• Avissar, R., and R. A. Pielke, 1989: A parameterization of heterogeneous land surfaces for atmospheric numerical models and its impact on regional meteorology. Mon. Wea. Rev.,117, 2113–2136.

  • Barlow, W., 1996: An examination of the Kansas derecho on July 1, 1994. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 527–530.

  • Bernstein, B. C., and R. H. Johnson, 1994: A dual-Doppler radar study of an OK PRE-STORM heat burst event. Mon. Wea. Rev.,122, 259–273.

  • Blackadar, A., 1957: Boundary layer wind maxima and their significance for the growth of nocturnal inversion. Bull. Amer. Meteor. Soc.,38, 283–290.

  • Colman, B. R., 1990a: Thunderstorms above frontal surfaces in environments without positive CAPE. Part I: 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.

  • Cooper, S. R., and M. L. Bentley, 1996: A case study of the 8–9 July 1993 Great Plains derecho. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 531–535.

  • Cotton, W. R., and R. A. Anthes, 1989: Storm and Cloud Dynamics. Academic Press, 878 pp.

  • ——, G. D. Alexander, R. Hertenstein, R. L. Walko, R. L. McAnelly, and M. Nicholls, 1995: Cloud venting—A review and some new global annual estimates. Earth-Sci. Rev.,39, 169–206.

  • Davies, H. C., 1983: Limitations of some common lateral boundary schemes used in regional NWP models. Mon. Wea. Rev.,111, 1002–1012.

  • Gal-Chen, T., and R. C. J. Sommerville, 1975: On the use of a coordinate transformation for the solution of the Navier–Stokes equations. J. Comput. Phys.,17, 209–228.

  • Grasso, L. D., 1996: Numerical simulation of the May 15 and April 26, 1991 tornadic thunderstorms. Colorado State University Atmospheric Science Paper 596, 151 pp. [Available from Department of Atmospheric Science, Colorado State University, Fort Collins, CO 80523.].

  • Hill, G. E., 1974: Factors controlling the size and spacing of cumulus clouds as revealed by numerical experiments. J. Atmos. Sci.,31, 646–673.

  • Johns, R. H., and W. D. Hirt, 1987: Derecho: Widespread convectively induced windstorms. Wea. Forecasting,2, 32–49.

  • Klemp, J. B., and R. B. Wilhelmson, 1978: The simulation of three-dimensional convective storm dynamics. J. Atmos. Sci.,35, 1070–1096.

  • Knupp, K. R., 1988: Downdrafts within High Plains cumulonimbi. Part II: Dynamics and thermodynamics. J. Atmos. Sci.,45, 3965–3982.

  • ——, and W. R. Cotton, 1985: Precipitating convective cloud downdraft structure: An interpretive survey. Rev. Geophys.,23, 183–215.

  • Kreis, K. W., 1995: Synoptic and mesoscale severe weather: A two day case study of a derecho and local hailstorms. M.A. thesis, Dept. of Marine, Earth and Atmospheric Science, North Carolina State University at Raleigh, 93 pp. [Available from Dept. of Marine, Earth and Atmospheric Science, North Carolina State University at Raleigh, 1125 Jordan Hall, Box 8208, Raleigh, NC 27695.].

  • Laing, A. G., 1997: The large scale environment of mesoscale convective complexes: Comparison with other deep convective weather systems. Preprints, 22d Conf. on Hurricanes and Tropical Meteorology, Fort Collins, CO, Amer. Meteor. Soc., 415–416.

  • ——, and J. M. Fritsch, 1997: The global population of mesoscale convective complexes. Quart. J. Roy. Meteor. Soc.,123, 389–405.

  • Lilly, D. K., 1962: On the numerical simulation of buoyant convection. Tellus,14, 148–172.

  • Louis, J. F., M. Tiedke, and J.-F. Geleyn, 1982: A short history of the PBL parameterization at the ECMWF. Proc. Workshop on Planetary Boundary Layer Parameterization, Shinfield Park, Reading, United Kingdom, European Centre for Medium-Range Weather Forecasts, 59–79.

  • Loveland, T. R., J. W. Merchant, D. O. Ohlen, and J. F. Brown, 1991:Development of a land-cover characteristics database for the conterminous U.S. Photo. Eng. Remote Sens.,57, 1453–1463.

  • Maddox, R. A., 1979: Synoptic and meso-α scale aspects of flash flood events. Bull. Amer. Meteor. Soc.,60, 115–123.

  • ——, 1980: Mesoscale convective complexes. Bull. Amer. Meteor. Soc.,61, 1374–1387.

  • ——, 1983: Large-scale meteorological conditions associated with midlatitude, mesoscale convective complexes. Mon. Wea. Rev.,111, 1475–1493.

  • Mahrer, Y., and R. A. Pielke, 1977: A numerical study of the airflow over irregular terrain. Beitr. Phys. Atmos.,50, 98–113.

  • McAnelly, R. L., J. E. Nachamkin, W. R. Cotton, and M. E. Nicholls, 1997: Upscale evolution of MCSs: Doppler radar analysis and analytical investigation. Mon. Wea. Rev.,125, 1083–1110.

  • McNider, R. T., and R. A. Pielke, 1981: Diurnal boundary-layer development over sloping terrain. J. Atmos. Sci.,38, 2198–2212.

  • McNulty, R. P., 1995: Severe and convective weather: A central region forecasting challenge. Wea. Forecasting,10, 187–202.

  • Messinger, F., and A. Arakawa, 1976: Numerical Methods in Atmospheric Models. GARP Publication Series, Vol. I, World Meteorological Organization, 64 pp.

  • Moller, A. R., C. A. Doswell, M. P. Foster, and G. R. Woodall, 1994:The operational recognition of supercell thunderstorm environments and storm structures. Wea. Forecasting,9, 327–347.

  • National Climate Data Center, 1985: Storm Data. Vol. 25, No. 5, 66 pp.

  • Pielke, R. A., and Coauthors, 1992: A comprehensive meteorological modeling system—RAMS. Meteor. Atmos. Phys.,49, 69–91.

  • Przybylinski, R. W., and D. M. DeCaire, 1985: Radar signatures associated with the derecho, a type of mesoscale convective system. Preprints, 14th Conf. on Severe Local Storms, Indianapolis, IN, Amer. Meteor. Soc., 228–231.

  • Rochette, S. M., J. T. Moore, P. S. Market, F. H. Glass, and D. L. Ferry, 1995: Heavy rain events associated with elevated thunderstorms in the midwest. Postprints, Fourth National Winter Weather Workshop, Kansas City, MO, NOAA, 25-1–25-12.

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

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

  • Schmidt, J. M., 1991: Numerical and observational investigation of long-lived, MCS-induced, severe surface wind events: The derecho. Colorado State University Atmospheric Science Paper 479, 196 pp. [Available from Department of Atmospheric Science, Colorado State University, Fort Collins, CO 80523.].

  • ——, and W. R. Cotton, 1989: A High Plains squall line associated with severe surface winds. J. Atmos. Sci.,46, 281–302.

  • ——, and ——, 1990: Interactions between upper and lower tropospheric gravity waves on squall line structure and maintenance. J. Atmos. Sci.,47, 1205–1222.

  • Smagorinsky, J., 1963: General circulation experiments with the primitive equations. Part I: The basic experiment. Mon. Wea. Rev.,91, 99–164.

  • Staudenmaier, M. J., Jr., 1993: A multiscale study on the development and evolution of the 7–8 July 1991 derecho. M.A. thesis, Dept. of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, 160 pp. [Available from Dept. of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, 1225 W. Dayton St., Madison, WI 53706.].

  • Trembeck, C. J., and R. Kessler, 1985: A surface temperature and moisture parameterization for use in mesoscale numerical models. Preprints, Seventh Conf. on Numerical Weather Prediction, Montreal, PQ, Canada, Amer. Meteor. Soc., 355–358.

  • Trier, S. B., and D. B. Parsons, 1993: Evolution of environmental conditions preceding the development of a nocturnal mesoscale convective complex. Mon. Wea. Rev.,121, 1078–1098.

  • Tripoli, G. J., and W. R. Cotton, 1989: Numerical study of an observed orogenic mesoscale convective system. Part 1: Simulated genesis and comparison with observations. Mon. Wea. Rev.,117, 273–304.

  • Walko, R. L., W. R. Cotton, M. P. Meyers, and J. Y. Harrington, 1995a: New RAMS cloud microphysics parameterization. Part I: The single-moment scheme. Atmos. Res.,38, 29–62.

  • ——, C. J. Tremback, and R. F. A. Hertenstein, 1995b: RAMS—The Regional Atmospheric Modeling System: Version 3b user’s guide. Aster, 117 pp. [Available from Aster Inc., P.O. Box 466, Fort Collins, CO 80522.].

  • Weisman, M. L., 1993: The genesis of severe, long-lived bow echoes. J. Atmos. Sci.,50, 645–670.

  • View in gallery

    Plan position indicator (PPI) from Limon, CO, radar at 0330 UTC on 13 May. Contours begin at 15.0 dBZ and are at 7.5-dBZ intervals. Shading begins at 15.0-dBZ and increases at 15.0-dBZ intervals.

  • View in gallery

    Radar summary at 0535 UTC on 13 May.

  • View in gallery

    High wind reports between 0534 and 1130 UTC from Storm Data (National Climate Data Center 1985).

  • View in gallery

    Surface analysis at 0000 UTC on 13 May.

  • View in gallery

    Location of the grids at 2000 UTC: (a) grids 1, 2, and 3; (b) grids 3 and 4.

  • View in gallery

    Equivalent potential temperature (shaded) (K), condensate mixing ratio (contoured) (g kg−1), and horizontal wind vectors at σ = 48 m and 0000 UTC on grid 2. The dashed vertical line along long = 97.5°W indicates the position of the cross section depicted in Fig. 7.

  • View in gallery

    Vertical cross section through long = 97.5°W of potential temperature (K) at 0000 UTC on grid 2.

  • View in gallery

    Temperature (shaded) (°C), condensate mixing ratio (contoured) (g kg−1), and horizontal wind vectors at σ = 48 m and 0000 UTC on grid 4.

  • View in gallery

    Vertical cross section through lat = 37.5°N of vertical velocity (shaded) (m s−1) and equivalent potential temperature (contoured) (K) at 0000 UTC on grid 4.

  • View in gallery

    Vertical vorticity (shaded) (×1000 s−1) and pressure perturbation (contoured) (hPa) at σ = 3381 m and 0000 UTC on grid 4.

  • View in gallery

    Vertical cross section through lat = 37.6°N of condensate mixing ratio (shaded) (g kg−1), hail mixing ratio (solid contours) (g kg−1), and rain mixing ratio (dashed contours) (g kg−1) at 0000 UTC on grid 4.

  • View in gallery

    Condensate mixing ratio (g kg−1) and horizontal wind vectors at σ = 48 m on grid 4 at (a) 0200 UTC, (b) 0400 UTC, and (c) 0600 UTC.

  • View in gallery

    Time series of winds at σ = 48 m on grid 4. The closed circles represent the maximum speed in the horizontal domain, and the open circles the maximum speed found in convective outflow. After 0300 UTC they coincide.

  • View in gallery

    Vertical cross section through long = 97.5°W of potential temperature (K) at 0600 UTC on grid 2.

  • View in gallery

    Vertical profile at lat = 38°N, long = 98°W at times 0000 and 0600 UTC from grid 2 of (a) meridional wind (m s−1) and (b) equivalent potential temperature (K).

  • View in gallery

    Vertical profile of observations interpolated to lat = 38°N, long = 98°W at times 0000 and 1200 UTC of (a) meridional wind (m s−1) and (b) equivalent potential temperature (K).

  • View in gallery

    Temperature (°C) at σ = 48 m at 0600 UTC on grid 4.

  • View in gallery

    Vertical cross section through lat = 38.1°N of vertical velocity (shaded) (m s−1) and equivalent potential temperature (contoured) (K) at 0600 UTC on grid 4.

  • View in gallery

    Perturbation pressure (shaded) (hPa), vertical velocity (contoured) (m s−1), and horizontal wind vectors at σ = 48 and 0600 UTC on grid 4.

  • View in gallery

    Vertical vorticity (×1000 s−1) (shaded), perturbation pressure (hPa) (contoured), and horizontal wind vectors at σ = 3381 m and 0600 UTC on grid 4.

  • View in gallery

    Conceptual model of the (a) daytime and (b) nighttime phases of the MCS. The three planes are located at the surface, 700-m, and 2500-m height. The shaded field in each plane is the condensate mixing ratio and the thin arrows are ground-relative winds. The thick arrows represent the main branches of the updraft and of the downdraft. The plot on the left of each sketch depicts the vertical profile of equivalent potential temperature. In (b) the locations of the upward π-gradient force and condensate loading that act on the up–down downdraft are noted.

  • View in gallery

    Diagram illustrating properties of idealized updraft and downdraft parcels at (a) daytime and (b) nighttime. The trajectory location with respect to the cloud is shown on the left. The top right sketches are skew T diagrams, illustrating thermodynamic properties of the environment and of the parcels, whose trajectories are represented by dots. The bottom-right sketches are vertical profiles of total buoyancy (including vapor effects and condensate loading) along the parcels’ paths.

  • View in gallery

    Time series (0530–0600 UTC) along the path of a nighttime up–down downdraft trajectory of (a) height, vertical velocity, and horizontal speed; (b) terms of the vertical momentum equation (5); and (c) parcel and ambient potential temperature.

  • View in gallery

    (a) Pressure perturbation (hPa) from the model; (b) sum of forcing terms of the pressure perturbation equation; (c) vertical vorticity forcing; (d) horizontal vorticity forcing; (e) deformation forcing; (f) fluid extension forcing; (g) linear forcing; (h) buoyancy forcing at 0600 UTC and σ = 3381 m. All forcing fields are in 10−4 s−2.

  • View in gallery

    (Continued)

  • View in gallery

    Time series (0530–0600 UTC) for a nighttime up–down downdraft parcel. (a) Terms of the zonal momentum equation; (b) terms of the meridional momentum equation.

  • View in gallery

    Pressure perturbation (hPa) at (a) 0549 UTC and σ = 396 m and (b) 0557 UTC and σ = 48 m. The thick line is the projection of the trajectory of a nighttime trajectory from 0530 to 0600 UTC. The circle marks the earliest position of the parcel, and the square marks the position of the parcel at the time the pressure field is depicted.

  • View in gallery

    (a) Pressure perturbation (hPa) obtained from the model (shaded) (hPa) and sum of forcing terms of the pressure perturbation equation (contoured) (10−4 s−2); (b) fluid extension term of the pressure perturbation equation (10−4 s−2) at 0600 UTC and σ = 48 m.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 31 31 5
PDF Downloads 22 22 3

Multiscale Evolution of a Derecho-Producing Mesoscale Convective System

View More View Less
  • 1 Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado
© Get Permissions
Full access

Abstract

In this paper the authors address one type of severe weather: strong straight-line winds. The case of a mesoscale convective system that developed in eastern Colorado on 12–13 May 1985 was studied. The system formed in the afternoon, was active until early morning, and caused strong winds during the night.

A multiscale nonhydrostatic full physics simulation was performed to formulate a conceptual model of the main airflow branches of the system, and to gain understanding of the physical processes involved in the strong wind generation in this storm. Four telescopically nested grids covering from the synoptic-scale down to cloud-scale circulations were used. A Lagrangian model was employed to follow trajectories of parcels that took part in the updraft and downdraft, and balances of forces were computed along the trajectories.

The strong nocturnal winds were caused by downdrafts reaching the surface and by a dynamically forced horizontal pressure gradient force. The most important branch of the downdraft had an “up–down” trajectory. Parcels originated close to the ground, were lifted up by a strong upward-directed pressure gradient force, and became colder than their surroundings as they ascended in a stable environment. Then, as they went through the precipitation shaft, they sank due to negative buoyancy enhanced by condensate loading. The upward pressure gradient force was partially related to midlevel perturbation vorticity in the storm.

* Current affiliation: NOAA/Forecast Systems Laboratory, Boulder, Colorado.

Corresponding author address: Dr. Lígia R. Bernardet, NOAA/ERL, 325 Broadway R/E/FS1, Boulder, CO 80303-3328.

Email: ligia@fsl.noaa.gov

Abstract

In this paper the authors address one type of severe weather: strong straight-line winds. The case of a mesoscale convective system that developed in eastern Colorado on 12–13 May 1985 was studied. The system formed in the afternoon, was active until early morning, and caused strong winds during the night.

A multiscale nonhydrostatic full physics simulation was performed to formulate a conceptual model of the main airflow branches of the system, and to gain understanding of the physical processes involved in the strong wind generation in this storm. Four telescopically nested grids covering from the synoptic-scale down to cloud-scale circulations were used. A Lagrangian model was employed to follow trajectories of parcels that took part in the updraft and downdraft, and balances of forces were computed along the trajectories.

The strong nocturnal winds were caused by downdrafts reaching the surface and by a dynamically forced horizontal pressure gradient force. The most important branch of the downdraft had an “up–down” trajectory. Parcels originated close to the ground, were lifted up by a strong upward-directed pressure gradient force, and became colder than their surroundings as they ascended in a stable environment. Then, as they went through the precipitation shaft, they sank due to negative buoyancy enhanced by condensate loading. The upward pressure gradient force was partially related to midlevel perturbation vorticity in the storm.

* Current affiliation: NOAA/Forecast Systems Laboratory, Boulder, Colorado.

Corresponding author address: Dr. Lígia R. Bernardet, NOAA/ERL, 325 Broadway R/E/FS1, Boulder, CO 80303-3328.

Email: ligia@fsl.noaa.gov

1. Introduction

The mechanisms for the development of persistent straight line winds associated with convection (derechos) are not yet well understood. A number of cases have been documented in the literature, but the dynamics that differentiate these cases from “ordinary” downburst cases and from severe storms that do not produce strong winds is unclear. It is important to understand these systems since, just in the United States, an average of 17 cases occur in each warm season (Johns and Hirt 1987).

Insight into these systems can be gained by looking into past cases. The climatology of derecho events shows that they usually occur in the warm season, on the cold side of a thermal boundary, usually a weak stationary summertime front (Johns and Hirt 1987). Several case studies also point to the existence of a thermal boundary and to the nocturnal character of the systems (Barlow 1996; Cooper and Bentley 1996; Kreis 1995; Staudenmaier 1993).

Derecho events are often associated with mesoscale convective complexes (Maddox 1980), which have been shown to be predominantely nocturnal (Laing and Fritsch 1997), and to form on the cold side of frontal discontinuities (Maddox 1983) when a low-level jet (LLJ) is present (Maddox 1983; Laing and Fritsch 1997).

Although no systematic study has been conducted to verify this, the papers above highlight an important relation between derechos and elevated storms, characterized by having their source of updraft air located above a stable planetary boundary layer (PBL). Elevated storms do not necessarily happen at night, but can occur whenever the PBL in a region is stable and a source of positively buoyant air exists at an upper level. This can happen in a localized region, for example, when a convective system leaves behind a cold pool. The next storm will not feed in that cool air, and may become an elevated thunderstorm if a source of positively buoyant air is present, as was observed by Schmidt and Cotton (1989) in the 2–3 August 1981 derecho case over the Dakotas.

McNulty (1995) stated that thunderstorms not based on the boundary layer are one of the most challenging forecast problems for the National Weather Service. One reason is that elevated thunderstorms have been identified as severe weather producers. Rochette et al. (1995) performed a composite study of seven heavy-rain producing mesoscale convective systems (MCSs) that developed in the Midwest and that had characteristics of elevated thunderstorms. They developed in different times of the day, but all formed on the cold side of a weak frontal system, a region in which the lower atmosphere was stable. Their analysis of the mesoscale environment for this type of storm showed that the source of instability for the systems was the LLJ, which advected high equivalent potential temperature (θe) air over the genesis region, and which is commonly observed in association with elevated storms (e.g., Laing 1997; Trier and Parsons 1993). They also identified some upper-level support for the system, namely a shortwave and cyclonic vorticity advection at 500 hPa and a region of divergence at 200 hPa. Convection over a stable layer also accounted for 25% of the cases of flash floods in the United States studied by Maddox (1979). Colman (1990a,b) did a climatology of elevated thunderstorms in the United States. He noted that they usually form on the north side of a thermal boundary and that storms associated with a stationary front occurred more often at 1200 UTC than at 0000 UTC. He examined several preconvective thermodynamic profiles and noted that in some cases parcels up to 250 hPa above the surface had very little convective available potential energy (CAPE), and that the instability in which the storms fed was not convective but symmetric.

In this paper we will use the 12–13 May 1985 case to explore multiscale aspects of derecho development and to investigate the role of the stable boundary layer in derecho maintenance. The outline of this paper is as follows. In section 2 a review of the observations is presented. A description of the configuration of the numerical model used in this study is presented in section 3. The evolution of the simulation is presented in section 4, with emphasis on both the mesoscale environment and the cloud-scale results. Section 5 describes the techniques used to analyze the simulation results: a pressure perturbation decomposition and a Lagrangian model. In section 6 an analysis of circulation branches in the convective system is presented, together with the mechanism for strong wind development in this case. The main conclusions are synthesized in section 7.

2. Case description

In this section the evolution of the MCS that was studied is described using observations from surface and upper-air stations, radar, and special observations recorded in Storm Data (National Climate Data Center 1985). This MCS was chosen because it was an event of severe straight-line winds, that transformed from a daytime to a nighttime system.

The strong winds in this case were first recorded at 0600 UTC1 of 13 May 1985, but the convective system itself formed in the afternoon. At 0035 UTC only light precipitation existed over southeast Colorado (CO). At 0130 UTC the Limon radar showed two strong cells and by 0330 UTC (Fig. 1) several cells had formed in southeast CO, aligned approximately in the east–west direction. A region of stratiform precipitation was located to the north of the convective line, while the cells were moving toward the northeast.

As the system moved farther toward the northeast into Kansas (KS), it exited the coverage of the Limon radar. Its evolution can be followed using National Weather Service manually digitized radar summaries. At 0535 UTC (Fig. 2) the cells that had formed in CO had entered KS, and by 0835 UTC (not shown) the cells were located in northern KS. A tornado watch box was issued by the National Weather Service at 0735 UTC. At 1035 UTC some cells of the MCS were moving over the Kansas–Nebraska border. Convection started to weaken after sunrise.

The swath of damage left behind by this system can be seen in Fig. 3. Reports of strong winds began at 0530 UTC in Syracuse (KS) with 35 m s−1 winds, and progressed toward the northeast. The last reports came from the Kansas–Nebraska border at 1015 UTC, and were of 40 m s−1 gusts. One weak tornado was reported associated with this MCS in Scott City (KS) at 0605 UTC.

At 0000 UTC a semistationary surface front extended from the Texas (TX) panhandle through Oklahoma (OK) toward the east (Fig. 4). A low pressure center of 1000 hPa was located on the border of the Texas panhandle and New Mexico. A dryline extended south from the center of low pressure, separating the moist air to the east from the dry air to the west. The winds turned cyclonically around the low, therefore flow in southeast CO was from the northeast and in KS from the east. A weak center of high pressure was located over Nebraska.

At upper levels, a center of low geopotential was present at 850 hPa over the surface low, causing southwest winds over the MCS genesis region. At higher levels a closed low was not present, but a trough whose axis tilted westward with height existed. Therefore the mean wind in upper levels was from the southwest, and the MCS developed on the downwind side of a trough, a region prone to large-scale upward motion.

During the evening, the position of the surface low remained unaltered, but it deepened slightly and reached 998 hPa by 1200 UTC. The semistationary front moved slightly north and by 1200 UTC was along the KS–OK border. The high pressure center migrated to Minnesota and contributed to the establishment of easterly winds over KS, which fed the convective system.

At the 850- and 700-hPa levels the low was closed at 1200 UTC, the trough had moved east, and southerlies were present at those levels over KS. The winds at 850 hPa over OK increased from 10 m s−1 to 20 m s−1 from 0000 UTC to 1200 UTC. The thermodynamic profile over western KS at 1200 UTC was characterized by a very stable layer with northeast winds from the surface (920 hPa) to 800 hPa, and a layer of intermediate conditional instability from 800 to 720 hPa aloft, in contrast to the thermodynamic profile observed in the same location during the afternoon, that displayed a 200-hPa-deep well mixed boundary layer. The winds were from the southwest above 820 hPa and strengthened during the night. This evolution in the vertical profile has the signature of a low-level jet, and will be discussed further in section 5.

In summary, we will study an MCS that formed in the late afternoon in southeast CO and propagated toward the northeast during the night, generating strong winds. The nocturnal system evolved on top of a stable boundary layer, in the presence of a low-level jet, which characterizes it as an elevated system. In the next section, the model used to study this case will be described, and in section 4 modeling results will be presented.

3. Model description

The simulation was performed using the nonhydrostatic, compressible option of version 3b of the Regional Atmospheric Modeling System (RAMS) developed at Colorado State University. This model is very versatile and may be set up in many ways, depending on the intended application. In this section we will limit the description to the parts of the numerics and physics that were actually used in our simulations.

a. Experimental setup

Four interactive telescopically nested grids were used in this simulation, to capture circulations from the synoptic scale down to cloud scale. The coarsest three grids were initialized in the beginning of the simulation, 1200 UTC on 12 May 1985, while the fourth grid was initialized 8 h into the run, when convection started to occur explicitly on grid 3. All grids used a rotated polar-stereographic projection, in which the pole of the projection is located near the center of the domain of the coarsest grid (latitude = 42.1°N, longitude = 107.1°W).

The horizontal grid spacings used in grids 1–4 were 80, 40, 10, and 2 km, respectively. The type of grid used in the model is Arakawa-C (Messinger and Arakawa 1976). The location of the grids at the time they were initialized is displayed in Fig. 5.

All grids had 38 vertical levels, and the top of the domain was at a height of 21 km. The vertical coordinate used in the model follows the terrain and is called sigma (σ) (Gal-Chen and Sommerville 1975). The height of each σ level is given by
i1520-0493-126-11-2991-e1
where zb is the topographical height of a grid point, zt is the height of the top of the domain, and z is the physical height of a vertical level.

The vertical grid spacing for a column located over a zero terrain elevation was 100 m near the ground, stretching at a rate of 1.1, up to a maximum Δz of 1000 m. Therefore, above a height of 10 km, the vertical grid spacing was a constant 1000 m. In grid points located over elevated terrain the vertical grid spacing was less than that, as described by (1).

Below we list briefly some of the numerical and physical characteristics of the model. A more general description can be found in Pielke et al. (1992), Walko et al. (1995a,b) and references therein.

b. Model initialization, numerics, and physics

Grids 1–3 were initialized at 1200 UTC of 12 May 1985 and grid 4 at 2000 UTC. The model was integrated for 24 h. The data to initialize the model were obtained from the National Center for Atmospheric Research and consisted of three files: (i) surface observations; (ii) rawindsonde data; (iii) model data in isobaric surfaces [called pressure data, from time zero, used to initialize the National Meteorological Center (now known as National Centers for Environmental Prediction) model]. A United States Geological Survey topography dataset was also used to initialize the model.

The model was integrated using a hybrid scheme, which consists of leapfrog time differencing for the momentum and continuity equations and forward differencing for the thermodynamic and moisture equations. Advection was done using a second-order scheme. To save computer time, the time step was split. A short time step was used to compute terms in the prognostic equations that are responsible for the generation of acoustic waves (Klemp and Wilhelmson 1978). All other processes were computed using longer time steps of 90.0, 45.0, 22.5, and 7.5 s in grids 1–4, respectively.

The horizontal boundaries of grid 1 and the vertical boundary of all grids were updated using a nudging technique (Davies 1983), in which an extra term is added to the prognostic equations of the model. This extra term forces the variables computed in the model to relax to the observed state. Nudging was applied to five grid points in each horizontal border and to six grid points in the upper part of the domain. The lower boundary was supplied by the model’s surface parameterization, which employs the scheme of Louis et al. (1981) to compute the surface fluxes of heat, moisture, and momentum. The surface in each grid box was subdivided in percentages of water, vegetation, shaded soil, and bare soil. The fluxes were computed separately for each type of surface and then averaged. The soil model (Tremback and Kessler 1985) used 11 levels (going down to half a meter in depth) on which temperature and moisture were prognosed. The vegetation model (Avissar and Pielke 1989) was run using a variable vegetation initialization (Loveland et al. 1991), but soil moisture was initialized horizontally homogeneous. Top soil temperature was initialized to be identical to the atmospheric temperature in each grid column.

Grids 3 and 4 moved to follow the track of the MCS. A grid movement direction and speed was specified subjectively to best follow the convective system. Grid movement was accomplished by a redefinition of the grid location. Information in locations no longer covered by the fine grid was lost, and information in new regions of the fine grid was interpolated from the parent grid.

The radiation scheme used was developed by Mahrer and Pielke (1977). Both shortwave and longwave parameterizations take into account water vapor but ignore the effects of condensed water. The scheme takes into account the longitudinal variation of shortwave radiation reaching the top of the atmosphere. The tendencies of the thermodynamic equation due to radiation were updated every 15 min.

Eddy diffusion was parameterized using the Smagorinsky (1963) deformation-K scheme. The vertical diffusion scheme was modified by Hill (1974) to include a dependency on the Brunt–Väisälä frequency and by Lilly (1962) to include a dependency on the Richardson number. A minimal horizontal diffusion was used for numerical purposes even when physical diffusion was small or zero.

For the first 8 h of simulation, a simple microphysics parameterization was used in the three grids running. Only two water species had their mixing ratio prognosed: vapor and liquid. All water in excess of 100% relative humidity was instantaneously condensed, but precipitation was not allowed. For the remaining time of the simulation, a bulk microphysics parameterization (Walko et al. 1995a) was used in all four grids, with prognostic equations for the mixing ratios of rain, pristine ice, snow, aggregates, graupel, and hail. Cloud water was diagnosed and pristine ice mixing ratio was prognosed. No cumulus parameterization or warm bubbles were employed.

4. Evolution of the simulated MCS

In this section the multiscale evolution of the simulation of the 12–13 May 1985 MCS that led to the strong winds is presented. The description begins at 0000 UTC of 13 May, 12 h into the run, when the initial convection from the MCS had already begun.

a. MCS at 0000 UTC

1) Large–scale environment

At 0000 UTC the stationary front stretched along the OK–KS border (Fig. 6). Moist air was carried from the south by the southerlies and pooled over southern OK, accentuating the moisture contrast along the front. A strong east–west dewpoint gradient was present over TX, indicating the position of the dryline.

South of the front, surface winds were from the south, and north of the front they were from the east. Surface winds in southeast CO, the genesis region of the MCS, were southeast. They were driven by the synoptic patterns and enhanced by a mountain–plains solenoid that developed during the day on the eastern side of the Rocky Mountains. This solenoid has been identified before (e.g., Tripoli and Cotton 1989) as important in the generation of deep convection in eastern CO.

A north–south cross section of potential temperature (θ) on grid 2 is shown in Fig. 7. At 0000 UTC, before sunset, a well-mixed boundary layer with a height of 1.3 km was present over the whole domain. The boundary layer cooled toward the north. The position of the front is not easily identifiable in the potential temperature plot indicating that the front was better defined as a moisture contrast and wind shift than as a temperature contrast at this time.

Deep convection was already present in southeast CO at 0000 UTC. At the surface (Fig. 6), the condensate mixing ratio field was associated with the developing MCS and with convection in the western slope of the Rocky Mountains. Some condensation was also present in northern CO, associated with the upslope flow in that location.

2) MCS on grid 4 at 0000 UTC

In this section the characteristics of the MCS on the cloud scale are examined using grid 4 results.

Figure 8 shows the condensate mixing ratio at the surface at 0000 UTC. The incipient MCS, which developed slightly earlier in the model than in the observations, was composed of several cells. Two strong cells could be seen at 0000 UTC. The southern cell was the product of the merger of two previously independent cells, triggered by vertical motion that developed on the edge of each cell’s cold pool. The cold pools were quite strong: 4°C colder than the environment. This can be attributed to the large condensate mixing ratio at the ground, which had a maximum of 4.2 g kg−1, corresponding to a rain rate of 111.6 mm h−1.

The edge of the cold pool was very well defined on the east and southeast sides of the storm (Fig. 8) and poorly defined toward the back of the storm. The sharp gradient on the east side of the storm was due to the convergence of the environmental flow with the storm outflow. The environmental flow was from the southeast at 7.0 m s−1 and inflow to the storm occurred along its eastern leading edge. The outflow diverged from the center of the cold pool with a maximum speed of 17.6 m s−1 (Fig. 8). Divergence was stronger from the north part of the cold pool, and spread toward the north and back of the storm, and also against the southeast inflow.

To examine more closely the characteristics of the strongest cell, the next figures will be presented on a subdomain of grid 4. Another view of the cold pool is presented in Fig. 9, an east–west cross section through the coldest part of the pool, showing θe and vertical motion. The θe of the cold pool was 4 K lower than the environment’s. The vertical motion at low levels was located at the leading edge of the cold pool, indicating that triggering of vertical motion by the cold pool was an important mechanism in the dynamics of the storm at this time. The values of θe present in the cloud (∼326 K) were also present in the lower boundary layer ahead of the storm, suggesting that the source of air for the storm was located close to the surface. Although θe is not strictly a conservative variable due to ice physics, precipitation, and mixing, it may be used as a qualitative tracer.

The pressure perturbation2 pattern at the surface was composed of two mesohighs associated with condensational cooling due to the previously distinct storms, and one large mesolow ahead of the two mesohighs. At midlevels, a pressure perturbation minimum of 2 hPa lower than the surroundings was present in the center of the storm (Fig. 10), collocated with the northern part of a band of high relative vorticity (5.5 × 10−3 s−1), and extending ahead of it.

A cross section through lat = 37.6°N reveals the distribution of some microphysical species within the storm (Fig. 11). More than 8.0 g kg−1 of condensate existed in the core of the cloud at a height of 5 km. The liquid species were cloud water (not shown) and rain. Rain, only present below 2 km, was a large contribution to total condensate at low levels, and was the only microphysical species to reach the ground. Cloud water existed in the core of the updraft, up to a height of 9 km, where the temperature fell below −40°C, and homogeneous freezing took place. The most important species in the core of the cloud was hail, which is the only mixed-phase species in the parameterization. Melting and shedding of hail are the most important sources of rainwater. Although in smaller quantities, snow, aggregates, graupel, and pristine ice were important constituents of the mid- and upper-level anvil. The anvil’s top was at a height of 11 km and (using a threshold of 0.2 g kg−1) extended mainly to the frontal flank of the storm.

b. Evolution of the storm from 0000 to 0600 UTC

From 0000 to 0600 UTC the cell described in the last section persisted as a long-lived evolving entity. The first important evolution was the consolidation of the merger of the two cells that occurred before 0000 UTC. The northern cell in the merger became the dominant one, and the southern cell lost strength, so the system became more round, with what looked like an appendage on the southern end (0200 UTC) (Fig. 12a). Between 0200 and 0600 UTC, the cell grew in horizontal dimensions, to become more elongated in the north–south direction (Figs. 12b,c). The surface condensate mixing ratios for the dominant cell increased during this period, from a maximum of 4.2 g kg−1 (111.0 mm h−1) at 0000 UTC to a maximum of 5.4 g kg−1 (114 mm h−1) at 0600 UTC. The dominant cell moved from 252° with a speed of 12.6 m s−1. New cells were triggered along the convergence line created by the outflow from the main cell, on its southwest flank. This caused the convective region of the MCS to become aligned in the east–west direction by 0300 UTC.

A comparison with the radar observations (Fig. 1) shows that the model captured successfully the presence of three cells of deep convection on the CO–KS border. The area of light rain reaching the surface to the north of the line of convective cells is present but cannot be seen in Fig. 12 because of the contour interval used. The location of the anvil also compared well with the satellite image, which indicated deep clouds on the CO–KS border.

A time series of the strongest surface winds sampled every 30 min in grid 4 is shown in Fig. 13. The outflow winds increased almost steadily from 15.5 m s−1 at sunset (0230 UTC) to 25.0 m s−1 at 0330 UTC. The winds then remained high, and reached a maximum at 0600 UTC with 26.5 m s−1, decreasing after 0930 UTC. The observed strong winds reported by Storm Data started only at 0530 UTC and lasted through 1030 UTC, therefore the model was a few hours early in the production of strong winds.

c. MCS at 0600 UTC

1) Large-scale environment

By 0600 UTC (midnight local time) the temperatures had cooled considerably, both north and south of the front, which remained quasi-stationary along the KS–OK border. Significant changes occurred in the lower troposphere over northern TX, OK, and KS. Stronger southerly winds developed, causing northward advection of high-θe air.

Figure 14, a cross section of θ, shows that the PBL had become more stable. The development of the LLJ is easily seen in Fig. 15a, which shows a vertical profile of meridional winds at lat = 38°N and long = 98°W. Between 0000 and 0600 UTC, meridional winds increased at all heights due to the eastward movement of the upper-level trough. At 1200 m above ground level (AGL), the meridional wind increased by 8.5 m s−1. This had a large impact in θe, which increased by 19K at this level (Fig. 15b). The large θe increase occurred between the surface and 2500 m AGL. Above that level, the θe profile remained unaltered. Figure 15 can be compared with Fig. 16, which shows observed rawindsonde data interpolated to a model column at the same latitude and longitude. A comparison shows that the observed environment was well represented by the model, with a local maximum increase in meridional winds and θe between 1000 and 1500 m. The figures are useful only for a qualitative comparison, however, because there are no observations at 0600 UTC, so the 1200 UTC data were used.

2) MCS on grid 4 at 0600 UTC

At 0600 UTC, the MCS was composed of three cells aligned east–west in eastern KS (Fig. 12c). A cold pool was not present at this time. A close-up of the temperature field (Fig. 17) in the region of the leading, dominant cell shows that an area of warm temperatures was associated with the main downdraft of the dominant cell. Therefore, cold-pool dynamics was not important as a mechanism for storm propagation at this time. The downdraft was collocated with the pocket of high temperatures, which could therefore be attributed to subsidence warming. Similar warm spots have been observed by other authors (e.g., Bernstein and Johnson 1994) in situations of convection over a stable boundary layer.

One interesting characteristic of this cell can be seen in a cross section of θe (Fig. 18). In the convective cloud, a very high θe core was present at midlevels (∼328 K). At the surface, a region of relatively low θe was present (∼324 K), both ahead of and behind the storm. The low θe ahead of the storm, in low levels, indicated that surface air was not being ingested by the storm. This situation contrasts with that obtained at 0000 UTC, in which similar values of θe were found on the surface ahead of the storm and in the core of the storm. At this time, values of θe consistent with those found in the core of the storm existed at the top of the shallow stable layer, suggesting that the storm had an elevated source of air.

Strong surface outflow (26 m s−1) (Fig. 19) existed ahead of the region where the downdraft reached the ground. The pattern of the outflow was not as symmetric as it was at 0000 UTC. Instead, the prevalent outflow was from the northwest, converging with the southeast environmental winds ahead of the storm. Some of the divergent winds also had a northeast direction. Outflow to the north was almost nonexistent at this time. This pattern is consistent with that found in the 2–3 August 1981 derecho case documented by Schmidt and Cotton (1989) and modeled by Schmidt (1991).

Figure 19 also shows the pressure perturbation and the surface winds. The strongest outflow originated on a mesohigh located on the northwest side of the storm. One might find surprising that the mesohigh at the surface roughly corresponds to the pocket of warm air, because in an MCS the mesohigh is commonly associated with air cooled by evaporation. However, the main contribution to this center of high pressure was dynamic and not hydrostatic, as will be discussed in section 6.

At midlevels (3381 m), the core of the storm was dominated by a center of low pressure 4 hPa lower than the surroundings (Fig. 20) and a mesocyclone with vorticity of 1.0 × 10−2 s−1. The figure also indicates the development of a midlevel rear inflow jet, causing the storm to resemble a bow-echo high-precipitation supercell, which has been previously identified in the literature as a component of derecho events (Moller et al. 1994).

The transition between a daytime and a nighttime regime described above was mainly determined by radiative cooling and its influence on PBL characteristics. The radiative cooling became more important after sunset, when shortwave radiation at the surface ceased. The time series of the strongest surface winds (Fig. 13) shows that the strongest winds began steadily increasing at 0200 UTC. Considering that sunset in central KS at this time of the year occurs at 0230 UTC, a relationship is expected between the environment, the storm structure, and the strong outflow. This relationship will be explored in the next sections.

5. Analysis framework

In this section the techniques used to examine the source of air for the updrafts and downdrafts of the modeled convective system, and the physical processes that led to the strong surface winds are outlined. Subsection a introduces the decomposition of the pressure perturbation into six forcing terms. Subsections b and c go over the Lagrangian (parcel) model that was used to advect parcels backward in time, to assess the origin of the updraft and downdraft air. Along the trajectories, properties of the parcel and balances of forces were computed, which aided understanding of the path the parcels took and the physical mechanisms involved.

a. Pressure perturbation decomposition

To shed light on the origins of the mesolow located at midlevels of the storm, the pressure perturbation decomposition method formulated by Rotunno and Klemp (1982) was employed:
i1520-0493-126-11-2991-e2

In the equations above, the bar quantities represent a horizontally averaged reference state, while the prime quantities are deviations from that state. Here (u, υ, w) are the three components of velocity, p is pressure, ρ is density, and g is the acceleration due to gravity. This equation relates the Laplacian of the pressure perturbation divided by the density with six forcing terms. The first term is the buoyancy, associated with the vertical gradient of perturbation density. The second term is the linear contribution, associated with low pressure downshear from the updraft and high pressure upshear from it. The remaining four terms are nonlinear terms associated with perturbation fluid extension, deformation, vertical vorticity, and horizontal vorticity. The only terms that can account for a reduction in pressure are PBOU, PLIN, PVVORT, and PHVORT, since PEXT and PDEF are the square of quantities.

The pressure perturbation can be approximated by
p2p
and qualitative understanding of the contributions to the pressure perturbation may be gained by analyzing the different terms on the rhs of (2). Since the relation described by (4) is just an approximation, and the set of original equations used to derive (2) differs from the RAMS set of equations (it is incompressible, inviscid, and does not consider water loading), an exact match between the pressure perturbation from RAMS and the one from (4) is not expected. As with any other analysis that resorts to simplification of the original model, the goal is merely to reproduce the original system well enough so that conclusions from the simplified system can be extended to the more complex one.

The computation of the terms described above for the mesocyclone of the storm will be presented in section 6. The mean horizontal wind used in PLIN was a redefined reference state representative of the storm’s environment, obtained through the horizontal average of an area of 2688 km2 ahead of the MCS on grid 4. The subject of redefinition of the reference state will be revisited in section 5c.

b. Trajectory model

A Lagrangian model (LM) was used to compute backward parcel trajectories in the convective system. The LM was adapted from previous versions developed by Cotton et al. (1995) and Grasso (1996).

The winds used by the LM to advect the parcels were obtained from the Eulerian model (RAMS). The analysis described here focused on two periods referred to as daytime and nighttime. In each period, trajectories were followed for 30–45 min. When RAMS was run for each period, the values of the wind components in the Eulerian grid 4 were recorded every 15 s, which corresponds to twice the time step on grid 4. The time step to advect the parcels was also set to 15 s. The LM took into account eventual grid movements of grid 4 by computing the position of the parcel with respect to a stationary reference frame and not with respect to grid 4. The trajectories discussed in the following sections are representative of a large set of trajectories that compose the updraft or the downdraft.

c. Downdraft dynamics

In RAMS, a time-independent, horizontally homogeneous reference state in hydrostatic balance with zero vertical motion and no condensate is defined:
i1520-0493-126-11-2991-eq1
where the reference state variables have subscript 0, deviations from the reference state have primes and total quantities (sum of the reference state and perturbation) have no indices. Here θυ = θ(1 + 0.61rυ) is virtual potential temperature, θ = T(p/p)(R/cp) is potential temperature, π = cp(p/p)(R/cp) is Exner function, T is temperature, rυ is vapor mixing ratio, rc is condensate mixing ratio, p is a reference pressure, R is the gas constant for dry air, and cp is the specific heat for dry air at constant pressure.
Using the definitions above, employing the Boussinesq approximation, and using the Km (eddy diffusion coefficient) parameterization to represent the eddy diffusion, the vertical momentum equation becomes (Cotton and Anthes 1989)
i1520-0493-126-11-2991-e5
where d/dt = ∂/∂t + u(∂/∂x) + υ(∂/∂y) + w(∂/∂z) is the material derivative.

The first term on the rhs is the vertical π-gradient force (PGFz). The effects of vapor and temperature in lowering the parcel’s density have been included in the virtual potential temperature, which appears in the second term on the rhs, called moist buoyancy (BUO). The effects of condensate loading have been included in the third term, and eddy diffusion effects are expressed by the last term (DIFz).

In RAMS, the distinction between buoyancy (BUO) and pressure gradient force (PGFz) in the vertical momentum equation is not obvious because they depend on the choice of reference state, which is arbitrary. However, the total acceleration (dw/dt) and, in particular, the sum of PGFz and BUO do not depend on the reference state. Here (dw/dt) is the same for any choice of decomposition of θυ and π, as long as the reference state is in hydrostatic balance and deviations are small enough so their product can be neglected.

In RAMS the reference state is chosen at the time of model initialization (1200 UTC of 5/12 in this case) as the model column with the lowest topography in grid 1 (in this simulation a column over the Pacific Ocean). This reference state is therefore not representative of the storm environment, and therefore physically meaningless. To discuss the individual PGFz and BUO as provided by the equation above, a new buoyancy acceleration (BUOnew) has been defined to be a physically meaningful quantity. An average of 726 model columns in grid 4 ahead of the MCS was taken to provide a sounding representative of the environment of the convective system at each analysis time. This sounding was called the new reference state. As the parcel ascended or descended, virtual potential temperature perturbations were computed with respect to this new reference state, and (BUOnew) was obtained.

The new PGFz (PGFnewz) to balance the newly defined buoyancy was then simply diagnosed:
i1520-0493-126-11-2991-eq2

With this method BUOnew and PGFnewz are guaranteed to be self-consistent and physically meaningful quantities that can be compared with observations and used to gain understanding of the physical processes that determine the parcels’ trajectories.

In the next section the vertical and horizontal momentum equations described below will be discussed for selected trajectories:
i1520-0493-126-11-2991-e6
where (PGFx, PGFy), (CORx, CORy), and (DIFx, DIFy) are the π-gradient force, Coriolis accelerations, and diffusion in the x and y directions, respectively. Note that it is not necessary to compute PGFNEWx and PGFNEWy since the arbitrary reference state is horizontally homogeneous and makes no contribution to the horizontal components of the PGF.

6. Analysis results

In the interest of providing general conclusions, individual parcel trajectories will not be discussed. Instead, the results obtained with the Lagrangian model will be summarized in two conceptual models and just the history of one parcel representing the nighttime downdraft will be discussed.

a. Updrafts and downdrafts

The main airflow branches in the system for daytime and nighttime are shown in Fig. 21. During the day (Fig. 21a), the updraft originated within the PBL. The air flowed from the south, driven by the low-level trough. As it approached the cloud, it was displaced upward by a PGF, reached the level of free convection, and continued to ascend, constituting the updraft. The downdraft also originated within the boundary layer, but at lower levels, so it reached the cloud coming from the east, encountered the precipitation shaft, became colder due to evaporation, and quickly reached saturation, descending due to negative buoyancy and condensate loading.

During the night-phase of the system (Fig. 21b), the main flow branches were different. There were two sources of updraft air. The bulk of the updraft came from the south, flowing on top of the stable PBL, as indicated by previous studies (Trier and Parsons 1993;Laing 1997). However, part of the updraft air came from low-level easterly winds, located within the stable PBL.

The nighttime downdraft air originated at low levels ahead of the cloud, ascended on its leading edge, reached a height of about 700 m, then sank on the back of the cloud, turning cyclonically following the mesocyclone. This up–down downdraft component had also been identified by Knupp (1988) in ordinary cumulonimbus storms in the high plains.

The distribution of buoyancy along the flow branches discussed above is shown in Figs. 22a,b for daytime and nighttime, respectively. During the daytime, the updraft was initially neutrally buoyant, becoming positively buoyant when condensation started to occur. The downdraft was composed of air cooled by evaporation that encountered an environment progressively warmer as it descended, and could therefore maintain its negative buoyancy and reach the ground colder than its environment, forming a cold pool.

At night, the branch of the updraft that originated above the PBL ascended with positive buoyancy. However, the branch that started within the PBL ascended under negative buoyancy conditions until it exited the PBL and became positively buoyant. The up–down downdraft also ascended with negative buoyancy, and then descended with negative buoyancy and strong loading.

The balance of forces along an up–down downdraft trajectory helps elucidate the causes of the upward displacement in conditions of negative buoyancy. Figure 23a shows the time series of vertical velocity, height, and horizontal speed along such a trajectory. The parcel ascended from the beginning of the trajectory until 0545 UTC and descended subsequently. The temporal evolution of the terms of the vertical momentum Eq. (5) is shown in Fig. 23b. The initial upward displacement is caused by a PGF, and buoyancy got progressively negative as the ascent through the stable layer took place. Figure 23c shows the temporal evolution of the parcel’s and the environment’s potential temperature. As the parcel ascended, its potential temperature increased by 1.4 K because 0.5 g kg−1 of vapor condensed and latent heat was released. But the environment’s potential temperature increased faster due to the stability of the layer, therefore the parcel became 1.9 K colder than the environment on the apex of its path.

As the parcel crossed the precipitation field, below cloud base, it became loaded with rain droplets (Fig. 23b). Loading was a large contributor to the downward acceleration on the second part of the trajectory. The figure also shows that when the parcel got near the ground, diffusion became very large and positive, acting to slow down the downdraft.

To help elucidate the physical mechanisms associated with the PGF that pulls updraft and downdraft parcels off the stable boundary layer, the nature of the midlevel mesolow (Fig. 20) was examined using the pressure perturbation decomposition described in section 5a. Since the mesolow slopes westward with height, it creates an upward PGF at low levels.

Agreement between the pressure perturbation from the model (Fig. 24a) and the one retrieved from Eq. (3) (Fig. 24b) is best in the region of the mesolow, and loses quality to the southwest of the mesolow. The forcing terms of the pressure perturbation equation are shown in Figs. 24c–h. Perturbation horizontal vorticity is the most important term associated with the mesolow, pointing to the importance of the horizontal shear of the updraft in lowering the pressure. Shear associated with the vertical decrease of the easterly inflow ahead of the storm is also an important component of the horizontal vorticity. Perturbation vertical vorticity associated with the mesocyclone is the second most important term. The third term in importance is the linear term, which induces a low pressure perturbation downshear of the updraft. Buoyancy acts only to produce positive pressure perturbation at these levels. In the region of the mesolow, the contribution from buoyancy is not very large however. As seen in section 5, deformation and fluid extension cannot contribute toward negative pressure.

b. Strong horizontal winds

The strong surface winds during the night were from the northwest, and diverged out of a center of high pressure located at the surface (Fig. 19). The high pressure was collocated with the downdraft, and was dynamically caused by a deceleration of the downward motion. A center of hydrostatically induced low pressure was located at the surface east of the region of the strong winds, ahead of the band of surface updraft. This low pressure, however, did not act to accelerate the outflow winds responsible for the derecho. As can be seen in Fig. 19, the winds over the low pressure center were from the southeast, according to the environmental winds at the surface. When the parcels associated with the environmental southeasterlies reached the western part of the low, they were decelerated, but their direction was not reversed. This flow met the northwest downdraft outflow to the west of the pressure gradient associated with the hydrostatic low. Parcels that descended in the downdraft did not reach the gradient associated with the hydrostatic low and therefore were not influenced by it.

To illustrate this idea, one up–down trajectory may be taken as an example. The parcel started its descent at 0545 UTC, and between 0546 and 0558 its horizontal speed increased from 4 to 25 m s−1 (Fig. 23a).

Figures 25a,b show the terms of the horizontal momentum equation for this trajectory. The parcel accelerated due to a PGF, which had peak values of 0.055 m s−2 in the x direction and −0.055 m s−2 in the y direction. If those were the only forces present, the parcel would have accelerated considerably during the 30 min its trajectory lasted. The Coriolis acceleration gave a very small contribution to the horizontal balance of forces, as is expected on this small scale. But diffusion acted strongly to reduce the wind speed, and therefore peak net accelerations were 0.025 m s−2 in the x direction and −0.035 m s−2 in the y direction, enough to accelerate the parcel by 21 m s−1 in 12 min.

The horizontal PGF can be seen again in Figs. 26a,b, which show horizontal cross sections of the perturbation pressure field at 0549 and 0557 UTC, respectively. Superimposed on the Eulerian field is the projection of the trajectory onto the horizontal plane. The cross sections were taken at the height of the parcel at each time: 396 m in Fig. 26a and 48 m in Fig. 26b. This figure supports the idea that the parcel was accelerated by the pressure gradient associated with the high pressure on the northern part of the cell, and not by the mesolow ahead of the storm.

The origin of the horizontal PGF was investigated using the pressure perturbation decomposition described in section 5 [Eq. (3)]. Figure 27a shows a comparison between the pressure perturbation computed by the model and the Laplacian of the pressure perturbation computed using Eq. (2). Although the derivation of this form of the pressure perturbation equation assumes an inviscid, incompressible fluid, the decomposition represents well the model’s pressure perturbation gradients. An inspection of the different forcing terms (not shown) indicates that the largest contribution to the southern part of the surface high pressure (which is associated with the strong winds) comes from the fluid extension term, related to the deceleration of the downdraft (Fig. 27b).

7. Summary

A multiscale numerical simulation of the 12 and 13 May Kansas derecho case was performed to investigate the physical mechanisms that may contribute to the development of strong winds in derecho events.

One interesting aspect of this case is that the MCS formed in the afternoon in the presence of a well-mixed boundary layer, but the strong winds occurred only at night. There was a distinct difference between the daytime and nighttime environments that the system encountered, which had a marked influence on the system dynamics.

Even though the system developed on the cold side of a front, the daytime boundary layer on that side was deep and well mixed. In this situation, the downdraft processes can be understood using previously developed and well-understood theory, as summarized by Knupp and Cotton (1985). As air approaches the cloud base, it is cooled by evaporation of raindrops. The cold air is negatively buoyant, sinks, and its temperature perturbation becomes more negative as it descends. As it reaches the ground, a cold pool forms. On the leading edge of the cold pool there is convergence, boundary layer air is forced up, reaches the level of free convection, and becomes part of the updraft (Rotunno et al. 1988).

After sunset, the MCS was exposed to a different environment. The boundary layer on the cold side of the front became very stable and a strong elevated low-level jet developed and brought high equivalent potential temperature to the region where the MCS was active. The LLJ, which was not the focus of this paper, is thought to be due to boundary layer processes (Blackadar 1957; McNider and Pielke 1981) combined with the eastward advection of the midtropospheric trough and associated winds to the area. This jet was crucial to the maintenance of the MCS during the night. Due to the strong low-level stability, only a small amount of PBL air made it up to the level of free convection, and most of the updraft air originated within the warm, moist LLJ.

It was in this new environment that the strong surface winds occurred. The strongest winds were from the northwest, and originated in the region where the downdraft reached the ground, which was also a region of high surface pressure. The horizontal winds were accelerated by a horizontal pressure gradient, located at the edge of the center of high pressure. At this time, a cold pool was not present and the high pressure was not related to hydrostatic effects. As the air descended in the downdraft and became saturated, evaporation processes became irrelevant and warming due to compression became important. The progressively colder environment that the parcels encountered as they descended reduced their negative buoyancy, and they actually reached the ground warmer than their surroundings. The pressure perturbation decomposition analysis revealed that the high pressure was mainly derived from compression effects associated with the deceleration of the downdraft.

Understanding the strong horizontal winds is therefore tied to understanding the downdraft mechanism. To address that, a balance of forces was computed along a typical downdraft trajectory, which originated near the surface and followed an up–down downdraft path. Results showed that the initial upward displacement of the parcel was caused by an upward-directed pressure gradient force, associated with a midlevel pressure deficit, itself associated with the midlevel horizontal and vertical vorticity perturbation in the storm. This upward displacement within the stable boundary layer caused the parcel to develop strong negative buoyancy. Additional downward acceleration occurred as the parcel crossed the precipitation field and became loaded with hydrometeors.

This work suggests that an identifiable feature of derechos may be a bow-echo-shaped cell with an embedded mesocyclone that exists on top of a nocturnal boundary layer or some other stable layer. If confirmed in other case studies, identification of such convective organization may contribute to forecasting of derecho events. This forecast guidance is in apparent contradiction with the climatology by Johns and Hirt (1987), in which strong convective instability was identified as a precursor to derecho development. This discrepancy results from a difference in timescales. Johns and Hirt’s study used the last operational sounding available before derecho development to identify environmental characteristics that led to derecho development. In many cases, this was the 0000 UTC sounding that preceded a nocturnal event. As shown in this research, at 0000 UTC surface parcels were conditionally unstable, but the actual derecho environment had a stable PBL. Therefore, the forecast guidance proposed in this paragraph is valid for shorter-term forecasts, or nowcasting.

There is observational support in the literature for the relation between derechos and bow-echo high-precipitation supercells (Moller et al. 1994). However, several studies (e.g., Przybylinski and DeCaire 1985) have related derecho events with bow-echo squall lines. Although the term bow echo is sometimes used loosely, it is important to distinguish the difference between bow-echo supercells and bow-echo squall lines, the latter being composed of a series of independent cells, of the bow-echo type or not. Weisman (1993) performed and analyzed an idealized numerical simulation of a bow-echo squall line that generated strong surface winds. The model was initialized with a horizontally homogeneous state, characterized by a well-mixed PBL and by a large amount of CAPE. His results point to the low-level shear, the cold pool, and the buoyancy-generated vorticity at the leading edge of the cold pool as important factors in defining the morphology and the longevity of the squall line. The large CAPE and the presence of a well-mixed PBL significantly differentiate the environment described by Weisman from the one present in the case studied in this paper. This might be at the heart of the morphologic distinction between bow-echo squall lines and the system studied here. A more extensive observational and numerical database would be necessary to verify this assertion.

Finally, since some derechos develop in an environment with a stable layer, it is possible that gravity wave dynamics plays a role in their development, although that mechanism was not obvious in this case. In general, gravity waves generated by convection propagate away from their source. However, under certain shear profiles, they stay in phase with convection (Schmidt and Cotton 1990), providing a source of lifting for the stable boundary layer air. In the 12–13 May case, during the evening, the troposphere was very stable at low levels and moderately stable at midlevels. Although this environment does not constitute a good low-level duct for gravity waves, due to the absence of a neutral layer on top of the surface stable layer, there is evidence (McAnelly et al. 1997) of heating profiles that generate gravity waves with very small horizontal wavenumbers, which are long lived because their vertical propagation is negligible. Further studies that focus on the possibility of a gravity wave propagating within or on top of the stable boundary layer would be of great interest.

Acknowledgments

Limon radar data were provided by Ray McAnelly. Brenda Thompson helped with the editing and Judy Sorbie-Dunn drafted some of the figures. The authors appreciate the suggestions made by the reviewers, and acknowledge the helpful comments of Drs. Michael Montgomery and Richard Johnson of Colorado State University. Financial support for this research was provided by CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior), Brazil, and by the National Science Foundation under Grant ATM-9420045.

REFERENCES

  • Avissar, R., and R. A. Pielke, 1989: A parameterization of heterogeneous land surfaces for atmospheric numerical models and its impact on regional meteorology. Mon. Wea. Rev.,117, 2113–2136.

  • Barlow, W., 1996: An examination of the Kansas derecho on July 1, 1994. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 527–530.

  • Bernstein, B. C., and R. H. Johnson, 1994: A dual-Doppler radar study of an OK PRE-STORM heat burst event. Mon. Wea. Rev.,122, 259–273.

  • Blackadar, A., 1957: Boundary layer wind maxima and their significance for the growth of nocturnal inversion. Bull. Amer. Meteor. Soc.,38, 283–290.

  • Colman, B. R., 1990a: Thunderstorms above frontal surfaces in environments without positive CAPE. Part I: 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.

  • Cooper, S. R., and M. L. Bentley, 1996: A case study of the 8–9 July 1993 Great Plains derecho. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 531–535.

  • Cotton, W. R., and R. A. Anthes, 1989: Storm and Cloud Dynamics. Academic Press, 878 pp.

  • ——, G. D. Alexander, R. Hertenstein, R. L. Walko, R. L. McAnelly, and M. Nicholls, 1995: Cloud venting—A review and some new global annual estimates. Earth-Sci. Rev.,39, 169–206.

  • Davies, H. C., 1983: Limitations of some common lateral boundary schemes used in regional NWP models. Mon. Wea. Rev.,111, 1002–1012.

  • Gal-Chen, T., and R. C. J. Sommerville, 1975: On the use of a coordinate transformation for the solution of the Navier–Stokes equations. J. Comput. Phys.,17, 209–228.

  • Grasso, L. D., 1996: Numerical simulation of the May 15 and April 26, 1991 tornadic thunderstorms. Colorado State University Atmospheric Science Paper 596, 151 pp. [Available from Department of Atmospheric Science, Colorado State University, Fort Collins, CO 80523.].

  • Hill, G. E., 1974: Factors controlling the size and spacing of cumulus clouds as revealed by numerical experiments. J. Atmos. Sci.,31, 646–673.

  • Johns, R. H., and W. D. Hirt, 1987: Derecho: Widespread convectively induced windstorms. Wea. Forecasting,2, 32–49.

  • Klemp, J. B., and R. B. Wilhelmson, 1978: The simulation of three-dimensional convective storm dynamics. J. Atmos. Sci.,35, 1070–1096.

  • Knupp, K. R., 1988: Downdrafts within High Plains cumulonimbi. Part II: Dynamics and thermodynamics. J. Atmos. Sci.,45, 3965–3982.

  • ——, and W. R. Cotton, 1985: Precipitating convective cloud downdraft structure: An interpretive survey. Rev. Geophys.,23, 183–215.

  • Kreis, K. W., 1995: Synoptic and mesoscale severe weather: A two day case study of a derecho and local hailstorms. M.A. thesis, Dept. of Marine, Earth and Atmospheric Science, North Carolina State University at Raleigh, 93 pp. [Available from Dept. of Marine, Earth and Atmospheric Science, North Carolina State University at Raleigh, 1125 Jordan Hall, Box 8208, Raleigh, NC 27695.].

  • Laing, A. G., 1997: The large scale environment of mesoscale convective complexes: Comparison with other deep convective weather systems. Preprints, 22d Conf. on Hurricanes and Tropical Meteorology, Fort Collins, CO, Amer. Meteor. Soc., 415–416.

  • ——, and J. M. Fritsch, 1997: The global population of mesoscale convective complexes. Quart. J. Roy. Meteor. Soc.,123, 389–405.

  • Lilly, D. K., 1962: On the numerical simulation of buoyant convection. Tellus,14, 148–172.

  • Louis, J. F., M. Tiedke, and J.-F. Geleyn, 1982: A short history of the PBL parameterization at the ECMWF. Proc. Workshop on Planetary Boundary Layer Parameterization, Shinfield Park, Reading, United Kingdom, European Centre for Medium-Range Weather Forecasts, 59–79.

  • Loveland, T. R., J. W. Merchant, D. O. Ohlen, and J. F. Brown, 1991:Development of a land-cover characteristics database for the conterminous U.S. Photo. Eng. Remote Sens.,57, 1453–1463.

  • Maddox, R. A., 1979: Synoptic and meso-α scale aspects of flash flood events. Bull. Amer. Meteor. Soc.,60, 115–123.

  • ——, 1980: Mesoscale convective complexes. Bull. Amer. Meteor. Soc.,61, 1374–1387.

  • ——, 1983: Large-scale meteorological conditions associated with midlatitude, mesoscale convective complexes. Mon. Wea. Rev.,111, 1475–1493.

  • Mahrer, Y., and R. A. Pielke, 1977: A numerical study of the airflow over irregular terrain. Beitr. Phys. Atmos.,50, 98–113.

  • McAnelly, R. L., J. E. Nachamkin, W. R. Cotton, and M. E. Nicholls, 1997: Upscale evolution of MCSs: Doppler radar analysis and analytical investigation. Mon. Wea. Rev.,125, 1083–1110.

  • McNider, R. T., and R. A. Pielke, 1981: Diurnal boundary-layer development over sloping terrain. J. Atmos. Sci.,38, 2198–2212.

  • McNulty, R. P., 1995: Severe and convective weather: A central region forecasting challenge. Wea. Forecasting,10, 187–202.

  • Messinger, F., and A. Arakawa, 1976: Numerical Methods in Atmospheric Models. GARP Publication Series, Vol. I, World Meteorological Organization, 64 pp.

  • Moller, A. R., C. A. Doswell, M. P. Foster, and G. R. Woodall, 1994:The operational recognition of supercell thunderstorm environments and storm structures. Wea. Forecasting,9, 327–347.

  • National Climate Data Center, 1985: Storm Data. Vol. 25, No. 5, 66 pp.

  • Pielke, R. A., and Coauthors, 1992: A comprehensive meteorological modeling system—RAMS. Meteor. Atmos. Phys.,49, 69–91.

  • Przybylinski, R. W., and D. M. DeCaire, 1985: Radar signatures associated with the derecho, a type of mesoscale convective system. Preprints, 14th Conf. on Severe Local Storms, Indianapolis, IN, Amer. Meteor. Soc., 228–231.

  • Rochette, S. M., J. T. Moore, P. S. Market, F. H. Glass, and D. L. Ferry, 1995: Heavy rain events associated with elevated thunderstorms in the midwest. Postprints, Fourth National Winter Weather Workshop, Kansas City, MO, NOAA, 25-1–25-12.

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

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

  • Schmidt, J. M., 1991: Numerical and observational investigation of long-lived, MCS-induced, severe surface wind events: The derecho. Colorado State University Atmospheric Science Paper 479, 196 pp. [Available from Department of Atmospheric Science, Colorado State University, Fort Collins, CO 80523.].

  • ——, and W. R. Cotton, 1989: A High Plains squall line associated with severe surface winds. J. Atmos. Sci.,46, 281–302.

  • ——, and ——, 1990: Interactions between upper and lower tropospheric gravity waves on squall line structure and maintenance. J. Atmos. Sci.,47, 1205–1222.

  • Smagorinsky, J., 1963: General circulation experiments with the primitive equations. Part I: The basic experiment. Mon. Wea. Rev.,91, 99–164.

  • Staudenmaier, M. J., Jr., 1993: A multiscale study on the development and evolution of the 7–8 July 1991 derecho. M.A. thesis, Dept. of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, 160 pp. [Available from Dept. of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, 1225 W. Dayton St., Madison, WI 53706.].

  • Trembeck, C. J., and R. Kessler, 1985: A surface temperature and moisture parameterization for use in mesoscale numerical models. Preprints, Seventh Conf. on Numerical Weather Prediction, Montreal, PQ, Canada, Amer. Meteor. Soc., 355–358.

  • Trier, S. B., and D. B. Parsons, 1993: Evolution of environmental conditions preceding the development of a nocturnal mesoscale convective complex. Mon. Wea. Rev.,121, 1078–1098.

  • Tripoli, G. J., and W. R. Cotton, 1989: Numerical study of an observed orogenic mesoscale convective system. Part 1: Simulated genesis and comparison with observations. Mon. Wea. Rev.,117, 273–304.

  • Walko, R. L., W. R. Cotton, M. P. Meyers, and J. Y. Harrington, 1995a: New RAMS cloud microphysics parameterization. Part I: The single-moment scheme. Atmos. Res.,38, 29–62.

  • ——, C. J. Tremback, and R. F. A. Hertenstein, 1995b: RAMS—The Regional Atmospheric Modeling System: Version 3b user’s guide. Aster, 117 pp. [Available from Aster Inc., P.O. Box 466, Fort Collins, CO 80522.].

  • Weisman, M. L., 1993: The genesis of severe, long-lived bow echoes. J. Atmos. Sci.,50, 645–670.

Fig. 1.
Fig. 1.

Plan position indicator (PPI) from Limon, CO, radar at 0330 UTC on 13 May. Contours begin at 15.0 dBZ and are at 7.5-dBZ intervals. Shading begins at 15.0-dBZ and increases at 15.0-dBZ intervals.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 2.
Fig. 2.

Radar summary at 0535 UTC on 13 May.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 3.
Fig. 3.

High wind reports between 0534 and 1130 UTC from Storm Data (National Climate Data Center 1985).

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 4.
Fig. 4.

Surface analysis at 0000 UTC on 13 May.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 5.
Fig. 5.

Location of the grids at 2000 UTC: (a) grids 1, 2, and 3; (b) grids 3 and 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 6.
Fig. 6.

Equivalent potential temperature (shaded) (K), condensate mixing ratio (contoured) (g kg−1), and horizontal wind vectors at σ = 48 m and 0000 UTC on grid 2. The dashed vertical line along long = 97.5°W indicates the position of the cross section depicted in Fig. 7.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 7.
Fig. 7.

Vertical cross section through long = 97.5°W of potential temperature (K) at 0000 UTC on grid 2.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 8.
Fig. 8.

Temperature (shaded) (°C), condensate mixing ratio (contoured) (g kg−1), and horizontal wind vectors at σ = 48 m and 0000 UTC on grid 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 9.
Fig. 9.

Vertical cross section through lat = 37.5°N of vertical velocity (shaded) (m s−1) and equivalent potential temperature (contoured) (K) at 0000 UTC on grid 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 10.
Fig. 10.

Vertical vorticity (shaded) (×1000 s−1) and pressure perturbation (contoured) (hPa) at σ = 3381 m and 0000 UTC on grid 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 11.
Fig. 11.

Vertical cross section through lat = 37.6°N of condensate mixing ratio (shaded) (g kg−1), hail mixing ratio (solid contours) (g kg−1), and rain mixing ratio (dashed contours) (g kg−1) at 0000 UTC on grid 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 12.
Fig. 12.

Condensate mixing ratio (g kg−1) and horizontal wind vectors at σ = 48 m on grid 4 at (a) 0200 UTC, (b) 0400 UTC, and (c) 0600 UTC.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 13.
Fig. 13.

Time series of winds at σ = 48 m on grid 4. The closed circles represent the maximum speed in the horizontal domain, and the open circles the maximum speed found in convective outflow. After 0300 UTC they coincide.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 14.
Fig. 14.

Vertical cross section through long = 97.5°W of potential temperature (K) at 0600 UTC on grid 2.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 15.
Fig. 15.

Vertical profile at lat = 38°N, long = 98°W at times 0000 and 0600 UTC from grid 2 of (a) meridional wind (m s−1) and (b) equivalent potential temperature (K).

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 16.
Fig. 16.

Vertical profile of observations interpolated to lat = 38°N, long = 98°W at times 0000 and 1200 UTC of (a) meridional wind (m s−1) and (b) equivalent potential temperature (K).

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 17.
Fig. 17.

Temperature (°C) at σ = 48 m at 0600 UTC on grid 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 18.
Fig. 18.

Vertical cross section through lat = 38.1°N of vertical velocity (shaded) (m s−1) and equivalent potential temperature (contoured) (K) at 0600 UTC on grid 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 19.
Fig. 19.

Perturbation pressure (shaded) (hPa), vertical velocity (contoured) (m s−1), and horizontal wind vectors at σ = 48 and 0600 UTC on grid 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 20.
Fig. 20.

Vertical vorticity (×1000 s−1) (shaded), perturbation pressure (hPa) (contoured), and horizontal wind vectors at σ = 3381 m and 0600 UTC on grid 4.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 21.
Fig. 21.

Conceptual model of the (a) daytime and (b) nighttime phases of the MCS. The three planes are located at the surface, 700-m, and 2500-m height. The shaded field in each plane is the condensate mixing ratio and the thin arrows are ground-relative winds. The thick arrows represent the main branches of the updraft and of the downdraft. The plot on the left of each sketch depicts the vertical profile of equivalent potential temperature. In (b) the locations of the upward π-gradient force and condensate loading that act on the up–down downdraft are noted.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 22.
Fig. 22.

Diagram illustrating properties of idealized updraft and downdraft parcels at (a) daytime and (b) nighttime. The trajectory location with respect to the cloud is shown on the left. The top right sketches are skew T diagrams, illustrating thermodynamic properties of the environment and of the parcels, whose trajectories are represented by dots. The bottom-right sketches are vertical profiles of total buoyancy (including vapor effects and condensate loading) along the parcels’ paths.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 23.
Fig. 23.

Time series (0530–0600 UTC) along the path of a nighttime up–down downdraft trajectory of (a) height, vertical velocity, and horizontal speed; (b) terms of the vertical momentum equation (5); and (c) parcel and ambient potential temperature.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 24.
Fig. 24.

(a) Pressure perturbation (hPa) from the model; (b) sum of forcing terms of the pressure perturbation equation; (c) vertical vorticity forcing; (d) horizontal vorticity forcing; (e) deformation forcing; (f) fluid extension forcing; (g) linear forcing; (h) buoyancy forcing at 0600 UTC and σ = 3381 m. All forcing fields are in 10−4 s−2.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 25.
Fig. 25.

Time series (0530–0600 UTC) for a nighttime up–down downdraft parcel. (a) Terms of the zonal momentum equation; (b) terms of the meridional momentum equation.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 26.
Fig. 26.

Pressure perturbation (hPa) at (a) 0549 UTC and σ = 396 m and (b) 0557 UTC and σ = 48 m. The thick line is the projection of the trajectory of a nighttime trajectory from 0530 to 0600 UTC. The circle marks the earliest position of the parcel, and the square marks the position of the parcel at the time the pressure field is depicted.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

Fig. 27.
Fig. 27.

(a) Pressure perturbation (hPa) obtained from the model (shaded) (hPa) and sum of forcing terms of the pressure perturbation equation (contoured) (10−4 s−2); (b) fluid extension term of the pressure perturbation equation (10−4 s−2) at 0600 UTC and σ = 48 m.

Citation: Monthly Weather Review 126, 11; 10.1175/1520-0493(1998)126<2991:MEOADP>2.0.CO;2

1

To convert from UTC to CST subtract 6 h. Therefore, 0600 UTC is 0000 CST.

2

Pressure perturbation is the total pressure in a given grid point minus the reference state pressure in that grid point. By subtracting the reference state, the influence of topography on the pressure is eliminated. The actual millibar value of the pressure perturbation is meaningless, since the reference state is arbitrary, but the pressure perturbation gradient does represent meteorological features.

Save