Cold Core Tornadoes: A Forecasting Technique

 

 Michael W. McDonald

Meteorological Service of Canada

Prairie Storm Prediction Centre, Winnipeg, Manitoba

 

ABSTRACT

 

Cold core funnels and tornadoes are not an uncommon occurrence in the Canadian prairies during the summer season. A cold core environment is one that exhibits a vertically stacked low from the surface to 500 mb bounded by an area with a 700 mb temperature equal to or colder than 0C. In the summer of 2000 there were 30 such cold core cases over the Prairies producing 21 reported funnel clouds and 11 tornadoes. There was likely a significantly higher number of cold core events but since 21 of the 30 cold core cases occurred over the sparsely populated northern Prairies, many additional events likely went unobserved. The limited resolution of numerical models and GRIB data and the mesoscale low level boundary interactions that ultimately account for the development of cold core funnels and tornadoes makes forecasting such events very difficult. A forecasting technique utilizing actual and model data focusing on low level moisture, instability, lift and convergence is applied to a case on 11 June 2000 that generated a cold core tornado in Saskatchewan. This technique does not determine where cold core events will occur but rather if they do occur, the area this technique defines would be the most favoured area for development.       

                                              _________________________

  

1.  Introduction

 

       Forecasting the development of cold core funnels and tornadoes have been one of the more difficult tasks facing the severe weather meteorologists during the summer season across the Prairies. It is often easier to conceptually anticipate supercell-producing tornadoes given a more energetic synoptic environment associated with these storms than the more benign regime associated with a cold core environment. During the summer season 2000 (May 15 to Sept 15) there were 30 cold core days across the Canadian Prairies. To be classed a cold core day there must be present a vertically stacked low from the surface to 500 mb with 700 mb temperatures of  0 C or colder. There were 11 reported tornadoes associated with these cold core regimes and 21 reported funnel clouds.  Although all of these tornadoes, as is the case with the vast majority of cold core events, are F0 in intensity they can, and have caused property damage and injuries. What makes forecasting these phenomena difficult is the mesoscale interactions that actually contribute to cold core tornado formation. The initial circulation and parent convective cloud associated with the tornado development forms in response to the collision of mesoscale surface boundaries, including convergence zones, shear zones, and outflow boundaries. Their incipient Doppler velocity signatures are very small and difficult to distinguish from normal background noise and low level reflectivities are quite weak (0 - 30 dBz).  This paper will review the processes involved with cold core tornadoes and examine a forecasting technique evaluated on 3 tornado producing cold core days during the summer of 2000 in delineating areas that would be prone to cold core tornado development. A full case study will also be included stepping through the technique process. 

 

 

2.  Cold core processes and methodology

 

       Even though non-supercell tornadoes have been documented for a number of years, our knowledge concerning them is rather diffuse.  The majority of tornadoes reported over the Prairies tend to be small, weak (F0-F1), and short-lived (5-10 min average lifetime). In a cold core tornado scenario, wall clouds are rarely observed; the funnel usually emerges from rapidly growing convective cloud displaying a well defined cloud base, indicative of a strong updraft. This lack of a pronounced cloud lowering suggests the lack of a mesocyclone and possible differences in the generating mechanisms in relation to its supercell tornado sibling. Bluestein (1985) coined the term “landspout” to define this type of tornado that develops in a benign synoptic environment under a rapidly developing convective cloud. In addition, the initial circulation is not associated with a thunderstorm mesocyclone or rear flank downdraft, and it develops from the ground up to cloud base along colliding mesoscale boundaries. The typical synoptic environment of a cold core landspout event is characterized by weak dynamic forcing with weak speed and directional wind shears (0-6 km), especially when compared to the typical mesocyclone-producing synoptic environment. The initial circulation is formed in the boundary layer near the intersection of two mesoscale surface boundaries where vertical vorticity is concentrated. The circulation intensifies due to vertical stretching as it becomes collocated with strong updrafts associated with rapidly developing convective clouds. A source of low level vorticity appears to be critical in the formation of these “landspouts”. The forecaster should be wary of semi-stationary convergence or shear zones associated with moderate low level vorticity and moisture convergence. In such cases it appears that boundary layer moisture convergence may help to create a potentially unstable, mesoscale area capable of producing landspouts. In addition, the concentration of low level shear and vorticity allows for the development of the incipient circulations  (Brady and Szoke, 1988). Figure 1 shows the schematic model of the life cycle of a “landspout”. In the beginning stage the horizontal shear across the convergence boundary has resulted in shearing instabilities labeled A, B, and C. At the same time, cumulus clouds have formed over the boundary owing to the forced uplift. The clouds continue to develop as the small vortices propagate along the boundary in the middle stage. In the final stage, vortex C has collocated with the updraft of a towering cumulus and developed into a tornado under the influence of vortex stretching. Since landspouts form from the ground up, technically all landspouts are tornadoes. It is the strength of the updraft and low level circulation that ultimately determines whether the condensation funnel can be detected on the ground.

 

 

Figure 1. Schematic model of the life cycle of the "landspout". The black line is the convergence boundary. Low level vortices are labeled A, B, and C.  (Wakimoto and Wilson 1989)

In most cases the initial, low level circulations are small, and therefore are extremely difficult to detect on Doppler as their velocity signatures are too small to distinguish from normal background noise. The convective cloud is rapidly developing aloft when the tornado forms, and therefore low level reflectivities are often quite weak (0-30 dBz). This makes for radar detection of these tornadoes nearly impossible.  Dissipation of these tornadoes occurs rapidly as precipitation loading induces a downdraft, thus vertically decoupling the mechanisms by which the vortex had maintained itself.   

 

 

3.       Forecasting technique

 

         Certainly the skill in forecasting cold core events is quite limited given the weak dynamical forcing and convectively driven mechanisms associated with such events. Since mesoscale processes dominate the eventual development of cold core tornadic events, synoptic scale model data offers little in the prediction of these events.  The current resolution of the GEM model cannot as of yet accurately depict the mesoscale dynamics associated with cold core tornadoes, so the forecaster is left to interpret synoptic data and acquired knowledge to ascertain possible clues in forecasting these events. What must be stressed here is that because of these limitations, this technique does not define that cold core funnels or tornadoes will occur, but if they do occur they would likely occur in the formulated area. Even if we can define an area which may be prone to cold core funnels, it would still be difficult to discriminate which cloud element would produce a funnel and which will not. Since enhanced updraft and low level vorticity seems to be favoured conditions for cold core tornadic events, attention should be focused on cells depicting stronger reflectivity on radar, indicative of these areas of enhanced updraft and vorticity.

         For this paper, a cold core case was defined as one that exhibited a vertically stacked low from the surface through 500 mb bounded by an area with 700 mb temperatures colder than 0C. For the 2000 summer season across the Canadian Prairies (May 15 to Sept. 15), 30 such cold core cases were identified. 21 were centred over the northern Prairies provinces while 9 were centred over the southern Prairies. Table 1 breaks down the reported funnels and tornadoes associated with these cold core events by province for the 2000 summer season. Since the vast majority of the cold core days were over the northern half of the Prairie provinces, a region dominated by sparsely populated boreal forest, the total number of reported funnels and tornadoes seem rather low, but in reality there were likely considerably more events than reported simply because there was no-one to see them. This is no more strikingly evident than in Manitoba where they had the most cold core days of the three Prairie provinces but only 1 reported funnel cloud. 11 of the 13 cold core days were centred over the vastly uninhabited areas of northern Manitoba. So there were likely a larger number of funnel clouds or tornadoes associated with these cold core events but went undetected by the human eye.   

 

 

 

Cold core Days

Reported funnels

Reported tornadoes

Alberta

11

11

5

Saskatchewan

6

9

6

Manitoba

13

1

0

Total

30

21

11

 

Table 1.  2000 Summer season (May 15 to Sept. 15) cold core events across the Canadian Prairies.                 

 

           For the subject of this paper, 3 tornado producing cold core events in the summer of 2000, which a full data set of numerical model and actual data were available, numerous fields were examined to see what was common in each case to see if a defined area could be found that incorporated the region that had a reported cold core tornado. Table 2 outlines the best fields used in assessing the area to delineate the potential for cold core funnels or tornadoes. Examining the data, 19 of the 21 reported funnel clouds and all 11 of the reported tornadoes occurred in the eastern quadrant of the vertical low. This is not all too surprising, as typically with vertical lows, the eastern quadrant is drier and more unstable, conducive to convective development, while stratiform cloud and precipitation is common in the western and northern quadrants of vertical lows.

 

·         Surface   -   within 100 km of isobaric troughs.

                                      -   within 100 km of main moisture (Td) axis.

 

·      700 mb     -   temperature < 0C

 

·        850 mb     -  within 100 km of main troughs.

                                       -  within 100 km of main vorticity axis.

                                       -  vertical velocity of  >= 2 mbars/sec.

 

·        Lifted Index   -   LI  <=  0.

 

·        CAPE   -    >=  300 J/kg

 

·        Moisture Flux Convergence (MFC) - within 100 km of main MFC axis.

 

 

               Table 2.  Meteorological fields used in cold core tornado forecasting technique.

 

 

4.       Case Study  -  11 June 2000.

 

         The above mentioned fields are applied to a case on 11 June 2000 where a cold core tornado was reported near Radisson, Saskatchewan around 2100 UTC. For each parameter, the area satisfying the prescribed limits are shaded. The shaded area on the composite map comprises the area where the union of all of the satisfying parameters are present. The location of the tornado is depicted by “*T” on the various charts.        

 

a)       1200 UTC GEM model on 11 June 2000 for T+0 and T+12hr time frame. Of note is the drier air on the eastern quadrant of the vertical low leading to the potential for convection and conversely the stratiform cloud and precipitation depicted on the western quadrant of the low.

 

   

 

b)  2100 UTC surface pressure and winds.  Surface trough depicted by dark dashed line.

 

 

 

c)       2100 UTC surface dew point temperature. Moisture axis depicted by dark dashed line.

 

 

 

d)      2100 UTC 850 mb heights and wind. Main trough depicted by dark dashed line.

 

 

e)       2100 UTC 850 mb vorticity. Main vorticity axis depicted by dark dashed line.

 

 

 

 

f)        2100 UTC 850 mb vertical velocity. Area shaded depicts upward velocities >= 2 mbars/sec.

 

 

g)      2100 UTC Lifted index. Area shaded depicts LI <= 0.

 

 

 

 

h)      2100 UTC Buoyant energy. Area shaded depicts CAPE >= 300 J/kg.

 

 

i)        2100 UTC Moisture flux convergence.  The shaded area depicts the axis of greatest moisture convergence.

 

 

 

j)        2100 UTC Composite chart. Area shaded depicts location common to all of these parameters.

       700 mb 0C isotherm is indicated by the dashed line separating cold core characteristics to the west of 

       the line from the more thermodynamically favoured  supercell regime east of the line. Of note is the

       report of several supercell derived tornadoes in the Minot, North Dakota area at the same time as the

       cold core tornado in Saskatchewan. 

 

 

 

5.       Discussion and summary   

 

         Forecasting the development of cold core events is certainly a difficult procedure. There are basic similarities between the cold core tornado and its supercell brethren. Moisture, instability and lift are necessary ingredience of both but on different scales. The cold core tornado is usually preceded by a low level circulation likely resulting from shearing instabilities and subsequently develops into a deeper circulation under the influence of a vigourous updraft. The vorticity of the cold core tornado originates at low levels as small vortices produced by the shearing instability along a low level convergence boundary. These small, low-level vortices can move along a convergence boundary (e.g. a surface trough) and may become collocated with the updraft of a building towering cumulus, developing into a tornado under the influence of vortex stretching. These cold core tornadoes thus occur in the early updraft dominated stages of the convective process with dissipation occurring as precipitation loading induces a downdraft, opposing the upward spiraling motion of the vortex with downward motion and low level divergence. Cold core events occur in synoptically weak dynamical situations and the mechanisms favouring development work at such a small scale, the current resolution of the numerical models and GRIB data cannot determine whether funnels will occur or not. In this forecast technique, the main emphasis is to try to discriminate what areas would potentially be a favoured area for cold core development. The key here is this technique does not conclude that cold core events will occur in the defined area but rather if they do occur they will likely occur in this area. This technique focused on the main low level convergent zones and the instability and moisture within these main zones for possible development. An arbitrary value of 100 km on either side of the main convergent zones was used to account for the mesoscale boundaries within these main synoptic boundaries. The mesoscale boundaries that should be monitored for possible tornadogenesis are: boundaries having strong horizontal shear across them, the collision point of two or more boundaries, or areas of rapidly developing convection near a suspected low level circulation. The examination of the three cold core tornado cases for 2000 showed all of the tornadoes occurred on the drier eastern quadrant of a vertically stacked low near main low level convergent, instability and moisture axes. For the 11 June 2000 case this technique also depicted an area over North Dakota that had potential for tornadogensis. Several supercell tornadoes where reported in the Minot, North Dakota area at that time in the more dynamically enriched atmosphere in that region. This technique will be tested again in the coming summer seasons to assess the validity of these parameters and hopefully assist the forecaster in at least focusing on a smaller area when trying to determine where cold core events would more likely occur.                         

 

 

6.       References

 

Bluestein, H. B., 1985: The formation of a “landspout” in a “broken-line” squall line in Oklahoma.

       Preprints, 14th Conference on Severe Local Storms, Indianapolis, Amer. Meteor. Soc., 267-270.

 

Brady, R. H., and E. J. Szoke, 1988: The landspout - A common type of northeast Colorado tornado.

       Preprints, 15th Conf. on Severe Local Storms, Baltimore, Amer. Meteor. Soc.., 312-315.

 

Brady, R. H., and E. J. Szoke, 1989: A case study of nonmesocyclone tornado development in northeast

       Colorado: Similarities to waterspout formation. Mon. Wea. Rev., 117, 843-856.

 

Cooley, J. R., 1978: Cold air funnel clouds. Mon. Wea. Rev., 106, 1368-1372.

 

Golden, J. H., 1971: Waterspouts and tornadoes over south Florida. Mon. Wea. Rev., 99, 146-154.

 

Wakimoto, R. M., and J. W. Wilson, 1989: Non-supercell tornadoes. Mon. Wea. Rev., 117, 1113-1140.