1. Introduction
Several efforts have been made to acquire and interpret near-ground pressure and wind measurements in or near a tornado. These efforts are motivated by the desire to understand the complex and violent nature of a tornado’s interaction with the surface. The earliest examinations of tornado measurements were conducted on serendipitous tornado or near-tornado encounters with statically positioned weather stations or barometers (e.g., Tepper and Eggert 1956; Fujita 1958; Ward 1961). A list of formally documented pressure deficits from these types of encounters is provided in Table 1, populated, in part, from historical listings in Table 16.2 of Davies-Jones and Kessler (1974). A pressure deficit may be defined as the difference between a predefined or ambient pressure value and the lowest observed pressure measurement within a relative time frame corresponding to the passage of a tornado. These historical pressure deficit measurements in tornadoes ranged from 5 to 192 hPa. Estimates of peak winds in tornadoes are less common since typical measuring systems for wind are more vulnerable to failure than those for pressure. Estimates of peak near-ground wind speeds as provided in Davies-Jones and Kessler (1974) are based mostly on damage and thus have a large degree of uncertainty given known complexities in estimating wind speed from damage (Doswell et al. 2009).
More recent efforts to understand tornado and near-tornado environments have used specially engineered tornado probes (e.g., Bluestein 1983; Winn et al. 1999) or data collected during unplanned tornado encounters with mobile weather stations (Blair et al. 2008). However, because of the inherent logistical difficulties of placing near-ground instrumentation in the path of a tornado, only a few such measurements in tornadoes exist (Table 1). Additionally, Lewellen et al. (1997) concluded from very high-resolution tornado modeling that pressure measurements in a tornado should be made with a temporal resolution of at least a few hertz to sufficiently resolve the characteristics of a tornado. However, only the pressure observations of Winn et al. (1999) from the pre-2002 record meet this criterion. Given these shortcomings, our understanding of the lowest levels of a tornado is limited.
More recent laboratory simulations of tornado-like vortices examined the pressure and force distribution on the individual faces of a cubical model (e.g., Mishra et al. 2008; Sengupta et al. 2008). These studies indicate that wind loads produced by tornadoes greatly exceed the loads produced by equivalent straight-line boundary layer wind profiles, including those of microbursts. Wind data collected from tornadoes would be most helpful for validating these simulations of wind loads. In the absence of these difficult-to-obtain data, structural engineers sometimes estimate wind loads on structures using the maximum pressure deficit within tornadoes to estimate wind speeds using cyclostrophic balance, even with its degraded suitability in estimating near-surface wind speeds. Thus, establishing a range of peak pressure deficits associated with tornadoes benefits this area of research.
Rapid development in computer technology has permitted the ability to numerically simulate tornadolike vortices. These simulations have ranged from vortex formation associated with a modeled supercell (Grasso and Cotton 1995; Wicker and Wilhelmson 1995; Adlerman and Droegemeier 2000; Finley et al. 2002) or nonsupercell thunderstorm (Lee and Wilhelmson 1997) to small-domain, very high-resolution vortex modeling (e.g., Lewellen 1993; Lewellen et al. 1997, 2000; Lewellen and Lewellen 2007; Le et al. 2008). The high-resolution models have the advantage of being able to represent highly detailed turbulent vortex structures, including the vortex interaction with the surface, and have been a primary tool for advancing our understanding of vortex dynamics, especially in the corner flow region (the area at the base of the vortex where the flow turns upward into the vortex). Results from these simulations have shown that vortex surface pressure and wind distributions depend on both the structure of the vortex aloft and the detailed structure of the corner flow region. Lewellen et al. (2000) showed that low-level cross sections of pressure through the tornado core agreed with their laboratory counterparts for high swirl-ratio two-cell vortices supporting multivortex structures and for low swirl ratio single-cell vortices. However, medium swirl-ratio two-cell vortex circulations can also produce surface pressure distributions very similar to a single-cell vortex pressure configuration (Lewellen et al. 1997, 2000) in cases where the dynamics of the corner flow region preclude the central downdraft from reaching the surface. These results suggest that single-cell vortex structures cannot be directly inferred from pressure measurements alone.
Nine near-ground pressure profiles from tornadoes obtained from 2002 to 2008 using Hardened In-situ Tornado Pressure Recorder (HITPR) probes and a mobile mesonet are documented in this paper. From 2002 to 2004, five separate tornado intercepts were made using one—in some cases two—HITPR probes (Samaras and Lee 2004; Lee et al. 2004; Wurman and Samaras 2004). Video probes were developed in 2004 to provide both visual evidence of tornado structure and a potential means of deriving estimates of winds speeds using stereo photogrammetry. In 2007, the Tactical Weather Instrumented Sampling in Tornadoes Experiment (TWISTEX) began. The goal of the project was to collect in situ observations within tornadoes using HITPR probes and to gather meteorological data within a few kilometers outside of the tornado using a mobile mesonet. During May 2008, four tornado intercept datasets were collected. One intercept was conducted as planned on a mature tornado, with a coordinated deployment of instrumentation both in and near the tornado. However, in the remaining three tornadoes, tornadic circulations either developed away from the primary area of concentrated storm rotation or the circulations were displaced from their anticipated path and impacted the mobile mesonet stations unexpectedly. These types of events hereafter will be referred to as “unplanned tornado encounters.” This study presents background on the instrumentation and methodology, followed by analysis of the measurements obtained in these nine cases. Further, the pressure and wind measurements presented herein are combined with the previously described measurements to present an extensive listing of near-ground tornado measurements taken to date.
2. Methodology
A variety of instruments were used to sample the nine tornadoes described in section 3. These included both in situ and mobile instrumentation arrays that are described in the following subsections.
a. In situ instrumentation
Two types of in situ instrumentation—a HITPR probe (Fig. 3a) and a video probe (Fig. 3b)—were available for deployment in tornadoes. For cases where the deployment team had sufficient time, two HITPR probes—labeled HITPR 1 and HITPR 2—were deployed. Both in situ surface probes are aerodynamically shaped and engineered to withstand the harsh tornadic environment. The HITPR probes are outfitted with sensors that measure temperature, pressure, and relative humidity recorded at 10 Hz. Details of the design and engineering of the HITPR may be found in Samaras and Lee (2004). All HITPR pressure data underwent quality control inspection and were bias checked with the mobile mesonet instrumentation when available. Because of instrument equilibration time limitations in a number of the cases, time series of temperature and relative humidity are not shown.
The video probes are outfitted with seven cameras each to provide a tornado-relative reference for the sampling position of the HITPRs while providing a full 360° field of view of the evolving/translating near-ground flow as the tornado passes. Six Korea Technology & Communications (KT&C) charge-coupled device (CCD) cameras, model KPC-650, are positioned horizontally, each spanning a 60° horizontal view with 420 lines of vertical resolution. The seventh camera is positioned vertically. Video from all cameras is simultaneously recorded at 30 frames per second. Deployment of two video probes may allow for derivation of the three-dimensional velocity of individual debris tags using stereo photogrammetric analysis (e.g., Forbes and Bluestein 2001; Rasmussen et al. 2003). Given the time constraints in probe deployment and the priority given to the HITPR probes, only one video probe was deployed and only in two of the nine cases presented herein. Therefore, all direct wind observations reported in this paper were obtained using the mobile mesonet stations.
b. Mobile mesonet stations
Three vehicles were outfitted with instrumentation to measure temperature, relative humidity, atmospheric pressure, and wind velocity based on the station configuration and instrumentation presented by Straka et al. (1996), with GPS used to record position and vehicle velocity (mobile mesonet shown in Fig. 3c). Some instrumentation differences exist between those specified in Straka et al. and newer mesonet stations, such as those used by TWISTEX and Texas Tech University’s Atmospheric Science Group, which share the same instrumentation (see Table 1 of Hirth et al. 2008). Mesonet station data were recorded every two seconds. When deployed in and close to the hook echo region of a supercellular storm, these data can provide thermodynamic and kinematic characteristics of the flow field bounding a tornado or tornadogenesis region (Markowski et al. 2002; Grzych et al. 2007; Hirth et al. 2008). Note that because the temperature and relative humidity instrument response time is too large with respect to the intercept time window of the narrow tornadoes directly sampled, no mesonet temperature and relative humidity time series are presented.
Owing to inaccuracies in the anemometry during significant vehicle accelerations, velocity data were removed in a similar manner as that employed by Markowski et al. (2002), Markowski (2002b), and Grzych et al. (2007). The mesonet datasets were also quality controlled for spurious meteorological readings and vehicle headings. Biases were removed, making use of intercomparisons between mesonet stations for extensive periods when the caravan was in relatively uniform meteorological conditions and predominantly in transit. Additionally, all mobile mesonet pressure measurements were reduced to sea level using Eq. (2.31) from Wallace and Hobbs (1977). In this calculation, the U.S. Geological Survey 1 arc s elevation data were used to provide an estimation of the elevation where each pressure measurement was taken.
In light of our primary objective, the mobile mesonet was positioned to sample the rear-flank downdraft (RFD) outflow and RFD gust front regions of tornadic supercells. This was successfully achieved on multiple occasions during the project (Finley and Lee 2008; Lee et al. 2008). However, the mobile mesonet stations also had unplanned tornado encounters on three occasions. Given the remarkable rarity of tornado encounters with research-caliber measuring equipment, we felt the scientific significance of these data justified their formal documentation.
3. Cases and results
Near-ground measurements in tornadoes were collected on nine occasions. Background and discussion of these events are described in detail in the following subsections. In each case the measured pressure trace is shown and the vortex structure is noted if possible. Time series of wind speed and direction are also presented where available. For all cases, time-to-space conversions were performed using estimates of each tornado’s translation speed. The translation speeds were estimated, where possible, by using the nearest Weather Surveillance Radar-1988 Doppler (WSR-88D) radar data and computing the speed of the tornadic vortex signature (Brown et al. 1978). In the absence of radar-based speeds, or to corroborate the radar estimates when available, tornado translation estimates were made using the pressure and wind measurements, available video, and previously documented translation speeds. For all cases, the tornado was observed to traverse the instrumentation at a near-constant velocity. A summary of the measurements obtained, in addition to previously documented tornado measurements, is given in Table 1.
Because surface roughness greatly affects near-surface turbulence and can make interpretation of velocity and pressure data more problematic, the surrounding terrain’s roughness category (Cook 1985) was estimated using visual observations, video from the deployment site, and aerial satellite photos for each case. We found that all cases may be classified as Cook (1985) category 1 (flat terrain), except for 10 May 2008, which we classify as category 4 (dense woodland). These roughness classifications represent a compromise between the empirically derived roughness values listed in Wieringa (1993) and Simiu and Scanlan (1996).
In addition to providing estimates of surface roughness in each case, the storm-relative location of each intercepted tornado was determined. In seven of the nine cases, the tornado occurred within the region of the supercell low-level mesocyclone and toward the tip of the hook echo (Stout and Huff 1953) as viewed from the radar. This region is classically recognized as a location for mesocyclonic tornadogenesis (e.g., Brandes 1978; Lemon and Doswell 1979; Davies-Jones 1982; Markowski 2002a). In the remaining two cases, the tornado occurred away from the primary area of concentrated storm rotation (Quinter and Beloit, Kansas). These latter cases correspond to two of the three unplanned tornado encounters. In the other unplanned encounter (10 May 2008), the tornado deviated from its anticipated path. The storm-relative intercept locations are provided on the radar imagery in Fig. 4 (radar animations are available on the supplemental Web page http://dx.doi.org/10.1175/2010MWR3201.s1). It is important to note that spatial differences may exist between the actual storm-relative intercept locations and the radar-indicated positions in Fig. 4. This is due to the radar image time being different from the intercept time (i.e., the radar-indicated storm position is displaced from its deployment-relative position at the time of intercept). Instances also arise where the combination of long deployment distance from the nearest WSR-88D and large storm tilt leads to the actual low-level storm-relative position differing from that depicted on radar. For example, in the Quinter, Kansas, case the lowest WSR-88D elevation scan sampled the storm at an elevation of approximately 2200 m AGL.
a. 7 May 2002, Mullinville, Kansas
A supercell that produced a series of large tornadoes was intercepted in northwest Kiowa County shortly before 0000 UTC. The teams positioned themselves to the east of the supercell with sufficient time to conduct a well-coordinated deployment of probes. Two HITPR probes were placed in the projected path of this tornado, approximately 13 km north of the town of Mullinville, Kansas (Fig. 4a). The tornado was initially cone shaped with condensation about half way to the surface with a prominent dust swirl near the ground (Fig. 5a) and gradually transitioned into a large tornado while encountering the HITPR 1 probe (Fig. 5b).
The tornado’s central axis passed closest to the HITPR 1 probe at 0000:07 UTC. A postevent survey conducted by members of the deployment team found that the tornado’s central axis took an oblique path relative to the probes (Fig. 6a). The HITPR 1 probe sampled the outer edge of the tornado, whereas HITPR 2 was just outside of it. Shortly after traversing the probes, the tornado produced F-3 damage (NCDC 2002).
The HITPR probes 1 and 2 measured maximum pressure deficits of 22 and 10 hPa, respectively (Fig. 7a), as the tornado translated at an estimated 5.7 m s−1. Note that the maximum recorded pressure deficit in the pressure time series plots may not represent the largest pressure deficit inside the tornado. A complete pressure history from HITPR 1 is found in Fig. 15 of Samaras and Lee (2004). The tornado’s structure cannot be determined with confidence based solely on the pressure measurements because of the oblique passage of the tornado with respect to the HITPR probes. However, by combining the pressure trace with visual observations, a confident tornado structure determination can be made. The pressure trace shown in Fig. 7a suggests a broad two-cell vortex with high swirl ratio, and the rapid pressure fluctuations superposed on the pressure trace also suggest weak secondary vortices on the periphery of the tornado. This characterization is consistent with the visual appearance of the tornado as captured by the video probe, which confirms the presence of subtornado-scale vortices within the parent circulation shortly before the circulation transected the HITPR probes (Fig. 5a). Observations by the deployment team also confirmed secondary vortices within the parent tornadic circulation as the tornado passed over the probes. This evidence of multiple vortices supports the notion that the northern portion of a high swirl-ratio tornado with two-cell vortex structure passed over the HITPR probes.
b. 15 May 2003, Stratford, Texas
The deployment team intercepted a supercell with cyclic tornadogenesis approximately 18 km northwest of Stratford, Texas (Fig. 4b). One HITPR probe was deployed on Highway 287 in the path of a large tornado. On the basis of an assessment of surface scouring from the tornado at the probe’s deployment location and on observations at close range by Samaras, we believe the HITPR 1 probe took a near-direct transect of the tornado (Fig. 6b). This tornado was later given a rating of F-3 (NCDC 2003a). The tornado translational speed was 15 m s−1 as documented in Wurman and Samaras (2004).
The HITPR 1 surface pressure profile shown in Fig. 7b reveals a complex vortex structure. On the basis of this surface pressure time series and supporting observations from the Doppler on Wheels (DOW; Wurman et al. 1997) located more than 11 km from the HITPR tornado sampling position (Wurman and Samaras 2004), we offer one possible explanation for the complex pressure profile. A general drop and rise in surface pressure is evident over a 2.6-min period, with a maximum deficit of 41 hPa. Embedded within the pressure profile bowl of the low-level mesocyclone between −1200 and 600 m (from −80 to 40 s) are several pressure fluctuations occurring on different time scales. Two vortices are apparent in the pressure time series: one from −1000 to −400 m (from −67 to −27 s) and another from −400 to 400 m (from −27 to +27 s). Within the pressure trace of the stronger, or more directly sampled, vortex are two very short time-scale surface pressure deficits—one at 50 m (3 s) and the other at 250 m (16 s)—suggestive of multiple vortices within this second circulation.
We surmise the second of the two broad internal pressure departures was associated with the visible and Doppler-indicated tornado (Wurman and Samaras 2004). Because the first of these vortices had a smaller surface pressure deficit, the vortex was either weaker or the vortex (of unknown strength) was sampled obliquely. DOW observations in Wurman and Samaras (2004) support the second of these vortices as being much stronger; although, insufficient beam resolution prevents this determination from being definitive. We believe the second of these vortices also possessed embedded secondary vortices, given the very distinct small time-scale pressure perturbations. The deployment team members confirm the presence of multiple vortices embedded within the parent tornadic circulation during the HITPR 1 intercept. To summarize, visual and radar evidence supports the assertion that the pressure trace captured a cascade of vortex scales from the low-level mesocyclone to the tornado and secondary vortex scales.
Using the DOW-measured velocity and the HITPR 1 pressure profile, Wurman and Samaras (2004) estimated a tornado core flow diameter between 400 and 450 m while traversing the HITPR 1 probe, as inferred from Fig. 7b. A DOW-measured peak wind speed difference of approximately 106 m s−1 was estimated horizontally across the tornado at the time of the HITPR 1 probe intercept. Secondary vortices within this tornado could not be sufficiently resolved, so the true peak wind speeds in the tornado as it passed over HITPR 1 are unknown.
c. 24 June 2003 (Case 1), Manchester, South Dakota
A supercell with cyclic tornadogenesis was intercepted on Highway 14 west of Manchester, South Dakota, with tornadogenesis occurring approximately 6 km south-southwest of the town (Lee et al. 2004). The deployment team positioned themselves in the projected path of the tornado, approximately 3 km north of Manchester (Fig. 4c), where the first of three probes (HITPR 1) was deployed (Figs. 8a and 8b). The team deployed the remaining two probes as they progressively retreated northward to elude the approaching tornado, and data from one of these probes (HITPR 2) are discussed in the following subsection. A post-event damage survey conducted upon retrieval of the HITPR 1 probe indicated this probe was positioned near the center of the field scouring produced by the tornado. From this evidence, we surmise HITPR 1 took a near-direct transect of the tornado (Fig. 6c). Damage assessment by the National Weather Service indicated the F-4 damage swath extended to a farmstead approximately 30 m north of where the probe was placed (NCDC 2003b).
The HITPR 1 probe measured a large surface pressure deficit of 100 hPa as the central axis of the tornado traversed the probe (Fig. 7c) at 0046:52 UTC with an estimated translational speed of 9.4 m s−1 as documented in Lee et al. 2004. The shape of this profile suggests a low swirl-ratio tornado with a single-cell vortex structure (Snow et al. 1980) or a medium swirl-ratio tornado with a two-cell vortex structure (Lewellen et al. 1997, 2000). Although video was taken by the deployment team (Fig. 8), it was not helpful in refining the determination of the vortex structure.
Lee et al. (2004) used two different applications of the HITPR 1 pressure data to estimate the maximum cyclostrophic wind speed at the deployment site to be 92 and 98 m s−1. We concur with Lee et al. on the problems with applying cyclostrophic balance to maximum wind speed estimation in the corner flow region, which is highly influenced by surface effects and resultant radial inflow. However, a few tens of meters above the surface where the radial inflow has relaxed considerably, application of cyclostrophic balance for estimating wind speed, while still quite approximate, would be more appropriate. Additionally, Lewellen et al. (1997) concluded from high-resolution tornado simulations that the maximum surface pressure deficit is a gross underestimate of the maximum pressure deficit occurring at a height of about 30 m above the surface. Given the Lewellen et al. findings, the 100-hPa pressure deficit recorded at the surface should be considered a lower bound on the actual maximum pressure deficit, which very likely was a few tens of meters above the surface. Thus, at or just above the elevation where the pressure deficit is maximized, the cyclostrophically derived maximum wind speed estimates based on a 100-hPa pressure deficit will have more validity and could very well be conservative. Note that the 92–98 m s−1 wind speed range places these estimates at a level capable of causing EF-5 damage on the Enhanced Fujita (EF) scale (e.g., Potter 2007). Just how close these wind speeds extend to the surface is uncertain; however, Lewellen et al. (2000) showed peak-time-averaged vortex tangential speeds occurring in the corner flow region for high and low swirl corner flows. Recent extensions of this work looking at conditions leading to near-surface vortex intensification by Lewellen and Lewellen (2007) also indicate peak wind speeds and pressure deficits no more than just a few tens of meters above the surface.
d. 24 June 2003 (Case 2), Manchester, South Dakota
A second intercept of the same tornado occurred 3 min and 10 s after case 1 at 0050:02 UTC. As the deployment team moved northward to stay in front of the projected path of the tornado, the second of three HITPR probes (HITPR 2) was deployed at a location approximately 4.5 km north of Manchester, South Dakota, or 1370 m north of the HITPR 1 intercept location (Fig. 4c). Determining the location of the HITPR 2 probe relative to the tornado was difficult because of the lack of vegetative scouring. However, based on concurrent nearby visual observations, the deployment team was confident the probe sampled at least a portion of the tornado (Fig. 6c). Although few damage indicators were present at the deployment location, the tornado caused F-4 damage to a farmstead, as discussed in the previous section, about three minutes prior to this intercept.
The HITPR 2 probe measured a surface pressure deficit of 54 hPa as at least some portion of the tornado traversed the probe (Fig. 7d). The tornado’s translation speed slowed down considerably from the speed given for the HITPR 1 intercept. A translation speed of 2.3 m s−1 was estimated using available video in conjunction with the HITPR 2 pressure measurements. Nearby videos taken of the tornado from both north and south at the time it crossed over the probe show a full condensation funnel to the ground with no apparent secondary vortices. At approximately 0055:47 UTC, about 3 min and 45 s after intercepting the HITPR 2 probe, the tornado rapidly transitioned into a small rope tornado. The tornado dissipated about 1 min and 15 s later, at approximately 0057:02 UTC.
The shape of the HITPR 2 pressure profile in Fig. 7d is once again suggestive of either a low swirl-ratio single-cell vortex or a medium swirl-ratio two-cell vortex; however, given the uncertainty in the HITPR 2 probe’s position relative to the center of the tornado path, the structure of the vortex cannot be conclusively determined from the pressure data. Nearby video taken by two of the authors from both north (Fig. 9) and south at and near the time the tornado crossed over the probe shows a full condensation funnel to the ground with no visible secondary vortices. There was a strong vertical updraft jet, especially along the west side of the narrow tornado base with a laminar condensation funnel of near-uniform width extending well above the surface (Fig. 9b). These visual observations are suggestive of a single-cell vortex, given the lack of a visual vortex breakdown point in the funnel.
e. 11 June 2004, Webb, Iowa
A developing supercell was intercepted approximately 5 km west of Webb in northwest Iowa (Fig. 4d). The crew positioned east-northeast of the storm’s mesocyclone and attendant tornado, which allowed sufficient time to conduct a well-coordinated deployment of probes. This tornado was later given a Fujita scale rating of F-3 (NCDC 2004).
Two HITPR probes (HITPR 1 and 2) and one video probe were deployed linearly, with the video probe positioned farthest to the north (Fig. 6d), followed by HITPR probes 1 and 2, respectively. Photos of the tornado at the time of probe placement and at the time of tornado passage over both HITPR and video probes are shown in Fig. 10. Videos taken by the video probe revealed that the tornado took an oblique passage relative to the probes, with the video probe and the HITPR 1 probes sampling the southern edge of the tornado and HITPR 2 likely sampling the southern edge of the tornado (Fig. 6d).
The HITPR 1 and 2 probes measured surface pressure deficits of 26 and 18 hPa, respectively, (Fig. 7e) as the tornado moved at an estimated speed of 7.7 m s−1. Because of the oblique transect of the HITPR probes relative to the tornado and the inconclusive video evidence, it is impossible to ascertain the true structure of the vortex in this case.
The lack of damage by the tornado made determining its strength difficult. However, two utility poles adjacent to the probe deployment site were snapped as the tornado crossed the road. Attempts were made to determine the wind speed of lofted debris captured by the video probe. To do so, we used the known specifications of each camera, provided in section 2, in an attempt to utilize the technique outlined in Forbes and Bluestein (2001). However, we determined that small errors in estimates of each debris tag’s size or distance resulted in large errors in the calculated wind speed. Therefore, estimates of debris-derived tornadic wind speeds using photogrammetry are not provided in this paper. These limitations emphasize the utility of deploying two video probes in future deployments for stereo photogrammetric analysis.
f. 10 May 2008, Broken Bow, Oklahoma
Members of the TWISTEX crew intercepted a supercell thunderstorm approximately 6 km north of Broken Bow, Oklahoma (Fig. 4e). The crew observed the storm as it approached from the northwest, but tornadogenesis did not appear imminent. As the storm approached, the crew noted that the supercell was moving more sharply to the right of its former course, placing them near the projected path of the low-level mesocyclone. The crew drove south on Highway 259, attempting to position south of the low-level mesocyclone before it crossed the highway. With considerable tree cover in this region hampering the visual observation of the storm’s features, TWISTEX crews could not position south of the mesocyclone on Highway 259 before the mesocyclone reached this road. Thus, the two mobile mesonet stations, M2 and M3, had an unplanned tornado encounter with a developing tornadic circulation while the mesonet was traveling south on Highway 259 (Fig. 6e). An experienced storm chaser in the area confirmed that a tornado developed over and just east of Highway 259 (R. Hill 2008, personal communication), and we estimated the tornado’s translational speed to be 12 m s−1.
A pressure deficit of approximately 5 hPa and a wind gust of 40 m s−1 at 3 m AGL were measured by M2 (Figs. 11a and 11b). M3 noted a smaller 3-m-AGL pressure deficit of approximately 3.5 hPa but a substantially higher wind gust near 50 m s−1. Both wind gusts were observed approximately 200–300 m south of, or roughly 10 s after measuring, the peak pressure deficit. We hypothesize that the mobile mesonet drove through a developing tornado, as evidenced by the recorded pressure deficit and the generally east-southeasterly flow on the north side of the circulation and westerly flow on the south side of the circulation. This idea is supported by the visual observations of tornadogenesis over Highway 259, coinciding with the approximate mesonet location in time. Unfortunately, a short segment of the anemometry data from the station immediately north of the assumed circulation center point was removed because of vehicle acceleration exceeding data quality control limits. The strong winds just south of the assumed circulation center were likely in the interface region between the developing tornado and an intense small-scale internal RFD outflow surge. An internal RFD outflow surge may be defined as a coherent pulse in the RFD outflow that is located behind the initial RFD gust front (Finley and Lee 2004, 2008; Lee et al. 2004; Marquis et al. 2008a,b). The heavy tree cover in the area produced high-frequency and significant amplitude fluctuations in wind magnitude and direction, and likely masked a considerable portion of the tornadic circulation. On the basis of the mesonet and eyewitness evidence, tornadogenesis was concurrent with this internal RFD outflow surge and occurred over Highway 259 just as the mesonet was passing through this location.
g. 23 May 2008, Quinter, Kansas
A supercell with cyclic tornadogenesis moving nearly due north was tracked by the TWISTEX crew at approximately 2100 UTC, south of Quinter, Kansas. Mesonet sampling of the RFD outflow and RFD gust front subsequently commenced as the storm moved north toward Quinter. At 2144 UTC, with a tornado in progress approximately 2–3 km to the north of the mesonet location, a much larger tornado quickly formed to the west-northwest of the mesonet. This tornado moved in a northeasterly direction initially, passing within 1 km to the northwest and north of the lead mesonet station, M1 (Fig. 4f). Afterward, the tornado motion was nearly due north. Owing to the large storm tilt with height for this storm and because of the time difference between the radar scan time and intercept time with this rapidly moving storm, the low-level storm-relative deployment position was closer to the actual low-level mesocyclone location than that indicated in Fig. 4f.
M1 and M2 were positioned rather close to the large tornado and experienced several RFD surges (Finley and Lee 2008). With this very large tornado ongoing (estimated width of 1.6 km; NCDC 2008a), an unplanned tornado encounter occurred as a narrow tornado moved over M2 (Fig. 6f; C. Collura 2008, personal communication). This tornado was positioned along an internal RFD outflow surge boundary and translated at an estimated speed of 20 m s−1. The RFD gust front was well northeast of the mesonet position when this internal boundary was sampled by the instrumentation and visually observed (debris tracers and onset of power line–utility pole damage). In a storm-relative framework, the concurrent tornadoes—one associated with the low-level mesocyclone and the other associated with the internal RFD surge boundary—best resemble the tornadic supercell schematics of Brandes (1978, his Fig. 18) and Davies-Jones (1982, his Fig. 3b), except the gust front tornadogenesis region was along a secondary internal boundary. Upon tornado passage, a 3-m-AGL pressure deficit of approximately 14 hPa was measured by M2 at 2144:16 UTC, while M1 on the northern periphery of this circulation and located 344 m north of M2 experienced a pressure deficit of approximately 6 hPa at nearly the same time (Figs. 11c and 11d). The coarse temporal resolution of the pressure data relative to the time scale of the tornado sampled prevents a conclusive determination of vortex structure. Immediately following the tornado’s passage, M2’s pressure trace was rather unsteady for a brief period, likely because of small vortices shed from nearby upstream downed utility poles and power lines or from trauma incurred by the mesonet station during tornado passage.
In addition to the 3-m-AGL pressure deficits, both mobile mesonet stations recorded significant wind gusts (Figs. 11c and 11d). Both mobile mesonet stations were traveling at approximately 10 m s−1 prior to the intercept and quickly decelerated as the wind speed ramped up considerably. As a result, a small portion of the wind measurements immediately prior to the intercept did not pass quality control, as indicated in Figs. 11c and 11d. At this time, M1 was deliberately driven off the road to avoid falling utility lines in high winds (the wind measurements at this time did not pass quality control criteria because of M1’s deceleration). A 46 m s−1 3-m-AGL wind gust was measured by M2 at 2144:18 UTC, two seconds after the maximum pressure deficit measured during the tornado passage. Trauma to the M2 aerovane curtailed wind data collection after this time. Visual observations at M2’s location indicated the surface winds backed just before tornado passage and veered upon passage. Note that adjustments to the M1 wind speed data in Figs. 11c and 11d after 2144:23 UTC were made to correct for the station’s inclined ditch position. Given the importance of keeping a consistent reference state for M1 (i.e., the anemometer at a consistent effective height above road level, as opposed to above ditch level), we chose to apply an adjustment factor to M1’s speeds. This adjustment was made possible through a fortuitous positioning of M3, which was parked 20–30 m away while the teams were still in vigorous RFD outflow. M3 had access to free-stream (unobstructed) flow and was not in a position to block the flow encountering M1. From the overlapping period, an adjustment factor of 1.25 was determined from the wind speed time series (i.e., the M1 wind speeds were increased by 25%).
h. 29 May 2008, Tipton, Kansas
After departing initially targeted storms near Kearney, Nebraska, the TWISTEX crews journeyed south to intercept a tornadic supercell with deployment taking place 9 km northwest of Tipton, Kansas, on Highway 181. A very short-lived tornado was first observed by the deployment teams at 0117:48 UTC. The tornado that would eventually be sampled developed shortly thereafter, at 0118:35 UTC, and subsequently passed over the in situ probe array at 0122:45 UTC (Fig. 4g) at an estimated speed of 14.6 m s−1. The instrumentation used on this day included two mobile mesonet stations, two HITPR probes, and one video probe.
M2 and M3 were arrayed south of the tornado, and the in situ probes were deployed linearly in the path of the tornado (Fig. 6g). Unfortunately, the southernmost HITPR probe failed to record measurements while sampling the tornado. Thus, only the in situ pressure measurements from HITPR 1 exist for this case. A video capture of the tornado at the time of probe placement (Fig. 12) gives a visual indication of the tornado’s structure approximately 1-km upstream and 1 min prior to passing over the in situ probes. At the time of intercept, the tornado had an elevated condensation funnel with a debris-filled near-surface circulation. This tornado was later assigned an EF-scale rating of EF-1 (NCDC 2008b).
M3 recorded an average wind speed of 39 m s−1 for a 20-s period and a peak speed of 44 m s−1 as the tornado passed just to its north (Figs. 7f, 11e, and 11f). Additionally, M3 measured a 3-m-AGL pressure deficit of 5.5 hPa. These observations are compared to pressure measurements from the HITPR 1 probe located 235 m to the north of M3 (Figs. 7f and 11f) and positioned within the tornado, south of the tornado’s central axis. The probe recorded broad radial pressure gradients and a surface pressure deficit of 15 hPa that remained quasi steady for a short time, corresponding to the passage of the tornado’s central axis. These pressure attributes suggest the tornado was a high swirl-ratio two-cell vortex with an axial downdraft present. Videos from the video probe show subtle evidence of radial flow directed internally outward with an annulus of debris around an apparently hollow interior. These videos also revealed secondary vortices with this tornado after passage. We believe the high-frequency pressure perturbations in Figs. 7f and 11f along the bottom plateau of the pressure trace were caused by weak secondary vortices passing over or near the HITPR 1 probe. Additional analysis of the thermodynamic and kinematic characteristics of the flow field in the proximate tornado environment and RFD outflow can be found in Lee et al. (2008) and is the subject of ongoing analysis.
i. 29 May 2008, Beloit, Kansas
About one hour after the first intercept, at approximately 0217 UTC, an unplanned tornado encounter occurred when a small and apparently anticyclonic tornado passed over the mesonet approximately 13 km north of Beloit, Kansas (Figs. 4h and 6h). This vortex was in a storm-relative position consistent with anticyclonic gust front tornadoes (Brandes 1993), and the tornado was assigned an EF-scale rating of EF-0 (NCDC 2008b). M2 and M3 were stationary and positioned facing west roughly 6 m apart. By this time, the sun had set and the mesonet had abandoned coordinated data gathering attempts for the evening.
M3 measured a surface pressure deficit of approximately 13 hPa and a maximum wind gust of about 40 m s−1 (Figs. 11g and 11h) as the tornado translated at an estimated speed of 5 m s−1. Video from M3 shows evidence of the tornado passing over M2 with the rear suspension of the M2 vehicle becoming fully unloaded and shifting a small distance northward during the time of peak wind speed. However, the pressure sensor on M2 unfortunately dropped a few readings during vortex passage, likely related to the rapid pressure drop. Since the temporal resolution of the pressure data was coarse relative to the scale of the tornado sampled, we cannot make a determination of vortex structure in this case. Given the mesonet sampling rate, we surmise the 40 m s−1 reading did not accurately capture the peak wind speeds during tornado passage.
Interestingly, Fig. 11g shows the wind direction from M3 switched from westerly to southerly just prior to the tornado passage, indicative of the RFD gust front passing or, more properly, retreating over the mesonet. During tornado passage the wind direction rapidly changed to the east. Analysis of video at this time suggests the tornado moved from south to north, with the center of the vortex passing just east of the teams. From the video and wind direction time series, we believe the mesonet sampled an anticyclonic tornado.
4. Conclusions
This paper has presented nine cases from 2002 to 2008 where near-surface pressure observations, and in some cases velocity measurements, were taken within tornadoes. In seven of these events, the intercepted tornadoes were within the radar-indicated or visually identified location of the supercell low-level mesocyclone. In the remaining two cases, the intercepted tornadoes occurred outside of this region and were located along either the RFD gust front or an internal RFD outflow surge boundary. These measurements add to the small collection of observations previously documented from in and near tornadoes (Table 1). Peak near-ground pressure deficits ranged from 5 to 100 hPa, with maximum instantaneous 3-m wind speeds of 40–50 m s−1 in the three cases where mobile mesonet data were available. Much higher winds no more than a few tens of meters above the surface, possibly in the range associated with EF-5 intensity, were estimated from surface pressure measurements in the 24 June 2003 Manchester, South Dakota, tornado.
In addition to reporting these near-ground measurements, we have attempted to infer vortex structure for each tornado where possible, based on surface pressure measurements, available videos, and eyewitness accounts. The observed tornadoes ranged from low-swirl single-cell vortices or medium-swirl two-celled vortices to high swirl-ratio two-cell multiple vortex tornadoes. The low-swirl single-cell and medium-swirl two-cell vortex structures are difficult to differentiate with only the pressure traces given their similar surface pressure signal. Observational documentation of vortex structure provides essential ground-truth evidence for comparison with high-resolution vortex simulations. Both observational and modeling research on tornado structures is essential for improving our understanding of tornado damage patterns and variations in intensity.
Our results show similarities to the previously documented measurements of tornadoes listed in Table 1, except for the 21 August 1904 report and the Blair et al. (2008) report from the EF-2 Tulia, Texas, tornado of 21 April 2007 (NCDC 2007). These pressure measurements nearly double the largest other measurement in the sample, taken on 24 June 2003 in a near-direct transect of a tornado causing F-4 damage at the measurement site. As stated in Blair et al. (2008), the accuracy of the 192-hPa near-ground pressure deficit from 21 August 1904, measured with an aneroid barometer, is unsubstantiated because of issues such as dynamic pressure effects and the harsh conditions under which the measurement was observed.
In an attempt to make a comparison between the 24 June 2003 Manchester, South Dakota, tornado pressure observations and those reported by Blair et al. for the 21 April 2007 Tulia, Texas, tornado, cyclostrophic wind speed approximations were made. As noted previously, although these estimated wind speeds may not be appropriate near the surface, they may have more validity within a few tens of meters above the surface. Using a 95-hPa pressure drop associated with the tornadic signal in the Manchester tornado pressure trace (Fig. 7c) and air density appropriate for the Manchester tornado case results in a cyclostrophic wind speed of approximately 95 m s−1, consistent with that found in Lee et al. (2004). Similarly, using a 170-hPa pressure drop associated with the Tulia, Texas, tornado in Blair et al. (2008, their Fig. 11) and air density appropriate for that case results in a cyclostrophic wind speed of approximately 135 m s−1. Ignoring tornado translation speeds for simplicity (note that the Tulia, Texas, tornado was moving approximately 10 m s−1 faster than the Manchester tornado), if we assume damage indicators should roughly scale with these wind speed estimates, then an apparent inconsistency exists between the F-4 site damage at the Manchester tornado deployment site and the EF-2 damage adjacent to the Tulia tornado observing location. The wind speed estimate for the Manchester tornado case agrees very well with F-4 site damage and falls well within the wind speed range associated with an F-4 tornado (Fujita 1981). Perhaps the Tulia tornado, during a short stage of great intensity, fortuitously managed to miss potential damage targets. However, it is hard to rectify the site damage, including the damage to the observing vehicle, given the wind speeds that should have accompanied such a large pressure gradient (even at the 2–3-m level). Although we view this 194-hPa pressure drop in the Tulia tornado with some apprehension because of the inconsistencies cited, we agree with Blair et al. that acquiring more near-ground tornado measurements, especially in violent tornadoes, is needed to better understand what the maximum pressure deficit is in high-end events.
In addition to presenting a comparative range of observed pressure deficits for a wide variety of tornado sizes and intensities, the results herein emphasize the need for high spatial and temporal resolution sampling in and near tornadoes. Measurements obtained using mobile mesonet stations are coarse compared to the scale of the tornadoes sampled. We agree with Lewellen et al. (1997) that a sampling frequency of a least a few hertz is required, and we recommend a sampling frequency of at least 10 Hz for in situ sampling missions. To improve comparisons made to laboratory and numerical simulations, enhanced horizontal and vertical spatial resolution in measurements from future tornado sampling missions is desirable. Because of the usual inability of mobile Doppler radar to reliably scan below approximately 20 m AGL, stereo photogrammetry could be used as a potential method to obtain estimates of near-ground wind velocities in some tornadoes. Estimating wind velocities using photogrammetry has been done extensively in the past; however, there are a number of limiting factors associated with photogrammetric analysis (Forbes and Bluestein 2001). Innovations in anemometry are being actively pursued and will represent a critical part of future TWISTEX in situ deployments. Additionally, increasing the number of HITPR style probes, aligned linearly, will better resolve the pressure distribution in a tornado.
In the quest for obtaining measurements in close proximity to a tornado, researchers undoubtedly place themselves in locations of heightened risk. Though operational safety was given highest priority during the field campaigns, two of these events clearly illustrate the unpredictability of tornadogenesis away from the classic tornadogenesis position in supercells. In the vertical-vorticity-rich environments of some supercells, quickly forming RFD gust front tornadoes and satellite tornadoes near the low-level mesocyclone periphery can be a significant threat to teams working nearby. With the considerable number of mobile platforms involved in large tornado-related field experiments, situational awareness becomes imperative in reducing the number of direct encounters. These challenges underscore the value of continuing research to improve our understanding of multiple potential tornadogenesis mechanisms in supercells while employing refined operations strategies that utilize the latest mobile communication technologies.
Efforts will continue in future TWISTEX field projects to collect high-resolution measurements of the tornadic flow field near the surface. In situ observational goals are guided by the desire to obtain information that can guide structural engineering interests, aid in assessing damage potential, and help to verify or lead to adjustments in the wind estimates used in the EF scale.
Acknowledgments
Partial support for the research was provided by NOAA Grants NA06OAR4600230 and NA08OAR4600887 and by the National Geographic Society. The authors would like to acknowledge all past participants of the TWISTEX field project as well as Chris Collura, Jeff Piotrowski, Amos Magliocco, Doug Kiesling, Roger Hill, and Jerry Funfsinn for their video contributions to this research. The manuscript was substantially improved thanks to the constructive comments of Dr. David Schultz, Dr. David Lewellen, and two anonymous reviewers. We are also grateful to Matt Grzych, Patrick Skinner, and Daryl Herzmann for their technical assistance.
REFERENCES
Adlerman, E. J., and K. K. Droegemeier, 2000: A numerical simulation of cyclic tornadogenesis. Preprints, 20th Conf. on Severe Local Storms, Orlando, FL, Amer. Meteor. Soc., 603–606.
Blair, S. F., D. R. Deroche, and A. E. Pietrycha, 2008: In situ observations of the 21 April 2007 Tulia, Texas tornado. Electron. J. Severe Storms Meteor., 3 , 1–27.
Bluestein, H. B., 1983: Measurements in the vicinity of severe thunderstorms and tornadoes with TOTO: 1982–1983 results. Preprints, 13th Conf. on Severe Local Storms, Tulsa, OK, Amer. Meteor. Soc., 89–92.
Brandes, E. A., 1978: Mesocyclone evolution and tornadogenesis: Some observations. Mon. Wea. Rev., 106 , 995–1011.
Brandes, E. A., 1993: Tornadic thunderstorm characteristics determined with Doppler radar. The Tornado: Its Structure, Dynamics, Prediction and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 143–159.
Brown, R. A., L. R. Lemon, and D. W. Burgess, 1978: Tornado detection by pulsed Doppler radar. Mon. Wea. Rev., 106 , 29–39.
Church, C. R., and J. T. Snow, 1993: Laboratory models of tornadoes. The Tornado: Its Structure, Dynamics, Prediction and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 19–39.
Church, C. R., J. T. Snow, G. L. Baker, and E. M. Agee, 1979: Characteristics of tornado-like vortices as a function of swirl ratio: A laboratory investigation. J. Atmos. Sci., 36 , 1755–1766.
Cook, N. J., 1985: The Designer’s Guide to Wind Loading of Building Structures, Part 1: Background, Damage Survey, Wind Data and Structural Classification. Butterworths, 371 pp.
Davies-Jones, R. P., 1982: Tornado dynamics. Thunderstorm Morphology and Dynamics, E. Kessler, Ed., Vol. II, Thunderstorms: A Social, Scientific, and Technological Documentary, U.S. Government Printing Office, 297–361.
Davies-Jones, R. P., 1986: Tornado dynamics. Thunderstorm Morphology and Dynamics, E. Kessler, Ed., Vol. II, Thunderstorms: A Social, Scientific, and Technological Documentary, University of Oklahoma Press, 197–236.
Davies-Jones, R. P., and E. Kessler, 1974: Tornadoes. Weather and Climate Modification, W. N. Hess, Ed., John Wiley and Sons, 552–595.
Doswell III, C. A., H. E. Brooks, and N. Dotzek, 2009: On the implementation of the Enhanced Fujita scale in the USA. Atmos. Res., 93 , 554–563.
Finley, C. A., and B. D. Lee, cited. 2004: High resolution mobile mesonet observations of RFD surges in the June 9 Basset, Nebraska supercell during Project ANSWERS 2003. [Available online at http://ams.confex.com/ams/11aram22sls/techprogram/paper_82005.htm].
Finley, C. A., and B. D. Lee, cited. 2008: Mobile mesonet observations of an intense RFD and multiple RFD gust fronts in the May 23 Quinter, Kansas tornadic supercell during TWISTEX 2008. [Available online at http://ams.confex.com/ams/24SLS/techprogram/paper_142133.htm].
Finley, C. A., W. R. Cotton, and R. A. Pielke, cited. 2002: Tornadogenesis in a simulated HP supercell. [Available online at http://ams.confex.com/ams/SLS_WAF_NWP/techprogram/paper_47216.htm].
Forbes, G. S., and H. B. Bluestein, 2001: Tornadoes, tornadic thunderstorms, and photogrammetry: A review of the contributions by T. T. Fujita. Bull. Amer. Meteor. Soc., 82 , 73–96.
Fujita, T., 1958: Tornado cyclone: Bearing system of tornadoes. Proc. Seventh Conf. on Radar Meteorology, Miami Beach, FL, Amer. Meteor. Soc., K31–K38.
Fujita, T., 1981: Tornadoes and downbursts in the context of generalized planetary scales. J. Atmos. Sci., 38 , 1511–1534.
Grasso, L. D., and W. R. Cotton, 1995: Numerical simulation of a tornado vortex. J. Atmos. Sci., 52 , 1192–1203.
Grzych, M. L., B. D. Lee, and C. A. Finley, 2007: Thermodynamic analysis of supercell rear-flank downdrafts from Project ANSWERS. Mon. Wea. Rev., 135 , 240–246.
Hirth, B. D., J. L. Schroeder, and C. C. Weiss, 2008: Surface analysis of the rear-flank downdraft outflow in two tornadic supercells. Mon. Wea. Rev., 136 , 2344–2363.
Le, K., F. L. Haan Jr., W. A. Gallus Jr., and P. P. Sarkar, 2008: CFD simulations of the flow field of a laboratory-simulated tornado for parameter sensitivity studies and comparison with field measurements. Wind Struct., 11 , 75–96.
Lee, B. D., and R. B. Wilhelmson, 1997: The numerical simulation of nonsupercell tornadogenesis. Part II: Evolution of a family of tornadoes along a weak outflow boundary. J. Atmos. Sci., 54 , 2387–2415.
Lee, B. D., C. A. Finley, and P. Skinner, cited. 2004: Thermodynamic and kinematic analysis of multiple RFD surges for the 24 June 2003 Manchester, SD cyclic tornadic supercell during Project ANSWERS 2003. [Available online at http://ams.confex.com/ams/11aram22sls/techprogram/paper_82000.htm].
Lee, B. D., C. A. Finley, and T. M. Samaras, cited. 2008: Thermodynamic and kinematic analysis near and within the Tipton, KS tornado on May 29 during TWISTEX 2008. [Available online at http://ams.confex.com/ams/24SLS/techprogram/paper_142078.htm].
Lee, J. J., T. M. Samaras, and C. R. Young, 2004: Pressure measurements at the ground in an F-4 tornado. Preprints, 22nd Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., 15.3. [Available online at http://ams.confex.com/ams/11aram22sls/techprogram/paper_81700.htm].
Lemon, L. R., and C. A. Doswell III, 1979: Severe thunderstorm evolution and mesocyclone structure as related to tornadogenesis. Mon. Wea. Rev., 107 , 1184–1197.
Lewellen, D. C., and W. S. Lewellen, 2007: Near-surface intensification of tornado vortices. J. Atmos. Sci., 64 , 2176–2194.
Lewellen, W. S., 1993: Tornado vortex theory. The Tornado: Its Structure, Dynamics, Prediction and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 19–39.
Lewellen, W. S., D. C. Lewellen, and R. I. Sykes, 1997: Large-eddy simulation of a tornado’s interaction with the surface. J. Atmos. Sci., 54 , 581–605.
Lewellen, W. S., D. C. Lewellen, and J. Xia, 2000: The influence of a local swirl ratio on tornado intensification near the surface. J. Atmos. Sci., 57 , 527–544.
Markowski, P. M., 2002a: Hook echoes and rear-flank downdrafts: A review. Mon. Wea. Rev., 130 , 852–876.
Markowski, P. M., 2002b: Mobile mesonet observations on 3 May 1999. Wea. Forecasting, 17 , 430–444.
Markowski, P. M., J. M. Straka, and R. N. Rasmussen, 2002: Direct surface thermodynamic observations within the rear-flank downdrafts of nontornadic and tornadic supercells. Mon. Wea. Rev., 130 , 1692–1721.
Marquis, J., Y. Richardson, J. Wurman, P. Markowski, and D. Dowell, 2008a: Mobile radar observations of tornadic supercells with multiple rear-flank gust fronts. Preprints, 24th Conf. on Severe Local Storms, Savannah, GA, Amer. Meteor. Soc., 19.3. [Available online at http://ams.confex.com/ams/24SLS/techprogram/paper_142112.htm].
Marquis, J., Y. Richardson, J. Wurman, and P. Markowski, 2008b: Single- and dual- Doppler analysis of a tornadic vortex and surrounding storm-scale flow in the Crowell, Texas, supercell of 30 April 2000. Mon. Wea. Rev., 136 , 5017–5043.
Mishra, A. R., D. L. James, and C. W. Letchford, 2008: Physical simulation of a single-celled tornado-like vortex, Part B: Wind loading on a cubical model. J. Wind Eng. Ind. Aerodyn., 96 , 1258–1273.
NCDC, 2002: Storm Data. Vol. 44, No. 5, 147 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801].
NCDC, 2003a: Storm Data. Vol. 45, No. 6, 420 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801].
NCDC, 2003b: Storm Data. Vol. 45, No. 6, 272 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801].
NCDC, 2004: Storm Data. Vol. 46, No. 6, 106 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801].
NCDC, 2007: Storm Data. Vol. 49, No. 4, 313 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801].
NCDC, 2008a: Storm Data. Vol. 50, No. 5, 198 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801].
NCDC, 2008b: Storm Data. Vol. 50, No. 5, 189 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801].
Potter, S., 2007: Fine-tuning Fujita. Weatherwise, 60 , 64–71.
Rasmussen, E. N., R. Davies-Jones, and R. L. Holle, 2003: Terrestrial photogrammetry of weather images acquired in uncontrolled circumstances. J. Atmos. Oceanic Technol., 20 , 1790–1803.
Samaras, T. M., and J. J. Lee, 2004: Pressure measurements within a large tornado. Preprints, Eighth Symp. on Integrated Observing and Assimilation Systems for Atmosphere, Oceans and Land Surface, Seattle, WA, Amer. Meteor. Soc., 4.9. [Available online at http://ams.confex.com/ams/84Annual/techprogram/paper_74267.htm].
Sengupta, A., F. L. Haan, P. P. Sarkar, and V. Balaramudu, 2008: Transient loads on buildings in microburst and tornado winds. J. Wind Eng. Ind. Aerodyn., 96 , 2173–2187.
Simiu, E., and R. H. Scanlan, 1996: Wind Effects on Structures: Fundamentals and Applications to Design. John Wiley and Sons, 688 pp.
Snow, J. T., C. R. Church, and B. J. Barnhart, 1980: An investigation of the surface pressure fields beneath simulated tornado cyclones. J. Atmos. Sci., 37 , 1013–1026.
Stout, G. E., and F. A. Huff, 1953: Radar records Illinois tornadogenesis. Bull. Amer. Meteor. Soc., 34 , 281–284.
Straka, J. M., E. N. Rasmussen, and S. E. Fredrickson, 1996: A mobile mesonet for finescale meteorological observations. J. Atmos. Oceanic Technol., 13 , 921–936.
Tepper, M., and W. E. Eggert, 1956: Tornado proximity traces. Bull. Amer. Meteor. Soc., 37 , 152–159.
Wallace, J. M., and P. V. Hobbs, 1977: Atmospheric Science: An Introductory Survey. Academic Press, 467 pp.
Ward, N. B., 1961: Radar and surface observations of tornadoes of May 4, 1961. Proc. Ninth Weather Radar Conf., Kansas City, MO, Amer. Meteor. Soc., 175–180.
Ward, N. B., 1972: The exploration of certain features of tornado dynamics using a laboratory model. J. Atmos. Sci., 29 , 1194–1204.
Wicker, L. J., and R. B. Wilhelmson, 1995: Simulation and analysis of tornado development and decay within a three-dimensional supercell thunderstorm. J. Atmos. Sci., 52 , 2675–2703.
Wieringa, J., 1993: Representative roughness parameters for homogeneous terrain. Bound.-Layer Meteor., 63 , 323–363.
Winn, W. P., S. J. Hunyady, and G. D. Aulich, 1999: Pressure at the ground in a large tornado. J. Geophys. Res., 104 , 22067–22082.
Wurman, J., and T. Samaras, 2004: Comparison of in-situ pressure and DOW Doppler winds in a tornado and RHI vertical slices through 4 tornadoes during 1996-2004. Preprints, 22nd Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., 15.4. [Available online at http://ams.confex.com/ams/11aram22sls/techprogram/paper_82352.htm].
Wurman, J., J. M. Straka, E. N. Rasmussen, M. Randall, and A. Zahrai, 1997: Design and deployment of a portable, pencil-beam, pulsed, 3-cm Doppler radar. J. Atmos. Oceanic Technol., 14 , 1502–1512.
Schematic diagram of a (a) one-cell vortex, (b) two-cell vortex, and (c) two-cell vortex with secondary vortices (Davies-Jones 1986).
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Radial profiles of time-averaged surface static pressure from laboratory simulations of tornadoes with different swirl ratios. The values of swirl ratio associated with profiles (a)–(g) are 0.0, 0.14, 0.29, 0.42, 0.60, 1.17, and 1.79, respectively. Figure is from Snow et al. (1980) and Church and Snow (1993).
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
(a) HITPR 1 (and 2) probe (diameter = 0.51 m), (b) video probe (diameter = 0.76 m) and (c) mobile mesonet (three cars on left).
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Location of tornado intercepts (white star) relative to the nearest WSR-88D radar-indicated storm and equivalent radar reflectivity factor (dBZ) for (a) 7 May 2002, (b) 15 May 2003, (c) 24 Jun 2003 cases 1 and 2, (d) 11 Jun 2004, (e) 10 May 2008, (f) 23 May 2008, (g) 29 May 2008 case 1, and (h) 29 May 2008 case 2. Temporal differences exist between the intercept location and the radar-indicated storm position.
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Tornado on 7 May 2002 (a) 157 s before and (b) 117 s before traversing the instrumented probes. Arrows identify secondary vortices. Time of tornado intercept is 0000:07 UTC.
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Schematic diagrams of instrumentation deployment relative to the estimated tornado track (gray swaths) for (a) 7 May 2002, (b) 15 May 2003, (c) 24 Jun 2003 cases 1 and 2, (d) 11 Jun 2004, (e) 10 May 2008, (f) 23 May 2008, (g) 29 May 2008 case 1, and (h) 29 May 2008 case 2. A range of uncertainty in the placement of the HITPR 2 probe is indicated in (c).
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Time series of surface pressure deficits, normalized to the determined location of the tornado’s central axis, from all successful HITPR probe deployments, including (a) 7 May 2002, (b) 15 May 2003, (c) 24 Jun 2003 case 1, (d) 24 Jun 2003 case 2, (e) 11 Jun 2004, and (f) 29 May 2008 case 1. In (c)–(e), 0 s corresponds to the maximum pressure deficit. In (a),(b),(f) a 10-s moving average was applied to time series to determine the approximate time-relative center point within the primary pressure deficit. M3 3-m-AGL pressure measurements are included in (f).
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Tornado on 24 Jun 2003 (a) 130 s before and (b) 86 s before passing over the HITPR 1 probe. Time of tornado intercept is 0046:52 UTC.
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Tornado on 24 Jun 2003 (a) 62 s before and (b) 15 s after passing over the HITPR 2 probe. Time of tornado intercept is 0050:02 UTC.
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Tornado on 11 Jun 2004 (a) 150 s before and (b) at the time of intercept. Time of tornado intercept is 1923:46 UTC.
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Wind speed (m s−1) and wind direction (°) at 3 m AGL vs time for (a) 10 May 2008, (c) 23 May 2008, (e) 29 May 2008 case 1, and (g) 29 May 2008 case 2. Pressure deficit (hPa) and wind speed (m s−1) at 3 m AGL vs time for (b) 10 May 2008, (d) 23 May 2008, (f) 29 May 2008 case 1, and (h) 29 May 2008 case 2. Figures are normalized as in Fig. 7. HITPR 1 surface pressure measurements are included in (f).
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Tornado on 29 May 2008 case 1 approximately 1 km upstream and 60 s prior to intercepting the instrumented probes.
Citation: Monthly Weather Review 138, 7; 10.1175/2010MWR3201.1
Summary of historical near-ground tornado measurements. Regular font indicates serendipitous measurements adopted from Davies-Jones and Kessler (1974). Italic font indicates other formally documented measurements (9 Jun 1995: Winn et al. 1999; 22 Apr 2007: Blair et al. 2008). Bold font indicates measurements presented herein.
* Supplemental information related to this paper is available at the Journals Online Web site: http://dx.doi.org/10.1175/2010MWR3201.s1.