P1.1 Environment Characteristics Associated with Tornado Events Near Closed Cold Core 500 Mb Lows

P1.1 ENVIRONMENT CHARACTERISTICS ASSOCIATED WITH TORNADO EVENTS NEAR CLOSED COLD CORE 500 MB LOWS

Jared L. Guyer * NOAA/NWS Storm Prediction Center, Norman, Oklahoma

Jonathan M. Davies Private Meteorologist, Wichita, Kansas

1. INTRODUCTION 2. METHODOLOGY

Previous studies have discussed tornado Tornado time, location, and F-scale damage events that occur in relatively close proximity to closed rating was determined from NWS Storm Data for cold core 500 mb lows (hereafter C500L). Modest tornadoes in the central and eastern United States that measures of low level moisture (namely in terms of satisfied the criterion for C500L tornadoes in DG04. surface dewpoints) and instability can render some For each tornado case, RUC (Benjamin et al. 2004) 00- tornado events near C500L difficult to forecast. Miller hr forecast soundings were gathered coincident with (1972) referred to these events as “type D” patterns. each tornado report. RUC soundings have been shown Subsequent studies by Goetsch (1988), Davies (1993), by Thompson et al. (2003) and Davies (2004) to serve and McDonald (2000) have examined severe as a reasonable proxy for direct observations in the thunderstorms and tornadoes east of the Rocky mesoscale supercell tornado environment. The RUC Mountains associated with “cold core” or “small” (non- soundings were gathered for the nearest available time tropical) tornadic supercell patterns. and location in relation to the tornado event, typically More recently, a preliminary study by Davies within 0-1 hr preceding the tornado and 100 km of the and Guyer (2004, hereafter DG04) examined the tornado location. The C500L tornado database included synoptic scale patterns (also see Davies 2006) and 39 tornadic cases. climatological tendencies associated with C500L A database of 21 non-tornadic C500L null tornadoes in the central and eastern United States over cases was gathered from days when/where the general a five year period (1999-2003). A C500L was defined synoptic pattern and environmental characteristics as at least one closed 30 m contour at 500 mb (Bell and appeared favorable for the possibility of C500L Bosart 1989), with 500 mb temperatures of -10°C or associated tornadoes. In absence of a specific tornado colder. Most common during the transitional spring and report, RUC proximity soundings for C500L null cases fall months, tornadoes that occur within 320 km were gathered based on the most likely location for cold (approximately 200 statute miles) of a C500L were the core tornadoes identified in DG04, which was typically focus of DG04. In close proximity to, and typically east coincident with a surface intersection/wind shift within or southeast of the C500L, DG04 found that tornadoes 320 km of the 500 mb low center (Fig. 1). Although the tend to occur along a boundary intersection focus point, sample size of the C500L tornado and null case events oftentimes in the form of a surface occlusion and/or are relatively small, they appear to provide meaningful wind shift axis (Fig. 1). DG04 noted that C500L tornado statistical results. environments are often characterized by marginal boundary layer moisture (surface dewpoints as low as 3. C500L TORNADOES VS. TRADITIONAL 48-55 °F) and relatively weak instability (e.g. CAPE of SUPERCELL TORNADOES 1000 J/kg or less). Readers of this manuscript are encouraged to consult DG04 and Davies (2006) for Comparisons of environmental characteristics additional discussion on the preliminary patterns and were made between the C500L tornado database and a climatology of C500L tornado events. As a follow-up to version of an existing RUC-derived database by Davies DG04, this study focuses on the thermodynamic and (2004) consisting of 532 “traditional” supercell kinematic characteristics associated with C500L tornadoes. This “traditional” database excluded C500L tornadoes. This includes a comparison of C500L tornado events, tornadoes associated with tornado environments vs. a diverse dataset of traditional hurricane/tropical systems, and “landspout” non- supercell tornado environments (section 3) and C500L supercell tornadoes. Table 1 provides a comparison tornado null events (section 4). between C500L tornadoes and typical supercell tornadoes showing median and 25th/75th percentiles for various thermodynamic and kinematic parameters.

3.1 Instability * Corresponding author address: Jared L. Guyer NOAA/NWS Storm Prediction Center, National Weather In comparison to traditional supercell tornado Center, 120 David L. Boren Blvd, Suite 2300, Norman, environments, 500L tornadoes are associated with OK 73072; e-mail: [email protected] considerably lesser amounts of convective instability. Lowest 100 mb MLCAPE values with C500L events

Fig. 1. Composite diagram adapted from Davies and Guyer (2004) with examples of surface features associated with an increased likelihood of C500L tornado events. Most likely area for C500L tornadoes denoted by green oval.

0-6 0-1 100 100 EL th 100 mb 0-3 km km km 0-1 km Sig Tor 75 Percentile SBCAPE mb ML mb ML Height Median MLCAPE MLCAPE Bulk Bulk SRH Parameter th (J/kg) LCL LFC (m 25 Percentile (J/kg) (J/kg) Shear Shear (m2/s2) (new) (m) (m) AGL) (kt) (kt) 1069 1756 160 48 25 131 1091 1431 9606 0.6 C500L 755 1221 108 33 17 76 915 1108 8455 0.2 Tornadoes 391 765 80 20 12 33 764 966 7483 0.1

Typical 2729 3450 117 53 30 253 1405 2159 12457 2.1 Supercell 1787 2457 76 45 22 162 1045 1681 11649 1.1 Tornadoes 1015 1456 39 37 15 90 770 1265 10545 0.5 Table 1. Tabular comparison of the median (in bold) and 25th/75th percentiles for environmental parameters for C500L tornadoes vs. typical supercell tornadoes. were roughly one-third, with surface-based (SB) CAPE of cases with 15 kts or less of 0-6 km shear. Low level values roughly one-half, of the broader supercell 0-1 km shear computations exhibit similar tendencies tornado dataset (Table 1). The median 100 mb toward lower values for C500L cases. C500L tornadoes MLCAPE was 755 J/kg and the SBCAPE 1221 J/kg for were associated with a median 0-1 km bulk shear of 17 C500L tornadoes. In contrast, the median 100 mb kt, with 0-1 km Storm Relative Helicity (SRH) median of MLCAPE for typical supercell tornado environments was 76 m2/s2. 1787 J/kg, with a median SBCAPE of 2457 J/kg. 3.3 Additional Characteristics 3.2 Vertical Shear Although the distribution of ML LCL heights C500L tornado environments are associated associated with C500L tornadoes were slightly lower with weaker vertical shear as compared to traditional than other supercell tornadoes, considerable overlap supercell tornado environments, with respect to both existed with little discrimination between the C500L and deep layer and low level (i.e. 0-1 km) shear. The traditional supercell tornado datasets. However, a median deep layer (0-6 km) bulk shear was 33 kt for greater separation was observed for ML LFC heights. C500L tornadoes, with 25th and 75th percentile values of The median ML LFC height for C500L tornadoes was 20 kt and 48 kt, respectively. This is likely attributable to 1108 m (3635 ft), as compared to the 1681 m (5515 ft) the weaker mid tropospheric winds closer to a vertically- median for typical supercell tornadoes. Parallel to the stacked closed low, in addition to the low topped nature observations of (moisture and) buoyancy being (i.e. EL height median of 8455 m or 27740 ft - Table 1) “concentrated” in the low levels, the typical LCL-LFC of C500L events. With emphasis that C500L tornado height separation was found to oftentimes be only 200- events can occur in rather weak deep layer vertical 400 m (650-1300 ft) for C500L tornado environments. shear, nearly one-third of the C500L cases occurred with 0-6 km bulk shear of 30 kts or less, including 20% 0-1 km th 0-3 km 50 mb 50 mb 0-1 km 75 Percentile 0-1 km SBCAPE 0-1 km Bulk SB LCL Median SBCAPE MLCAPE ML LFC SRH th SB VGP (J/kg) SB EHI 2 2 Shear (m) 25 Percentile (J/kg) (J/kg) (m) (m /s ) (kt) 266 0.57 1756 1221 1.1 1282 131 25 774 C500L 210 0.43 1221 915 0.7 948 76 17 560 Tornadoes 159 0.24 765 654 0.2 713 33 12 393

177 0.29 1303 913 0.4 1264 98 22 845 C500L 132 0.18 759 727 0.2 1104 38 16 716 Null 99 0.16 495 340 0.0 993 9 8 492 Table 2. Tabular comparison of the median (in bold) and 25th/75th percentiles for potential discriminators between C500L tornadoes and C500L null tornado events.

4. C500L TORNADOES VS. C500L NULL TORNADO tornado cases (Fig. 3). The median 0-3 km SBCAPE EVENTS was 210 J/kg for tornado cases, with 75% of events in excess of 160 J/kg. In contrast, null cases featured less Statistical analysis revealed a relative buoyancy “concentrated” in the lowest 0-3 km layer, with emphasis on thermodynamic and kinematic a median of 132 J/kg and 75% of cases ≤175 J/kg 0-3 computations derived from the lowest few kilometers for km SBCAPE. the discrimination of C500L tornado events vs. nontornadic null cases. A tabular summary above (Table 2) SBCAPE compares the environmental characteristics of C500L tornado events vs. null cases, including median, and 3000 25th/75th percentiles associated with each parameter. 2500 2260 4.1 Lifted Parcel Discussion 2000 1756 1500

Findings in this study stress the importance of J/kg 1411 1303 1221 choosing the “best” lifted parcel choice based on the 1000 meteorological situation, whether it be a surface-based 765 759 631 (SB) parcel or mixed layer (ML) parcel calculation of 500 495 347 mean conditions through a given layer. Likely owing to the more modest buoyancy concentrated at lower 0 C500L Tornado C500L Null levels, discrimination skill typically decreased when using a lowest 100 mb ML calculation in C500L environments. This suggests that the commonly used Fig. 2. Box and whiskers diagram of surface-based lowest 100 mb MLCAPE is too deep a mean lifted layer (SB) CAPE (J/kg) for C500L tornadoes vs. null events. in C500L situations. There was 25-60% more total Median with box denoting the 25th/75th percentiles, with CAPE and 50-90% more low-level CAPE (0-3 km) when outer whiskers represent the 10th/90th percentiles of using SB or 50 mb ML parcels. This not only applies to values. CAPE calculations directly, but also derived parameters

(i.e. EHI, VGP etc.) as well. Statistically, surface-based 0-3 km SBCAPE (SB) parcel calculations were found to be best, followed by a 50 mb parcel. Based on the findings in this study, 350 314 the authors caution against using 100 mb ML parcel 300 calculations (whether it be CAPE or derived fields) in 266 C500L scenarios, with a preference for a SB or lowest 250 220 210 50 mb ML lifted parcel. 200 177 J/kg 150 159 4.2 Instability 124 132 100 99 78 The median surface SBCAPE associated with 50 C500L tornadoes was 1221 J/kg, with 25th/75th 0 percentiles of 766-1756 J/kg, which was approximately C500L Tornado C500L Null 30-40% greater than C500L null events (Fig. 2) This study found that 0-3 km SBCAPE was the best discriminator between C500L tornadoes and C500L null Fig. 3. Same as Fig. 2, except 0-3 km SBCAPE.

4.3 Vertical Shear

As discussed in section 3.2, C500L tornado was observed in LCL height (Fig. 8), LFC heights events tend to occur in relatively weak deep layer shear, (Davies 2004) were lower (median of 948 m/3110 ft) with a median 0-6 km bulk shear of 33 kt. For C500L with C500L tornadoes (Fig. 9). tornadoes vs. null events, 0-6 km bulk shear exhibited virtually no skill in distinguishing between the two 0-1 km Vorticity Generation Parameter (VGP) environments (not shown). Low level shear computations also exhibited little discrimination, with 0.70 only slightly higher values of 0-1 km bulk shear (Fig. 4) 0.62 0.60 and 0-1 km Storm Relatively Helicity (SRH) for C500L 0.57 0.50 tornado cases (Fig. 5). 0.48 0.43 0.40

0-1 km Bulk Shear 0.30 0.29 0.24 0.20 35 0.18 0.16 30 30 0.10 0.09 28 0.07 25 25 0.00 22 C500L Tornado C500L Null 20

kt 17 16 15 Fig. 6. Same as Fig. 2, except 0-1 km Vorticity 12 10 Generation Parameter (VGP). 9 8 5 5

0 0-1 km Energy-Helicity Index (EHI) C500L Tornado C500L Null 2.0 1.8 1.7 Fig. 4. Same as Fig. 2, except 0-1 km bulk shear (kt). 1.6 1.4 1.2 0-1 km Storm Relative Helicity (SRH) 1.1 1.1 1.0 0.8 250 0.7 0.6 0.4 0.4 200 193 0.2 0.2 0.2 172 0.0 0.0 0.0 0.0 150 C500L Tornado C500L Null 131 m2/s2 100 98 76 Fig. 7. Same as Fig. 2, except 0-1 km Energy-Helicity 50 33 38 Index (EHI).

2 9 0 C500L Tornado C500L Null 50 mb ML LCL Height

1800 1663 Fig. 5. Same as Fig. 2, except 0-1 km Storm Relative 1600 2 2 Helicity (SRH - m /s ). 1400 1420

1200

1000 1034

m 889 4.4 Additional/derived parameters 800 831 722 600 603 646 528 It was found that the Vorticity Generation 400 447

Parameter (VGP – Rasmussen and Blanchard 1998) 200 computed for the 0-1 km layer was a good discriminator 0 between C500L tornado events and null cases (Fig. 6). C500L Tornado C500L Null The median 0-1 km VGP for tornadoes was 0.43, with 75% of cases greater than 0.24. Null events were associated with a median 0-1 km VGP of 0.18, with 71% Fig. 8. Same as Fig. 2, except 50 mb ML LCL Height of events below 0.24. It was also found that 0-1 km (m). Energy-Helicity Index (EHI – Rasmussen 2003) values provided some discrimination between C500L tornadoes vs. null cases (Fig. 7). While relatively little variance 50 mb ML LFC Height MXACCL

2250 0.25 0.23 2076 2000 0.21 0.20 1750 0.17 0.16 1500 0.15 0.15 1448

m 0.13

1250 1282 1264 m/s2 0.12 0.10 0.11 1104 1000 993 0.08 948 0.07 0.05 750 713 650 636 500 0.00 C500L Tornado C500L Null C500L Tornado C500L Null

Fig. 9. Same as Fig. 2, except 50 mb ML LFC Height Fig. 10. Same as Fig. 2, except MXACCL (m/s2). (m).

4.5 Vertical Distribution of CAPE and Maximum weaker vertical shear through a deep layer, and lower Parcel Acceleration (MXACCL) storm tops (EL heights). Comparing traditional supercell tornado environments with C500L tornado cases, Because CAPE is located closer to the C500L events cases tended to have considerably higher ground in most C500L settings, an experimental amounts of 0-3 km SBCAPE, often the result of a “fatter” computation (MXACCL) was devised to estimate the positive CAPE area located relatively low in the vertical maximum vertical parcel acceleration due to vertical profile (e.g. closer to the ground, see Davies 2006). buoyancy distribution through increasing depths in a This suggests that low-level parcel accelerations and sounding. This was found to be potentially useful for resultant stretching may be important for tornado discriminating between C500L tornado environments development in events associated with C500L cases. It and C500L tornado null cases. MXACCL is somewhat was found that C500L tornado cases tended to higher 0- similar to CAPE density (Blanchard 1998), but rather 1 km VGP and 0-1 EHI values, and slightly higher 0-1 than dividing the total CAPE by the entire depth from km shear and 0-1 km SRH, in comparison to null C500L LFC to EL, CAPE from the LFC to ascending 25 mb cases. increments in the sounding is summed and stored in a Additionally, it was found that CAPE table. Then each accumulated CAPE value in the table calculations using a shallower mixed-layer lifted parcel is divided by the corresponding depth it represents. (e.g. the lowest 50 mb), or surface-based (SB) parcels, This yields a series of average accelerations (units m/s2) appear better suited for assessing thermodynamic from LFC to a particular 25 mb level as one ascends characteristics in tornado events associated with C500L. higher in the sounding. The largest of these average Given the relatively shallow moisture layer indicated by accelerations is called MXACCL. Although the RUC soundings in many such events, the lowest experimental, the potential advantage of MXACCL over 100 mb often may be too deep a mixed layer to properly CAPE density is that it more directly considers represent buoyancy and associated environment acceleration through portions of a sounding where characteristics in such settings. CAPE increases rapidly with height, rather than averaging the acceleration through the entire depth of Acknowledgements. The authors would like to the LFC to EL layer. thank Steve Weiss (SPC) for his thorough review of this The MXACCL, using a surface-based (SB) manuscript. parcel appeared to exhibited some skill in distinguishing between C500L tornado environments and C500 null cases. The median value for MXACCL in C500L 6. REFERENCES tornado events was found to be 0.16 m/s2, while the median MXACCL for C500L null tornado events was Bell, G. D., and L. F. Bosart, 1989: A 15-year 0.13 m/s2 (Fig. 10). climatology of Northern Hemisphere 500 mb closed cyclone and anticyclone centers. Mon. Wea. Rev., 117, 2142-2163. 5. CONCLUSIONS Benjamin, S. G., D. Dévényi, S. S. Weygandt, K. J. Within the spectrum of tornadic supercells, Brundage, J. M. Brown, G. A. Grell, D. Kim, B. E. tornadoes associated with C500L are typically Schwartz, T. G. Smirnova, T. L. Smith, and G. S. characterized by lower amounts of convective instability, Manikin, 2004: An Hourly Assimilation–Forecast Cycle: The RUC. Mon. Wea. Rev., 132, 495–518.

Blanchard, D.O., 1998: Assessing the vertical distribution of convective available potential energy. Wea. Forecasting, 13, 870-877.

Davies, J. M., 1993: Small tornadic supercells in the central plains. Preprints, 17th Conf. on Severe Local Storms, St. Louis, MO, Amer. Meteor. Soc., 305-309.

Davies, J. M., 2004: Estimations of CIN and LFC associated with tornadic and nontornadic supercells. Wea. Forecasting, 19, 714-726.

Davies, J.M., and J.L. Guyer, 2004: A preliminary climatology of tornado events with closed cold core 500 mb lows in the central and eastern United States. Preprints, 22nd Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc. (CD ROM).

Davies, J. M., 2006: Tornadoes with cold core 500 mb lows. Wea. Forecasting, accepted for publication late 2006 or early 2007.

Goetsch, E. H., 1988: Forecasting cold core severe weather outbreaks. Preprints, 15th Conf. on Severe Local Storms, Baltimore, MD, Amer. Meteor. Soc., 468- 471.

McDonald, M., 2000: Cold core tornadoes: A forecasting technique. Internal report, Prairie Storm Prediction Centre, Environment Canada, 7 pp.

Miller, R. C., 1972: Notes on analysis and severe-storm forecasting procedures of the Air Force Global Weather Central, AWS Tech. Rep. 200 (rev.), Air Weather Service, Scott AFB, IL, 190 pp.

Rasmussen, E. N., and Blanchard, D. O., 1998: A baseline climatology of sounding-derived supercell and tornado forecast parameters. Wea. Forecasting: 13, 1148-1164.

Rasmussen, E. N., 2003: Refined supercell and tornado forecast parameters. Wea. Forecasting: 18, 530-535.

Thompson, R. L., R. Edwards, J. A. Hart, K. L. Elmore, and P. Markowski, 2003: Close proximity soundings within supercell environments obtained from the Rapid Update Cycle. Wea. Forecasting: 18, 1243–1261.