APRIL 2001 NOTES AND CORRESPONDENCE 261 NOTES AND CORRESPONDENCE Estimating Downburst-Related Maximum Surface Wind Speeds by Means of Proximity Soundings in New South Wales, Australia BART GEERTS University of Wyoming, Laramie, Wyoming 7 April 2000 and 5 September 2000 ABSTRACT A regional climatology of strong wind gusts associated with thunderstorms is presented, and the ability to estimate gust strength from ambient conditions is tested. Strong wind events were selected for 10 stations in New South Wales, Australia, from anemograph records and coincident thunderstorm reports. Most events took place between midafternoon and late evening and during the warmer months of the year, which is broadly consistent with the occurrence of severe thunderstorms in general. One sounding-based index, designed to predict the strength of microbursts, proves to be of limited value in predicting the magnitude of strong convective gusts, even of short-lived gusts. A modi®ed index that combines the microburst index with upper-level wind speed is more useful. 1. Introduction ment rate generally decreases as the downburst ap- proaches the ground (Wakimoto 1985). Foster (1958) Extreme wind gusts associated with thunderstorms used proximity soundings to estimate the effect of evap- are dif®cult to forecast and there is a dearth of opera- orative cooling: a downdraft was assumed to start at the tional methods that reliably forecast the magnitude of ``level of free sinking,'' where the mean moist adiabat thunderstorm wind speeds. In part, the problem stems (average between parcel and environment at 500±700 from the range of mechanisms producing strong con- hPa) intersected the ambient temperature. He also as- vective gusts. Other than tornadoes, which are not ad- sumed that this downdraft remained saturated to the dressed in this study, these are mainly the dif¯uence of ground. He found a poor correlation between gust downbursts at the surface, but also density current prop- speeds estimated from soundings and those observed in agation and vertical transfer of horizontal momentum. 100 thunderstorm events; the estimated values generally Almost all of the kinetic energy in a downdraft converts exceeded the observed ones. After careful validation and to horizontal kinetic energy so the maximum gust speed adjustment of proximity sounding data, Pino and Moore near the edge of the downburst, upon impact, should be (1989) repeated Foster's procedure, but they found an about the same as the maximum downdraft, however even weaker correlation for 47 storms that produced differences are possible depending on the depth of the both extreme gusts and hail. Wakimoto (1985) dem- out¯ow and the downdraft diameter. Negative convec- onstrated that Foster's method (i.e., moist-adiabatic de- tive-scale accelerations are largely due to negative buoy- scent) tends to overestimate the strength of ``dry'' mi- ancy, which in turn is mostly explained by evaporative crobursts, because the sinking air parcel may run out of cooling, melting, and precipitation loading (Houze water before reaching the ground, and he agreed that 1993, p. 36). Evaporative cooling dominates especially this method is also unlikely to accurately predict the when the layer below cloud base is deep and well mixed strength of a ``wet microburst'' (i.e., a small downburst (Wakimoto 1985), and the incipient downdraft is rich associated with a thunderstorm) (Wakimoto 1985, p. in liquid/frozen water (Srivastava 1985). Entrainment 1141). Kessinger et al. (1988) estimated that water load- below the downdraft base can further accelerate the ing accounts for ;20% of the observed downward ac- downdraft if the ambient wet-bulb temperature is lower celeration in downbursts associated with a Colorado than that of the downdraft parcel; however, the entrain- thunderstorm. Srivastava (1987) showed that the melt- ing of graupel/hail may be a signi®cant energy source, and both observations (Wakimoto and Bringi 1988) and Corresponding author address: Dr. B. Geerts, Dept. of Atmo- numerical simulations (Proctor 1989) have demonstrat- spheric Sciences, University of Wyoming, Laramie, WY 82071. ed that microbursts tend to start at the melting level, E-mail: [email protected] rather than at Foster's level of free sinking. q 2001 American Meteorological Society 262 WEATHER AND FORECASTING VOLUME 16 Density current propagation and downward transfer porating upper-level wind speed into a sounding-based of horizontal momentum are related to convective predictor of maximum surface winds near nontornadic downdrafts; that is, they are thunderstorm out¯ow phe- thunderstorms are explored in section 4. The climatol- nomena. A convectively generated cold pool builds due ogy of extreme gusts in NSW is presented ®rst. to moist-adiabatic descent and entrainment of midtro- pospheric low-u air. It then spreads at a speed propor- e 2. The data and their climatology tional to the square roots of its temperature de®cit and its depth. Deep, intense cold pools may be long lived The data gathering objective was to select strong gusts and generate waves that may cause strong gusts (Weck- associated with thunderstorms. First anemographs were werth and Wakimoto 1992). Downdrafts do carry the examined and only gust events of 20.5 m s21 (40 kt) or ambient momentum from the incipient and entraining more were selected. This threshold is twice as high as levels to the surface, so momentum transfer from the the one that de®nes a microburst (Wakimoto 1985). lower (e.g., Brandes 1977) to middle troposphere is a Anemographs were mainly from Dynes anemometers potentially large source of strong surface winds, yet its that continuously record wind speed and direction and contribution to strong winds underneath thunderstorms have response times of about 2 s. The gust climatology is poorly understood (Johns and Doswell 1992, p. 596). is not restricted to a certain season (as in Harnack et al. Long swaths of nonrotating damaging wind can be 1997); however, severe thunderstorms are rare in the produced through a combination of mesoscale and mi- colder months in NSW (Grif®ths et al. 1993; Geerts and croscale processes in bow echoes (Johns and Hirt 1987; Noke-Raico 1995). Gusts were attributed to deep con- Weisman 1993; Przybylinski 1995). Strong surface vection if either a thunderstorm was reported at the sta- winds in bow echoes result from a combination of down- tion with the anemometer (``locally'' or ``in the vicin- ward momentum transfer, cold-pool dif¯uence, and a ity'') within an hour of the gust event, or, in the absence horizontal perturbation pressure gradient, creating a of hourly reports, if the strong winds were short-lived rear-in¯ow jet that may descend to the surface (Weisman and marked by sudden, large, but unsustained changes 1992). In most storm systems damaging winds result in wind speed and/or direction. Rather than using ob- from out¯ow processes, but in the case of supercell jective criteria, events were selected manually, some- storms in¯ow winds may also be damagingly strong times with the aid of corresponding surface charts, to (Johns and Doswell 1992). In mesoscale convective sys- exclude wind maxima associated with strong frontal dis- tems/complexes, mesoscale circulations induced by hor- turbances without deep convection, but to include gusts izontal surface pressure gradients may contribute to due to lines of thunderstorms found in prefrontal strong surface winds (Vescio and Johnson 1992; troughs. Deep convection commonly occurs in a pre- Schmidt and Cotton 1989). frontal trough in southeastern Australia (Speer and Strong wind gusts associated with isolated deep con- Geerts 1994; Kraus et al. 2000). During the warmer vection in a weakly sheared environment are almost months (when most strong gust events occurred) cold always due to downbursts, especially microbursts (Fu- fronts rarely penetrate into inland NSW, but strong jita 1990; Atkins and Wakimoto 1991; Proctor and winds, sometimes exceeding 21 m s21, may affect the Bowles 1992). Based on the ideas of Proctor (1989) and coastal strip in association with a ``southerly burster'' Wolfson (1990), McCann (1994) developed a micro- (Colquhoun et al. 1985). Most events included lasted burst-gust prediction index, which he named WlNDEX less than 20 min, with a maximum of 33 min. There (Wind Index). WlNDEX captures the environmental was no strict duration limit of events used in the selec- conditions leading to dry or wet microbursts, but it does tion process, although most long-duration events were not estimate extreme surface winds in thunderstorms in interpreted as frontal and therefore were excluded. None general, nor does it gauge storm probability. WlNDEX of the extreme wind events were associated with a re- was not designed to estimate gust strength in more or- ported tornado, which does not exclude the possibility ganized storms, where processes other than downdraft that some gust events were due to rotating wind. spreading may be important. For instance, downbursts Nine anemometers in NSW and one in the Australian may occur in close proximity to horizontal or vertical Capital Territory were used. Their locations are shown vortex circulations, and sometimes a divergent micro- in Fig. 1. At least 20 yr of data were used for each burst wind pattern cannot be readily discriminated from station with a maximum of 33 yr; the periods overlapped sheared or rotating winds (Kessinger et al. 1988). but did not match. This paper describes the climatology of strong, con- Events shorter than 10 min and/or involving signif- vectively generated wind gusts in New South Wales icant directional change (.908) at the beginning of the (NSW), Australia. Only gusts observed in the vicinity event were classi®ed as short-lived. A total of 476 of thunderstorms are included, but the characteristics of events emerged from the selection process, of which the parent thunderstorms (size, type, longevity, etc.) are 171 (36%) were de®ned as short-lived. The frequencies unknown.