1. Introduction
Improved knowledge of how deep, moist convection is initiated is of fundamental importance to the U.S. National Weather Service (NWS) and other operational forecasting groups. Improved nowcasts (i.e., 0–3-h forecasts) and short-term forecasts (6–12 h) of convective initiation and the mode of early convection could prove invaluable to severe local storm forecasters to help focus their attention on rapidly evolving mesoscale weather scenarios.
It is our impression that dramatic warm season weather forecast failures are often the result of an inability to anticipate the initiation of convection, or forecasting convection that fails to develop. Incorrect convective initiation forecasts cause quantitative precipitation forecast errors, and large errors in temperature forecasts are induced by the failure to anticipate the development of large, rain-cooled airmasses and cloudiness. On some occasions during the spring, large regions are forecast to have a moderate or high risk of severe thunderstorms, but the skies remain clear. Though statistical techniques may ultimately provide reliable forecasts of precipitation amounts and coverage, successful mesoscale prediction with numerical atmospheric forecast models will require improved understanding and an accurate explicit or parameterized representation of convective initiation. This forecast challenge stems from a fundamental lack of knowledge regarding the processes that allow or prevent the initiation of deep, moist convection.
In spite of the difficulties of collecting representative in situ measurements on the scales of individual incipient clouds, special mesoscale observations collected by a variety of fixed and mobile instrumented platforms have provided important insights regarding the environments of developing convection. The capability to detect clear air boundary layer structures near developing moist convection is offered by single or multiple ground-based Doppler radars (Eymard 1984; Parsons et al. 1991; Wilson et al. 1994) and airborne Doppler radars (Wakimoto et al. 1996). Research aircraft have provided in situ observations of airflow and thermal properties near boundaries that are believed to play a central role in the initiation of convection (e.g., Ziegler and Hane 1993). The subjective interpretation of geostationary satellite data provides fundamentally important guidance on the initiation of convection along boundaries (Purdom 1982), while cloud-mesoscale models have played an increasingly important role by providing complete and internally consistent datasets with which to test various hypotheses. This combination of observations and models has led previous investigators to conclude that storm initiation is closely linked to boundary layer convergence lines.
The primary effect of a convergence line is to deepen the moist layer locally and provide a region potentially favorable for deep convection. The initiation of moist convection has been investigated along a variety of boundaries in differing geographical regions, including Florida sea breezes (Wakimoto and Atkins 1994; Fankhauser et al. 1995); mountain-induced ridge-top and lee convergence zones in New Mexico (Raymond and Wilkening 1982) and Colorado (Banta 1984), respectively;Colorado Front Range convergence zones (Wilson and Schreiber 1986); and drylines (Hane et al. 1993; Ziegler et al. 1997; Hane et al. 1997; Atkins et al. 1998) as well as the intersection of an east–west baroclinic zone with a dryline (Bluestein et al. 1990) on the southern U.S. plains. Koch and Ray (1997) use Weather Surveillance Radar-1988 Doppler (WSR-88D) data to document the initiation of thunderstorms along sea breezes, the Piedmont front, and other boundaries in North Carolina. Moist boundary layer air may be elevated to its lifted condensation level (LCL), forming a “forced” cumulus, while additional forced lifting may bring enough air to the level of free convection (LFC) to form an “active” deep, moist convective cloud (Stull 1985). Deep convection may form at the intersection of a convergence line with horizontal convective rolls where enhanced updrafts are present (Wilson et al. 1992; Atkins et al. 1995) as well as at collision points of thunderstorm outflows with other outflows or sea breezes (Kingsmill 1995).
The sensitivity of boundary layer cumulus convection to wind shear and thermal stratification effects and their mesoscale variability is receiving increased attention by researchers. Storm initiation, organization, and lifetime may be enhanced when the convective clouds move at a velocity similar to that of the convergence line (Wilson and Megenhardt 1997). The horizontal mesoscale variation of moisture may dictate which of several simultaneously preexisting mesoscale boundaries initiate deep convection in a given case (e.g., Weaver et al. 1994). The initiation of convection is very sensitive to small-scale variability in the boundary layer thermodynamics (Mueller et al. 1993; Weckwerth et al. 1996;Crook 1996). The existence of adiabatic updraft plumes in the boundary layer raises the question of whether near-surface air feeds developing boundary layer cumulus clouds (Rennó and Williams 1995).
Thunderstorms often form near drylines of the southern U.S. plains (Rhea 1966; Bluestein and Parker 1993), but their initiation is difficult to forecast. According to an “ingredients based” approach to thunderstorm forecasting (Johns and Doswell 1992; McNulty 1995), the coincidence of low convective inhibition (CIN) with high convective available potential energy (CAPE) and deep tropospheric wind shear along a well-defined mesoscale boundary with strong low-level convergence strongly suggest a high likelihood for the initiation of severe convection. During the 1994 and 1995 field phases of the Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX) (Rasmussen et al. 1994), deep moist dryline convection did not develop in several cases despite very favorable ingredients. In each of these cases where deep convection did not develop, small values of CIN at the dryline suggested that rising air parcels could easily attain the LCL and LFC and grow into deep convection (Colby 1984). Since severe storms developed along several drylines observed during the 1991 Central Oklahoma Profiler Studies project (COPS-91) (e.g., Hane et al. 1993), a comparison of the COPS and VORTEX cases affords the opportunity to begin exploring the necessary requirements for initiating deep dryline convection.
In accord with a key U.S. Weather Research Program (USWRP) objective of refining quantitative precipitation forecasts via improved observations and knowledge of boundary layer processes (Emanuel et al. 1995), the present study focuses on the impact of boundary layer evolution on the initiation or inhibition of deep moist convection along the dryline. Employing selected cases from the 1991, 1994, and 1995 field projects, we examine the connections between the development of shallow and deep moist convection, the intensity of the horizontal thermal gradients and the vertical circulations that accompany cloud initiation, and the realization of convective instability. Section 2 presents the data sources and analysis techniques, while section 3 presents the case studies. Discussion of the results in section 4 is followed by conclusions in section 5.
2. Mesoscale data analysis
In several cases during COPS and VORTEX, in situ aircraft observations were concentrated along dryline segments to document the potential impact of prestorm environmental conditions on convective initiation and storm development. The National Oceanic and Atmospheric Administration (NOAA) P-3 aircraft measured environmental variables across the dryline during the daytime during each experiment, concentrating on stepped traverses but including horizontal box survey patterns in the dryline environment. The National Center for Atmospheric Research (NCAR) Electra joined the P-3 to probe drylines during VORTEX-95, adding a clear air remote sensing capability. Both the P-3 and the Electra were based at Will Rogers World Airport in Oklahoma City, Oklahoma, during the COPS and VORTEX projects.
Proceeding from objective techniques described in Ziegler and Hane (1993) and Hane et al. (1993), the P-3 stepped traverse data were analyzed to document the properties of the boundary layer across the dryline. In the first analysis step, the P-3 horizontal wind measurements were filtered to provide a smooth input divergence profile for kinematically calculating vertical motion (Mohr et al. 1986). Prefiltering of the input east–west (u) wind component1 time series was accomplished with a four-pass application of a triangular weighting function varying linearly from a value of one at the current scan point to zero over a lag time of ±12 s,2 while boundary points of each leg were smoothed with a simple elliptic filter. A one-pass application of the Barnes (1964) scheme was used to spatially interpolate all data, including the prefiltered u component, to a regular Cartesian grid that was oriented in the east–west direction and centered on the sets of flight legs in a given stepped traverse pattern. The analysis grid had a horizontal grid spacing of 0.2 km and a vertical grid spacing of 0.1 km. Vertical velocities were derived from the interpolated u-component field by upward integration of the Boussinesq continuity equation from w = 0 at ground level.
The Barnes (1964) weighting function takes the form w = exp(−r2/κ), where r is the distance in the x–z plane separating a datum and a grid point and κ is a smoothing parameter. Since κ is constant, the filtering is uniform or isotropic in the plane. The frequency response Do of the Barnes filter (0 ⩽ Do ⩽ 1) may be expressed as Do = exp[−κ*(π/λ*)2] (Koch et al. 1983), where κ* = κ/L2, L = 2Δ, and λ* = λ/L. Picking Δ = 0.3 km to represent the coarsest (vertical) data spacing in the stepped traverses, and selecting κ* = 0.439, we obtain a theoretical filtering response (i.e., 100 × Do) of about 50% at λ = 1.5 km and ∼1% at the Nyquist wavelength λ = L.
Photogrammetric cloud analyses were performed using the time-lapse 16-mm color film or video camera systems flown on the P-3 during the COPS and VORTEX projects, respectively. Advantages of the photogrammetric cloud analysis over visible satellite imagery, which was also used, include the following: 1) improved resolution of smaller cumuli and more precise east–west positioning of clouds than possible from satellite imagery, 2) measurements of cloud base and top heights, 3) detailed imaging of overall cumulus cloud morphology. Although image data were gathered through most of a given mission, the photogrammetric analyses were restricted to the images recorded on stepped traverse legs. Presented in detail in appendix A, the photogrammetric technique is based on known information of the position and absolute east–west speed of the P-3, the measured extent of cloud boundaries in the horizontal and vertical image directions, and several known or estimated properties of the camera lens and recording systems (Holle 1988). The output of the photogrammetric analysis are the positions and ranges of bounding boxes that just enclose individual clouds or cloud clusters. In cases of complex cloud fields, only the closest clouds (i.e., relatively unmasked by intervening cumuli) are recorded. Due to a gap in the image record between 2220 and 2300 UTC (all times are UTC) on 15 May 1991 (about 12% of the combined image record), the properties of the cloud field were approximated from a detailed weather event log during that time interval.
Two additional sources of in situ pressure, temperature, dewpoint, and horizontal wind measurements were available. Atmospheric soundings were obtained from surface vehicles employing the Mobile Cross-chain Loran Atmospheric Sounding System (M-CLASS), based on CLASS technology developed by NCAR (Lauritzen et al. 1987) and converted by the National Severe Storms Laboratory (NSSL) for mobile operation in vans (e.g., Rust et al. 1990). Values of CAPE and CIN were computed for each sounding based on the virtual temperature buoyancy (Doswell and Rasmussen 1994) and the lowest 50-mb average parcel. Doswell and Rasmussen (1994) point out that the virtual warming effect of water vapor may be sufficient to substantially reduce (or even eliminate) CIN and significantly augment small values of CAPE. The difference between the balloon rise rate and its rise rate in still air, roughly 5 m s−1, was used to estimate mesoscale vertical motion to within approximately ±0.5 m s−1. A total of two M-CLASS vans were used in COPS-91, while as many as four M-CLASS vans were deployed during VORTEX. A “mobile mesonet” of as many as 15 vehicles (including M-CLASS vans) were equipped with instruments to obtain profiles of thermodynamic variables across the dryline during VORTEX (Straka et al. 1996). At a nominal traverse speed of 25 m s−1, the 6-s sampling rate of a mobile mesonet corresponded to a spatial sampling interval of 150 m.
The information from photogrammetry was depicted on individual soundings to estimate the typical environments of observed cumulus clouds. Linear interpolation of the observed cloud base and top heights to the observed pressure–height profile of the sounding was used to estimate the corresponding pressure values. Equating cloud-base pressure to the LCL pressure or condensation pressure p* and following Betts (1984), the corresponding LCL temperature TLCL was inferred from a piecewise-linear regression of measured TLCL versus p* along a single mid–boundary layer aircraft traverse through the dryline for each case.3 After plotting a given cloud-base point, a moist-adiabatic profile was constructed and plotted upward from the cloud base to the estimated cloud-top pressure. Although moist-adiabatic conditions may rarely be achieved in cumuli due to mixing of cloudy and environmental air (e.g., Pruppacher and Klett 1978), the most buoyantly unstable profile could be visualized in this manner.
3. Results
a. 15 May 1991 dryline (COPS-91)
Special mesoscale field observations for the 15 May 1991 dryline case during the COPS-91 experiment are described by Hane et al. (1993). These observations represent the most extensive data collection ever obtained for a single dryline, with aircraft traverses and mobile soundings being obtained at approximately 1- to 2-h intervals from midmorning through early evening in the eastern Texas panhandle region. Mesoscale modeling studies of this case have investigated the mechanisms of deep convective initiation (Ziegler et al. 1997), the sensitivity of dryline formation to the spatial soil moisture distribution (Shaw et al. 1997), and the impact of land use patterns on moist convection (Pielke et al. 1997).
The dryline sharpened just east of Amarillo, Texas, during midmorning and moved slowly eastward, becoming nearly stationary during the late afternoon just west of McLean, Texas, in the eastern Texas panhandle. By midafternoon the dryline was oriented approximately north–south around 100.5°W longitude (the east–west analysis origin) (Fig. 1a), and convection had begun to form along its extent from the northern Texas panhandle into western Kansas and northeastern Colorado (Fig. 1b). A photogrammetric cloud analysis based on all east–west traverses (Fig. 1c) reveals a high spatial density of cumulus clouds through the afternoon along and just east of the dryline.
The dryline was probed by east–west stepped traverses of the P-3 aircraft east of Amarillo, Texas. During the traverse from 1532 to 1646 (Fig. 2a), the dryline was already marked by pronounced east–west gradients of water vapor mixing ratio (qv) and virtual potential temperature (θv); the deceleration, stagnation, and lifting of westerly flow; and a weak moisture bulge. Photogrammetric data indicated the presence of a shallow stratocumulus layer east of the dryline, and a small high-based (i.e., much higher than the moist boundary layer depth) cumulus west of the dryline. By the 1649–1804 traverse (Fig. 2b), the convergence, lifting, vertical wind shear, and horizontal gradients of the nearly stationary dryline had become much better defined, while the stratocumulus just east of the dryline was in the process of dissipating. By the 1807–1935 traverse (Fig. 2c), gradients had weakened as convective boundary layer (CBL) growth east of the quasi-stationary dryline transported low-level moist, (virtually) potentially cool air upward. Mainly high-based, small cumulus–fractus and cumulus had developed around the surface dryline location. By the 2014–2208 traverse (Fig. 2g), the dryline had moved eastward 5 km to x = −55 km. Hence, the original dryline had nearly dissipated by midafternoon due to deep CBL growth and vertical mixing to the east of the dryline. Widespread, high-based cumulus and towering cumulus clouds had developed from the surface dryline location eastward for a distance of over 55 km.
By the 2220–2305 traverse (Fig. 2h), the dryline had reformed at x = −22 km, with additional concentrated zones of moisture convergence and lifting at x = −14 km, x = −3 km, and x = 11 km [i.e., “multiple drylines,” as in Hane et al. (1993) and Crawford and Bluestein (1997)]. A deep surge of easterly winds and moisture at low levels (i.e., “moisture surge” or simply“surge”) was present east of the mesoscale boundary at x = −3 km. Since the surge boundary is embedded in easterly flow while the flow veers and becomes westerly around the dryline, westward advection would cause the surge boundary to approach the dryline. Towering cumuli were concentrated over an interval of 30-km width to the east of the surface dryline location, and a towering cumulus with a lowered base (i.e., lower than adjacent and previous bases) had developed directly above the boundary at x = 11 km. By the 2307–0006 traverse (Fig. 2i), the westward moisture surge and eastward-propagating dryline had collided, dramatically intensifying horizontal gradients, moisture convergence, and mesoscale lift at the dryline and assisting the development of towering cumulus east of the dryline. A small cumulus above the analysis domain top of 3 km at x = −3 km was embedded in the intense mesoscale updraft, which lifted mostly hot, dry air from west of the dryline, explaining the rather high cloud base. Deep, swelling cumuli had developed east of the dryline, with a pronounced eastward tilt in response to the prevailing vertical wind shear in the cloud layer. Due to steadily decreasing θυ east of the dryline and surge, the collision of the two boundaries is an occlusion process.
During the P-3 maneuvers, M-CLASS soundings from NSSL-1 and NSSL-2 were obtained roughly 30 km west and east of the dryline, respectively. The 2311 NSSL-1 sounding (Fig. 3a) reveals a hot, dry, deeply well-mixed layer to about 600 mb west of the dryline, while the 2307 NSSL-2 sounding (Fig. 3b) shows the relatively cool, moist conditions east of the dryline. The NSSL-2 sounding displays a pronounced temperature inversion around 500 mb with nearly dry-adiabatic lapse rates above, suggesting that the western boundary layer top has been lifted about 100 mb while deeper tropospheric subsidence has capped the elevated residual layer to the east of the dryline. The inferred moist adiabats of unmixed cumuli and towering cumulus clouds are generally cooler than the environmental temperature profile, with deeper tops being restricted by the elevated inversion layer. The boundary layer moisture deepens and increases and the low-level winds back slightly at the NSSL-2 sounding site by 0057 (Fig. 3c).
Severe thunderstorms developed during the early evening along the dryline, producing small tornadoes near Garden City, Kansas, and strong tornadoes and large hail near Laverne, Oklahoma, and Shamrock, Texas (USDOC 1991). Small cumulonimbi developed just east of the dryline during late afternoon and early evening (e.g., Fig. 4a), preceding the storm over northeastern Wheeler County, Texas (Fig. 4b). The Shamrock storm was in its early development phase at 0101 on 16 May (Fig. 4b), had intensified and subsequently propagated southwestward4 toward the dryline by 0136 on 16 May (Fig. 4c), and had produced an F3 tornado beginning 8 miles south of Shamrock from 0217 to 0310. The Laverne tornado (0135–0211) had just formed by 0135 (Fig. 4c). The Laverne, Wheeler County, and Shamrock storms all initiated to the east of the dryline (e.g., note moist southeasterly flow in Figs. 4b,c to the west of the storms), subsequently organizing and ultimately moving northeastward away from the dryline in accord with the scenario described by Rhea (1966) and Bluestein and Parker (1993).
An analysis of the daytime evolution of CAPE and CIN reveals a difficult short-term forecasting scenario in which there is actually a slight, gradual reduction of instability from morning through midday east of the dryline as heating erodes the inhibition to convection (Fig. 5). The P-3 observations suggest that vertical mixing forces a drying trend in the lowest 50-mb layer during midday east of the dryline (Figs. 2c and 2g), accounting for the decreasing CAPE in spite of increasing surface temperatures. From around 2000 to 2307 at the Shamrock, Texas, sounding site, CAPE dramatically increases from under 1000 J kg−1 to over 2500 J kg−1. The westward-propagating moist and unstable air mass documented in Figs. 2h,i replenishes the supply of unstable air and increases the potential for severe convection despite slowly increasing CIN values. Our mesoanalysis suggests that a trough in the southeastern Texas Panhandle (Fig. 1a) delineated the western edge of the moisture surge observed in the P-3 data. The increased instability between the 2307 and 0057 NSSL-2 soundings (Figs. 3b,c) supports the hypothesis that strong, deep, localized moisture convergence and mesoscale lifting developed near the dryline during the early evening, the former possibly assisting the increase of CAPE and the latter helping to release the growing convective instability and initiate the intense storms. In a later section we will present further analyses of the P-3 data, which support the hypothesis of strengthening convergence near the dryline.
A study by Bluestein et al. (1990) provides added support to the notion that surges of moist, low θυ air near drylines may aid in initiating deep convection. Bluestein et al. (1990) performed an west-to-east traverse through a surge boundary (as depicted in their Fig. 8) on 28 May 1985, measuring decreasing temperature, increasing absolute humidity, and backing winds. We computed θυ from their data assuming a surface pressure (estimated from their soundings) of 950 mb, noting a sensitivity of 0.1 K mb−1 of error in assumed surface pressure (negligibly dependent on dewpoint temperature). Since pressure perturbations across such boundaries are only of order 0.1 mb, the implied uncertainty of the θυ gradient from the surface pressure uncertainty is negligible. A θυ drop of 1.5 K was computed over a 6-km west-to-east traverse of the surge based on their “mobile” measurements during their 2221–2241 traverse, yielding a θυ gradient comparable to values reported for an Oklahoma dryline by Ziegler and Hane (1989) and determined from the near-surface P-3 pass for the 15 May surge in the present study (Fig. 2h).
b. 7 June 1994 dryline (VORTEX-94)
The 7 June 1994 dryline was nearly stationary during the late afternoon and was oriented approximately north–south around 99.7°W longitude (the assumed east–west analysis origin) during the period of intensive VORTEX observation in extreme northwestern Oklahoma (Fig. 6a). Spatially isolated, high-based cumuli developed near the dryline (Fig. 6b) and were detected from the P-3 by photogrammetric analysis5 (Fig. 6c). The dryline was probed by east–west stepped traverses of the P-3 aircraft near Buffalo, Oklahoma. The first two stepped traverses of the P-3, performed in the periods 1934–2031 and 2030–2103 (not shown), revealed weak westerly wind shear and diffuse horizontal gradients of qυ and θυ across the dryline. High-based, weak cumulus convection had developed just west of the dryline within a 20-km-wide zone containing several localized 1 m s−1 updrafts in the boundary layer.
During the subsequent traverse from 2215 to 2245 (Fig. 7a), the dryline was marked by pronounced east–west gradients of moisture and θυ and the deceleration, stagnation, and lifting of westerly flow. A few scattered, high-based cumuli and towering cumuli had developed over and to the west of the dryline. By the 2248–2328 traverses (Fig. 7b), the dryline had moved eastward by 3 km and had become more sharply defined as cumulus coverage decreased. By the 2332–0001 traverses (Fig. 7c), the dryline had begun retreating to the west as the vertical circulation decreased in depth and cumuli dissipated.
Soundings from four M-CLASS systems were obtained in a quasi-linear west–east array across the dryline between 2100 and 2200. The NSSL-2 sounding (Fig. 8a) reveals a hot, dry, almost homogeneously well-mixed CBL air mass with westerly winds to the west of the dryline, while warm, moist air and southerly low-level winds are in place east of the dryline as documented in the soundings from NSSL-4 (Fig. 8b), NSSL-3 (Fig. 8c), and NCAR (Fig. 8d). With the exception of two rather deep towering cumuli, other cumulus clouds are largely restricted by the environmental temperature profile (Figs. 8b,c). These intense spatial gradients of stability parameters across the dryline (Fig. 9) have not previously been observed to the authors’ knowledge, although they have been successfully simulated by Ziegler et. al. (1997) and Shaw et al. (1997) using a research mesoscale model employing fine spatial resolution. Although CIN decreased to zero at the dryline in the presence of strong convergence and large CAPE, yet deep cumuli did not develop, we note the nearly 200-mb-deep layer of warm, dry air and substantial wind shear through which a boundary layer parcel would have to rise undiluted to initiate a storm. Bluestein et al. (1987) present another example of a dryline sounding (their Fig. 2f) in which CAPE and CIN values appeared to be favorable for convective initiation, yet where the LCL was within an elevated dry layer and storms did not develop. Although deep convection in such an environment is unlikely, storms could conceivably be initiated given localized mesoscale convergence extreme enough to dominate small-scale turbulent mixing and force large volumes of moist boundary layer air up to the LFC.
The dryline was probed by east–west traverses of the NSSL-2 mobile mesonet during midafternoon about 15 km south of a low-level traverse of the P-3. The horizontal profiles of absolute humidity from NSSL-2 and the P-3 are very similar despite slight differences of north–south position and timing of the two traverses (Fig. 10). The corresponding θυ profiles have different amplitudes, and surface gradients at the dryline are weaker at the surface than aloft, suggesting the existence of a superadiabatic layer near the surface east of the dryline.6 The θυ gradient at the location of the moisture increase marking the dryline (i.e., the 3 g kg−1 increase of qυ from x = −20 to x = −15 km) is rather weak at the surface, consistent with surface observations of dryline passage by Crawford and Bluestein (1997), but has a value of 1°C per 8 km east of the dryline. However, the θυ gradient as low as 150 m above ground level (AGL) (value of 1.5°C per 4 km) is considerable. A superadiabatic layer with peak amplitude near the dryline is confirmed by the M-CLASS soundings (Figs. 8b,c). The horizontal θυ and qυ gradients in the boundary layer at distances greater than 5 km east of the dryline (i.e., ∼1°–2°C and 4 g kg−1 per 30 km, respectively) are remarkably constant in time in the presence of afternoon heating, based on comparisons with other traverses (not shown), suggesting a nearly constant thermal solenoid intensity and a degree of horizontal uniformity of vertical mixing.
An additional three mobile mesonets (i.e., Probe-1, Cam-2, Cam-3) performed traverses in concert with NSSL-2, each confirming the existence of the steplike increases of moisture observed by NSSL-2 and the P-3 (not shown). In addition, some north–south variation of moisture is suggested during traverses in the interval between 2030 and 2130 (not shown). Convergence and veering of the surface wind from a southerly to westerly direction was noted to coincide with the moisture decreases on all traverses. On many of these traverses, decreases of virtual temperature occurred as moisture content increased.
c. 6 May 1995 dryline (VORTEX-95)
The 6 May 1995 dryline sharpened and remained nearly stationary during the midafternoon over the eastern Texas panhandle (Fig. 11a), just east of the east–west analysis origin, Amarillo, Texas. Isolated cumuli and cumulus bands developed in the vicinity of the dryline by midafternoon (Fig. 11b), and numerous small, shallow cumulus were documented by photogrammetric analysis at close range to the east–west stepped traverses flown by the P-3 (Fig. 11c). Atkins et al. (1998) correlated satellite-derived cloud locations with data from the Amarillo WSR-88D radar and the sensitive ELDORA Doppler radar flown on the NCAR Electra, showing that large isolated cumulus clouds developed along the dryline in regions of enhanced convergent airflow and mesoscale rising motion (e.g., Fig. 11c).
The dryline was probed by east–west stepped traverses of the P-3 aircraft southeast of Amarillo, Texas. During the traverse from 2145 to 2213 (Fig. 12a), the dryline was marked by pronounced east–west gradients of moisture and θυ and the deceleration, stagnation, and lifting of westerly flow. Photogrammetric data indicated the presence of shallow cumuli both along and east of the dryline. By the 2216–2246 traverses (Fig. 12b), the surface dryline had moved westward about 5 km as localized upward motion, abruptly deeper moisture, and cumulus activity around x = 55 suggest the presence of a persisting horizontal boundary at the top of the moist layer (i.e., “elevated dryline”). By the 2248–2319 traverse (Fig. 12c), the surface dryline had sharpened and propagated eastward, while cumulus developed within regions of rising motion around x = 40. By the 2322–2351 traverse (Fig. 12d), the surface dryline had moved 3 km westward as four areas of cumuli developed within mesoscale updrafts. By the 2354–0046 traverse (Fig. 12i), the dryline had continued moving westward while cumulus activity had largely dissipated.
Soundings from Amarillo, Texas (AMA), and NSSL-4 were obtained roughly 30 km west and east of the dryline, respectively. The 2300 AMA sounding (Fig. 13a) reveals a warm, dry, deeply well-mixed boundary layer west of the dryline, while the 2250 NSSL-4 sounding (Fig. 13b) shows the relatively cool, moist conditions east of the dryline. The NSSL-4 sounding displays a very strong temperature inversion at p ∼ 590 mb, and a balloon rise rate of ∼3–4 m s−1 in the elevated inversion layer (not shown) implies mesoscale downdrafts stronger than −1 m s−1 (refer to discussion in section 2). The thermal lapse rates from the NSSL-4 sounding also suggest pronounced subsidence in the free atmosphere from about 550 to 640 mb. The strong subsidence revealed by the P-3 analyses after 2322 (Figs. 12d,i) is consistent with the descending motions inferred from the suppressed balloon rise rate in the elevated inversion layer. The inferred moist adiabat of unmixed cumulus and towering cumulus clouds are generally cooler than the environmental temperature profile, with deeper tops being restricted by the elevated inversion layer (Fig. 13b). Bluestein et al. (1987) present a dryline sounding (their Fig. 2b) that is similar in many respects to our NSSL-4 sounding, including a shallow CAPE-bearing layer above the CBL that was capped by an elevated layer of high CIN and inferred subsidence.
An east–west traverse of the dryline was performed by the Probe-3 mobile mesonet during midafternoon about 15 km north of a low-level traverse flown by the P-3. The shapes of the horizontal profiles of absolute humidity from Probe-3 [5 g kg−1 (4 km)−1] and the P-3 are very similar despite slight differences of north–south position and timing of the traverses and a known hysteresis of the chilled mirror dewpoint sensor on the P-37 (Fig. 14). Horizontal moisture gradients of such extreme magnitudes have been reported in classic studies (e.g., NSSP Staff 1963; Schaefer 1973). The corresponding θυ profiles suggest the existence of a local temperature maximum around the dryline, with cooler temperatures to the east of the dryline. As in the case of the 7 June 1994 traverses (Fig. 10), the surface horizontal θυ difference at the location of the peak moisture gradient on 6 May is less than half the value of the corresponding gradient aloft at ∼150 m AGL, emphasizing that surface observations may not be very representative of the boundary layer state. The surface θυ difference across the dryline, 0.5°C (2 km)−1, is consistent with the largest surface θυ difference value reported by Crawford and Bluestein (1997).
4. Discussion
a. Cumulus and storm development relative to boundary layer circulations
The dryline is collocated with a mesoscale updraft in the (western) rising branch of a secondary circulation, while additional roll-like secondary circulations are usually present on either side of the dryline (e.g., Atkins et al. 1998). Most clouds were detected within north–south distances of 10 km of the P-3 traverses on both 15 May 1991 (Fig. 1c) and 6 May 1995 (Fig. 11c), with somewhat greater spacings in the north–south direction on 7 June 1994 (Fig. 6c). Since the inferred shapes of both the 8 June 1974 dryline (Koch and McCarthy 1982) and the 6 May 1995 dryline exhibited a wavelike character with a wavelength of about 30 km and an amplitude of less than 5 km, and since most of the observed clouds were concentrated over distances less than this wavelength, the east–west positioning errors of individual clouds relative to circulation bands is probably on the order of a few kilometers. Assuming two-dimensional boundary layer structures in the east–west direction (i.e., no meridional variation), the cross-sectional analyses suggest a close correspondence between individual clouds and regions of mesoscale lifting (Figs. 2, 7, and 12).
The photogrammetrically measured cumulus clouds were counted within 10-km-wide bins relative to the surface dryline location for each of the case days, with results summarized in Fig. 15. On each of the three days, cumulus cloud frequency achieves a maximum value within 25 km of the dryline, while clouds tend to concentrate in the range −10 ⩽ x ⩽ 40 km. Although samples on the individual case days do not readily conform, the distribution composed of the sum of the three case days is approximately Gaussian in shape. Following the Central Limit theorem of statistics (e.g., Lapin 1975), grouped samples taken from a population (e.g., all cumulus clouds along drylines) will converge toward a Gaussian distribution whose peak approximates the population mean increasingly well as the total sample size increases. With cautious interpretation due to the rather small cloud sample size, our combined cloud frequency data suggest that the peak frequency of the cloud population should be located around approximately x = 15 km east of the dryline. This finding is consistent with Rhea (1966), who documented the strong tendency of storms to initiate very close to the dryline.
The presence of strong mesoscale moisture convergence at the surface is an important factor for increasing the likelihood of convective initiation. Indeed, profiles of horizontally averaged moisture convergence and vertical motion over the analysis cross sections reveal high values of surface moisture convergence and large near-surface vertical updraft gradients (Fig. 16). On 15 May 1991, deep layer maximum updrafts exceeding 0.3 m s−1 are 50% more intense than on 7 June 1994 and triple the peak magnitudes on 6 May 1995. Moisture convergence displays very large vertical variations characterized by an exponentially decreasing magnitude above a nearly constant boundary layer of varying depth, with the deepest and most intense moisture convergence on 15 May 1991. The presence of the relatively deep and intense boundary layer circulations on 15 May 1991 produced the most effective forcing of deep convection of the three case days. It is the presence of strong deep layer convergence in the present cases, rather than strong surface convergence alone, that appears to be the more effective predictor of deep convective initiation. Our data are consistent with the notion that organized mesoscale updrafts are required to lift parcels through their LCL and LFC to initiate deep, moist dryline convection. Since Rhea (1966) documented that storms also organize and intensify close to the dryline, it is reasonable to speculate that mesoscale lifting along the dryline is important for both the initiation and the growth of storms.
b. Cumulus and storm development relative to parcel stability
It is usually assumed that deep convection will develop either along or in the vicinity of strong mesoscale boundaries in unstable environments if CIN decreases to zero. Another common assumption, a variant on the first, is that convection will be initiated if CIN becomes smaller than the kinetic energy of the mesoscale updraft that provides the forcing. From the forecasting perspective, it is problematic that the “proximity” soundings used to gauge the level of instability are typically no closer than a few tens to a few hundreds of kilometers from the area of boundary layer convergence, thus far enough removed from the area of highest convective potential to be unrepresentative. Moreover, the boundary layer updraft is in large measure an externally forced secondary mesoscale circulation, rather than being a turbulent thermal “bubble” of air possessing local buoyancy or high initial vertical momentum that may be invoked to locally “break the capping inversion” or exceed the residual CIN.
To illustrate the notion that vanishing CIN increases the likelihood of convective inititation, Colby (1984) infers that measured CIN values must decrease substantially below 16 J kg−1 by afternoon heating for initiation of deep convection in boundary layer updrafts of several meters per second magnitude. On the other hand, Rhea (1966) qualifies the prediction of convective initiation using CIN, noting that, “no existing stable layer was found to be suppressing thunderstorm development . . . in the first 50 mi to the east of the . . . dryline location,” despite the overwhelming tendency of deep convection to develop along the dryline. Conversely, the 7 June 1994 and 6 May 1995 NSSL-4 soundings (Figs. 8b and 13b, respectively) possess zero CIN,8 yet only very shallow convection develops along those drylines where the secondary circulations increase the strength and depth of boundary layer lifting during the late afternoon. Furthermore, the CIN on 15 May 1991 is ∼100 J kg−1 around 1900 when cumuli begin to form, is as low as ∼19 J kg−1 during the late afternoon as swelling and towering cumuli become more widespread, and more than doubles to over 40 J kg−1 by 0100 before the development of tornadic deep convection (Figs. 3b, 4, and 5). Note that although CIN should be zero in areas experiencing deep convection (implying a necessary condition for the occurrence of deep convection), our results are consistent with the findings of Bluestein et al. (1987) and Mueller et al. (1993) in suggesting that vanishing CIN is not a sufficient condition for initiating storms.
Ziegler et al. (1997) conducted mesoscale model simulations of drylines and dryline convection, including the 15 May 1991 case, showing that intense moisture convergence could locally reduce CIN to zero as mesoscale updrafts lifted moist air through the LCL and LFC and initiated deep convection. Crook and Moncrieff (1988) had previously shown that deep moist convection was initiated along convergence lines as CIN was reduced to zero. The secondary dryline circulation is solenoidally forced, with an active frontogenesis process leading to enhanced thermal and airflow gradients and a well-defined mesoscale updraft along the dryline (Ziegler and Hane 1993; Ziegler et al. 1995). The mesoscale updraft along the dryline achieves a peak value around the top of the moist layer, and the updraft vanishes around the level of the deep CBL, which projects over the surface dryline location from the west. Since the depth of the thermal solenoid field is around the top of the moist layer, as judged by the extent of the low θυ air east of the dryline [Figs. 2, 7, 10, 12, and 14; see also Ziegler and Hane (1993) and Hane et al. (1997)], the height and intensity of the maximum updraft scales with the moist boundary layer depth.
The simulated soundings where deep convection forms feature a nearly constant absolute humidity below cloud base and a saturated, absolutely unstable layer between the base and top of a developing cumulus (Ziegler et al. 1997). While the 7 June 1994 and 6 May 1995 NSSL-4 soundings possess zero CIN due to high boundary layer absolute humidity, as the simulated“cloud initiation” soundings, they also possess rather warm, dry, and stable layers above the moist boundary layer that inhibit the formation and deepening of cumulus clouds. We will explore the apparent role of insufficiently deep and strong mesoscale lifting on convective inhibition in section 4c.
The photogrammetric analyses suggest that a systematic lowering of the cloud-base height occurs from west to east across the dryline. Cumuli are typically high based to the west of the dryline due to lifting by thermals or organized secondary circulations and the nearly homogeneously mixed conditions, while cloud bases are progressively lower with increasing distance east of the dryline (Figs. 2, 7, and 12). The gradual lowering of observed cloud bases are consistent with increasing values of condensation pressure p* (Betts and Ball 1995) in response to the progressively more moist and virtually cooler conditions east of the dryline.
A rising air parcel in the convective boundary layer would need to remain unmixed with drier environmental air to achieve its LCL and LFC as predicted by parcel theory. Such a restriction on mixing would limit warming and drying of the parcel, which in turn would limit increases of the LCL and LFC values for the parcel following the motion.9 In the event that mesoscale lifting was sufficient to achieve parcel saturation, thus forming a cumulus cloud, but insufficiently deep to lift the air parcel through its LFC, mixing and its effect on parcel buoyancy forcing would become important. Inspection of the individual adiabatic cumulus cloud profiles (Figs. 3b, 8b,c, and 13b) suggests the following: 1) The lower levels of all cumuli are convectively stable and hence “forced,” while the tops of many of the cumuli are convectively unstable (“active”) as based herein on the assumption of moist-adiabatic ascent to graphically depict individual in-cloud temperature profiles; 2) mixing between cumuli and their environment, which internally cools the cloud by evaporation and forces the within-cloud lapse rates substantially to the left of the moist parcel adiabat, is suppressing deep growth by forcing cloudy air parcels toward an equilibrium state with the local environment. The side-looking camera images (Figs. 2, 7, and 12) appear to confirm these sounding-based inferences of cumulus dynamics, in that they reveal convective turret structures at the tops of a few cumuli with other cumuli having smoother cloud-top shapes, suggesting the presence or absence, respectively, of thermally buoyant upward accelerations.
c. Nowcasting the destabilization of proximity dryline soundings
The application of parcel theory to determine convective instability of a sounding adopts the simplest possible Lagrangian approach in which the naturally occurring parcel displacements are neglected, and it is implicitly assumed a priori that the airflow needed to force air parcels through the LCL and LFC exists. A common finding of dryline analyses to date, including the present study, is that the secondary circulation at the dryline may be shallow and therefore incapable of lifting boundary layer air through the LCL and LFC. A similar restriction on convective initiation has been demonstrated from field observations of boundary layer airflow in cloud-prone regions in studies of Florida sea breezes (Atkins et al. 1995; Fankhauser et al. 1995; Kingsmill 1995; Weckwerth et al. 1996; Wilson and Megenhardt 1997) and convergence boundaries in northeast Colorado (Wilson et al. 1992; Crook 1996). Recent mesoscale model simulations that explicitly resolved large cumulus clouds and storms demonstrate the need for the joint occurrence of deep vertical displacements and sufficiently large p* (e.g., high values of low-level absolute humidity) for convective initiation along the dryline (Ziegler et al. 1997).
Mesoscale lifting along the dryline produces moistening and virtual cooling at and above the top of the moist boundary layer, thereby enhancing the horizontal gradients of virtual temperature and absolute humidity. The intensification of thermal gradients increases the rate of eastward advection of the hot, dry boundary layer, which in turn opposes the vertical advective cooling and moistening caused by the mesoscale updraft. Typical widths of drylines and their updraft regions are in the 1–10-km range, while the updraft depth is on the order of the deep, often well-mixed convective boundary layer to the west of the dryline [Fig. 17; see also Fig. 14 of Ziegler and Hane (1993)]. Since westerly vertical wind shear is commonly present, owing mainly to larger-scale baroclinicity or local solenoidal effects or some combination, the lower levels of the mesoscale updraft ingest and lift moist boundary layer air from the east as a combination of drier air from the west and moist air from below are processed through the upper levels (see also Fig. 9 of Ziegler et al. 1995). Since the u component achieves a magnitude of U ∼ 10 m s−1 while the maximum updraft is of order 1 m s−1 at the dryline location, individual air trajectories are rather steeply inclined toward the east in upper levels. From around the top of the moist layer down toward the surface, the horizontal wind typically becomes southerly or southeasterly as the u component changes sign and U becomes small compared to values either west of the dryline or aloft. Air parcels have residence times in the mesoscale updraft on the order of 10 min, imposing the strong temporal constraint that the LCL and LFC need to be achieved before the air parcel exits the zone of lifting.
To assess the relative roles of the horizontal and vertical transports of boundary layer air for cloud formation, the proximity soundings in the various cases were modified using our simple conceptual model of the dryline environment (Fig. 17) and a prognostic kinematic numerical approach described in appendix B. The kinematic model is based on conservation equations of potential temperature and water vapor, considering horizontal and vertical advection by the mesoscale flow while neglecting turbulent vertical mixing due to unresolved scales of motion (i.e., length scales less than 600 m). It is hypothesized that the vertical fluxes due to the mesoscale updraft (i.e., “mesoscale fluxes”) dominate over the turbulent fluxes associated with smaller scales of motion (Raymond and Wilkening 1980; Pielke et al. 1991), under which circumstances the small-scale chaotic, random parcel displacements from turbulence may be neglected compared to the systematic mesoscale lifting. In addition, the kinematic model is initialized by observed soundings, which implicitly carry the full effects of turbulent vertical mixing in the CBL. By considering both the thermal gradients and the relative strengths of the horizontal wind and the mesoscale updraft through the boundary layer, the ability of an air parcel to become water saturated before exiting the east side of the mesoscale updraft can be assessed.
Results of the sounding modifications with the simple kinematic model, along with key kinematic parameter values for each case day, are presented in Fig. 18. In the following discussion we define hLCL as the height of the LCL (referred to as CBZ in Fig. 18), while hwmax is defined in appendix B. On 15 May 1991 the dimensionless ratio Hlcl ≡ hwmax/hLCL = 2.2 (km AGL)/[2.0 (km MSL) − 0.7] ≈ 1.7 (Fig. 18a), while Hlcl = 1.0/0.8 ≈ 1.3 on 6 May 1995 (Figs. 18c). Since Hlcl > 1 in both the 15 May 1991 and 6 May 1995 cases, horizontal advection does not significantly warm and dry the mesoscale updraft until lifted air has attained its LCL. In other words, air parcels are likely to achieve the LCL provided that Hlcl significantly exceeds unity. The modified 7 June 1994 sounding (Fig. 18b), which shows excellent agreement with the 2143 NSSL-4 sounding released at the dryline, does not achieve saturation due to the drying effects of the relatively deep westerly wind-bearing layer in that case. Since Hlcl = 1.5/2.8 ≈ 0.54 on 7 June 1994, the LCL is not attained before a lifted air parcel detrains from the mesoscale updraft.
If lifting through a layer deeper than the LCL continued beyond the onset of saturation, deep convection would be triggered by the formation of an absolutely convectively unstable layer with its base around the LCL (Crook and Moncrieff 1988; Ziegler et al. 1997). Following the same reasoning as used to develop and apply the quantity Hlcl, a second useful dimensionless ratio is Hlfc ≡ hwmax/hLFC, where hLFC is the height of the LFC. For example, the estimated values of Hlfc based on the modified soundings in Fig. 18 are as follows: 15 May 1991 (2.2/1.3 ≈ 1.7 from Fig. 18a), 7 June 1994 (1.5/3.4 ≈ 0.44), and 6 May 1995 (1.0/3.5 ≈ 0.29). Considering the magnitude of the lapse rate in the layer overlying the predicted LCL and the likelihood that cumulus circulations are mixed with environmental air, the modified 15 May 1991 sounding with both Hlcl > 1 and Hlfc > 1 would have the highest probability of initiating deep convection of the three cases considered.
Local circulations could support cumulus formation even though the larger mesoscale environmental flow does not favor the attainment of saturation. An example is the clear air Doppler radar analysis of the 6 May 1995 dryline by Atkins et al. (1998), who found that localized updrafts having deeper vertical circulations than along the segments of dryline between updraft cores sometimes initiated cumulus clouds. If the local circulation strengthened and became more erect, thereby decreasing the westerly horizontal influx of dry air, a lifted parcel would spend more time in the mesoscale updraft and its peak value of relative humidity prior to detrainment would increase. With weakening of the horizontal inflow below some threshold value, saturation could be achieved. For example, setting U to zero in the 7 June 1994 case (Fig. 18d) yields a predicted LCL that is consistent with the observed cloud base of 3.7 km (i.e., cloud above “subgrid-scale moist plume” in Fig. 7a). Note that setting U to zero in the kinematic model is not equivalent to parcel theory, since a layer of air of finite depth is being lifted for a finite time period through an updraft profile. It is concluded that initiation of forced or active cumulus convection requires that the magnitude of the horizontal flux of dry air from the west be locally negligible in relation to the vertical flux of moist air in the mesoscale updraft below the LCL or LFC, respectively.
d. Nowcasting the development of deep convergence along drylines
To apply the sounding modification technique described in section 4c, a knowledge of the depth and intensity of boundary layer convergence near the dryline is required. Although the suite of mesoscale observations available to forecasters does not resolve small-scale features such as drylines explicitly, it is possible to infer a range of likely divergence profiles at the dryline from wind profile measurements representing the“eastern” and “western” soundings in Fig. 17. Recalling the simple sounding modification technique in appendix B and neglecting variations along the dryline compared to across-dryline gradients, key environmental parameters include the depths of the western and eastern boundary layers (AGL), the profile of the difference of the normal component of the low-level horizontal wind across the dryline (i.e., must consider local dryline orientation), and the width of the dryline.
One method of estimating local conditions on either side of the dryline is to modify appropriate regional morning soundings to reflect afternoon surface conditions (e.g., McGinley 1986). One example of this sounding modification technique for the dryline environment is the classic study of Rhea (1966). A second useful information source for estimating boundary layer profiles is from output of the National Centers for Environmental Prediction’s (NCEP) Rapid Update Cycle model (Benjamin et al. 1998). Regional boundary layer wind measurements from WSR-88D velocity azimuth display (VAD) scans and the NOAA Wind Profiler Demonstration Network would be useful in gauging the evolution of the depth and intensity of secondary circulations that produce convergence at the dryline. Since neither mesoscale observations nor operational model forecasts can resolve the dryline itself (Ziegler et al. 1997), it is necessary to consider varying dryline widths in the range of order 1–10 km. Applying the estimated normal velocity component profiles over an assumed dryline width and vertically integrating the resulting profile of the normal divergence component, the vertical motion profile at the dryline can then be estimated. In principle the forecaster can then take the following actions: 1) modify the eastern proximity sounding following the method outlined in appendix B; 2) compute Hlcl and Hlfc (as described in section 4c). This approach should be sufficient to estimate the likelihood of shallow or deep convective initiation, subject to the representativeness of the estimated eastern and western soundings and the assumed dryline width.
5. Conclusions
The processes that force the initiation of deep convection along the dryline have been inferred from special mesoscale observations obtained during the COPS-91, VORTEX-94, and VORTEX-95 field projects. Observations from aircraft, mobile CLASS soundings, and mobile mesonets define the fields of airflow, absolute humidity, and virtual temperature in the boundary layer across the dryline on the 15 May 1991, 7 June 1994, and 6 May 1995 case days. Film and video cloud images obtained by time-lapse cameras on the P-3 are used to reconstruct the mesoscale distribution of cumulus by photogrammetric methods, allowing inferences concerning the environmental conditions accompanying cloud formation or suppression.
The results of the present study are consistent with the classic notion that the dryline is a favored zone for cumulus cloud formation. The combined cloud distributions for the three cases examined approximate a Gaussian shape, suggesting a peak cloud frequency 15 km east of the dryline based on the Central Limit theorem. Cumuli were concentrated within the interval from 10 km west to 40 km east of the dryline. Intense, deep mesoscale moisture convergence is inferred in cloudy regions, with mesoscale subsidence or a lack of deep vertical motion in cloud-free regions. Our results document the modulating effect of westerly wind shear on convective initiation in the mesoscale updrafts at the dryline, suggesting that moist boundary layer air parcels must be lifted to their LCL and LFC prior to leaving the mesoscale updraft to form deep convection. Clouds are high based to the west of the dryline, and bases progressively lower with distance east of the dryline as higher humidities and cooler temperatures promote lower LCLs in the boundary layer.
Nowcasting convective initiation from proximity environmental soundings according to conventional practice must be strongly qualified due to the assumptions of deep lifting and lack of mixing invoked by parcel theory. On two of the case days, 7 June 1994 and 6 May 1995, deep moist convection did not develop despite a proximity sounding with a zero value of CIN (note: stability parameter value as computed using virtual temperature and lifting the lowest 50-mb average parcel). Since vertical circulations must be deep and strong enough to initiate clouds, the presence of strong surface convergence is also not an unambiguous predictor for convective initiation. For example, surface convergence values during the late afternoon on 6 May 1995 were comparably strong relative to the other cases, yet weak, shallow mesoscale updrafts resulted from a very shallow convergence layer just above the surface capped by divergence in higher levels. Though mesoscale lifting was occasionally just deep enough to attain the LCL in each of the cases, the mesoscale updrafts were not deep enough to lift moist boundary layer air through the LFC and thereby initiate storms. In all cases, it was inferred that mixing between the cumulus cloud and its environment was diluting the buoyancy of the cloudy updraft and suppressing deep growth. The presence of intense, deep mesoscale lifting was inferred to be necessary to overcome the retarding effects of mixing on storm initiation.
To assess the relative roles of the horizontal and vertical transports of boundary layer air for cloud formation, the proximity soundings in the various cases were modified using a simple conceptual model of the dryline environment and a prognostic kinematic numerical scheme. Using a simplified set of conservation equations for heat and water vapor composed of horizontal and vertical advection terms, the proximity soundings were modified according to the measured mean mesoscale updraft and horizontal wind profiles at the dryline. The technique was directly verified by successfully approximating the observed dryline sounding in the 7 June 1994 case. Cloud formation is predicted when the vertical mesoscale moisture flux predominates below the LCL, and deep convection is predicted if strong mesoscale lifting is deeper than the LFC. Equivalently, cloud or storm formation is predicted if the dimensionless ratio of the maximum updraft height to the LCL or LFC height is greater than unity, respectively. It is suggested that deep convection is likely to follow the joint attainment of the LCL and LFC if the lapse rate of the top of the lifted portion of the sounding is significantly larger than the predicted moist adiabatic value. Results generally suggest that a modification of proximity soundings to account for mesoscale lift, the across-dryline differences of environmental thermal stratification, and westerly wind shear effects can improve the diagnosis of the mesoscale dryline environment and the prediction of convective initiation at the dryline.
Our conclusions about the nature of the convective initiation process are necessarily limited by the lack of spatially detailed measurements of boundary layer airflow, moisture, and temperature around the time and in the location where storms ultimately form. For example, we have speculated that locally deep, intense mesoscale convergence could conceivably initiate storms at isolated locations along the dryline in a given case, even though larger-scale conditions are hostile to more widespread convective initiation. The depth and amount of precipitable boundary layer water vapor is another critically important, but poorly resolved, quantity required to forecast convective initiation. To address such observational deficiencies near a variety of (slowly moving) boundaries including stationary fronts, warm fronts, decayed thunderstorm outflow boundaries, and drylines, the authors have recently joined with NSSL colleague J. Schneider and other scientists to propose a “Thunderstorm Initiation Mobile Experiment” (TIMEx). Community discussions of the proposed field study are taking place via planning meetings and an interactive World Wide Web site (http://www.nssl.noaa.gov/srad/timex). To assist the NWS, we intend for TIMEx to provide a unique set of observations that provide the basis for conceptual models of the convective initiation process and help improve the accuracy and specificity of storm forecasts.
Acknowledgments
The staff of the NOAA Aircraft Operations Center capably operated the P-3 aircraft and its data systems during COPS-91 and VORTEX. We are indebted to the numerous COPS and VORTEX scientists and volunteers whose dedication and skill in planning and executing the scientific aspects of the P-3 missions and operating the M-CLASS sounding systems and mobile mesonets produced a very useful dataset. Discussions with Ronald Holle were critical to the successful application of photogrammetric analysis in this study. Reviews of early versions of the manuscript by Robert Maddox and David Schultz and the comments offered by three anonymous reviewers were very helpful. Discussions with Robert Maddox and Jeff Trapp motivated the application of the Barnes scheme to the objective analysis of P-3 data. Daniel Geiszler analyzed the mobile mesonet data for the 7 June 1994 case under the mentorship of the lead author (CLZ) for his special project in the 1995 Research Experience for Undergraduates (REU) program at the Oklahoma Weather Center. We gratefully acknowledge the skillful and tireless efforts of John Cortinas, William Beasley, Jeanne Schneider, Jerry Straka, and Cindy Machacek for their work to coordinate the REU program, Sonia Lasher-Trapp offered assistance with access to the 7 June 1994 mobile mesonet data. Funding and support for VORTEX field operations were provided by NOAA and the National Severe Storms Laboratory, the Center for the Analysis and Prediction of Storms at the University of Oklahoma (OU) under Grant NSF ATM 912-0009, and the OU Graduate College. The National Science Foundation also provided major funding for the REU program.
REFERENCES
Atkins, N. T., R. M. Wakimoto, and T. M. Weckwerth, 1995: Observations of the sea-breeze front during CaPE. Part II: Dual-Doppler and aircraft analysis. Mon. Wea. Rev.,123, 944–969.
——, R. M. Wakimoto, and C. L. Ziegler, 1998: Observations of the fine-scale structure of a dryline during VORTEX 95. Mon. Wea. Rev.,126, 525–550.
Banta, R. M., 1984: Daytime boundary-layer evolution over mountainous terrain. Part I: Observations of the dry circulations. Mon. Wea. Rev.,112, 340–356.
Barnes, S. L., 1964: A technique for maximum detail in numerical weather map analysis. J. Appl. Meteor.,3, 396–409.
Benjamin, S. G., J. M. Brown, K. J. Brundage, B. Schwartz, T. Smirnova, T. L. Smith, L. L. Morone, and G. J. DiMego, 1998: The operational RUC-2. Preprints, 16th Conf. on Weather Analysis and Forecasting, Phoenix, AZ, Amer. Meteor. Soc., 249–252.
Betts, A. K., 1984: Boundary layer thermodynamics of a High Plains severe storm. Mon. Wea. Rev.,112, 2199–2211.
——, and J. H. Ball, 1995: The FIFE surface diurnal cycle climate. J. Geophys. Res.,100, 25 679–25 693.
Bluestein, H. B., and S. S. Parker, 1993: Modes of isolated, severe convective storm formation along the dryline. Mon. Wea. Rev.,121, 1354–1372.
——, E. W. McCaul Jr., G. P. Byrd, R. L. Walko, and G. R. Woodall 1987: Forecasting and nowcasting cumulus convection with soundings released from a storm-intercept vehicle. Proc. Symp. on Mesoscale Analysis and Forecasting, Vancouver, BC, Canada, International Association of Meteorology and Atmospheric Physics, ESA SP-282, 135–139.
——, ——, and ——, 1990: An observational study of splitting convective clouds. Mon. Wea. Rev.,118, 1359–1370.
Colby, F. P., Jr., 1984: Convective inhibition as a predictor of convection during AVE-SESAME II. Mon. Wea. Rev.,112, 2239–2252.
Crawford, T. M., and H. B. Bluestein, 1997: Characteristics of dryline passage during COPS-91. Mon. Wea. Rev.,125, 463–477.
Crook, N. A., 1996: Sensitivity of moist convection forced by boundary layer processes to low-level thermodynamic fields. Mon. Wea. Rev.,124, 1767–1785.
——, and M. W. Moncrieff, 1988: The effect of large-scale convergence on the generation and maintenance of deep moist convection. J. Atmos. Sci.,45, 3606–3624.
Doswell, C. A., III, and E. N. Rasmussen, 1994: The effect of neglecting the virtual temperature correction in CAPE calculations. Wea. Forecasting,9, 625–629.
Emanuel, K., and Coauthors, 1995: Report of the First Prospectus Development Team of the U.S. Weather Research Program to NOAA and the NSF. Bull. Amer. Meteor. Soc.,76, 1194–1208.
Eymard, L., 1984: Radar analysis of a tropical convective boundary layer with shallow cumulus clouds. J. Atmos. Sci.,41, 1380–1393.
Fankhauser, J. C., N. A. Crook, J. Tuttle, L. J. Miller, and C. G. Wade, 1995: Initiation of deep convection along boundary layer convergence lines in a semitropical environment. Mon. Wea. Rev.,123, 291–313.
Hane, C. E., C. L. Ziegler, and H. B. Bluestein, 1993: Investigation of the dryline and convective storms initiated along the dryline. Bull. Amer. Meteor. Soc.,74, 2133–2145.
——, H. B. Bluestein, T. M. Crawford, M. E. Baldwin, and R. M. Rabin, 1997: Severe thunderstorm development in relation to along-dryline variability: A case study. Mon. Wea. Rev.,125, 231–251.
Holle, R. L., 1988: Photogrammetry of thunderstorms. Instruments and Techniques for Thunderstorm Observation and Analysis, Vol. 3, Thunderstorms: A Social, Scientific, and Technological Documentary. 2d ed. E. Kessler, Ed., University of Oklahoma Press, 51–63.
Johns, R. H., and C. A. Doswell III, 1992: Severe local storms forecasting. Wea. Forecasting,7, 588–612.
Kingsmill, D. E., 1995: Convection initiation associated with a sea-breeze front, a gust front, and their collision. Mon. Wea. Rev.,123, 2913–2933.
Koch, S. E., and J. McCarthy, 1982: The evolution of an Oklahoma dryline. Part II: Boundary-layer forcing of mesoconvective systems. J. Atmos. Sci.,39, 237–257.
——, and C. A. Ray, 1997: Mesoanalysis of summertime convergence zones in central and eastern North Carolina. Wea. Forecasting,12, 56–77.
——, M. DesJardins, and P. J. Kocin, 1983: An interactive Barnes objective map analysis scheme for use with satellite and conventional data. J. Climate Appl. Meteor.,22, 1487–1503.
Lapin, L. L., 1975: Statistics: Meaning and Method. Harcourt Brace Jovanovich, 591 pp.
Lauritzen, D., Z. Malekmadani, C. Morel, and R. McBeth, 1987: The Cross-chain LORAN Atmospheric Sounding System (CLASS). Preprints, Sixth Symp. Meteorological Observations and Instrumentation, New Orleans, LA, Amer. Meteor. Soc., 340–343.
McGinley, J., 1986: Nowcasting mesoscale phenomena. Mesoscale Meteorology and Forecasting, P. Ray, Ed., Amer. Meteor. Soc., 657–688.
McNulty, R. P., 1995: Severe and convective weather: A central region forecasting challenge. Wea. Forecasting,10, 187–202.
Mohr, C. G., L. J. Miller, R. L. Vaughan, and H. W. Frank, 1986: The merger of mesoscale datasets into a common Cartesian format for efficient and systematic analysis. J. Atmos. Oceanic Technol.,3, 143–161.
Mueller, C. K., J. W. Wilson, and N. A. Crook, 1993: The utility of sounding and mesonet data to nowcast thunderstorm initiation. Wea. Forecasting,8, 132–146.
NSSP Staff, 1963: Environmental and thunderstorm structures as shown by National Severe Storms Project observations in spring 1960 and 1961. Mon. Wea. Rev.,91, 271–292.
Parsons, D. B., M. A. Shapiro, R. M. Hardesty, R. J. Zamora, and J. M. Intrieri, 1991: The finescale structure of a West Texas dryline. Mon. Wea. Rev.,119, 1242–1258.
Pielke, R. A., G. A. Dalu, J. S. Snook, T. J. Lee, and T. G. F. Kittel, 1991: Nonlinear influence of mesoscale land use on weather and climate. J. Climate,4, 1053–1069.
——, T. J. Lee, J. H. Copeland, J. L. Eastman, C. L. Ziegler, and C. A. Finley, 1997: Use of USGS-provided data to improve weather and climate simulations. Ecol. Appl.,7, 3–21.
Pruppacher, H. R., and J. D. Klett, 1978: Microphysics of Clouds and Precipitation. D. Reidel, 714 pp.
Purdom, J. F. W., 1982: Subjective interpretation of geostationary satellite data for nowcasting. Nowcasting, K. Browning, Ed., Academic Press, 149–162.
Rasmussen, E. N., J. M. Straka, R. Davies-Jones, C. A. Doswell III, F. H. Carr, M. D. Eilts, and D. R. MacGorman, 1994: Verification of the Origins of Rotation in Tornadoes Experiment: VORTEX. Bull. Amer. Meteor. Soc.,75, 995–1006.
Raymond, D. J., and M. H. Wilkening, 1980: Mountain-induced convection under fair weather conditions. J. Atmos. Sci.,37, 2693–2706.
——, and ——, 1982: Flow and mixing in New Mexico mountain cumuli. J. Atmos. Sci.,39, 2211–2228.
Rennó, N. O., and E. R. Williams, 1995: Quasi-Lagrangian measurements in convective boundary layer plumes and their implications for the calculation of CAPE. Mon. Wea. Rev.,123, 2733–2742.
Rhea, J. O., 1966: A study of thunderstorm formation along drylines. J. Appl. Meteor.,5, 58–63.
Rust, W. D., R. P. Davies-Jones, D. W. Burgess, R. A. Maddox, L. C. Showell, T. C. Marshall, and D. K. Lauritsen, 1990: Testing a mobile version of a Cross-chain Loran Atmospheric Sounding System (M-CLASS). Bull. Amer. Meteor. Soc.,71, 173–180.
Schaefer, J. T., 1973: The motion and morphology of the dryline. NOAA Tech. Memo ERL NSSL-66, 81 pp. [NTIS COM-74-10043.].
Shaw, B. L., R. A. Pielke, and C. L. Ziegler, 1997: A three-dimensional numerical simulation of a Great Plains dryline. Mon. Wea. Rev.,125, 1489–1506.
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.
Stull, R. B., 1985: A fair-weather cumulus cloud classification scheme for mixed-layer studies. J. Climate Appl. Meteor.,24, 49–56.
USDOC, 1991: Storm Data. Vol. 33, No. 5, Department of Commerce, 260 pp.
Wakimoto, R. M., and N. T. Atkins, 1994: Observations of the sea-breeze front during CaPE. Part I: Single-Doppler, satellite, and cloud photogrammetric analysis. Mon. Wea. Rev.,122, 1092–1114.
——, W.-C. Lee, H. B. Bluestein, C.-H. Liu, and P. H. Hildebrand, 1996: ELDORA observations during VORTEX 95. Bull. Amer. Meteor. Soc.,77, 1465–1481.
Weaver, J. F., J. F. W. Purdom, and E. J. Szoke, 1994: Some mesoscale aspects of the 6 June 1990 Limon, Colorado, Tornado case. Wea. Forecasting,9, 45–61.
Weckwerth, T. M., J. W. Wilson, and R. M. Wakimoto, 1996: Thermodynamic variability within the convective boundary layer due to horizontal convective rolls. Mon. Wea. Rev.,124, 769–784.
Wilson, J. W., and W. E. Schreiber, 1986: Initiation of convective storms at radar-observed boundary-layer convergence lines. Mon. Wea. Rev.,114, 2516–2536.
——, and D. L. Megenhardt, 1997: Thunderstorm initiation, organization, and lifetime associated with Florida boundary layer convergence lines. Mon. Wea. Rev.,125, 1507–1525.
——, G. B. Foote, N. A. Crook, J. C. Fankhauser, C. G. Wade, J. D. Tuttle, and C. K. Mueller, 1992: The role of boundary layer convergence zones and horizontal rolls in the initiation of thunderstorms: A case study. Mon. Wea. Rev.,120, 1785–1815.
——, T. M. Weckwerth, J. Vivekanandan, R. M. Wakimoto, and R. W. Russell, 1994: Boundary layer clear-air radar echoes: Origin of echoes and accuracy of derived winds. J. Atmos. Oceanic Technol.,11, 1184–1206.
Ziegler, C. L., and C. E. Hane, 1993: An observational study of the dryline. Mon. Wea. Rev.,121, 1134–1151.
——, W. J. Martin, R. A. Pielke, and R. L. Walko, 1995: A modeling study of the dryline. J. Atmos. Sci.,52, 263–285.
——, T. J. Lee, and R. A. Pielke, 1997: Convective initiation at the dryline: A modeling study. Mon. Wea. Rev.,125, 1001–1026.
APPENDIX A
Photogrammetric Cloud Analysis Technique
Photogrammetric cloud analyses were performed following Holle (1988) using the right-side-mounted time-lapse 16-mm color film or video camera systems flown on the P-3 during the COPS and VORTEX projects, respectively. Left-side camera images were inspected but were not utilized to maximize the statistical independence of the individual sampled clouds. Examples of the images used are presented in Figs. 2, 7, and 12 in the text. Nose camera images such as shown in Fig. 7 were also not analyzed. The overall quality of both the film and videotape footage is considerably better than the sample images, which are obtained from scanned video frames (i.e., requiring a film-to-video transfer in the COPS case). The photogrammetric technique is based on known information of the position and absolute east–west speed of the P-3, the measured position of cloud boundaries in the horizontal and vertical image directions, and several known or estimated properties of the camera lens and recording systems. The output of the photogrammetric analysis is the position and range of a vertically and east–west-oriented bounding box that just encloses individual clouds or cloud clusters. In cases of complex cloud fields, only the closest clouds relatively unmasked by intervening cumuli are analyzed.
Interpretation of the film and video camera images required a knowledge of the time of each frame. For the video camera system flown during VORTEX-94 and -95, time (UTC) was manually preset into the P-3 data system to within ±1 s of the WWV10 standard and digitally recorded as a time stamp on each video frame. Since a time stamp was not included on the film from the COPS-91 case, it was necessary to determine beginning and ending times corresponding to reference frames near the beginning and end of the time-lapse movie segment. The determination of Ts and Te, the beginning and ending reference times, respectively, required careful matching of the time of onset of recorded banking maneuvers to visible rapid appearances or disappearances of ground. For Nmax equal to the number of frames from Ts to Te and for time (s), the time between frames R = (Te − Ts)/Nmax = 4.97 s. The time (s) of an arbitrary 16-mm movie frame was computed from T = To + NR, where N is the frame number from the beginning reference time To, followed by conversion toUTC. The Ts and Te values are accurate to within ±R/2 or approximately ±2.5 s, while the error in R is negligible due to the length of the footage and the large number of frames.
A cloud’s horizontal (along track) and vertical positions and the transit time across the image frame are fundamental photogrammetric measurables. The method of crossed diagonals is used to locate the center of a movable measurement grid at the center of the projected or displayed image viewed either on paper, if from film, or through transparency, if viewed via video monitor. After the measuring grid is rotated about the image center to align the horizontal centerline with the horizon, values of the vertical distance from the horizontal line that divides the image in half to the cloud feature, y, can be determined. The time (or frame number) Tc at which a cloud edge crosses the vertical centerline effectively locates the vertical cloud edges according to the known aircraft location. Measurements are obtained only during periods of rather smooth, straight, and level flight, easily detected as a straight and level path of a cloud tag across the grid, and possible deviations of the camera orientation due to minor roll, pitch, or yaw maneuvers of the P-3 are neglected.
The P-3 employed a 10-mm focal length lens on all flights, allowing the determination of other camera parameters using Eq. (A.1). By direct measurement of individual frames, the following 16-mm movie parameters were obtained: 1) dimensions of 10 mm horizontally × 7.7 mm vertically (±0.1 mm); 2) α = 53.1° from Eq. (A.1). In comparison, the Sony DXC151A video camera has a specified imaging element size of 2/3 in. (16.93 mm), implying an 11-mm imaging circle and a horizontal image dimension of 8.8 mm due to the 4/3 aspect ratio of television.
Corrections were determined for the video system to account for apparent differences between the manufacturer-specified and actual (i.e., apparent) video image sizes. Using a 400′:1" aerial photographic survey of the northwest quadrant of Will Rogers World Airport and selected video frames obtained while either on the taxiway or crossing the end of the runway before landing, individual landmarks on both sides of the video frame (e.g., hanger edges, water towers) were located on the aerial survey. After plotting rays corresponding to the viewing angle that emanated from the known camera location, constructing a line normal to one ray that intersected the other ray, and measuring the three sides of the right triangle thus formed, the horizontal effective viewing angle of the video system was computed using trigonometry. Averaging the α values obtained from the sine and cosine relations yielded α = 40.9° ± 0.2°. Inserting this viewing angle estimate into Eq. (A.1) and assuming f = 10 mm, the corresponding horizontal image dimension is d = 7.5 mm.
Since the parameters Tc, α, and y either require hand–eye measurement or are based on manufacturer specifications of uncertain precision, and since these parameters are subject to error, a sensitivity check was performed on the photogrammetric calculations. The error of the east–west cloud position is ±VR/2 (120 m s−1 × 2.5 s) or roughly ± 300 m in the COPS-91 case and ±120 m in the VORTEX cases. Two error levels were considered for α (±0.2°, ±1°) and y (±1 mm, ±2.5 mm) to allow for error sources beyond those due to precision of the measuring tools alone. The α error level for the 16-mm film system should be small compared to that for the video system that is considered in these calculations. For fixed y = 0, rh < 30 km, and α within ±1°, the range error of less than about ±0.8 km is negligible. For fixed α = 40.9° and y within ±2.5 mm, the height error is less than about ±100 m within 15 km and less than ±300 m within 30 km. As an independent check, the heights of clearly discernible ground targets (m MSL) within 5-km range (i.e., y < 0) were computed by the photogrammetric technique and showed excellent agreement with the known altitudes of the target locations (i.e., errors of order ±10 m).
APPENDIX B
Simple Numerical Transport Model for Modifying Input Soundings
Neglecting horizontal gradients of all variables on either side of the dryline, the proximity soundings are used as proxy for conditions along the lateral boundaries of the updraft in the dryline zone (Fig. 17). Either the P-3 measurements or a proximity sounding on the west side of the dryline is assumed to characterize the deep, dry boundary layer. On input of the observed profiles of θ and qυ of the proximity sounding into the model, equivalent to displacing the proximity sounding to the center of the updraft with the low-level inflow (Fig. 17), Eqs. (B.1) and (B.2) are integrated forward in time using upwind (uncentered) space and forward time differences and a time step of 20 s for a period τ not exceeding the advective timescale Lx/Uwest. In other words, integration is stopped at some t < τ if water saturation is achieved. In tests where Uwest is set to zero, τ is set to the vertical advective timescale formed by dividing the updraft depth by the mean updraft value. Small-scale noise generated by the nonregular grid spacing is effectively controlled by the smoothing inherent in the upwind differencing scheme employed by the kinematic model.
Surface and boundary layer observations and cloudiness on 15 May 1991. (a) Surface state at 2100 UTC, with solid-contoured above mean sea level (MSL) pressure (mb) and dashed-contoured dewpoint temperature (°C); (b) visible satellite imagery at 2100 UTC; (c) map representation of photogrammetric estimates of north–south position (i.e., range from a P-3 leg) and east–west extent (i.e., relative to P-3 location) of all clouds or cloud areas (gray line segments), with 2307–2315 UTC P-3 traverse across dryline at ∼150 m above ground level (AGL). Surface station model in (a) includes winds (half barb = 2.5 m s−1, full barb = 5 m s−1), temperature (°C) at upper left, dewpoint temperature (°C) at lower left, and MSL pressure (mb) at upper right. Selected NCAR PAM mesonet sites are denoted by filled squares. The symbols N1, N2, and A in (c) locate the 2311 NSSL-1 sounding, and the NSSL-2 soundings at Shamrock, TX, and Alanreed, TX (the westernmost NSSL-2 sounding site), respectively, while the plotted traverse was chosen to depict conditions between the soundings presented in Fig. 3. The P-3 temperature is about 1.5°C cooler than the local surface temperature assuming a dry-adiabatic lapse rate. The heavy dashed line in (a) denotes a convergence line. Gray filled and open white rectangles in (a) and (b), respectively, depict the location of east–west traverses of the NOAA P-3 aircraft as discussed in the text.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Surface and boundary layer observations and cloudiness on 15 May 1991. (a) Surface state at 2100 UTC, with solid-contoured above mean sea level (MSL) pressure (mb) and dashed-contoured dewpoint temperature (°C); (b) visible satellite imagery at 2100 UTC; (c) map representation of photogrammetric estimates of north–south position (i.e., range from a P-3 leg) and east–west extent (i.e., relative to P-3 location) of all clouds or cloud areas (gray line segments), with 2307–2315 UTC P-3 traverse across dryline at ∼150 m above ground level (AGL). Surface station model in (a) includes winds (half barb = 2.5 m s−1, full barb = 5 m s−1), temperature (°C) at upper left, dewpoint temperature (°C) at lower left, and MSL pressure (mb) at upper right. Selected NCAR PAM mesonet sites are denoted by filled squares. The symbols N1, N2, and A in (c) locate the 2311 NSSL-1 sounding, and the NSSL-2 soundings at Shamrock, TX, and Alanreed, TX (the westernmost NSSL-2 sounding site), respectively, while the plotted traverse was chosen to depict conditions between the soundings presented in Fig. 3. The P-3 temperature is about 1.5°C cooler than the local surface temperature assuming a dry-adiabatic lapse rate. The heavy dashed line in (a) denotes a convergence line. Gray filled and open white rectangles in (a) and (b), respectively, depict the location of east–west traverses of the NOAA P-3 aircraft as discussed in the text.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Surface and boundary layer observations and cloudiness on 15 May 1991. (a) Surface state at 2100 UTC, with solid-contoured above mean sea level (MSL) pressure (mb) and dashed-contoured dewpoint temperature (°C); (b) visible satellite imagery at 2100 UTC; (c) map representation of photogrammetric estimates of north–south position (i.e., range from a P-3 leg) and east–west extent (i.e., relative to P-3 location) of all clouds or cloud areas (gray line segments), with 2307–2315 UTC P-3 traverse across dryline at ∼150 m above ground level (AGL). Surface station model in (a) includes winds (half barb = 2.5 m s−1, full barb = 5 m s−1), temperature (°C) at upper left, dewpoint temperature (°C) at lower left, and MSL pressure (mb) at upper right. Selected NCAR PAM mesonet sites are denoted by filled squares. The symbols N1, N2, and A in (c) locate the 2311 NSSL-1 sounding, and the NSSL-2 soundings at Shamrock, TX, and Alanreed, TX (the westernmost NSSL-2 sounding site), respectively, while the plotted traverse was chosen to depict conditions between the soundings presented in Fig. 3. The P-3 temperature is about 1.5°C cooler than the local surface temperature assuming a dry-adiabatic lapse rate. The heavy dashed line in (a) denotes a convergence line. Gray filled and open white rectangles in (a) and (b), respectively, depict the location of east–west traverses of the NOAA P-3 aircraft as discussed in the text.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Objective analyses of P-3 stepped traverse measurements with cloud observations on 15 May 1991. Analyses (left column): (a) 1532–1646; (b) 1649–1804; (c) 1807–1935; (g) 2014–2208; (h) 2220–2305; (i) 2307–0006. Contours are water vapor mixing ratio (g kg−1), while vector velocity in the cross section is scaled at upper left. Cloud images in the right column are either VIS imagery (time in parentheses) or views from P-3 side-looking cameras obtained at the location indicated by a heavy black dot in the indicated analysis (viewing direction in parentheses): (d) VIS image (1600); (e) VIS image (1800); (f) VIS image (1900); (j) 2122 (south); (k) 2153 (south); (l) 2327 (north); (m) 2348 (north). The margin between white and light gray fill in (a), (b), (c), and (g)–(i) is θv = 308, 309, 311, and 312 K, respectively, while gray fill increases darkness in 1 K increments. The x-distance scale of an unlabeled panel is the scale of the next lower labeled panel.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Objective analyses of P-3 stepped traverse measurements with cloud observations on 15 May 1991. Analyses (left column): (a) 1532–1646; (b) 1649–1804; (c) 1807–1935; (g) 2014–2208; (h) 2220–2305; (i) 2307–0006. Contours are water vapor mixing ratio (g kg−1), while vector velocity in the cross section is scaled at upper left. Cloud images in the right column are either VIS imagery (time in parentheses) or views from P-3 side-looking cameras obtained at the location indicated by a heavy black dot in the indicated analysis (viewing direction in parentheses): (d) VIS image (1600); (e) VIS image (1800); (f) VIS image (1900); (j) 2122 (south); (k) 2153 (south); (l) 2327 (north); (m) 2348 (north). The margin between white and light gray fill in (a), (b), (c), and (g)–(i) is θv = 308, 309, 311, and 312 K, respectively, while gray fill increases darkness in 1 K increments. The x-distance scale of an unlabeled panel is the scale of the next lower labeled panel.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Objective analyses of P-3 stepped traverse measurements with cloud observations on 15 May 1991. Analyses (left column): (a) 1532–1646; (b) 1649–1804; (c) 1807–1935; (g) 2014–2208; (h) 2220–2305; (i) 2307–0006. Contours are water vapor mixing ratio (g kg−1), while vector velocity in the cross section is scaled at upper left. Cloud images in the right column are either VIS imagery (time in parentheses) or views from P-3 side-looking cameras obtained at the location indicated by a heavy black dot in the indicated analysis (viewing direction in parentheses): (d) VIS image (1600); (e) VIS image (1800); (f) VIS image (1900); (j) 2122 (south); (k) 2153 (south); (l) 2327 (north); (m) 2348 (north). The margin between white and light gray fill in (a), (b), (c), and (g)–(i) is θv = 308, 309, 311, and 312 K, respectively, while gray fill increases darkness in 1 K increments. The x-distance scale of an unlabeled panel is the scale of the next lower labeled panel.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
(Continued) An open rectangle in (d), (e), and (f) locates the adjacent stepped traverse. The “cloud box” boundaries are thick if imaged, and thin either if the cloud extended outside the image area or if the cloud base was located by its shadow. Cloud boundaries are black if horizontal range Rh is less than or equal to 10 km and gray if Rh > 10 km. A cloud base above the analysis domain top ztop is denoted by a line segment just above ztop. Thin quasi-horizontal curves are the individual flight legs.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
(Continued) An open rectangle in (d), (e), and (f) locates the adjacent stepped traverse. The “cloud box” boundaries are thick if imaged, and thin either if the cloud extended outside the image area or if the cloud base was located by its shadow. Cloud boundaries are black if horizontal range Rh is less than or equal to 10 km and gray if Rh > 10 km. A cloud base above the analysis domain top ztop is denoted by a line segment just above ztop. Thin quasi-horizontal curves are the individual flight legs.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
(Continued) An open rectangle in (d), (e), and (f) locates the adjacent stepped traverse. The “cloud box” boundaries are thick if imaged, and thin either if the cloud extended outside the image area or if the cloud base was located by its shadow. Cloud boundaries are black if horizontal range Rh is less than or equal to 10 km and gray if Rh > 10 km. A cloud base above the analysis domain top ztop is denoted by a line segment just above ztop. Thin quasi-horizontal curves are the individual flight legs.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Skew T–logp diagrams of soundings on either side of the dryline on 15–16 May 1991: (a) NSSL-1 (2311);(b) NSSL-2 (2307); (c) NSSL-2 (0057). Soundings indicated at upper right are denoted by solid, black curves, with sounding parameters indicated at lower left. Dashed gray curves in (a) and (c) are the (reference) NSSL-2 (2307) sounding. The (x, y) sounding coordinates (km) relative to the analysis origin are indicated at upper left. Open triangles and horizontal segments joined by moist-adiabatic curves (gray) denote the bases and tops, respectively, of individual, photogrammetrically measured cumuli. The parameters CBP (cloud base pressure) and CBZ (cloud base height) have units of mb and km MSL, respectively. Height scale has units of km MSL. Due to noisy Loran data, NSSL-1 winds in (a) are not present at heights between 750 and 500 mb while NSSL-2 winds in (c) are not plotted at pressures less than 800 mb.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Skew T–logp diagrams of soundings on either side of the dryline on 15–16 May 1991: (a) NSSL-1 (2311);(b) NSSL-2 (2307); (c) NSSL-2 (0057). Soundings indicated at upper right are denoted by solid, black curves, with sounding parameters indicated at lower left. Dashed gray curves in (a) and (c) are the (reference) NSSL-2 (2307) sounding. The (x, y) sounding coordinates (km) relative to the analysis origin are indicated at upper left. Open triangles and horizontal segments joined by moist-adiabatic curves (gray) denote the bases and tops, respectively, of individual, photogrammetrically measured cumuli. The parameters CBP (cloud base pressure) and CBZ (cloud base height) have units of mb and km MSL, respectively. Height scale has units of km MSL. Due to noisy Loran data, NSSL-1 winds in (a) are not present at heights between 750 and 500 mb while NSSL-2 winds in (c) are not plotted at pressures less than 800 mb.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Skew T–logp diagrams of soundings on either side of the dryline on 15–16 May 1991: (a) NSSL-1 (2311);(b) NSSL-2 (2307); (c) NSSL-2 (0057). Soundings indicated at upper right are denoted by solid, black curves, with sounding parameters indicated at lower left. Dashed gray curves in (a) and (c) are the (reference) NSSL-2 (2307) sounding. The (x, y) sounding coordinates (km) relative to the analysis origin are indicated at upper left. Open triangles and horizontal segments joined by moist-adiabatic curves (gray) denote the bases and tops, respectively, of individual, photogrammetrically measured cumuli. The parameters CBP (cloud base pressure) and CBZ (cloud base height) have units of mb and km MSL, respectively. Height scale has units of km MSL. Due to noisy Loran data, NSSL-1 winds in (a) are not present at heights between 750 and 500 mb while NSSL-2 winds in (c) are not plotted at pressures less than 800 mb.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Deep dryline convection on the evening of 16 May 1991. (a) Photo looking north at 0040, with NSSL-2 in foreground and developing Wheeler County, TX, storm in background (courtesy C. Hane, NSSL); (b)–(c) WSR-88D radar reflectivity from KTLX (Oklahoma City) at 0101 and 0136, respectively. The location and viewing direction of photo (“P”) in (a) are indicated in (b), while times and locations of tornado touchdowns (“T”) are indicated in (c). Selected 1-min PAM mesonet observations are depicted in (b) and (c) as in Fig. 1.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Deep dryline convection on the evening of 16 May 1991. (a) Photo looking north at 0040, with NSSL-2 in foreground and developing Wheeler County, TX, storm in background (courtesy C. Hane, NSSL); (b)–(c) WSR-88D radar reflectivity from KTLX (Oklahoma City) at 0101 and 0136, respectively. The location and viewing direction of photo (“P”) in (a) are indicated in (b), while times and locations of tornado touchdowns (“T”) are indicated in (c). Selected 1-min PAM mesonet observations are depicted in (b) and (c) as in Fig. 1.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Deep dryline convection on the evening of 16 May 1991. (a) Photo looking north at 0040, with NSSL-2 in foreground and developing Wheeler County, TX, storm in background (courtesy C. Hane, NSSL); (b)–(c) WSR-88D radar reflectivity from KTLX (Oklahoma City) at 0101 and 0136, respectively. The location and viewing direction of photo (“P”) in (a) are indicated in (b), while times and locations of tornado touchdowns (“T”) are indicated in (c). Selected 1-min PAM mesonet observations are depicted in (b) and (c) as in Fig. 1.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Evolution of CAPE and CIN as derived from NSSL-2 M-CLASS soundings on 15–16 May 1991. The labels A and S denote the NSSL-2 sounding locations in the east–west (W) direction relative to the towns of Alanreed and Shamrock, TX, respectively, located by the symbols A and N2 in Fig. 1c.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Evolution of CAPE and CIN as derived from NSSL-2 M-CLASS soundings on 15–16 May 1991. The labels A and S denote the NSSL-2 sounding locations in the east–west (W) direction relative to the towns of Alanreed and Shamrock, TX, respectively, located by the symbols A and N2 in Fig. 1c.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Evolution of CAPE and CIN as derived from NSSL-2 M-CLASS soundings on 15–16 May 1991. The labels A and S denote the NSSL-2 sounding locations in the east–west (W) direction relative to the towns of Alanreed and Shamrock, TX, respectively, located by the symbols A and N2 in Fig. 1c.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 1 but for 7 June 1994 case. (a) Surface state at 2200 UTC, with filled black squares locating Oklahoma mesonet observations; (b) visible GOES-8 satellite image at 2231 UTC; (c) photogrammetric cloud observations and P-3 measurements along ∼150 m AGL legs at 2114–2125 (western S–N leg), 2137–2148 (eastern N–S leg), and 2215–2223 (east–west leg). Westernmost three soundings in Fig. 8 are located in (c). The circled cross in (a) locates the NCAR M-CLASS sounding taken near Alva, OK.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 1 but for 7 June 1994 case. (a) Surface state at 2200 UTC, with filled black squares locating Oklahoma mesonet observations; (b) visible GOES-8 satellite image at 2231 UTC; (c) photogrammetric cloud observations and P-3 measurements along ∼150 m AGL legs at 2114–2125 (western S–N leg), 2137–2148 (eastern N–S leg), and 2215–2223 (east–west leg). Westernmost three soundings in Fig. 8 are located in (c). The circled cross in (a) locates the NCAR M-CLASS sounding taken near Alva, OK.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 1 but for 7 June 1994 case. (a) Surface state at 2200 UTC, with filled black squares locating Oklahoma mesonet observations; (b) visible GOES-8 satellite image at 2231 UTC; (c) photogrammetric cloud observations and P-3 measurements along ∼150 m AGL legs at 2114–2125 (western S–N leg), 2137–2148 (eastern N–S leg), and 2215–2223 (east–west leg). Westernmost three soundings in Fig. 8 are located in (c). The circled cross in (a) locates the NCAR M-CLASS sounding taken near Alva, OK.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 2, but for 7 June 1994 case. Analyses (left column): (a) 2215–2245; (b) 2248–2328; (c) 2332–0001. Cloud images (right column): (d) 2225:40 (east); (e) 2236:52 (west); (f) 2242:02 (north); (g) 2310:03 (east). The margin between white and light gray fill is θυ = 321 K, while gray fill increases darkness in 1 K increments.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 2, but for 7 June 1994 case. Analyses (left column): (a) 2215–2245; (b) 2248–2328; (c) 2332–0001. Cloud images (right column): (d) 2225:40 (east); (e) 2236:52 (west); (f) 2242:02 (north); (g) 2310:03 (east). The margin between white and light gray fill is θυ = 321 K, while gray fill increases darkness in 1 K increments.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 2, but for 7 June 1994 case. Analyses (left column): (a) 2215–2245; (b) 2248–2328; (c) 2332–0001. Cloud images (right column): (d) 2225:40 (east); (e) 2236:52 (west); (f) 2242:02 (north); (g) 2310:03 (east). The margin between white and light gray fill is θυ = 321 K, while gray fill increases darkness in 1 K increments.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 3 but for 7 June 1994 case: (a) NSSL-2 at 2131; (b) NSSL-4 at 2143; (c) NSSL-3 at 2131; (d) NCAR at 2118. Soundings are located (in order of presentation) in Figs. 6c and 6a. Dashed gray curve in (a) is NSSL-4 sounding.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 3 but for 7 June 1994 case: (a) NSSL-2 at 2131; (b) NSSL-4 at 2143; (c) NSSL-3 at 2131; (d) NCAR at 2118. Soundings are located (in order of presentation) in Figs. 6c and 6a. Dashed gray curve in (a) is NSSL-4 sounding.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 3 but for 7 June 1994 case: (a) NSSL-2 at 2131; (b) NSSL-4 at 2143; (c) NSSL-3 at 2131; (d) NCAR at 2118. Soundings are located (in order of presentation) in Figs. 6c and 6a. Dashed gray curve in (a) is NSSL-4 sounding.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
East–west profiles of CAPE and CIN across the dryline as derived from M-CLASS soundings on 7 June 1994.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
East–west profiles of CAPE and CIN across the dryline as derived from M-CLASS soundings on 7 June 1994.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
East–west profiles of CAPE and CIN across the dryline as derived from M-CLASS soundings on 7 June 1994.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Intercomparison of east–west profiles of qυ (solid) and θυ (dashed) across the dryline on 7 June 1994. The 2149–2159 P-3 (black) and 2136–2158 NSSL-2 mobile mesonet (gray) traverses are depicted. The P-3 traverse is at ∼150 m AGL.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Intercomparison of east–west profiles of qυ (solid) and θυ (dashed) across the dryline on 7 June 1994. The 2149–2159 P-3 (black) and 2136–2158 NSSL-2 mobile mesonet (gray) traverses are depicted. The P-3 traverse is at ∼150 m AGL.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Intercomparison of east–west profiles of qυ (solid) and θυ (dashed) across the dryline on 7 June 1994. The 2149–2159 P-3 (black) and 2136–2158 NSSL-2 mobile mesonet (gray) traverses are depicted. The P-3 traverse is at ∼150 m AGL.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 1 but for 6 May 1995 case: (a) Surface state at 2200; (b) visible GOES-8 satellite image at 2202; (c) photogrammetric cloud observations and P-3 measurements at ∼150 m AGL (2216–2223 leg). In (c), the heavy curve denotes dryline location as inferred from clear air convergence and local maximum radar reflectivity; contours are 0.5 and 1.0 m s−1 vertical velocity values at 0.7 km AGL using the 2219–2229 NCAR Electra ELDORA radar analysis leg [adapted from Fig. 12a of Atkins et al. (1998)]. The AMA and NSSL-4 (N4) soundings in Fig. 13 are located in (c).
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 1 but for 6 May 1995 case: (a) Surface state at 2200; (b) visible GOES-8 satellite image at 2202; (c) photogrammetric cloud observations and P-3 measurements at ∼150 m AGL (2216–2223 leg). In (c), the heavy curve denotes dryline location as inferred from clear air convergence and local maximum radar reflectivity; contours are 0.5 and 1.0 m s−1 vertical velocity values at 0.7 km AGL using the 2219–2229 NCAR Electra ELDORA radar analysis leg [adapted from Fig. 12a of Atkins et al. (1998)]. The AMA and NSSL-4 (N4) soundings in Fig. 13 are located in (c).
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 1 but for 6 May 1995 case: (a) Surface state at 2200; (b) visible GOES-8 satellite image at 2202; (c) photogrammetric cloud observations and P-3 measurements at ∼150 m AGL (2216–2223 leg). In (c), the heavy curve denotes dryline location as inferred from clear air convergence and local maximum radar reflectivity; contours are 0.5 and 1.0 m s−1 vertical velocity values at 0.7 km AGL using the 2219–2229 NCAR Electra ELDORA radar analysis leg [adapted from Fig. 12a of Atkins et al. (1998)]. The AMA and NSSL-4 (N4) soundings in Fig. 13 are located in (c).
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig 2, but for 6 May 1995 case. Analyses (left column): (a) 2145–2213; (b) 2216–2246; (c) 2248–2319; (d) 2322–2351; (i) 2354–0046. Cloud images (right column): (e) 2209:49 (south); (f) 2240:10 (north); (g) 2304:03 (north); (h) 2325:12 (north); (j) 0001:06 (south). The margin between white and light gray fill is θυ = 309 K, while gray fill increases darkness in 1 K increments.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig 2, but for 6 May 1995 case. Analyses (left column): (a) 2145–2213; (b) 2216–2246; (c) 2248–2319; (d) 2322–2351; (i) 2354–0046. Cloud images (right column): (e) 2209:49 (south); (f) 2240:10 (north); (g) 2304:03 (north); (h) 2325:12 (north); (j) 0001:06 (south). The margin between white and light gray fill is θυ = 309 K, while gray fill increases darkness in 1 K increments.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig 2, but for 6 May 1995 case. Analyses (left column): (a) 2145–2213; (b) 2216–2246; (c) 2248–2319; (d) 2322–2351; (i) 2354–0046. Cloud images (right column): (e) 2209:49 (south); (f) 2240:10 (north); (g) 2304:03 (north); (h) 2325:12 (north); (j) 0001:06 (south). The margin between white and light gray fill is θυ = 309 K, while gray fill increases darkness in 1 K increments.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 3 but for the 6 May 1995 case: (a) Amarillo (2300); (b) NSSL-4 (2250). Soundings are located in Fig. 11c. Dashed gray curve in (a) is NSSL-4 sounding.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 3 but for the 6 May 1995 case: (a) Amarillo (2300); (b) NSSL-4 (2250). Soundings are located in Fig. 11c. Dashed gray curve in (a) is NSSL-4 sounding.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 3 but for the 6 May 1995 case: (a) Amarillo (2300); (b) NSSL-4 (2250). Soundings are located in Fig. 11c. Dashed gray curve in (a) is NSSL-4 sounding.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 10 but for the 6 May 1995 case. The 2322–2331 P-3 (black) and 2323–2330 Probe-3 mobile mesonet (gray) traverses are depicted.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 10 but for the 6 May 1995 case. The 2322–2331 P-3 (black) and 2323–2330 Probe-3 mobile mesonet (gray) traverses are depicted.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Same as in Fig. 10 but for the 6 May 1995 case. The 2322–2331 P-3 (black) and 2323–2330 Probe-3 mobile mesonet (gray) traverses are depicted.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Histogram of cloud frequencies in 10-km-wide intervals across the dryline for all case days. One cloud event means that one cloud area was contained within or overlapped into the specified distance interval (i.e., one or more clouds per interval).
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Histogram of cloud frequencies in 10-km-wide intervals across the dryline for all case days. One cloud event means that one cloud area was contained within or overlapped into the specified distance interval (i.e., one or more clouds per interval).
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Histogram of cloud frequencies in 10-km-wide intervals across the dryline for all case days. One cloud event means that one cloud area was contained within or overlapped into the specified distance interval (i.e., one or more clouds per interval).
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Vertical profiles of horizontally averaged moisture flux convergence [(a), (c), (e)] and vertical velocity [(b), (d), (f)] over the cross-sectional analysis domains in the 15 May 1991 (row 1), 7 June 1994 (row 2), and 6 May 1995 (row 3) cases. Moisture flux convergence is the east–west component.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Vertical profiles of horizontally averaged moisture flux convergence [(a), (c), (e)] and vertical velocity [(b), (d), (f)] over the cross-sectional analysis domains in the 15 May 1991 (row 1), 7 June 1994 (row 2), and 6 May 1995 (row 3) cases. Moisture flux convergence is the east–west component.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Vertical profiles of horizontally averaged moisture flux convergence [(a), (c), (e)] and vertical velocity [(b), (d), (f)] over the cross-sectional analysis domains in the 15 May 1991 (row 1), 7 June 1994 (row 2), and 6 May 1995 (row 3) cases. Moisture flux convergence is the east–west component.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Conceptual model of dryline environment during afternoon and early evening, showing dryline position in relation to cumulus clouds and airflow streamlines. The lower heavy dashed curve denotes the extent of the moist convective boundary layer, while the upper heavy dashed curve locates the deep, dry convective boundary layer (west of dryline), and the elevated residual layer (east of dryline and above moist layer). The gray dashed curve locates the surface of zero westerly wind component. The vertical gray lines locate proximity and dryline soundings described in the text. The heavy dashed streamline denotes a buoyantly accelerated cloudy air parcel trajectory.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Conceptual model of dryline environment during afternoon and early evening, showing dryline position in relation to cumulus clouds and airflow streamlines. The lower heavy dashed curve denotes the extent of the moist convective boundary layer, while the upper heavy dashed curve locates the deep, dry convective boundary layer (west of dryline), and the elevated residual layer (east of dryline and above moist layer). The gray dashed curve locates the surface of zero westerly wind component. The vertical gray lines locate proximity and dryline soundings described in the text. The heavy dashed streamline denotes a buoyantly accelerated cloudy air parcel trajectory.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
Conceptual model of dryline environment during afternoon and early evening, showing dryline position in relation to cumulus clouds and airflow streamlines. The lower heavy dashed curve denotes the extent of the moist convective boundary layer, while the upper heavy dashed curve locates the deep, dry convective boundary layer (west of dryline), and the elevated residual layer (east of dryline and above moist layer). The gray dashed curve locates the surface of zero westerly wind component. The vertical gray lines locate proximity and dryline soundings described in the text. The heavy dashed streamline denotes a buoyantly accelerated cloudy air parcel trajectory.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
M-CLASS soundings (black) modified by a combination of horizontal and vertical transport using the simplified model described in the text: (a) 0000 UTC 16 May 1991; (b) 2112 UTC 7 June 1994 (uwest > 0); (c); c) 2250 UTC 6 May 1995; (d) 2112 UTC 7 June 1994 (uwest = 0). In (b) and (d), the modified sounding is compared to the 2143 UTC sounding (gray) released at the dryline location by NSSL-4. CAPE and CIN have units of J kg−1, while CBP and CBZ have units of mb and km (MSL), respectively. The parameter hmax (the height of the maximum updraft) has units of km (AGL); the definitions of all listed model parameters are given in the text. Parameter subscripting is omitted for clarity.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
M-CLASS soundings (black) modified by a combination of horizontal and vertical transport using the simplified model described in the text: (a) 0000 UTC 16 May 1991; (b) 2112 UTC 7 June 1994 (uwest > 0); (c); c) 2250 UTC 6 May 1995; (d) 2112 UTC 7 June 1994 (uwest = 0). In (b) and (d), the modified sounding is compared to the 2143 UTC sounding (gray) released at the dryline location by NSSL-4. CAPE and CIN have units of J kg−1, while CBP and CBZ have units of mb and km (MSL), respectively. The parameter hmax (the height of the maximum updraft) has units of km (AGL); the definitions of all listed model parameters are given in the text. Parameter subscripting is omitted for clarity.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
M-CLASS soundings (black) modified by a combination of horizontal and vertical transport using the simplified model described in the text: (a) 0000 UTC 16 May 1991; (b) 2112 UTC 7 June 1994 (uwest > 0); (c); c) 2250 UTC 6 May 1995; (d) 2112 UTC 7 June 1994 (uwest = 0). In (b) and (d), the modified sounding is compared to the 2143 UTC sounding (gray) released at the dryline location by NSSL-4. CAPE and CIN have units of J kg−1, while CBP and CBZ have units of mb and km (MSL), respectively. The parameter hmax (the height of the maximum updraft) has units of km (AGL); the definitions of all listed model parameters are given in the text. Parameter subscripting is omitted for clarity.
Citation: Weather and Forecasting 13, 4; 10.1175/1520-0434(1998)013<1106:TIOMCA>2.0.CO;2
The u component is approximately the dryline-normal wind component, since long (∼100 km) dryline segments are often oriented on average approximately north–south. The normal wind component in cases where undulations or bulges occur along dryline segments would have both zonal and meridional components.
The 12-s smoothing period corresponds to a horizontal distance of roughly 1.4 km at a nominal ground speed of 120 m s−1.
Betts (1984) showed that an aircraft traverse or a balloon or aircraft sounding through the convective boundary layer is sufficient to characterize the thermal mixed layer properties in the vicinity of the in situ measurements.
The term “propagation” refers to the storm movement due to the combined effects of northeastward advection and rapid redevelopment on the southwest storm flank.
Photogrammetric analysis was used to estimate the heights of cloud bases when visible in images. Additionally, cloud shadows visible in Fig. 6b correspond to a cloud height of 3.7 km AGL given a computed local sun angle of ∼37° and a roughly estimated distance of 5 km from the tip of a cloud to the edge of the corresponding cloud shadow.
There must be a superadiabatic layer near the surface as long as the turbulent convective boundary layer is active, since it is the unstable stratification that drives the turbulent vertical fluxes. Ziegler et al. (1997) present additional examples of superadiabatic layers in observed and modeled soundings during afternoon east of or along the 15 May, 16 May, and 26 May 1991 drylines (Figs. 8, 16, and 18), and explain on pp. 1015–1016 that a mesoscale horizontal influx of potentially cool air toward the boundary offsets the vertical fluxes and helps maintain a minimum virtual potential temperature above the surface layer.
The chilled mirror dewpointer system flown on the P-3 during COPS and VORTEX could not respond to dewpoint temperature drops exceeding roughly 5°C km−1 (Ziegler and Hane 1993).
The virtual temperature correction decreases the computed CIN value over that value defined by the area enclosed by the union of the environmental and lifted parcel temperature profiles.
Our present dataset is regrettably too limited to allow an assessment of mixing following the motion of boundary layer and cumulus air parcels. Enhanced mesoscale observations that define the 4D fields of small mesoscale boundary layer airflow and absolute humidity, together with in situ thermodynamic parcel samples targeted to specific volumes of rising air, would provide a means to evaluate the bulk impact of mixing, if any. The concluding section of this paper offers remarks about preliminary plans to obtain such measurements.
Call letters of government-operated radio station broadcasting time standard.
