• Benjamin, S. G., and Coauthors, 2002: RUC20—The 20-km version of the Rapid Update Cycle. NWS Tech. Procedure Bull. 490, 29 pp.

  • Benjamin, S. G., and Coauthors, 2004: An hourly assimilation–forecast cycle: The RUC. Mon. Wea. Rev., 132 , 495518.

  • Bluestein, H. B., 1985: The formation of a “landspout” in a “broken-line” squall line in Oklahoma. Preprints, 14th Conf. on Severe Local Storms, Indianapolis, IN, Amer. Meteor. Soc., 267–270.

  • Bothwell, P. D., , Hart J. A. , , and Thompson R. L. , 2002: An integrated three-dimensional objective analysis scheme in use at the Storm Prediction Center. Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., J117–J120.

  • Brady, R. H., , and Szoke E. J. , 1989: A case study of nonmesocyclone tornado development in northeast Colorado: Similarities to waterspout formation. Mon. Wea. Rev., 117 , 843856.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Burgess, D. W., , Donaldson R. J. Jr., , and Desrochers P. R. , 1993: Tornado detection and warning by radar. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geophys. Union, 203–221.

    • Crossref
    • Export Citation
  • Caruso, J. M., , and Davies J. M. , 2005: Tornadoes in nonmesocyclone environments with pre-existing vertical vorticity along convergence boundaries. Natl. Wea. Assoc. Electron. J. Operational Meteor., 2005-EJ4 [Available online at http://www.nwas.org/ej/.].

    • Search Google Scholar
    • Export Citation
  • Colby, F. P., 1984: Convective inhibition as a predictor of convection during AVE-SESAME-2. Mon. Wea. Rev., 112 , 22392252.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Craven, J. P., , and Brooks H. E. , 2004: Baseline climatology of sounding-derived parameters associated with deep moist convection. Natl. Wea. Dig., 28 , 1324.

    • Search Google Scholar
    • Export Citation
  • 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.

  • Davies, J. M., 2004: Estimations of CIN and LFC associated with tornadic and nontornadic supercells. Wea. Forecasting, 19 , 714726.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies, J. M., , and Johns R. H. , 1993: Some wind and instability parameters associated with strong and violent tornadoes. 1. Wind shear and helicity. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geophys. Union, 573–582.

    • Crossref
    • Export Citation
  • Davies-Jones, R. P., , Burgess D. , , and Foster M. , 1990: Test of helicity as a tornado forecast parameter. Preprints, 16th Conf. on Severe Local Storms, Kananaskis Park, AB, Canada, Amer. Meteor. Soc., 588–592.

  • Doswell C. A. III, , , and Rasmussen E. N. , 1994: The effect of neglecting the virtual temperature correction on CAPE calculations. Wea. Forecasting, 9 , 625629.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hart, J. A., , and Korotky W. , 1991: The SHARP workstation v1.50 user’s guide. NOAA/National Weather Service, 30 pp. [Available from NWS Eastern Region Headquarters, 630 Johnson Ave., Bohemia, NY 11716.].

  • Johns, R. H., , Davies J. M. , , and Leftwich P. W. , 1993: Some wind and instability parameters associated with strong and violent tornadoes. 2. Variations in the combinations of wind and instability parameters. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geophys. Union, 583–590.

    • Crossref
    • Export Citation
  • Lee, B. D., , and Wilhelmson R. B. , 1997: The numerical simulation of nonsupercell tornadogenesis. Part II: Evolution of a family of tornadoes along a weak outflow boundary. J. Atmos. Sci., 54 , 23872415.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, B. D., , and Wilhelmson R. B. , 2000: The numerical simulation of nonsupercell tornadogenesis. Part III: Parameter tests investigating the role of CAPE, vortex sheet strength, and boundary layer vertical shear. J. Atmos. Sci., 57 , 22462261.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , Straka J. M. , , and Rasmussen E. N. , 2002: Direct surface thermodynamic observations within rear-flank downdrafts of nontornadic and tornadic supercells. Mon. Wea. Rev., 130 , 16921721.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moncrief, M., , and Miller M. J. , 1976: The dynamics and simulation of tropical cumulonimbus and squall lines. Quart. J. Roy. Meteor. Soc., 102 , 373394.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pietrycha, A. E., , and Manross K. L. , 2003: WSR-88D analysis of vortices embedded within a surface low pressure trough and subsequent convection initiation. Preprints, 31st Int. Conf. on Radar Meteorology, Seattle, WA, Amer. Meteor. Soc., 835–838.

  • Pietrycha, A. E., , and Rasmussen E. N. , 2004: Finescale surface observations of the dryline: A mobile mesonet perspective. Wea. Forecasting, 19 , 10751088.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., 2003: Refined supercell and tornado forecast parameters. Wea. Forecasting, 18 , 530535.

  • Rasmussen, E. N., , and Blanchard D. O. , 1998: A baseline climatology of sounding-derived supercell and tornado forecast parameters. Wea. Forecasting, 13 , 11481164.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Thompson, R. L., , Edwards R. , , Hart J. A. , , Elmore K. L. , , and Markowski P. , 2003: Close proximity soundings within supercell environments obtained from the Rapid Update Cycle. Wea. Forecasting, 18 , 12431261.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., , and Wilson J. W. , 1989: Non-supercell tornadoes. Mon. Wea. Rev., 117 , 11131140.

  • Wakimoto, R. M., , and Atkins N. T. , 1996: Observations on the origin of rotation: The Newcastle tornado during VORTEX 94. Mon. Wea. Rev., 124 , 384407.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    Tornadoes in (a) a high LCL environment over northwest KS on 29 Jun 2000 near Bird City, KS (photo courtesy J. Reed Photography) and (b) a small SRH and high LCL environment over south-central KS on 27 Aug 2004 near Wellington, KS (photo courtesy R. Renfro).

  • View in gallery

    Lower portion of a skew T–logp diagram of a RUC analysis sounding associated with the high-based tornadic supercell on 29 Jun 2000 shown in Fig. 1a. Temperature profiles (thick solid curves), dewpoint profiles (thick dashed curves), lifted parcel ascents above LCL using a mixed-layer lowest 100-hPa parcel (heavy dotted curves), and positive CAPE (hatched) are shown with important features labeled. For viewing simplicity, the virtual temperature correction is not shown.

  • View in gallery

    As in Fig. 2, but for RUC analysis profiles associated with high-based nontornadic storms that developed on preexisting boundaries in (a) southeast CO on 8 Jun 2004 and (b) west-central TX on 19 May 2003.

  • View in gallery

    Box-and-whisker plots of 0–2-km lapse rate (°C km−1) for F1 and greater intensity tornadoes (338 total cases) in the RUC sounding database used for this study. These are categorized by (a) small compared with typical values of 0–1-km SRH values as defined in the text and (b) high compared with typical LCL heights as defined in the text, using a mixed-layer lowest 100-hPa parcel. Boxes represent the middle 50% of cases in each category, whiskers extend to the 90th and 10th percentiles, and horizontal bars indicate median values.

  • View in gallery

    Box-and-whisker plots showing (a) LFC height (m AGL), (b) CIN (J kg−1), and (c) 0–3-km CAPE (J kg−1) categorized by selected nontornadic and F1–F4 tornadic cases from RUC sounding database in this study, using mixed-layer lowest 100-hPa parcels. All these cases (314 total) had steep 0–2-km lapse rates (≥7°C km−1) along with small SRH and/or high LCL heights as defined in the text. Conventions are as in Fig. 4. Tornadoes of F0 intensity are not shown.

  • View in gallery

    Parameter fields of (a) 0–1-km SRH (m2 s−2), (b) 0–3-km lapse rate (°C km−1), and (c) LFC height (m AGL) from the SPC mesoanalysis at 2300 UTC 27 Aug 2004 over the KS–OK–MO area. Surface features are shown in (a) along with selected parameter values at Winfield, KS, near the surface boundary intersection. Wind barbs in (a) estimate storm motion in knots. Heavy dots in (b) denote axes of steepest lapse rates. (d) The areas where 0–3-km lapse rates ≥8°C km−1 overlap LFC heights ≤2000 m AGL, and tornado tracks and intensities for 0000–0100 UTC 28 Aug 2004.

  • View in gallery

    (a), (b) Video captured by the author showing tornadoes in south-central NE on 24 May 2004. Several tornadoes, including the one in (a), had nonmesocyclone visual characteristics as described in text. The tornado on the right in (b) had mesocyclone visual characteristics, including a rear-flank downdraft (indicated by RFD) and a wall cloud.

  • View in gallery

    Parameter fields of (a) 0–1-km SRH (as in Fig. 6a, with surface features), (b) 0–3-km lapse rate (as in Fig. 6b), and (c) 0–3-km CAPE (heavy lines, J kg−1) from SPC mesoanalysis at 2000 UTC 24 May 2004 over the NE–KS–MO area. Surface vertical vorticity (thin lines, 10−4 s−1) is also shown in (c) with times signs indicating the location of significant vorticity maxima. Wind barbs in (c) indicate surface wind direction and speed (kt). (d) The area where 0–3-km lapse rates ≥8°C km−1 overlap 0–3-km CAPE ≥ 50 J kg−1, and tornado tracks and intensities (F1+ intensities labeled) at 2030–2230 UTC.

  • View in gallery

    Parameter fields of (a) 0–3-km lapse rate (as in Figs. 6b and 8b) and (b) 0–3-km CAPE and surface vorticity (as in Fig. 8c) from SPC mesoanalysis at 1800 UTC 8 Jun 2004 over eastern CO. Surface features are also shown in (b). (c) Composite radar reflectivity at 1926 UTC with relevant features labeled.

  • View in gallery

    Box-and-whisker plot (conventions as in Fig. 4) of LCL height (m AGL; lowest 100-hPa mixed-layer parcels) from all tornado cases (520 total) in the RUC database used for this study, categorized by location west or east of −98.5° longitude.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 154 154 36
PDF Downloads 112 112 26

Tornadoes in Environments with Small Helicity and/or High LCL Heights

View More View Less
  • 1 Wichita, Kansas
© Get Permissions
Full access

Abstract

Recent studies have suggested that supercell tornado environments are usually associated with large 0–1-km storm-relative helicity (SRH) and relatively low lifting condensation levels (LCL heights). However, occasional tornadoes of significance occur in environments having characteristics that appear less supportive of supercell tornadoes, including small SRH values and/or relatively high LCL heights. Such tornadoes, whether associated with supercell or nonsupercell processes (more precisely termed mesocyclone and nonmesocyclone processes), present a challenge for forecasters. This empirical study uses a database of soundings derived from the Rapid Update Cycle model to examine thermodynamic characteristics of F1 and greater intensity tornado events associated with small SRH and/or high LCL heights. Results strongly suggest that many such tornado events are associated with steep lapse rates in the lowest few kilometers above ground. The low level of free convection heights, small convective inhibition, and sizable convective available potential energy below 3 km were also found to be of possible importance. These thermodynamic characteristics combined would likely reduce resistance to upward accelerations, potentially enhancing ascent for low-level parcels entering thunderstorm updrafts and, hence, low-level stretching. From prior research, if preexisting boundaries were available to provide surface vertical vorticity for stretching, such thermodynamic characteristics could be an important component of tornado events that involve nonmesocyclone processes. These same thermodynamic characteristics may also offer clues for the investigation of mesocyclone tornado events that do not fit well with accepted tornado forecasting parameters from prior studies.

Corresponding author address: Jonathan M. Davies, 3206 N. Westwind Bay, Wichita, KS 67205-2528. Email: jdavies1@cox.net

Abstract

Recent studies have suggested that supercell tornado environments are usually associated with large 0–1-km storm-relative helicity (SRH) and relatively low lifting condensation levels (LCL heights). However, occasional tornadoes of significance occur in environments having characteristics that appear less supportive of supercell tornadoes, including small SRH values and/or relatively high LCL heights. Such tornadoes, whether associated with supercell or nonsupercell processes (more precisely termed mesocyclone and nonmesocyclone processes), present a challenge for forecasters. This empirical study uses a database of soundings derived from the Rapid Update Cycle model to examine thermodynamic characteristics of F1 and greater intensity tornado events associated with small SRH and/or high LCL heights. Results strongly suggest that many such tornado events are associated with steep lapse rates in the lowest few kilometers above ground. The low level of free convection heights, small convective inhibition, and sizable convective available potential energy below 3 km were also found to be of possible importance. These thermodynamic characteristics combined would likely reduce resistance to upward accelerations, potentially enhancing ascent for low-level parcels entering thunderstorm updrafts and, hence, low-level stretching. From prior research, if preexisting boundaries were available to provide surface vertical vorticity for stretching, such thermodynamic characteristics could be an important component of tornado events that involve nonmesocyclone processes. These same thermodynamic characteristics may also offer clues for the investigation of mesocyclone tornado events that do not fit well with accepted tornado forecasting parameters from prior studies.

Corresponding author address: Jonathan M. Davies, 3206 N. Westwind Bay, Wichita, KS 67205-2528. Email: jdavies1@cox.net

1. Introduction

Numerous studies have examined wind and thermodynamic parameters associated with tornadic environments (e.g., Davies-Jones et al. 1990; Davies and Johns 1993; Johns et al. 1993; Rasmussen and Blanchard 1998; Rasmussen 2003; Craven and Brooks 2004). These have helped refine characteristics and parameters that are useful for mesocyclone tornado forecasting, such as sizable storm-relative helicity (SRH; Davies-Jones et al. 1990) and low lifting condensation level heights (LCL; e.g., Rasmussen and Blanchard 1998; Thompson et al. 2003, hereafter T03). Widely used composite calculations, such as the energy–helicity index (EHI; Hart and Korotky 1991; Davies 1993) and significant tornado parameter (STP; see T03), also incorporate one or more of these parameters. However, occasional tornadoes of strong intensity (F2+) or approaching this intensity (F1) occur in environments not well supported by these characteristics and parameters, and it is these events that are often the most challenging for forecasters. Two examples are shown in Fig. 1. The first was a long-lived tornado (Fig. 1a) documented by multiple photographers in open country associated with a radar-indicated supercell over northwest Kansas on 29 June 2000 in an environment with LCL heights and cloud bases that were quite high (around 2000 m AGL). Because the tornado moved mainly through farmland and open country, it was officially rated F1, but was impressive visually and possibly stronger in intensity. The second was a tornado that produced F2 damage in south-central Kansas (Fig. 1b) on 27 August 2004 where both SRH and LCL heights seemed unfavorable for tornadoes. Such events are the motivation for this study.

Studies, such as T03 and Davies (2004, hereafter D04), have used model-derived soundings to examine environments associated with tornadic and nontornadic supercells, showing that such estimations can be useful for both forecasting and research. For this empirical study, the database from D04 was expanded using profiles from the Rapid Update Cycle (RUC; Benjamin et al. 2004) model and time–location constraints similar to those used in D04. Instead of focusing exclusively on storms that were supercells as in D04, some storms, including tornadic ones, were allowed into the database even though they were difficult to define as supercells based on radar data, or appeared to be nonsupercell in nature (Wakimoto and Wilson 1989, hereafter WW89). The storm associated with the tornado in Fig. 1b is one such example. In this respect, the database had similarities to the one used in Craven and Brooks (2004), where cases were not categorized as supercell or nonsupercell. Supercell and nonsupercell tornado terminology can be ambiguous when tornadoes from nonsupercell processes occur along the flanking line or gust front of a supercell away from the mesocyclone. Therefore, for clarity in this study, tornadoes associated with supercell mesocyclones will be called mesocyclone tornadoes, and tornadoes not associated with mesocyclones will be called nonmesocyclone tornadoes, as in Brady and Szoke (1989, hereafter BS89).

The database of RUC profiles was then examined for tornado cases that did not fit into established SRH and LCL parameter ranges typically associated with tornadoes (e.g., T03). Rather than focusing strictly on F2 and greater intensity tornadoes, F1 tornadoes were included as in D04 because they are responsible for a significant number of injuries (e.g., 207 in 2003–04) and property damage according to the National Oceanic and Atmospheric Administration publication Storm Data. The resulting “atypical” tornado soundings were then studied for features and characteristics that might set them apart from those meeting more established tornado-forecasting criteria.

Low-level thermodynamic characteristics such as lapse rates in the lowest few kilometers, convective inhibition (CIN; Colby 1984), level of free convection (LFC) height, and convective available potential energy (CAPE; Moncrief and Miller 1976) below 3 km AGL were examined, comparing atypical and typical tornado soundings from the database. Discussion in BS89, and modeling work by Lee and Wilhelmson (1997, 2000), hinted at the importance of low-level thermodynamic settings to nonmesocyclone tornadoes, but little additional investigation has been done in this area. Many nonmesocyclone tornadoes occur in high LCL settings (e.g., Caruso and Davies 2005), as do some tornadoes associated with mesocyclones in supercell environments (e.g., the tornado shown in Fig. 1a).

The following section will describe the database used and define atypical SRH and LCL environments for purposes of this study. Section 3 will discuss parameters and characteristics that were examined. Results from the database investigation will be presented in section 4, and will be compared with results from a database independent of this study in section 5. These thermodynamic results will be applied to three brief case investigations in section 6. The final section is a concluding discussion.

2. Database and definition of events with small SRH and/or high LCL heights

The database of RUC profiles from D04 (covering the years 2001–mid-2003) was expanded and doubled in size for use in this study using additional RUC-20 profiles from 2003 and 2004 (Benjamin et al. 2002). The resulting database of RUC analysis soundings was associated with storms chosen randomly from severe or tornado-warned events (similar to D04), with 94% of the storms linked to verified severe thunderstorm or tornado reports from Storm Data. All profiles were located within 100 km of, and 90 min prior to, radar-warned storms in the storm inflow air mass (see D04 for more details). Of 1049 profiles in the database, 209 were from the RUC-2 (2001–2002), and 840 were from the RUC-20 (2002–2004). Table 1 summarizes the profiles grouped by nontornadic cases and by tornado intensity.

As noted earlier, mesocyclone–nonmesocyclone categorization was not used for profiles in this investigation, which is different than D04 but similar to Craven and Brooks (2004). One reason was ambiguities with mesocyclone–nonmesocyclone categorization using available data in cases located a significant distance from radar. The desire to examine a realistic spectrum of severe and tornadic cases encountered by forecasters was also a major factor for including storms that were difficult to categorize or probably nonmesocyclone in nature. It should also be highlighted that the distinction between mesocyclone and nonmesocyclone processes can be unclear and confusing in tornado cases that involve supercell thunderstorms but tornadogenesis processes that may be nonmesocyclone in nature, as in Wakimoto and Atkins (1996). Due in part to an emphasis on supercell storms in the initial database from D04, it can be noted that roughly 80% of the profiles were associated with storms that at some point in their life cycle appeared to have supercell (mesocyclone) characteristics according to definitions in either T03 or D04.

Box-and-whisker diagrams from T03 (see their Figs. 7 and 11) were used to determine atypical values of 0–1-km SRH (SRH0–1) and mixed-layer LCL associated with significant tornadoes. Values falling outside the box-and-whisker ranges for significant tornadoes in T03 were found to be SRH0–1 < 75 m2 s−2 and LCL > 1300 m AGL, called “small” SRH and “high” LCL values in this study. These contrast with more typical tornadic values of SRH0–1 ≥ 75 m2 s−2 and LCL ≤ 1300 m AGL. Table 2 summarizes F1 and greater intensity tornado cases in the database for this study that were associated with these “less favorable“ SRH and/or LCL characteristics, composing roughly one-third (36%) of the total F1+ intensity cases. Tornadoes rated F1 were included and considered important (as in D04) because they are responsible for significant property damage (more than $320 million during 2003–04 according to Storm Data) as well as injuries and occasional deaths (e.g., eight deaths and 207 injuries during 2003–04).

The profiles associated with tornadoes in Table 2 do not appear to be supported by environmental parameter values and ranges from prior studies (e.g., Rasmussen and Blanchard 1998; T03) that are commonly used in tornado forecasting. As such, they represent tornado events that would likely prove difficult to forecast. These cases motivated a search for other detectable characteristics that might be relevant.

3. Characteristics and parameters examined

A subjective and visual examination of the atypical tornadic RUC profiles in Table 2 suggested that certain low-level thermodynamic characteristics set them apart from more typical tornadic profiles. More specifically, lapse rates in the lowest few kilometers appeared to be consistently steep, and were often present in combination with a relatively low LFC height and small CIN (see D04). The LFC and CIN characteristics were generally the result of a relatively large mixing ratio in the lowest 1–2 km, and the absence of a temperature inversion above this layer.

A RUC profile associated with the high-based tornadic supercell shown in Fig. 1a serves as a good example of these characteristics and is shown in Fig. 2. Note that the lapse rate in the lowest 2–3 km of this sounding was quite steep, between 9° and 10°C km−1, which is very close to the dry-adiabatic rate, without any temperature inversion above. This is similar to comments by BS89, who noted that steep and unstable low-level lapse rate conditions favoring dust devils appeared to be associated with nonmesocyclone tornado formation. Also note that, even though the mixed-layer LCL on the profile in Fig. 2 was quite high (near 2100 m AGL, matching visual observations of the storm), the dewpoint and mixing ratio were relatively large through the lowest 1.5 km. Using the lowest 100-hPa mixed-layer lifted parcels, this resulted in an LFC of roughly 2200 m AGL located immediately above the LCL, with small CIN (<10 J kg−1). The presence of CAPE below 3 km (CAPE0–3; Rasmussen 2003), with values of 50–60 J kg−1, also reflected significant low-level moisture.

This thermodynamic configuration in low levels, with an environmental lapse rate approaching the dry-adiabatic rate, would result in no resistance to rapid parcel ascent in the lowest few kilometers if locally intense surface heating was ongoing. Furthermore, the combination of low LFC heights, small CIN, and CAPE within the lowest 3 km when using mixed-layer lifted parcels suggests the presence of substantial low-level relative humidity in a surface-based layer, and also the absence of any low-level temperature inversion that would strongly inhibit rising parcels. This contrasts with steep low-level lapse rate settings that have less low-level moisture (see the example in Fig. 3a), resulting in smaller relative humidity and much higher LFC heights with little or no CAPE0–3 for mixed-layer parcels (cf. Fig. 3a and Fig. 2), even with little or no CIN present. An additional example in Fig. 3b had more moisture in the low levels than Fig. 3a along with steep low-level lapse rates through 2 km AGL, but also a temperature inversion above that (near 650 hPa). The resulting warm layer and CIN (>40 J kg−1) in Fig. 3b would likely slow parcels rising from below, with high LFC heights (near 3500 m AGL) above the inversion and no CAPE0–3 present. Comparing these environmental examples (Figs. 3a and 3b, both associated with nontornadic thunderstorms developing on preexisting boundaries) with the high-based tornadic environment example in Fig. 2 suggests the importance of the combination of several thermodynamic characteristics mentioned above.

These characteristics might help to enhance low-level stretching beneath thunderstorm updrafts, a process that could have an impact on tornadogenesis, particularly if it occurred over a preexisting boundary providing a source of vertical vorticity (WW89 and BS89). It is worth noting that the thermodynamic profile in Lee and Wilhelmson (1997, see their Fig. 2), which was used as the basis for model simulations of nonmesocyclone tornadoes, had very similar characteristics. The similarity in these profiles suggests the possibility that this thermodynamic combination could also contribute to tornado development with mesocyclones in high LCL environments. In the case of Figs. 1a and 2, the supercell formed near a northeast–southwest-oriented surface trough and wind shift, and the tornado occurred 75–90 min after the first echo on radar. With steep low-level lapse rates present for rapid parcel ascent, sizable deep-layer shear (20–25 m s−1, not shown) likely offered support for the supercell characteristics of this atypical, high-based tornadic mesocyclone and storm.

From the above observations, it was decided to compute and explore lapse rates in the lowest 2–3 km, along with low-level moisture variables such as LFC height, CIN, and CAPE in the lowest 3 km for profiles in the RUC database used for this study. Low-level lapse rates were computed by simply subtracting the model-derived temperature at the top of the desired layer from the surface temperature and dividing by depth in kilometers, with resulting units of degrees Celsius per kilometer. The height of the LFC (m AGL) and the CIN (J kg−1) were also computed as in D04, and CAPE0–3 (J kg−1) was computed as in Rasmussen (2003). All thermodynamic parameter computations were performed using the lowest 100-hPa mixed-layer lifted parcels, similar to T03 and D04, and utilized virtual temperature correction (Doswell and Rasmussen 1994).

As in D04, observed surface temperature, dewpoint, and wind information at the same time and location were saved with each profile and used to modify the profiles in the lowest 150 hPa when they differed significantly from the raw model profile information (see D04 for details). This modification was performed to make the lowest levels of model-derived soundings representative of observed surface conditions.

4. Results from RUC database

a. F1–F4 tornadoes

Median values of selected thermodynamic parameters are shown in Table 3, organized by small versus typical SRH0–1 values and high versus typical LCL heights as defined in section 2. (Median values are used in Table 3 because variables such as CAPE and CIN had extreme values in a few cases, rendering mean values less representative of the true distribution.) For brevity, only lapse rates from the 0–2-km layer are given. Note that in the small SRH0–1 and high LCL tornado cases, the median 0–2-km lapse rates were markedly steeper, by 2°C km−1 or more, with the difference between groupings statistically significant above the 95% confidence using a Student’s t test for equal and unequal variances. Results were similar for lapse rates in the 0–3-km layer (not shown), with slightly less difference (by 0.2°–0.4°C) in the median values between groupings. It should be noted that these lapse rates were steeper than those found by Craven and Brooks (2004) that were associated with a more typical cross section of tornado cases characterized by large 0–1-km shear values and relatively low LCL heights, similar to T03.

The distribution of 0–2-km lapse rate values in Figs. 4a and 4b also shows that tornadic profiles having small SRH0–1 values (<75 m2 s−2) or high LCL heights (>1300 m AGL) were associated with much steeper low-level lapse rates. If the higher LCL height cases alone were being examined, one might argue that the steep lapse rates were only a reflection of larger differences between temperature and dewpoint that occur with warmer surface temperatures in higher LCL environments, and not a relevant characteristic associated with tornadic high LCL storms. However, the fact that the same signal was evident when viewing the tornado cases by SRH (regardless of LCL height; Fig. 4a) suggests some significance to this association. More than a third of the Table 2 profiles were associated with both small SRH values and high LCL heights.

Low-level lapse rates are steep over large areas with surface heating during many afternoons of the year, particularly in late spring and summer, so the above finding may not at first appear particularly relevant or useful. However, as suggested in the prior section, additional characteristics may also have relevance, such as relatively low LFC heights, small CIN, and notable CAPE0–3. In Table 3, median values of LFC height, CIN, and CAPE0–3 are shown. Note that median LFC heights were relatively low (<2000 m AGL; see D04) for all tornado case groupings shown in the table, including those associated with small SRH and/or high LCL heights. Also note that median CIN values were smaller (<15 J kg−1) for the small SRH and/or high LCL cases, and that low-level CAPE was significant for the same cases, greater than the median value of 64 J kg−1 from tornado cases in Rasmussen (2003).

The fact that these additional low-level thermodynamic characteristics were present when low-level lapse rates were also steep may be important, as suggested in the prior section. This issue is investigated in the next section.

b. Tornadic and nontornadic cases with steep lapse rates

The results in the prior section are of little value unless they also suggest some difference between low-level thermodynamic environments for tornadic and nontornadic cases when steep low-level lapse rates are present. This was explored by examining all cases from the database in this study, tornadic and nontornadic, having 0–2-km lapse rates equal to or steeper than 7°C km−1. This lapse rate value was chosen from Fig. 4, where 7°C km−1 showed good separation between tornadoes with small SRH and/or high LCL heights and those associated with more typical tornadic environments. Figure 5 shows box-and-whisker plots comparing LFC, CIN, and CAPE0–3 for the nontornadic (224 cases) and F1–F4 tornadic cases (90 cases) that were associated with both relatively steep 0–2-km lapse rates and small SRH and/or high LCL heights. (Tornado cases that were F0 in intensity are not shown in Fig. 5 in order to emphasize the differences between the two groupings.) Note that the tornadic cases with steep low-level lapse rates tended to have somewhat lower LFC heights, less CIN, and more low-level CAPE, with differences of roughly one quartile between categories, particularly with CAPE0–3. Though not an exceptionally strong signal with considerable overlap between groupings, Fig. 5 does suggest that steep low-level lapse rates accompanied by lower LFC heights, small CIN, and significant low-level CAPE may be relevant to environments capable of supporting some atypical tornado events.

5. Results from an independent database of observed soundings

Exploration of an independent database of observed soundings was considered desirable to determine if any of the characteristics and signals found in the RUC database used for this study were also present for small SRH and/or high LCL tornado cases from observed sounding profiles. This was possible using results from a large database of 0000 UTC rawinsonde soundings generously provided by the lead author of Craven and Brooks (2004). This section will briefly summarize those results.

The collection of observed sounding computations was from an unpublished database (J. Craven 2004, personal communication) covering a wide time frame (the years 1957–96). The proximity criteria were the same as with the smaller database of observed soundings examined in Craven and Brooks (2004), with severe events located within 185 km of the sounding release location and the 6-h period centered on 0000 UTC. These results were grouped by significant tornadoes (F2–F5 intensity) and significant hail/wind events (>5 cm hail and/or >33 m s−1 wind gust). After some quality control was applied, the resulting database contained parameter computations from 1531 soundings associated with significant tornadoes, and 3454 soundings associated with nontornadic severe reports.

The observed soundings associated with tornadoes were categorized and examined in the same way as the F1–F4 tornadoes from the RUC database in Table 3 from the previous section, and are shown in Table 4. For tornadoes associated with small SRH and/or high LCL heights, it can be seen that median 0–2-km lapse rates were noticeably steeper (by 1.5°–2°C km−1) than with tornadoes associated with more typical SRH or LCL values, which is quite similar to Table 3 for the RUC database. However, the results regarding LFC, CIN, and CAPE0–3 were not similar, with median LFC heights 500–600 m higher than in Table 3, median CIN values 40–60 J kg−1 larger, and median CAPE0–3 50–70 J kg−1 smaller, although some CAPE0–3 (15–60 J kg−1) was indicated for all categories. Apart from these latter results, Table 4 at least confirms that low-level lapse rates were considerably steeper for atypical tornado cases associated with small SRH and/or high LCL heights, which is the same as the results from the RUC database in section 4a.

When nontornadic and tornadic events associated with steep low-level lapse rates in the observed sounding database were compared (not shown), similar to the comparison in section 4b, there was little discernible difference between the mean or median values of LFC height, CIN, and CAPE0–3. This echoed the contradictions between Tables 3 and 4 noted above with the same parameters, and may be related to issues resulting from the broader proximity criteria used (nearly twice as far in distance and 4 times further removed in time than with the RUC database). This result suggests caution in interpreting and applying the LFC, CIN, and low-level CAPE results from section 4. Nevertheless, the strength of the low-level lapse rate signal in both databases implies that this particular characteristic may have subjective value to forecasters in combination with other parameters in some atypical tornado forecast settings, an idea that will be touched upon with three brief case studies in the next section.

6. Case investigations

In this section, two cases from 2004 involving F1–F2 tornadoes in small SRH and/or high LCL environments are briefly examined from the standpoint of low-level thermodynamic factors discussed in prior sections. These are compared with one nontornadic case. Mesoanalysis graphics are used from the Storm Prediction Center [SPC; Bothwell et al. (2002); information available online in real time at http://www.spc.noaa.gov/exper/mesoanalysis/] with parameter fields computed using mixed-layer parcels as performed in this study.

Lapse rate for the 0–3-km depth, instead of 0–2 km, is the only product currently available online from SPC regarding low-level lapse rate fields. Although 0–2-km lapse rates were found to be slightly steeper in the atypical RUC tornado cases discussed in section 4a, results between the 0–2- and 0–3-km layers were similar (also noted in section 4a). Therefore, the 0–3-km fields were considered to be a reasonable surrogate for 0–2-km lapse rate values in the cases presented here. It should be mentioned that there can be settings where lapse rates taper off rapidly above 2 km (e.g., Fig. 3b in section 3), rendering lapse rates in the 0–3-km layer somewhat less representative of shallower depths. This emphasizes that the most relevant depth is neither clearly defined nor a fixed layer, and will probably depend on the setting.

a. Case 1–27 August 2004, south-central Kansas

On the evening of 27 August 2004, a few tornadoes occurred in south-central Kansas south of Wichita, Kansas (Fig. 6d), including an F2 tornado that lasted nearly 20 min (see Fig. 1b from section 1). This tornado occurred in an environment exhibiting relatively small SRH values and high LCL heights, near a weak surface low on a weak quasi-stationary front oriented northeast–southwest (Fig. 6a).

Figure 6a shows the SRH0–1 field for 2300 UTC from the SPC mesoanalysis, 60–90 min before the tornadoes, which suggests little in the way of mesocyclone tornado potential over south-central Kansas where SRH0–1 was small (<50 m2 s−2) and LCL heights were relatively high (>1500 m AGL). Deep-layer shear in this area (not shown) was less than 20 m s−1. However, near and parallel to the slow-moving surface front, an axis of strong surface heating and steep 0–3-km lapse rates (>8°C km−1, shown in Fig. 6b) overlapped an area of low LFC height values (<2000 m AGL; Fig. 6c) over southern Kansas. In this same area, CIN was small (<25 J kg−1, not shown) and CAPE0–3 was large (50–100 J kg−1, not shown) near the Kansas–Oklahoma border northeast of the surface low.

The F2 tornado near Wellington, Kansas, occurred at the intersection of the front and an outflow boundary from thunderstorm cells farther north, a location likely rich in surface vertical vorticity just northeast of the surface low. Radar data from Wichita (not shown; see Caruso and Davies 2005) indicated that the tornado developed in association with a new and rapidly developing echo located at the boundary intersection, just southwest of ongoing radar echoes and without a prior mesocyclone signature aloft. It is possible that the local environment, in this case featuring steep low-level lapse rates, low LFC heights (see area highlighted in Fig. 6d), small CIN, and significant CAPE0–3, contributed to the strong tornado that occurred at the boundary intersection from what appeared to be nonmesocyclone processes.

b. Case 2–24 May 2004, south-central Nebraska

On this day, the author observed and photographed several tornadoes (Fig. 7) with a supercell thunderstorm in south-central Nebraska southwest of Hebron, Nebraska (see Fig. 8d for tornado locations). The storm developed near a quasi-stationary front and moved east-southeast. The first tornadoes observed had the visual appearance of “landspouts” (Bluestein 1985) along the flanking line of the storm, visible as dirt debris drawn upward into short condensation funnels extending from a flat cloud base (Fig. 7a). A later, longer-lived tornado was the strongest observed (F1–F2 intensity) and was accompanied by a rear-flank downdraft and other visual supercell characteristics at the same time another tornado was in progress a short distance to its southeast (Fig. 7b).

Early to midafternoon mesoanalysis graphics from SPC showed the SRH0–1 in south-central Nebraska near the Kansas border to be relatively small (<75 m2 s−2; Fig. 8a) compared with that in northwest Missouri (>200 m2 s−2; Fig. 8a) where an F2 tornado occurred with a different supercell. However, steep 0–3-km lapse rates were present in southern Nebraska on the SPC mesoanalysis (8°–9°C km−1; Fig. 8b) along an axis extending northward from west-central Kansas, and CAPE0–3 was also sizable (>50 J kg−1; Fig. 8c) in the general area where the tornadoes occurred. Also, CIN was small (<25 J kg−1, not shown) and LFC heights low (<2000 m AGL, not shown) in this same area. Figure 8d indicates where the steepest low-level lapse rates intersected significant low-level CAPE and other desirable low-level thermodynamic characteristics over south-central Nebraska and northern Kansas at 2000 UTC just prior to the tornadoes.

This area of enhanced low-level lapse rates, significant low-level CAPE, small CIN, and low LFC heights was coincident with maximized surface vertical vorticity (see the vorticity analysis in Fig. 8c), and possibly contributed to the rapid development of several tornadoes in an environment seemingly characterized by relatively small SRH. It is interesting that visual observations of the tornadoes (detailed radar data were not examined) suggested both mesocyclone and nonmesocyclone characteristics. Although SRH was relatively small, the 0–6-km deep-layer shear (not shown) was marginal to adequate for supercells (15–18 m s−1) according to results in T03. The damaging tornado in northwest Missouri was associated with a more typical tornado environment involving larger SRH just north of a warm frontal boundary (Fig. 8a) and 22–25 m s−1 of deep-layer shear (not shown). It may be notable that the Missouri supercell took longer from its inception on radar to produce tornadoes (2.5 h) than the southern Nebraska storm (roughly 1 h) that was in a steeper low-level lapse rate environment.

c. Case 3–8 June 2004, southeast Colorado

In this nonmesocyclone case (associated with the environment in Fig. 3a from section 3), a thunderstorm developed directly on a preexisting boundary seen in surface, satellite, and radar data, but did not produce tornadoes.

This isolated hail-producing storm (Fig. 9c) formed over a stationary wind shift line in southeast Colorado on 8 June 2004 in a steep low-level lapse rate environment. From the SPC mesoanalysis graphic (Fig. 9a) at 1800 UTC, 0–3-km lapse rates were very steep (>9°C km−1) in the vicinity of the surface boundary, which was rich with vertical vorticity, as seen in Fig. 9b. A composite reflectivity image (Fig. 9c) shows that the storm was situated directly on this northeast to southwest oriented boundary, which is seen as a fine line in the radar depiction. However, Fig. 9b also indicates that there was no low-level CAPE anywhere in the area where the boundary-related storm occurred. Correspondingly, LFC heights were quite high (>3000 m AGL; see Fig. 3a in section 3), even though CIN appeared very small (near 0) where the storm developed. The fact that LCL heights were also very high in the same area (>3000 m AGL) due to shallow low-level moisture (Fig. 3a) would suggest the potential for considerable mixing, with significant dilution of buoyancy for rising parcels below the cloud base, a probable negative factor for nonmesocyclone tornadoes.

In the above cases, it is impossible to explain why tornadoes did or did not occur. Apart from the thermodynamic factors in events such as the 8 June 2004 nonmesocyclone case, it is possible that developing thunderstorms may not be optimally oriented along boundaries to take advantage of preexisting vertical vorticity. Local-scale boundary characteristics such as strength and orientation of vertical vortex sheets (Lee and Wilhelmson 2000) are small-scale factors beyond the scope of the investigations above. However, the cases presented in this section are consistent with the general results from the RUC database examined in this study. Furthermore, the 24 May 2004 case suggests that mesocyclone and nonmesocyclone tornado processes are at times difficult to distinguish and may even overlap to some degree. In summary, the cases discussed above do suggest potential application of the results in this study to some short-term tornado forecasting and nowcasting situations, as well as avenues for further research, mentioned in the concluding section.

7. Discussion

Nonmesocyclone tornadogenesis processes were identified and discussed by WW89 and BS89. These tornadoes develop through low-level stretching of enhanced vertical vorticity in pockets along slow-moving or stationary preexisting wind shift boundaries via thunderstorm updrafts positioned over the boundary. Apart from noting significant amounts of total CAPE (e.g., BS89; Lee and Wilhelmson 2000) and the presence of steep low-level lapse rates (e.g., BS89), prior studies have not discussed specific low-level thermodynamic environment characteristics that might enhance and augment near-surface stretching processes important to nonmesocyclone tornadoes. This empirical study strongly suggests that steep lapse rates in approximately the lowest 2–3 km are an important factor. Though less clear, other low-level thermodynamic factors such as the absence of a low-level temperature inversion (small CIN) above a layer with enough moisture to generate relatively low LFC heights (near or just above the LCL) and significant low-level CAPE with mixed-layer parcels may also be important. When the rate of change of temperature in the lowest few kilometers approaches the dry-adiabatic rate with little convective inhibition above a significantly moist layer, rising parcels would likely encounter no resistance in low levels and accelerate rapidly with limited reduction of buoyancy. Local heating might even generate superadiabatic conditions to enhance parcel ascent further, as suggested by dust devils observed along dryline moisture discontinuities in Pietrycha and Rasmussen (2004). When these processes take place beneath a thunderstorm updraft directly along a vorticity-rich boundary, tornado development may be possible if the updraft and boundary-related vertical vorticity are oriented properly.

It was also suggested in WW89 that nonmesocyclone processes could have an impact on some mesocyclone tornadoes. As mentioned in section 2, the distinction between mesocyclone and nonmesocyclone environments and processes is blurred by cases such as those described by Wakimoto and Atkins (1996), who provided documentation of an F3 tornado event that was associated with a supercell, but not linked to the storm’s midlevel mesocyclone. Table 2 from section 2 suggests that occasional significant tornadoes can occur with storms, some of which are supercells, in environments that do not exhibit accepted mesocyclone tornado parameter characteristics (e.g., relatively large SRH and low LCL heights). Results from this study imply that thermodynamic characteristics with the potential for rapid low-level parcel ascent may contribute to some tornadoes occurring with supercell mesocyclones in what appear to be otherwise unfavorable environments (e.g., relatively high LCL heights and/or relatively small SRH values). Visual observations from the 24 May 2004 case in section 6 also suggest that mesocyclone and nonmesocyclone tornado processes may, in some cases, be overlapping or ongoing concurrently in the same environmental setting.

Low LCL heights have been recognized by several recent researchers (e.g., Rasmussen and Blanchard 1998; T03) as important to most significant mesocyclone tornado environments. It is thought that low LCL heights and cloud bases are indicative of large low-level humidity that would reduce the likelihood of near-surface outflow and cold pool production that could interfere with tornadogenesis (Markowski et al. 2002). If this is true, a relevant question would focus on how tornadoes can occur in high LCL environments. It has been documented that tornadoes from nonmesocyclone processes can develop when storm updrafts are relatively young but growing rapidly (e.g., BS89; Burgess et al. 1993). Such observations suggest that tornadoes in environments with steep low-level lapse rates may occur relatively early in a storm’s life cycle before production of extensive precipitation and cool outflow. There were indications in the tornadic cases discussed in section 6 that this was true, regardless of whether the tornadoes were associated with mesocyclone characteristics. This early potential for tornadoes relative to storm cycle in steep lapse rate environments may suggest one explanation for tornado occurrences in high LCL settings.

There is also a tendency for tornadoes in the high plains of the United States to be associated with higher LCL heights, as seen in Fig. 10 where cases west of −98.5° longitude approximate surface elevations higher than 0.5 km MSL. Surface heating over higher elevations adjacent to drier air from the desert Southwest results in larger spreads between low-level temperature and dewpoint, and therefore higher LCL heights, but also contributes to steeper low-level lapse rates. Based on the empirical findings in this paper, the invasion of adequate moisture and low-level CAPE into the high plains when steep lapse rates are also present could help explain why tornadoes can occur with higher LCL heights in this region. From Fig. 10, forecasters should be alert that both mesocyclone and nonmesocyclone tornadoes in the high plains of the United States can happen with LCL heights that appear higher than typical values suggested by studies such as Rasmussen and Blanchard (1998) and T03.

The low-level thermodynamic characteristics discussed in this study are probably of limited value as a direct suggestion of tornado potential in small SRH and/or high LCL environments. However, they may be useful to forecasters in a subjective sense when combined with the assessment of preexisting boundaries having enhanced vertical vorticity, as touched upon in section 6. From those cases, and cases such as those examined in Caruso and Davies (2005), slow-moving wind shift boundaries (such as weak fronts or surface troughs), typically oriented from north to south or northeast to southwest, appear to be a common surface pattern associated with nonmesocyclone tornado events. This boundary orientation was also associated with some atypical tornado cases that had mesocyclone characteristics (e.g., case 2 in section 6, and the case shown in Figs. 1a and 2), supported by marginal amounts of horizontal vorticity and deep-layer shear in the local environment. An awareness of boundary locations and thermodynamic characteristics in similar cases may help with short-term recognition of tornado potential in occasional atypical settings for forecasters, as noted in Caruso and Davies (2005). Furthermore, in settings with nonmesocyclone characteristics, this awareness can be enhanced through careful radar assessment of low-level boundaries (within detectable distances) to locate preexisting low-level circulations that may be relevant to tornado potential (Pietrycha and Manross 2003).

Finally, this study suggests many questions and areas for future research. What processes are really involved when tornadoes occur in small SRH or high LCL environments? Through what depth (sub-LCL, sub-LFC, or some other layer) are steep low-level lapse rates most relevant and optimal for increased parcel ascent beneath thunderstorm updrafts? To what degree can intense low-level stretching compensate for meager amounts of environmental horizontal streamwise vorticity in some mesocyclone tornado cases? Why do some tornadoes in small SRH and/or high LCL environments seem to be tied directly to preexisting boundaries, and other tornadoes less so, particularly those associated with some supercell mesocyclones? Why does cool outflow not appear to be a major factor in tornadic high LCL environments where the deeper subcloud layer would suggest the potential for rapid evaporative cooling when precipitation falls? To what degree do mesocyclone and nonmesocyclone processes overlap (if at all) in small SRH and/or high LCL tornado events associated with supercell storms and mesocyclones? The list of questions is long and offers the opportunity for modelers and field researchers to explore tornadoes associated with environments that do not match characteristics emphasized in much current tornado forecast training and research.

Acknowledgments

The author gratefully recognizes Jeff Craven (NWS Jackson, Mississippi) for sharing the results from his large database of observed soundings associated with severe convection. Thanks to Earl Barker (Harris Corporation, Omaha, Nebraska) for programming assistance and encouragement, and Al Pietrycha (NWS, Goodland, Kansas), John Stoppkotte (NWS, North Platte, Nebraska), and Mike Umscheid (NWS, Dodge City, Kansas) for useful discussions and ideas. In addition to helpful comments and suggestions from two anonymous reviewers, valuable input was received from Corey Mead and Rich Thompson at SPC. Discussion with Erik Rasmussen (CIMMS) was helpful and much appreciated. Finally, thanks to Chuck Doswell (CIMMS) for his insightful review and comments pertaining to this paper.

REFERENCES

  • Benjamin, S. G., and Coauthors, 2002: RUC20—The 20-km version of the Rapid Update Cycle. NWS Tech. Procedure Bull. 490, 29 pp.

  • Benjamin, S. G., and Coauthors, 2004: An hourly assimilation–forecast cycle: The RUC. Mon. Wea. Rev., 132 , 495518.

  • Bluestein, H. B., 1985: The formation of a “landspout” in a “broken-line” squall line in Oklahoma. Preprints, 14th Conf. on Severe Local Storms, Indianapolis, IN, Amer. Meteor. Soc., 267–270.

  • Bothwell, P. D., , Hart J. A. , , and Thompson R. L. , 2002: An integrated three-dimensional objective analysis scheme in use at the Storm Prediction Center. Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., J117–J120.

  • Brady, R. H., , and Szoke E. J. , 1989: A case study of nonmesocyclone tornado development in northeast Colorado: Similarities to waterspout formation. Mon. Wea. Rev., 117 , 843856.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Burgess, D. W., , Donaldson R. J. Jr., , and Desrochers P. R. , 1993: Tornado detection and warning by radar. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geophys. Union, 203–221.

    • Crossref
    • Export Citation
  • Caruso, J. M., , and Davies J. M. , 2005: Tornadoes in nonmesocyclone environments with pre-existing vertical vorticity along convergence boundaries. Natl. Wea. Assoc. Electron. J. Operational Meteor., 2005-EJ4 [Available online at http://www.nwas.org/ej/.].

    • Search Google Scholar
    • Export Citation
  • Colby, F. P., 1984: Convective inhibition as a predictor of convection during AVE-SESAME-2. Mon. Wea. Rev., 112 , 22392252.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Craven, J. P., , and Brooks H. E. , 2004: Baseline climatology of sounding-derived parameters associated with deep moist convection. Natl. Wea. Dig., 28 , 1324.

    • Search Google Scholar
    • Export Citation
  • 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.

  • Davies, J. M., 2004: Estimations of CIN and LFC associated with tornadic and nontornadic supercells. Wea. Forecasting, 19 , 714726.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies, J. M., , and Johns R. H. , 1993: Some wind and instability parameters associated with strong and violent tornadoes. 1. Wind shear and helicity. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geophys. Union, 573–582.

    • Crossref
    • Export Citation
  • Davies-Jones, R. P., , Burgess D. , , and Foster M. , 1990: Test of helicity as a tornado forecast parameter. Preprints, 16th Conf. on Severe Local Storms, Kananaskis Park, AB, Canada, Amer. Meteor. Soc., 588–592.

  • Doswell C. A. III, , , and Rasmussen E. N. , 1994: The effect of neglecting the virtual temperature correction on CAPE calculations. Wea. Forecasting, 9 , 625629.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hart, J. A., , and Korotky W. , 1991: The SHARP workstation v1.50 user’s guide. NOAA/National Weather Service, 30 pp. [Available from NWS Eastern Region Headquarters, 630 Johnson Ave., Bohemia, NY 11716.].

  • Johns, R. H., , Davies J. M. , , and Leftwich P. W. , 1993: Some wind and instability parameters associated with strong and violent tornadoes. 2. Variations in the combinations of wind and instability parameters. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geophys. Union, 583–590.

    • Crossref
    • Export Citation
  • Lee, B. D., , and Wilhelmson R. B. , 1997: The numerical simulation of nonsupercell tornadogenesis. Part II: Evolution of a family of tornadoes along a weak outflow boundary. J. Atmos. Sci., 54 , 23872415.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, B. D., , and Wilhelmson R. B. , 2000: The numerical simulation of nonsupercell tornadogenesis. Part III: Parameter tests investigating the role of CAPE, vortex sheet strength, and boundary layer vertical shear. J. Atmos. Sci., 57 , 22462261.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., , Straka J. M. , , and Rasmussen E. N. , 2002: Direct surface thermodynamic observations within rear-flank downdrafts of nontornadic and tornadic supercells. Mon. Wea. Rev., 130 , 16921721.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Moncrief, M., , and Miller M. J. , 1976: The dynamics and simulation of tropical cumulonimbus and squall lines. Quart. J. Roy. Meteor. Soc., 102 , 373394.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pietrycha, A. E., , and Manross K. L. , 2003: WSR-88D analysis of vortices embedded within a surface low pressure trough and subsequent convection initiation. Preprints, 31st Int. Conf. on Radar Meteorology, Seattle, WA, Amer. Meteor. Soc., 835–838.

  • Pietrycha, A. E., , and Rasmussen E. N. , 2004: Finescale surface observations of the dryline: A mobile mesonet perspective. Wea. Forecasting, 19 , 10751088.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., 2003: Refined supercell and tornado forecast parameters. Wea. Forecasting, 18 , 530535.

  • Rasmussen, E. N., , and Blanchard D. O. , 1998: A baseline climatology of sounding-derived supercell and tornado forecast parameters. Wea. Forecasting, 13 , 11481164.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Thompson, R. L., , Edwards R. , , Hart J. A. , , Elmore K. L. , , and Markowski P. , 2003: Close proximity soundings within supercell environments obtained from the Rapid Update Cycle. Wea. Forecasting, 18 , 12431261.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., , and Wilson J. W. , 1989: Non-supercell tornadoes. Mon. Wea. Rev., 117 , 11131140.

  • Wakimoto, R. M., , and Atkins N. T. , 1996: Observations on the origin of rotation: The Newcastle tornado during VORTEX 94. Mon. Wea. Rev., 124 , 384407.

    • Crossref
    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

Tornadoes in (a) a high LCL environment over northwest KS on 29 Jun 2000 near Bird City, KS (photo courtesy J. Reed Photography) and (b) a small SRH and high LCL environment over south-central KS on 27 Aug 2004 near Wellington, KS (photo courtesy R. Renfro).

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 2.
Fig. 2.

Lower portion of a skew T–logp diagram of a RUC analysis sounding associated with the high-based tornadic supercell on 29 Jun 2000 shown in Fig. 1a. Temperature profiles (thick solid curves), dewpoint profiles (thick dashed curves), lifted parcel ascents above LCL using a mixed-layer lowest 100-hPa parcel (heavy dotted curves), and positive CAPE (hatched) are shown with important features labeled. For viewing simplicity, the virtual temperature correction is not shown.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 3.
Fig. 3.

As in Fig. 2, but for RUC analysis profiles associated with high-based nontornadic storms that developed on preexisting boundaries in (a) southeast CO on 8 Jun 2004 and (b) west-central TX on 19 May 2003.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 4.
Fig. 4.

Box-and-whisker plots of 0–2-km lapse rate (°C km−1) for F1 and greater intensity tornadoes (338 total cases) in the RUC sounding database used for this study. These are categorized by (a) small compared with typical values of 0–1-km SRH values as defined in the text and (b) high compared with typical LCL heights as defined in the text, using a mixed-layer lowest 100-hPa parcel. Boxes represent the middle 50% of cases in each category, whiskers extend to the 90th and 10th percentiles, and horizontal bars indicate median values.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 5.
Fig. 5.

Box-and-whisker plots showing (a) LFC height (m AGL), (b) CIN (J kg−1), and (c) 0–3-km CAPE (J kg−1) categorized by selected nontornadic and F1–F4 tornadic cases from RUC sounding database in this study, using mixed-layer lowest 100-hPa parcels. All these cases (314 total) had steep 0–2-km lapse rates (≥7°C km−1) along with small SRH and/or high LCL heights as defined in the text. Conventions are as in Fig. 4. Tornadoes of F0 intensity are not shown.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 6.
Fig. 6.

Parameter fields of (a) 0–1-km SRH (m2 s−2), (b) 0–3-km lapse rate (°C km−1), and (c) LFC height (m AGL) from the SPC mesoanalysis at 2300 UTC 27 Aug 2004 over the KS–OK–MO area. Surface features are shown in (a) along with selected parameter values at Winfield, KS, near the surface boundary intersection. Wind barbs in (a) estimate storm motion in knots. Heavy dots in (b) denote axes of steepest lapse rates. (d) The areas where 0–3-km lapse rates ≥8°C km−1 overlap LFC heights ≤2000 m AGL, and tornado tracks and intensities for 0000–0100 UTC 28 Aug 2004.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 7.
Fig. 7.

(a), (b) Video captured by the author showing tornadoes in south-central NE on 24 May 2004. Several tornadoes, including the one in (a), had nonmesocyclone visual characteristics as described in text. The tornado on the right in (b) had mesocyclone visual characteristics, including a rear-flank downdraft (indicated by RFD) and a wall cloud.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 8.
Fig. 8.

Parameter fields of (a) 0–1-km SRH (as in Fig. 6a, with surface features), (b) 0–3-km lapse rate (as in Fig. 6b), and (c) 0–3-km CAPE (heavy lines, J kg−1) from SPC mesoanalysis at 2000 UTC 24 May 2004 over the NE–KS–MO area. Surface vertical vorticity (thin lines, 10−4 s−1) is also shown in (c) with times signs indicating the location of significant vorticity maxima. Wind barbs in (c) indicate surface wind direction and speed (kt). (d) The area where 0–3-km lapse rates ≥8°C km−1 overlap 0–3-km CAPE ≥ 50 J kg−1, and tornado tracks and intensities (F1+ intensities labeled) at 2030–2230 UTC.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 9.
Fig. 9.

Parameter fields of (a) 0–3-km lapse rate (as in Figs. 6b and 8b) and (b) 0–3-km CAPE and surface vorticity (as in Fig. 8c) from SPC mesoanalysis at 1800 UTC 8 Jun 2004 over eastern CO. Surface features are also shown in (b). (c) Composite radar reflectivity at 1926 UTC with relevant features labeled.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Fig. 10.
Fig. 10.

Box-and-whisker plot (conventions as in Fig. 4) of LCL height (m AGL; lowest 100-hPa mixed-layer parcels) from all tornado cases (520 total) in the RUC database used for this study, categorized by location west or east of −98.5° longitude.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF928.1

Table 1.

A summary of RUC profiles associated with nontornadic and tornadic cases in the database used for this study.

Table 1.
Table 2.

A summary of atypical F1–F4 tornadoes associated with small SRH and/or high LCL heights from the database examined in this study, as defined in the text.

Table 2.
Table 3.

A summary of selected median parameter values for F1–F4 tornadoes (338 total cases) categorized by small (<75 m2 s−2) compared with more typical values of 0–1-km SRH values, and by high (>1300 m AGL) compared with more typical LCL heights from the database examined in this study. All computations used mixed-layer lowest 100-hPa lifted parcels.

Table 3.
Table 4.

Summary of selected median parameter values for F2–F5 tornadoes (1531 total cases) categorized and computed as in Table 3, but from observed sounding database (J. Craven 2004, personal communication) provided to the author and used as a comparison tool in this study.

Table 4.
Save