On forecasting severe storms in Alberta using environmental sounding data

Thermodynamic and dynamic parameters computed from observed sounding data are examined to determine whether they can aid in forecasting the potential for severe weather in Alberta. The primary focus is to investigate which sounding parameters can provide probabilistic guidance to distinguish between Significant Tornadoes (F2 to F4), Weak Tornadoes (F0 and F1), and Non-Tornado severe hail storms (≥ 3 cm diameter hail but no reported tornado). The observational data set contains 87 thunderstorm events from 1967 to 2000 within 200 km of Stony Plain, Alberta. Three tornadic thunderstorms with F-scale ratings of F3 and F4 are examined in more detail. A secondary focus is to determine whether sounding data can be used to predict 24 hour snowfall amounts (specifically amounts ≥ 10 cm). Snowfall data covered all of Alberta east of the mountains from October 1990 to April 1993. The major findings were: (a) Significant Tornadoes tended to have stronger environmental bulk wind shear values than Weak Tornadoes or Non-Tornado storms, with a shear magnitude in the 900-500 mb layer exceeding 3 m s-1 km-1. Combining the 900-500 mb shear with the 900-800 mb shear increased the probabilistic guidance for the likelihood of Significant Tornado occurrence. (b) Values of storm-relative helicity showed skill in distinguishing Significant Tornadoes from both Weak Tornadoes and Non-Tornadoes. Significant Tornadoes tended to occur with 0-3 km storm-relative helicity >140 m2 s-2 whereas Weak Tornadoes were typically formed with values between 30 and 150 m 2 s-2. (c) The amount of precipitable water showed statistically significant differences between Significant Tornadoes and the other two groups. Significant Tornadoes had values exceeding 21 mm. Combining precipitable water values with the 900-500 mb shear increased the probabilistic guidance for the potential of Significant Tornadoes. (d) Values of thermal buoyancy, storm convergence, and height of the lifted condensation level provided no skill in discriminating between the three storm categories. (e) The Edmonton tornado case, unlike the Holden and Pine Lake cases, did not feature a prominent synoptic scale moisture front. (f) Observed snowfall amounts showed a roughly linear dependence on the 850 mb temperature, supporting a moisture conservation theory.