• Adams, M. E., S. Businger, and M. L. Kaplan, 1995: Terrain-induced midtropospheric frontogenesis and jet streak formation during STORM-FEST IOP-17. Preprints, Seventh Conf. on Mountain Meteorology, Breckenridge, CO, Amer. Meteor. Soc., 96–100.

    • Search Google Scholar
    • Export Citation
  • Bernard, A., D. Demaiffe, N. Mattielli, and R. S. Punongbayan, 1991:: Anhydrite-bearing pumices from Mount Pinatubo: Further evidence for the existence of sulphur-rich silicic magmas. Nature, 354 , 139140.

    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., 1993: Observations and Theory of Weather Systems. Vol. II, Synoptic-Dynamic Meteorology in Midlatitudes, Oxford University Press, 594 pp.

    • Search Google Scholar
    • Export Citation
  • Caccia, J-L., M. Crochet, and K. Saada, 1997: ST radar evaluation of the standard deviation of the air vertical velocity perturbed by the local orography. J. Atmos. Solar-Terr. Phys, 59 , 11271131.

    • Search Google Scholar
    • Export Citation
  • Carbone, R. E., 1982: A severe frontal rainband. Part I: Stormwide hydrodynamic structure. J. Atmos. Sci, 39 , 258279.

  • Clark, T. L., W. D. Hall, and R. M. Banta, 1994: Two- and three-dimensional simulations of the 9 January 1989 severe Boulder windstorm: Comparison with observations. J. Atmos. Sci, 51 , 23172343.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., and C. F. Mass, 1995: The structure and evolution of cold surges east of the Rocky Mountains. Mon. Wea. Rev, 123 , 25772610.

    • Search Google Scholar
    • Export Citation
  • Crook, N. A., 1988: Trapping of low-level internal gravity waves. J. Atmos. Sci, 45 , 15331541.

  • Darby, L. S., W. D. Neff, and R. M. Banta, 1999: Multiscale analysis of a meso-β frontal passage in the complex terrain of the Colorado Front Range. Mon. Wea. Rev, 127 , 20622081.

    • Search Google Scholar
    • Export Citation
  • Davis, C. A., 1997: Mesoscale anticyclonic circulations in the lee of the central Rocky Mountains. Mon. Wea. Rev, 125 , 28382855.

  • Durran, D. R., 1986a: Mountain waves. Mesoscale Meteorology and Forecasting, P. S. Ray, Ed., Amer. Meteor. Soc., 472–492.

  • Durran, D. R., 1986b: Another look at downslope windstorms. Part I: The development of analogs to supercritical flow in an infinitely deep, continuously stratified fluid. J. Atmos. Sci, 43 , 25272543.

    • Search Google Scholar
    • Export Citation
  • Durran, D. R., 1990: Mountain waves and downslope winds. Atmospheric Processes over Complex Terrain, Meteor. Monogr., No. 45, Amer. Meteor. Soc., 59–81.

    • Search Google Scholar
    • Export Citation
  • Fulton, R., D. S. Zrnić, and R. J. Doviak, 1990: Initiation of a solitary wave family in the demise of a nocturnal thunderstorm density current. J. Atmos. Sci, 47 , 319337.

    • Search Google Scholar
    • Export Citation
  • Fritts, D. C., and G. D. Nastrom, 1992: Sources of mesoscale variability of gravity waves. Part II: Frontal, convective, and jet stream excitation. J. Atmos. Sci, 49 , 111127.

    • Search Google Scholar
    • Export Citation
  • Hartjenstein, G., and R. Bleck, 1991: Factors affecting cold-air outbreaks east of the Rocky Mountains. Mon. Wea. Rev, 119 , 22802292.

    • Search Google Scholar
    • Export Citation
  • Hogg, D. C., and and Coauthors, 1983: An automated profiler of temperature, wind, and humidity in the troposphere. J. Climate Appl. Meteor, 22 , 807831.

    • Search Google Scholar
    • Export Citation
  • Jin, Y., S. E. Koch, Y-L. Lin, F. M. Ralph, and C. Chen, 1996: Numerical simulations of an observed gravity current and gravity waves in an environment characterized by complex stratification and shear. J. Atmos. Sci, 53 , 35703588.

    • Search Google Scholar
    • Export Citation
  • Keller, T. L., 1994: Implications of the hydrostatic assumption on atmospheric gravity waves. J. Atmos. Sci, 51 , 19151929.

  • Koch, S. E., 1984: The role of an apparent mesoscale frontogenetical circulation in squall line initiation. Mon. Wea. Rev, 112 , 20902111.

    • Search Google Scholar
    • Export Citation
  • Kropfli, R. A., and and Coauthors, 1995: Cloud physics studies with 8 mm wavelength radar. Atmos. Res, 35 , 299313.

  • Kuettner, J. P., P. A. Hildebrand, and T. L. Clark, 1987: Convection waves: Observations of gravity wave systems over convectively active boundary layers. Quart. J. Roy. Meteor. Soc, 113 , 445467.

    • Search Google Scholar
    • Export Citation
  • Lee, T. J., R. A. Pielke, R. C. Kessler, and J. Weaver, 1989: Influence of cold pools downstream of mountain barriers on downslope winds and flushing. Mon. Wea. Rev, 117 , 20412058.

    • Search Google Scholar
    • Export Citation
  • Lilly, D. K., and E. J. Zipser, 1972: The Front Range windstorm of 11 January 1972—A meteorological narrative. Weatherwise, 25 , 5663.

    • Search Google Scholar
    • Export Citation
  • Martner, B. E., and J. D. Marwitz, 1982: Wind characteristics in southern Wyoming. J. Appl. Meteor, 21 , 18151827.

  • Marwitz, J. D., and P. J. Dawson, 1984: Low-level airflow in southern Wyoming during wintertime. Mon. Wea. Rev, 112 , 12461262.

  • May, P. T., K. P. Moran, and R. G. Strauch, 1989: The accuracy of RASS temperature profiles. J. Appl. Meteor, 28 , 13291335.

  • Mecikalski, J. R., and J. S. Tilley, 1992: Cold surges along the Front Range of the Rocky Mountains: Development of a classification scheme. Meteor. Atmos. Phys, 48 , 249271.

    • Search Google Scholar
    • Export Citation
  • Nance, L. B., 1997: On the inclusion of compressibility effects in the Scorer parameter. J. Atmos. Sci, 54 , 362367.

  • Nance, L. B., and D. R. Durran, 1997: A modeling study of nonstationary trapped mountain lee waves. Part I: Mean-flow variability. J. Atmos. Sci, 54 , 22752291.

    • Search Google Scholar
    • Export Citation
  • Nastrom, G. D., M. R. Peterson, J. L. Green, K. S. Gage, and T. E. VanZandt, 1990: Sources of gravity wave activity seen in the vertical velocities observed by the Flatland VHF radar. J. Appl. Meteor, 29 , 783792.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., R. M. Hardesty, M. A. Shapiro, and R. E. Cupp, 1988:: Doppler lidar observations of a downslope windstorm. Mon. Wea. Rev, 116 , 22652275.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., P. T. May, B. B. Stankov, and M. A. Shapiro, 1991: Radio acoustic sounding system observations of an arctic front. J. Appl. Meteor, 30 , 881892.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., P. T. May, and M. A. Shapiro, 1992: Radio acoustic sounding system (RASS) and wind profiler observations of lower- and middle-tropospheric weather systems. Mon. Wea. Rev, 120 , 22982313.

    • Search Google Scholar
    • Export Citation
  • Politovich, M. K., and B. C. Bernstein, 1995: Production and depletion of supercooled liquid water in a Colorado winter storm. J. Appl. Meteor, 34 , 26312648.

    • Search Google Scholar
    • Export Citation
  • Post, M. J., and R. E. Cupp, 1990: Optimizing a pulsed Doppler lidar. Appl. Opt, 29 , 41454158.

  • Post, M. J., A. Weickmann, K. R. Healy, R. J. Willis, and C. Grund, 1996:: Comparison of Mount Pinatubo and El Chichon volcanic events: Lidar observations at 10.6 and 0.69 μm. J. Geophys. Res, 101 , 39293940.

    • Search Google Scholar
    • Export Citation
  • Pratte, J. F., and L. Kaimal, 1986: Surface mesonetwork operator's manual. 2d version. 121 pp. [Available from NOAA/ERL/Forecast Systems Laboratory, Boulder, CO 80303.].

    • Search Google Scholar
    • Export Citation
  • Queney, P., 1948: The problem of airflow over mountains: A summary of theoretical studies. Bull. Amer. Meteor. Soc, 29 , 1626.

  • Ralph, F. M., M. Crochet, and S. V. Venkateswaran, 1992: A study of mountain lee waves using clear-air radar. Quart. J. Roy. Meteor. Soc, 118 , 597627.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., C. Mazaudier, M. Crochet, and S. V. Venkateswaran, 1993: Doppler sodar and radar wind-profiler observations of gravity-wave activity associated with a gravity current. Mon. Wea. Rev, 121 , 444463.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., P. J. Neiman, D. W. van de Kamp, and D. C. Law, 1995: Using spectral moment data from NOAA's 404-MHz radar wind profilers to observe precipitation. Bull. Amer. Meteor. Soc, 76 , 17171739.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., P. J. Neiman, and T. L. Keller, 1999: Deep-tropospheric gravity waves created by leeside cold fronts. J. Atmos. Sci, 56 , 29863009.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, R. M., and and Coauthors, 1992: The Winter Icing and Storms Project (WISP). Bull. Amer. Meteor. Soc, 73 , 951974.

  • Rasmussen, R. M., B. C. Bernstein, M. Murakami, G. Stossmeister, and J. Reisner, 1995: The 1990 Valentine's Day arctic outbreak. Part I: Mesoscale and microscale structure and evolution of a Colorado Front Range shallow upslope cloud. J. Appl. Meteor, 34 , 14811511.

    • Search Google Scholar
    • Export Citation
  • Schaefer, J. T., 1986: The dryline. Mesoscale Meteorology and Forecasting, P. Ray, Ed., Amer. Meteor. Soc., 549–570.

  • Schubert, W. H., S. K. Cox, P. E. Ciesielski, and C. M. Johnson-Pasqua, 1987: Operation of a ceilometer during the FIRE marine stratocumulus experiment. Atmospheric Science Paper 420, Colorado State University.

    • Search Google Scholar
    • Export Citation
  • Scorer, R. S., 1949: Theory of lee waves of mountains. Quart. J. Roy. Meteor. Soc, 75 , 4156.

  • Shapiro, M. A., 1984: Meteorological tower measurements of a surface cold front. Mon. Wea. Rev, 112 , 16341639.

  • Shapiro, M. A., and D. Keyser, 1990: Fronts, jet streams, and the tropopause. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. Holopainen, Eds., Amer. Meteor. Soc., 167–191.

    • Search Google Scholar
    • Export Citation
  • Shapiro, M. A., T. Hampel, D. Rotzoll, and F. Mosher, 1985: The frontal hydraulic head: A micro-α-scale (∼1 km) triggering mechanism for mesoconvective weather systems. Mon. Wea. Rev, 113 , 11661183.

    • Search Google Scholar
    • Export Citation
  • Shutts, G. J., and A. Broad, 1993: A case study of lee waves over the Lake District in northern England. Quart. J. Roy. Meteor. Soc, 119 , 377408.

    • Search Google Scholar
    • Export Citation
  • Simpson, J. E., and R. E. Britter, 1980: A laboratory model of an atmospheric mesofront. Quart. J. Roy. Meteor. Soc, 106 , 485500.

  • Smith, R. B., 1979: The influence of mountains on the atmosphere. Advances in Geophysics, Vol. 21, Academic Press, 87–230.

  • Steenburgh, W. J., and C. F. Mass, 1994: The structure and evolution of a simulated Rocky Mountain lee trough. Mon. Wea. Rev, 122 , 27402761.

    • Search Google Scholar
    • Export Citation
  • Tripoli, S., A. Buzzi, and A. Speranza, 1990: Orographic cyclogenesis. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. Holopainen, Eds., Amer. Meteor. Soc., 107–127.

    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., 1982: The life cycle of thunderstorm gust fronts as viewed with Doppler radar and rawinsonde data. Mon. Wea. Rev, 110 , 10601082.

    • Search Google Scholar
    • Export Citation
  • Weber, B. L., D. B. Wuertz, D. C. Law, A. S. Frisch, and J. M. Brown, 1992: Effects of small-scale vertical motion on radar measurements of wind and temperature profiles. J. Atmos. Oceanic Technol, 9 , 193209.

    • Search Google Scholar
    • Export Citation
  • Wesley, D. A., R. M. Rasmussen, and B. C. Bernstein, 1995: Snowfall associated with a terrain-generated convergence zone during the Winter Icing and Storms Project. Mon. Wea. Rev, 123 , 29572977.

    • Search Google Scholar
    • Export Citation
  • Westwater, E. R., and F. O. Guiraud, 1980: Ground-based microwave radiometric retrieval of precipitable water vapor in the presence of clouds with high liquid content. Radio Sci, 15 , 947957.

    • Search Google Scholar
    • Export Citation
  • Wolfe, D., and and Coauthors, 1995: An overview of the mobile profiler system: Preliminary results from field tests during the Los Angeles free-radical study. Bull. Amer. Meteor. Soc, 76 , 523534.

    • Search Google Scholar
    • Export Citation
  • Wuertz, D. B., B. L. Weber, R. G. Strauch, A. S. Frisch, C. G. Little, D. A. Merritt, K. P. Moran, and D. C. Welsh, 1988: Effects of precipitation on UHF wind profiler measurements. J. Atmos. Oceanic Technol, 5 , 452464.

    • Search Google Scholar
    • Export Citation
  • Wurtele, M. G., R. D. Sharman, and T. L. Keller, 1987: Analysis and simulations of a troposphere–stratosphere gravity wave model. Part I. J. Atmos. Sci, 44 , 32693281.

    • Search Google Scholar
    • Export Citation
  • Young, G. S., and R. H. Johnson, 1984: Meso- and microscale features of a Colorado cold front. J. Climate Appl. Meteor, 23 , 13151325.

    • Search Google Scholar
    • Export Citation
  • View in gallery
    Fig. 1.

    Base map of northeastern Colorado showing terrain contours (m), the 22 surface mesonet locations (•); the heavily instrumented remote sensing site at Platteville, CO (▴); the Doppler lidar site (▪), the operational SAO surface sites (★), and the operational rawinsonde site at DEN (+). The Continental Divide is portrayed as a bold dashed line. The location of downtown Denver (○) is also shown. The domain of this base map is shown in Figs. 2b, 3b, and 5–7

  • View in gallery
    Fig. 2.

    The 700-mb geopotential height (dam, thin solid) and potential temperature (K, bold dashed) analysis at (a) 1200 UTC 9 Mar 1993 and (b) 0000 UTC 10 Mar. The 700-mb wind vector flags are 25 m s−1, full barbs are 5 m s−1, and half-barbs are 2.5 m s−1. Wind vectors with solid-dot heads in (b) are wind-profiler observations. The 700-mb geopotential height (potential temperature) rounded to the nearest decameter (K) is plotted to the right (left) of the corresponding wind vector. The bold dotted line AA′ in (b) is the projection line for the cross section in Fig. 4. The profiler observation at Platteville, CO, in (b) is labeled PLT, and the two profiler measurements in Nebraska are at Merriman (north) and McCook (south). The outer dashed box in (b) corresponds to the regional mesoscale domain shown in Figs. 5–7, while the inner dashed box portrays the fine mesoscale domain in Figs. 1 and 8. The thin dotted line is the Continental Divide

  • View in gallery
    Fig. 3.

    The sea level pressure (mb, thin solid) and 3-h pressure change [shading: light, ≤−2 mb (3 h)−1; dark, ≥2 mb (3 h)−1] analysis at (a) 1200 UTC 9 Mar 1993, and (b) 0000 UTC 10 Mar. The surface wind barbs, the projection line AA′, the dashed boxes, and the Continental Divide are the same as in Fig. 2. The arctic front in each panel is represented by the bold line with closely spaced frontal symbols; all other fronts have widely spaced conventional symbols. The dryline is shown as a bold dashed line

  • View in gallery
    Fig. 4.

    Cross section of potential temperature (K, solid) and alongfront wind speed directed toward 125° (m s−1, bold dashed), along line AA′ of Figs. 2b and 3b at 0000 UTC 10 Mar 1993. The light-shaded region is the low-level cross-front flow (directed toward 195°) greater than 8 m s−1. Thin-dashed lines are frontal boundaries and the tropopause. The letter J marks the jet stream core aloft. Rawinsonde, wind-profiler, and surface sites are marked with long-bold, long-thin, and short vertical tick marks, respectively. Selected rawinsonde and wind profiler wind vectors are shown. Wind flags and barbs are the same as in Fig. 2

  • View in gallery
    Fig. 5.

    Regional mesoscale analysis of sea level pressure (mb, thin solid) at (a) 1800 UTC 9 Mar 1993, (b) 2100 UTC 9 Mar, (c) 0000 UTC 10 Mar, and (d) 0300 UTC 10 Mar. Frontal boundaries and surface wind barbs are the same as in Fig. 3. Terrain is shaded: light, 1500–2500 m; dark,>2500 m. The domain of this analysis is shown in Figs. 2b and 3b. The dashed box corresponds to the fine mesoscale domain in Figs. 1 and 8

  • View in gallery
    Fig. 6.

    The same as in Fig. 5 except for surface potential temperature (θ; K)

  • View in gallery
    Fig. 7.

    The same as in Fig. 5 except for 3-h surface pressure change (mb) ending at the designated times

  • View in gallery
    Fig. 8.

    Frontal isochrone analysis at the surface between (a) 1800 and 2130 UTC 9 Mar, (b) 2130 UTC 9 Mar and 0030 UTC 10 Mar, and (c) 0030 and 0500 UTC 10 Mar. The bold isochrones with conventional frontal symbols (as in Fig. 3) at 2000 UTC 9 Mar, 2230 UTC 9 Mar and 0330 UTC 10 Mar correspond to the times of the plotted data. Surface wind barbs (as in Fig. 2) and potential temperatures (K, rounded to the nearest degree) at these times are plotted. The lidar site and remote sensing site are marked with the symbols ▪ and ▴, respectively. Terrain is shaded: light, 2000–3000 m; dark,>3000 m. The domain of this analysis is shown in Figs. 2b, 3b, and 5–7

  • View in gallery
    Fig. 9.

    Mesonet time series traces (5-min resolution) of wind direction (DIR; °, bold), wind speed gust (SPD; m s−1), 1-h wind vectors (as in Fig. 2), temperature (T; °C), dewpoint temperature (Td; °C), and surface pressure (PRS; mb) between 1600 UTC 9 Mar and 1200 UTC 10 Mar 1993 at (a) BRI and (b) LTN. The vertical dashed lines labeled 1, 2, and 3 mark the initial arctic frontal advance, the frontal retreat, and the secondary frontal surge, respectively. See Fig. 1 for site locations

  • View in gallery
    Fig. 10.

    Doppler lidar PPI scan of radial velocity (m s−1) at 2135 UTC 9 Mar 1993. The elevation angle of the scan was 1° above horizontal. The bold white arrow denotes the direction of the downslope flow (i.e., from the northwest). Frontal boundaries are the same as in Fig. 3. Because the elevation of the PPI scan is a function of range, the lower-tropospheric frontal position east of the lidar is not valid at a single level. Rather, it is valid within a narrow layer between ∼550 and 675 m above ground. For reference, the lidar location and the surface frontal position at 2130 UTC (based on mesonet observations) are shown on the isochrone analyses in Figs. 8a and 8b

  • View in gallery
    Fig. 11.

    Isotach analysis of the horizontal wind (m s−1) in the plane of the cross section, derived from the Doppler lidar west–east RHI scan of radial wind velocity at 2345 UTC 9 Mar 1993. The downslope jet and mountain profile are shown in light and dark gray, respectively. The lidar is located at x = 0 km. When elevation scan angles are high (i.e., near x = 0) and vertical air motions are large, the horizontal wind derivations become less reliable. Upward motion was observed directly above the lidar during this scan (not shown); hence, the horizontal wind immediately upstream of the lidar is underestimated and the horizontal wind immediately downstream of the lidar is overestimated for steep scanning angles

  • View in gallery
    Fig. 12.

    (a), (b) Time–height sections and (c) time series of data from PLT between 1500 UTC 9 Mar and 1000 UTC 10 Mar 1993. (a) Wind profiles and RASS virtual potential temperature analysis (θυ; K, solid) based on the 15-min-averaged profiles measured by the 924-MHz wind profiler. The dashed isentropes are based on hourly RASS θυ profiles from the 404-MHz profiler. The 6-min resolution surface winds from PLT are included. Every other wind profile and surface wind is shown. The complete time series of wind profiles were subjected to one pass of a 1–2–1 temporal Hann filter. Wind flags and barbs are the same as in Fig. 2. Bold dots show cloud-base observations measured by the ceilometer. (b) Wind profiles and zonal wind speed analysis (m s−1, solid) based on the 15-min-averaged profiles measured by the 924-MHz profiler. The 0 m s−1 isotach is bold. Wind profiles, surface winds, and wind flags and barbs are as in (a). (c) Time series traces of data from the surface meteorological site [T = temperature (°C); Td = dewpoint temperature (°C); PRS = surface pressure (mb)] and from the dual-channel radiometer [VPR = integrated precipitable water vapor (cm); LQD = integrated liquid water (mm)]. The vertical dashed lines are the same as in Fig. 9. Bold lines on the top and bottom of each figure frame show the temporal domain of the PLT profiler analyses in Figs. 13 and 14

  • View in gallery
    Fig. 13.

    Time–height section of 3-min resolution data from the 924-MHz wind profiler at PLT between 1945 UTC 9 Mar and 0000 UTC 10 Mar 1993: (a) wind profiles, zonal wind speed analysis (m s−1, solid), and vertical-velocity shading (light, ≤−0.5 m s−1; dark, ≥0.5 m s−1), and (b) wind profiles and vertical-velocity analysis (m s−1, solid). The 6-min resolution surface winds from PLT are included. The profiles of horizontal and vertical wind were subjected to one pass of a 1–2–1 temporal Hann filter. The 0 m s−1 isotach in (a) is bold. Wind flags and barbs are the same as in Fig. 2. The bold dot at 2012 UTC and 2.4 km MSL in (b) is a rope cloud observation from the PLT ceilometer. The vertical dashed lines are as in Fig. 9

  • View in gallery
    Fig. 14.

    The same as in Fig. 13 except between 0045 and 0300 UTC 10 Mar 1993

  • View in gallery
    Fig. 15.

    Sequence of four visible satellite images at (a) 1830, (b) 1930, (c) 2030, and (d) 2130 UTC on 9 Mar 1993. The arctic frontal positions are based on the 3-h regional mesoscale analyses in Figs. 5–7 and on the isochrone analyses in Fig. 8. The frontal symbols are the same as in Fig. 3. The bold white dot in each frame marks the heavily instrumented remote sensing site at PLT

  • View in gallery
    Fig. 16.

    Photograph of lenticular wave clouds (ACSLs) and a shallow frontal rope cloud (rope) at ∼2100 UTC 9 Mar 1993 looking east-southeast (∼120°) from the Doppler lidar

  • View in gallery
    Fig. 17.

    Ka-band radar west–east RHI scan (oriented 255°–75°) of reflectivity [dBZ; ranging between −30 dBZ (light) and −10 dBZ (dark)] at 2003 UTC 9 Mar 1993 at PLT. The frontal position, which is based on measurements from the various remote sensing and in situ instruments at PLT, is also shown. The radar is located at x = 0 km

  • View in gallery
    Fig. 18.

    (a) Vertically stacked time series of vertical velocity (vertical bars, scale at upper right) measured by the 924- and 404-MHz wind profilers at PLT between 1824 and 2100 UTC 9 Mar 1993. Every other 3-min profile measured by the 924-MHz profiler below 3.5 km MSL is included in the time series, while every 6-min profile from the 404-MHz system is used above that. Light shading under the curves represents upward motion. The profiles of vertical wind were subjected to one pass of a 1–2–1 temporal Hann filter. Clouds measured directly above the collocated Ka-band cloud-sensing radar with sequential (∼2 min resolution) RHI scans are shown in dark shading starting at 1924 UTC. Only those clouds that exhibited a significant spatiotemporal history are shown. Surface winds (barbs are as in Fig. 2) from PLT are included. The vertical dashed line marks the initial arctic frontal passage, and the horizontal dashed line denotes the mean height of the tropopause. The position of the arctic front (bold solid) is based on the isotach analysis in Fig. 13a. The five bold dots near the bottom of the frame mark the times of individual vertical-velocity profiles shown in Fig. 19, and the two short thick tick marks correspond to the times of the thermodynamic soundings and Scorer parameter analyses shown in Figs. 20–22. (b) Time series trace of integrated liquid water (mm) at PLT

  • View in gallery
    Fig. 19.

    Profiles of Hann-filtered vertical velocity measured by the 924- and 404-MHz wind profilers at PLT at 1918, 1954, 2000, 2012, and 2018 UTC 9 Mar 1993. The time-averaged, mountain wave–induced vertical-velocity profile is shown in bold gray shade. The height range of the front-induced ACSLs observed by the Ka-band radar, from their initial sighting at 1957 UTC to the time of the last vertical-velocity profile at 2018 UTC, is shown with the thin dashed lines. The depth of the well-mixed boundary layer, based on Fig. 12, is also shown

  • View in gallery
    Fig. 20.

    Skew T–logp hybrid soundings at 1930 (dashed) and 2030 (solid) UTC 9 Mar 1993. The thermodynamic profiles below 4.5 km MSL were derived from Platteville's RASS θυ analysis in Fig. 12a at 1930 and 2030 UTC. Above the upper range of the RASS observations (i.e., above ∼4.5 km MSL), the profiles were completed by temporally interpolating the 1200 UTC 9 Mar and 0000 UTC 10 Mar Denver rawinsonde soundings to 1930 and 2030 UTC. The wind profiles below 5 km MSL were generated from Platteville's 924-MHz 15-min-averaged profiles at 1930 and 2030 UTC. The upper portions of these profiles were completed using the hourly averaged wind profiles from the Platteville 404-MHz profiler at 1930 and 2030 UTC. The thermodynamic and wind profiles were smoothed vertically using five passes of a 1–2–1 Hann filter after the data were transformed onto a uniform 250-m grid. Winds are the same as in Fig. 2. Every other wind measurement on the 250-m grid is shown

  • View in gallery
    Fig. 21.

    Vertical profiles of (a) front-relative, cross-front wind speed directed toward 70° (uf; m s−1), and (b) the Scorer parameter (l2; km−2) from Eq. (2) based on the uf profiles in (a) and the θυ profiles in Fig. 20. The bold and thin solid lines correspond to 1930 (prefrontal) and 2030 UTC (postfrontal) 9 Mar 1993, respectively

  • View in gallery
    Fig. 22.

    Vertical profiles of (a) cross-mountain wind speed directed toward 90° (um; m s−1), and (b) the Scorer parameter (l2; km−2) from Eq. (2) based on the um profiles in (a) and the θυ profiles in Fig. 20. The bold and thin solid lines correspond to 1930 (prefrontal) and 2030 UTC (postfrontal) 9 Mar 1993, respectively

  • View in gallery
    Fig. 23.

    (a) Vertical-velocity profile (m s−1, bold solid) of the resonant wave mode analytically derived from Eq. (3) based on the prefrontal, front-relative l2 profile at 1930 UTC 9 Mar 1993 (Fig. 21b). The vertical-velocity profile observed by the profilers at 2012 UTC (the time of frontal passage) is also shown (thin solid). Because the front moved rapidly across PLT, only one observational profile was available for comparison with the analytic curve. (b) Analytic vertical-velocity profile (bold) of the resonant wave mode based on the prefrontal, mountain-relative l2 profile at 1930 UTC 9 Mar (Fig. 22b). The time-averaged, mountain wave–induced vertical-velocity profile measured by the 924- and 404-MHz wind profilers is shown in bold gray shade (as in Fig. 19). The phase of the analytic curves (i.e., ascent in the troposphere) was chosen to correspond to the observed profiles. The magnitude of these curves was scaled by matching the value of the tropospheric upward-motion peak in these curves with the maximum tropospheric upward motion observed above ∼4 km MSL by the profilers. The horizontal wavelength (λx) of the front-relative and mountain-relative analytic solutions is shown

  • View in gallery
    Fig. 24.

    Conceptual representation of nonclassical frontal propagation in northeastern Colorado. (a) The initial arctic frontal advance (1800 and 2100 UTC 9 Mar 1993) and its nonclassical eastward retreat (0000 UTC 10 Mar), and the secondary frontal surge (0300 UTC). (b) The primary airstreams contributing to the nonclassical frontal propagation. The bold, southeastward-pointing streamlines portray the warm mountain-top northwesterlies that interacted with the Continental Divide to produce amplified mountain wave activity and strong downslope flow in eastern Colorado, and the bold streamlines in eastern Colorado highlight the anticyclonic gyre and northeasterly flow that was generated as the cold, postfrontal northwesterly flow encountered the Cheyenne Ridge. The interaction of these opposing airstreams played a dominant role in causing the oscillatory behavior of the arctic front. Deflected-gap flow on the warm side of the front (thin streamlines) may have also contributed to this behavior. Dotted line XX′ is a projection line for the cross section in Fig. 25. Terrain is shaded: light, 1500–2500 m; dark,>2500 m

  • View in gallery
    Fig. 25.

    Schematic summary of the structure and generation of leeside, mountain-forced and frontally forced gravity waves during the initial frontal advance along line XX′ in Fig. 24b. The shallow arctic air mass and mountain wave–induced downslope windstorm are shown as medium and light gray-shaded areas, respectively. A thermally direct vertical circulation (gray-shaded arrows) straddled the leading edge of the density-current-like front, with a rope cloud (black cloud) situated in the shallow updraft. Sensible heating (wiggly arrows) in the arctic air and weak cold advection of isentropes (thin dashed lines) in the warm air are also shown, resulting in a temporal decrease in the thermodynamic contrast across the front (see equation). Streamlines (thin solid) portray the high-amplitude, vertically propagating and vertically trapped mountain waves forced by flow over the high terrain, and they also portray the vertically propagating gravity waves excited by flow over the obstacle-like front. The frontal gravity waves created lenticular wave clouds aloft (gray-shaded lenses). The remote sensing site at PLT and the lidar site are shown for reference

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 76 26 0
PDF Downloads 27 12 0

Observations of Nonclassical Frontal Propagation and Frontally Forced Gravity Waves Adjacent to Steep Topography

Paul J. NeimanNOAA/Environmental Technology Laboratory, Boulder, Colorado

Search for other papers by Paul J. Neiman in
Current site
Google Scholar
PubMed
Close
,
F. Martin RalphNOAA/Environmental Technology Laboratory, Boulder, Colorado

Search for other papers by F. Martin Ralph in
Current site
Google Scholar
PubMed
Close
,
Robert L. WeberNOAA/Environmental Technology Laboratory, Boulder, Colorado

Search for other papers by Robert L. Weber in
Current site
Google Scholar
PubMed
Close
,
Taneil UttalNOAA/Environmental Technology Laboratory, Boulder, Colorado

Search for other papers by Taneil Uttal in
Current site
Google Scholar
PubMed
Close
,
Louisa B. NanceCooperative Institute for Research in Environmental Sciences and NOAA/ETL, Boulder, Colorado

Search for other papers by Louisa B. Nance in
Current site
Google Scholar
PubMed
Close
, and
David H. LevinsonUnited States Department of the Interior—BLM, Missoula, Montana

Search for other papers by David H. Levinson in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Through the integrated analysis of remote sensing and in situ data taken along the Front Range of Colorado, this study describes the interactions that occurred between a leeside arctic front and topographically modulated flows. These interactions resulted in nonclassical frontal behavior and structure across northeastern Colorado. The shallow arctic front initially advanced southwestward toward the Front Range foothills, before retreating eastward. Then, a secondary surge of arctic air migrated westward into the foothills. During its initial southwestward advance, the front exhibited obstacle-like, density-current characteristics. Its initial advance was interrupted by strong downslope northwesterly flow associated with a high-amplitude mountain wave downstream of the Continental Divide, and by a temporal decrease in the density contrast across the front due to diurnal heating in the cold air and weak cold advection in the warm air. The direction and depth of flow within the arctic air also influenced the frontal propagation.

The shallow, obstacle-like front actively generated both vertically propagating and vertically trapped gravity waves as it advanced into the downslope northwesterly flow, resulting in midtropospheric lenticular wave clouds aloft that tracked with the front. Because the front entered a region where strong downslope winds and mountain waves extended downstream over the high plains, the wave field in northeastern Colorado included both frontally forced and true mountain-forced gravity waves. A sequence of Scorer parameter profiles calculated from hourly observations reveals a sharp contrast between the prefrontal and postfrontal wave environments. Consequently, analytic resonant wave mode calculations based on the Scorer parameter profiles reveal that the waves supported in the postfrontal regime differed markedly from those supported in the prefrontal environment. This result is consistent with wind profiler observations that showed the amplitude of vertical motions decreasing substantially through 16 km above mean sea level (MSL) after the shallow frontal passage.

Corresponding author address: Paul J. Neiman, NOAA/Environmental Technology Laboratory, Mail Code R/E/ET7, 325 Broadway, Boulder, CO 80303. Email: Paul.J.Neiman@noaa.gov

Abstract

Through the integrated analysis of remote sensing and in situ data taken along the Front Range of Colorado, this study describes the interactions that occurred between a leeside arctic front and topographically modulated flows. These interactions resulted in nonclassical frontal behavior and structure across northeastern Colorado. The shallow arctic front initially advanced southwestward toward the Front Range foothills, before retreating eastward. Then, a secondary surge of arctic air migrated westward into the foothills. During its initial southwestward advance, the front exhibited obstacle-like, density-current characteristics. Its initial advance was interrupted by strong downslope northwesterly flow associated with a high-amplitude mountain wave downstream of the Continental Divide, and by a temporal decrease in the density contrast across the front due to diurnal heating in the cold air and weak cold advection in the warm air. The direction and depth of flow within the arctic air also influenced the frontal propagation.

The shallow, obstacle-like front actively generated both vertically propagating and vertically trapped gravity waves as it advanced into the downslope northwesterly flow, resulting in midtropospheric lenticular wave clouds aloft that tracked with the front. Because the front entered a region where strong downslope winds and mountain waves extended downstream over the high plains, the wave field in northeastern Colorado included both frontally forced and true mountain-forced gravity waves. A sequence of Scorer parameter profiles calculated from hourly observations reveals a sharp contrast between the prefrontal and postfrontal wave environments. Consequently, analytic resonant wave mode calculations based on the Scorer parameter profiles reveal that the waves supported in the postfrontal regime differed markedly from those supported in the prefrontal environment. This result is consistent with wind profiler observations that showed the amplitude of vertical motions decreasing substantially through 16 km above mean sea level (MSL) after the shallow frontal passage.

Corresponding author address: Paul J. Neiman, NOAA/Environmental Technology Laboratory, Mail Code R/E/ET7, 325 Broadway, Boulder, CO 80303. Email: Paul.J.Neiman@noaa.gov

1. Introduction

The behavior of shallow cold fronts along the eastern slope of the Rocky Mountains, and along other north–south-oriented mountain ranges, presents special forecast challenges partly due to the sharp contrast between warm downslope flow and leeside cold surges, as well as the distortion of fronts by terrain effects. Although studies of individual weather phenomenon, such as downslope windstorms or leeside fronts, promote a greater understanding and can help improve forecasting techniques, the conditions on any given day can be influenced strongly by several competing processes. The goal of this paper is to illustrate the complex evolutions that can occur when several phenomena interact during the passage of a leeside cold front across Colorado's Front Range corridor, including modulation of frontal propagation by terrain-induced flows and the generation of gravity waves by the front itself.

This study is made possible through use of a uniquely well-suited set of observations gathered on 9 and 10 March 1993 during the multiyear Winter Icing and Storms Project (WISP) that was carried out in eastern Colorado and southern Wyoming (Rasmussen et al. 1992). The key platforms used here were deployed by the National Oceanic and Atmospheric Administration's (NOAA) Environmental Technology Laboratory (ETL) and Forecast Systems Laboratory (FSL), and included two wind profilers, a Doppler lidar, a cloud radar, and a surface mesonetwork, which are described later in detail. The event of 9–10 March was characterized by strong mountain wave–induced downslope (i.e., westerly component) flow, a shallow arctic front that oscillated in a nonclassical west–east fashion across northeastern Colorado, a subsequent secondary cold surge, and the generation of frontally forced gravity waves as the downslope flow impinged upon the shallow front. The term nonclassical is used to describe the propagation characteristics of this front because it did not exhibit a steady forward motion typical of most fronts, though this behavior is not necessarily unusual in regions of steep terrain such as in eastern Colorado (e.g., Neiman et al. 1991; Darby et al. 1999).

To diagnose the key physical processes that contributed to this complex event it is useful to consider four types of cool-season, terrain-induced flows that are common in this region: 1) leeside frontal surges, 2) lee vortices, 3) gap flows, and 4) mountain-induced gravity waves (i.e., mountain waves). The blocking effect of North America's Rocky Mountains accelerates cold fronts southward along the eastern side of the high terrain (e.g., Hartjenstein and Bleck 1991; Mecikalski and Tilley 1992; Colle and Mass 1995). The leading edges of these orographically trapped leeside fronts often acquire density-current characteristics during their southward migration across eastern Colorado (e.g., Young and Johnson 1984; Shapiro 1984; Shapiro et al. 1985; Neiman et al. 1991). The postfrontal, northerly component airstream flowing across the west–east-oriented Cheyenne Ridge in southeastern Wyoming can induce an anticyclonic mesoscale vortex or gyre to the lee (south) of this ridge, subsequently driving the front westward toward the Front Range foothills (Wesley et al. 1995; Davis 1997). This postfrontal gyre generates regional cloud and precipitation distributions that differ from what would be expected solely from synoptic-scale forcing. The weather of Colorado's Front Range urban corridor can also be affected by gap flow through the “wind corridor” in south-central Wyoming (Martner and Marwitz 1982; Marwitz and Dawson 1984). The Continental Divide here is at its lowest point between Montana and New Mexico, and the mountain ranges straddling this west–east corridor constrict in a north–south sense. Therefore, low-level westerly flow is often channeled through this gap, turning anticyclonically around the Medicine Bow and Laramie mountains situated at the southern flank of the wind corridor (these physiographic features are shown in Fig. 2 of Marwitz and Dawson 1984), thus contributing to strong northwesterly flow in northern Colorado east of the high terrain. Finally, when westerly component flow traverses the Rockies, mountain waves can be generated and downslope winds can become strong (e.g., Queney 1948; Lilly and Zipser 1972; Smith 1979; Durran 1986a; Neiman et al. 1988; Durran 1990).

Perhaps the most challenging forecast problem related to terrain-induced flows in eastern Colorado is the interaction of leeside frontal surges with downslope windstorms. Depending on the orientation, propagation direction, and depth of the leeside front relative to the mountains, the front and mountain waves can interact in a variety of ways. For example, Lee et al. (1989) used two-dimensional, geostrophic, numerical simulations to show that downslope windstorms induced by large-amplitude mountain waves are significantly affected by the presence, depth, and intensity of cold air along the eastern slope of the Rockies. Their study showed that the ability of this cold air to be flushed eastward away from the mountains depends on the direction and depth of the geostrophic flow within the cold air, that is, a decrease in the upslope component in the cold air and a decrease in its depth increases the likelihood of flushing. In their simulations, the strength of the mountain wave–induced downslope flow is of secondary importance in flushing out the cold air. To date, operational numerical models are limited in their ability to resolve the structural and dynamical details of both mountain waves and shallow, terrain-trapped cold surges because of the relatively coarse spatial resolution of these models and the uncertainties associated with initializing them. These limitations can adversely impact local forecasts, as well as forecasts of larger-scale, terrain-induced phenomena [e.g., lee cyclones (Tripoli et al. 1990); lee troughs (Steenburgh and Mass 1994); lee frontogenesis (Adams et al. 1995); and drylines (Schaefer 1986)] that affect regions far downstream of the high terrain.

Studies by Kuettner (1987), Ralph et al. (1993), and Jin et al. (1996) have shown that flow impinging on atmospheric boundaries can create gravity waves analogous to those associated with mountains. Wind-profiling radars and commercial aircraft have observed enhanced vertical-velocity variance at high altitudes that are attributed to gravity waves launched by surface cold fronts (e.g., Nastrom et al. 1990; Fritts and Nastrom 1992). Recently, attention has turned to the forcing of gravity waves by orographically trapped, leeside cold fronts. Since these fronts can take on the characteristics of obstacle-like density currents, the interaction of these fronts with the impinging cross-mountain flow can lead to the generation of deep-tropospheric gravity waves, as has been shown recently by Ralph et al. (1999). Their observational and companion numerical analysis shows vertically trapped lee waves that extended upward from shallow leeside fronts into the upper troposphere, where they produced a distinctive signature in the water vapor channel satellite imagery.

In the following sections, we will show how the evolution of the leeside arctic cold front of 9–10 March 1993 was modulated by terrain-induced mesoscale flows, and how this front generated both vertically trapped and vertically propagating gravity waves through the depth of the troposphere.

2. Observing systems

Locations of the key observing systems within the inner experimental domain that were used in this study are shown in Fig. 1, and their observing characteristics are summarized in Table 1. The NOAA/FSL 22-station surface mesonet (Pratte and Kaimal 1986) measured standard meteorological parameters every 5 min. Additional ground-based, in situ observing networks that were utilized include the national-scale Surface Airway Observation (SAO; 1-hourly) network and upper-air rawinsonde network (12-hourly). Two surface stations that measured standard meteorological parameters at 2- and 6-min intervals complemented the heavily instrumented remote sensing site at Platteville, Colorado (PLT). NOAA's Geostationary Operational Environment Satellite-7 satellite provided visible, infrared, and water vapor imagery.

Remote sensors, deployed mostly by NOAA/ETL, played a central role in the WISP program, and most were clustered at PLT (Fig. 1). A 404-MHz wind profiler at PLT measured hourly averaged vertical profiles of the horizontal wind from 0.5 to 16.25 km above ground level (AGL) at 250-m vertical resolution with its two orthogonal beams directed 16.3° off vertical and one vertically pointing beam. The vertical beam also provided direct measurements of vertical air motions in the precipitation-free troposphere and lower stratosphere every 6 min. This profiler was part of the larger 28-station National Profiler Network across the central United States operated and maintained by NOAA (Ralph et al. 1995). NOAA/ETL deployed a 924-MHz wind profiler at PLT that operated with four orthogonally directed beams offset 21° from vertical and a fifth beam directed vertically (Wolfe et al. 1995). This five-beam system is more effective at measuring the full three-dimensional wind than a three-beam system due to the redundancy in measurements (Wuertz et al. 1988; Weber et al. 1992). On the other hand, this configuration increases the time needed to determine each wind profile, and it assumes spatial homogeneity on larger scales than the three-beam system. Vertical profiles of the horizontal and vertical wind were calculated from the 924-MHz profiler using 15-min and 3-min averaging periods. These profiles extended from 0.12 to ∼5 km AGL with ∼100-m range resolution. The wind profilers described above were combined with an acoustic source in a configuration known as a radio acoustic sounding system (RASS; May et al. 1989). Vertical profiles of virtual temperature (Tυ) were calculated from the relationship between the speed of sound and Tυ at each range gate, and then transformed into virtual potential temperature (θυ) using the method described by Neiman et al. (1992). The approximate upper ranges of the RASS θυ profiles measured with the 924- and 404-MHz profilers were 1.1 and 3.4 km AGL, respectively.

NOAA/ETL operated additional remote sensors at PLT: a scanning Ka-band radar, a dual-channel radiometer, and a laser ceilometer. The Ka-band radar (Kropfli et al. 1995) operates at a frequency of 8.66 mm and is very sensitive to cloud particles (i.e., sensitivity of ∼−30 dBZ at 10-km range). Its narrow beamwidth (0.5°) and small pulse length (0.25 μs) provide reflectivity and radial velocity depictions of clouds with very high spatial resolution. The dual-channel radiometer remotely measures column-integrated precipitable water vapor and liquid water every 2 min from the brightness temperatures obtained in the 20.6- and 31.6-GHz wavelength bands (Westwater and Guiraud 1980; Hogg et al. 1983). The laser ceilometer measures the height of cloud base with 50-m resolution up to ∼3.7 km AGL at 30-s intervals (Schubert et al. 1987). Because the laser beam cannot penetrate deeply into clouds, it does not measure cloud thickness and cannot detect overlying cloud layers.

NOAA/ETL's 10.6-μm scanning CO2 Doppler lidar (Post and Cupp 1990) was situated 42 km west-southwest of the heavily instrumented PLT site and ∼34 km east of the Continental Divide (Fig. 1). Its narrow beamwidth (1.8 m at 20-km range) and lack of sidelobes provided detailed observations (≤300 m spatial and ≤1 min temporal resolutions) of radial velocity and backscatter within its ∼30 km observing radius. Because of the large amount of volcanic aerosol created by the eruption of Mount Pinatubo in 1991 (Bernard et al. 1991), the lidar measurements extended well into the stratosphere (Post et al. 1996), thus providing a unique opportunity to observe the detailed structure and evolution of terrain-induced flows through a deep layer of the atmosphere. The lidar was operating between 1630 UTC 9 March and 0110 UTC 10 March 1993.

3. Background meteorology for 9–10 March 1993

The background meteorological conditions of 9–10 March are shown in Figs. 2–7 to establish the larger-scale context for the detailed analyses and interpretations of nonclassical frontal propagation, tropospheric frontal structure, and frontally forced gravity waves presented later in sections 4–6.

a. Synoptic-scale perspective

On the synoptic scale, a 700-mb cyclonic disturbance migrated southeastward from southwestern Canada to the Dakotas between 1200 UTC 9 March and 0000 UTC 10 March (Fig. 2). The center of this disturbance was initially north of the analysis domain at 1200 UTC (Fig. 2a). A trailing cold-frontal baroclinic zone extended outward from the cyclone center into the northwesterly current across Montana and Idaho at 1200 UTC and across Nebraska and Wyoming at 0000 UTC. This 700-mb baroclinic zone remained north and east of PLT where prefrontal downslope west-northwesterly flow predominated. The corresponding sea level pressure (SLP) analyses (Fig. 3) show the cyclonic disturbance intensifying at the surface over the northern United States during the 12-h period, and a pressure rise—fall couplet of greater than 2 mb (3 h)−1 flanking the cyclone at both times.

At 1200 UTC 9 March, the cold-frontal baroclinic zone over Montana at 700 mb (Fig. 2a) extended to the surface in southern Wyoming (Fig. 3a). The air behind this front, which originated over the Canadian arctic, was characterized by northwesterly flow. The arctic air had spilled westward over the Continental Divide in western Wyoming and Montana, but the western boundary of the arctic air became realigned along the divide in Wyoming during the following hours. The region of synoptic-scale pressure rise behind the front and cyclone showed no evidence of multiple centers, either spatially or from a time series perspective. Farther south and east, the back edge of a retreating air mass of continental polar origin extended southward from the Dakotas to Texas and then eastward to Louisiana.

By 0000 UTC 10 March (Fig. 3b), the arctic front at the surface had advanced southward to southeastern Colorado and northern Oklahoma. It was followed by a pressure rise center of >2 mb (3 h)−1 that had separated from the synoptic-scale pressure rise center located behind the cyclone. Comparison of the 0000 UTC SLP and 700-mb panels (Figs. 3b and 2b, respectively) reveals the arctic air over Kansas and eastern Colorado was shallow [<∼3 km above mean sea level (MSL)]. At this time, a distinctly separate frontal surge east of the Rockies preceded the arctic front and penetrated southward into northern Texas and southeastern Oklahoma. Farther east, the back edge of the continental polar air mass had retreated eastward to Iowa, Missouri, and Arkansas by 0000 UTC 10 March (Fig. 3b).

A north–south cross section of potential temperature (θ) and alongfront wind speed (u) at 0000 UTC 10 March (Fig. 4) highlights the vertical continuity of the arctic front and the southward protrusion of the front below ∼700 mb across eastern Colorado and western Kansas. Enhanced baroclinicity, static stability, and vertical wind shear associated with the front extended upward from the surface near Granada, Colorado (GDA), to the vertically depressed tropopause at ∼380 mb over Bismark, North Dakota (BIS). Northeasterly flow immediately behind the surface intersection of the shallow front backed to northwesterly within the cold air. Both are typical signatures of an orographically trapped cold surge east of the Rocky Mountains. A west-northwesterly jet stream core of ∼57 m s−1 was situated at ∼350 mb on the warm edge of the arctic front aloft.

b. Regional mesoscale evolution

Three-hourly surface analyses of SLP, θ, and 3-h pressure tendency highlight the 9-h regional mesoscale evolution of the arctic front and adjacent boundaries and air masses between 1800 UTC 9 March and 0300 UTC 10 March (Figs. 5–7). During this 9-h period, the arctic front migrated southward (east of the high terrain) from northeastern Colorado to the Texas Panhandle.1 A nose of high pressure, and a 3-h pressure rise center of 3–4 mb that separated from its parent synoptic-scale rise center, propagated southward behind the front (Figs. 5, 7). Postfrontal northwesterly flow in eastern Wyoming and western Nebraska encountered the west–east-oriented Cheyenne Ridge in southeastern Wyoming, resulting in an anticyclonic gyre and 5–15 m s−1 northerly to northeasterly flow across northeastern Colorado. From northeastern Colorado southward, the flow immediately behind the front was directed along the isallobaric wind vector. The arctic front's θ contrast of 8–12 K was confined to a narrow <50 km wide transition during the period 1800–0000 UTC (Figs. 6a–c). These signatures are all consistent with those associated with orographically trapped cold surges east of the Rockies (e.g., Hartjenstein and Bleck 1991; Mecikalski and Tilley 1992; Colle and Mass 1995). A secondary cold surge was evident in northeastern Colorado at 0300 UTC, but this surge was confined to Denver's Front Range urban corridor, as evinced by the small spatial scale of the pressure rise center [>3 mb (3 h)−1] and its proximity to the high terrain (Fig. 7d). Though comparable finescale structures and evolutions may have occurred in adjacent areas, they could not be resolved due to limited observations there.

Strong downslope northwesterly flow was observed east of Colorado's high terrain on the warm side of the arctic front between 1800 UTC 9 March and 0300 UTC 10 March (Figs. 5–7) with gusts >20–25 m s−1, primarily before 0000 UTC. The presence of a low pressure trough collocated with a thermal ridge along the eastern slope of the high terrain, and a strong cross-mountain pressure gradient to its west (Figs. 5 and 6), suggest that significant mountain wave activity aloft was present and contributed to strong downslope flow in the region. Doppler lidar observations of a leeside downslope jet and wind-profiler observations of deep-tropospheric vertical-velocity perturbations (both are shown and discussed later) support this interpretation. The prefrontal northwesterly flow in extreme northern Colorado (to the east of the high terrain) may have been enhanced by strong prefrontal westerly flow that was funneled through the gap in southern Wyoming2 and deflected anticyclonically around the southern periphery of this gap. Due to the limited number of observations within and downstream of the gap, we were not able to quantitatively assess the role of gap flow in enhancing the prefrontal northwesterly flow in northern Colorado. However, an earlier observational study based on detailed research flight-level data within the boundary layer in the vicinity of the gap (Marwitz and Dawson 1984) showed distinct anticyclonic curvature to the flow through the gap, and northwesterly flow extending downstream into northern Colorado.

As the arctic front migrated southeastward across the Great Plains, wind profilers in Nebraska (locations shown Fig. 2b) documented temporal and vertical changes in wind velocity (not shown) that were typical of many cold frontal passages (e.g., Shapiro and Keyser 1990; Neiman et al. 1992). However, orographic trapping of the arctic air by the Rocky Mountains, the development of an anticyclonic gyre within the postfrontal northwesterly flow around the Cheyenne Ridge, prefrontal gap flow through southern Wyoming, and the presence of downslope winds resulted in nonclassical frontal behavior across northeastern Colorado. Observations of the front by the suite of sensors in northeastern Colorado showed horizontal, vertical, and temporal variations in the standard meteorological parameters that were substantially different from those taken far from the steep terrain to the north and east of Colorado. The interaction of this front with Colorado's Front Range and its orographically forced downslope northwesterly flow is the focus of the following three sections.

4. Nonclassical frontal propagation in northeastern Colorado

The NOAA/FSL surface mesonet documented the arctic front's nonclassical west–east propagation characteristics, and a subsequent secondary cold surge, across northeastern Colorado east of the Continental Divide, as is highlighted in three frontal isochrone analyses (Fig. 8). These analyses show the initial southwestward frontal advance toward the Front Range foothills between 1800 and 2130 UTC 9 March (Fig. 8a), its eastward retreat away from the high terrain during the period 2130 UTC 9 March to 0030 UTC 10 March (Fig. 8b), and a secondary westward surge of arctic air into the foothills between 0030 and 0500 UTC 10 March (Fig. 8c).

The arctic front initially progressed southwestward toward the foothills, except in the west-central portion of the domain after 2030 UTC (Fig. 8a). The leading edge of the front moved across PLT at ∼2012 UTC 9 March with an average phase velocity of 7.0 ± 1 m s−1 from 70°, but it did not pass the lidar site located 42 km west-southwest of PLT. The front, which contained a θ contrast of 10 K in <30 km, separated north-northeasterly flow on its cold side from northwesterly downslope flow on its warm side. Similar warm-side values of θ between the foothills and the plains suggest that the prefrontal airstream extending eastward from the divide was warmed adiabatically by descending turbulent flow associated with a mountain wave. A northwestward decrease in θ was observed on the warm side of the arctic front (e.g., Fig. 8a). Surface mesonet traces in this region (e.g., Figs. 9a and 12c) show evidence of a weak cold-frontal passage (i.e., a pressure rise that coincided with an abrupt increase in northwesterly flow and a cessation of the diurnal increase in temperature) from northwest to southeast between 1515 and 1930 UTC. Because the foothills sites also observed these signatures, it is likely that weak cold advection migrated eastward across the divide and then descended the steep eastern slope of the Rockies and onto the plains. Net 700-mb cooling of 2 K within the downslope flow at Denver between 1200 UTC 9 March and 0000 UTC 10 March (Fig. 2), and a 4-K decrease in θ within the downslope-induced thermal ridge between 1800 and 0000 UTC (Fig. 6), support this assertion.

Between 2130 UTC 9 March and 0030 UTC 10 March, the arctic front retreated eastward over the northern two-thirds of the domain (Fig. 8b). During this period, an anticyclonic gyre within the arctic air mass developed south of the retreat zone over the Denver area and the west–east-oriented Palmer Ridge. Though the Cheyenne Ridge can produce an anticyclonic gyre of varying size and intensity along Colorado's northern Front Range that is dependent on the environmental static stability and wind-profile characteristics, its initial effects are observed much farther north than the Denver area. Therefore, this gyre may have resulted, at least in part, from the blocking action associated with the advancing (retreating) downslope flow (arctic front) to the north. The arctic front retreated eastward beyond PLT at ∼2310 UTC 9 March. Between 2000 and 2230 UTC (Figs. 8a and 8b), sensible heating of 2–6 K was observed in the arctic air in the northeastern portion of the mesonet, thus resulting in a decrease of the potential temperature or density contrast across the front.

The secondary surge of arctic air propagated westward across the mesonet between 0030 and 0500 UTC 10 March (Fig. 8c), passing PLT at ∼0130 UTC 10 March with a phase velocity of ∼6.8 m s−1 from 80°. Stronger easterly component flow was observed behind the secondary surge than behind the initial arctic front, thus possibly contributing to the westward advance of the secondary surge into the foothills. The secondary surge originated behind the initial arctic front, as evidenced by the fact that the sites that remained in the cold air following the initial frontal passage observed a distinct transition with the second surge (see also Fig. 9b). Note that the first isochrone of the secondary surge at 0030 UTC 10 March (Fig. 8c) marks a distinctly different boundary than the final isochrone of the initial retreating front at the same time (Fig. 8b). Values of θ were up to 10 K colder behind the second surge at 0330 UTC 10 March than behind the initial front. This behavior suggests a reinforcing pulse of arctic air entered the mesonet domain from the cold reservoir over the plains, similar to that observed in other studies (e.g., Politovich and Bernstein 1995; Rasmussen et al. 1995). Because this small-scale secondary surge along the northern Front Range originated immediately south (downstream) of the Cheyenne Ridge, it is also plausible that the ridge played a direct role in its initiation. Davis (1997) describes a similar Front Range feature with a small spatiotemporal scale that was initiated by weak diurnal heating along the Cheyenne Ridge.

Mesonet time series traces from Brighton, Colorado (BRI), and Littleton, Colorado (LTN) (Fig. 9), highlight temporal transitions from two regions that exhibited distinctly different frontal propagation. The traces from BRI were typical of those from other sites that experienced the initial frontal advance, its subsequent retreat, and then the secondary surge. The initial frontal passage at ∼2035 UTC 9 March was accompanied by large and rapid changes of all variables. The restoration of warmer and drier downslope flow occurred at ∼0005 UTC 10 March, though the temperature subsequently decreased in this downslope regime due to diurnal cooling. The moist secondary surge passed BRI at ∼0205 UTC 10 March. The traces from LTN were typical of those from other sites that observed the initial frontal passage and secondary surge, but not the intermediate frontal retreat. The initial frontal passage at 2110 UTC 9 March was similar in character to that observed at BRI. However, the dry downslope flow did not return over LTN, given that the winds did not revert to dry northwesterly. The secondary surge at 0405 UTC 10 March exhibited characteristics similar to those at BRI, except that the thermodynamic transition at LTN was less pronounced because the cold air behind the initial frontal passage had remained in place.

The shallow arctic air remained east of NOAA/ETL's Doppler lidar while it was recording data between 1630 and 0110 UTC 9–10 March. By the time the secondary surge migrated westward beyond the site at 0320 UTC 10 March, the lidar was no longer operating. Nevertheless, it was instrumental in documenting, and aided in refining, the westernmost position of the arctic front between ∼2050 and 2140 UTC 9 March (Fig. 8a) with range–height indicator (RHI) and plan-position indicator (PPI) scans. A low-angle (1°) PPI scan at 2135 UTC 9 March (Fig. 10) documented large horizontal gradients of radial velocity [>10 (m s−1) km−1] associated with the arctic front ∼23 km east of the lidar, and strong downslope northwesterly flow (15–20 m s−1) impinging upon the sharp frontal interface. The cold air did not contain significant easterly component flow, and it was also quite shallow (∼700 m deep; shown and discussed in section 5), so it is not surprising that the front's eastward retreat had commenced by this time (e.g., Lee et al. 1989). The frequent RHI and PPI scans during the initial frontal advance and subsequent retreat also provided detailed documentation of the evolving cross-mountain downslope flow. A representative west–east-oriented RHI scan at 2345 UTC 9 March (Fig. 11) documented a prominent downslope jet (∼20 m s−1) below 4.5 km MSL, with weaker flow in the middle troposphere. This jet extended at least 15 km east of the lidar within 2 km of the ground. The presence of a vertically confined downslope jet, with weak flow aloft, indicates the presence of a mountain wave and leeside subsidence (e.g., Durran 1986b; Neiman et al. 1988; Clark et al. 1994).

5. Detailed tropospheric observations of the arctic front and secondary surge

The wind profilers and RASS at PLT measured wind and θυ that documented the depth and timing of the initial arctic frontal passage, its subsequent retreat, and the secondary surge. The time–height analyses in Fig. 12 are based on 15-min-averaged profiles of data taken by the 924-MHz five-beam profiler/RASS, though the upper portion of the θυ analysis in Fig. 12a was also based on data from the collocated 404-MHz RASS. The analyses in Figs. 13 and 14 are based on 3-min-averaged profiles taken by the 924-MHz profiler. The 3-min RASS θυ profiles were too noisy to analyze.

Prior to the initial arctic frontal passage at 2012 UTC 9 March, the time–height analyses based on the 15-min profiles (Figs. 12a,b) showed downslope northwesterly flow (10–18 m s−1) in a weakly stratified regime. A low-level wind direction shift from westerly to stronger northwesterly at ∼1730 UTC coincided with the premature cessation of low-level diurnal heating, and signaled the onset of weak cold advection on the warm side of the arctic front. The arctic front exhibited a sharp 10-K drop in θυ and a rapid wind shift to northerly flow in the lowest ∼700 m. Surface traces of this transition (Fig. 12c) mirrored those at BRI (Fig. 9a). Integrated water vapor (Fig. 12c) increased with the frontal passage, largely in response to the moist conditions in the low-level cold air. Above the front, 2–3 K of cooling and a wind shift from northwesterly to westerly (Fig. 12a) coincided with frontally forced ascent shown in Fig. 13, thus suggesting that the frontal passage modified the character of the thermodynamic and flow fields immediately above the shallow cold dome. The ceilometer observed clouds above 3.7 km MSL while the shallow cold air resided over PLT. The integrated liquid water trace (Fig. 12c) reveals that these clouds contained water droplets. Downslope westerly flow of warmer air descended to the surface at ∼2310 UTC as the front retreated eastward beyond PLT. The secondary surge passed PLT at 0130 UTC 10 March and was characterized by a period of cooling, ascending with time, which was considerably more gradual than that associated with the initial frontal passage. Changes in the surface variables were similar to those at BRI. Upslope easterly flow behind the secondary surge forced a layer of low (<2.8 km MSL) stratus after 0630 UTC 10 March. The easterly component flow behind this surge was stronger (−10 m s−1), and extended deeper into the troposphere (2.5 km MSL), than behind the initial front (−5 m s−1 and 2.2 km MSL). Therefore, it is not surprising that the secondary surge advanced westward into the foothills, whereas the initial front retreated eastward away from the high terrain. These observations are consistent with the two-dimensional, geostrophic modeling results of Lee et al. (1989) that show a dependence between shallow cold air flushing eastward away from the Continental Divide and the direction and depth of flow within the cold air, that is, a decrease in the upslope component in the cold air and a decrease in its depth increases the likelihood of flushing.

To obtain a finer-scale temporal depiction of the three frontal transitions than the 15-min profiles could provide, 3-min profiles of the horizontal and vertical wind measured by the 924-MHz profiler were analyzed (Figs. 13,14). Figure 13a shows a scale-contracted gradient in zonal wind associated with the initial frontal passage, followed by an elevated density-current-like head (∼1000 m deep) and vertically depressed wake region, between 2012 and 2040 UTC. Northerly flow was confined within this shallow cold dome. Figure 13b contains a thermally direct vertical-velocity couplet with the front that is also consistent with a density current, including rapid ascent (>2.5 m s−1) forced by strong convergence3 (∼56 × 10−4 s−1) at the frontal interface (leading to a narrow rope cloud that was observed by the collocated ceilometer at 2012 UTC and 2.4 km MSL) and postfrontal subsidence (<−1.5 m s−1) maximized within the elevated head. The 2-min surface data contained a sharp, density-current-like drop in temperature of 10°C (Fig. 12c). These observations are consistent with previous studies (e.g., Young and Johnson 1984; Shapiro 1984; Shapiro et al. 1985; Neiman et al. 1991) that show the leading edge of orographically trapped leeside cold fronts often acquire density-current characteristics during their southward migration across eastern Colorado.

A well-established theoretical relationship for the speed of density currents entering an environment with no opposing prefrontal flow (e.g., Fulton et al. 1990; Bluestein 1993) is given by
Cρkgdθυwθυcθυc1/2
where g = 9.8 m s−2 is the gravitational acceleration; d is the mean height of the cold air; θυw and θυc are the vertically integrated virtual potential temperatures in the warm and cold air, respectively; and k approximates the internal Froude number. In an idealized framework for steady flow that is inviscid, unstratified, and infinitely deep, k = 2. However, mixing across the two-fluid interface and surface friction reduces this value in the real atmosphere to ∼0.8–1.1 (e.g., Carbone 1982; Wakimoto 1982; Koch 1984), or an average of 0.95. The meteorological variables used in Eq. (1) were ascer-tained from Figs. 12a and 13a (i.e., θυw = 303.2 K, θυc = 296.3 K, d = 708 m) and yield a phase speed of 12.1 m s−1. Laboratory experiments (e.g., Simpson and Britter 1980) demonstrate that this ground-relative phase speed should be offset by ∼60% of the speed of the opposing prefrontal flow. Subtracting six-tenths of the layer-mean prefront wind component directed toward the front in the lowest 708 m (i.e., 60% of 8.4 m s−1) from Cρ gives a phase speed of 7.1 m s−1. This predicted value is very close to what was observed (7.0 ± 1.0 m s−1) and provides additional support for the density-current interpretation.

Because the phase speed of the density-current-like leeside arctic front was linked to the opposing downslope flow, and the strength of the downslope flow is modulated by mountain wave activity, a direct connection can be made between the propagation characteristics of the arctic front and the intensity of the mountain wave activity and its downstream influence. The downslope flow opposing the frontal motion at PLT, far removed from the high terrain, was ∼8.4 m s−1, whereas the lidar observed opposing downslope flow at least twice that magnitude within the low-level downslope jet close to the foothills (Fig. 11). Therefore, it is not surprising that the front stalled as it approached the high terrain of northeastern Colorado (Fig. 8). Similarly, the temporal decrease in density contrast across the front (Figs. 8a,b) arising from sensible heating in the cold air and weak cold advection in the warm air also likely contributed to the cessation of its westward motion. The upcoming section on frontally forced gravity waves will focus on the initial period when the arctic front advanced southwestward as a density current.

The time–height analysis in Fig. 13 also highlighted the eastward retreat of the initial front beyond PLT between 2230 and 2310 UTC. During this period, low-level cold northerly flow (∼10 m s−1) at PLT was replaced by warm downslope westerly flow that descended with time from ∼2.3 km MSL to the surface. A thermally direct vertical circulation accompanied this transition, with subsidence and ascent of 1–2 m s−1 within the cold northerlies and warm westerlies, respectively.

The secondary surge is illustrated in the time–height sections of Fig. 14. Unlike the initial frontal passage, the secondary surge did not possess density-current attributes despite the presence of a prominent upward-motion plume exceeding 2 m s−1. Rather, the wind shifted gradually from northwesterly to easterly across the boundary, and an elevated head was not observed. By entering representative meteorological values from the surge environment (i.e., θυw = 300.8 K, θυc = 295.2 K, d = 904 m) into Eq. (1), and assuming a layer-mean opposing flow of 5.2 m s−1, a predicted density-current phase speed of 9.2 m s−1 is obtained. This value is much larger than the observed phase speed of 6.8 m s−1, and provides additional evidence that the secondary surge was not a density current.

6. Frontally forced gravity waves

As just shown, the arctic front exhibited density-current attributes during its initial migration across PLT. Density currents can act as obstacles to the ambient flow they displace, resulting in the generation of gravity waves (e.g., Ralph et al. 1993; Jin et al. 1996; Ralph et al. 1999). Ralph et al. (1999) described the generation of vertically trapped gravity waves by density-current-like shallow cold surges propagating southward east of the Rockies. These gravity waves, which produced distinctive signatures in the water vapor channel satellite imagery, arose largely because blocking by the Rockies changed the frontal orientation and disrupted thermal wind balance near the terrain-modified fronts. This imbalance manifested itself as strong prefrontal, front-relative, front-normal flow. The enhanced convergence of the prefrontal, front-normal flow with the obstacle-like fronts then excited the gravity waves. The structure, orientation, and propagation of the terrain-modified arctic front during its initial advance on 9 March were similar to those of the cold surges shown in Ralph et al. (1999). Therefore it is likely that blocking by the Rockies also disrupted thermal wind balance to create frontally forced gravity waves on 9 March. We will show evidence that the gravity waves generated by the initial arctic frontal passage on this day produced both vertically propagating and vertically trapped modes.

a. Observations

Sequential satellite imagery (Fig. 15) establishes that the western or upstream edge of clouds in eastern Colorado was roughly collocated with the shallow southward-propagating arctic front, thus suggesting a link between the clouds and the obstacle-like front propagating into ambient northwesterly flow. These clouds were present over PLT in the postfrontal regime, consistent with the ceilometer observations in Fig. 12a. Companion infrared satellite imagery (not shown) revealed that these cloud tops approached ∼8 km MSL. An east-southeastward-pointing photograph taken from the lidar site at ∼2100 UTC 9 March (Fig. 16), together with the sequential imagery (Fig. 15), demonstrate the clouds were a cluster of midtropospheric standing lenticular wave clouds (ACSLs). Because these clouds moved with the advancing front, the term standing is relative to the front. The satellite imagery (Fig. 15) indicates that the ACSLs extended downstream from the front for a distance of 50–125 km as individual cloud sheets. This relatively large horizontal scale, and the lenticular character of the clouds, are typical of vertically propagating gravity waves (e.g., Smith 1979; Durran 1986a). Hence, the sequential imagery and photograph strongly suggest that the front itself was exciting vertically propagating gravity waves in the troposphere in response to the forced ascent of prefrontal flow over the front. The photo also shows a shallow rope cloud that marked the strong updraft forced by convergence at the advancing edge of the frontal head (Fig. 13b). This rope cloud was situated roughly beneath the western edge of the ACSLs, though the limited depth perception that the photo offers introduces ambiguity in documenting the relative positions of these two cloud features. Real-time visual observations were less ambiguous.

NOAA's Ka-band Doppler radar was optimally positioned at PLT to document the evolution of these frontally forced wave clouds. At 2003 UTC 9 March 1993, these clouds were at ∼5.5 km MSL and their upwind or western edge was situated 2 km west of PLT and 5 km ahead of the front (Fig. 17). The ACSL exhibited a 560-m upward deflection above the shallow front. This upward deflection coincided with the position of the strong updraft (>2 m s−1) forced by convergence at the frontal head (Fig. 13b). The ACSL extended eastward beyond the maximum range of the radar.

The analysis shown in Fig. 18, which combines Ka-band cloud observations taken directly over the radar at ∼2 min intervals with wind-profiler-observed vertical velocities, surface winds, the arctic front's position, and a time series of integrated liquid water, emphasizes the wave induced vertical motions through a layer extending into the lower stratosphere and the locations of the clouds relative to these vertical motions and the shallow front. Prior to the frontal passage, large-amplitude vertical motions of 1–2 m s−1 extended upward from ∼3.5 km MSL into the stratosphere. The disturbed character of the stratosphere, and the large-amplitude vertical motions in the troposphere (independent of the front), are characteristic signatures of leeside mountain waves (e.g., Queney 1948; Smith 1979; Durran 1986a; Ralph et al. 1992; Caccia et al. 1997). Within this deep layer of prefrontal ascent, wave clouds were observed between 6.6 and 8.9 km MSL from the time the radar began operating at 1924 UTC until 1957 UTC; these wave clouds were quite likely associated with the mountain waves. The wave clouds in this layer do not appear in the satellite imagery because they were subvisible cirrus clouds composed only of small ice particles (Fig. 18b) (Kropfli et al. 1995). The layer below ∼4 km MSL in Fig. 18a marked the well-mixed prefrontal boundary layer (Fig. 12), and hence, the more rapid prefrontal vertical-velocity variability in this layer likely resulted primarily from turbulent-scale boundary layer processes.

Strong upward motion (∼1.5–2.8 m s−1) associated with the front is clearly depicted in Fig. 18a between 1954 and 2024 UTC from the lowest range gate upward to at least 6 km MSL.4 The ACSLs observed between 1957 and 2024 UTC from 5.2 and 5.9 km MSL coincided with this strong ascent. Comparison of these Ka-band and wind-profiler analyses with Queney's (1948) idealized cross sections of vertically propagating gravity waves provides additional evidence that the front was exciting this class of waves. In particular, the Ka-band analyses in Figs. 17 and 18a show the western terminus of the ACSLs situated upstream of the advancing obstacle-like front, and the wind-profiler analysis in Fig. 18a shows tropospheric ascent upward to at least 6 km MSL commencing 12–18 min prior to frontal passage (i.e., 5.0–7.5 km ahead of the front, given a frontal phase speed of 7 m s−1); these analyses are consistent with the idealized sections portraying tropospheric ascent residing upstream of the obstacle.

After 2030 UTC, the ACSLs lowered and thickened as the vertical motions transitioned to neutral or weakly downward (Fig. 18a). Given that clouds form in response to the Lagrangian time history of vertical displacements rather than to instantaneous vertical motions, and because the ACSLs in this case had advected downstream from their generation region above the front, this portion of the ACSL cloud sheet existed in the absence of strong ascent. The integrated liquid water trace (Fig. 18b) reveals that these ACSLs contained liquid-phase cloud particles. The amplitude of vertical motions decreased substantially through 16 km after 2024 UTC and remained smaller through the end of the period of interest on 10 March (not shown). Though changes in ambient conditions upstream of the arctic front could have accounted for the rapid decrease in gravity wave–induced vertical-velocity activity, it is more likely that this decrease was linked to the sharp frontal passage because the vertical-velocity transition immediately followed the frontal passage and both transitions were similarly abrupt. It should be noted that the environment that initially supported the mountain waves and frontal waves had changed abruptly because of the frontal passage (see the θ/u analyses in Fig. 12), thus suggesting further that the frontal passage played a direct role in modulating the vertical-velocity field above the radar. The relationship between the frontal passage and the subsequent decrease in mountain and frontal gravity wave activity over the shallow arctic air mass is quantified in the following subsection.

Superimposing five representative vertical-velocity profiles between 1918 and 2018 UTC from Fig. 18a onto a single graph (Fig. 19) helps distinguish the vertical-velocity structures associated with the mountain wave from those associated with the arctic front. The steady, sinusoid character of these profiles in the troposphere and stratosphere, including the region aloft near the time of the arctic frontal passage (i.e., starting at 2000 UTC), provides strong evidence this structure resulted primarily from mountain wave activity. The mean mountain wave profile was not depicted within the ∼2.5 km deep boundary layer because vertical motions in the boundary layer are typically influenced by turbulent scale processes. The profiles between 2000 and 2018 UTC show a departure from the mountain wave signature in the lower and middle troposphere that is tied to the ascent associated with the front. At 2000 UTC this upward motion (≤∼1 m s−1) extended upward from the lowest range gate to the mountain wave regime at 5.5 km MSL. Twelve minutes later, this ascent was stronger (≥∼1.5 m s−1) and entered the mountain wave sinusoid at ∼8 km MSL. The frontal ACSLs appeared at about the time the deepening upward motion associated with the front crossed the level of the cloud base. These clouds thickened as the upward motion continued to deepen, and then thinned as this region of ascent dropped below cloud base (Figs. 18a and 19). These observations strongly suggest the upward-motion anomaly (relative to the mountain wave sinusoid) extending into the middle troposphere was associated with vertically propagating gravity wave activity forced by flow over the front. Within this frontal-wave anomaly, the profiles at 2012 and 2018 UTC each contained a distinct spike of large upward motion (>2.5 m s−1) centered at ∼3 km MSL that is consistent with a shallow updraft forced by low-level convergence at the leading edge of a shallow, sharp front (e.g., Young and Johnson 1984; Shapiro 1984; Shapiro et al. 1985). Hence, the structure and behavior of the profiles in Fig. 19 are apparently the result of a superposition of the updraft forced by convergence at the front's leading edge and the wave-induced ascent associated with flow over the front and the Rocky Mountains.

b. Application of two-dimensional linear theory

1) Basic theoretical background

To explore the wave environment for flow over the shallow arctic front and the Continental Divide, vertical profiles of the Scorer parameter were calculated from hourly hybrid soundings at PLT and Denver, Colorado, between 1730 UTC 9 March and 0330 UTC 10 March 1993. Scorer parameter diagnostics will be presented for the soundings at 1930 and 2030 UTC 9 March (Fig. 20), which bracket the arctic frontal passage at 2012 UTC. Given that both the arctic front and the Continental Divide in this case are nearly two-dimensional, it is reasonable to consider the two-dimensional, Boussinesq version of the Scorer parameter (e.g., Scorer 1949; Smith 1979):
l2NU2U2Uz2
where z is the altitude, U is the cross-barrier flow, and N is the Brunt–Väisälä frequency. The l2 profiles for the frontal wave environment (Fig. 21b) were determined using vertical profiles of the component of the front-relative flow orthogonal to the advancing front (i.e., U = uf; Fig. 21a). The mean frontal motion across PLT was 7.0 m s−1 from 70°. Similarly, the l2 profiles for the mountain wave environment (Fig. 22b) were calculated using vertical profiles of the component of the flow perpendicular to the north–south-oriented Continental Divide (i.e., U = um; Fig. 22a).

Numerous investigators (e.g., Scorer 1949; Smith 1979; Crook 1988; Ralph et al. 1992; Shutts and Broad 1993; Nance and Durran 1997) have shown that the shape of the l2 profile is an important indicator in determining the most favored wave modes. Specifically, a rapid decrease of l2 with height favors wave trapping, whereas an l2 profile that does not decrease significantly with height allows wave energy to propagate vertically. In the present case, the prefrontal l2 profiles at 1930 UTC for flow over the front and divide (Figs. 21b and 22b, respectively) generally decreased with height up to 8 km MSL and then increased with height aloft to values larger than those at low levels, thus indicating conditions favorable for “leaky” trapped lee waves (Wurtele et al. 1987). The subtle differences between these two l2 profiles arose solely from differences in the cross-barrier wind profiles (Figs. 21a and 22a); the thermodynamic profiles were identical. Only 1 h later, the front-relative and mountain-relative l2 profiles at 2030 UTC (Figs. 21b and 22b, respectively) had evolved dramatically below 4 km MSL in response to thermodynamic and kinematic changes associated with the shallow frontal passage.

To quantitatively assess the role of this evolving environment on wave structure and response, an analytic method that utilizes the l2 profiles in Figs. 21b and 22b was employed. For steady-state flow over a two-dimensional obstacle, the linearized vertical-structure relationship is
2ŵz2l2k2xŵ
where ŵ = ŵ(kx, z) is the Fourier amplitude of vertical velocity, kx is the horizontal wavenumber in the downstream direction, and l2 is the Boussinesq Scorer parameter presented in Eq. (2). The l2 profiles are assumed to represent conditions in the unperturbed environment upstream of the obstacle. Analytic solutions for the resonant wave mode (i.e., ŵ = 0 at the ground) based on the front-relative and mountain-relative l2 profiles were obtained using the numerical technique described in Nance (1997). This method predicts the vertical profile of vertical velocity associated with the nonhydrostatic resonant mode, and it also predicts the horizontal wavenumber of this mode, from which the horizontal wavelength is calculated via λx = 2π/kx. The phase and amplitude of the analytic curves was chosen to correspond to the observed vertical-velocity profiles.

2) Analytic solutions for front-relative flow

The analytic vertical-velocity profile in the prefrontal environment for flow over the arctic front is shown in Fig. 23a. This front-relative profile was derived from the prefrontal sounding at 1930 UTC, which was truly an upstream sounding relative to the frontal obstacle. The sinuous character and vertical wavelength of this analytic profile is quite similar to what was observed during the frontal passage at 2012 UTC, thus suggesting that a resonant (i.e., vertically trapped) wave mode was triggered by the advancing arctic front. However, the sign change of vertical velocity in the lower stratosphere indicates that some of the wave energy was able to propagate vertically. Hence, this solution represents a leaky trapped mode. Similar behavior has been described in analytic and numerical treatments of trapped lee waves (e.g., Wurtele et al. 1987; Keller 1994). The horizontal wavelength of this wave solution is λx = 39.8 km. The analytic curves based on the earlier prefrontal soundings (not shown) were structurally similar to the analytic curve based on the 1930 UTC sounding, thus suggesting that the front-relative wave environment did not change notably in the prefrontal regime.

Though the comparison between the predicted and observed profiles is quite good, differences exist (Fig. 23a). Most notably, the predicted curve contains comparatively weak upward motion below 4.5 km MSL in the layer where the shallow and narrow updraft forced by frontal convergence was observed. This disparity reflects the fact that the analytic method cannot simulate the frontal updraft. Above 6.5 km MSL, the analytic solution slightly underestimated the magnitude of the observed vertical velocity. This minor discrepancy quite likely arose because the analytic method is unable to predict vertically propagating gravity waves that were excited by the front (see section 6a). Examination of idealized vertically propagating gravity waves and closer inspection of the available observations lend further support to the interpretation that the presence of these waves contributed to the minor discrepancy described above. Specifically, using two-dimensional hydrostatic linear theory applied to obstacle flows (e.g., Smith 1979), the vertical wavelength of vertically propagating gravity waves can be defined as λz = 2π/l. Vertically averaging the front-relative l2 profile within the troposphere at 1930 UTC (Fig. 21b) yields a λz of 18.1 km for this class of waves, thus giving downward motion in the lower stratosphere where subsidence was underpredicted, and upward motion through the troposphere where ascent was underestimated. In addition, Fig. 18a shows evidence that the downward-motion branch of a frontally forced, vertically propagating gravity wave was observed aloft. Immediately following the shallow frontal passage at 2012 UTC, the wind profiler observed a temporally descending sign change of vertical velocity to downward between 12 and 9.5 km MSL, and an increase in the magnitude of subsidence between 12 and 13.2 km MSL. The spatiotemporal behavior of this subsidence, which was short lived (12–18 min), is consistent with Queney's (1948) idealized section of a vertically propagating gravity wave but in a moving frame of reference relative to the front, assuming a large λz of order 18 km.

After the frontal passage, the front-relative wave environment changed dramatically below 4 km MSL (Fig. 21b), such that the resonant wave mode starting at 2030 UTC (not shown) exhibited a vertical-velocity structure typical of a purely vertically trapped wave, that is, strong ascent in the lower troposphere decaying exponentially with height aloft. Between 1930 and 2030 UTC, the horizontal wavelength of the resonant mode supported by the environment decreased from 39.8 to 13.5 km, the latter being more typical of trapped waves. There was no evidence of a train of frontally forced trapped waves of either (or any) wavelength in the satellite imagery or from the ground-based observations. The extreme nonstationarity associated with the migrating front did not allow for the generation of a train of trapped waves that could propagate downstream intact. Rather, the front most likely generated a single trapped wave that was not able to persist in the unfavorable postfrontal environment.

3) Analytic solutions for mountain-relative flow

The mountain-relative l2 profile in the prefrontal environment at 1930 UTC 9 March (Fig. 22b) supported a resonant wave mode possessing a horizontal wavelength of 25.1 km and a sinusoidal vertical-velocity pattern (Fig. 23b) that resembled the leaky analytic frontal wave solution (Fig. 23a). The analytic vertical-velocity curve at 1930 UTC also resembled the analytic curves that were calculated from the earlier prefrontal soundings (not shown), thus suggesting that the mountain-relative wave environment did not change fundamentally in the prefrontal regime. The comparison between the observed and predicted prefrontal vertical-velocity profiles for the mountain waves (Fig. 23b) is not as favorable (though still realistic) as that for the frontal waves (Fig. 23a). The fact that the mountain-relative l2 profile in the prefrontal environment was calculated from a sounding that was downstream relative to the topographic obstacle quite likely contributed significantly to this enhanced disparity. In addition, the analytic method could not account for vertically propagating mountain waves, thus providing an additional source of error. Given that PLT is 76 km downstream of the divide, the vertically propagating mode only contributes to the wave signature above PLT near or above the tropopause in the region where the amplitude of the analytic vertical motions was too small.

Following the frontal passage, the atmosphere no longer supported mountain-induced resonant waves, since the analytic method could not converge on a solution using the hourly l2 profiles starting at 2030 UTC. Because the mountain-relative analytic results changed fundamentally only with the frontal passage, it is unlikely that this change was simply tied to problems associated with using soundings downstream of the topographic obstacle. Rather, this fundamental change is consistent with the wind profiler observations that showed the amplitude of vertical motions decreasing substantially through 16 km MSL after the frontal passage, and it is consistent with the two-dimensional modeling work of Lee et al. (1989) that shows how shallow cold pools east of the divide can reduce the amplitude of leeside mountain-wave activity aloft.

7. Summary and discussion

Through the integrated analysis of remote sensing and in situ data taken along the Front Range of Colorado on 9–10 March 1993, this study has described the primary interactions between a leeside arctic front and the topographically modulated flows over and around the Rocky Mountains. A summary of these interactions and their significance is outlined below and depicted in Figs. 24 and 25.

The large-scale evolution of this event was characterized by the southeastward migration of a lower-tropospheric cyclone from southwestern Canada into the Dakotas. A trailing arctic cold front extended outward from the cyclone center into the northwesterly current. As is often the case, the air mass behind this front acquired an anticyclonic gyre as it encountered the Cheyenne Ridge in southeastern Wyoming, turning the postfrontal flow from northwesterly in eastern Wyoming to north-northeasterly in northeastern Colorado (Fig. 24b). Thereafter, the arctic air became orographically trapped along the steep eastern slope of the Rocky Mountains (Fig. 24a).

The arctic front propagated in a nonclassical west–east fashion across northeastern Colorado (Fig. 24a). It initially advanced southwestward toward the Front Range foothills and then retreated eastward away from the high terrain. Shortly thereafter, a secondary surge of arctic air migrated westward into the foothills. The term nonclassical is used to describe the arctic front's initial west–east oscillation because the front did not exhibit typical quasi-steady forward motion, though its behavior is not necessarily unusual in regions of steep terrain such as in eastern Colorado (e.g., Neiman et al. 1991; Darby et al. 1999). Topographic modulation of the prefrontal flow ahead of the leeside arctic front played a significant role in impacting the propagation characteristics of the front (Figs. 24b, 25). Specifically, the initial southwestward progress of the front toward the foothills was interrupted by strong prefrontal downslope northwesterly flow primarily associated with mountain wave activity downstream of the Continental Divide. The direction and depth of flow within the postfrontal arctic air also influenced the frontal propagation in northeastern Colorado. The cold air behind the initial arctic front was shallow (<1 km deep) and contained north-northeasterly flow, whereas the arctic air behind the secondary cold surge possessed deeper and stronger upslope-component flow. These observations are consistent with the two-dimensional, geostrophic modeling results of Lee et al. (1989) that show a dependence between shallow cold air flushing eastward away from the Continental Divide and the direction and depth of flow within the cold air, that is, a decrease in the upslope component in the cold air and a decrease in its depth increases the likelihood of flushing.

The initial arctic frontal passage exhibited density-current characteristics (Fig. 25), that is, a scale-contracted transition in the kinematic and thermodynamic fields, a thermally direct vertical circulation with a strong updraft and rope cloud at the leading edge of the front, and an elevated frontal head followed by a vertically depressed wake region. In addition, the front propagated across the Platteville, Colorado, remote sensing site at a speed consistent with density-current theory. These observations are consistent with previous studies (e.g., Young and Johnson 1984; Shapiro 1984; Shapiro et al. 1985; Neiman et al. 1991) that show the leading edge of orographically trapped leeside cold fronts often acquire density-current characteristics during their southward migration across eastern Colorado. Because the arctic front on 9 March behaved as a density current, its phase speed was linked to both the opposing downslope northwesterly flow and to the front's density contrast. Since the strength of the opposing downslope flow was stronger closer to the mountains, the front stalled as it approached the mountains. Sensible heating in the arctic air and weak cold advection in the warm downslope flow decreased the density contrast across the front with time, thus also contributing to the cessation of its westward motion. Neither the retreating arctic front nor the secondary surge subsequently possessed density-current attributes.

During its initial southwestward advance the density-current-like arctic front acted as an obstacle to the ambient downslope northwesterly flow, resulting in the generation of vertically propagating gravity waves that produced lenticular wave clouds aloft that tracked with the front (Fig. 25). The front may have also excited a vertically trapped or resonant wave, as will be summarized in the following paragraph. Wesley et al. (1995) and Davis (1997) show that regional cloud and precipitation distributions in northeastern Colorado can be created by postfrontal flow deflected around the Cheyenne Ridge, but that these distributions are generally confined to low levels. In our case, this deflected postfrontal flow resulted in a density-current-like front in northeastern Colorado that acted as an obstacle to the prefrontal downslope flow and excited deep-tropospheric gravity waves and lenticular wave clouds. Because the front entered a region where strong downslope winds and mountain waves extended downstream over the high plains, the wave field over Platteville included both frontally forced, and true mountain-forced, gravity waves. It should be noted that the relative positions of the frontal wave clouds and front were not constant (e.g., Fig. 15), which is not surprising given that the phase relationship between the frontal waves and mountain waves changed with time as the front propagated through the mountain wave environment (Fig. 25).

A sequence of Scorer parameter profiles calculated from hourly wind profiler/RASS measurements at Platteville reveals a sharp contrast between the prefrontal and postfrontal wave environments. Analytic resonant wave mode calculations based on the Scorer parameter profiles reveal that the waves supported in the postfrontal regime differed markedly from those supported in the prefrontal environment, as is illustrated in the cross-section conceptualization in Fig. 25. Specifically, during the 1-h period straddling the frontal passage, the horizontal resonant wavelength of the analytic frontal wave solution decreased from 39.8 to 13.5 km, and the vertical structure of the analytic vertical-velocity profile also changed markedly. There was no evidence of a train of frontally forced trapped waves of either (or any) wavelength in the satellite imagery or from the ground-based observations. The extreme nonstationarity associated with the migrating front did not allow for the generation of a train of trapped waves that could propagate downstream intact. Rather, the front most likely generated a single trapped wave that was not able to persist in the unfavorable postfrontal environment. For mountain-relative flow, the atmosphere supported vertically trapped, resonant mountain waves at the start of the 1-h period straddling the frontal passage but not at the end of this period. This result is consistent with wind-profiler observations that showed the amplitude of vertical motions decreasing substantially through 16 km MSL after the frontal passage. This result also supports the two-dimensional modeling work of Lee et al. (1989) that shows how shallow cold pools east of the divide can reduce the amplitude of leeside mountain wave activity aloft.

The behavior of shallow cold fronts along the eastern slope of the Rockies, and along other large mountain ranges, presents special forecast challenges, especially when several types of terrain-induced flows simultaneously impact the frontal evolution. Although studies of individual weather phenomenon promote a greater understanding that can help improve forecasting techniques, it is also illustrative to show the complex evolutions that can occur when several phenomena interact in tandem. As such, this study presented not only an in-depth analysis of a shallow, leeside arctic cold front, but it also described how the front was impacted by, and interacted with, a variety of terrain-modulated flows and how these interactions excited gravity waves over the front that extended through the troposphere. To date, operational numerical models are limited in their ability to resolve the structural and dynamical details of terrain-modulated flows and fronts, and their mutual interactions, because of the relatively coarse spatial resolution of these models and the uncertainties associated with initializing them. These limitations can, and do, result in simulations that do not adequately reflect the complexities of the real atmosphere. Through comprehensive observational studies such as the one presented here, dynamical and physical understanding of these complicated structures and evolutions can be advanced, and model deficiencies can be overcome, resulting in improved forecasts within, and downstream of, mountainous terrain.

Acknowledgments

We extend special thanks to the two anonymous reviewers, and to Lisa Darby of NOAA/ETL and Ed Szoke of NOAA/FSL, for providing many insightful comments and suggestions that enhanced the scope and quality of this manuscript. Brooks Martner, Sergey Matrosov, and Allen White participated in fruitful roundtable discussions regarding the interpretation of Ka-band radar and laser ceilometer observations and analyses. Michelle Ryan prepared many of the Ka-band analyses used in these discussions. Jim Adams and Al Romero provided exceptional drafting services.

REFERENCES

  • Adams, M. E., S. Businger, and M. L. Kaplan, 1995: Terrain-induced midtropospheric frontogenesis and jet streak formation during STORM-FEST IOP-17. Preprints, Seventh Conf. on Mountain Meteorology, Breckenridge, CO, Amer. Meteor. Soc., 96–100.

    • Search Google Scholar
    • Export Citation
  • Bernard, A., D. Demaiffe, N. Mattielli, and R. S. Punongbayan, 1991:: Anhydrite-bearing pumices from Mount Pinatubo: Further evidence for the existence of sulphur-rich silicic magmas. Nature, 354 , 139140.

    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., 1993: Observations and Theory of Weather Systems. Vol. II, Synoptic-Dynamic Meteorology in Midlatitudes, Oxford University Press, 594 pp.

    • Search Google Scholar
    • Export Citation
  • Caccia, J-L., M. Crochet, and K. Saada, 1997: ST radar evaluation of the standard deviation of the air vertical velocity perturbed by the local orography. J. Atmos. Solar-Terr. Phys, 59 , 11271131.

    • Search Google Scholar
    • Export Citation
  • Carbone, R. E., 1982: A severe frontal rainband. Part I: Stormwide hydrodynamic structure. J. Atmos. Sci, 39 , 258279.

  • Clark, T. L., W. D. Hall, and R. M. Banta, 1994: Two- and three-dimensional simulations of the 9 January 1989 severe Boulder windstorm: Comparison with observations. J. Atmos. Sci, 51 , 23172343.

    • Search Google Scholar
    • Export Citation
  • Colle, B. A., and C. F. Mass, 1995: The structure and evolution of cold surges east of the Rocky Mountains. Mon. Wea. Rev, 123 , 25772610.

    • Search Google Scholar
    • Export Citation
  • Crook, N. A., 1988: Trapping of low-level internal gravity waves. J. Atmos. Sci, 45 , 15331541.

  • Darby, L. S., W. D. Neff, and R. M. Banta, 1999: Multiscale analysis of a meso-β frontal passage in the complex terrain of the Colorado Front Range. Mon. Wea. Rev, 127 , 20622081.

    • Search Google Scholar
    • Export Citation
  • Davis, C. A., 1997: Mesoscale anticyclonic circulations in the lee of the central Rocky Mountains. Mon. Wea. Rev, 125 , 28382855.

  • Durran, D. R., 1986a: Mountain waves. Mesoscale Meteorology and Forecasting, P. S. Ray, Ed., Amer. Meteor. Soc., 472–492.

  • Durran, D. R., 1986b: Another look at downslope windstorms. Part I: The development of analogs to supercritical flow in an infinitely deep, continuously stratified fluid. J. Atmos. Sci, 43 , 25272543.

    • Search Google Scholar
    • Export Citation
  • Durran, D. R., 1990: Mountain waves and downslope winds. Atmospheric Processes over Complex Terrain, Meteor. Monogr., No. 45, Amer. Meteor. Soc., 59–81.

    • Search Google Scholar
    • Export Citation
  • Fulton, R., D. S. Zrnić, and R. J. Doviak, 1990: Initiation of a solitary wave family in the demise of a nocturnal thunderstorm density current. J. Atmos. Sci, 47 , 319337.

    • Search Google Scholar
    • Export Citation
  • Fritts, D. C., and G. D. Nastrom, 1992: Sources of mesoscale variability of gravity waves. Part II: Frontal, convective, and jet stream excitation. J. Atmos. Sci, 49 , 111127.

    • Search Google Scholar
    • Export Citation
  • Hartjenstein, G., and R. Bleck, 1991: Factors affecting cold-air outbreaks east of the Rocky Mountains. Mon. Wea. Rev, 119 , 22802292.

    • Search Google Scholar
    • Export Citation
  • Hogg, D. C., and and Coauthors, 1983: An automated profiler of temperature, wind, and humidity in the troposphere. J. Climate Appl. Meteor, 22 , 807831.

    • Search Google Scholar
    • Export Citation
  • Jin, Y., S. E. Koch, Y-L. Lin, F. M. Ralph, and C. Chen, 1996: Numerical simulations of an observed gravity current and gravity waves in an environment characterized by complex stratification and shear. J. Atmos. Sci, 53 , 35703588.

    • Search Google Scholar
    • Export Citation
  • Keller, T. L., 1994: Implications of the hydrostatic assumption on atmospheric gravity waves. J. Atmos. Sci, 51 , 19151929.

  • Koch, S. E., 1984: The role of an apparent mesoscale frontogenetical circulation in squall line initiation. Mon. Wea. Rev, 112 , 20902111.

    • Search Google Scholar
    • Export Citation
  • Kropfli, R. A., and and Coauthors, 1995: Cloud physics studies with 8 mm wavelength radar. Atmos. Res, 35 , 299313.

  • Kuettner, J. P., P. A. Hildebrand, and T. L. Clark, 1987: Convection waves: Observations of gravity wave systems over convectively active boundary layers. Quart. J. Roy. Meteor. Soc, 113 , 445467.

    • Search Google Scholar
    • Export Citation
  • Lee, T. J., R. A. Pielke, R. C. Kessler, and J. Weaver, 1989: Influence of cold pools downstream of mountain barriers on downslope winds and flushing. Mon. Wea. Rev, 117 , 20412058.

    • Search Google Scholar
    • Export Citation
  • Lilly, D. K., and E. J. Zipser, 1972: The Front Range windstorm of 11 January 1972—A meteorological narrative. Weatherwise, 25 , 5663.

    • Search Google Scholar
    • Export Citation
  • Martner, B. E., and J. D. Marwitz, 1982: Wind characteristics in southern Wyoming. J. Appl. Meteor, 21 , 18151827.

  • Marwitz, J. D., and P. J. Dawson, 1984: Low-level airflow in southern Wyoming during wintertime. Mon. Wea. Rev, 112 , 12461262.

  • May, P. T., K. P. Moran, and R. G. Strauch, 1989: The accuracy of RASS temperature profiles. J. Appl. Meteor, 28 , 13291335.

  • Mecikalski, J. R., and J. S. Tilley, 1992: Cold surges along the Front Range of the Rocky Mountains: Development of a classification scheme. Meteor. Atmos. Phys, 48 , 249271.

    • Search Google Scholar
    • Export Citation
  • Nance, L. B., 1997: On the inclusion of compressibility effects in the Scorer parameter. J. Atmos. Sci, 54 , 362367.

  • Nance, L. B., and D. R. Durran, 1997: A modeling study of nonstationary trapped mountain lee waves. Part I: Mean-flow variability. J. Atmos. Sci, 54 , 22752291.

    • Search Google Scholar
    • Export Citation
  • Nastrom, G. D., M. R. Peterson, J. L. Green, K. S. Gage, and T. E. VanZandt, 1990: Sources of gravity wave activity seen in the vertical velocities observed by the Flatland VHF radar. J. Appl. Meteor, 29 , 783792.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., R. M. Hardesty, M. A. Shapiro, and R. E. Cupp, 1988:: Doppler lidar observations of a downslope windstorm. Mon. Wea. Rev, 116 , 22652275.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., P. T. May, B. B. Stankov, and M. A. Shapiro, 1991: Radio acoustic sounding system observations of an arctic front. J. Appl. Meteor, 30 , 881892.

    • Search Google Scholar
    • Export Citation
  • Neiman, P. J., P. T. May, and M. A. Shapiro, 1992: Radio acoustic sounding system (RASS) and wind profiler observations of lower- and middle-tropospheric weather systems. Mon. Wea. Rev, 120 , 22982313.

    • Search Google Scholar
    • Export Citation
  • Politovich, M. K., and B. C. Bernstein, 1995: Production and depletion of supercooled liquid water in a Colorado winter storm. J. Appl. Meteor, 34 , 26312648.

    • Search Google Scholar
    • Export Citation
  • Post, M. J., and R. E. Cupp, 1990: Optimizing a pulsed Doppler lidar. Appl. Opt, 29 , 41454158.

  • Post, M. J., A. Weickmann, K. R. Healy, R. J. Willis, and C. Grund, 1996:: Comparison of Mount Pinatubo and El Chichon volcanic events: Lidar observations at 10.6 and 0.69 μm. J. Geophys. Res, 101 , 39293940.

    • Search Google Scholar
    • Export Citation
  • Pratte, J. F., and L. Kaimal, 1986: Surface mesonetwork operator's manual. 2d version. 121 pp. [Available from NOAA/ERL/Forecast Systems Laboratory, Boulder, CO 80303.].

    • Search Google Scholar
    • Export Citation
  • Queney, P., 1948: The problem of airflow over mountains: A summary of theoretical studies. Bull. Amer. Meteor. Soc, 29 , 1626.

  • Ralph, F. M., M. Crochet, and S. V. Venkateswaran, 1992: A study of mountain lee waves using clear-air radar. Quart. J. Roy. Meteor. Soc, 118 , 597627.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., C. Mazaudier, M. Crochet, and S. V. Venkateswaran, 1993: Doppler sodar and radar wind-profiler observations of gravity-wave activity associated with a gravity current. Mon. Wea. Rev, 121 , 444463.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., P. J. Neiman, D. W. van de Kamp, and D. C. Law, 1995: Using spectral moment data from NOAA's 404-MHz radar wind profilers to observe precipitation. Bull. Amer. Meteor. Soc, 76 , 17171739.

    • Search Google Scholar
    • Export Citation
  • Ralph, F. M., P. J. Neiman, and T. L. Keller, 1999: Deep-tropospheric gravity waves created by leeside cold fronts. J. Atmos. Sci, 56 , 29863009.

    • Search Google Scholar
    • Export Citation
  • Rasmussen, R. M., and and Coauthors, 1992: The Winter Icing and Storms Project (WISP). Bull. Amer. Meteor. Soc, 73 , 951974.

  • Rasmussen, R. M., B. C. Bernstein, M. Murakami, G. Stossmeister, and J. Reisner, 1995: The 1990 Valentine's Day arctic outbreak. Part I: Mesoscale and microscale structure and evolution of a Colorado Front Range shallow upslope cloud. J. Appl. Meteor, 34 , 14811511.

    • Search Google Scholar
    • Export Citation
  • Schaefer, J. T., 1986: The dryline. Mesoscale Meteorology and Forecasting, P. Ray, Ed., Amer. Meteor. Soc., 549–570.

  • Schubert, W. H., S. K. Cox, P. E. Ciesielski, and C. M. Johnson-Pasqua, 1987: Operation of a ceilometer during the FIRE marine stratocumulus experiment. Atmospheric Science Paper 420, Colorado State University.

    • Search Google Scholar
    • Export Citation
  • Scorer, R. S., 1949: Theory of lee waves of mountains. Quart. J. Roy. Meteor. Soc, 75 , 4156.

  • Shapiro, M. A., 1984: Meteorological tower measurements of a surface cold front. Mon. Wea. Rev, 112 , 16341639.

  • Shapiro, M. A., and D. Keyser, 1990: Fronts, jet streams, and the tropopause. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. Holopainen, Eds., Amer. Meteor. Soc., 167–191.

    • Search Google Scholar
    • Export Citation
  • Shapiro, M. A., T. Hampel, D. Rotzoll, and F. Mosher, 1985: The frontal hydraulic head: A micro-α-scale (∼1 km) triggering mechanism for mesoconvective weather systems. Mon. Wea. Rev, 113 , 11661183.

    • Search Google Scholar
    • Export Citation
  • Shutts, G. J., and A. Broad, 1993: A case study of lee waves over the Lake District in northern England. Quart. J. Roy. Meteor. Soc, 119 , 377408.

    • Search Google Scholar
    • Export Citation
  • Simpson, J. E., and R. E. Britter, 1980: A laboratory model of an atmospheric mesofront. Quart. J. Roy. Meteor. Soc, 106 , 485500.

  • Smith, R. B., 1979: The influence of mountains on the atmosphere. Advances in Geophysics, Vol. 21, Academic Press, 87–230.

  • Steenburgh, W. J., and C. F. Mass, 1994: The structure and evolution of a simulated Rocky Mountain lee trough. Mon. Wea. Rev, 122 , 27402761.

    • Search Google Scholar
    • Export Citation
  • Tripoli, S., A. Buzzi, and A. Speranza, 1990: Orographic cyclogenesis. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. Holopainen, Eds., Amer. Meteor. Soc., 107–127.

    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., 1982: The life cycle of thunderstorm gust fronts as viewed with Doppler radar and rawinsonde data. Mon. Wea. Rev, 110 , 10601082.

    • Search Google Scholar
    • Export Citation
  • Weber, B. L., D. B. Wuertz, D. C. Law, A. S. Frisch, and J. M. Brown, 1992: Effects of small-scale vertical motion on radar measurements of wind and temperature profiles. J. Atmos. Oceanic Technol, 9 , 193209.

    • Search Google Scholar
    • Export Citation
  • Wesley, D. A., R. M. Rasmussen, and B. C. Bernstein, 1995: Snowfall associated with a terrain-generated convergence zone during the Winter Icing and Storms Project. Mon. Wea. Rev, 123 , 29572977.

    • Search Google Scholar
    • Export Citation
  • Westwater, E. R., and F. O. Guiraud, 1980: Ground-based microwave radiometric retrieval of precipitable water vapor in the presence of clouds with high liquid content. Radio Sci, 15 , 947957.

    • Search Google Scholar
    • Export Citation
  • Wolfe, D., and and Coauthors, 1995: An overview of the mobile profiler system: Preliminary results from field tests during the Los Angeles free-radical study. Bull. Amer. Meteor. Soc, 76 , 523534.

    • Search Google Scholar
    • Export Citation
  • Wuertz, D. B., B. L. Weber, R. G. Strauch, A. S. Frisch, C. G. Little, D. A. Merritt, K. P. Moran, and D. C. Welsh, 1988: Effects of precipitation on UHF wind profiler measurements. J. Atmos. Oceanic Technol, 5 , 452464.

    • Search Google Scholar
    • Export Citation
  • Wurtele, M. G., R. D. Sharman, and T. L. Keller, 1987: Analysis and simulations of a troposphere–stratosphere gravity wave model. Part I. J. Atmos. Sci, 44 , 32693281.

    • Search Google Scholar
    • Export Citation
  • Young, G. S., and R. H. Johnson, 1984: Meso- and microscale features of a Colorado cold front. J. Climate Appl. Meteor, 23 , 13151325.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Base map of northeastern Colorado showing terrain contours (m), the 22 surface mesonet locations (•); the heavily instrumented remote sensing site at Platteville, CO (▴); the Doppler lidar site (▪), the operational SAO surface sites (★), and the operational rawinsonde site at DEN (+). The Continental Divide is portrayed as a bold dashed line. The location of downtown Denver (○) is also shown. The domain of this base map is shown in Figs. 2b, 3b, and 5–7

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 2.
Fig. 2.

The 700-mb geopotential height (dam, thin solid) and potential temperature (K, bold dashed) analysis at (a) 1200 UTC 9 Mar 1993 and (b) 0000 UTC 10 Mar. The 700-mb wind vector flags are 25 m s−1, full barbs are 5 m s−1, and half-barbs are 2.5 m s−1. Wind vectors with solid-dot heads in (b) are wind-profiler observations. The 700-mb geopotential height (potential temperature) rounded to the nearest decameter (K) is plotted to the right (left) of the corresponding wind vector. The bold dotted line AA′ in (b) is the projection line for the cross section in Fig. 4. The profiler observation at Platteville, CO, in (b) is labeled PLT, and the two profiler measurements in Nebraska are at Merriman (north) and McCook (south). The outer dashed box in (b) corresponds to the regional mesoscale domain shown in Figs. 5–7, while the inner dashed box portrays the fine mesoscale domain in Figs. 1 and 8. The thin dotted line is the Continental Divide

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 3.
Fig. 3.

The sea level pressure (mb, thin solid) and 3-h pressure change [shading: light, ≤−2 mb (3 h)−1; dark, ≥2 mb (3 h)−1] analysis at (a) 1200 UTC 9 Mar 1993, and (b) 0000 UTC 10 Mar. The surface wind barbs, the projection line AA′, the dashed boxes, and the Continental Divide are the same as in Fig. 2. The arctic front in each panel is represented by the bold line with closely spaced frontal symbols; all other fronts have widely spaced conventional symbols. The dryline is shown as a bold dashed line

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 4.
Fig. 4.

Cross section of potential temperature (K, solid) and alongfront wind speed directed toward 125° (m s−1, bold dashed), along line AA′ of Figs. 2b and 3b at 0000 UTC 10 Mar 1993. The light-shaded region is the low-level cross-front flow (directed toward 195°) greater than 8 m s−1. Thin-dashed lines are frontal boundaries and the tropopause. The letter J marks the jet stream core aloft. Rawinsonde, wind-profiler, and surface sites are marked with long-bold, long-thin, and short vertical tick marks, respectively. Selected rawinsonde and wind profiler wind vectors are shown. Wind flags and barbs are the same as in Fig. 2

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 5.
Fig. 5.

Regional mesoscale analysis of sea level pressure (mb, thin solid) at (a) 1800 UTC 9 Mar 1993, (b) 2100 UTC 9 Mar, (c) 0000 UTC 10 Mar, and (d) 0300 UTC 10 Mar. Frontal boundaries and surface wind barbs are the same as in Fig. 3. Terrain is shaded: light, 1500–2500 m; dark,>2500 m. The domain of this analysis is shown in Figs. 2b and 3b. The dashed box corresponds to the fine mesoscale domain in Figs. 1 and 8

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 6.
Fig. 6.

The same as in Fig. 5 except for surface potential temperature (θ; K)

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 7.
Fig. 7.

The same as in Fig. 5 except for 3-h surface pressure change (mb) ending at the designated times

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 8.
Fig. 8.

Frontal isochrone analysis at the surface between (a) 1800 and 2130 UTC 9 Mar, (b) 2130 UTC 9 Mar and 0030 UTC 10 Mar, and (c) 0030 and 0500 UTC 10 Mar. The bold isochrones with conventional frontal symbols (as in Fig. 3) at 2000 UTC 9 Mar, 2230 UTC 9 Mar and 0330 UTC 10 Mar correspond to the times of the plotted data. Surface wind barbs (as in Fig. 2) and potential temperatures (K, rounded to the nearest degree) at these times are plotted. The lidar site and remote sensing site are marked with the symbols ▪ and ▴, respectively. Terrain is shaded: light, 2000–3000 m; dark,>3000 m. The domain of this analysis is shown in Figs. 2b, 3b, and 5–7

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 9.
Fig. 9.

Mesonet time series traces (5-min resolution) of wind direction (DIR; °, bold), wind speed gust (SPD; m s−1), 1-h wind vectors (as in Fig. 2), temperature (T; °C), dewpoint temperature (Td; °C), and surface pressure (PRS; mb) between 1600 UTC 9 Mar and 1200 UTC 10 Mar 1993 at (a) BRI and (b) LTN. The vertical dashed lines labeled 1, 2, and 3 mark the initial arctic frontal advance, the frontal retreat, and the secondary frontal surge, respectively. See Fig. 1 for site locations

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

 Fig. 10.
Fig. 10.

Doppler lidar PPI scan of radial velocity (m s−1) at 2135 UTC 9 Mar 1993. The elevation angle of the scan was 1° above horizontal. The bold white arrow denotes the direction of the downslope flow (i.e., from the northwest). Frontal boundaries are the same as in Fig. 3. Because the elevation of the PPI scan is a function of range, the lower-tropospheric frontal position east of the lidar is not valid at a single level. Rather, it is valid within a narrow layer between ∼550 and 675 m above ground. For reference, the lidar location and the surface frontal position at 2130 UTC (based on mesonet observations) are shown on the isochrone analyses in Figs. 8a and 8b

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 11.
Fig. 11.

Isotach analysis of the horizontal wind (m s−1) in the plane of the cross section, derived from the Doppler lidar west–east RHI scan of radial wind velocity at 2345 UTC 9 Mar 1993. The downslope jet and mountain profile are shown in light and dark gray, respectively. The lidar is located at x = 0 km. When elevation scan angles are high (i.e., near x = 0) and vertical air motions are large, the horizontal wind derivations become less reliable. Upward motion was observed directly above the lidar during this scan (not shown); hence, the horizontal wind immediately upstream of the lidar is underestimated and the horizontal wind immediately downstream of the lidar is overestimated for steep scanning angles

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 12.
Fig. 12.

(a), (b) Time–height sections and (c) time series of data from PLT between 1500 UTC 9 Mar and 1000 UTC 10 Mar 1993. (a) Wind profiles and RASS virtual potential temperature analysis (θυ; K, solid) based on the 15-min-averaged profiles measured by the 924-MHz wind profiler. The dashed isentropes are based on hourly RASS θυ profiles from the 404-MHz profiler. The 6-min resolution surface winds from PLT are included. Every other wind profile and surface wind is shown. The complete time series of wind profiles were subjected to one pass of a 1–2–1 temporal Hann filter. Wind flags and barbs are the same as in Fig. 2. Bold dots show cloud-base observations measured by the ceilometer. (b) Wind profiles and zonal wind speed analysis (m s−1, solid) based on the 15-min-averaged profiles measured by the 924-MHz profiler. The 0 m s−1 isotach is bold. Wind profiles, surface winds, and wind flags and barbs are as in (a). (c) Time series traces of data from the surface meteorological site [T = temperature (°C); Td = dewpoint temperature (°C); PRS = surface pressure (mb)] and from the dual-channel radiometer [VPR = integrated precipitable water vapor (cm); LQD = integrated liquid water (mm)]. The vertical dashed lines are the same as in Fig. 9. Bold lines on the top and bottom of each figure frame show the temporal domain of the PLT profiler analyses in Figs. 13 and 14

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 13.
Fig. 13.

Time–height section of 3-min resolution data from the 924-MHz wind profiler at PLT between 1945 UTC 9 Mar and 0000 UTC 10 Mar 1993: (a) wind profiles, zonal wind speed analysis (m s−1, solid), and vertical-velocity shading (light, ≤−0.5 m s−1; dark, ≥0.5 m s−1), and (b) wind profiles and vertical-velocity analysis (m s−1, solid). The 6-min resolution surface winds from PLT are included. The profiles of horizontal and vertical wind were subjected to one pass of a 1–2–1 temporal Hann filter. The 0 m s−1 isotach in (a) is bold. Wind flags and barbs are the same as in Fig. 2. The bold dot at 2012 UTC and 2.4 km MSL in (b) is a rope cloud observation from the PLT ceilometer. The vertical dashed lines are as in Fig. 9

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 14.
Fig. 14.

The same as in Fig. 13 except between 0045 and 0300 UTC 10 Mar 1993

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 15.
Fig. 15.

Sequence of four visible satellite images at (a) 1830, (b) 1930, (c) 2030, and (d) 2130 UTC on 9 Mar 1993. The arctic frontal positions are based on the 3-h regional mesoscale analyses in Figs. 5–7 and on the isochrone analyses in Fig. 8. The frontal symbols are the same as in Fig. 3. The bold white dot in each frame marks the heavily instrumented remote sensing site at PLT

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 16.
Fig. 16.

Photograph of lenticular wave clouds (ACSLs) and a shallow frontal rope cloud (rope) at ∼2100 UTC 9 Mar 1993 looking east-southeast (∼120°) from the Doppler lidar

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 17.
Fig. 17.

Ka-band radar west–east RHI scan (oriented 255°–75°) of reflectivity [dBZ; ranging between −30 dBZ (light) and −10 dBZ (dark)] at 2003 UTC 9 Mar 1993 at PLT. The frontal position, which is based on measurements from the various remote sensing and in situ instruments at PLT, is also shown. The radar is located at x = 0 km

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 18.
Fig. 18.

(a) Vertically stacked time series of vertical velocity (vertical bars, scale at upper right) measured by the 924- and 404-MHz wind profilers at PLT between 1824 and 2100 UTC 9 Mar 1993. Every other 3-min profile measured by the 924-MHz profiler below 3.5 km MSL is included in the time series, while every 6-min profile from the 404-MHz system is used above that. Light shading under the curves represents upward motion. The profiles of vertical wind were subjected to one pass of a 1–2–1 temporal Hann filter. Clouds measured directly above the collocated Ka-band cloud-sensing radar with sequential (∼2 min resolution) RHI scans are shown in dark shading starting at 1924 UTC. Only those clouds that exhibited a significant spatiotemporal history are shown. Surface winds (barbs are as in Fig. 2) from PLT are included. The vertical dashed line marks the initial arctic frontal passage, and the horizontal dashed line denotes the mean height of the tropopause. The position of the arctic front (bold solid) is based on the isotach analysis in Fig. 13a. The five bold dots near the bottom of the frame mark the times of individual vertical-velocity profiles shown in Fig. 19, and the two short thick tick marks correspond to the times of the thermodynamic soundings and Scorer parameter analyses shown in Figs. 20–22. (b) Time series trace of integrated liquid water (mm) at PLT

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 19.
Fig. 19.

Profiles of Hann-filtered vertical velocity measured by the 924- and 404-MHz wind profilers at PLT at 1918, 1954, 2000, 2012, and 2018 UTC 9 Mar 1993. The time-averaged, mountain wave–induced vertical-velocity profile is shown in bold gray shade. The height range of the front-induced ACSLs observed by the Ka-band radar, from their initial sighting at 1957 UTC to the time of the last vertical-velocity profile at 2018 UTC, is shown with the thin dashed lines. The depth of the well-mixed boundary layer, based on Fig. 12, is also shown

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 20.
Fig. 20.

Skew T–logp hybrid soundings at 1930 (dashed) and 2030 (solid) UTC 9 Mar 1993. The thermodynamic profiles below 4.5 km MSL were derived from Platteville's RASS θυ analysis in Fig. 12a at 1930 and 2030 UTC. Above the upper range of the RASS observations (i.e., above ∼4.5 km MSL), the profiles were completed by temporally interpolating the 1200 UTC 9 Mar and 0000 UTC 10 Mar Denver rawinsonde soundings to 1930 and 2030 UTC. The wind profiles below 5 km MSL were generated from Platteville's 924-MHz 15-min-averaged profiles at 1930 and 2030 UTC. The upper portions of these profiles were completed using the hourly averaged wind profiles from the Platteville 404-MHz profiler at 1930 and 2030 UTC. The thermodynamic and wind profiles were smoothed vertically using five passes of a 1–2–1 Hann filter after the data were transformed onto a uniform 250-m grid. Winds are the same as in Fig. 2. Every other wind measurement on the 250-m grid is shown

Citation: Monthly Weather Review 129, 11; 10.1175/1520-0493(2001)129<2633:OONFPA>2.0.CO;2

Fig. 21.