Although the exact physical mechanism responsible for HBs is unclear, it is likely that the local rise in temperature, decrease in dew point temperature and pressure, and sometimes increased wind speed, are the result of air originating above the boundary layer, subsiding dry adiabatically and then diverging as it reaches the surface. This chain of events is usually associated with decaying areas of convective precipitation and nearly dry-adiabatic lapse rates in mid-levels. Knowledge of the mesoscale environment in which heat bursts may form can help forecasters anticipate heat burst potential on a given day. In this paper we suggest that knowledge of the environment in which heat bursts tend to occur, combined with mesonet and WSR-88D data, is needed in order for a forecaster to detect, follow, and possibly forecast the evolution of heat burst events.
To investigate the HB's spatial characteristics, a two-pass Barnes objective analysis was performed. The Barnes scheme produced gridded fields of mass divergence, equivalent potential temperature (theta-e), and altimeter setting from Mesonet observations of temperature, relative humidity, pressure and winds. Perturbations of these variables were defined as the difference between gridded values and the mean value at a given time. Thus, positive perturbations of altimeter setting represented relatively higher pressure while negative perturbations of theta-e were associated with drier air. The mean altimeter setting values at 0000, 0100, 0300 and 0500 UTC were 1003.1, 1003.1, 1003.7, and 1003.3 mb, respectively. The corresponding mean values of thata-e were 344.6, 341.1, 335.4 and 334.0 K.
In addition to Mesonet data, both WSR-88D data from Oklahoma City (KTLX; in Level II format) and Frederick (KFDR; in NIDS format) were examined to determine both: 1) the evolution of reflectivity and 2) velocity-based parameters, such as mid-level convergence, associated with this event. Finally, soundings taken at 1200 and 0000 UTC over Oklahoma and Texas were used to estimate the environment in which the convection associated with the HBs formed and evolved.
Regional rawinsonde observations at 0000 UTC showed a dry, warm, deep mixed layer west of the dryline, with a weak capping inversion remaining intact above the moist layer east of the dryline. CAPE values over northern TX and southern OK remained above 2000 J kg-1. A moderately strong southerly low-level jet (near 850 hPa) was also present east of the dryline, with wind speeds near 15 ms-1. Mid-tropospheric lapse rates (700-300 hPa) were 7o C km-1 both east and west of the dryline. Considerable veering of the winds was evident through 6 km in the moist sector, as evidenced by the 0000 UTC sounding from Norman, OK (OUN). Despite the relatively large amount of deep layer shear (0-6 km), no supercell or bowing storm structures were observed during the evening.
By 0100 UTC, the mesohigh existed between Grandfield (GRAN) and Medicine Park (MEDI), with a mesolow located about 75 km to the northwest (Fig 3b). As a result of these pressure perturbation couplets, a mesoscale pressure gradient force was most likely an important driver for the strong, sustained winds across southwest Oklahoma. However, HBs were observed within this pressure gradient from Altus (ALTU) and HOBA eastward to near MEDI. As a result of the HBs, wind damage and markedly lower values of theta-e were observed over this region.
During the next hour, the pressure perturbation couplet moved slightly eastward. The pressure difference between the couplet had increased from 6 hPa to 8 hPa, and severe winds in excess of 25 ms-1 were reported in the region of strong pressure gradient. This HB was observed by MEDI, Fort Cobb (FTCB) and Apache (APAC). Of the three sites, FTCB observed the most dramatic thermodyanamic perturbations, including a 6oC (~10oF) temperature increase and an 18 oC (~30 oF) dewpoint decrease over a 20 min period (Fig 2a).
By 0300 UTC, an elongated mesohigh and mesolow couplet was located over southern and central OK (Fig 3c). Near this time,winds exceeding severe criteria and temperatures near 40 oC were observed in the region of large pressure gradient near Chickasha (CHIC). Effects of the HB were observed progressively from south (ACME) to north (El Reno, ELRE). Some of the most extreme wind damage occurred over this area. The short-lived nature of the temperature extrema and wind perturbations over these sites suggest that these were produced by individual downdrafts.
Between 0400 and 0600 UTC the meoshigh and mesolow moved slowly eastward as the nearby convection continued to weaken (Fig 3d). However, throughout this period reports of high winds and anomalously warm temperatures continued to persist. These kinematic and thermodynamic perturbations were observed by the Washington (WASH) (Fig 2b), Norman (NORM), Ketchum Ranch (KETC), and Pauls Valley (PAUL) mesonet sites. These sites marked the easternmost extent of HB characteristics. In general, the area affected by HBs tended to experience southwesterly winds, whereas east of this area the flow was southerly. This discontinuity marked the final eastward progression of the HB. A low-level inversion was intact east of the discontinuity, where as to the west of the discontinuity, a deep nearly-dry-adiabatic lapse rate, conducive to deep momentum transport, was present. Intense mesoscale subsidence influences on the theta- e field over much of central and western OK was striking, with theta-e values mostly less than 330 K.
Although this event is similar to previously studied HBs (i.e. presence of steep lower and mid-tropospheric lapse rates, weakening convection, temperature, moisture, pressure and wind anomalies), a physically-based predictive conceptual model utilizing real-time mesoscale observations is lacking. Previous authors (Johnson et. al. 1989) have postulated that descending air, forced by evaporative cooling beneath the convective anvil, was responsible for adiabatic warming near the surface. In turn, this heating generates a region of lowered surface pressures beneath the anvil. A pressure gradient is then induced between the rain-cooled mesohigh and the subsidence-induced mesolow or wake depression. These processes may indeed explain much of the 22- 23 May 1996 HB. However, while it is likely that subsidence associated with nearby convection was key to this event, it is not clear what physical mechanism was responsible for organizing the subsidence on the larger mesoscale spatial and temporal scales. It is also not clear to what degree, if any, the low-level jet was responsible for the unusually damaging winds.
Numerical simulations of HB events may aid the development of a predictive methodology that is useful to forecasters. In the interim, environments characterized by steep lower and mid-tropospheric lapse rates, dry mid-tropospheric air, shallow surface- based inversions, and weakening nocturnal convection should be considered candidates for HB production.
