• Anthes, R. A., and T. T. Warner, 1978: Development of hydrodynamical models suitable for air pollution and other mesometeorological studies. Mon. Wea. Rev.,106, 1045–1078.

  • Atkins, N. T., R. M. Wakimoto, and C. L. Ziegler, 1998: Observations of the finescale structure of a dryline during VORTEX 95. Mon. Wea. Rev.,126, 525–550.

  • Barnes, S. L., F. Caracena, and A. Marroquin, 1996: Extracting synoptic-scale diagnostic information from mesoscale models: The eta model, gravity waves, and quasigeostrophic diagnostic. Bull. Amer. Meteor. Soc.,77, 519–528.

  • Blackadar, A. K., 1979: High resolution models of the planetary boundary layer. Advances in Environmental Science and Engineering, J. R. Pfafflin and E. N. Ziegler, Eds., Vol. 1, Gordon and Breach, 50–85.

  • Bluestein, H. B., 1986: Fronts and jet streaks: A theoretical perspective. Mesoscale Meteorology and Forecasting, P. S. Ray, Ed., Amer. Meteor. Soc., 173–215.

  • ——, and T. M. Crawford, 1997: Mesoscale dynamics of the near-dryline environment: Analysis of data from COPS-91. Mon. Wea. Rev.,125, 2161–2175.

  • Braun, S. A., and R. A. Houze Jr., 1997: The evolution of the 10–11 June PRE-STORM squall line: Initiation, development of rear inflow, and dissipation. Mon. Wea. Rev.,125, 478–504.

  • Browning, K. A., 1990: Organization of clouds and precipitation in extratropical cyclones. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. O. Holopainen, Eds., Amer. Meteor. Soc., 129–153.

  • Byers, H. R., and R. R. Braham Jr., 1949: The Thunderstorm. U.S. Govt. Printing Office, 287 pp.

  • Carbone, R. E., 1982: A severe frontal rainband. Part I: Stormwide hydrodynamic structure. J. Atmos. Sci.,39, 258–279.

  • ——, J. W. Conway, N. A. Crook, and M. W. Moncrieff, 1990: The generation and propagation of a nocturnal squall line. Part I: Observations and implications for mesoscale predictability. Mon. Wea. Rev.,118, 26–49.

  • Castle, J. A., J. D. Locatelli, J. E. Martin, and P. V. Hobbs, 1996: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part IV: The evolution of a drytrough on 8–9 March 1992. Mon. Wea. Rev.,124, 1591–1595.

  • Colman, B. R., 1990a: Thunderstorms above frontal surfaces in environments without positive CAPE. Part I: A climatology. Mon. Wea. Rev.,118, 1103–1121.

  • ——, 1990b: Thunderstorms above frontal surfaces in environments without positive CAPE. Part II: Organization and instability mechanisms. Mon. Wea. Rev.,118, 1123–1144.

  • Crook, N. A., and M. W. Moncrieff, 1988: The effects of large-scale convergence on the generation and maintenance of deep moist convection. J. Atmos. Sci.,45, 3606–3624.

  • ——, 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.

  • Doswell, C. A., III, 1987: The distinction between large-scale and mesoscale contribution to severe convection: A case study example. Wea. Forecasting,2, 3–16.

  • Dudhia, J., 1993: A nonhydrostatic version of the Penn State–NCAR mesoscale model: Validation tests and simulation of an Atlantic cyclone and cold front. Mon. Wea. Rev.,121, 1493–1513.

  • Grell, G. A., J. Dudhia, and D. R. Stauffer, 1994: A description of the fifth-generation Penn State/NCAR mesoscale model (MM5). NCAR Tech. Note NCAR/TN-398+STR, National Center for Atmospheric Research, Boulder, CO, 138 pp. [Available from National Center for Atmospheric Research, Information Support Services, P.O. Box 3000, Boulder, CO 80307-3000.].

  • Hobbs, P. V., and P. O. G. Persson, 1982: The mesoscale and microscale structure and organization of clouds and precipitation in midlatitude cyclones. Part V: The substructure of narrow cold-frontal rainbands. J. Atmos. Sci.,39, 280–295.

  • ——, J. D. Locatelli, and J. E. Martin, 1990: Cold fronts aloft and the forecasting of precipitation and severe weather east of the Rocky Mountains. Wea. Forecasting,5, 613–626.

  • ——, ——, and ——, 1996: A new conceptual model for cyclones generated in the lee of the Rocky Mountains. Bull. Amer. Meteor. Soc.,77, 1169–1178.

  • Hoskins, B. J., I. Draghici, and H. C. Davies, 1978: A new look at the ω-equation. Quart. J. Roy. Meteor. Soc.,104, 31–38.

  • ——, M. E. McIntyre, and R. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc.,111, 877–946.

  • Huschke, R. E., Ed., 1959: Glossary of Meteorology. Amer. Meteor. Soc., 638 pp.

  • Kain, J. S., and J. M. Fritsch, 1993: Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 165–170.

  • Keyser, D., and T. N. Carlson, 1984: Transverse ageostrophic circulations associated with elevated mixed layers. Mon. Wea. Rev.,112, 2465–2478.

  • ——, B. D. Schmidt, and D. G. Duffy, 1992: Quasigeostrophic vertical motions diagnosed from along- and cross-isentrope components of the Q vector. Mon. Wea. Rev.,120, 731–741.

  • Koch, S. E., and J. McCarthy, 1982: The evolution of an Oklahoma dryline. Part II: Boundary-layer forcing of meso-convective systems. J. Atmos. Sci.,39, 237–257.

  • ——, A. Aksakal, and J. T. McQueen, 1997: The influence of mesoscale humidity and evapotranspiration fields on a model forecast of a cold frontal squall line. Mon. Wea. Rev.,125, 58–83.

  • Lanicci, J. M., and T. T. Warner, 1997: A case study of lid evolution using analyses of observational data and a numerical model simulation. Wea. Forecasting,12, 228–252.

  • Lemone, M. A., M. Zhou, C.-H. Moeng, D. H. Lenschow, L. J. Miller, and R. L. Grossman, 1999: An observational study of wind profiles in the baroclinic convective mixed layer. Bound.-Layer Meteor.,90, 47–82.

  • Locatelli, J. D., J. E. Martin, J. A. Castle, and P. V. Hobbs, 1995a: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part III: The development of a squall line associated with weak cold-frontogenesis aloft. Mon. Wea. Rev.,123, 2641–2662.

  • ——, ——, and P. V. Hobbs, 1995b: Development and propagation of precipitation cores on cold fronts. Atmos. Res.,38, 177–206.

  • ——, M. T. Stoelinga, R. D. Schwartz, and P. V. Hobbs, 1997: Surface convergence induced by cold fronts aloft and pre-frontal surges. Mon. Wea. Rev.,125, 2808–2820.

  • ——, ——, and P. V. Hobbs, 1998: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part V: Thermodynamic and dual-Doppler radar analysis of a squall line associated with a cold front aloft. Mon. Wea. Rev.,126, 860–875.

  • Lynn, B. H., W.-K. Tao, and P. J. Wetzel, 1998: A study of landscape-generated deep moist convection. Mon. Wea. Rev.,126, 928–942.

  • Martin, J. E., 1999: Quasigeostrophic forcing of ascent in the occluded sector of cyclones and the trowal airstream. Mon. Wea. Rev.,127, 70–88.

  • ——, J. D. Locatelli, P. V. Hobbs, P.-Y. Wang, and J. A. Castle, 1995:Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part I: A synoptic-scale rainband associated with a lee dryline and lee trough. Mon. Wea. Rev.,123, 241–264.

  • Miller, L. J., M. A. LeMone, W. Blumen, R. L. Grossman, N. Gamage, and R. J. Zamora, 1996: The low-level structure and evolution of a dry arctic front over the central United States. Part I: Mesoscale observations. Mon. Wea. Rev.,124, 1648–1675.

  • Neiman, P. J., and R. M. Wakimoto, 1999: The interaction of a Pacific cold front with shallow air masses east of the Rocky Mountains. Mon. Wea. Rev.,127, 2102–2127.

  • ——, F. M. Ralph, M. A. Shapiro, B. F. Smull, and D. Johnson, 1998:An observational study of fronts and frontal mergers over the continental United States. Mon. Wea. Rev.,126, 2521–2554.

  • Newton, C. W., 1950: Structure and mechanism of the prefrontal squall line. J. Meteor.,7, 210–222.

  • Ogura, Y., H.-M. Juang, K.-S. Zhang, and S.-T. Soong, 1982: Possible triggering mechanisms for severe storms in SESAME-AVE IV (9–10 May 1979). Bull. Amer. Meteor. Soc.,63, 503–515.

  • 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.

  • Reisner, J., R. M. Rasmussen, and R. T. Bruintjes, 1998: Explicit forecasting of supercooled liquid water in winter storms using the MM5 mesoscale model. Quart. J. Roy. Meteor. Soc.,124, 1071–1107.

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

  • Rotunno, R., J. B. Klemp, and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci.,45, 463–485.

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

  • Shapiro, M. A., 1982: Mesoscale weather systems of the central United States. CIRES Tech. Rep., 78 pp. [Available from Cooperative Institute in Environmental Sciences, University of Colorado/NOAA, Boulder, CO 80309].

  • Shaw, B. L., R. A. Pielke, and C. L. Ziegler, 1997: A three-dimensional numerical simulation of a Great Plains dryline. Mon. Wea. Rev.,125, 1489–1506.

  • Steenburgh, W. J., and C. F. Mass, 1994: The structure and evolution of a simulated Rocky Mountain lee trough. Mon. Wea. Rev.,122, 2740–2761.

  • Sun, W.-Y., and C.-C. Wu, 1992: Formation and diurnal variation of the dryline. J. Atmos. Sci.,49, 1606–1619.

  • Sutcliffe, R. C., 1947: A contribution to the problem of development. Quart. J. Roy. Meteor. Soc.,73, 370–383.

  • Thorpe, A. J., M. J. Miller, and M. W. Moncrieff, 1982: Two-dimensional convection in nonconstant shear: a model of midlatitude squall lines. Quart. J. Roy Meteor. Soc.,108, 739–762.

  • Trier, S. B., D. B. Parsons, and J. E. H. Clark, 1991: Environment and evolution of a cold-frontal mesoscale convective system. Mon. Wea. Rev.,119, 2429–2455.

  • Wang, P. Y., J. E. Martin, J. D. Locatelli, and P. V. Hobbs, 1995: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part II: Arctic fronts. Mon. Wea. Rev.,123, 1328–1344.

  • Weisman, M. L., W. C. Skamarock, and J. B. Klemp, 1997: The resolution dependence of explicitly modeled convective systems. Mon. Wea. Rev.,125, 527–548.

  • 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.

  • ——, and C. K. Mueller, 1993: Nowcasts of thunderstorm initiation and evolution. Wea. Forecasting,8, 113–131.

  • Zhang, D. L., and R. A. Anthes, 1982: A high-resolution model of the planetary boundary layer—Sensitivity tests and comparisons with SESAME-79 data. J. Appl. Meteor.,21, 1594–1609.

  • Ziegler, C. L., T. J. Lee, and R. A. Pielke Sr., 1997: Convective initiation at the dryline: A modeling study. Mon. Wea. Rev.,125, 1001–1026.

  • View in gallery

    (a) Surface analysis valid at 0600 UTC 9 Mar 1992, with sea level pressure (solid contours, hPa), frontal positions (heavy solid lines), dryline position (heavy dashed line), and station data plotted according to the standard convention (temperature and dewpoint in °C). (b) The 500-hPa analysis valid at 0500 UTC 9 Mar 1992, with geopotential heights (thin solid contours, dam), temperature (heavy solid contours, °C), and station data plotted according to the standard convention. Dashed line indicates the leading edge of strong cold advection

  • View in gallery

    Infrared satellite image valid at 0400 UTC 9 Mar 1992, with an overlaid contour plot of radar reflectivity from a composite of low-elevation scans of all radars in the vicinity. Solid and dashed contours represent reflectivity values of 20 and 40 dBZ, respectively. Precipitation features between white arrows and black arrows are the CFA rainband and the pre–dry trough rainband, respectively

  • View in gallery

    Results from the 18-h model simulation on the 25-km domain, valid at 0600 UTC 9 Mar 1992. (a) Surface analysis of sea level pressure (solid contours, hPa), with subjectively analyzed frontal positions (heavy solid lines) and dryline position (heavy dashed line). (b) The 500-hPa analysis with geopotential heights (thin solid contours, dam) and temperature (heavy solid contours, °C). Dashed line indicates the leading edge of strong cold advection

  • View in gallery

    Simulated infrared satellite image based on temperature at optical depth = 1 (temperature scale shown at bottom) from the 16-h forecast on the 8.3-km domain, valid at 0400 UTC 9 Mar 1992, with an overlaid contour plot of simulated reflectivity factor based on rain and snow mixing ratios averaged through the lowest 3 km from the model simulation. Solid and dashed contours represent reflectivity values of 10 and 30 dBZ, respectively. Precipitation features between white arrows and black arrows are the simulated CFA rainband and pre–dry trough rainband, respectively

  • View in gallery

    The Q-vector plot at 750 hPa valid at 0000 UTC 9 Mar 1992, based on output from the 75-km model simulation at 12 h. Gray-shaded bands labeled C and P are locations of initiation of the CFA and pre–dry trough rainbands, respectively, in the model simulation. Also shown in (a) are geopotential height (heavy solid contours, dam) and temperature (thin solid contours, °C); and in (b), Q-vector divergence (contour interval of 1 × 10−14 m kg−1 s−1, with heavy dashed, thin solid, and heavy solid contours denoting positive divergence, zero divergence, and convergence, respectively). Note, upward motion is implied by convergence (heavy solid contours)

  • View in gallery

    Quasigeostrophic vertical velocity (ω) partitions at 750 hPa valid at 0000 UTC 9 Mar 1992, based on output from the 75-km model simulation at 12 h [contour interval of 1 dPa s−1, with heavy dashed, thin solid, and heavy solid contours denoting positive (downward), zero, and negative (upward) ω, respectively]. Gray-shaded bands labeled C and P are locations of initiation of the CFA and pre–dry trough rainbands, respectively, in the model simulation. Separate panels depict QG vertical velocity due to (a) Q-vector forcing, (b) topographic lower boundary condition, (c) Ekman lower boundary condition, and (d) the sum of (a), (b), and (c)

  • View in gallery

    Ertel potential vorticity (solid contours, PVU) and wind vectors (scale at upper right) on the 310-K isentropic surface valid at 0000 UTC 9 Mar 1992, based on output from the 75-km model simulation at 12 h. Gray-shaded bands labeled C and P are locations of initiation of the CFA and pre–dry trough rainbands, respectively, in the model simulation, and L is the location of the simulated low pressure center at the surface

  • View in gallery

    Vertical cross section valid at 0500 UTC 9 Mar from the 8.3-km model simulation at 17 h, along the line A–B in Fig. 4, showing potential temperature (thin solid contours, K) and precipitation (rain + snow) mixing ratio (heavy solid lines, contour interval 0.1 g kg−1). Circulation vectors in the plane of the cross section shown in (a) and (c) (scale at lower right). Gray shading is equivalent potential temperature in (a), CAPE in (b), and Ertel potential vorticity in (c); units and shading values are shown underneath each panel. Also shown in (b) are two 16-h trajectories from 1300 UTC 8 Mar to 0500 UTC 9 Mar. Arrowheads mark hourly positions. Trajectories terminate exactly in the cross section, and remained within 150 km of the cross section throughout the trajectory integration period. Heavy dashed contours in (c) are cloud water mixing ratio (contour interval 0.1 g kg−1)

  • View in gallery

    Vertical cross section valid at 0100 UTC 9 Mar from the 8.3-km model simulation at 13 h, along the line C–D in Fig. 4, showing potential temperature (thin solid contours, K), circulation vectors in the plane of the cross section (scale at lower right), and equivalent potential temperature (gray shading, legend at bottom of figure). Also shown are the surface location of the dryline (D) and the two baroclinic zones described in the text (heavy dashed lines marked BZ1 and BZ2)

  • View in gallery

    Vertical cross section from the 8.3-km model simulation, along the line C–D in Fig. 4, at (a) 9, (b) 13, and (c) 17 h into the simulation, showing potential temperature (thin solid contours, K), the along-cross-section component of the gradient of potential temperature (gray shading, legend at bottom), net ascent of 3-h trajectories ending at every grid point below approximately 700 hPa (hPa, medium contours, dashed are positive/descent, solid are negative/ascent), and precipitation (rain + snow) mixing ratio (heavy solid lines, contour interval 0.1 g kg−1). Also shown are the surface location of the dryline (D) and the two baroclinic zones described in the text (heavy dashed lines marked BZ1 and BZ2)

  • View in gallery

    Vertical cross section valid at 0100 UTC 9 Mar from the 8.3-km model simulation at 13 h, along the line C–D in Fig. 4, showing potential temperature (thin solid contours, K), circulation vectors in the plane of the cross section (scale at lower right), and CAPE (gray shading, legend at bottom). Also shown are the surface location of the dryline (D) and the two baroclinic zones described in the text (heavy dashed lines marked BZ1 and BZ2)

  • View in gallery

    Vertical cross section valid at 0100 UTC 9 Mar from the 8.3-km model simulation at 13 h, along the line C–D in Fig. 4, showing potential temperature (thin solid contours, K) and vertical velocity (ω) fields [contour interval of 1 dPa s−1, with heavy dashed, thin solid, and heavy solid contours denoting positive (downward), zero, and negative (upward) ω, respectively]. (a) The QG vertical velocity based on the QG diagnosis from the 75-km simulation; (b) full vertical velocity from the 8.3-km simulation, unsmoothed; and (c) full vertical velocity from the 8.3-km simulation, smoothed to the same scale as the input fields in the 75-km QG diagnosis

  • View in gallery

    Vertical cross sections from the 8.3-km model simulation (a) along line C–D in Fig. 4 at 13 h and (b) along line C–D in Fig. 4 at 17 h, showing potential temperature (thin solid contours, K), cloud water mixing ratio (heavy dashed lines, contour interval 0.1 g kg−1), circulation vectors in the plane of the cross section (scale at lower right), and Ertel potential vorticity (gray shading, legend at bottom)

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 27 27 5
PDF Downloads 11 11 3

Structure and Evolution of Winter Cyclones in the Central United States and Their Effects on the Distribution of Precipitation. Part VI: A Mesoscale Modeling Study of the Initiation of Convective Rainbands

View More View Less
  • 1 Department of Atmospheric Sciences, University of Washington, Seattle, Washington
© Get Permissions
Full access

Abstract

A cyclonic storm that moved over the central United States on 8–9 March 1992 developed two convective rainbands, namely, a pre–dry trough rainband and a cold front aloft (CFA) rainband. This study extends the results of previous investigations of these two rainbands by examining their initiation with the use of a nested-grid mesoscale model simulation with spatial resolution down to 8.3 km. The model simulation reproduced the synoptic-scale setting in which the rainbands developed, as well as the mesoscale processes that initiated the rainbands.

The pre–dry trough rainband was produced by the gradual ascent of a convectively unstable airstream above a gently sloping warm-frontal zone east of the dryline. After sufficient lifting, the instability was released through upright convection. The gradual ascent is well estimated by quasigeostrophic diagnosis, but the location and timing of the rainband are very sensitive to the convective stability characteristics within the airstream.

The CFA rainband was initiated by a Pacific cold front that occluded with the warm-frontal surface. This mesoscale occlusion process produced a narrow region of enhanced ascent at the dryline, which resulted in the lifting of the western edge of an air mass with high convective available potential energy. The lower-tropospheric mesoscale occlusion process was not resolved by a quasigeostrophic vertical velocity diagnosis. Also, although an upper-level front and tropopause fold were present, the CFA was separate from that feature.

Corresponding author address: Peter V. Hobbs, Department of Atmospheric Sciences, University of Washington, Box 351640, Seattle, WA 98195-1640.

Email: phobbs@atmos.washington.edu

Abstract

A cyclonic storm that moved over the central United States on 8–9 March 1992 developed two convective rainbands, namely, a pre–dry trough rainband and a cold front aloft (CFA) rainband. This study extends the results of previous investigations of these two rainbands by examining their initiation with the use of a nested-grid mesoscale model simulation with spatial resolution down to 8.3 km. The model simulation reproduced the synoptic-scale setting in which the rainbands developed, as well as the mesoscale processes that initiated the rainbands.

The pre–dry trough rainband was produced by the gradual ascent of a convectively unstable airstream above a gently sloping warm-frontal zone east of the dryline. After sufficient lifting, the instability was released through upright convection. The gradual ascent is well estimated by quasigeostrophic diagnosis, but the location and timing of the rainband are very sensitive to the convective stability characteristics within the airstream.

The CFA rainband was initiated by a Pacific cold front that occluded with the warm-frontal surface. This mesoscale occlusion process produced a narrow region of enhanced ascent at the dryline, which resulted in the lifting of the western edge of an air mass with high convective available potential energy. The lower-tropospheric mesoscale occlusion process was not resolved by a quasigeostrophic vertical velocity diagnosis. Also, although an upper-level front and tropopause fold were present, the CFA was separate from that feature.

Corresponding author address: Peter V. Hobbs, Department of Atmospheric Sciences, University of Washington, Box 351640, Seattle, WA 98195-1640.

Email: phobbs@atmos.washington.edu

1. Introduction

The term convective rainband, as used in this study, refers to a band of precipitation that is convective in nature, but synoptic scale in length because it is organized by fronts or baroclinic zones associated with midlatitude cyclones. Convective rainbands are a common precipitation feature in the central United States, particularly in winter and spring. Hobbs et al. (1996) described a conceptual model for cyclones in the central United States, referred to as the Structurally Transformed by Orography Model (STORM), and have described two characteristic types of convective rainbands that are closely connected to the frontal structure in this conceptual model. These are (i) the pre–dry trough rainband, which forms when the potentially unstable layer of air east of the dryline is lifted to saturation in a gradually ascending airstream within a warm frontal zone that intersects the surface at the dryline and slopes toward the northeast; and (ii) the CFA rainband, which forms as a cold front moves over the Rocky Mountains and occludes with the warm-frontal structure east of the dryline, generating a region of lifting at the location of the cold front aloft (CFA) within the warm occluded structure. Although the STORM model has its roots in studies of cyclones in the Genesis of Atlantic Lows Experiment on the East Coast in 1986, many of the components of the STORM conceptual model were synthesized from studies of cyclones observed during the Storm-scale Operational and Research Meteorology—Fronts Experiment Systems Test (STORM-FEST) in the central United States in 1992, and, in particular, from studies of the case of 8–9 March 1992. The latter case produced robust examples of both a pre–dry trough rainband (Martin et al. 1995) and a CFA rainband (Locatelli et al. 1995a, 1998). Therefore, we chose the 8–9 March 1992 case for the detailed numerical modeling study described in the present paper.

A complete examination of the relationships between the baroclinic structures within STORM-type cyclones and attendant convective rainbands is beyond the scope of a single modeling study. Thus, the specific focus of this paper is the relationship between the synoptic-scale baroclinic evolution of the 8–9 March 1992 storm and the initiation of the pre–dry trough rainband and the CFA rainband.

The initiation of convection is a complex scientific and forecasting problem that has been studied extensively, particularly with regard to convective storms in the central United States. It has been shown that a wide variety of meteorological structures and processes can provide the lifting necessary to initiate deep convection. Indeed, convection often initiates from a combination of lifting mechanisms acting on different time and space scales. Doswell (1987) has provided clear and practical definitions of the terms large scale and mesoscale. Large-scale processes are those that are well understood in terms of quasigeostrophic (QG) theory. We will use the term synoptic scale interchangeably with large scale. Mesoscale processes occupy a scale between large scale and microscale, and cannot be understood without considering processes on both of those other scales. Because frontal boundaries associated with midlatitude cyclones are considered to be synoptic scale in length but mesoscale in cross-front structure (Bluestein 1986), we define frontal-scale structures and processes to be that subset of mesoscale structures and processes that are specifically associated with frontal boundaries. Convective scale is the scale of individual convective storms, approximately equivalent to meso-γ scale.

In the absence of strong synoptic- or frontal-scale forcing, convective storms can usually be traced to preexisting meso-β or meso-γ-scale convergence lines within the boundary layer (Byers and Braham 1949; Purdom 1976; Wilson and Schreiber 1986; Wilson and Mueller 1993). Using satellite imagery, Purdom (1976) showed that convergence lines from land–sea-breeze circulations and previous convective outflows often initiate convection, especially at the intersection of two or more such convergence lines. Boundary layer convergence lines that initiate convection have also been shown to have other origins, such as terrain-induced flows [e.g., the Denver cyclone, Crook et al. (1990)], land contrast circulations (Lynn et al. 1998), and boundary layer roll circulations (Trier et al. 1991).

An additional important component with regard to convective initiation in the southern plains of the United States is a lower-tropospheric moisture boundary known as the dryline (Rhea 1966; Schaefer 1986), which marks the often sharp transition between warm moist air originating from the Gulf of Mexico and warmer but drier air that has subsided off the high terrain of Mexico and the southwestern United States. The dryline is a location of enhanced convergence, due to its typical collocation with a synoptic-scale Rocky Mountain lee trough (Martin et al. 1995), as well as an “inland sea breeze” effect that is driven by a surface moisture contrast (Sun and Wu 1992; Shaw et al. 1997). However, most studies of convective initiation at the dryline have found that a combination of dryline convergence and some other source of convergence provides a preferred location for the development of deep convection. Examples of these other sources of convergence include the intersection of boundary layer rolls with the dryline (Ziegler et al. 1997; Atkins et al. 1998) and the interaction of an outflow boundary from previous convection with the dryline (Carbone et al. 1990).

When the additional ingredient of synoptic- or frontal-scale forcing is present, the nature of the lifting mechanism that leads to convective initiation becomes even more complex. In the case of a frontal passage, convection can be initiated by the front itself or by interactions between the front and any of the phenomena discussed above. For example, Braun and Houze (1997) found that the well known 10–11 June 1985 Oklahoma–Kansas squall line was initiated by a cold front at its northern end and by two separate outflow boundaries at its middle and southern ends. Of particular interest for the present study is the idea that convection can be initiated by a dryline being overtaken by a cold front from the west, a process that has been discussed by many investigators (e.g., Koch and McCarthy 1982; Ogura et al. 1982; Shapiro 1982; Schaefer 1986; Neiman et al. 1998; Neiman and Wakimoto 1999). Locatelli et al. (1995a) found that the CFA rainband associated with the 8–9 March 1992 case formed at approximately the same time that a Pacific cold front overtook a dryline. An important question that will be addressed in the present study is: how do the frontal structures and vertical velocity patterns behave before, during, and after a dryline is overtaken by a Pacific cold front?

The role of large-scale forcing in convective initiation has been a matter of debate in the literature. Part of this debate stems from the difficulty in defining exactly what is meant by large scale. Koch and McCarthy (1982), Keyser and Carlson (1984), and Doswell (1987) have argued that the relatively weak vertical velocities estimated from either quasigeostrophic or semigeostrophic diagnostic equations are insufficient to provide the necessary lifting for convective initiation. On the other hand, Crook and Moncrieff (1988) define large scale as an order of magnitude greater than convective scale (i.e., somewhere between Doswell’s large scale and convective scale, or what might more commonly be referred to as meso-β scale), and they show that, with an appropriate sounding, a very modest meso-β-scale ascent of 8 cm s−1 is sufficient to initiate convection in just a few hours. Crook and Moncrieff make the important point that the key to convective initiation (assuming that instability is present) is sufficient parcel ascent, which can be attained either by a strong vertical velocity over a short time or by a weak vertical velocity over a longer time. The question of what was the relevant spatial scale for the vertical velocity that initiated the convective rainbands in the 8–9 March 1992 case will be addressed in the present study.

This study will also address Martin et al.’s (1995) hypothesis that the pre–dry trough rainband in the 8–9 March 1992 case was of an elevated nature and driven by the release of convective instability aloft. Colman (1990a,b) found that most wintertime thunderstorms and some nonwinter thunderstorms in the United States develop as elevated convection above a frontal interface. He found that these elevated storms form in situations with little or no convective available potential energy (CAPE; i.e., 500 J kg−1 or less) for parcels at any level. Therefore, Colman concluded that processes other than buoyant upright convection are key to the development of these thunderstorms. The present study will attempt to reconcile these ideas with regard to the pre–dry trough rainband of 8–9 March 1992, which generated several reports of lightning and thunderstorms (Martin et al. 1995).

The remainder of this paper is organized as follows. Section 2 describes the observed event for comparison with the model simulation that is presented in section 3. Section 4 analyzes the large-scale (quasigeostrophic) forcing as it pertains to the timing and location of the initiation of the convective rainbands that are the focus of this study. Section 5 analyzes the initiation of the rainbands in more detail using the highest-resolution (8.3 km) model simulation. A discussion of the results and a summary of major conclusions are presented in sections 6 and 7, respectively.

2. Observed storm structure

The cyclonic storm of 8–9 March 1992, and associated mesoscale weather phenomena, have been analyzed in detail by Martin et al. (1995), Wang et al. (1995), Locatelli et al. (1995a, 1998), Castle et al. (1996), Miller et al. (1996), Neiman et al. (1998), and Lemone et al. (1999). However, pertinent features are described here, primarily for comparison with our model simulation.

During the day on 8 March 1992, a surface lee cyclone developed in eastern Colorado as an upper-level short-wave trough moved northeastward from northwestern Mexico around a long-wave, upper-level cutoff low centered over Arizona. Attendant with the surface cyclogenesis were three significant features that organized precipitation: an arctic front1 that moved southward across the northern Great Plains, a lee trough and associated dryline that developed south of the low pressure center, and a Pacific cold front that crossed the Rocky Mountains and approached the low pressure center and lee trough from the west. The lee trough marked the surface location of a warm-frontal structure that sloped upward toward the northeast [referred to by Martin et al. (1995) as the dry trough], in which a warm dry southwesterly air current originating from over the Mexican Plateau rose above the cooler but moister air mass east of the lee trough. By 0600 UTC 9 March (Fig. 1a), the arctic front had reached eastern Colorado and western Kansas, and provided a thermal boundary on which the low pressure center developed and propagated to the northeast. East of the low center, the arctic front remained stationary; west of the low center, it moved rapidly southward around the low center. The lee trough began to migrate eastward, occupying a position just east of 100°W at this time. The Pacific cold front was analyzed on the National Meteorological Center (now known as the National Centers for Environmental Prediction) operational surface analysis at this time (not shown) as a surface cold front in essentially the same location as the lee trough in our analysis (Fig. 1a). It is not shown as a surface cold front in our analysis because the Pacific cold front had already occluded with the warm-frontal surface east of the lee trough, and the leading edge of the Pacific polar air mass occurred aloft and farther east than the surface trough. Note that the 500-hPa position of the leading edge of cold advection (Fig. 1b) is east of the surface trough. This is in agreement with Locatelli et al. (1995a) and Neiman et al.’s (1998) depiction of a lee trough and a CFA farther east on their surface analyses.

The two rainbands that are the focus of this study occurred in the region south of the stationary arctic front and east of the lee trough. At 0600 UTC 9 March, the pre–dry trough rainband appears as two disconnected bands in radar and satellite imagery (Fig. 2), but the time sequence of radar and satellite images in Martin et al. (1995) demonstrates its coherent rainband structure and temporal continuity. The radar signature of the CFA rainband at 0600 UTC 9 March (Fig. 2) is approximately 100–200 km ahead of the lee trough position (Fig. 1a), and approximately 50–100 km behind the leading edge of the baroclinic zone at 500 hPa (Fig. 1b). However, these analyses are not valid at precisely the same time, and there is some uncertainty in the location of the CFA due to the spacing of soundings. A third precipitation feature, which is not a focus of this study, was a general band of cloudiness and snowfall that extended from northeastern Colorado to northern Iowa. This was associated with the arctic front.

3. Model simulation

The model used in this study is the Fifth-Generation Pennsylvania State University–National Center for Atmospheric Research Mesoscale Model (MM5). The MM5 is a sigma-coordinate, primitive-equation model originally developed by Anthes and Warner (1978). More recently it has been upgraded with nonhydrostatic dynamics (Dudhia 1993) and multilevel nesting capabilities (Grell et al. 1994). Key physical parameterizations that were chosen for this study are a mixed-phase bulk microphysical parameterization similar to that described by Reisner et al. (1998), with cloud water, cloud ice, rain, and snow as prognostic variables; the Kain and Fritsch (1993) cumulus parameterization scheme; and a high-resolution planetary boundary layer scheme (Blackadar 1979; Zhang and Anthes 1982).

Because this is a multiscale study of the processes that led to the initiation of convective rainbands, a multinested model domain configuration was employed. Results from all model domains will be discussed. The outermost domain is a 75-km grid that covers southern Canada, the United States, Mexico, and the northeastern Pacific Ocean out to 140°W. Within this domain is a two-way nested 25-km grid that covers the region from central Minnesota to the southern tip of Texas in the north–south direction, and from the Pacific coast of California to the Mississippi–Alabama border in the east–west direction. Finally, an 8.3-km grid was one-way nested inside the 25-km grid, covering a region from central South Dakota to central Texas in the north–south direction, and from the Colorado–Utah border to central Missouri in the east–west direction. The two outer domains were run with both the explicit microphysical parameterization and the cumulus parameterization, whereas the 8.3-km domain was run with explicit precipitation only, in order to elucidate the precise lifting mechanisms that led to the initiation of convection. Weisman et al. (1997) have demonstrated that for squall-line-type storms 4 km is approximately the maximum grid spacing at which resolved processes alone can develop a cold pool fast enough to realistically produce a squall-line-type convective system. However, in the present case, we are more concerned with the initiation of larger-scale convective rainbands, which we postulate to be strongly forced by frontal dynamics, rather than solely by self-induced cold-pool lifting within a uniform environment. Therefore, an 8.3-km domain with explicit microphysics and no cumulus parameterization was deemed suitable for this study. While details of the simulated convection would likely have been improved with either a higher-resolution domain or a cumulus parameterization in a lower-resolution domain; this configuration was employed as a compromise between resource limitations and the need to explicitly resolve the convective initiation.

All three domains were run with 27 vertical sigma layers, with vertical resolution ranging from Δσ = 0.01 (Δp ≅ 9 hPa) near the surface to Δσ = 0.50 (Δp ≅ 43 hPa) in the middle and upper troposphere. Initial conditions were generated by updating a coarse-resolution first guess with all available surface and upper-air observations, including special observations from the STORM-FEST project. All three domains were initialized at 1200 UTC 8 March 1992, and model integrations were performed for a period of 24 h.

The simulated sea level pressure pattern at 0600 UTC 9 March 1992 (Fig. 3a) is quite similar to the observed pattern (Fig. 1a), with a low pressure center straddling the Colorado–Kansas border, an arctic front extending from Idaho to Lake Michigan, and a lee trough extending south from the low center into central Texas. The arctic front is somewhat farther north and the lee trough somewhat farther east in the simulation. At 500 hPa (Fig. 3b), the leading edge of the baroclinic zone is similar to that observed (Fig. 1b), though it has not progressed as far eastward into Oklahoma in the simulation.

The model-simulated cloud and precipitation fields at 1800 UTC 9 March (Fig. 4) compare favorably with the corresponding observed fields (Fig. 2). The dominant feature is the cloud mass north of the arctic front, which produced heavy snow. The CFA convective rainband has just formed in eastern Kansas and Oklahoma, although it is characterized by a series of individual cells rather than the more continuous line observed. The pre–dry trough rainband can also be seen in the model simulation in northwestern Missouri and western Iowa. The convective nature of both of these simulated rainbands will be more clearly demonstrated in vertical cross sections in a later section of this paper. Another observed feature that is roughly duplicated by the model is the area of convection in eastern Oklahoma, although that is not a focus of this study.

4. Synoptic-scale diagnosis

To investigate the relationship between synoptic-scale forcing and the convective rainbands, our analysis of this model simulation begins with an examination of QG forcing and vertical velocity response. Locatelli et al. (1995a) performed limited QG diagnosis of this case, examining the forcing term in the QG omega equation in the vicinity of the CFA. They found that the CFA was frontogenetical and inferred from this an upward motion pattern at its leading edge, but they did not solve the omega equation to quantitatively ascertain the QG vertical velocity field. The following analysis extends the work of Locatelli et al. (1995a) by quantitatively solving the QG omega equation. The analysis was performed on output from the 75-km model domain as follows. First, the height and temperature fields were interpolated from the model’s sigma coordinate system to pressure levels with a vertical spacing of 30 hPa. Due to the fact that the area of interest includes substantial topography, a step-mountain lower boundary was defined such that the lowest grid point in each column is the lowest above-ground point. To remove small-scale variability in the height and temperature fields that is not consistent with QG scaling a low-pass filter was applied to these fields, with 50% attenuation occurring at wavelengths of 300 km, similar to that suggested by Barnes et al. (1996). Using the smoothed pressure level data, Q vectors (Hoskins et al. 1978) and their divergence were calculated at each grid point, and a simple relaxation method was used to invert the Q-vector form of the omega equation and obtain vertical velocities. The upper boundary condition was defined as zero vertical velocity at 75 hPa, and the lateral boundary conditions were set to a smoothed value of the full vertical velocity. Latent heating was not included in the diagnosis because the focus of the analysis was to examine the mechanisms that led to the development of convection, rather than the mechanisms that sustained convection.

It was of particular interest in this case to examine the effects of the two lower boundary conditions that are often employed in solving the omega equation, namely, the Ekman condition and the topographic condition. One would anticipate both to be significant in a situation where there is both a strong surface cyclone and substantial topography. These boundary conditions were defined in the standard way [e.g., as in the appendix of Hoskins et al. (1985)], with the Ekman condition supplying a vertical velocity that is roughly proportional to the surface value of geostrophic relative vorticity, and the topographic condition supplying a vertical velocity that is roughly proportional to upslope forcing. The upslope forcing is defined as Vsfc · htpg, where Vsfc is a frictionally adjusted geostrophic wind (i.e., the surface geostrophic wind, but reduced in magnitude and turned to the left by a constant amount) and htpg is the height of the topography. Both boundary conditions were employed at all grid points composing the irregular lower boundary in the step-mountain pressure coordinate grid. Two simplifications should be noted here. The first is the assumption that the frictionally altered surface wind is proportional to the surface geostrophic wind (Vsfc = aVg + bk × Vg). This affects both the constant of proportionality in the Ekman vertical velocity, and the strength of the upslope forcing in the topographic condition. Although this is a good approximation in most situations, there are situations where it is deficient. For example, it is known that the surface wind is often turned more than 90° from geostrophic in the region west of a lee trough (Bluestein and Crawford 1997), in which case the factor a would be negative. The second simplification is the location of the lower boundary. In principle, the Ekman condition applies at the top of the boundary layer, whereas the topographic condition applies near the surface. For the sake of simplicity, and because initial experiments indicated that topographic forcing was generally stronger than Ekman forcing, the ground surface was chosen as the location where both boundary conditions apply.

The 750-hPa level was chosen for displaying the QG diagnosis because it is close to the vertical location where both of the convective rainbands were initiated, and because it is low enough to capture the effects of the topographic and Ekman forcing, without actually intersecting the topography. Three significant features stand out in the Q-vector distribution at 0000 UTC 9 March 1992 (Fig. 5a). The first is the region of strong frontogenesis within the warm-frontal zone over southern Minnesota and South Dakota. In this region, the large Q vectors are perpendicular to the isotherms and are, therefore, due to an increase in the magnitude of the temperature gradient (frontogenesis), rather than rotation of the temperature gradient. The frontogenesis results from strong geostrophic deformation associated with the col-type height pattern in the presence of a substantial baroclinic zone associated with the arctic front and overlying quasi-stationary front. The second significant feature is the region of large Q vectors in the vicinity of the closed cyclonic circulation over Colorado. Note that these Q vectors are largely parallel to the isotherms, indicating that they are associated with rotation of the vector thermal gradient rather than with frontogenesis. This is not surprising, in view of the strong cyclonic rotation in the presence of a substantial thermal gradient. These patterns are similar to those of along- and cross-isentrope components of the Q vector seen in a developing idealized baroclinic wave (Keyser et al. 1992) and in observed occluding cyclones (Martin 1999). The third significant feature is the Q-vector pattern in the vicinity of the baroclinic zone over the Texas and Oklahoma panhandles that is associated with the advancing Pacific cold front. The Q vectors are weak but oriented generally toward warmer air, indicating frontogenesis. Finally, the “warm sector” is generally associated with weak Q-vector forcing.

Noteworthy regions of Q-vector convergence (which induces upward motion) can be seen in southern Minnesota and northeastern Nebraska immediately equatorward of the arctic/quasi-stationary front, as well as in northeastern Colorado (Fig. 5b). Interestingly, the strongest Q-vector forcing is associated with rotation of the thermal gradient in the vicinity of Colorado (where heavy snow was occurring at this time) rather than with the region of frontogenesis farther northeast. However, of primary interest for this study is the fact that neither of the convective rainbands (shading in Fig. 5b) were initiated in regions of significant Q-vector convergence;in fact, the CFA rainband actually initiated partly in a region of weak Q-vector divergence.

Due to the linear nature of the omega equation, the forcing and boundary conditions can be partitioned and inverted in a piecewise manner, with the sum of the separate solutions equaling the total solution that is obtained by inverting all forcing and boundary conditions at once. In this way, the effect of different forcing elements and boundary conditions can be examined separately. In the present case, the Q-vector forcing was inverted without boundary conditions, and then each boundary condition (Ekman and topographic) was inverted separately without interior forcing. The QG vertical velocity field at 750 hPa due to Q-vector forcing only (Fig. 6a) reflects a smoothed version of the Q-vector divergence (Fig. 5b), although some differences are apparent, due to the vertical influence of somewhat different forcing above and below 750 hPa. The primary upward QG vertical velocity is associated with the snowband along the arctic/quasi-stationary front, whereas only weak upward air motion prevails in the vicinity of the pre–dry trough rainband, and little or no upward vertical velocity is seen in the vicinity of the nascent CFA rainband.

Inversion of the two boundary forcing terms yielded vertical velocity fields with maximum magnitude near the ground (as expected), but also with maximum noise near the ground. These vertical velocity fields attenuated and became smoother with height, and are shown here at the 750-hPa level. Although the center of the surface cyclone was located in the lee of the Rockies, a larger-scale surface cyclone roughly straddled the Rocky Mountain crest, yielding a quadrupole pattern in the topographic forcing centered over Colorado, with upslope motion to the northeast and southwest, and downslope to the northwest and southeast (Fig. 6b). Note that, compared to the dynamical forcing (Fig. 6a), the upslope motion actually made a comparable or greater contribution to the upward motion in northeastern Colorado where the heaviest snow was falling. However, in the region where the CFA rainband developed, weak downslope motion prevailed. The pre–dry trough rainband developed farther east over the plains, where topographic forcing was negligible. Ekman forcing also contributed to significant upward vertical velocity over eastern Colorado (Fig. 6c), coincident with the position of the surface low pressure center at this time (not shown).

The summation of all three vertical velocity forcing mechanisms (Q-vector divergence, topographic forcing, and Ekman forcing; Fig. 6d) shows that the CFA rainband actually developed in a region of little or no upward QG vertical velocity. The pre–dry trough rainband, although situated in a region of upward QG vertical velocity, was oriented roughly perpendicular to the long axis of the elongated region of strongest upward motion. In summary, the QG vertical velocity field provided little indication of precisely where either of these rainbands was likely to develop.

In addition to the QG diagnosis it is of interest to examine the development of these rainbands in terms of potential vorticity (PV) dynamics (Hoskins et al. 1985), since the basic ideas of this framework are not constrained to the low Rossby number regime, as they are in QG theory. In this study, a PV analysis is useful only to the extent that it can explain the vertical velocity that led to the initiation of the two rainbands. Within the PV framework, vertical velocity occurs when a PV anomaly passes above (or below) the point of interest or, in other words, when a PV anomaly is embedded in vertically sheared flow—the so-called vacuum cleaner effect described by Hoskins et al. (1985). This effect can be qualitatively ascertained by examining the advection of PV by the thermal wind, analogous to the advection of vorticity by the thermal wind in the Sutcliffe (1947) development rule.

The tropopause level is of particular significance in PV dynamics, because deviations in the height of the tropopause result in significant PV anomalies due to the large vertical gradient of PV across the tropopause. Maps of PV on isentropic surfaces that intersect the tropopause at midlatitudes are often used to track dynamically important tropopause-level PV anomalies (Hoskins et al. 1985). On the 310-K isentropic PV map at 0000 UTC 9 March, the prominent feature is the lobe of high PV that is situated immediately upwind of the developing surface low pressure center and is wrapping around a long-wave trough that covers the southwestern United States. An estimate of PV advection by the thermal wind can be made by comparing the upper-level PV distribution (Fig. 7) with a proxy for the thermal wind, namely, isotherms at a midtropospheric level (Fig. 5a). Such a comparison suggests upward motion over northeastern Colorado, and downward motion over the Oklahoma panhandle, consistent with the QG diagnosis. However, there are no details in the upper-level PV pattern that indicate the specific locations and orientations of the convective rainbands. The possibility remains that smaller-scale preexisting PV anomalies in the middle or lower troposphere played a role in rainband initiation. This is one of the issues that will be examined in the next section with the aid of cross sections from the 8.3-km model run.

5. Rainband initiation

a. The pre–dry trough rainband

The initiation of the rainbands on the mesoscale can be examined by means of vertical cross sections derived from the highest-resolution (8.3 km) model simulation, the locations of which are shown in Fig. 4. Cross section A–B was chosen to illustrate the formation of the pre–dry trough rainband. It passes through the rainband and is oriented approximately parallel to trajectories of air parcels that entered into the updraft of this rainband. At 0500 UTC 9 March, shortly after the initiation of the pre–dry trough rainband, a stable layer that resembles a gently sloping warm front is evident in the potential temperature (θ) distribution in cross section A–B (Fig. 8a). This sloping stable layer intersects the surface near the 200-km position and rises to 800 hPa at the 1100-km position. It is separate from the arctic front, which can be seen below 900 hPa in the rightmost 200 km of the cross section. It is also entirely east of the surface dryline position (which is outside of the cross section, to the west). Thus, there is a warm-frontal surface overlying the region east of the dryline that is typically considered to be the warm sector of cyclones in this region. Although this warm-frontal structure slopes toward the northeast, it intersects the ground along the more north–south-oriented position of the dryline. For this reason, it has been termed the dry trough (Martin et al. 1995; Hobbs et al. 1996). Associated with the warm-frontal feature is a gradual ascent toward the northeast. Although somewhat noisy, the circulation vectors generally depict this ascent. This ascending airstream from the south is responsible for transporting high-θe air northward and upward in an elongated lobe that detaches from the ground and rises up to 700 hPa. The upper side of this lobe is convectively unstable. Therefore, continued lifting of this layer eventually releases the instability, resulting in the formation of the pre–dry trough rainband at the farthest forward extent of the high-θe lobe, as is seen in the cross section. This process was hypothesized by Martin et al. (1995) in their observational analysis of this case, and it is now confirmed by the present model simulation.

Further support for this hypothesis is provided by showing CAPE in cross section A–B (Fig. 8b). Normally, CAPE is calculated as a single value from a sounding by lifting an air parcel initially at or near the surface (or some average of parcels in a near-surface layer) up to the equilibrium level. It may also be depicted as a horizontal contour plot showing the surface-based CAPE for all soundings in a horizontal domain. The depiction of CAPE in a cross section requires some explanation. At each grid point in the full 3D model domain, the value of CAPE is calculated as the positive area in a sounding that would be achieved by a parcel, initially at the level of the grid point in question, that is lifted to its equilibrium level. Thus, unshaded regions in the figure are composed of air parcels that would achieve little or no buoyancy if lifted. It is clear from this depiction that the pre–dry trough rainband was generated from an elevated source, since at the location of the rainband only air between 820 and 570 hPa has significant CAPE. The pre–dry trough rainband may be one class of convection that contributes to the high incidence in the central United States of elevated convection above frontal surfaces that is seen in the climatological study of Colman (1990a).

Two 17-h backward trajectories that terminate in the updraft of the pre–dry trough rainband were calculated (Fig. 8b). These trajectories show that the parcels that produced the pre–dry trough rainband underwent gradual, steady ascent over a period of many hours, until the point in time when the convective instability in the layer in which they were embedded was released. Note that the QG vertical velocity in the region between Texas and Missouri (Fig. 6d) is characterized by weak values of only a few dPa s−1 (μbar s−1). However, the hourly QG estimates of vertical velocity along trajectory number 1 (shown in Table 1) are comparable to the full vertical velocities, and the net ascent of trajectory number 2 estimated from QG vertical velocity along the trajectory during the 16-h period prior to convective initiation is close to the actual net ascent of the trajectory during the same period. The convective inhibition is seen to decrease steadily during the lifetime of the trajectory, reaching zero at the time that convection breaks out. Therefore, in the case of the pre–dry trough rainband, the ascent that leads to the formation of the rainband is well described by QG diagnosis, because it is due to slow frontal lifting over large distances. Since it acts over a long period of time, it is able to bring about the release of convective instability and generate a mesoscale rainband.

In the previous section, the development of the rainbands was discussed in terms of PV dynamics, but the PV distribution was shown only on a single isentropic surface. To examine the relationship between tropospheric PV anomalies and the initiation of the pre–dry trough rainband, the PV distribution in cross section A–B (Fig. 8c) is examined. There is a region of enhanced PV immediately upstream of the location where the rainband was initiated. However, there are also sporadic regions of enhanced PV within the entire warm-frontal zone, and each of these lies directly beneath a region where cloud has formed. All of these PV anomalies are generated by latent heat release in the ascending airstream and are more accurately described as an effect of the gradual ascending motion, rather than as preexisting features that caused the rainband.

b. The CFA rainband

Cross section C–D (location shown in Fig. 4) was chosen to illustrate the formation of the CFA rainband. It is an east–west cross section that is roughly perpendicular to the CFA, the CFA rainband, and the dryline. Several key features are evident in a plot of θ, θe, and circulation vectors along cross section C–D valid at 0100 UTC 9 March, approximately 1 h prior to the initiation of the CFA rainband in the simulation (Fig. 9). First, a pool of cooler air, capped by a stable layer, resides at low levels in the right half of the cross section, bounded above approximately by the 298-K isentropic surface. However, this air is also characterized by high moisture and, therefore, carries high values of θe. The stable layer is the same layer that was seen in cross section A–B (Fig. 8a), and the high-θe air is the same air mass that was transported northward and aloft in that cross section. The cool moist air mass is the familiar capped unstable layer that is very common east of the dryline in the central United States, especially in spring and summer (Shapiro 1982; Schaefer 1986; Lanicci and Warner 1997; Neiman et al. 1998; Neiman and Wakimoto 1999). Above this cool moist air mass, between 650 and 1000 km in the cross section, is a warm dry air mass, characterized by relatively high θ and relatively low θe values. This is air that had previously subsided off the Mexican Plateau and flowed northeastward above the low-level cool moist air. Finally, a mass of cool dry Pacific polar air, occupying most of the troposphere, is approaching from the west, as evidenced by the strong westerly component to the circulation vectors and lower values of both θ and θe. The Pacific polar air is characterized by moisture values that are intermediate between those of the moist air east of the dryline and the dry air from the southwest. As this air mass descends the east slope of the Rocky Mountains, it approaches the other two air masses in the form of two distinct baroclinic zones, the leading edges of which are designated BZ1 and BZ 2 in Fig. 9.

The first is a midtropospheric baroclinic zone, which passes over the cool low-level air mass in a warm-occlusion manner. Ascent can be seen within and ahead of BZ1. However, the warm air ahead of the baroclinic zone in the midtroposphere is not unstable, and convection is not produced. With regard to the unstable layer of air in the lower troposphere beneath BZ1, Locatelli et al. (1995a, 1997) hypothesized that the passage of cold air aloft can produce a moving surface pressure pattern that induces an isallobaric wind change, convergence, and vertical velocity underneath the CFA, sufficient to trigger convection in an unstable environment. Neiman et al. (1998) observed a surface pressure trough and wind shift associated with the CFA passage over the cool moist air mass east of the dryline in the 8–9 March 1992 case. However, in the present modeling study, the surface pressure signature associated with BZ1 is not strong enough to induce significant convergence and vertical velocity. Convergence values of around 1 × 10−5 s−1 are present beneath BZ1 within 1 km of the surface, but this translates to only 1 cm s−1 vertical velocity at 1 km above the surface. In summary, BZ1 does not participate in the initiation of the CFA rainband.

At low levels, the advance of the Pacific polar air mass is led by baroclinic zone BZ2, which is somewhat behind the advance of BZ1. Upon encountering the dryline, baroclinic zone BZ2 occludes with the cool air mass to the east, as evidenced by the V-shaped isentropes at the location of the dryline; this forces a narrow column of ascent directly above the point of occlusion. This narrow column of ascent and associated plume of high-θe air are similar to those shown by Shapiro (1982) and Neiman and Wakimoto (1999). The air that is lifted is of a unique character: the weak vertical gradient of θ in that column suggests that the air is of neutral dry stability and therefore easily lifted. Additionally, with low-level moisture in the column and the advance of lower-θe air aloft, the column is potentially unstable. It is this occlusion process, between the low-level advance of the Pacific polar air mass (led by BZ1) and the cool moist air mass to the east, that initiates the CFA rainband.

The occlusion process can be seen more clearly in a time series of cross sections through the developing CFA rainband (Fig. 10). The middle panel (Fig. 10b) is at the same time and location as cross section C–D in Fig. 9. The first cross section (Fig. 10a) is 4 h earlier and 100 km south of the original cross section C–D, and the third cross section (Fig. 10c) is 4 h later and 100 km north of the original cross section C–D. The reason for moving the cross section is to approximately follow the northward movement of the low-level air mass in which the CFA rainband developed. Each panel shows potential temperature, the along-cross-section component of the potential temperature gradient, and 3-h parcel ascent as a proxy for vertical velocity. Because vertical velocity tends to reflect the noisy signature of gravity waves, a method was devised to show a more meaningful depiction of parcel ascent that filters out short-timescale oscillatory vertical motion. At each time and at each grid point below ∼650 hPa, a 3-h backward trajectory was computed and the net ascent of that trajectory (in hPa) was defined as the parcel ascent for that grid point at that time. It is this quantity that is shown in the cross section (Fig. 10).

At the first time (2100 UTC 8 Mar; Fig. 10a), the baroclinic zone BZ1 has already occluded with the dryline, whereas the surface baroclinic zone BZ2 has not yet reached the dryline. Weak ascent can be seen ahead of BZ2, and descent behind it. The western end of the cool moist air mass (i.e., the dryline) is characterized by a thermal gradient that opposes that of approaching BZ2 and forces a separate small region of ascent at its western end (at the 470-km horizontal position in the cross section). Four hours later (Fig. 10b), BZ2 has occluded with the opposing thermal gradient at the dryline, combining the separate ascent regions into a single column of significant ascent nearly 200 hPa deep, similar to the process described in the observational study of Neiman and Wakimoto (1999). In the final panel (Fig. 10c), BZ1 has moved farther east and aloft. BZ2 further occludes with the cool air mass to the east and forms a new CFA, as can be seen by the clearly tipped-forward leading edge of eastward thermal gradient associated with BZ2. At this time, the CFA rainband has developed (as seen by the precipitation column), and the ascent pattern that initiated the rainband in the previous panel is now cluttered by precipitation-induced downdrafts. From this time on, the rainband continues to track eastward with the CFA associated with BZ2.

In contrast to the pre–dry trough rainband, the CFA rainband is more of a surface-based convective system in terms of the source of its air. Examination of CAPE in the environment into which the rainband propagates (Fig. 11) shows highest values at the surface, although potentially buoyant air was available through a depth of 1.5–2.0 km.

It is of interest to examine the extent to which the occlusion process and the associated narrow region of lifting are evident in the QG-diagnosed vertical velocity described in the previous section. To that end, QG vertical velocity (including the contributions from Q-vector forcing and both boundary conditions) is shown in cross section C–D at 0100 UTC 9 March (Fig. 12a), with the region of shading indicating the column of lifting seen in Fig. 10b. It is clear that filtering of the input fields, and the inherent low-pass scale selection that occurs upon inversion of the QG omega equation, results in a smoothing of the structural details that were seen in previous figures (e.g., there is no signature of separate ascent regions associated with the BZ1 and BZ2 boundaries seen in Fig. 9). The vertical velocity field due to Q-vector forcing only (not shown) hints at an occluded structure, insofar as it has a slightly tipped-forward transition between descent and ascent, as well as a tipped-forward axis of maximum ascent. With the topographic boundary condition included, however, strong downslope forcing favors descent near the surface, producing a tipped-backward vertical velocity structure. Furthermore, the key region of maximum parcel ascent that initiated the rainband (in the model simulation) is coincident with near-zero to slightly negative values in the QG vertical velocity field, demonstrating that the QG diagnosis cannot capture this key feature. This is explained by the smallness in scale of the column of maximum parcel ascent, compared to the smallest resolvable scale of the QG diagnosis method. The unsmoothed full vertical velocity distribution (Fig. 12b), while noisy, clearly shows a region of significant ascent in a position that is consistent with the net parcel ascent maximum. However, if the full vertical velocity field in the model is filtered to the same degree as the input fields to the QG diagnosis, the result (Fig. 12c) is qualitatively very similar to the QG vertical velocity distribution (Fig. 12a). The only significant difference is that in the smoothed full vertical velocity, the region of ascent slightly undercuts the region of descent, creating a slightly tipped-forward structure. The stronger descent near the surface in the QG solution is due to the topographic boundary condition, suggesting that the diagnosis method is either overemphasizing the downslope forcing, or misplacing the location of the transition from downslope to upslope forcing, which should be approximately at the lee trough location.

Finally, we examine the initiation of the CFA rainband in terms of the potential vorticity distribution. The potential vorticity field in cross section C–D at 0100 UTC 9 March (Fig. 13) shows that there is a small surface-based PV anomaly at the location of the ascent region that initiates the CFA rainband (at the 460-km position in the cross section), but it was recently generated by nocturnal surface cooling in conjunction with enhanced vertical vorticity at the lee trough. The more significant feature is the lobe of PV extending down from the stratosphere within an upper-level front–tropopause fold. A reasonable hypothesis within the PV framework is that the advection of this PV feature was responsible for generating the lower-tropospheric lifting seen in the cross section. However, an important point that the cross section does not illustrate is that the upper-level front–tropopause fold is oriented from southwest to northeast, whereas the dryline is oriented from north to south. If a cross section farther to the south and at a later time is examined (Fig. 13b), the occlusion and associated vertical motion can still be seen in the lower troposphere at the 540-km position, but the tropopause fold and upper-tropospheric PV anomaly are now much farther to the west, out of the cross section. Therefore, as was the case with the pre–dry trough rainband, the lifting mechanism that initiates the rainband is better understood in terms of frontal interaction and frontally induced ascent (in this case, a lower-tropospheric occlusion process) than it is in terms of vertical velocity induced by the advection of one or more identifiable preexisting PV anomalies.

6. Discussion

In section 1, the term convective rainband was defined as a band of precipitation that is convective in nature, but synoptic scale in length. In referring to the precipitation features in this study, we have used this term because it is consistent with the established use of the term rainband to refer to frontally forced precipitation structures within an extratropical cyclone (e.g., Browning 1990). We use the term convective rainband rather than the term squall line, because, according to the Glossary of Meteorology (Huschke 1959), a squall line specifically refers to a “nonfrontal” line of convection. This is consistent with the view that squall lines are self-propagating mesoscale phenomena that are not critically dependent on frontal or baroclinic forcing for their initiation or maintenance (Newton 1950; Thorpe et al. 1982; Rotunno et al. 1988).

In some cases, the connection between a convective rainband and frontal features is obvious, because the band is collocated with a surface front. For instance, a convective rainband can be driven by a surface cold front when the lifted warm-sector air is conditionally neutral or unstable, producing a band of convection that is tied to the surface position of the cold front; examples of this situation are the “frontal squall line” observed and modeled by Koch et al. (1997) or the vigorous narrow cold-frontal rainbands analyzed by Hobbs and Persson (1982), Carbone (1982), and Locatelli et al. (1995b). However, an important point that Hobbs et al. (1990, 1996) emphasized is that convective rainbands east of the Rocky Mountains can also be driven by frontal features aloft, the structures and positions of which are often not apparent from conventional surface frontal analyses.

In this study, the initiation of the convective rainbands in the 8–9 March 1992 storm has been examined in terms of its relationship to the baroclinic dynamics acting on both the large scale and the mesoscale. Discussions of the findings for each of the two rainbands examined in this study are given below.

a. The pre–dry trough rainband

The pre–dry trough rainband was found to be of the elevated convective storm type. It developed as a result of the gradual lifting of a convectively unstable layer above the warm-frontal surface east of the dryline, as hypothesized by Martin et al. (1995). As such, its initiation is well understood in terms of QG vertical velocity induced by weak frontogenesis within the warm-frontal structure east of the dryline. Therefore, this particular class of convective rainband is an exception to Doswell’s (1987) statement that the origins of the lift required to initiate convection are not likely to be found in large-scale ascent. It should be noted that this type of situation might be loosely classified as an “overrunning” situation, which Doswell recognized as an exception to his rule.

The direct connection between the pre–dry trough rainband and large-scale ascent, however, does not necessarily translate into an easier forecast problem. Because this band developed in an airstream with strong horizontal velocity and weak vertical velocity, it is likely that the precise timing and location of the release of convective instability were very sensitive to details of the vertical stability profile in the layer being lifted. Therefore, even if the process is understood and the synoptic-scale structure is well forecasted by numerical models, small errors in the vertical stability profile would result in large errors in forecasting the pre–dry trough rainband.

When examined in terms of PV dynamics, the pre–dry trough rainband was not connected to the advection of any significant identifiable preexisting PV anomaly, either at the tropopause or within the troposphere. However, due to condensational heating, enhanced PV developed above the entire length of the warm-frontal upgliding airstream. This likely contributed to enhanced ascent along parcel trajectories, but did not specifically determine the location where the rainband developed.

Colman (1990a,b) found that most cases of thunderstorms that developed above frontal surfaces were characterized by little or no (⩽500 J kg−1) positive CAPE. Based on this finding, he concluded that some process other than buoyant upright convection was occurring in these elevated thunderstorms. However, in the present study, the pre–dry trough rainband produced by the model was of an upright convective nature and was able to develop from air with less than 550 J kg−1 maximum CAPE. Therefore, it appears that a careful distinction must be made between “little positive CAPE” and “no positive CAPE,” because in the present case, a little CAPE (less than 550 J kg−1) appears to have been sufficient to support upright convection in a frontal upgliding situation.

b. The CFA rainband

The CFA rainband was initiated by a column of ascent that maximized as a surge of Pacific polar air, which descended east of the Rocky Mountains, occluded with the warm-frontal structure at the dryline. The initiation of convection by this occlusion process has been shown and discussed by previous investigators (Koch and McCarthy 1982; Ogura et al. 1982; Shapiro 1982; Schaefer 1986; Neiman et al. 1998; Neiman and Wakimoto 1999). These and other studies have firmly established the finescale nature of the dryline. Parsons et al. (1991) even emphasized the gravity current–like structure of the dryline. Koch and McCarthy (1982) and Schaefer (1986) originally discussed the contribution of an approaching cold front to the initiation of convection at the dryline only in terms of large-scale balanced frontogenetical circulations. In contrast, the case study by Shapiro (1982), as well the recent studies of Neiman et al. (1998) and Neiman and Wakimoto (1999), have also emphasized the finescale structure and lifting associated with an approaching Pacific cold front. The relevant scale on which key dynamical processes occur likely lies somewhere between the scale of balanced frontogenesis theory and the scale of gravity current dynamics, which might best be described by the term frontal scale, as defined in the introduction. Neiman and Wakimoto describe the narrow ascent plumes forced by the Pacific cold front and dryline as “kinematically forced.” The present study confirms the conclusions of Shapiro (1982), Neiman et al. (1998), and Neiman and Wakimoto (1999), namely, that when a Pacific cold front occludes with a dryline, previously separate frontal-scale ascent regions associated with the two boundaries join to form an enhanced narrow region of ascent that can initiate convection.

Most of the investigators referred to in the previous paragraph hesitated to apply the term occlusion when describing the process of a Pacific cold front overtaking a dryline, presumably reserving that term for the process of a classically defined cold front overtaking a classically defined warm front within the context of the Norwegian cyclone model. However, we see no compelling reason to confine the term occlusion to the classical Norwegian cyclone model, particularly since it so aptly describes the interaction between a Pacific cold front and a dryline. This is because the dryline marks the surface position of what is essentially a warm front that slopes to the east or northeast. Therefore, we have not hesitated to use the term occlusion here. Shapiro (1982) introduced the term frontal–dryline merger to refer to this same process, and this term has been adopted by Neiman et al. (1998) and Neiman and Wakimoto (1999). However, the term merger has not been consistently applied in the literature. In fact, Neiman et al. reserved the term merger to refer to the overtaking of one front by another when both fronts slope in the same direction, to contrast that process with an occlusion, which refers to the overtaking of a warm front by an oppositely sloped cold front. Thus, if Neiman et al.’s definitions of occlusion and merger are adopted, the overtaking of the eastward-sloping warm-frontal boundary at the dryline by a westward sloping Pacific cold front is more aptly termed a frontal-dryline occlusion than a frontal–dryline merger.

Although the QG-diagnosed vertical velocity in the vicinity of this occlusion process hinted at an occluded-like structure, there was actually zero QG vertical velocity at the location of maximum parcel ascent. This demonstrates the frontal-scale nature of the initiating mechanism for the CFA rainband. In this regard, the initiation of the CFA rainband is consistent with the assertion of Doswell (1987) that large-scale dynamics are not sufficient to explain the triggering of convection. However, the initiation of the CFA rainband was a direct result of the interaction of a Pacific cold front with a dryline and the associated pool of cooler air to its east, all of which are often referred to as “synoptic scale” features. The apparent contradiction arises from the fact that the term synoptic scale is often loosely taken to include fronts, which are synoptic scale in length but mesoscale in width (Bluestein 1986).

Locatelli et al. (1995a) analyzed this same CFA rainband using the STORM-FEST observational dataset, as well as a coarse-grid (45 km) MM4 model simulation. From those analyses, they hypothesized that the formation of the CFA rainband was due to two complementary processes. The first was the juxtaposition of broad QG ascent and a narrow region of instability east of the dryline. The second was surface convergence in response to the movement of the CFA over the cool but moist air mass east of the dryline [see Locatelli et al. (1997) for a detailed description of this process]. The results of the present high-resolution modeling study yield somewhat different conclusions regarding the specific mechanisms for the formation of the CFA rainband. The 8.3-km model simulation (particularly Fig. 11) shows that the CFA rainband was initiated when a narrow region of ascent (∼50 km) developed at the western end of a broad (>400 km wide) region of instability. Second, the CFA associated with baroclinic zone 1 showed no evidence of inducing appreciable vertical velocity at the surface as it moved east of the dryline, and baroclinic zone 2 initiated a rainband at the time it occluded with the dryline, not after it overran the dryline and became a CFA. The possibility remains that this process may have acted at a later time, either in association with BZ1 or BZ2, to contribute to convective maintenance.

The eastward movement of the Pacific polar air mass in the form of two separate surges, with the lower-tropospheric surge lagging behind, was not specifically found in previous observational studies of this case (Locatelli et al. 1995a; Neiman et al. 1998). However, there was evidence in the analyses of both studies suggesting that such a structure existed. Locatelli et al.’s (1995a) analysis of this case (specifically their Figs. 5c and 5d) show that the CFA rainband initially developed approximately 100 km behind their analyzed position of the CFA at 500 hPa. Therefore, it is possible that they were analyzing the location of a baroclinic zone that corresponds to BZ1 in the model simulation. Neiman et al.’s (1998) time–height section from soundings at Norman, Oklahoma (their Fig. 12), also suggests the presence of two Pacific cold front surges, with a midtropospheric baroclinic zone arriving at 0200 UTC 9 March, followed by a lower-tropospheric baroclinic zone arriving at 0600 UTC 9 March. In addition, we have seen a two-surge structure in some of our modeling investigations of other cases. It would be reasonable to hypothesize that the Rocky Mountains can perturb or retard the passage of the Pacific cold front in the lower troposphere in a manner that results in a delayed surge of cold air at lower levels. Steenburgh and Mass (1994) found a similar structure in their modeling study of a Rocky Mountain lee trough. They attributed the formation of the trailing near-surface cold air to downward mixing from the cold air mass aloft. Ogura et al. (1982) have also suggested that a thermal trough in the lee of the Rockies inhibits the eastward propagation of surface cold fronts in the daytime, but weakens at night allowing the surface cold front to advance. A problem that requires further investigation is that the present study and the examples cited above are inconsistent with regard to whether the CFA rainband develops with the passage of the first elevated baroclinic zone or the second lower baroclinic zone.

We have examined the lifting that initiates the CFA rainband qualitatively in the context of PV dynamics. In a cross section normal to the rainband at its northern end (Fig. 13a), a substantial tropopause fold and associated high-PV air extended down into the middle troposphere to within 100 km of the location where the CFA rainband initiated. In this cross section, the CFA could be analyzed as the lower-tropospheric terminus of an upper-level front–tropopause fold structure, implicating the advection of this PV anomaly in the development of the rainband. However, in a cross section farther south (Fig. 13b), the lower-tropospheric occlusion process looks similar to that in the northern cross section, including a narrow region of ascent, but the upper-level front–tropopause fold are now well to the west (out of the cross section). These results cast doubt on the importance of a preexisting tropopause-level PV anomaly in the initiation of the CFA rainband. In addition, these results, together with those of other recent observational analyses of frontal evolution in the central United States (Locatelli et al. 1998; Neiman et al. 1998;Neiman and Wakimoto 1999), clearly show that CFAs in the central United States develop as part of a warm-occluded structure in the lower troposphere, rather than as an isolated upper-level front–tropopause fold extending into the lower troposphere. Thus, the fundamental importance of an occlusion process should be added to the carefully worded definition of a CFA given by Hobbs et al. (1996).

We have not addressed here the question of maintenance of the CFA rainband, partly due to the limited resolution of the model simulation, and partly due to the failure of the simulation to propagate the rainband at the observed speed once it developed. In some ways the question of maintenance is more interesting than the question of initiation, because the predominant paradigm for mesoscale squall lines is that they are essentially self-propagating systems that depend only on the shear and thermodynamic conditions of the environment rather than on frontal structures associated with synoptic-scale cyclones. There is an obvious difficulty in applying this paradigm to convective rainbands that are 1000 km or more in length and are collocated along their entire length with an advancing baroclinic zone above the surface. An alternative to this paradigm is that there is a dynamical connection between the CFA and its associated convective rainband that augments (or even supersedes) the cold-pool/self-propagation mechanism. Additional work on the 8–9 March 1992 case, as well on other cases that conform to the STORM conceptual model, is under way to examine this possible dynamical connection.

7. Conclusions

A multiscale numerical modeling study of the 8–9 March 1992 cyclone in the central United States was carried out to investigate the initiation of convective rainbands that had been observationally analyzed by previous investigators. Specifically, the pre–dry trough rainband (Martin et al. 1995) and the CFA rainband (Locatelli et al. 1995a, 1998) were studied. The double-nested model configuration reproduced both the synoptic-scale evolution of the cyclone and the mesoscale initiation of the rainbands.

The pre–dry trough rainband formed as a result of gradual ascent of a potentially unstable layer of air (and the eventual release of that instability) within the warm-frontal zone that is coincident with the dryline and slopes toward the northeast. This release of potential instability was explicitly resolved by the simulation’s innermost (8.3 km) domain, and confirmed the hypothesis of Martin et al. (1995). The elevated convection ensued despite modest values of CAPE (<550 J kg−1) in the ascending airstream. The magnitude of vertical velocity within the ascending airstream was consistent with that estimated from QG diagnosis, but the precise timing of the release of the instability (and thus the location of the rainband) is likely very sensitive to the details of the vertical sounding within the ascending airstream, making the pre–dry trough rainband a challenging forecast problem.

The CFA rainband was initiated by a frontal-scale occlusion process between an advancing Pacific polar air mass and the cool moist air mass east of the dryline. The Pacific polar air mass advanced toward the east in the form of two separate baroclinic zones, the first in the midtroposphere and the second at the surface. Both baroclinic zones occluded with the dryline in a warm-occluded fashion, continuing eastward aloft above the cool moist air mass east of the dryline as a CFA. The second (lower) baroclinic initiated the CFA rainband when it occluded with the dryline. Separate frontal-scale ascent plumes associated with this second baroclinic zone and with the dryline combined to form a single, more vigorous ascent plume that lifted potentially unstable air and initiated convection, in a process similar to that shown observationally for a different case by Neiman and Wakimoto (1999).

On the synoptic scale, the strongest forcing of vertical motion, in terms of quasigeostrophic diagnosis, occurred near the arctic front (due to frontogenesis) and near the center of cyclonic circulation in the lower troposphere (due to rotation of the temperature gradient). However, patterns of quasigeostrophic forcing and vertical motion gave little indication of the locations of initiation of the two convective rainbands. The pre–dry trough rainband formed in a broad region of weak upward motion south of the arctic front, and the CFA rainband formed close to the transition between upward and downward QG vertical motion along the Pacific cold front.

The initiation of the rainbands was also examined qualitatively in terms of PV dynamics. Neither rainband appeared to be connected to the advection of one or more identifiable preexisting PV anomalies at the tropopause or within the troposphere. The CFA rainband developed immediately east of a downward-protruding tropopause fold–positive PV anomaly at its northern end, but was much farther from this upper-level feature at its southern end. Both rainbands developed near regions where diabatic heating produced tropospheric PV anomalies, but these anomalies developed concurrently with the rainbands and did not bear a causal relationship to the rainbands.

Acknowledgments

We thank David Schultz and John Nielsen-Gammon for helpful comments during the review process. This research was supported by Grant ATM-9632580 from the National Science Foundation’s Division of Atmospheric Sciences, Mesoscale Dynamic Meteorology Program. The MM5 model code, as well as the computer resources and technical support necessary for running the model, were provided by the National Center for Atmospheric Research, which is sponsored by the National Science Foundation.

REFERENCES

  • Anthes, R. A., and T. T. Warner, 1978: Development of hydrodynamical models suitable for air pollution and other mesometeorological studies. Mon. Wea. Rev.,106, 1045–1078.

  • Atkins, N. T., R. M. Wakimoto, and C. L. Ziegler, 1998: Observations of the finescale structure of a dryline during VORTEX 95. Mon. Wea. Rev.,126, 525–550.

  • Barnes, S. L., F. Caracena, and A. Marroquin, 1996: Extracting synoptic-scale diagnostic information from mesoscale models: The eta model, gravity waves, and quasigeostrophic diagnostic. Bull. Amer. Meteor. Soc.,77, 519–528.

  • Blackadar, A. K., 1979: High resolution models of the planetary boundary layer. Advances in Environmental Science and Engineering, J. R. Pfafflin and E. N. Ziegler, Eds., Vol. 1, Gordon and Breach, 50–85.

  • Bluestein, H. B., 1986: Fronts and jet streaks: A theoretical perspective. Mesoscale Meteorology and Forecasting, P. S. Ray, Ed., Amer. Meteor. Soc., 173–215.

  • ——, and T. M. Crawford, 1997: Mesoscale dynamics of the near-dryline environment: Analysis of data from COPS-91. Mon. Wea. Rev.,125, 2161–2175.

  • Braun, S. A., and R. A. Houze Jr., 1997: The evolution of the 10–11 June PRE-STORM squall line: Initiation, development of rear inflow, and dissipation. Mon. Wea. Rev.,125, 478–504.

  • Browning, K. A., 1990: Organization of clouds and precipitation in extratropical cyclones. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. O. Holopainen, Eds., Amer. Meteor. Soc., 129–153.

  • Byers, H. R., and R. R. Braham Jr., 1949: The Thunderstorm. U.S. Govt. Printing Office, 287 pp.

  • Carbone, R. E., 1982: A severe frontal rainband. Part I: Stormwide hydrodynamic structure. J. Atmos. Sci.,39, 258–279.

  • ——, J. W. Conway, N. A. Crook, and M. W. Moncrieff, 1990: The generation and propagation of a nocturnal squall line. Part I: Observations and implications for mesoscale predictability. Mon. Wea. Rev.,118, 26–49.

  • Castle, J. A., J. D. Locatelli, J. E. Martin, and P. V. Hobbs, 1996: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part IV: The evolution of a drytrough on 8–9 March 1992. Mon. Wea. Rev.,124, 1591–1595.

  • Colman, B. R., 1990a: Thunderstorms above frontal surfaces in environments without positive CAPE. Part I: A climatology. Mon. Wea. Rev.,118, 1103–1121.

  • ——, 1990b: Thunderstorms above frontal surfaces in environments without positive CAPE. Part II: Organization and instability mechanisms. Mon. Wea. Rev.,118, 1123–1144.

  • Crook, N. A., and M. W. Moncrieff, 1988: The effects of large-scale convergence on the generation and maintenance of deep moist convection. J. Atmos. Sci.,45, 3606–3624.

  • ——, 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.

  • Doswell, C. A., III, 1987: The distinction between large-scale and mesoscale contribution to severe convection: A case study example. Wea. Forecasting,2, 3–16.

  • Dudhia, J., 1993: A nonhydrostatic version of the Penn State–NCAR mesoscale model: Validation tests and simulation of an Atlantic cyclone and cold front. Mon. Wea. Rev.,121, 1493–1513.

  • Grell, G. A., J. Dudhia, and D. R. Stauffer, 1994: A description of the fifth-generation Penn State/NCAR mesoscale model (MM5). NCAR Tech. Note NCAR/TN-398+STR, National Center for Atmospheric Research, Boulder, CO, 138 pp. [Available from National Center for Atmospheric Research, Information Support Services, P.O. Box 3000, Boulder, CO 80307-3000.].

  • Hobbs, P. V., and P. O. G. Persson, 1982: The mesoscale and microscale structure and organization of clouds and precipitation in midlatitude cyclones. Part V: The substructure of narrow cold-frontal rainbands. J. Atmos. Sci.,39, 280–295.

  • ——, J. D. Locatelli, and J. E. Martin, 1990: Cold fronts aloft and the forecasting of precipitation and severe weather east of the Rocky Mountains. Wea. Forecasting,5, 613–626.

  • ——, ——, and ——, 1996: A new conceptual model for cyclones generated in the lee of the Rocky Mountains. Bull. Amer. Meteor. Soc.,77, 1169–1178.

  • Hoskins, B. J., I. Draghici, and H. C. Davies, 1978: A new look at the ω-equation. Quart. J. Roy. Meteor. Soc.,104, 31–38.

  • ——, M. E. McIntyre, and R. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc.,111, 877–946.

  • Huschke, R. E., Ed., 1959: Glossary of Meteorology. Amer. Meteor. Soc., 638 pp.

  • Kain, J. S., and J. M. Fritsch, 1993: Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 165–170.

  • Keyser, D., and T. N. Carlson, 1984: Transverse ageostrophic circulations associated with elevated mixed layers. Mon. Wea. Rev.,112, 2465–2478.

  • ——, B. D. Schmidt, and D. G. Duffy, 1992: Quasigeostrophic vertical motions diagnosed from along- and cross-isentrope components of the Q vector. Mon. Wea. Rev.,120, 731–741.

  • Koch, S. E., and J. McCarthy, 1982: The evolution of an Oklahoma dryline. Part II: Boundary-layer forcing of meso-convective systems. J. Atmos. Sci.,39, 237–257.

  • ——, A. Aksakal, and J. T. McQueen, 1997: The influence of mesoscale humidity and evapotranspiration fields on a model forecast of a cold frontal squall line. Mon. Wea. Rev.,125, 58–83.

  • Lanicci, J. M., and T. T. Warner, 1997: A case study of lid evolution using analyses of observational data and a numerical model simulation. Wea. Forecasting,12, 228–252.

  • Lemone, M. A., M. Zhou, C.-H. Moeng, D. H. Lenschow, L. J. Miller, and R. L. Grossman, 1999: An observational study of wind profiles in the baroclinic convective mixed layer. Bound.-Layer Meteor.,90, 47–82.

  • Locatelli, J. D., J. E. Martin, J. A. Castle, and P. V. Hobbs, 1995a: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part III: The development of a squall line associated with weak cold-frontogenesis aloft. Mon. Wea. Rev.,123, 2641–2662.

  • ——, ——, and P. V. Hobbs, 1995b: Development and propagation of precipitation cores on cold fronts. Atmos. Res.,38, 177–206.

  • ——, M. T. Stoelinga, R. D. Schwartz, and P. V. Hobbs, 1997: Surface convergence induced by cold fronts aloft and pre-frontal surges. Mon. Wea. Rev.,125, 2808–2820.

  • ——, ——, and P. V. Hobbs, 1998: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part V: Thermodynamic and dual-Doppler radar analysis of a squall line associated with a cold front aloft. Mon. Wea. Rev.,126, 860–875.

  • Lynn, B. H., W.-K. Tao, and P. J. Wetzel, 1998: A study of landscape-generated deep moist convection. Mon. Wea. Rev.,126, 928–942.

  • Martin, J. E., 1999: Quasigeostrophic forcing of ascent in the occluded sector of cyclones and the trowal airstream. Mon. Wea. Rev.,127, 70–88.

  • ——, J. D. Locatelli, P. V. Hobbs, P.-Y. Wang, and J. A. Castle, 1995:Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part I: A synoptic-scale rainband associated with a lee dryline and lee trough. Mon. Wea. Rev.,123, 241–264.

  • Miller, L. J., M. A. LeMone, W. Blumen, R. L. Grossman, N. Gamage, and R. J. Zamora, 1996: The low-level structure and evolution of a dry arctic front over the central United States. Part I: Mesoscale observations. Mon. Wea. Rev.,124, 1648–1675.

  • Neiman, P. J., and R. M. Wakimoto, 1999: The interaction of a Pacific cold front with shallow air masses east of the Rocky Mountains. Mon. Wea. Rev.,127, 2102–2127.

  • ——, F. M. Ralph, M. A. Shapiro, B. F. Smull, and D. Johnson, 1998:An observational study of fronts and frontal mergers over the continental United States. Mon. Wea. Rev.,126, 2521–2554.

  • Newton, C. W., 1950: Structure and mechanism of the prefrontal squall line. J. Meteor.,7, 210–222.

  • Ogura, Y., H.-M. Juang, K.-S. Zhang, and S.-T. Soong, 1982: Possible triggering mechanisms for severe storms in SESAME-AVE IV (9–10 May 1979). Bull. Amer. Meteor. Soc.,63, 503–515.

  • 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.

  • Reisner, J., R. M. Rasmussen, and R. T. Bruintjes, 1998: Explicit forecasting of supercooled liquid water in winter storms using the MM5 mesoscale model. Quart. J. Roy. Meteor. Soc.,124, 1071–1107.

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

  • Rotunno, R., J. B. Klemp, and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci.,45, 463–485.

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

  • Shapiro, M. A., 1982: Mesoscale weather systems of the central United States. CIRES Tech. Rep., 78 pp. [Available from Cooperative Institute in Environmental Sciences, University of Colorado/NOAA, Boulder, CO 80309].

  • Shaw, B. L., R. A. Pielke, and C. L. Ziegler, 1997: A three-dimensional numerical simulation of a Great Plains dryline. Mon. Wea. Rev.,125, 1489–1506.

  • Steenburgh, W. J., and C. F. Mass, 1994: The structure and evolution of a simulated Rocky Mountain lee trough. Mon. Wea. Rev.,122, 2740–2761.

  • Sun, W.-Y., and C.-C. Wu, 1992: Formation and diurnal variation of the dryline. J. Atmos. Sci.,49, 1606–1619.

  • Sutcliffe, R. C., 1947: A contribution to the problem of development. Quart. J. Roy. Meteor. Soc.,73, 370–383.

  • Thorpe, A. J., M. J. Miller, and M. W. Moncrieff, 1982: Two-dimensional convection in nonconstant shear: a model of midlatitude squall lines. Quart. J. Roy Meteor. Soc.,108, 739–762.

  • Trier, S. B., D. B. Parsons, and J. E. H. Clark, 1991: Environment and evolution of a cold-frontal mesoscale convective system. Mon. Wea. Rev.,119, 2429–2455.

  • Wang, P. Y., J. E. Martin, J. D. Locatelli, and P. V. Hobbs, 1995: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part II: Arctic fronts. Mon. Wea. Rev.,123, 1328–1344.

  • Weisman, M. L., W. C. Skamarock, and J. B. Klemp, 1997: The resolution dependence of explicitly modeled convective systems. Mon. Wea. Rev.,125, 527–548.

  • 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.

  • ——, and C. K. Mueller, 1993: Nowcasts of thunderstorm initiation and evolution. Wea. Forecasting,8, 113–131.

  • Zhang, D. L., and R. A. Anthes, 1982: A high-resolution model of the planetary boundary layer—Sensitivity tests and comparisons with SESAME-79 data. J. Appl. Meteor.,21, 1594–1609.

  • Ziegler, C. L., T. J. Lee, and R. A. Pielke Sr., 1997: Convective initiation at the dryline: A modeling study. Mon. Wea. Rev.,125, 1001–1026.

Fig. 1.
Fig. 1.

(a) Surface analysis valid at 0600 UTC 9 Mar 1992, with sea level pressure (solid contours, hPa), frontal positions (heavy solid lines), dryline position (heavy dashed line), and station data plotted according to the standard convention (temperature and dewpoint in °C). (b) The 500-hPa analysis valid at 0500 UTC 9 Mar 1992, with geopotential heights (thin solid contours, dam), temperature (heavy solid contours, °C), and station data plotted according to the standard convention. Dashed line indicates the leading edge of strong cold advection

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 2.
Fig. 2.

Infrared satellite image valid at 0400 UTC 9 Mar 1992, with an overlaid contour plot of radar reflectivity from a composite of low-elevation scans of all radars in the vicinity. Solid and dashed contours represent reflectivity values of 20 and 40 dBZ, respectively. Precipitation features between white arrows and black arrows are the CFA rainband and the pre–dry trough rainband, respectively

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 3.
Fig. 3.

Results from the 18-h model simulation on the 25-km domain, valid at 0600 UTC 9 Mar 1992. (a) Surface analysis of sea level pressure (solid contours, hPa), with subjectively analyzed frontal positions (heavy solid lines) and dryline position (heavy dashed line). (b) The 500-hPa analysis with geopotential heights (thin solid contours, dam) and temperature (heavy solid contours, °C). Dashed line indicates the leading edge of strong cold advection

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 4.
Fig. 4.

Simulated infrared satellite image based on temperature at optical depth = 1 (temperature scale shown at bottom) from the 16-h forecast on the 8.3-km domain, valid at 0400 UTC 9 Mar 1992, with an overlaid contour plot of simulated reflectivity factor based on rain and snow mixing ratios averaged through the lowest 3 km from the model simulation. Solid and dashed contours represent reflectivity values of 10 and 30 dBZ, respectively. Precipitation features between white arrows and black arrows are the simulated CFA rainband and pre–dry trough rainband, respectively

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 5.
Fig. 5.

The Q-vector plot at 750 hPa valid at 0000 UTC 9 Mar 1992, based on output from the 75-km model simulation at 12 h. Gray-shaded bands labeled C and P are locations of initiation of the CFA and pre–dry trough rainbands, respectively, in the model simulation. Also shown in (a) are geopotential height (heavy solid contours, dam) and temperature (thin solid contours, °C); and in (b), Q-vector divergence (contour interval of 1 × 10−14 m kg−1 s−1, with heavy dashed, thin solid, and heavy solid contours denoting positive divergence, zero divergence, and convergence, respectively). Note, upward motion is implied by convergence (heavy solid contours)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 6.
Fig. 6.

Quasigeostrophic vertical velocity (ω) partitions at 750 hPa valid at 0000 UTC 9 Mar 1992, based on output from the 75-km model simulation at 12 h [contour interval of 1 dPa s−1, with heavy dashed, thin solid, and heavy solid contours denoting positive (downward), zero, and negative (upward) ω, respectively]. Gray-shaded bands labeled C and P are locations of initiation of the CFA and pre–dry trough rainbands, respectively, in the model simulation. Separate panels depict QG vertical velocity due to (a) Q-vector forcing, (b) topographic lower boundary condition, (c) Ekman lower boundary condition, and (d) the sum of (a), (b), and (c)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 7.
Fig. 7.

Ertel potential vorticity (solid contours, PVU) and wind vectors (scale at upper right) on the 310-K isentropic surface valid at 0000 UTC 9 Mar 1992, based on output from the 75-km model simulation at 12 h. Gray-shaded bands labeled C and P are locations of initiation of the CFA and pre–dry trough rainbands, respectively, in the model simulation, and L is the location of the simulated low pressure center at the surface

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 8.
Fig. 8.

Vertical cross section valid at 0500 UTC 9 Mar from the 8.3-km model simulation at 17 h, along the line A–B in Fig. 4, showing potential temperature (thin solid contours, K) and precipitation (rain + snow) mixing ratio (heavy solid lines, contour interval 0.1 g kg−1). Circulation vectors in the plane of the cross section shown in (a) and (c) (scale at lower right). Gray shading is equivalent potential temperature in (a), CAPE in (b), and Ertel potential vorticity in (c); units and shading values are shown underneath each panel. Also shown in (b) are two 16-h trajectories from 1300 UTC 8 Mar to 0500 UTC 9 Mar. Arrowheads mark hourly positions. Trajectories terminate exactly in the cross section, and remained within 150 km of the cross section throughout the trajectory integration period. Heavy dashed contours in (c) are cloud water mixing ratio (contour interval 0.1 g kg−1)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 9.
Fig. 9.

Vertical cross section valid at 0100 UTC 9 Mar from the 8.3-km model simulation at 13 h, along the line C–D in Fig. 4, showing potential temperature (thin solid contours, K), circulation vectors in the plane of the cross section (scale at lower right), and equivalent potential temperature (gray shading, legend at bottom of figure). Also shown are the surface location of the dryline (D) and the two baroclinic zones described in the text (heavy dashed lines marked BZ1 and BZ2)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 10.
Fig. 10.

Vertical cross section from the 8.3-km model simulation, along the line C–D in Fig. 4, at (a) 9, (b) 13, and (c) 17 h into the simulation, showing potential temperature (thin solid contours, K), the along-cross-section component of the gradient of potential temperature (gray shading, legend at bottom), net ascent of 3-h trajectories ending at every grid point below approximately 700 hPa (hPa, medium contours, dashed are positive/descent, solid are negative/ascent), and precipitation (rain + snow) mixing ratio (heavy solid lines, contour interval 0.1 g kg−1). Also shown are the surface location of the dryline (D) and the two baroclinic zones described in the text (heavy dashed lines marked BZ1 and BZ2)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 11.
Fig. 11.

Vertical cross section valid at 0100 UTC 9 Mar from the 8.3-km model simulation at 13 h, along the line C–D in Fig. 4, showing potential temperature (thin solid contours, K), circulation vectors in the plane of the cross section (scale at lower right), and CAPE (gray shading, legend at bottom). Also shown are the surface location of the dryline (D) and the two baroclinic zones described in the text (heavy dashed lines marked BZ1 and BZ2)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 12.
Fig. 12.

Vertical cross section valid at 0100 UTC 9 Mar from the 8.3-km model simulation at 13 h, along the line C–D in Fig. 4, showing potential temperature (thin solid contours, K) and vertical velocity (ω) fields [contour interval of 1 dPa s−1, with heavy dashed, thin solid, and heavy solid contours denoting positive (downward), zero, and negative (upward) ω, respectively]. (a) The QG vertical velocity based on the QG diagnosis from the 75-km simulation; (b) full vertical velocity from the 8.3-km simulation, unsmoothed; and (c) full vertical velocity from the 8.3-km simulation, smoothed to the same scale as the input fields in the 75-km QG diagnosis

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Fig. 13.
Fig. 13.

Vertical cross sections from the 8.3-km model simulation (a) along line C–D in Fig. 4 at 13 h and (b) along line C–D in Fig. 4 at 17 h, showing potential temperature (thin solid contours, K), cloud water mixing ratio (heavy dashed lines, contour interval 0.1 g kg−1), circulation vectors in the plane of the cross section (scale at lower right), and Ertel potential vorticity (gray shading, legend at bottom)

Citation: Monthly Weather Review 128, 10; 10.1175/1520-0493(2001)129<3481:SAEOWC>2.0.CO;2

Table 1.

Pressure, vertical (ω), velocity, and covective inhibition along trajectory number 1 in Fig. 8b. Actual trajectory ascent from hours 0–16 was 129 hPa. Estimated trajectory ascent for the same period from integration of QG ω along the trajectory: 111 hPa

Table 1.

1

It should be noted that east of the low pressure center the feature referred to here as the arctic front is equivalent to the more complex “hybrid front” analyzed by Neiman et al. (1998), which they showed had previously formed from a merger between a true arctic front, a quasistationary polar front, and a maritime warm front.

Save