its inherent sampling limitations, may detect a mixture of these kinds of vortices. Operators should also be aware that only in very rare instances can the WSR-88D actually observe the actual tornado, again, owing to the sampling limitations of the radar (the tornado must be very large and/or very close to the radar). In most instances, a TVS is actually the signature of an intermediate-scale vortex, observed as a gate-to-gate velocity couplet. See this presentation by Stumpf (1998) for figures.
j) Radar-observable vortex signatures which are associated with tornadoes can occur with a variety of storm types. These range from the classic Great Plains supercell (with large horizontal and vertical extent) as well as supercells with small horizontal extent (mini supercells), supercells with small vertical extent (low-topped supercells), or both (low-topped mini supercells). Tornadoes and radar-observable vortex signatures have also been observed with storms embedded within tropical cyclone rain bands ("TC-mesos"), along the leading edge and comma head of bow-echo squall lines, and with rapidly-developing convection (non-supercell tornadoes, landspouts, waterspouts). Do not be misled into believing that all supercells are the same - like the classic big isolated supercells more common to the Central and Southern Plains. Be aware that many varieties exist, including some that probably have not yet been observed. NSSL maintains a WSR-88D tornado case-study Web page that contains the description (with figures) of a number of these typical and atypical tornadic storm cases.
k) Because the WSR-88D provides only discrete horizontal samples of the atmosphere (1 azimuthal resolution; 1 km and 250 m range resolution for reflectivity and velocity respectively), storm-scale vortices can only be depicted in a degraded sense (Wood and Brown 1997). Factors include vortex core diameter to beam width radius ratio, strength of rotation in the vortex, and the offset between the vortex centroid and the centroid of the radar beam. A particular vortex of a given diameter and rotational velocity could be viewed by the radar in a number of configurations given its range from the radar and the vortex/beam centroid offsets. And, if a vortex is shaped asymmetrically, changes in viewing angle will also alter its radar depiction. Consider that these sampling limitations will reduce the velocity estimate of the vortex. Consult Wood and Brown (1997) for information depicting the degree of velocity degradation in radar-sampled vortices.
l) At extended ranges, the radar horizon prevents sampling below the mid-altitudes of mesocyclones. Thus, the radar may observe mid-altitude rotation that is strong for storms at extended ranges, but the radar cannot determine if the low-altitude rotation is strong or even exists. Users should employ the use of spotter reports, or data from another radar sampling the signature from a closer range. At near ranges, the "cone-of-silence" effect will prevent sampling of vortices above a certain altitude, and only a portion of the vortex can be diagnosed for warnings. Forecasters should use data from other WSR-88Ds at farther ranges to sample the mid- and high-altitude data being missed in the cone-of-silence.
m) Many algorithm-detected radar-observable vortex signatures (both mid-altitude and low-altitude) are NOT associated with tornadoes on the ground. Bear in mind that in some instances, atmospheric vortices can be too small (owing to sampling limitations), or hidden by radar data artifacts (such as range-folded data). Radar algorithms cannot detect these unobservable vortices. Also, some vortex detections may be the result of dealiasing errors, leading to false detections. The user should examine velocity images along with algorithm output at all times.
n) When issuing a warning based on radar, remember the total time involved includes: viewing and analyzing the radar vortex signature (this can take anywhere from 1 minute to 6 minutes if you are using algorithm products as guidance, as they are generated at the end of a volume scan), mechanically composing the warning message (2-3 min., or 1 min if using AWIPS), and disseminating the warning (1 min. or more). With the possible lapse of 3 to 10 minutes of time, the location of the mesocyclone or Tornadic Vortex Signature (TVS) that triggered the decision to issue the warning could have moved a considerable distance. Thus, this translated distance of the signature needs to be taken into account when locations are mentioned in the warning (especially when using algorithm overlays for location guidance). This translated distance also needs to be considered for warnings in downstream counties.
o) Adaptable parameter sets are being provided with the Build 10 TDA that correspond to a variety of storm types. It is important to understand that storm types are a factor of the storm's mesoscale or near-storm environment (NSE), and NOT due to the region of the U.S. that the storm is occurring. The NSE should be closely monitored during warning operations so that the proper adaptable parameter sets are always used. Keeping adaptable parameter settings at some site-selected default value because of regional expectations of a certain storm type (i.e., mini supercells always occur in the Northeast) may result in poor algorithm performance if the prevailing NSE does not correspond to the default settings. For example, if an NSE supportive of large and tall ("Oklahoma-style") supercells is occurring in New York, use the TDA adaptable parameter developed for these types of storms.
p) Based on algorithm studies and field surveys, the OSF now offers improvement to performance of the current WSR-88D Mesocyclone Algorithm (88D-MA) by allowing local radar sites the option to lower the Threshold Pattern Vector (TPV) (1). This change appears to improve the probability of detecting tornadic vortices and increases the lead times for warnings in certain storm situations. (2)
q) Guidelines for warnings based on established thresholds for shear have limitations at far ranges. A series of mesocyclone strength nomograms (which display rotational velocity as a function of range) for three different vortex diameters are given in Figure1, Figure 2, and Figure 3. The variation of the slopes of the lines on the nomograms illustrate this range-dependency problem. The slopes become progressively steeper for the 2 nm and especially the 1 nm diameter nomograms. Meteorologists who use these nomograms should recognize that there can be considerable overlap in the strength categories for a given rotational velocity for different types of storms (smaller diameter mini supercells, for example).
