• Albers, S. C., J. A. McGinley, D. L. Birkenheuer, and J. R. Smart, 1996: The Local Analysis and Prediction System (LAPS): Analysis of clouds, precipitation, and temperature. Wea. Forecasting,11, 273–287.

  • Beebe, R. G., 1958: An instability line development as observed by the tornado research airplane. J. Meteor.,15, 278–282.

  • Black, T. L., 1988: The step-mountain eta coordinate regional model: A documentation. NOAA/NWS Rep., 47 pp. [Available from the National Centers for Environmental Prediction, Room 204, 5200 Auth Road, Camp Springs, MD 20746.].

  • ——, 1994: The new NMC mesoscale eta model: Description and forecast examples. Wea. Forecasting,9, 265–278.

  • Bluestein, H. B., and G. R. Woodall, 1990: Doppler-radar analysis of a low-precipitation severe storm. Mon. Wea. Rev.,118, 1640–1664.

  • ——, and S. Parker, 1993: Modes of isolated convective storm formation along the dryline. Mon. Wea. Rev.,121, 1354–1372.

  • ——, E. W. McCaul Jr., G. P. Byrd, and G. R. Woodall, 1988: Mobile sounding observations of a tornadic storm near the dryline: The Canadian, Texas, storm of 7 May 1986. Mon. Wea. Rev.,116, 1790–1804.

  • ——, ——, ——, ——, G. Martin, S. Keighton, and L. C. Showell, 1989: Mobile sounding observations of a thunderstorm near the dryline: The Gruver, Texas, storm complex of 25 May 1987. Mon. Wea. Rev.,117, 244–250.

  • ——, ——, ——, R. L. Walko, and R. P. Davies-Jones, 1990: An observational study of splitting convective clouds. Mon. Wea. Rev.,118, 1359–1370.

  • Brock, F., G. Saum, and S. Semmer, 1986: Portable Automated Mesonet II. J. Atmos. Oceanic Technol.,3, 573–582.

  • ——, K. C. Crawford, R. L. Elliott, G. W. Cuperus, S. J. Stadler, H. L. Johnson, and M. D. Eilts, 1995: The Oklahoma mesonet: A technical overview. J. Atmos. Oceanic Technol.,12, 5–19.

  • Carlson, T. N., and F. H. Ludlam, 1968: Conditions for the development of severe local storms. Tellus,20, 203–226.

  • ——, S. G. Benjamin, G. S. Forbes, and Y.-F. Li, 1983: Elevated mixed layers in the regional severe storm environment: Conceptual model and case studies. Mon. Wea. Rev.,111, 1453–1473.

  • Chadwick, R., 1988: The wind profiler demonstration network. Preprints, Symp. on Lower Tropospheric Profiling: Needs and Technologies, Boulder, CO, Amer. Meteor. Soc., 109–110.

  • Chang, J.-T., and P. J. Wetzel, 1991: Effects of spatial variations of soil moisture and vegetation on the evolution of a prestorm environment: A numerical case study. Mon. Wea. Rev.,119, 1368–1390.

  • Christian, T. W., and R. M. Wakimoto, 1989: The relationship between radar reflectivities and clouds associated with horizontal roll convection on 8 August 1982. Mon. Wea. Rev.,117, 1530–1544.

  • Crook, N. A., T. L. Clark, and M. W. Moncrieff, 1990: The Denver cyclone. Part I: Generation in low Froude number flow. J. Atmos. Sci.,47, 2725–2742.

  • Fankhauser, J. C., 1971: Thunderstorm environment interactions determined from aircraft and radar observations. Mon. Wea. Rev.,99, 171–192.

  • Fujita, T. T., 1958: Structure and movement of a dry front. Bull. Amer. Meteor. Soc.,39, 574–582.

  • Gal-Chen, T., 1978: A method for the initialization of the anelastic equations: Implications for matching models with observations. Mon. Wea. Rev.,106, 587–606.

  • Gurka, J. J., 1976: Satellite and surface observations of strong wind zones accompanying thunderstorms. Mon. Wea. Rev.,104, 1484–1493.

  • Hane, C. E., R. B. Wilhelmson, and T. Gal-Chen, 1981: Retrieval of thermodynamic variables within deep convective clouds: Experiments in three dimensions. Mon. Wea. Rev.,109, 564–576.

  • ——, C. L. Ziegler, and H. B. Bluestein, 1993: Investigation of the dryline and convective storms initiated along the dryline: Field experiments during COPS-91. Bull. Amer. Meteor. Soc.,74, 2133–2145.

  • Jorgensen, D. P., and M. A. LeMone, 1989: Vertical velocity characteristics of oceanic convection. J. Atmos. Sci.,46, 621–640.

  • Kaimal, J. C., and J. J. Finnigan, 1993: Atmospheric Boundary Layer Flows—Their Structure and Measurement. Oxford University Press, 289 pp.

  • Koch, S. E., 1982: The role of apparent gravity waves as a triggering mechanism for severe thunderstorms along a dryline. Preprints, 12th Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 573–576.

  • ——, and J. McCarthy, 1982: The evolution of an Oklahoma dryline. Part II: Boundary layer forcing of mesoconvective systems. J. Atmos. Sci.,39, 237–257.

  • LeMone, M. A., 1973: The structure and dynamics of horizontal roll vortices in the planetary boundary layer. J. Atmos. Sci.,30, 1077–1091.

  • Lewis, J. M., Y. Ogura, and L. Gidel, 1974: Large-scale influences upon the generation of a mesoscale disturbance. Mon. Wea. Rev.,102, 545–560.

  • Mahrt, L., 1976: Mixed layer moisture structure. Mon. Wea. Rev.,104, 1403–1407.

  • Marks, F., Jr., and R. Houze, 1987: Inner core structure of Hurricane Alicia from airborne Doppler radar observations. J. Atmos. Sci.,44, 1296–1317.

  • McCarthy, J., and S. E. Koch, 1982: The evolution of an Oklahoma dryline. Part I: A mesoscale and subsynoptic-scale analysis. J. Atmos. Sci.,39, 225–236.

  • McGuire, E. L., 1962: The vertical structure of three drylines as revealed by aircraft traverses. NSSP Rep. 7, 11 pp. [Available from National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.].

  • Mesinger, F., Z. I. Janjic, S. Nickovic, D. Gavrilov, and D. G. Deaven, 1988: The step-mountain coordinate: Model description and performance for cases of Alpine lee cyclogenesis and for a case of Appalachian redevelopment. Mon. Wea Rev.,116, 1493–1518.

  • Murray, R., and S. M. Daniels, 1953: Transverse flow in the entrance and exit to jet streams. Quart. J. Roy. Meteor. Soc.,79, 236–241.

  • NSSP Staff Members, 1963: Environmental and thunderstorm structures as shown by National Severe Storms Project observations in the spring of 1960 and 1961. Mon. Wea. Rev.,91, 271–292.

  • Palmén, E., and C. W. Newton, 1969: Atmospheric Circulation Systems: Their Structure and Physical Interpretation. International Geophysics Series, Vol.13, Academic Press, 603 pp.

  • 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.

  • Purdom, J. F. W., 1976: Some uses of high-resolution GOES imagery in the mesoscale forecasting of convection and its behavior. Mon. Wea. Rev.,104, 1474–1483.

  • Rabin, R. M., S. Stadler, P. J. Wetzel, D. J. Stensrud, and M. Gregory, 1990: Observed effects of landscape variability on convective clouds. Bull. Amer. Meteor. Soc.,71, 272–280.

  • Rhea, J. O., 1966: A study of thunderstorm formation along drylines. J. Appl. Meteor.,5, 58–63.

  • Rotunno, R., and J. B. Klemp, 1982: The influence of shear-induced pressure gradient on thunderstorm motion. Mon. Wea. Rev.,110, 136–151.

  • Rust, W. D., D. W. Burgess, R. A. Maddox, L. Showell, T. Marshall, and D. Lauritzen, 1990: Testing a mobile version of a Cross-chain LORAN Atmospheric (M-CLASS) Sounding System. Bull. Amer. Meteor. Soc.,71, 173–180.

  • Sanders, F., and D. O. Blanchard, 1993: The origin of a severe thunderstorm in Kansas on 10 May 1985. Mon. Wea. Rev.,121, 133–149.

  • Schaefer, J. T., 1974: A simulative model of dryline motion. J. Atmos. Sci.,31, 956–964.

  • ——, 1986: The dryline. Mesoscale Meteorology and Forecasting, P. S. Ray, Ed., Amer. Meteor. Soc., 549–572.

  • Segal, M., and R. W. Arritt, 1992: Nonclassical mesoscale circulations caused by surface sensible heat-flux gradients. Bull. Amer. Meteor. Soc.,73, 1593–1604.

  • ——, W. E. Shreiber, G. Kallos, J. R. Garratt, A. Rodi, J. Weaver, and R. A. Pielke, 1989: The impact of crop areas in northeast Colorado on midsummer mesoscale thermal circulations. Mon. Wea. Rev.,117, 809–825.

  • Shapiro, M. A., 1982: Mesoscale weather systems of the central United States. Cooperative Institute for Research in Environmental Sciences Rep., 78 pp. [Available from National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.].

  • Sutherland, R. A., 1986: Broadband and spectral emissivities (2–18 mm) of some natural soils and vegetation. J. Atmos. Oceanic Technol.,3, 199–202.

  • Szoke, E. J., M. L. Weisman, J. M. Brown, F. Caracena, and T. W. Schlatter, 1984: A subsynoptic analysis of the Denver tornadoes of 3 June 1981. Mon. Wea. Rev.,112, 790–808.

  • Treadon, R. E., 1993: The NMC eta model post processor: A documentation. NMC Office Note 394, 44 pp. [Available from the National Centers for Environmental Prediction, NOAA Science Center, Room 101, 5200 Auth Road, Camp Springs, MD 20746.].

  • Uccellini, L. W., and D. R. Johnson, 1979: The coupling of upper and lower tropospheric jet streaks and implications for the development of severe convective storms. Mon. Wea. Rev.,107, 682–703.

  • Wakimoto, R. M., and N. T. Atkins, 1994: Observations of the sea-breeze front during CaPE. Part I: Single-Doppler, satellite, and cloud photogrammetry analysis. Mon. Wea. Rev.,122, 1092–1114.

  • Weckwerth, T. M., and R. M. Wakimoto, 1992: The initiation and organization of convective cells atop a cold-air outflow boundary. Mon. Wea. Rev.,120, 2169–2187.

  • Wilson, J. W., and W. E. Schreiber, 1986: Initiation of convective storms by radar-observed boundary layer convergent lines. Mon. Wea. Rev.,114, 2516–2536.

  • ——, 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.

  • Zhao, Q., F. H. Carr, and G. B. Lesins, 1991: Improvement of precipitation forecasts by including cloud water in NMC’s eta model. Preprints, Ninth Conf. on Numerical Weather Prediction, Denver, CO, Amer. Meteor. Soc., 50–53.

  • 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, 1996: Simulations of deep, moist convection at the dryline with a 3-D mesoscale model. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 273–277.

  • View in gallery
    Fig. 1.

    Upper-air analyses at (a) 700 mb, 1200 UTC 26 May 1991; (b) 700 mb, 0000 UTC 27 May 1991; (c) 300 mb, 1200 UTC 26 May 1991; and (d) 300 mb, 0000 UTC 27 May 1991. Solid contours are heights (dam), dashed contours are temperatures (°C), and station models use conventional format (long wind barb, 5 m s−1; short barb, 2.5 m s−1; and flag, 25 m s−1).

  • View in gallery
    Fig. 2.

    Surface dewpoint (°C) analyses at (a) 1500 UTC 26 May 1991, (b) 1800 UTC 26 May 1991, (c) 2100 UTC 26 May 1991, and (d) 0000 UTC 27 May 1991. Dryline denoted by scalloped line, outflow boundary by dash–dotted line, and PAM stations by solid squares. Station models use conventional format with sea level pressure (mb and tenths above 1000) at upper right.

  • View in gallery
    Fig. 3.

    Wind profiler observations from Vici, Oklahoma (circled “V” in Fig. 2d), on 26 May 1991. Wind barbs have conventional format.

  • View in gallery
    Fig. 4.

    Skew T plots of National Weather Service soundings at 0000 UTC 27 May 1991 from (a) Amarillo, Texas, and (b) Norman, Oklahoma.

  • View in gallery
    Fig. 5.

    Vertical cross-sectional analyses from aircraft pattern flown from 2158 to 2323 UTC 26 May 1991 at the location marked by the dashed line in Fig. 2d. (a) Water vapor mixing ratio (g kg−1, solid) and aircraft track (dotted); (b) west-to-east component of the wind (m s−1) with positive values solid and negative dashed; (c) virtual potential temperature (excess above 310 K, solid) and shaded areas of vertical motion (see key); and (d) deviation perturbation pressure (10−1 mb) with positive contours solid, negative dashed, and 10 g kg−1 vapor mixing ratio heavy dotted contour.

  • View in gallery
    Fig. 6.

    Mobile-CLASS soundings at the times and locations noted in (a)–(d). Sounding locations are also noted in Fig. 2d. In (d) an aircraft sounding (thin solid lines) from the location in Fig. 2d denoted by circled “P” is included.

  • View in gallery
    Fig. 7.

    Relative humidity (%) as a function of pressure (mb) from mobile-CLASS soundings at the four noted times east of the dryline near Shamrock, Texas.

  • View in gallery
    Fig. 8.

    Visible images centered on the northeastern Texas panhandle from GOES-7 satellite on 26 May 1991 at (a) 2001 and (b) 2101 UTC.

  • View in gallery
    Fig. 8.

    (Continued) (c) Dryline locations at 2100 and 0000 UTC and tracks of storms initiated along dryline with severe reports (T—tornado and H—hail) are indicated.

  • View in gallery
    Fig. 9.

    Aircraft track from 2023 to 2152 UTC 26 May 1991 with wind speed and direction noted at 1-min intervals. Scalloped and dashed lines enclose area of dryline primary moisture gradients. Dash–dot–dot line indicates location of cloud line noted in text; “C” and “M” denote respective locations of Canadian and McLean PAM sites.

  • View in gallery
    Fig. 10.

    Dewpoint distribution in the dryline region based on (a) aircraft observations at 860 mb and (b) PAM time series. In (a) dewpoint temperature (°C) shown by solid contours is derived from aircraft measurements along track shown. Dryline is indicated by scalloped line and cloud line by dash–dot–dot line; location of first echo also noted.

  • View in gallery
    Fig. 11.

    Analyzed fields from aircraft sawtooth pattern at 860 mb; dryline and cloud line locations indicated as in Fig. 10a. Variable fields are (a) temperature (°C), (b) virtual potential temperature (K), and (c) horizontal divergence (10−4 s−1). In (c) divergence values less than −1.5 × 10−4 s−1 are shaded.

  • View in gallery
    Fig. 12.

    South-to-north wind component (m s−1) along a 10-min segment of the aircraft track whose location (dashed) is noted in Fig. 11c.

  • View in gallery
    Fig. 13.

    Aircraft-measured vertical air motion (m s−1) along 9-min segments of the aircraft pattern ahead of the dryline (a) in the area where thunderstorms developed and (b) 120 km to the south-southwest.

  • View in gallery
    Fig. 14.

    NDVI image from the NOAA-11 polar-orbiting satellite at 2105 UTC 27 May 1991. Brighter areas indicate higher vegetative density. Black areas are clouds or bodies of surface water.

  • View in gallery
    Fig. 15.

    Surface radiative temperature from GOES-7 infrared image on 26 May 1991 at (a) 1901 UTC and (b) 2101 UTC. Darker shades indicate warmer temperatures; white areas are cloud. Smoothed contours of temperature (°C) are included in cloud-free areas.

  • View in gallery
    Fig. 16.

    Water vapor mixing ratio (g kg−1) from aircraft measurements along the 10-min segment of track whose location is indicated in Fig. 11c.

  • View in gallery
    Fig. 17.

    Water vapor mixing ratio (g kg−1) in the lowest atmospheric layer from the mesoscale eta model at (a) 1800 UTC 26 May 1991 and (b) 0000 UTC 27 May 1991.

  • View in gallery
    Fig. 18.

    Vertical motion (10−2 Pa s−1) at 750 mb from the mesoscale eta model at 2100 UTC. The location of the observed cloud line is indicated by the dash–dotted line.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 363 154 22
PDF Downloads 178 68 8

Severe Thunderstorm Development in Relation to Along-Dryline Variability: A Case Study

Carl E. HaneNOAA/National Severe Storms Laboratory, Norman, Oklahoma

Search for other papers by Carl E. Hane in
Current site
Google Scholar
PubMed
Close
,
Howard B. BluesteinSchool of Meteorology, University of Oklahoma, Norman, Oklahoma

Search for other papers by Howard B. Bluestein in
Current site
Google Scholar
PubMed
Close
,
Todd M. CrawfordSchool of Meteorology, University of Oklahoma, Norman, Oklahoma

Search for other papers by Todd M. Crawford in
Current site
Google Scholar
PubMed
Close
,
Michael E. BaldwinGeneral Sciences Corporation, Laurel, Maryland

Search for other papers by Michael E. Baldwin in
Current site
Google Scholar
PubMed
Close
, and
Robert M. RabinNOAA/National Severe Storms Laboratory, Norman, Oklahoma

Search for other papers by Robert M. Rabin in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Long-lived thunderstorms were initiated during the afternoon of 26 May 1991 ahead of a dryline in northwestern Oklahoma. Various reasons for initiation in this particular along-dryline location are investigated through analysis of observations collected during the Cooperative Oklahoma Profiler Studies—1991 field program. Observing systems included in situ and radar instrumentation aboard a research aircraft, soundings from mobile laboratories, a mesonetwork of surface stations, meteorological satellites, and operational networks of surface and upper-air stations.

Elevated moistening east of the dryline revealed by soundings and aircraft observations in combination with thermal plume activity was apparently insufficient to promote sustained convection on this day without aid from an additional lifting mechanism. Satellite observations reveal scattered convection along the dryline by midafternoon and a convective cloud line intersecting the dryline at an angle in the area of most pronounced storm initiation, extending southwestward into the dry air. Another prominent feature on this day was a mesoscale bulge along the dryline extending northeastward into southwest Kansas. Deep convection was initiated along this bulge, but was in general short-lived.

Potential causes of the lifting associated with the cloud line that was apparently key to the preferred location for storm development in northwest Oklahoma were investigated: (a) a mesoscale circulation resulting from horizontal differences in radiative (temperature) properties of the underlying surface and (b) upward motion induced by an upper-level mesoscale disturbance. Analysis of vegetative and surface temperature distributions from satellite observations suggests a potential (more research is needed) link between surface characteristics and the development of the dryline bulge and observed cloud line through horizontal differences in vertical momentum transport. A run of the currently operational eta model indicates some skill in predicting dryline location and motion and predicts upward motion in the northern part of the region that was generally more convectively active, but shows no indication of upper-level support in the vicinity of the observed cloud line.

 Additional affiliation: NOAA/National Centers for Environmental Prediction, Camp Springs, Maryland.

 Current affiliation: NOAA/Storm Prediction Center, Norman, Oklahoma.

Corresponding author address: Dr. Carl E. Hane, National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.

Email: hane@nssl.nssl.uoknor.edu

Abstract

Long-lived thunderstorms were initiated during the afternoon of 26 May 1991 ahead of a dryline in northwestern Oklahoma. Various reasons for initiation in this particular along-dryline location are investigated through analysis of observations collected during the Cooperative Oklahoma Profiler Studies—1991 field program. Observing systems included in situ and radar instrumentation aboard a research aircraft, soundings from mobile laboratories, a mesonetwork of surface stations, meteorological satellites, and operational networks of surface and upper-air stations.

Elevated moistening east of the dryline revealed by soundings and aircraft observations in combination with thermal plume activity was apparently insufficient to promote sustained convection on this day without aid from an additional lifting mechanism. Satellite observations reveal scattered convection along the dryline by midafternoon and a convective cloud line intersecting the dryline at an angle in the area of most pronounced storm initiation, extending southwestward into the dry air. Another prominent feature on this day was a mesoscale bulge along the dryline extending northeastward into southwest Kansas. Deep convection was initiated along this bulge, but was in general short-lived.

Potential causes of the lifting associated with the cloud line that was apparently key to the preferred location for storm development in northwest Oklahoma were investigated: (a) a mesoscale circulation resulting from horizontal differences in radiative (temperature) properties of the underlying surface and (b) upward motion induced by an upper-level mesoscale disturbance. Analysis of vegetative and surface temperature distributions from satellite observations suggests a potential (more research is needed) link between surface characteristics and the development of the dryline bulge and observed cloud line through horizontal differences in vertical momentum transport. A run of the currently operational eta model indicates some skill in predicting dryline location and motion and predicts upward motion in the northern part of the region that was generally more convectively active, but shows no indication of upper-level support in the vicinity of the observed cloud line.

 Additional affiliation: NOAA/National Centers for Environmental Prediction, Camp Springs, Maryland.

 Current affiliation: NOAA/Storm Prediction Center, Norman, Oklahoma.

Corresponding author address: Dr. Carl E. Hane, National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.

Email: hane@nssl.nssl.uoknor.edu

1. Introduction

a. Dryline thunderstorms

The dryline, which in the United States is found most frequently in the Great Plains in the springtime months, is recognized to be an important factor in the initiation of severe thunderstorms. Rhea (1966), using synoptic-scale data, confirmed the dryline’s importance in thunderstorm initiation on a statistical basis through a climatological study. Bluestein and Parker (1993), using radar data over a 16-yr period, identified modes of isolated severe storm development along the dryline.

Forecasters of severe thunderstorms, in assessing the potential for storm initiation, take into consideration a number of factors including (a) the stability of the atmosphere (basically the magnitude of the lapse rate between low and high levels and the amount of available low-level moisture) in terms of its ability to support strong updrafts, (b) the presence of pronounced stable layers between low and midlevels that can prevent updrafts from penetrating above low to midlevels, and (c) areas where upward motion can be localized to aid in destabilization and carry moist low-level air to the condensation level with sufficient vigor to initiate deep moist convection. Examples of mechanisms that can produce localized upward motion include frontal circulations (e.g., see Palmén and Newton 1969), upper-level disturbances (e.g., Uccellini and Johnson 1979), dryline circulations (e.g., Koch and McCarthy 1982), outflow boundaries from previous convection (e.g., Gurka 1976), convergence areas resulting from orographic influences (e.g., Szoke et al. 1984; Crook et al. 1990), and circulations resulting from horizontal variation of heating and/or moistening properties of the underlying surface [e.g., sea breezes and circulations resulting from geographically varying land form or land use (Segal and Arritt 1992)].

On many springtime days in the central and southern Plains the only identifiable mechanism present to provide localized lifting is the dryline. This can be the case under synoptically “quiescent” conditions, when fronts and upper-level disturbances are absent, and when outflow boundaries from previous convection do not exist. Primary factors considered in forecasting deep convection then include the position and strength of the dryline and atmospheric stability [e.g., convective available potential energy (CAPE) and convective inhibition (CIN)] in the dryline area. Since convection is seldom initiated along the entire length of the dryline, an important forecasting problem is identifying preferred areas along the dryline for convective initiation. In some cases larger-scale features may dictate preferred locations (e.g., the dryline may be “capped” south of a certain latitude). In other cases the strength of the capping inversion may not be well known or it may be of marginal strength. Mechanisms that provide the needed localized upward motion in such situations might include mesoscale low pressure areas along the dryline (Bluestein et al. 1988), boundary layer circulations associated with cloud lines that intersect the dryline at an angle [e.g., boundary layer rolls (LeMone 1973) in the moist air], or gravity waves (Koch 1982).

b. Past work

Other studies have clearly shown the tendency for local development of thunderstorms along various types of boundaries [e.g., a front by Lewis et al. (1974), boundary intersection with “arc” clouds by Purdom (1976), a sea breeze front by Wakimoto and Atkins (1994), and internal gravity waves propagating from a gust front by Weckwerth and Wakimoto (1992)]. Investigations of convergence boundaries in Colorado have yielded new information on localization of convective initiation by boundary collisions (Wilson and Schreiber 1986) and by boundary layer horizontal convective rolls (Christian and Wakimoto 1989; Wilson et al. 1992).

Many examples of localized convective development along the dryline have been noted. McCarthy and Koch (1982) and Koch and McCarthy (1982) attributed perturbations along the dryline to gravity waves that emanated from an elevated source in the dry air. Investigations by Bluestein et al. (1989), Bluestein and Woodall (1990), and Bluestein and Parker (1993) all include examples of localized convective development along the dryline. In a study of tornadic storm development along a westward moving dryline Bluestein et al. (1988) argued that a local maximum in diabatic heating in the dry air led to pressure falls and backed winds near Canadian, Texas, where the storm formed. Bluestein et al. (1990) investigated the development of small cumulonimbi at the intersection of an outflow boundary and the dryline in southwestern Oklahoma. In that case, observations indicated local deepening of the moist layer owing to westward motion of the dryline just prior to convective initiation.

In an investigation of dryline structure near Midland, Texas, Parsons et al. (1991) reported that convection was initiated near and to the northeast of the intersection of a cold front and retrograding dryline. Schaefer (1986) has noted the intense convection that often develops as fronts approach/merge with the dryline. Sanders and Blanchard (1993) investigated a case in which storms were initiated over a very limited area of western Kansas just ahead of a dryline. A large area of potential instability existed over the entire region, but because of the presence of an extensive thermodynamic lid, convection developed only over a limited area where the lid was lifted. The importance of lid strength and lifting mechanisms was demonstrated in numerical experiments by Ziegler et al. (1996) that simulated convective initiation at the dryline on four different days in May 1991.

c. A field investigation

Field operations were conducted on 26 May 1991 during the Cooperative Oklahoma Profiler Studies—1991 (COPS-91) field program (Hane et al. 1993) to investigate the reason for localization of thunderstorm initiation along the dryline. Observing systems employed for this experiment included a research aircraft equipped with both in situ instrumentation and a Doppler radar, two mobile laboratories capable of rawinsonde release, a surface mesonetwork, the Profiler Demonstration Network (Chadwick 1988), and several ground-based Doppler radars. The observational array was first used to investigate the structure of the dryline before convection began and then to observe remotely with Doppler radar tornadic thunderstorms that developed along the dryline.

The experiment described here likely represents the most deliberate attempt to date to investigate the reasons for localized development of thunderstorms along the dryline. A key strategy in this attempt was to fly a research aircraft in a horizontal sawtooth pattern along a 190-km section of the dryline where conditions were thought to be favorable for thunderstorm development. In the section that follows, the observational systems utilized in the experiment will be described and an account will be given of the large-scale environment in which the mesoscale event took place. This will be followed by presentation of a representative vertical cross section through the dryline obtained by aircraft penetrations. Next, along-line variability in relation to thunderstorm initiation will be assessed through mobile sounding data, in situ aircraft data, surface mesonetwork data, and satellite measurements of several kinds. Finally, forecast fields for this day from the (currently) finest-scale operational mesoscale model will be discussed in relation to along-line variability.

2. Observing systems employed and the larger-scale environment

a. Observing systems

The primary experimental area for activities on this day included the Texas panhandle, western Oklahoma, and southwest Kansas. The primary observing platform was the National Oceanic and Atmospheric Administration P-3 research aircraft equipped with a large array of meteorological in situ sensors, an X-band Doppler radar, and a horizontally scanning C-band radar (Marks and Houze 1987). The aircraft was used (a) to locate the dryline by monitoring onboard sensor displays for moisture gradients, (b) for obtaining soundings through its altitude range, (c) for producing vertical cross sections by executing stepped traverse patterns, (d) for assessing horizontal structure and gradients in fields through flying various patterns at constant pressure, and (e) for collecting Doppler radar data from regions of precipitation. Two mobile laboratories from the National Severe Storms Laboratory (NSSL) were equipped with CLASS (Cross-chain Loran Atmospheric Sounding System) ballooning capability and continuously recording surface data systems (Rust et al. 1990). On 26 May these mobile labs were used to obtain series of soundings at 1-h intervals in the moist and dry air on either side of the dryline.

A network of the National Center for Atmospheric Research (NCAR) Portable Automatic Mesonetwork (PAM) stations (Brock et al. 1986) was deployed in the Texas panhandle and western Oklahoma (refer to Fig. 2a for locations). These stations filled large gaps within the National Weather Service (NWS) network of surface sites and provided 1-min resolution time series of standard meteorological variables as the dryline or other features translated over individual sites. Stations operating on this day within the Environmental Research Laboratories (ERL) Profiler Demonstration Network (Chadwick 1988) included three within the area of interest (all in Oklahoma). The profiler site closest to the area of interest was at Vici, Oklahoma, situated about 40 km east-southeast of the location of the initial radar echo of a storm that formed near the dryline. Satellite data of various kinds were very useful in this case, in which predominantly clear conditions early in the day yielded to fields of developing convective clouds in the afternoon. Both visible and infrared images from the GOES-7 geostationary satellite were available every 30 min. Higher-resolution images provided measurement of infrared surface temperature and normalized differential vegetation index (NDVI) (Rabin et al. 1990) and were available from the NOAA-11 polar-orbiting satellite with passes every 12 h.

b. Synoptic overview

The area of interest was largely uninfluenced by synoptic-scale disturbances. Figure 1a shows a regional 700-hPa (mb) analysis at 1200 UTC, about 9 h prior to the development of dryline thunderstorms. The flow over the region is generally west-southwesterly at 10–12 m s−1 and contains only very weak perturbations. A significant trough is far to the west near the California–Arizona border. The afternoon position of the dryline was in the eastern Texas panhandle along the eastern edge of a 700-mb thermal ridge that extended from west Texas northward to western Nebraska. During the day the thermal ridge strengthened, as can be seen from Fig. 1b, which shows the regional 700-mb analysis at 0000 UTC (27 May). This warming likely resulted from daytime heating of the high terrain in the southern Rockies coupled with horizontal advection and from subsidence east of the minor ridge (Fig. 1a) located in western New Mexico [note the 5°C increase at Albuquerque, New Mexico (ABQ)]. Carlson and Ludlam (1968) and Carlson et al. (1983) have examined the processes surrounding the development of the thermal lid resulting from advection of air heated over the Mexican plateau and southern Rockies. Thunderstorm development within such an environment usually requires intense heating of moist air near the surface and significant upward motion for ascending boundary layer parcels to break through the thermal lid. No synoptic-scale disturbances that might provide the necessary vertical motion were evident in the 0000 UTC analysis (Fig. 1b).

Higher in the atmosphere (300 mb) the morning (1200 UTC) analysis shows a quite uniform westerly to west-southwesterly flow of 15–25 m s−1 over the region (Fig. 1c). No major disturbances are evident in the flow. There is an axis of strong winds over central Arizona, central New Mexico, and the Texas panhandle; however, no pronounced jet streak is evident. By evening (0000 UTC 27 May) 300-mb winds over the area of primary interest (central New Mexico to central Oklahoma) weakened slightly (Fig. 1d), while flow to the north (Utah, Colorado, Kansas, and Nebraska) strengthened by 5–15 m s−1. These changes appear to result from a latitudinal shift in the strength of the geopotential gradient over a broad region rather than from the progression of wind maxima and minima within the flow.

Surface shelter-level (hereafter, just “surface”) conditions during the morning and afternoon over the region are depicted in Fig. 2 based upon information from NWS reporting stations and PAM sites. At midmorning (Fig. 2a) the dryline was located as shown, far west of its position later in the day (dryline position is analyzed here based on the eastern edge of the largest moisture gradient and does not necessarily follow a given isodrosotherm). An outflow boundary from earlier convection in Kansas was apparent in southwest Kansas. The air to the north of this boundary was decidedly cooler and at most locations drier.

Surface conditions at midday (1800 UTC) are depicted in Fig. 2b. The dryline advanced rapidly eastward in west Texas owing to strong surface heating and vertical mixing of the shallow moist layer [revealed by morning Amarillo sounding (not shown)]. Vertical mixing brings dry air with stronger westerly momentum toward the surface (Schaefer 1974), resulting in eastward dryline motion. The dryline advanced eastward also across the extreme southwest corner of Kansas, but in the northern Texas panhandle it remained farther to the west. The reason for lack of eastward motion in the latter area is likely due to retardation of heating (note slightly cooler surface temperatures), perhaps owing to land use (irrigation) practices in this area (skies were clear). There is a secondary area of east-to-west moisture decrease evident in the extreme eastern Texas panhandle ahead of the indicated dryline position. The location of sharpest moisture gradient remained farther west at 1800 UTC and is the reason for analyzing the dryline in the location shown. The outflow boundary in southwestern Kansas remained quasi-stationary and is identified by cooler and generally drier air to its north.

Surface moisture distribution and station plots at 2100 UTC (about 90 min after the first convective clouds formed along the dryline) are shown in Fig. 2c. Deep vertical mixing of moisture over virtually all of the Texas panhandle has occurred and the dryline is repositioned along the Texas–western Oklahoma border where a secondary sharp gradient had existed earlier (see Fig. 2b). Farther north the position of the dryline has shifted slightly either eastward or westward in several locations. The analyzed northward small-scale bulge of the dryline in the eastern Oklahoma panhandle is based upon aircraft measurements (described in next section).

At 0000 UTC the location of the dryline differs little from that 3 h earlier. The pronounced dryline bulge (again based upon aircraft measurements) noted on the 2100 UTC analysis has advanced into southwestern Kansas. Interestingly, the relatively moist area west of the dryline in the northern Texas panhandle is still evident. Moisture that had earlier (Fig. 2c) advected into southwestern Kansas from northwestern Oklahoma (moisture that might follow a more northerly trajectory if the outflow boundary were absent) was cut off by the mesoscale dryline bulge.

Profiler observations from the site at Vici, Oklahoma (V in Fig. 2d), provide a record of the changes in vector wind as a function of height during the period prior to convective initiation. The time–height section from Vici (Fig. 3) shows no evidence of pronounced wind maxima moving through the region. There is a slow increase in wind speed during the day in the 3–6-km layer and a slow decrease in the 1–2-km layer through 2200 UTC. Low-level winds back during the day, providing an improved environmental shear profile for supercell storms. Comparison of the profiler winds with the analyses in Fig. 1 shows good agreement except at 700 mb (0000 UTC) where profiler winds are considerably stronger. The Vici site was likely close enough to the thunderstorm complex that diversion of flow around the edges produced local increases in flow during the late afternoon. Observational evidence (Fankhauser 1971) strongly suggests that at least a partial obstacle effect is present in midlevels near isolated thunderstorms.

National Weather Service soundings from Amarillo, Texas (AMA), and Norman, Oklahoma (OUN, see Fig. 1b for locations), on the evening of 26 May (Fig. 4) illustrate the stratification in the dry and moist air, respectively, in the dryline region. At AMA (Fig. 4a) shallow low-level moisture present on the morning sounding (not shown) has mixed vertically, resulting in a mean mixing ratio in the lowest 100 mb of less than 5 g kg−1. The lapse rate is approximately dry adiabatic from the surface to 580 mb, consistent with vertical mixing through the layer, and wind speed and direction are quasi-uniform below 600 mb. In contrast, at OUN (Fig. 4b), east of the dryline, the mean mixing ratio in the lowest 100 mb exceeds 15 g kg−1. This moist layer is characterized by a strongly veering wind profile and is capped by a substantial inversion from 820 to 850 mb, consistent with the inhibition of deep convection except in areas of significant upward motion. Convective available potential energy in the air mass represented by this sounding is in the 2500–3000 J kg−1 range.

3. Representative finescale cross-dryline structure

a. Objectively analyzed fields

The P-3 aircraft on this day was used to help define mesoscale structures in and around the dryline. A series of vertically stacked legs quasi-perpendicular to the dryline (e.g., see Beebe 1958; Fujita 1958; McGuire 1962; NSSP Staff 1963; Ziegler and Hane 1993; Hane et al. 1993) was flown. Observations from this pattern were used to produce a vertical cross section of mesoscale dryline structure during the period 2158–2323 UTC in southwestern Kansas at the location noted by the dashed line in Fig. 2d. Some error in such analyses derives from the nonsteady nature of the flow coupled with the relatively long period needed to complete the pattern. However, the dryline was quasi-stationary during this time, and smoothing from the objective analysis removes variability on scales less than 1–2 km.

The distribution of water vapor mixing ratio (g kg−1) in the vertical cross section through the dryline is shown in Fig. 5a. The flight path of the aircraft is also indicated in the cross section (above terrain whose elevation averaged about 740 m). Data collected along the flight path were interpolated, using a Cressman weighting function, to grid points (spaced at 200 m in the horizontal and 100 m in the vertical) within the vertical cross section. The eastern edge of the dryline zone within the cross section is marked by the rapid east-to-west decrease in mixing ratio near the center of the figure. The depth of the moist layer (wherein mixing ratios decrease very little with height) east of the dryline is about 1 km. The billow or plume at the east edge of the dryline zone has been noted in other studies (e.g., Ziegler and Hane 1993) and has been hypothesized to result from a combination of vertical advection and turbulent mixing. The western edge of the dry bulge that extends northward into Kansas (as depicted in Fig. 2d) is revealed by higher mixing ratios in the left portion of the cross section. There is a gradual decrease in mixing ratio with height in the dry air, perhaps owing to entrainment of dry air from aloft accompanying rapid growth of the mixed layer (Mahrt 1976).

The east–west component of the wind (westerly wind is positive and earth-relative) in this cross section is shown in Fig. 5b. Strong convergence at the east edge of the dryline zone coincides with the large horizontal moisture gradient evident in Fig. 5a. The moist air east of the dryline is characterized by strong vertical shear, whose maximum is about 12 m s−1 over a 1200-m depth, and is primarily concentrated at the top of the moist layer (1.6–1.8 km). In the dry air westerly winds increase slowly with height.

The distribution of virtual potential temperature θv in Fig. 5c includes a horizontal gradient of θv (about 1 K over 8 km) at the dryline with lower values in the moist air. Therefore, as has been found in other cases (Parsons et al. 1991; Ziegler and Hane 1993), virtual potential temperature gradients are more strongly linked to temperature gradients than to moisture gradients. The θv gradient contributes toward development of a vertical mesoscale circulation (though this circulation may not be realized) normal to the dryline through solenoidal forcing with ageostrophic flow directed toward the heated air.

b. Derived fields

Also indicated in Fig. 5c are areas of ascent and descent. Vertical motion was derived kinematically from the horizontal dryline-normal wind field assuming zero vertical velocity at the lowest level and integrating the anelastic continuity equation upward. The largest area of upward motion (about 5 km wide) surrounded the region of maximum moisture gradient and contained the region exceeding 1 m s−1 indicated in Fig. 5c at X = 1–2 km. If the vertical motion and vapor mixing ratio fields are superimposed (not shown), the western edge of the primary region of upward motion (at X = −1 km) coincides with the western edge of the moist plume and the eastern edge of the upward motion area coincides with the location of the plume maximum (X = 4 km). It is therefore likely that the position and shape of the plume are strongly influenced by the upward motion. However, as is pointed out by Ziegler and Hane (1993), the airflow is primarily horizontal, implying that vertical turbulent mixing also plays a strong role in maintaining the plume. An area of downward motion approximately 6 km wide lies immediately to the east (downwind) adjacent to the area of upward motion. The western edge of the descent coincides with the moist plume peak and the eastern edge with the moisture “trough” (at X = 10 km).

The perturbation pressure field in the same vertical cross section was retrieved from the wind field following the methods outlined by Gal-Chen (1978) and Hane et al. (1981) Pressure gradients (Fig. 5d) were primarily horizontal with higher values on the moist (east) side of the dryline. Lower pressure west of the dryline and higher pressure to the east is consistent with the temperature structure coupled with quasi-hydrostatic balance. Near the top of the area of primary horizontal moisture gradient there is clearly a pressure couplet (denoted by “+” and “−” in Fig. 5d) that approximately straddles the area of primary updraft. This is interpreted to be a dynamically induced (nonhydrostatic) feature whose presence has been previously explained as an effect of ambient shear–updraft interaction by Rotunno and Klemp (1982) for the case of thunderstorm updrafts. The presence of this pressure couplet likely contributes significantly to vertical and horizontal accelerations in the area of strongest vertical motions at the dryline (e.g., downward acceleration at z = 1.8 km just west of the dryline zone’s eastern edge).

c. Initiation of convection

The foregoing examination of atmospheric fields points toward a conceptual structure much the same as that summarized by Ziegler and Hane (1993) for a northwestern Oklahoma dryline. The dryline is viewed as a mixing zone between dry and moist convective boundary layers where localized convergence produces vertical motion, and a combination of advection and turbulent mixing results in a nearly vertical plume of moist air extending from the surface to heights of several kilometers or more. A mesoscale circulation induced by horizontal density (virtual potential temperature) gradients is also intertwined in the redistribution of moisture and other variables. West winds carry moist air eastward from near the top of the plume, producing an elevated moist layer to the east of the dryline. The elevated moist layer is viewed as a potential source of moisture for the initial growth of clouds. If such clouds persist and produce precipitation, the potential exists for surface outflows and tapping of surface-based moisture in evolution of these clouds toward development of long-lasting storms.

The above-described mesoscale structure is presented to suggest that dryline vertical motions on this day were sufficient to produce convective clouds. Clouds and storms clearly did not develop all along the dryline. In the specific area where the flight pattern was flown from which vertical cross sections were derived (i.e., across the dryline bulge) a few deep convective clouds formed, but were short-lived. Therefore, it is concluded that other factors (e.g., moisture availability or strength of localized upward motion) must have varied along the dryline, and that influences favorable for cloud and storm development were superimposed upon the general structure in a few locations. In the section that follows a number of processes for localizing storm development along the dryline will be explored.

4. Along-dryline variability and storm initiation

a. Stratification

Sounding information along the dryline was quite limited on this day, since only two mobile laboratories were available and there was a need to have measurements in the dry air. Observations are therefore limited to hourly series of soundings in both dry and moist air at one point along the dryline. Additional sounding information in low levels was obtained through in situ instrumentation as the aircraft flew the vertical cross section pattern (Fig. 5). Locations of CLASS and aircraft in situ soundings are indicated in Fig. 2d as noted below.

A sounding in the dry air west of the dryline (location “a” in Fig. 2d) released at 2005 UTC is shown in Fig. 6a. Dewpoint depressions in low levels are in excess of 20°C, and the lapse rate is quasi-dry adiabatic from the surface to above 700 mb, consistent with turbulent mixing through a deep vertical layer. In Fig. 6b stratification in the moist air east of the dryline is shown (location “b” in Fig. 2d; note that at sounding release times locations b, c, and d were all east of the dryline) from a sounding released at the same time. The low-level moist layer extends upward to near 750 mb at the base of a stable layer. Above that conditions are very similar in the two soundings. As the dryline advanced slowly eastward, sounding release locations were adjusted to maintain 25–40-km distance from it. A sounding in the moist air 1 h later (Fig. 6c) reveals that moist air is present above the 700-mb level where it had been very dry. The dry-adiabatic lapse rate and quasi-constant mixing ratio between about 820 and 700 mb are indicative of turbulent mixing.

One interpretation of this structure is that moist air from low levels is advected and mixed upward at the dryline, then advected eastward while continuing to undergo vertical mixing. The layer below 850 mb may then be air that has not circulated upward at the dryline, but is maintained by horizontal flow with a near-zero east–west component. The increase in moisture aloft may have also been influenced by sounding releases located farther east at successive times; however, the distances between sites are relatively small. One hour later (Fig. 6d) the well-mixed moist layer is found at a slightly higher level, indicative of continued vertical mixing. Unfortunately, winds were missing from the soundings taken in the moist air, hindering interpretation of changes from one hour to the next. The upward growth of the moist layer is reemphasized in Fig. 7, in which relative humidity is plotted versus pressure for a series of four soundings east of the dryline. Examination of visible satellite images indicates that extensive cumulus cloud formation along the sector of the dryline sampled by these soundings began at about 2030 UTC. This coincides with the period of the most rapid increase in height of the elevated moist layer top (2004–2103 in Fig. 7). Though the dryline in this area appears to have been influencing the development of cumulus, there was apparently no organizing influence along this sector that would allow for long-lived deep convection.

A low-level sounding was obtained by the P-3 aircraft between about 2200 and 2315 UTC at the location marked “P” in Fig. 2d, just east of the pronounced bulge of the dryline in southwest Kansas. This abbreviated sounding is plotted in Fig. 6d with the CLASS sounding obtained at nearly the same time. Both soundings are estimated to have been taken about 25 km east of the dryline. Low-level temperatures are several degrees cooler than in the sounding obtained about 230 km to the south-southwest (38 km east of Shamrock, Texas). The dramatic difference, however, is in the moisture profile, which indicates much more low-level moisture in the northern location just ahead of the dryline. Based upon late afternoon surface winds at the stations nearest to this area [Dodge City, Kansas (DDC, see Fig. 2c), Medicine Lodge, Kansas (P28), and Gage, Oklahoma (GAG)], it appears that winds maintained a more easterly component here than farther south along the dryline. Thus, higher dewpoint air to the east could be advected westward to the dryline in the northern area. In addition, the northern location may have been near enough to the outflow boundary that pooling of moisture along the boundary contributed to the higher values. A comparison of sounding winds in the two areas is inconclusive because of missing data. The vertical cross section in Fig. 5b indicates an easterly component extending 700–800 m above the surface. The CLASS sounding in the Shamrock area from the same time period contains a 2–3 m s−1 easterly component in the lowest 350 m, but winds are missing above that. In terms of development of deeper convection, the northern area was more active. The timing and locations of these storms are discussed in the next section.

b. Development of cloud fields

Convective clouds began forming along the Texas–Oklahoma portion of the dryline between 1900 and 2000 UTC. A regional GOES-7 visible image at 2001 UTC is shown in Fig. 8a. In the eastern Texas panhandle there is a broken line of cumulus oriented approximately north-northeast to south-southwest along the dryline and during the period that moisture gradients were intensifying in that area. There is also another line of convective clouds (high-based) at an angle to the dryline extending southwestward from the northeast corner of the Texas panhandle. The extensive field of cumulus in the lower left portion of the figure is west of the dryline. No sustained deep convection developed in this region. The thunderstorm in southeastern Colorado (outside the area of intensive observation) appears to have developed from a cluster of clouds between 1800 and 1900 UTC in an area where the dryline was retreating westward.

The visible image at 2101 UTC is shown in Fig. 8b. Some eastward motion of the dryline in the eastern Texas panhandle is indicated by changes in the convective cloud field. The secondary convective cloud line that extends southwestward from the northeast corner of the Texas panhandle was more evident at this time. The line was about 300 km long, was solid to broken near its northeastern end, and contained scattered convective clouds farther west. Based upon limited sounding information west of the dryline, it is estimated that the bases of these clouds well west of the dryline were about 3.5–4 km above the surface.

Satellite images at later times reveal that the majority of convective clouds with small anvils that were north of the intersection of the secondary line and the dryline (Fig. 8b) had lifetimes on the order of 1 h. These small thunderstorms are represented in Fig. 8c by storm tracks labeled C, D, and E. The storm represented by track F was long-lasting and evolved into a severe cluster (severe reports are noted). The timing of convective initiation tends to be later farther north, leading to the impression that low-level convergence on the edge of the dry bulge was the initiating mechanism. Unfortunately, there are no data to fix the location of the bulge nose. It is possible that the storm represented by track F was initiated when the bulge nose encountered the remnants of the outflow boundary (Fig. 2) and was sustained owing to stronger low-level convergence in its environment. A thunderstorm that was initiated in southeast Colorado benefited from the pool of moisture in southwestern Kansas, and produced a tornado (Dowell 1996, personal communication) in the area southwest of Garden City (GCK). It is likely that this storm’s dissipation resulted from the cutoff of low-level moisture by the dryline bulge.

The cloud cluster at the intersection of the dryline and secondary cloud line (track B) produced radar detectable precipitation at 2115 UTC, reflectivity of 52 dBZ at 2130 UTC, and eventually a tornadic supercell. The motion represented by track B is a simplification of the complicated evolution that occurred. The original cell appears to have undergone multiple splits and new cells formed on the forward flank outflow (as revealed by WSR-88D radar).

c. Along-line differences in across-line gradients

A horizontal sawtooth pattern (ground-relative) was flown along a 215-km segment of the dryline at 860 mb between 2020 and 2155 UTC. The track of the aircraft through the Texas panhandle and northwestern Oklahoma within this pattern is shown in Fig. 9. Superimposed upon the track are wind speed and direction at 1-min intervals and the schematic position of the dryline as defined by locations of the approximate points along each leg where the dewpoint began to decrease rapidly (1.5°C in 20 km) to the west (scalloped) and where it leveled off again (dashed). Also shown on the plot is the approximate location of the secondary cloud line’s axis. Examination of the wind field reveals confluence between the moist and dry air all along the dryline (i.e., winds are generally southerly ahead of the dryline and southwesterly well behind it). The transition in wind direction occurs generally in the zone of maximum moisture gradient, but is not extremely abrupt as in the case of some drylines (Crawford and Bluestein 1995, personal communication). The east–west leg centered on 2100 UTC (across the cloud line) appears to contain the largest gradient in east–west wind. Speed convergence (again across the cloud line) is apparent in the north–south component of the wind between the 2101–2104 and the 2119–2122 UTC segments of east–west legs.

To examine in more detail the distribution of water vapor along this section of the dryline a subjective analysis of dewpoint temperature (°C) was carried out from the aircraft data (approximately 860 mb) and is shown in Fig. 10a. An additional leg (2239–2250) has been extracted from the stacked traverse pattern and added to the basic pattern to extend the analysis farther north. Dewpoints were averaged over 40-s intervals once per minute for the analysis. There is likely some distortion of the pattern compared to an instantaneous one owing to dryline motion and evolution (most potential for this is in the area of the northernmost leg). The decrease of moisture from east to west is seen to take place in either one or two steps (generally one step in the northern and two in the southern half). The northernmost three east–west legs are sufficiently long that the increase in vapor along the western edge of the mesoscale bulge is part of the analysis. Farther south there is a very dry area (dewpoints less than 3°C) that is likely an eastward extending finger of the dry air entering the central Texas panhandle, as depicted in Fig. 2c at the surface.

Since the dry air is just west of the surface mesonetwork station at Canadian (Fig. 10a), a time plot of dewpoint temperature at that location is instructive. In Fig. 10b dewpoint plots at Canadian and McLean (farther south, Fig. 10a) imply that the depiction of the very dry area as a local feature is correct. The sharp decrease in dewpoint from approximately 11° to 6°C from 2102 to 2121 UTC at Canadian does not occur at McLean. This small-scale dry “push” could not have reached northwestern Oklahoma in time to influence storm initiation, but could have later affected the evolution of existing storms.

The distribution of temperature from the 860-mb sawtooth pattern is shown in Fig. 11a. Temperatures are 2°–3°C warmer along the western edge of the dryline moisture gradient compared to values east of the dryline, consistent with more rapid surface and boundary layer heating along with enhanced turbulent mixing west of the dryline. The warmest air generally coincides with areas of pronounced dryness; for example, note that the warmest region in the southwestern part of the domain is collocated with the area of lowest dewpoints. The dry bulge extending north-northeastward into Kansas is clearly also warmer than its surroundings.

Virtual potential temperature over the region is plotted in Fig. 11b. Consistent with the findings of Parsons et al. (1991) and Ziegler and Hane (1993), there is a 1°–2°C change in virtual potential temperature over tens of kilometers across the dryline with warmer air on the dry side. Gradients are largest in the south portion of the analyzed domain and along the northern edge in Kansas. No direct relation between the along-dryline variation of across-line θv gradients and the location of storm development is evident. The decrease in θv to the northwest of the warm bulge places the western portion (in this figure) of the cloud line along the warm axis, oriented normal to the θv gradient. Therefore, a solenoidal circulation produced by this temperature gradient would place rising motion over the cloud line region.

An analysis of the horizontal divergence field (10−4 s−1) is shown in Fig. 11c. Also included are the locations of the dryline and secondary cloud line. The divergence field was estimated by time averaging (over 40 s) the u and v (west-to-east and south-to-north, respectively) wind components, subjectively analyzing each component in the horizontal plane, and graphically differentiating the component fields. This procedure has the effect of smoothing small-scale features such that wavelengths of 40 km and greater are emphasized, consistent with the north–south separation of east–west legs. Convergence is apparent along the entire length of the dryline zone and tends to locate in the area of maximum moisture gradient. The leading edge of the dryline is generally in a weakly convergent region. Double maxima in convergence in the two southernmost east–west legs coincide with the double moisture gradients. The maximum convergence in the northernmost east–west leg results from a very sharp change in the east–west wind component.

In the center of the domain, a strongly convergent region is aligned east-northeast to west-southwest, consistent with the location and orientation of the secondary cloud line. Most of the convergence in this region is due to the v component, as is emphasized in Fig. 12, which shows the v-component time series as the aircraft flew northward across the cloud line. A decrease in v component of 4.5 m s−1 over one minute is indicated. The picture that emerges from these observations is one where convective clouds develop along the cloud line in dry air and are advected across the dryline zone through a strongly convergent region. They encounter increasing low-level moisture as they move east-northeastward, and upon reaching the deep moisture east of the dryline they grow vigorously. The existence of the upward motion associated with the secondary cloud line appears to have been a key element in determining at what location along the dryline the most vigorous sustained deep convection formed.

d. Along-line variation in boundary layer character

Another possible influence on storm development lies in the character of the boundary layer along the dryline, both in the moist and dry air. It is conceivable, for example, that in certain locations, perturbations in the boundary layer might be more vigorous or larger than at other locations, thereby enhancing the development of deep convective clouds. To examine this question aircraft data from the sawtooth pattern are examined, with emphasis on vertical velocity measurements.

Vertical air motions from aircraft measurements require knowledge of the vertical motion of the aircraft relative to the ground. The latter variable was in this case calculated using an algorithm described by Jorgensen and LeMone (1989). Vertical air motions over several sections of the aircraft track in the moist air are shown in Fig. 13. The 9-min history shown in Fig. 13a is ahead of the dryline near the intersection of the dryline and cloud line, while Fig. 13b includes measurements taken ahead of the dryline about 120 km to the south-southwest (exact locations may be ascertained by reference to Fig. 9). Visual inspection reveals higher amplitude velocities in the north location; moreover, calculation of w2 yields values of 0.88 and 0.56 m2 s−2 for the north and south segments, respectively. A Fourier analysis of each segment (not shown) yields a relative peak in the spectrum at about 3.6 km (horizontal wavelength) for the north segment and 1.9 km for the south segment. There are stronger and larger perturbations in vertical velocity along the north segment perhaps because the north and south segments generally lie in convergent and divergent flow areas (see Fig. 11c), respectively. Convergence may lead to a deepening of the boundary layer, and it has been noted (Kaimal and Finnigan 1993) that perturbations producing deviations from a −5/3 slope (as is the case here) within the inertial subrange of the kinetic energy spectrum will scale with the boundary layer depth.

e. The mesoscale upward motion producing the cloud line

It is clear that the intersection point of the dryline and cloud line was a preferred location for development of deep convection. From a nowcasting point of view, it is of great interest to understand what mechanism produced the cloud line. The observing strategies employed on this day were not designed to examine the cloud line environment directly, since its existence was not evident to field personnel in real time. Therefore, the reason for the cloud line’s development cannot be determined with certainty. However, several possibilities were explored based upon the data sources available and findings here described.

It is possible that the development of the mesoscale upward motion that produced the cloud line is related to the vegetative and radiative temperature properties of the underlying surface. Segal and Arritt (1992) have reviewed the various causes of “nonclassical mesoscale circulations” (NCMCs) that include spatial differences in such factors as soil wetness and vegetation. The basic premise in these cases is that there are from various causes perturbed areas where the sensible heat flux to the atmosphere is sufficiently different from that in surrounding areas that temperature and pressure differences arise, resulting in the development of mesoscale circulations in vertical planes. In modeling studies these circulations are characterized by the strongest upward motion over the sharpest gradients and weaker sinking motion over surrounding broad areas. In investigations using mesoscale models, Chang and Wetzel (1991) and Ziegler et al. (1995, 1996) have explored the effect of variations in vegetation and soil moisture on the development of environments for convection.

Irrigated areas during the growing season are cited by Segal and Arritt (1992) as typical examples of regions for NCMC development. In the region of interest for this study concentrated areas of irrigation are found in the northern and southwestern portions of the Texas panhandle. One measure of the effects of this irrigation is in the density of green vegetation, which is proportional to the normalized differential vegetation index, available from satellite measurements. Rabin et al. (1990) have used NDVI data to assess the effects of landscape variability on development of convective clouds.

An NDVI image was produced from an individual overpass of a polar-orbiting satellite on 27 May at about 2100 UTC and is shown in Fig. 14. Cloud cover precludes such measurements and appears as a dark streak extending from lower left to upper right in the image. The granular relatively bright areas in southwestern Kansas and northern and southwestern portions of the Texas panhandle are agricultural fields, many of which are irrigated based on the map shown by Segal and Arritt (1992). The relatively darker elongated area that is oriented west-southwest to east-northeast across the center of the Texas panhandle is the broad Canadian River valley whose terrain and soil are generally not suitable for cropland. Much of the latter area is characterized by sandy soil, sagebrush, and low-growing evergreens.

Surface radiative temperature, which can also be measured remotely, is strongly related to the character (vegetation, soil type, etc.) of that surface. The surface temperature on 26 May was estimated in clear-sky regions from the received infrared radiance in the 11-μm channel on the GOES-7 geostationary satellite. It should be noted that the difference in surface temperature between areas with large differences in vegetative cover can be underestimated by as much as a few degrees.

In Fig. 15 the distribution of surface temperature from infrared satellite measurements at two times during the afternoon of 26 May is shown. The shading produced by rectangular areas of various brightness levels indicates the basic temperature value, whereas the contours have been added to provide enhanced recognition of gradients and absolute values. No contours are included near areas containing cloud (indicated by white), since either total or partial cloud cover within a given area results in cloud-top temperature mixture with the surface temperature measurement. Significant gradients in surface temperature are indicated; for example, the irrigated areas of the northern panhandle are considerably cooler than the more strongly heated Canadian River valley (differences of 4°C are typical). Cooler surface temperatures are evident at 1900 UTC (Fig. 15a) to the south of the Canadian Valley in the central panhandle. By 2100 UTC (Fig. 15b) a broken line of convective clouds has formed (indicated by brighter and cooler areas) in the central panhandle and extending to the northeast to intersect the dryline at the previously noted location.

The surface temperature undoubtedly strongly influences the temperature of the air just above, as is demonstrated, for example, by the retardation of eastward dryline motion in the northern Texas panhandle at 1800 UTC (Fig. 2b). It is possible that the irrigation and enhanced transpiration in the northern panhandle impedes sensible heating and destabilization, and thereby slows the vertical mixing that is essential for eastward dryline motion. It is also possible that a cause of the cloud line is upward motion that is part of a mesoscale circulation that results from spatial differences in vertical momentum transport. In this connection, Segal (1995, personal communication) has noted that convergence generated by modifying the background flow is all that is needed (changes in wind direction are not necessary). Mesoscale circulations between areas of contrasting surface temperature have been documented over cropland in northeastern Colorado by Segal et al. (1989).

Intense surface heating immediately to the northwest of the cloud line might allow for enhanced vertical mixing that results in surface winds with more westerly and less southerly component. Farther southeast, under this scenario, winds would mix vertically (but less efficiently) and have more southerly component, thereby producing convergence somewhere between the two regions. Further evidence of upward motion on the mesoscale is provided by the trace of water vapor mixing ratio as the aircraft flew northward (Fig. 9) across the cloud line (Fig. 16). Mixing ratios increase from a mean of 6 g kg−1 south of the cloud line to 7.5 g kg−1 to the north. The abrupt increase takes place over an estimated distance of 800 m. Higher mixing ratios to the north are consistent with rising air bringing higher mixing ratios from lower altitudes in an air current with southerly component. A valid concern is whether the upward motion, which would be strongest in the lower boundary layer, would extend sufficiently high to produce clouds along the western portion of the cloud line, where cloud bases are estimated to have been 3.5–4 km above the ground. Factors working in favor of cloud development here were the apparent local and intense nature of the low-level convergence, and the dry-adiabatic lapse rate below cloud base, wherein rising thermals initiated near the surface would not be impeded.

The aircraft patterns, deployment of sounding teams, etc., were planned for the purpose of maximizing information on cross-dryline gradients and varying conditions along the dryline itself rather than for probing features such as the linear mesoscale feature that produced the cloud line. Indeed, the presence and significance of the cloud line were not realized during the course of data collection. A stepped traverse pattern normal to the cloud line, for example, would have yielded valuable information on its structure and origin. Therefore, the link between the differential heating of varied landscapes and the development of the cloud line cannot be absolutely established.

f. Predictions of an operational mesoscale model

To explore the possible influence of the larger scale on the development of the cloud line and variability along the dryline, and more generally, to assess the ability of an operational model to predict the presence, location, and movement of the dryline on this day, a mesoscale model was employed. The eta model (Black 1988; Mesinger et al. 1988), developed at the National Meteorological Center (currently the National Centers for Environmental Prediction) was run after the fact using archived data for initialization. A mesoscale version of this model has been used operationally for guidance in mesoscale forecasts since mid-1994.

The configuration of the mesoscale eta model that was used in this investigation is nearly identical to that described by Black (1994). The model structure includes a grid network with 29-km resolution in the horizontal and 50 layers in the vertical over a domain that covers most of North America. The dynamical and physical processes within the model are basically as described by Black (1994), including the explicit cloud water parameterization (Zhao et al. 1991) that was made part of the operational mesoscale code. The initial conditions for this case were obtained by interpolating the initial fields used in the Nested Grid Model (NGM) run of 1200 UTC 26 May 1991. Boundary conditions were obtained by direct interpolation of the aviation run of the global spectral Medium-Range Forecast Model. The model was integrated to 36 h and output obtained via a postprocessor (Treadon 1993) on the model grid, as well as on a 0.5° latitude–longitude grid with 50-mb vertical spacing.

Model output fields at 1800 UTC 26 May and 0000 UTC 27 May (6- and 12-h forecasts, respectively) are shown in Fig. 17. Water vapor mixing ratio (g kg−1) is plotted for the lowest model atmospheric layer over the region of interest in order to identify the location and movement of the dryline. The model cannot resolve the actual gradients associated with the dryline, but concentrated east–west gradients of moisture are evident at both times depicted in Fig. 17. The implied dryline position and motion are reasonably accurate (by comparison with Fig. 2) with the exception of its retardation at midday in the northern Texas panhandle and its movement northeastward through southwestern Kansas during the afternoon. A possible reason for the latter forecast error is that the model initial conditions may have failed to adequately take into account the morning outflow boundary in Kansas. Airflow near this boundary consistently maintained an easterly component during the day, resulting in relatively cool, moist air over southwestern Kansas. In terms of aiding in forecasting storm development, the model position of the dryline was likely helpful in northwest Oklahoma, but misleading in southwestern Kansas.

The vertical motion at 750 mb predicted by the eta model for 2200 UTC is shown in Fig. 18. Contours are for the omega field (10−2 Pa s−1); therefore, a −2 label corresponds to approximately +2 cm s−1. The forecast vertical motion does indicate weak ascent over the dryline, with slightly larger values over southwestern Kansas. There appears to be no obvious relation between the distribution of model-predicted vertical motion and the location of the cloud line in the Texas panhandle. This result also holds for lower levels in the model output (not shown). If the existence of the cloud line were dependent upon larger-scale forcing, it would follow that the cloud line would be located entirely within an area of upward motion. Therefore, it appears that scales smaller than that resolved by the model were responsible for the cloud line circulation (assuming that the model forecast was accurate). It should also be noted that the version of the model used here does include predictions of temperature and moisture within one soil layer, but vegetative effects are excluded. Therefore, the model should not be expected to capture the heating patterns noted in the previous section.

There were in the model forecasts weak upper-level wind maxima that contributed to the patterns of upward motion over the area (e.g., see Murray and Daniels 1953; Shapiro 1982). These wind maxima moved over Vici in the forecasts, so that a limited comparison with observations can be made. For example, the model predicted a decrease in wind speed as wind maxima at 750 and 850 mb moved to the northeast of Vici during the afternoon. Profiler observations at Vici indicated steady or perhaps slightly increasing wind speeds during the same time period, casting doubt on the presence of these maxima over this area. Comparison of model and observed wind directions (not shown) at 500, 750, and 850 mb between 1800 and 0000 UTC showed good agreement except at 850 mb, where the model indicated little deviation from 200°, while the profiler measured pronounced backing from 225° to 180°. This backing may have been in response to lowering pressure to the northwest that occurred owing to boundary layer physical processes not accounted for in the model. Similarly, the previously mentioned outflow boundary over southwestern Kansas did not appear to be maintained in the model prediction, perhaps owing to insufficient resolution of boundary layer processes.

5. Summary and discussion

a. Conclusions

A dryline over the south-central Plains on 26 May 1991 provided an environment for the development of severe storms. Analysis of mesoscale data from various observing systems on this day has yielded insight into why storms formed at a particular location along the dryline, but not at others. On the larger scale, there were no apparent upper-air disturbances present to help explain along-dryline regional differences in convective activity. Synoptic surface analysis indicated highly variable motion of the dryline at different along-line locations, including in one area a pronounced mesoscale bulge of dry air toward the northeast. Deep convection was initiated along the east edge of the bulge, but was primarily short-lived. The differences in dryline motion appear to have been at least partially due to variations in radiative temperature of the underlying surface west of the dryline.

Satellite images indicated development of convective clouds along the dryline in the early afternoon. A short time later a second cloud line developed that intersected the dryline at a 45° angle in the northeast Texas panhandle and extended southwestward for 300 km into the dry air. It was just east of the cloud line–dryline intersection that deep convection was sustained. Convective clouds that developed along the secondary line in the dry air appear to have advected across the dryline zone and grew rapidly in the moist environment just east of the dryline. The development of upward motion that produced the cloud line therefore appears to have been a key factor in determining the location of thunderstorm development along the dryline.

Along-line differences in variable fields that might influence storm development were investigated by research aircraft through execution of a sawtooth pattern flown at 860 mb along a 200-km section of the dryline. There was no obvious relation between the strength of across-line moisture or virtual potential temperature gradients and the location of storm development. The horizontal divergence field included convergence at all locations along the dryline. Maximum convergence was west of the dryline leading edge and coincided with the location of the cloud line. A south-to-north aircraft leg across the cloud line yielded a dramatic drop in south-to-north wind. Thus, maximum convergence along the dryline in this case resulted from the superposition of convergence in the along-line (north–south) direction on the dryline across-line convergence. Inspection of vertical air motion time series and spectra obtained from the aircraft revealed that perturbations were larger and occurred on longer spatial scales in an area east of the dryline–cloud line intersection than in an area farther to the south.

Two explanations were explored for the cause of the mesoscale upward motion that produced the cloud line intersecting the dryline: 1) the upward motion was caused by radiative temperature characteristics of the underlying surface; 2) the upward motion was related to mesoscale features not evident on the synoptic scale such as upper-level wind maxima. Examination of available radiative data from satellites resulted in a potential scenario that links surface characteristics through vertical momentum transport arguments to the development of the cloud line. Without more data that directly probes the processes involved, the validity of the scenario is uncertain. The possibility of an upper-level mesoscale feature producing the needed upward motion was explored using a mesoscale version of the eta model. The model results showed some skill in predicting dryline motion and a broad area of upward motion where deep convective elements were in general observed to be more numerous, but indicated no distinct upward motion coincident with the cloud line.

b. Implications for forecasting and future research

It appears that forecasting the location of thunderstorm development along the dryline in cases similar to the one investigated here would be very difficult using current tools. The development of a northeastward bulge in the dryline may have been related to mesoscale processes producing the cloud line that intersected the dryline and likely led to initiation of deep convection farther north. This bulge, however, was not detectable by the synoptic surface network [the current Oklahoma Mesonetwork (Brock et al. 1995) likely would have resolved it had it occurred over Oklahoma]. The cloud line itself was observed by satellite, but less than 1 h before thunderstorm initiation. Determination of whether the suggested link of cloud lines and mesoscale bulges to surface effects is valid will depend upon future observational and modeling research that focuses more specifically on spatial distribution of radiative effects, vertical momentum transport, and boundary layer vertical motion. Additional studies of this kind combined with the accumulation of forecaster experience (if such a link is verified and is acting frequently) should yield a recognition of geographical areas and seasons prone to this type of development. The relation of land use to thunderstorm development should be explored vigorously because of its forecasting implications and since the issue of inadvertent weather modification is involved.

A mesoscale model would likely be a helpful tool in predicting the location of convective initiation, but only if its horizontal resolution were on the order of 10 km and if it could account for boundary layer processes in some detail. If indeed radiative properties related to land form and use are important in many cases of thunderstorm initiation, it will be necessary to include such factors as soil moisture, crop type and distribution, cloud cover, and added vertical resolution in the boundary layer. Enhanced resolution and representation of physical processes should help in maintaining outflow boundaries and dryline motion (including mesoscale bulges) in model predictions. Experiments of this kind have been carried out recently in a research mode by Ziegler et al. (1996). A model that incorporates small-scale data in real time such as the Local Analysis and Prediction System (Albers et al. 1996) might be used in the future in an operational mode.

Acknowledgments

We are grateful to M. Segal and R. Wakimoto for comments and suggestions on an earlier version of this paper. D. Jorgensen provided software needed for the work, and both he and B. Smull contributed valuable suggestions. The accomplishment of field experiments was made possible by scientific crews on the P-3 aircraft and in M-CLASS vehicles, including the following individuals: D. Bartels, H. Bluestein, M. Douglas, S. Fredrickson, C. George, C. Hane, D. Jorgensen, S. Koch, W. Martin, J. Meitin, T. Shepherd, L. Showell, T. Shuur, B. Smull, M. Stolzenburg, G. Stumpf, A. Watson, Q. Xu, and C. Ziegler. We are grateful to the flight crew of NOAA’s Office of Aircraft Operations for their support. M. Shapiro and M. Douglas provided valuable suggestions in the planning phase of experiments. Thanks are given to J. O’Bannon for preparation of figures. The PAM network was provided by the Atmospheric Technology Division of NCAR, which is sponsored by the National Science Foundation (NSF). PAM received partial support from the National Aeronautics and Space Administration through the efforts of S. Koch. A portion of the work was supported by Grants ATM-9019821 and ATM-9302379 to the University of Oklahoma from the NSF.

REFERENCES

  • Albers, S. C., J. A. McGinley, D. L. Birkenheuer, and J. R. Smart, 1996: The Local Analysis and Prediction System (LAPS): Analysis of clouds, precipitation, and temperature. Wea. Forecasting,11, 273–287.

  • Beebe, R. G., 1958: An instability line development as observed by the tornado research airplane. J. Meteor.,15, 278–282.

  • Black, T. L., 1988: The step-mountain eta coordinate regional model: A documentation. NOAA/NWS Rep., 47 pp. [Available from the National Centers for Environmental Prediction, Room 204, 5200 Auth Road, Camp Springs, MD 20746.].

  • ——, 1994: The new NMC mesoscale eta model: Description and forecast examples. Wea. Forecasting,9, 265–278.

  • Bluestein, H. B., and G. R. Woodall, 1990: Doppler-radar analysis of a low-precipitation severe storm. Mon. Wea. Rev.,118, 1640–1664.

  • ——, and S. Parker, 1993: Modes of isolated convective storm formation along the dryline. Mon. Wea. Rev.,121, 1354–1372.

  • ——, E. W. McCaul Jr., G. P. Byrd, and G. R. Woodall, 1988: Mobile sounding observations of a tornadic storm near the dryline: The Canadian, Texas, storm of 7 May 1986. Mon. Wea. Rev.,116, 1790–1804.

  • ——, ——, ——, ——, G. Martin, S. Keighton, and L. C. Showell, 1989: Mobile sounding observations of a thunderstorm near the dryline: The Gruver, Texas, storm complex of 25 May 1987. Mon. Wea. Rev.,117, 244–250.

  • ——, ——, ——, R. L. Walko, and R. P. Davies-Jones, 1990: An observational study of splitting convective clouds. Mon. Wea. Rev.,118, 1359–1370.

  • Brock, F., G. Saum, and S. Semmer, 1986: Portable Automated Mesonet II. J. Atmos. Oceanic Technol.,3, 573–582.

  • ——, K. C. Crawford, R. L. Elliott, G. W. Cuperus, S. J. Stadler, H. L. Johnson, and M. D. Eilts, 1995: The Oklahoma mesonet: A technical overview. J. Atmos. Oceanic Technol.,12, 5–19.

  • Carlson, T. N., and F. H. Ludlam, 1968: Conditions for the development of severe local storms. Tellus,20, 203–226.

  • ——, S. G. Benjamin, G. S. Forbes, and Y.-F. Li, 1983: Elevated mixed layers in the regional severe storm environment: Conceptual model and case studies. Mon. Wea. Rev.,111, 1453–1473.

  • Chadwick, R., 1988: The wind profiler demonstration network. Preprints, Symp. on Lower Tropospheric Profiling: Needs and Technologies, Boulder, CO, Amer. Meteor. Soc., 109–110.

  • Chang, J.-T., and P. J. Wetzel, 1991: Effects of spatial variations of soil moisture and vegetation on the evolution of a prestorm environment: A numerical case study. Mon. Wea. Rev.,119, 1368–1390.

  • Christian, T. W., and R. M. Wakimoto, 1989: The relationship between radar reflectivities and clouds associated with horizontal roll convection on 8 August 1982. Mon. Wea. Rev.,117, 1530–1544.

  • Crook, N. A., T. L. Clark, and M. W. Moncrieff, 1990: The Denver cyclone. Part I: Generation in low Froude number flow. J. Atmos. Sci.,47, 2725–2742.

  • Fankhauser, J. C., 1971: Thunderstorm environment interactions determined from aircraft and radar observations. Mon. Wea. Rev.,99, 171–192.

  • Fujita, T. T., 1958: Structure and movement of a dry front. Bull. Amer. Meteor. Soc.,39, 574–582.

  • Gal-Chen, T., 1978: A method for the initialization of the anelastic equations: Implications for matching models with observations. Mon. Wea. Rev.,106, 587–606.

  • Gurka, J. J., 1976: Satellite and surface observations of strong wind zones accompanying thunderstorms. Mon. Wea. Rev.,104, 1484–1493.

  • Hane, C. E., R. B. Wilhelmson, and T. Gal-Chen, 1981: Retrieval of thermodynamic variables within deep convective clouds: Experiments in three dimensions. Mon. Wea. Rev.,109, 564–576.

  • ——, C. L. Ziegler, and H. B. Bluestein, 1993: Investigation of the dryline and convective storms initiated along the dryline: Field experiments during COPS-91. Bull. Amer. Meteor. Soc.,74, 2133–2145.

  • Jorgensen, D. P., and M. A. LeMone, 1989: Vertical velocity characteristics of oceanic convection. J. Atmos. Sci.,46, 621–640.

  • Kaimal, J. C., and J. J. Finnigan, 1993: Atmospheric Boundary Layer Flows—Their Structure and Measurement. Oxford University Press, 289 pp.

  • Koch, S. E., 1982: The role of apparent gravity waves as a triggering mechanism for severe thunderstorms along a dryline. Preprints, 12th Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 573–576.

  • ——, and J. McCarthy, 1982: The evolution of an Oklahoma dryline. Part II: Boundary layer forcing of mesoconvective systems. J. Atmos. Sci.,39, 237–257.

  • LeMone, M. A., 1973: The structure and dynamics of horizontal roll vortices in the planetary boundary layer. J. Atmos. Sci.,30, 1077–1091.

  • Lewis, J. M., Y. Ogura, and L. Gidel, 1974: Large-scale influences upon the generation of a mesoscale disturbance. Mon. Wea. Rev.,102, 545–560.

  • Mahrt, L., 1976: Mixed layer moisture structure. Mon. Wea. Rev.,104, 1403–1407.

  • Marks, F., Jr., and R. Houze, 1987: Inner core structure of Hurricane Alicia from airborne Doppler radar observations. J. Atmos. Sci.,44, 1296–1317.

  • McCarthy, J., and S. E. Koch, 1982: The evolution of an Oklahoma dryline. Part I: A mesoscale and subsynoptic-scale analysis. J. Atmos. Sci.,39, 225–236.

  • McGuire, E. L., 1962: The vertical structure of three drylines as revealed by aircraft traverses. NSSP Rep. 7, 11 pp. [Available from National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.].

  • Mesinger, F., Z. I. Janjic, S. Nickovic, D. Gavrilov, and D. G. Deaven, 1988: The step-mountain coordinate: Model description and performance for cases of Alpine lee cyclogenesis and for a case of Appalachian redevelopment. Mon. Wea Rev.,116, 1493–1518.

  • Murray, R., and S. M. Daniels, 1953: Transverse flow in the entrance and exit to jet streams. Quart. J. Roy. Meteor. Soc.,79, 236–241.

  • NSSP Staff Members, 1963: Environmental and thunderstorm structures as shown by National Severe Storms Project observations in the spring of 1960 and 1961. Mon. Wea. Rev.,91, 271–292.

  • Palmén, E., and C. W. Newton, 1969: Atmospheric Circulation Systems: Their Structure and Physical Interpretation. International Geophysics Series, Vol.13, Academic Press, 603 pp.

  • 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.

  • Purdom, J. F. W., 1976: Some uses of high-resolution GOES imagery in the mesoscale forecasting of convection and its behavior. Mon. Wea. Rev.,104, 1474–1483.

  • Rabin, R. M., S. Stadler, P. J. Wetzel, D. J. Stensrud, and M. Gregory, 1990: Observed effects of landscape variability on convective clouds. Bull. Amer. Meteor. Soc.,71, 272–280.

  • Rhea, J. O., 1966: A study of thunderstorm formation along drylines. J. Appl. Meteor.,5, 58–63.

  • Rotunno, R., and J. B. Klemp, 1982: The influence of shear-induced pressure gradient on thunderstorm motion. Mon. Wea. Rev.,110, 136–151.

  • Rust, W. D., D. W. Burgess, R. A. Maddox, L. Showell, T. Marshall, and D. Lauritzen, 1990: Testing a mobile version of a Cross-chain LORAN Atmospheric (M-CLASS) Sounding System. Bull. Amer. Meteor. Soc.,71, 173–180.

  • Sanders, F., and D. O. Blanchard, 1993: The origin of a severe thunderstorm in Kansas on 10 May 1985. Mon. Wea. Rev.,121, 133–149.

  • Schaefer, J. T., 1974: A simulative model of dryline motion. J. Atmos. Sci.,31, 956–964.

  • ——, 1986: The dryline. Mesoscale Meteorology and Forecasting, P. S. Ray, Ed., Amer. Meteor. Soc., 549–572.

  • Segal, M., and R. W. Arritt, 1992: Nonclassical mesoscale circulations caused by surface sensible heat-flux gradients. Bull. Amer. Meteor. Soc.,73, 1593–1604.

  • ——, W. E. Shreiber, G. Kallos, J. R. Garratt, A. Rodi, J. Weaver, and R. A. Pielke, 1989: The impact of crop areas in northeast Colorado on midsummer mesoscale thermal circulations. Mon. Wea. Rev.,117, 809–825.

  • Shapiro, M. A., 1982: Mesoscale weather systems of the central United States. Cooperative Institute for Research in Environmental Sciences Rep., 78 pp. [Available from National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.].

  • Sutherland, R. A., 1986: Broadband and spectral emissivities (2–18 mm) of some natural soils and vegetation. J. Atmos. Oceanic Technol.,3, 199–202.

  • Szoke, E. J., M. L. Weisman, J. M. Brown, F. Caracena, and T. W. Schlatter, 1984: A subsynoptic analysis of the Denver tornadoes of 3 June 1981. Mon. Wea. Rev.,112, 790–808.

  • Treadon, R. E., 1993: The NMC eta model post processor: A documentation. NMC Office Note 394, 44 pp. [Available from the National Centers for Environmental Prediction, NOAA Science Center, Room 101, 5200 Auth Road, Camp Springs, MD 20746.].

  • Uccellini, L. W., and D. R. Johnson, 1979: The coupling of upper and lower tropospheric jet streaks and implications for the development of severe convective storms. Mon. Wea. Rev.,107, 682–703.

  • Wakimoto, R. M., and N. T. Atkins, 1994: Observations of the sea-breeze front during CaPE. Part I: Single-Doppler, satellite, and cloud photogrammetry analysis. Mon. Wea. Rev.,122, 1092–1114.

  • Weckwerth, T. M., and R. M. Wakimoto, 1992: The initiation and organization of convective cells atop a cold-air outflow boundary. Mon. Wea. Rev.,120, 2169–2187.

  • Wilson, J. W., and W. E. Schreiber, 1986: Initiation of convective storms by radar-observed boundary layer convergent lines. Mon. Wea. Rev.,114, 2516–2536.

  • ——, 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.

  • Zhao, Q., F. H. Carr, and G. B. Lesins, 1991: Improvement of precipitation forecasts by including cloud water in NMC’s eta model. Preprints, Ninth Conf. on Numerical Weather Prediction, Denver, CO, Amer. Meteor. Soc., 50–53.

  • 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, 1996: Simulations of deep, moist convection at the dryline with a 3-D mesoscale model. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 273–277.

Fig. 1.
Fig. 1.

Upper-air analyses at (a) 700 mb, 1200 UTC 26 May 1991; (b) 700 mb, 0000 UTC 27 May 1991; (c) 300 mb, 1200 UTC 26 May 1991; and (d) 300 mb, 0000 UTC 27 May 1991. Solid contours are heights (dam), dashed contours are temperatures (°C), and station models use conventional format (long wind barb, 5 m s−1; short barb, 2.5 m s−1; and flag, 25 m s−1).

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 2.
Fig. 2.

Surface dewpoint (°C) analyses at (a) 1500 UTC 26 May 1991, (b) 1800 UTC 26 May 1991, (c) 2100 UTC 26 May 1991, and (d) 0000 UTC 27 May 1991. Dryline denoted by scalloped line, outflow boundary by dash–dotted line, and PAM stations by solid squares. Station models use conventional format with sea level pressure (mb and tenths above 1000) at upper right.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 3.
Fig. 3.

Wind profiler observations from Vici, Oklahoma (circled “V” in Fig. 2d), on 26 May 1991. Wind barbs have conventional format.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 4.
Fig. 4.

Skew T plots of National Weather Service soundings at 0000 UTC 27 May 1991 from (a) Amarillo, Texas, and (b) Norman, Oklahoma.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 5.
Fig. 5.

Vertical cross-sectional analyses from aircraft pattern flown from 2158 to 2323 UTC 26 May 1991 at the location marked by the dashed line in Fig. 2d. (a) Water vapor mixing ratio (g kg−1, solid) and aircraft track (dotted); (b) west-to-east component of the wind (m s−1) with positive values solid and negative dashed; (c) virtual potential temperature (excess above 310 K, solid) and shaded areas of vertical motion (see key); and (d) deviation perturbation pressure (10−1 mb) with positive contours solid, negative dashed, and 10 g kg−1 vapor mixing ratio heavy dotted contour.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 6.
Fig. 6.

Mobile-CLASS soundings at the times and locations noted in (a)–(d). Sounding locations are also noted in Fig. 2d. In (d) an aircraft sounding (thin solid lines) from the location in Fig. 2d denoted by circled “P” is included.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 7.
Fig. 7.

Relative humidity (%) as a function of pressure (mb) from mobile-CLASS soundings at the four noted times east of the dryline near Shamrock, Texas.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 8.
Fig. 8.

Visible images centered on the northeastern Texas panhandle from GOES-7 satellite on 26 May 1991 at (a) 2001 and (b) 2101 UTC.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 8.
Fig. 8.

(Continued) (c) Dryline locations at 2100 and 0000 UTC and tracks of storms initiated along dryline with severe reports (T—tornado and H—hail) are indicated.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 9.
Fig. 9.

Aircraft track from 2023 to 2152 UTC 26 May 1991 with wind speed and direction noted at 1-min intervals. Scalloped and dashed lines enclose area of dryline primary moisture gradients. Dash–dot–dot line indicates location of cloud line noted in text; “C” and “M” denote respective locations of Canadian and McLean PAM sites.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 10.
Fig. 10.

Dewpoint distribution in the dryline region based on (a) aircraft observations at 860 mb and (b) PAM time series. In (a) dewpoint temperature (°C) shown by solid contours is derived from aircraft measurements along track shown. Dryline is indicated by scalloped line and cloud line by dash–dot–dot line; location of first echo also noted.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 11.
Fig. 11.

Analyzed fields from aircraft sawtooth pattern at 860 mb; dryline and cloud line locations indicated as in Fig. 10a. Variable fields are (a) temperature (°C), (b) virtual potential temperature (K), and (c) horizontal divergence (10−4 s−1). In (c) divergence values less than −1.5 × 10−4 s−1 are shaded.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 12.
Fig. 12.

South-to-north wind component (m s−1) along a 10-min segment of the aircraft track whose location (dashed) is noted in Fig. 11c.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 13.
Fig. 13.

Aircraft-measured vertical air motion (m s−1) along 9-min segments of the aircraft pattern ahead of the dryline (a) in the area where thunderstorms developed and (b) 120 km to the south-southwest.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 14.
Fig. 14.

NDVI image from the NOAA-11 polar-orbiting satellite at 2105 UTC 27 May 1991. Brighter areas indicate higher vegetative density. Black areas are clouds or bodies of surface water.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 15.
Fig. 15.

Surface radiative temperature from GOES-7 infrared image on 26 May 1991 at (a) 1901 UTC and (b) 2101 UTC. Darker shades indicate warmer temperatures; white areas are cloud. Smoothed contours of temperature (°C) are included in cloud-free areas.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 16.
Fig. 16.

Water vapor mixing ratio (g kg−1) from aircraft measurements along the 10-min segment of track whose location is indicated in Fig. 11c.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 17.
Fig. 17.

Water vapor mixing ratio (g kg−1) in the lowest atmospheric layer from the mesoscale eta model at (a) 1800 UTC 26 May 1991 and (b) 0000 UTC 27 May 1991.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Fig. 18.
Fig. 18.

Vertical motion (10−2 Pa s−1) at 750 mb from the mesoscale eta model at 2100 UTC. The location of the observed cloud line is indicated by the dash–dotted line.

Citation: Monthly Weather Review 125, 2; 10.1175/1520-0493(1997)125<0231:STDIRT>2.0.CO;2

Save