Volume 19, Issue 1 p. 91-110
Research article
Free Access

Mesoscale observations of an extended heat burst and associated wind storm in Central Oklahoma

Jeffrey B. Basara

Corresponding Author

Jeffrey B. Basara

Oklahoma Climatological Survey, Norman, OK, USA

School of Meteorology, University of Oklahoma, Norman, OK, USA

Oklahoma Climatological Survey, 120 David L. Boren Blvd., Suite 2900, Norman, Oklahoma 73072, USA.Search for more papers by this author
Mason D. Rowell

Mason D. Rowell

School of Meteorology, University of Oklahoma, Norman, OK, USA

Search for more papers by this author
First published: 07 July 2011

Abstract

On 13 May 2009, 13 surface-based observing stations within central and western Oklahoma recorded maximum wind speeds in excess of 22.5 m s−1, along with gusts exceeding 28 m s−1 in isolated locations, during a localized wind storm that lasted in excess for over an hour. These wind speeds were associated with an enhanced mesoscale pressure gradient that developed during the late evening of 12 May 2009 and early morning hours of 13 May 2009, which then slowly propagated from west to east. Ultimately, the event produced localized wind damage in northern portions of Oklahoma City. Analysis of the synoptic and mesoscale conditions present at each stage of the event revealed that specific parameters associated with heat bursts occurred during the event, but with the inclusion of varying mesoscale and microscale influences. As such, a close examination of the mesoscale pressure field indicated that a dynamically variable mesoscale convective system (MCS) produced a mesohigh/mesolow couplet that ultimately created the strong pressure gradient. The result was a prolonged period (in excess of 1 h) of enhanced wind speed values across a swath approximately 300 km in length and 50 km in width that were not directly associated with the anomalously warmer and drier conditions that occurred due to the heat burst. Copyright © 2011 Royal Meteorological Society

1. Introduction

The occurrence of localized high wind speed events within the United States and across the globe has been well documented for several years. With time and improved research, the more frequent wind event types have been assigned common classifications, each of which depend on a variety of different atmospheric and boundary layer processes. These include, but are not limited to, downslope winds (Beran, 1967), gust fronts (Mahoney, 1988), derechos (Johns and Hirt, 1987) and microbursts (wet and dry; Wakimoto, 1985).

Specifically within the Southern Plains of the United States, the combination of varying environmental conditions often yields the formation of intense convection and severe weather, including supercell thunderstorms and squall lines that generate tornadoes, damaging winds and large hail (Brooks et al., 2003; Doswell et al., 2005; Hocker and Basara, 2008a, 2008b). However, unique conditions within the synoptic or mesoscale can often result in variants of the more typical localized high wind events with associated parameters dissimilar to the classic types. Thus, in-depth analyses that thoroughly diagnose the characteristic dynamics associated with specific, unique cases often provide new insights into the atmospheric processes that yield such events.

One such case that has received comparatively less study than more common mesoscale weather events occurs when dry convective downdrafts penetrate a shallow stable layer near the surface, producing a localized thermal anomaly. These events, known as heat bursts (Johnson, 1983), have been associated with rapid increases in air temperature, rapid decreases in humidity, pressure jumps and strong, gusty winds (Williams, 1963; Sloan, 1966; Johnson, 1983; Johnson et al., 1989; McNulty, 1991; Bernstein and Johnson, 1994).

The mechanisms for the formation of heat bursts have been documented by a number of studies including Johnson (1983) and Johnson et al. (1989). The surface-based characteristics of heat bursts often begin with the decay of precipitation areas in association with stronger convective cells. Johnson et al. (1989) determined that strong downdrafts that pass through a deep, nearly dry adiabatic layer yield warming at the surface if (1) a low-level stable layer is present and (2) the downdraft is capable of penetrating through the equilibrium level to the surface. In such cases, the downdraft air will warm dry-adiabatically below the equilibrium layer and produce a localized thermal anomaly. Johnson (2001) noted that two possible scenarios exist that could lead to the development of heat bursts. These include the possibility of dry microbursts penetrating a near-surface stable layer, but also subsidence associated with the rear inflow jet that is often present in mesoscale convective systems (MCS; Smull and Houze, 1987; Rutledge et al., 1988; Stumpf et al., 1991; Bernstein and Johnson, 1994; Houze, 2004). The depth and relative strength of the stable layer near the surface is critical to the observed occurrences of heat bursts. Bernstein and Johnson (1994) noted that a main contributor to the relative rareness of heat burst events is likely due to the rare occurrence of a shallow stable layer with strong descent above it.

Using data from the Oklahoma Mesonet, a dense network of mesoscale observing stations in Oklahoma, McPherson et al. (2011) documented and analysed 207 heat burst events captured by the network from 1994 to 2009. The results found that heat bursts were most frequently observed in the western two-thirds of the state, were mainly nocturnal events and were most common during the warm season with June as the most active month. McPherson et al. (2011) also noted that nearly all of the events occurred in close proximity to weak radar echoes generally less than 30 dBz and yielded an average temperature increase of 5.3 °C, an average decrease in dew point temperature of 6.9 °C, an average peak wind speed of 18.2 m s−1, and an average thermodynamic perturbation of 72 min. The results further reinforce that with a mesoscale surface network in place, heat burst events and their evolution can be observed with increased frequency and resolution.

During the overnight hours of 12–13 May 2009, a heat burst event occurred across portions of western and central Oklahoma near an ongoing MCS. Over a period of several hours, a surface mesolow/mesohigh couplet developed and strengthened which resulted in an enhanced mesoscale pressure gradient. Subsequently, a severe windstorm developed and slowly propagated across portions of central Oklahoma, producing sustained winds in excess of 22.5 m s−1 and isolated gusts over 28 m s−1 across a swath nearly 300 km in length and 50 km in width (Figure 1). In addition, the severe windstorm passed over the northern sections of Oklahoma City (OKC) causing localized damage to vegetation and structures. Because the event occurred over a region containing dense surface observation networks comprised of stations within the Oklahoma Mesonet (McPherson et al., 2007) and the Oklahoma City Micronet (Basara et al., 2010), this study seeks to examine the mesoscale evolution of the event and the interactions between the convective and mesoscale features that produced the severe windstorm.

Details are in the caption following the image

The estimated occurrence of wind speed values of 18 and 22.5 m s−1 or greater on from 0000 to 1200 UTC on 13 May 2009 determined from Oklahoma Mesonet and Oklahoma City Micronet observations. The focus of this study is on the wind values found in the largest swath that impacted central Oklahoma

2. Data

Observation systems used in this study include the following: (1) three United States National Weather Service (NWS) Weather Surveillance Radar 88 Doppler (WSR-88D) radars (Crum and Alberty, 1993), (2) the upper-air observations of the NWS, (3) surface observing stations of the Oklahoma Mesonet, and, (4) surface observing stations of the Oklahoma City Micronet.

A partnership of Oklahoma State University and the University of Oklahoma and managed by the Oklahoma Climatological Survey (OCS), the Oklahoma Mesonet is a permanent mesoscale surface observing network of 120 remote, meteorological stations across Oklahoma (Brock et al., 1995; McPherson et al., 2007). Each station measures more than 20 environmental variables, including wind at 2 and 10 m, air temperature at 1.5 and 9 m, relative humidity, rainfall, pressure, solar radiation, and soil temperature and moisture at various depths. All sensors are mounted on or near a 10 m tower supported by three guy wires and powered via solar energy. Mesonet data are collected and transmitted to a central point every 5 min where they are quality controlled, distributed and archived (Shafer et al., 2000).

The Oklahoma City Micronet (OKCNET) is a network designed to improve atmospheric monitoring across the Oklahoma City Metropolitan area (Basara et al., 2010). This 40 station network consists of four Oklahoma Mesonet Stations (OKCE, OKCN, OKCW, and SPEN), as well as 36 stations mounted on traffic signals at a height of approximately 9 m and station spacing of approximately 3 km. At each traffic signal site, air temperature, humidity, pressure, rainfall, wind speed and wind direction are measured and transmitted every minute to a central facility 24 h per day, year-round, where they are quality controlled, distributed, and archived using the Oklahoma Mesonet infrastructure. The Oklahoma City Micronet includes a cluster of stations within the central business district as well as stations throughout the Metropolitan area.

3. Overview of the event

3.1. Synoptic conditions

Upper-air analyses from 0000 UTC (1900 local time; local time = UTC − 5 h) on 13 May 2009 (i.e. prior to the event) revealed mainly zonal flow at 500 hPa across the Southern Plains of the United States (Figure 2(a)) with a short-wave ridge over the eastern portions of the United States and a short-wave trough in the northwest portions of the United States. At 850 hPa (Figure 2(b)) a trough was located in the lee of the Rocky Mountains extending from the border between the United States and Mexico to a closed low over eastern Montana and the western Dakotas. At 1200 UTC on 13 May 2009, the upper air pattern shifted slightly to the east as the shortwave-trough at 500 hPa both deepened and translated across the northern United States (Figure 3(a)). Similarly, the 850 hPa trough shifted east as the 850 hPa low moved into south-central Canada (Figure 3(b)). In addition, a strong low-level jet was oriented from southwest to northeast extending from west Texas into Illinois.

Details are in the caption following the image

The (a) 500 hPa and (b) 850 hPa analyses on 0000 UTC on 13 May 2009. The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), height above sea level (m; upper right), and wind barbs (m s−1)

Details are in the caption following the image

The (a) 500 hPa and (b) 850 hPa analyses on 1200 UTC on 13 May 2009. The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), height above sea level (m; upper right), and wind barbs (m s−1)

Overall, the synoptic conditions changed little during the period spanning the duration of the mesoscale windstorm event. This is further reinforced via soundings launched from the KOUN upper air station located approximately 30 km south of Oklahoma City at 0000 and 1200 UTC on 13 May 2009 (Figure 4(a) and (b)). In both cases a nearly dry adiabatic layer extended from approximately 750 to 450 hPa. At the same time, a stable layer extended from the surface to approximately 750 hPa. Such profiles were consistent with those noted previously as being conducive toward heat burst formation (Johnson et al., 1983, 1989; Bernstein and Johnson, 1994; McPherson et al., 2011).

Details are in the caption following the image

The (a) 0000 UTC and (b) 1200 UTC soundings from the upper-air station in Norman, OK on 13 May 2009

3.2. Mesoscale evolution

At 0100 UTC on 13 May 2009, prior to the mesoscale windstorm event in central Oklahoma, observations from the Oklahoma Mesonet revealed a general east-west oriented pressure gradient across Oklahoma with the greatest values found towards the east (approximately 1012 hPa) and the lowest values in the west (approximately 1000 hPa in the Oklahoma Panhandle). As such, the surface winds were primarily southeasterly with temperature values ranging from 19 to 29 °C and dew point temperature values ranging from 16 to 23 °C within the study domain (Figure 5).

Details are in the caption following the image

Surface analysis on 0100 UTC on 13 May 2009. The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), pressure reduced to sea level (hPa; upper right), and wind barbs (m s−1). Pressure contours are displayed in black every 1 hPa

At approximately 2230 UTC on 12 May 2009, thunderstorms initiated along the dryline located in west Texas, intensified, and entered portions of southwest Oklahoma as a broken line of convective cells between 0100 and 0200 UTC (Figure 6). Within this time frame, the first detection of heat burst related conditions occurred at the Hollis Oklahoma Mesonet site (HOLL) in extreme southwest Oklahoma where temperatures values increased from 25.1 °C at 0125 UTC to 32.3 °C at 0150 UTC (Figure 7). This heat burst also corresponded with a peak wind gust of 22.5 m s−1 and a dew point temperature decrease of nearly 19 °C. However, no significant pressure perturbation was observed at the site or within the nearby mesoscale pressure field.

Details are in the caption following the image

Surface analysis and radar reflectivity from the KFDR radar on 0200 UTC on 13 May 2009. The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), pressure reduced to sea level (hPa; upper right), station identifier (middle right), maximum wind gust (lower right; m s−1), and wind barbs (m s−1). Pressure contours are displayed in black every 1 hPa This figure is available in colour online at wileyonlinelibrary.com/journal/met

Details are in the caption following the image

Five minute time series observations from the Hollis Mesonet site from 0000 to 0800 UTC on 13 May 2009 (7:00 pm on 12 May 2009 to 3:00 am on 13 May 2009 local time) including air temperature ( °C; top/black), dew point temperature ( °C; top/grey), average wind speed (m s−1; middle/grey), maximum wind gust (m s−1; middle/black), wind direction (diamonds), and pressure reduced to sea level (hPa; bottom/black)

By 0340 UTC, a more pronounced heat burst occurred further north and to the west of an isolated convective cell. Observations from the Butler Mesonet site (BUTL) in the vicinity of this cell revealed an overall temperature increase of over 9 °C during the heat burst (Figure 8), with the associated thermal anomaly impacting multiple stations in west-central Oklahoma (Figure 9) as well. Unlike the earlier heat burst near HOLL, significant pressure perturbations were evident in the surface observations at BUTL. As the convective cell passed over the site, the mean sea level pressure increased nearly 3 hPa before falling nearly 6 hPa over a 30 min period during the heat burst. As the heat burst subsided at BUTL, the pressure slowly rose over 4 hPa to ambient synoptic conditions. This pressure anomaly was also made evident within the mesoscale pressure field via a mesolow which rapidly developed and was co-located with the surface thermal perturbation (Figure 9). Simultaneously, a more pronounced mesohigh developed further south along the leading edge of the convective line.

Details are in the caption following the image

Five minute time series observations from the Butler Mesonet site from 0000 to 0800 UTC on 13 May 2009 (7:00 pm on 12 May 2009 to 3:00 am on 13 May 2009 local time) including air temperature ( °C; top/black), dew point temperature ( °C; top/grey), average wind speed (m s−1; middle/grey), maximum wind gust (m s−1; middle/black), wind direction (diamonds), and pressure reduced to sea level (hPa; bottom/black)

Details are in the caption following the image

Surface analysis and radar reflectivity from the KFDR radar on 0340 UTC on 13 May 2009. The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), pressure reduced to sea level (hPa; upper right), station identifier (middle right), maximum wind gust (lower right; m s−1), and wind barbs (m s−1). Pressure contours are displayed in black every 1 hPa and temperature contours greater than 24 °C are displayed in grey every 1 °C This figure is available in colour online at wileyonlinelibrary.com/journal/met

Though the more intense convection across southern Oklahoma consolidated into an MCS, the isolated northern convective cell slowly dissipated. The analysis at 0530 UTC revealed that a well-defined mesohigh with a maximum mean sea level pressure (MSLP) value of 1010 hPa developed along with the MCS and associated cold pool across south-central Oklahoma (Figure 10). At the same time, the thermal anomaly associated with the previously pronounced heat burst had weakened but remained co-located with the mesolow that had propagated northeast into central Oklahoma. Given the minimum observed MSLP of this 1001 hPa mesolow combined with the proximity of the prominent mesohigh, a strong pressure gradient developed yielding strong sustained wind speeds and gusts at locations embedded within the gradient.

Details are in the caption following the image

Surface analysis and radar reflectivity from the KTLX radar on 0530 UTC on 13 May 2009.The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), pressure reduced to sea level (hPa; upper right), station identifier (middle right), maximum wind gust (lower right; m s−1), and wind barbs (m s−1). Pressure contours are displayed in black every 1 hPa and temperature contours greater than 23 °C are displayed in grey every 1 °C This figure is available in colour online at wileyonlinelibrary.com/journal/met

Examination of the time series from the El Reno Mesonet site (ELRE), which was one such station embedded within the strong pressure gradient at 0530 UTC, demonstrates that while conditions consistent with a heat burst occurred at the site (i.e. increased temperature and decreased dew point temperatures), the more dramatic impacts of the event were associated with the pressure and wind observations (Figure 11). First, the surface pressure increased 4.5 hPa during a 30 min period followed by a rapid drop of 9.5 hPa over the following 60 min. As the pressure decreased, the surface winds increased with gusts over 25 m s−1 observed at intervals for nearly an hour while sustained wind speeds in excess of 15 m s−1 lasted for approximately 90 min.

Details are in the caption following the image

Five-minute time series observations from the El Reno Mesonet site from 0300 to 1100 UTC on 13 May 2009 (10:00 pm on 12 May 2009 to 6:00 am on 13 May 2009 local time) including air temperature ( °C; top/black), dew point temperature ( °C; top/grey), average wind speed (m s−1; middle/grey), maximum wind gust (m s−1; middle/black), wind direction (diamonds), and pressure reduced to sea level (hPa; bottom/black)

3.3. Peak intensity and high-resolution surface observations

The mesoscale pressure gradient became stronger as both the mesohigh and mesolow propagated to the east and strengthened to the observed respective intensities of 1012 and 1000 hPa as recorded by Oklahoma Mesonet stations at 0630 UTC (Figure 12). This period is representative of the most intense magnitudes for both the mesohigh located to the southeast of Oklahoma City and mesolow located to the northwest of Oklahoma City.

Details are in the caption following the image

Surface analysis and radar reflectivity from the KTLX radar on 0630 UTC on 13 May 2009.The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), pressure reduced to sea level (hPa; upper right), station identifier (middle right), maximum wind gust (lower right; m s−1), and wind barbs (m s−1). Pressure contours are displayed in black every 1 hPa and temperature contours greater than 23 °C are displayed in grey every 1 °C This figure is available in colour online at wileyonlinelibrary.com/journal/met

Due to the location relative to both the mesohigh and mesolow, the Oklahoma City Metropolitan area was directly impacted by the enhanced pressure gradient, which was captured in greater detail by the high-resolution Oklahoma City Micronet. Time series analyses from numerous sites across the northern portion of OKC revealed ambient wind speed values prior to the event ranged from 5 to 15 m s−1. However, as the mesohigh/mesolow couplet approached Oklahoma City, stations across the region noted a significant pressure rise of nearly 5 hPa from approximately 1006 hPa to over 1010 hPa. During this pressure rise, several stations including the KNW104 site in northwest Oklahoma City also demonstrated a wind shift from approximately 150° (southeast) to approximately 220° (southwest) as well as a decrease in the wind speed values (e.g. Figure 13).

Details are in the caption following the image

One minute time series observations from the KNW104 Oklahoma City Micronet site from 0300 to 1100 UTC on 13 May 2009 (10:00 pm on 12 May 2009 to 6:00 am on 13 May 2009 local time) including air temperature ( °C; top/black), dew point temperature ( °C; top/grey), average wind speed (m s−1; middle/grey), maximum wind gust (m s−1; middle/black), wind direction (diamonds), and pressure reduced to sea level (hPa; bottom/black)

As the enhanced pressure gradient moved over northern portions of Oklahoma City, all OKCNET stations recorded maximum pressure values followed by a rapid decrease to minimum values below the ambient conditions in place prior to the event. While variability existed, the maximum mean pressure decrease during the event was 7.5 hPa over a 70 min span (Table I). For a subset of stations located in northern portions of Oklahoma City, the maximum pressure was followed by a secondary maximum approximately 30 min later (Figure 14). For this subset of stations, pressure values declined at rates that exceeded, in the most dramatic cases, 10 hPa h−1 (Table I). Following the pressure minimum, conditions gradually returned to ambient conditions. On average, the characteristic event attributes, from initial pressure rise through the rapid decrease until the return to ambient conditions, spanned approximately 6.5 h (395 min; Table I).

Details are in the caption following the image

One-minute time series observations from the KNW105 Oklahoma City Micronet site from 0300 to 1100 UTC on 13 May 2009 (10:00 pm on 12 May 2009 to 6:00 am on 13 May 2009 local time) including air temperature ( °C; top/black), dew point temperature ( °C; top/grey), average wind speed (m s−1; middle/grey), maximum wind gust (m s−1; middle/black), wind direction (diamonds), and pressure reduced to sea level (hPa; bottom/black)

Table I. Mean sea level pressure anomalies observed by OKCNET sites during the 13 May 2009 windstorm
Site Maximum change (hPa) Primary interval (min) Secondary pressure time interval (min) Total rate of pressure decrease (hPa h−1) Secondary rate pressure decrease (hPa h−1) Pressure anomaly duration (min)
KNW101 6.5 42 9.3 388
KNW103 9.3 87 6.4 368
KNW104 11.6 84 8.3 389
KNW105 9 81 43 6.7 12.6 394
KNW106 8.4 90 46 5.6 11.0 395
KNW107 8.6 91 46 5.7 11.2 388
KNW108 7.7 101 4.6 397
KNW202 9 74 7.3 403
KNE101 9.9 97 6.1 395
KNE202 7.7 57 8.1 397
KNE103 8.1 54 9.0 401
KNE105 8.2 60 8.2 393
KCB101 7.6 48 9.5 394
KCB102 7.3 49 8.9 400
KCB103 7.9 50 9.5 395
KCB104 7.4 49 9.1 391
KCB105 7.4 48 9.3 401
KCB106 6.6 98 4.0 397
KCB107 7 45 9.3 396
KCB108 6.4 48 8.0 401
KCB109 6.7 48 8.4 401
KCB110 6.9 47 8.8 396
KSW101 7 105 55 4.0 7.6 390
KSW102 6.7 50 8.0 398
KSW103 7.8 57 8.2 398
KSW104 7.1 39 10.9 400
KSW105 6.3 98 3.9 398
KSW106 6.6 74 5.4 397
KSW107 6.8 82 5.0 394
KSW108 6.9 89 4.7 400
KSW109 6.4 74 5.2 404
KSW110 6.6 108 3.7 397
KSW111 6.6 72 5.5 397
KSW112 7.3 80 5.5 398
KSE101 6.6 78 5.1 391
KSE102 7 61 6.9 388
Mean 7.5 70 7.0 395
Standard deviation 1.1 21 2.0 6

A significant consequence of the increased pressure gradient was that wind speed values increased significantly from ambient values near 10 m s−1 to sustained values in excess of 20 m s−1 with gusts approaching 30 m s−1. Table II displays the maximum sustained wind speed and gust values during the event, revealing again that northern portions of Oklahoma City were most impacted. Time series observations from the KNW104 site (Figure 13), which measured the peak wind gust (29.2 m s−1) and greatest pressure drop (11.6 hPa), demonstrates the strong degree of correlation between increases in observed wind speeds and the negative pressure tendency. Further, the wind speed values were consistently anomalous relative to the ambient conditions for over an hour.

Table II. Wind, temperature, and humidity anomalies observed by OKCNET sites during the 13 May 2009 windstorm
Site Maximum wind speed (m s−1) Maximum wind gust (m s−1) Maximum temperature ( °C) Minimum dew point temperature ( °C) Thermal anomaly duration (min)
KNW101 23.7 25.2 26.8 9 279
KNW103 20 25.2 26.3 9.2 277
KNW104 24.3 29.2 25.3 9.9 269
KNW105 16.2 23.5 26 9.6 270
KNW106 15.8 20.9 25.5 9.8 259
KNW107 17.4 23.5 25.6 10 270
KNW108 16.1 20.4 25.9 9.4 264
KNW202 14.5 19.9 26.5 9.1 289
KNE101 20.3 25.8 26.7 9.7 263
KNE202 15.1 19.5 25.4 10.3 276
KNE103 15.9 20.6 25.7 9.7 269
KNE105 15.1 19.5 25.4 10.3 267
KCB101 15 19.7 25.3 10.2 267
KCB102 NA NA 25.4 9.9 269
KCB103 NA NA 25.4 9.8 275
KCB104 NA NA 25.2 9.9 261
KCB105 NA NA 25.1 10.3 276
KCB106 NA NA 25.2 10.4 279
KCB107 NA NA 25.4 10.1 270
KCB108 NA NA 25.5 10 274
KCB109 NA NA 25 10.5 270
KCB110 17.4 22.1 25.4 9.9 272
KSW101 17.9 23.1 23.5 11.4 267
KSW102 16.7 22.7 25.5 10.1 270
KSW103 19 24.1 25.7 10.3 278
KSW104 15.6 20.1 25.4 10.2 256
KSW105 12.1 16.4 25.4 10 268
KSW106 11.2 16.7 25.2 10.3 265
KSW107 9.7 13.6 24.5 10.9 259
KSW108 19.2 24 24.1 10.9 287
KSW109 10.4 13.7 24.7 11.3 260
KSW110 11.8 19.2 23 12.1 268
KSW111 11.5 15.2 24.1 11.2 257
KSW112 13 16.6 23.3 11.9 262
KSE101 11.1 15.8 24.7 10.9 278
KSE102 10.9 14.1 24.6 10.6 235
Mean 15.6 20.4 25.2 10.3 269
Standard deviation 3.8 4.0 0.8 0.7 10

Examination of the pressure and wind fields across Oklahoma City at 0628 UTC (Figure 15) indicated that both wind speed and direction between the mesohigh and mesolow strongly resembled that of pressure gradient driven flow, with the lowest observed pressure values (approximately 999 hPa) and strongest observed wind speed values (nearly 30 m s−1) matched at the KNW104 site during the event. Additionally, corresponding to the peak intensity of the event, the surface pressure gradient observed by OKCNET stations was 12 hPa over a distance of only 25 km, stretching from the southern areas of Oklahoma City into the northern portions. As such, Figure 15 also demonstrates that while severe wind gusts approaching 30 m s−1 were occurring over portions of north Oklahoma City, locations in south Oklahoma City experienced wind speed conditions consistent with the ambient synoptic environment (i.e. 5–10 m s−1). This implies a rapid cross-isobaric acceleration of air parcels from the mesohigh towards the mesolow via the enhanced pressure gradient ongoing at this time.

Details are in the caption following the image

Surface analysis centered on Oklahoma City at 0628 UTC on 13 May 2009. The station model includes observations of maximum wind gust (upper left; m s−1), pressure reduced to sea level (hPa; upper right), station identifier (lower right), and wind barbs (m s−1). Pressure contours are displayed in black every 1 hPa

Another unique facet of this case is that the temperature and humidity values recorded at all OKCNET sites seemingly demonstrated heat burst conditions. Approximately 2 h following the initial pressure rise, air temperature values increased and the humidity (dew point temperature) decreased. Furthermore, the thermal anomaly persisted, on average, nearly 4 h (e.g. Figures 13 and 14), with the maximum temperatures observed again for those sites located in northern portions of Oklahoma City (Table II). Given that the ambient air temperature prior to the event generally ranged from 20 to 21 °C, the impact of the thermal anomaly increased the surface temperatures during the overnight hours by 3–6 °C depending on location. Conversely, the dew point temperature, which ranged from 17 to 18 °C prior to the event, plummeted 7–9 °C depending on location. However, it is also critical to recognize certain noteworthy disparities that distinguish the aforementioned characteristics from the more classic heat burst events. Most importantly, (1) the onset of heat burst conditions lagged behind the windstorm, and, (2) the peak thermal anomaly occurred 1–2 h following the observed minimum pressure values and associated maximum observed wind speed values. Yet, while the thermal anomaly began after the pressure perturbations, both returned to ambient conditions at nearly the same time at all locations.

3.4. Dissipation of the windstorm

Once beyond Oklahoma City, the windstorm and associated pressure and thermal anomalies weakened in east-central Oklahoma. Figure 16 reveals that by 0930 UTC the main precipitation associated with both the mesohigh across southern Oklahoma and the heat burst and associated mesolow across central Oklahoma had dissipated. As a result, the mesohigh weakened such that it was no longer distinguishable from the background surface pressure field. A heat burst related thermal anomaly was still present across central Oklahoma to the northeast of Oklahoma City, however the mesolow had begun to fill and was no longer well defined. Overall, the weakening of both the mesohigh and mesolow led to a relaxation of the mesoscale pressure gradient and a decrease in the surface wind speeds.

Details are in the caption following the image

Surface analysis and radar reflectivity from the KTLX radar at 0930 UTC on 13 May 2009.The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), pressure reduced to sea level (hPa; upper right), station identifier (middle right), maximum wind gust (lower right; m s−1), and wind barbs (m s−1). Pressure contours are displayed in black every 1 hPa and temperature contours greater than 23 °C are displayed in grey every 1 °C This figure is available in colour online at wileyonlinelibrary.com/journal/met

By 1200 UTC, nearly 12 h after the beginning of the event, all precipitation had ceased and the key mesoscale features that produced the windstorm event, including the mesolow, mesohigh, and associated thermal anomalies had fully dissipated. As such, the surface pressure gradient and associated wind fields returned to ambient conditions maintained by synoptic-scale processes (Figure 17).

Details are in the caption following the image

Surface analysis and radar reflectivity from the KTLX radar at 1200 UTC on 13 May 2009.The station model includes observations of air temperature ( °C; upper left), dew point temperature ( °C; lower left), pressure reduced to sea level (hPa; upper right), station identifier (middle right), maximum wind gust (lower right; m s−1), and wind barbs (m s−1). Pressure contours are displayed in black every 1 hPa

4. Discussion

As a result of the swath associated with the 13 May 2009 windstorm event, a region approximately 300 km in length and 50 km in width enclosed localized damage at many locations, including severed tree limbs, downed power lines, and structural damage to buildings. Fortunately, the associated impacts were not severe enough to result in any injuries or fatalities.

Typically, mesoscale windstorm events in Oklahoma are associated with severe squall lines (Hocker and Basara, 2008b) or nonlinear MCS events termed ‘derechos’ (Johns and Hirt, 1987; Coniglio et al., 2004). In each of these cases, heavy precipitation is a primary mechanism for the generation of strong to severe surface winds. To appreciate fully the unique aspects of the 13 May 2009 windstorm, it should be realized that the most extreme conditions produced by the event over a period of several hours occurred in the absence of significant precipitation. Nevertheless, precipitation played multiple roles in the processes that ultimately produced the windstorm event.

For example, as the line of scattered thunderstorms entered southwest Oklahoma it consolidated into a MCS as it traversed the southern portion of the state. Along the leading edge of the strong convection, a mesohigh developed and intensified. The development of the mesohigh was consistent with past studies that demonstrated that rain-cooled convective downdrafts result in a cold-pool of air near the surface which is vital to the formation of a mature MCS (Coniglio and Stensrud, 2001; Corfidi, 2003; Engerer et al., 2008). As such, the primary mechanism for the formation of mesohighs associated with MCS evolution is via the hydrostatic response to the local cold pool which yields increased surface pressure values (Fujita, 1959; Johnson, 2001). Typically, the mesohigh develops along the leading edge of the convective precipitation (Fujita, 1955, 1963; Johnson and Hamilton, 1988; Houze et al., 1990) before eventually trailing the stronger precipitation located in the convective line (Loehrer and Johnson, 1995; Johnson, 2001).

In this case, Figure 9 reveals that the mesohigh was co-located along the line of intense convection at 0340 UTC, which later became a well-defined mesohigh behind the convective line by 0530 UTC (Figure 10). The influence of the mesohigh was also observed across central Oklahoma and in the vicinity of those locations directly impacted by the windstorm. For example, Figures 11, 13, and 14 all demonstrate that prior to the windstorm and thermal anomaly associated with the heat burst, the surface pressure at the stations rose significantly. Thus, stations such as the ELRE Mesonet site and the Oklahoma City Micronet locations first experienced a trend of increasing surface pressure as the northern flank of the mesohigh translated across the region prior to the increased surface winds.

The precipitation processes involved in this event also played a crucial role in the development of the observed mesolow. Early studies by Fujita (1955, 1963) noted the occurrence of a wake low associated with strong squall lines. Subsequent work by Smull and Houze (1987), Rutledge et al. (1988), and Stumpf et al. (1991) highlighted the importance of the rear inflow jet in concentrating subsidence and associated warming through the lower troposphere within the region trailing the cold pool. As a result, the hydrostatic response to this concentrated warming is an area of lower pressure values near the surface. Eventually, Houze et al. (1990) and Loehrer and Johnson (1995) developed conceptual models for symmetric and asymmetric MCS life cycles that include both a mesohigh and mesolow.

At first glance, this case appears to display the appearance of an asymmetric MCS. However, during peak intensity of the windstorm (Figure 12), the majority of the convection was well removed from the area and had weakened substantially. Moreover, the initial formation of the mesolow was associated with a heat burst to the west of an isolated cell on the northern flank of the convective line and exhibited a significant thermal anomaly at the surface throughout its lifetime. Using radar observations, Smith et al. (2010) examined the heat bursts on 13 May 2009 and noted that both mechanisms described by Johnson (2001) occurred during the heat burst occurrences. Thus, while the rear-inflow jet typically associated with mature MCSs played a role in the subsidence warming, dry microbursts may have also impacted the development of the mesolow.

Another key facet of the windstorm event was the strengthening of the pressure gradient to the intensities noted at approximately 0630 UTC (Figures 12 and 15). Both the mesolow and mesohigh strengthened as they translated to the east. In the case of the mesolow, minimum pressure values decreased within the core from approximately 1002–998 hPa as it moved from western Oklahoma to northern Oklahoma City. Concurrently, the surface pressure values associated with the mesohigh were observed at 1011 hPa when located in southwest Oklahoma at 0340 UTC and then increased to peak values of 1012 hPa to the southeast of Oklahoma City at 0630 UTC.

A second key aspect was the relative convergence of mesohigh/mesolow couplet. As the MCS became increasingly asymmetric, the axis of the mesohigh shifted further north within the stratiform precipitation trailing the convective line. Such a transition of the mesohigh within MCSs has been documented by Houze et al. (1990), Loehrer and Johnson (1995), and Adams-Selin and Johnson (2010). As a result of both the increasing mesohigh and mesolow intensities as well as the reduced spatial distance between their respective centres, the surface pressure gradient strengthened and the surface wind speeds increased in response.

Examination of the OKCNET data during the peak intensity of the windstorm further revealed several other important aspects of the event. For example, Johnson (2001) noted that often the most intense pressure gradients associated with mesohigh/mesolow couplets were located near the back edge of the stratiform precipitation: Figure 12 reveals that the 13 May 2009 windstorm was consistent with the previous results. Furthermore, Johnson (2001) noted that surface pressure gradients, such as that observed at 0630 UTC (approximately 12 hPa over a 25 km distance), are comparable to pressure gradients found in a moderate hurricane. However, the 13 May 2009 windstorm was relatively short-lived and Vescio and Johnson (1992) noted that air parcels travelling through gradients in mesoscale systems aren't embedded within the pressure gradient long enough to reach optimum velocities. Finally, Figure 15 confirms that in such events, the surface winds are nearly perpendicular to the isobars (Johnson, 2001), yielding gradient surface flow.

The results of this analysis also bring to light a misconception of the 13 May 2009 windstorm that may be deduced if only the time series data from impacted sites were analysed, especially those located in and the around the Oklahoma City Metropolitan area (e.g. Figures 13 and 14). Analysis of the OKCNET data revealed that the thermal anomaly occurred at all sites and followed the pressure and associated wind anomalies. Moreover, the differing strengths of the thermal anomalies between the northern and southern sites were relatively small when compared to those of the pressure and wind anomalies for similar transects. When combined with the spatial analyses of the event, the windstorm appears to be primarily due to the enhanced pressure gradient from the mesohigh/mesolow couplet and not as a result of specific convective downdrafts associated with the heat burst itself.

5. Conclusion

During the overnight hours of 12–13 May 2009, an unusual windstorm propagated across central Oklahoma and produced a swath of strong to severe winds approximately 300 km in length and 50 km in width. While damage from the 13 May 2009 windstorm event was limited, the windstorm was observed by a dense array of in situ surface stations that captured the evolution of the mesoscale surfaces features from initiation, through intensification and dissipation.

The results of this study demonstrated that the windstorm was produced by a combination of features and processes that included the following:
  • a line of convective storms formed into a mesoscale convective system (MCS) across southwest Oklahoma and produced a surface mesohigh;

  • an isolated convective cell on the northern flank of the MCS yielded a prolonged heat burst that produced both a localized warm thermal anomaly and a mesolow;

  • the mesohigh and mesolow intensified as each moved into central Oklahoma while the distance between the mesohigh and mesolow decreased due to increasing MCS asymmetry;

  • a strong pressure gradient developed between the mesohigh/mesolow couplet and then reached a peak intensity of approximately 12 hPa over a 25 km distance within the Oklahoma City Metropolitan area;

  • observations from the Oklahoma City Micronet during peak intensity revealed surface flow nearly perpendicular to the isobars, with peak wind gusts near 30 m s−1, and,

  • the pressure gradient weakened as the convective precipitation dissipated and thus both the mesolow and mesohigh also weakened.

Overall, the synoptic environment yielded a favourable thermodynamic profile conducive to the development of heat bursts. Examination of the time series obtained from numerous stations demonstrated that multiple heat bursts occurred across portions of western and central Oklahoma. However, the isolated convective storm to the north of the mature MCS produced a heat burst that was particularly long-lived not only in terms of its thermal signature, but also its impact on the surface pressure field. While it is not uncommon for an asymmetric MCS to feature both a mesohigh and mesolow, the strength of the pressure gradient was uncommon. As such, while convective downdrafts associated with the heat burst do not appear to have produced the windstorm from a first order perspective, the heat burst itself played a critical role in the final distribution of the surface pressure field and associated enhanced gradient that generated the pronounced windstorm. Furthermore, because the warm/dry downdraft was able to penetrate through the surface stable layer and produce the heat burst, the low-level thermal anomaly was stronger than typically expected with those wake lows associated with MCS events that also yield a stronger cold-pools near the surface. Ultimately, further research and examination of this event from a numerical perspective will be crucial towards determining the exact role that heat bursts play on mesolow intensification and whether similar events are likely to occur in the future.

Acknowledgements

Oklahoma's taxpayers fund the Oklahoma Mesonet through the Oklahoma State Regents for Higher Education and the Oklahoma Department of Public Safety. Funding for the Oklahoma City Micronet and support for this study was provided by the Oklahoma State Regents for Higher Education (equipment), the Office of the Vice President for Research at the University of Oklahoma (personnel and supplies), the Oklahoma Mesonet (personnel), the Oklahoma Climatological Survey (personnel and in-kind support), and the City of Oklahoma City (in-kind support). The authors would also like to thank Landon Harrison and Chase Thomason for their assistance in developing the focus of the study and analysis and Brad Illston for assistance with Figure 1.