• Banta, R. M., , L. D. Olivier, , and D. H. Levinson, 1993: Evolution of the Monterey Bay sea-breeze layer as observed by pulsed Doppler lidar. J. Atmos. Sci., 50 , 39593982.

    • Search Google Scholar
    • Export Citation
  • Banta, R. M., and Coauthors, 1997: Nocturnal cleansing flows in a tributary valley. Atmos. Environ., 31 , 21472162.

  • Banta, R. M., , L. D. Olivier, , P. Kaufman, , D. H. Levinson, , and C. J. Zhu, 1999: Wind flow patterns in the Grand Canyon as revealed by Doppler lidar. J. Appl. Meteor., 38 , 10691083.

    • Search Google Scholar
    • Export Citation
  • Banta, R. M., , L. S. Darby, , J. D. Fast, , J. O. Pinto, , C. D. Whiteman, , W. J. Shaw, , and B. D. Orr, 2004: Nocturnal low-level jet in a mountain basin complex. Part I: Evolution and implications to other flow features. J. Appl. Meteor., 43 , 13481365.

    • Search Google Scholar
    • Export Citation
  • Bedard Jr., A. J., , and P. Neilley, 1998: Project MCAT (Mountain Induced Clear Air Turbulence): Background, goals, instrumentation and methodologies. Preprints, Eighth Conf. on Mountain Meteorology, Flagstaff, AZ, Amer. Meteor. Soc., 163–171.

  • Brinkmann, W. A. R., 1974: Strong downslope winds at Boulder, Colorado. Mon. Wea. Rev., 102 , 592602.

  • 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
  • Clark, T. L., , W. D. Hall, , R. M. Kerr, , D. Middleton, , L. Radke, , F. M. Ralph, , P. J. Neiman, , and D. Levinson, 2000: Origins of aircraft-damaging clear-air turbulence during the 9 December 1992 Colorado downslope windstorm: Numerical simulations and comparison with observations. J. Atmos. Sci., 57 , 11051131.

    • Search Google Scholar
    • Export Citation
  • 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
  • Darby, L. S., , A. Gohm, , L. B. Nance, , S. Gabersek, , R. M. Banta, , and S. Sandberg, 2000: Foehn flow in the Austrian Alps interrupted by a cold front passage, Part I. Preprints, Ninth Conf. on Mountain Meteorology, Aspen, CO, Amer. Meteor. Soc., 79–82.

  • Darby, L. S., , R. M. Banta, , and R. A. Pielke Sr., 2002: Comparisons between mesoscale model terrain sensitivity studies and Doppler lidar measurements of the sea breeze at Monterey Bay. Mon. Wea. Rev., 130 , 28132838.

    • Search Google Scholar
    • Export Citation
  • Doyle, J. D., , and D. R. Durran, 2002: The dynamics of mountain-wave induced rotors. J. Atmos. Sci., 59 , 186201.

  • Doyle, J. D., and Coauthors, 2000: An intercomparison of model-predicted wave breaking for the 11 January 1972 Boulder windstorm. Mon. Wea. Rev., 128 , 901914.

    • Search Google Scholar
    • Export Citation
  • Eastman, J. L., , R. A. Pielke, , and D. J. McDonald, 1998: Calibration of soil moisture for large-eddy simulations over the FIFE area. J. Atmos. Sci., 55 , 11311140.

    • Search Google Scholar
    • Export Citation
  • Fast, J. D., , and L. S. Darby, 2004: An evaluation of mesoscale model predictions of down-valley and canyon flows and their consequences using Doppler lidar measurements during VTMX 2000. J. Appl. Meteor., 43 , 420436.

    • Search Google Scholar
    • Export Citation
  • Fingerhut, W. A., , and P. F. Lester, 1973: Lower turbulent zones associated with mountain lee waves. Department of Meteorology Paper No. 30, San Jose State University, 123 pp.

  • Flamant, C., and Coauthors, 2002: Gap flow in an Alpine Valley during a shallow south foehn event: Observations, numerical simulations, and hydraulic analog. Quart. J. Roy. Meteor. Soc., 128 , 11711210.

    • Search Google Scholar
    • Export Citation
  • Förchtgott, J., 1949: Rotor streaming in the lee of mountain ridges. Bull. Meteor. Czech., 4 , 1416.

  • Förchtgott, J., 1969: Evidence for mountain-sized lee eddies. Weather, 24 , 255260.

  • Gerbier, N., , and M. Berenger, 1961: Experimental studies of lee waves in the French Alps. Quart. J. Roy. Meteor. Soc., 87 , 1323.

  • Hoinka, K. P., 1985: Observation of the airflow over the Alps during a foehn event. Quart. J. Roy. Meteor. Soc., 111 , 199224.

  • Holmboe, J., , and H. Klieforth, 1957: Investigation of mountain lee waves and the air flow over the Sierra Nevada. Final Report, Dept. of Meteorology, UCLA, Contract No. AF19(604)-728, 290 pp.

  • Kuettner, J., 1959: The rotor flow in the lee of mountains. Geophysical Research Division Res. Note 6, Air Force Cambridge Research Center, 20 pp.

  • Kuettner, J., 1968: Lee waves in the Colorado Rockies. Weatherwise, 21 , 180195.

  • Lester, P. F., 1994: Turbulence: A New Perspective for Pilots. Jeppesen Sanderson Training Products, 212 pp.

  • Lester, P. F., 1995: Aviation Weather. Jeppesen Sanderson Training Products, 335 pp.

  • Lester, P. F., , and W. A. Fingerhut, 1974: Lower turbulent zones associated with mountain lee waves. J. Appl. Meteor., 13 , 5461.

  • Lilly, D. K., 1978: A severe downslope windstorm and aircraft turbulence event induced by a mountain wave. J. Atmos. Sci., 35 , 5977.

  • Lilly, D. K., , and W. Toutenhoofd, 1969: The Colorado Lee Wave Program. . Clear Air Turbulence and Its Detection, Y. Pao and A. Goldburg, Eds., Plenum Press, 232–245.

    • Search Google Scholar
    • Export Citation
  • Lyra, G., 1943: Theorie der stationaeren Leewellenstromung in freier Atmosphaere. Z. Angew. Math. Mech, 23 , 128.

  • National Transportation Safety Board, 2001: Aircraft accident Report: Uncontrolled descent and collision with terrain, United Airlines flight 585, Boeing 737-200, N999UA, 4 miles south of Colorado Springs Municipal Airport, Colorado Springs, Colorado, March 3, 1991. NTSB Rep. AAR-01/01, NTIS Rep. PB2001-910401, 214 pp. [Available online at http://www.ntsb.gov/publictn/2001/AAR0101.pdf.].

  • Neff, W. D., 1994: Mesoscale air quality studies with meteorological remote sensing systems. Int. J. Remote Sens., 15 , 393426.

  • Neiman, P. J., 1988: Doppler lidar observations of a downslope wind storm. Mon. Wea. Rev., 116 , 22652275.

  • Neiman, P. J., , F. M. Ralph, , R. L. Weber, , T. Utall, , L. B. Nance, , and D. H. Levinson, 2001: Observations of nonclassical frontal propagation and frontally forced gravity waves adjacent to steep topography. Mon. Wea. Rev., 129 , 26332659.

    • Search Google Scholar
    • Export Citation
  • Nicholls, J. M., 1973: The airflow over the mountains: Research 1958–1972. WMO Tech. Note 127, World Meteorological Organization, Geneva, Switzerland, 73 pp.

  • Pielke, R. A., and Coauthors, 1992: A comprehensive meteorological modeling system—RAMS. Meteor. Atmos. Phys., 49 , 6991.

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

  • Poulos, G. S., 1996: The interaction of katabatic winds and mountain waves. Ph.D. dissertation, Colorado State University, 297 pp.

  • Poulos, G. S., , and J. E. Bossert, 1995: An observational and prognostic numerical investigation of complex terrain dispersion. J. Appl. Meteor., 34 , 650669.

    • Search Google Scholar
    • Export Citation
  • Poulos, G. S., , J. E. Bossert, , T. B. McKee, , and R. A. Pielke, 2000: The interaction of katabatic flow and mountain waves. Part I: Observations and idealized simulations. J. Atmos. Sci., 57 , 19191936.

    • Search Google Scholar
    • Export Citation
  • Queney, P., 1955: Rotor phenomena in the lee of mountains. Tellus, 7 , 367371.

  • Queney, P., , G. Corby, , N. Gerbier, , H. Koshmieder, , and J. Zierep, 1960: The airflow over the mountains. WMO Tech. Note 34, World Meteorological Organization, Geneva, Switzerland, 135 pp.

  • Scorer, R. S., 1955: The theory of airflow over mountains—IV. Separation of flow from the surface. Quart. J. Roy. Meteor. Soc., 81 , 340350.

    • Search Google Scholar
    • Export Citation
  • U.S. General Accounting Office, 1993: Aviation safety: FAA can better prepare general aviation pilots for mountain flying risks. Rep. RCED-94-15, 72 pp. [Available online at http://archive.gao.gov/t2pbat4/150580.pdf.].

  • Vosper, S., , and S. Mobbs, 2002: Numerical simulations of lee-wave rotors. Preprints, 10th Conf. on Mountain Meteorology, Park City, UT, Amer. Meteor. Soc., CD-ROM, P15.3.

  • Weigel, A. P., , and M. W. Rotach, 2004: Flow structure and turbulence characteristics of the daytime atmosphere in a steep and narrow Alpine valley. Quart. J. Roy. Meteor. Soc., 130 , 26052627.

    • Search Google Scholar
    • Export Citation
  • Whelan, R. F., 2000: Exploring the Monster. Mountain Lee Waves: The Aerial Elevator. Wind Canyon Books, 169 pp.

  • Worthington, R. M., 2002: Mountain waves launched by convective activity within the boundary layer above mountains. Bound.-Layer Meteor., 103 , 469491.

    • Search Google Scholar
    • Export Citation
  • Zhong, S., , and J. D. Fast, 2003: An evaluation of the MM5, RAMS, and Meso-Eta models at subkilometer resolution using field campaign data in the Salt Lake Valley. Mon. Wea. Rev., 131 , 13011322.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Map of experiment region, centered on PAF, and the Colorado Springs Airport runways (boldface lines to the left of PAF). Terrain contours are every 100 m from 1100 to 2300 m MSL and every 200 m between 2300 and 4300 m MSL. The Doppler lidar was stationed at PAF. The 915-MHz radar wind profilers were stationed at PAF, WOC, and FCN. The light gray dashed box indicates the approximate location of the RAMS grid 4. The small gray box 12–14-km west of the lidar indicates the region over which the vertical profiles of lidar-measured horizontal winds shown in Fig. 7 were averaged.

  • View in gallery

    Topography of the RAMS model domain. (a) Topography for the outermost grid (grid 1). Terrain contours are every 200 m from 1200 to 3400 m MSL. (b) Topography for the innermost grid (grid 4). Terrain contours are every 100 m from 1800 to 4000 m MSL.

  • View in gallery

    Doppler lidar range–height radial velocity scans pointing toward the west along the 270° radial. West is to the right. Tick marks are in 2-km increments. The color bar at the bottom of each plot represents radial velocities in m s−1. Black arrows indicate the westerly component flow (positive radial velocities). Red arrows indicate the easterly component flow (negative radial velocities). The lidar is located at x = z = 0 and times are UTC. (a) Lee wave and rotor at 0120 UTC. (b) Deep lee wave and rotor at 0610 UTC. The red star indicates approximate position of the research aircraft as it traversed along the foothills at this time. (c) Lee wave and rotors at 0938 UTC, shortly before the cold front passage.

  • View in gallery

    Weather maps, with the standard symbols, obtained from NOAA’s National Climatic Data Center: (a) surface, 0000 UTC 1 Apr 1997, (b) surface, 0600 UTC 1 Apr 1997, (c) surface, 1200 UTC 1 Apr 1997, (d) 500 mb, 1200 UTC 1 Apr 1997.

  • View in gallery

    The 915-MHz radar wind profiler measurements for 1 Apr 1997. Time (UTC) runs from right to left. The wind barbs point into the wind. Long barbs represent winds of 10 kt and short barbs represent winds of 5 kt. The arrow on each plot indicates the profile containing the wind shift associated with the cold front passage: (a) FCN profiler, (b) PAF profiler, and (c) WOC profiler.

  • View in gallery

    (top) Comparison of the east–west wind component from an east–west cross section of RAMS grid 4 and (middle) the lidar east–west wind component obtained from an east–west range–height scan. The RAMS cross section has been cropped to match the lidar’s range. West is to the right. Note that lidar data within 2-km range of the lidar have been deleted because the elevation angle of the beam is too steep to calculate a horizontal wind in this region. Solid contours represent the westerly component flow and dashed contours represent the easterly component flow and contours are every 4 m s−1. The bold line is the zero contour. The two vertical lines on the lidar plot indicate the window from which vertical profiles of the horizontal wind were extracted for the time series shown in Fig. 7. Westerly wind speeds >16 m s−1 have heavy shading and the easterly component flow has light shading. Arrows serve as reminders as to which way the wind was flowing. (bottom) The modeled potential temperature (K) along the same x axis as the u-component plots, but with extended range to the west: (a) 0115, (b) 0330, (c) 0700, (d) 0800, and (e) 0915 UTC 1 Apr 1997.

  • View in gallery

    (Continued)

  • View in gallery

    (Continued)

  • View in gallery

    (Continued)

  • View in gallery

    (Continued)

  • View in gallery

    Time–height series of vertical profiles of the horizontal wind between 0000 and 1230 UTC 1 Apr 1997. Solid contours represent westerly component flow. Dashed contours represent the easterly component flow. Westerly flow greater than 8 m s−1 has been shaded dark gray. All easterly component flow (rotors before 1000 UTC and postfrontal flow after 1000 UTC) has been lightly shaded. The zero contour is boldface. These profiles represent conditions 12–14 km west of the lidar, nearly adjacent to the foothills.

  • View in gallery

    Surface meteorological measurements for 1 Apr 1997 at the PAF site: (a) temperature (°C; 2 m AGL), (b) mixing ratio (g kg−1; 2 m AGL), (c) wind direction (°; 10 m AGL), (d) wind speed (m s−1; 10 m AGL), and (e) pressure (mb; 1 m AGL). The vertical dashed line indicates the time of the front passage at PAF.

  • View in gallery

    Model output from grid 4, 1300 UTC 1 Apr 1997. West is to the right. Vectors indicate horizontal wind (see legend for speeds) while contours represent potential temperature (K) in increments of 1 K.

  • View in gallery

    Terrain map with WKA flight tracks overlayed. The time of the flight tracks is 0400–0642 UTC 1 Apr 1997.

  • View in gallery

    Time series of turbulence intensity on 1 Apr 1997 derived from the eddy dissipation rate calculated from wind sensors onboard the WKA. The categories of light, moderate, and severe apply to commercial aircraft. The shaded regions indicate when the WKA was executing western traverses. (Data courtesy of P. Neilley)

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 49 49 3
PDF Downloads 39 39 1

The Evolution of Lee-Wave–Rotor Activity in the Lee of Pike’s Peak under the Influence of a Cold Frontal Passage: Implications for Aircraft Safety

View More View Less
  • 1 National Oceanic and Atmospheric Administration/Earth System Research Laboratory, Boulder, Colorado
  • | 2 Earth Observing Laboratory, National Center for Atmospheric Research, Boulder, Colorado
© Get Permissions
Full access

Abstract

A lee-wave–rotor system interacting with an approaching cold front in the lee of Pike’s Peak near Colorado Springs, Colorado, on 1 April 1997 is studied observationally and numerically. Dynamical effects associated with the approaching cold front caused the amplification of the evolving lee wave and rotor, creating increasingly more hazardous flight conditions for nearby airports. The rapidly evolving winds measured by a Doppler lidar and 915-MHz wind profilers, and simulated by the Regional Atmospheric Modeling System (RAMS), produced light-to-moderate turbulence for a research aircraft making missed approaches at the Colorado Springs Airport during the wave amplification phase. As the cold front approached the foothills, the lee-wave–rotor system ended abruptly, reducing hazardous flight conditions.

The Doppler lidar’s detailed measurements of the lee-wave–rotor system allowed for an evaluation of RAMS ability to capture these complex wind features. Qualitative and quantitative comparisons between the lidar range–height measurements and model x–z cross sections are presented. In a broad sense, the numerical simulations were successful in the prediction of the prefrontal amplification and the postfrontal decay of the waves as measured by the lidar. RAMS also predicted observed wind reversals above the lee waves, which were indicators of breaking wave instability. At times RAMS performed poorly by over- or underpredicting the wind speeds in the lee wave, as well as the horizontal extent of the lee wave or rotor.

Corresponding author address: Lisa Darby, National Oceanic and Atmospheric Administration, CSD3, 325 Broadway, Boulder, CO 80305. Email: lisa.darby@noaa.gov

Abstract

A lee-wave–rotor system interacting with an approaching cold front in the lee of Pike’s Peak near Colorado Springs, Colorado, on 1 April 1997 is studied observationally and numerically. Dynamical effects associated with the approaching cold front caused the amplification of the evolving lee wave and rotor, creating increasingly more hazardous flight conditions for nearby airports. The rapidly evolving winds measured by a Doppler lidar and 915-MHz wind profilers, and simulated by the Regional Atmospheric Modeling System (RAMS), produced light-to-moderate turbulence for a research aircraft making missed approaches at the Colorado Springs Airport during the wave amplification phase. As the cold front approached the foothills, the lee-wave–rotor system ended abruptly, reducing hazardous flight conditions.

The Doppler lidar’s detailed measurements of the lee-wave–rotor system allowed for an evaluation of RAMS ability to capture these complex wind features. Qualitative and quantitative comparisons between the lidar range–height measurements and model x–z cross sections are presented. In a broad sense, the numerical simulations were successful in the prediction of the prefrontal amplification and the postfrontal decay of the waves as measured by the lidar. RAMS also predicted observed wind reversals above the lee waves, which were indicators of breaking wave instability. At times RAMS performed poorly by over- or underpredicting the wind speeds in the lee wave, as well as the horizontal extent of the lee wave or rotor.

Corresponding author address: Lisa Darby, National Oceanic and Atmospheric Administration, CSD3, 325 Broadway, Boulder, CO 80305. Email: lisa.darby@noaa.gov

1. Introduction

The complexity of the terrain in the Colorado Springs, Colorado, region, which is highlighted by over 2400 m of relief in the lee of the multisummit Pike’s Peak massif [4300 m above mean sea level (MSL)], can cause local winds to be highly variable in time and space. Downslope windstorms are a frequent occurrence and, due to the proximity of the Colorado Springs Airport (COS) to the Rocky Mountains, are potential aircraft hazards (Lester 1994, 1995). Flow reversals, or rotors, associated with strong downslope winds can be responsible for wind shear that produces light to severe turbulence for incoming jet aircraft. In an extreme example, rotor flows associated with a strong lee wave were considered as a possible reason for the downing of United Airlines Flight 585 while on approach to the Colorado Springs Airport in 1991 (National Transportation Safety Board 2001; Doyle and Durran 2002).

To assess how strong winds may affect incoming aircraft in the Colorado Springs region, the Federal Aviation Administration (FAA), in collaboration with the National Oceanic and Atmospheric Administration (NOAA) and the National Center for Atmospheric Research (NCAR), sponsored the Mountain-induced Clear Air Turbulence (MCAT) field experiment in the winter and spring of 1997 (Bedard and Neilley 1998).

A number of lee-wave events were sampled by a NOAA Doppler lidar during MCAT. This paper describes a single case on 1 April 1997, with an emphasis on the 0000–1200 UTC period. During this time, the Doppler lidar observed a variety of phenomena associated with lee waves and the propagation of a cold front through the Colorado Springs region. These phenomena included 1) distinct flow reversals of varying character near the foothills to the southwest and west of the airport, 2) amplification of lee waves, 3) sudden decay of lee waves near the surface, and 4) the shallow surge from a southward-moving cold front. The gustiness, shear, rotors, and flow reversals associated with the lee waves created light-to-moderate turbulence, as indicated by measurements taken on board the Wyoming King Air (WKA) deployed by NCAR scientists.

Using data acquired during MCAT and from high-resolution numerical modeling, our goal in this paper is twofold: 1) combine measurements from a research aircraft, a Doppler lidar, a meteorological tower, and radar wind profilers to elucidate the interaction between synoptic-scale weather patterns (specifically, a cold frontal passage), mesoscale winds, and aircraft turbulence; and 2) qualitatively and quantitatively evaluate mesoscale model results on the observational scale of a lidar.

Section 2 provides background information on lee waves and rotors. In section 3 the 1997 MCAT field project is described. The model description appears in section 4. Section 5 includes weather conditions, observations from the various instruments deployed, and the Regional Atmospheric Modeling System (RAMS) simulation results. Implications for aircraft, as shown by King Air measurements, are presented in section 6. Section 7 summarizes the paper.

2. Background

Downslope wind storms are inherently nonlinear phenomena, and as such are not amenable to study using most lee-wave theory (e.g., Lyra 1943; Queney 1955; Scorer 1955; Queney et al. 1960; Nicholls 1973). Kuettner (1959) used a two-dimensional theory of the hydraulic jump to explain high-reaching turbulent zones and found that for a given wind profile, low-level turbulence associated with waves was partly attributable to the dynamic effect of an upstream inversion and partly to heating in the leeside boundary layer. Doyle and Durran (2002) used a nonhydrostatic model for two-dimensional simulations using simple topography. In simulations without radiative forcing or surface temperature changes, they found that surface–boundary layer drag and lee-wave-induced adverse pressure gradients work together in the formation of low-level rotors. (In this context, adverse pressure gradient refers to a pressure gradient that creates a force in the direction opposite to that of the large-scale mountain-wave-inducing cross-barrier pressure gradient.) With the addition of surface heating they show that vertical rotor scale and turbulence intensity increase in agreement with the theory of Kuettner (1959) and field observations.

Much of what is known about downslope windstorms and their associated turbulence has first been learned via research aircraft flights [see, e.g., Brinkmann (1974); Kuettner (1968); Lester and Fingerhut (1974); Lilly (1978) for measurements taken near Boulder, Colorado, and Förchtgott (1949, 1969; Gerbier and Berenger (1961); Hoinka (1985) for analyses of measurements from various mountainous regions in Europe]. Since aircraft data are obtained in a continuous path through a nonstationary 3D flow field, the volumetric coverage is sparse. Thus, analyzing aircraft data alone can create some problems with regard to interpretation of the evolution of features and correlation with other fixed-location datasets. These problems can be somewhat alleviated when aircraft data are combined with scanning Doppler lidar data, which can provide high time and spatial resolution, and thus resolve the vertical structure and horizontal variability of the winds that occur during downslope wind storms (Neiman 1988), and clear-air turbulence (CAT) associated with a breaking lee wave (Clark et al. 2000). In the present study, the Doppler lidar provided a context for interpreting data from the WKA single-level flight tracks near the foothills of Colorado Springs.

One example of a nonlinear phenomenon associated with downslope windstorms and lee waves is the rotor (Lester and Fingerhut 1974). Today’s general aviation and commercial pilots are aware that the sight of lenticular clouds may indicate turbulence in the area, but are largely ignorant of the entire mountain-wave system (Lester 1994, 1995). Conceptually, the wave flow itself may be quite laminar, but the turbulent zone exists at levels below the wave crests and troughs. Topographical features only a few hundred meters in height can produce rotor and turbulence activity associated with orographically forced waves (Vosper and Mobbs 2002). Often, individual rotors are imbedded within the larger low-level turbulent zone. For our purposes a “rotor” is a turbulent circulation below mountain-wave crests, with horizontal vorticity generally aligned parallel to the mountain range. Other conceptual models of the wave–rotor system exist (e.g., Worthington 2002). Other field experiments have also studied rotors or the low-level turbulent zone associated with lee waves such as, The Sierra Wave Project and Jet Stream Project (Holmboe and Klieforth 1957; Kuettner 1959; Whelan 2000), and the studies of Fingerhut and Lester (1973) and Lester and Fingerhut (1974) within the Colorado Lee-Wave Observational Program (Lilly and Toutenhoofd 1969) near Boulder, Colorado. These studies confirm that severe turbulence is commonly found in the updraft on the upstream side of the rotor and that vertical extent and turbulence intensity increase with increasing wave amplitude.

In recent years, with the increase in computer capability, numerical models have been run with steadily increasing resolution, making forecasting applications of the lee-wave–rotor system feasible. As the time approaches when the model resolution approximates the sampling volume of some lidars, comparisons between high-resolution model results and Doppler lidar measurements will aid in evaluating the model results, as in Clark et al. (1994, 2000), and Fast and Darby (2004). This, of course, assumes that the mesoscale model formulation and initialization are sufficient to accurately represent the observed natural conditions. Previous studies have shown that high-resolution mesoscale models are capable of reproducing many aspects of complex-terrain flows (Poulos and Bossert 1995; Poulos 1996; Poulos et al. 2000; Zhong and Fast 2003; Fast and Darby 2004; Banta et al. 2004). In this paper, a complex wind event was chosen to test the capabilities of a mesoscale model by comparing the results with the high time- and spatial-resolution measurements of a Doppler lidar.

3. The MCAT field program

The MCAT field program was motivated by the National Transportation Safety Board (2001) and the U.S. General Accounting Office (1993) concerns regarding increased aircraft hazards in the vicinity of complex terrain. The FAA’s response to this need was MCAT. The field program was held from February through early April 1997, and focused on quantitative documentation of terrain-induced, low-level atmospheric hazards.

a. Local topography

Figure 1 shows the experiment region, with the Colorado Springs Airport near the center of the map. The runway locations are depicted by the bold lines next to the dot representing the Doppler lidar location at Peterson Air Force Base (PAF). For simplicity, reference to PAF includes Peterson Air Force Base and the Colorado Springs Airport, since they are adjoining properties. The complexity of the terrain surrounding PAF is shown in Fig. 1. The wind flow in the region is greatly influenced by Pike’s Peak to the west of PAF and Cheyenne Mountain to the west-southwest (Fig. 1). Pike’s Peak is 4300 m high and provides over 2400 m of relief above the plains to the east. Cheyenne Mountain is the nearest major topographical obstacle to PAF with a maximum elevation of 2915 m.

The Palmer Lake Divide is to the north of PAF. Although the Palmer Lake Divide is diminutive compared with the Rocky Mountains, this east–west-oriented divide impacts the propagation of cold fronts toward Colorado Springs from the north in two different ways. Colder air from the north is most often diverted eastward around the divide and then approaches PAF from the northeast. Occasionally colder air from the north passes through the gap between the foothills and the divide, and then approaches PAF from the northwest.

b. Instrumentation

1) Lidar

The NOAA/Earth System Research Laboratory (ESRL) Doppler lidar deployed for this experiment was the Transverse Atmospheric Pressure CO2 (TEACO2) system that simultaneously measures range-gated radial wind velocity and backscattered signal intensity with a nominal range of 30 km (Post and Cupp 1990). Doppler lidar is especially well suited for wind flow studies in complex terrain (e.g., Banta et al. 1993, 1997, 1999, 2004; Darby et al. 1999; Flamant et al. 2002). This lidar scans in a vertical plane (maintaining a constant azimuth angle while sweeping in elevation) and quasi-horizontally (maintaining a constant elevation angle while sweeping in azimuth). During MCAT, a 20-min-long sequence of survey scans was repeated at half-hour intervals, leaving additional time for targeted scanning of phenomena of interest to the program. An exception to this scanning strategy occurred when the University of WKA continuously made missed approaches along the north–south runway to obtain turbulence measurements, described below.

2) Wyoming King Air

During the research flights, the WKA would approach the Colorado Springs Airport either from the north or the south, execute a simulated approach and departure, head west to fly along the Front Range, and then turn eastward in preparation for the next simulated approach and departure. Instrumentation on board the WKA included standard meteorological measurements, as well as “state-of-the-aircraft measurements, including vertical acceleration, [which] provided measurements of the response of the aircraft to the turbulence encountered” (Bedard and Neilley 1998). Turbulence intensity was estimated from aircraft measurements by taking the cube root of the eddy dissipation. The values were assigned to categories of light, moderate, and severe turbulence as appropriate for large, commercial transport aircraft (Bedard and Neilley 1998; P. Neilley 1999, personal communication). The typical Doppler lidar scanning strategy during these flight legs was to perform small conical sector scans by varying the azimuth while maintaining a constant elevation angle. The elevation angle was chosen so measurements would be taken along the glide path of the approaching WKA.

3) 915-MHz radar wind profilers

The 915-MHz radar wind profilers were stationed at Fort Carson (FCN), Peterson Air Force Base (PAF), next to the Doppler lidar, and west of Colorado Springs (WOC; close to the foothills; Fig. 1). The three radar wind profilers deployed by ESRL for the experiment operated continuously, had vertical range gates of 100 m, and typically measured winds up to 4 km above ground level (AGL). [See Neff (1994) for examples of profiler data in complex terrain.] Surface meteorological parameters shown in the paper are from a meteorological tower deployed by ESRL, collocated with the PAF profiler. The 10-s samples of wind direction and speed (10 m AGL), temperature and humidity (2 m AGL), and pressure (1 m AGL) are shown.

4. Modeling description

a. Configuration

A simulation of the selected case night (2100 UTC 31 March–1500 UTC 1 April 1997) was completed using the RAMS, version 3b. RAMS is a primitive equation, terrain-following, mesoscale model (Pielke et al. 1992). A 3:1 nesting ratio was used with four grids, so that grid spacing was minimized over the region of interest (Fig. 2). The final grid configuration was chosen on the basis of the relevant scales of the atmospheric phenomena occurring, and available computational power. The phenomena of specific interest were 1) in the horizontal, a ∼5-km rotor circulation observed by Doppler lidar to the lee of Cheyenne Mountain and Pike’s Peak (see Figs. 1 and 3); 2) in the vertical, near the surface, the scale of the rotor and the approaching cold frontal layer of ∼1 km; and 3) at higher altitudes, the vertical wavelength of the terrain-induced gravity wave of ∼4 km. So, using between 6 and 10 grid points to numerically realize these scales, Δx and Δy were equal to 18 km, 6 km, 2 km and 666 m on grids 1–4, respectively. At the lowest model level Δz was set to 100 m, and Δz increased by a factor of 1.07 for each subsequent vertical level up to a maximum Δz of 500 m. With this configuration the model atmosphere was 18 km deep. The outermost model domain, grid 1 (Fig. 2a) was 630 × 684 km in horizontal extent, and subsequent grids were incrementally more focused on the study region. Grid 4 (Fig. 2b) was 50.6 × 26.6 km in horizontal extent.

The model was configured based on our experience numerically modeling complex terrain flows using RAMS and the following are some key setup parameters:

  • Long time step on grids 1, 2, 3, and 4, respectively: 36, 12, 6, and 3 s;
  • Nudging area: the six outermost grid points on grid 1;
  • Nudging altitude: 14 500 m and above;
  • Terrain dataset: 30 arc second (grids 1–3) and 3 arc second (grid 4) United States Geological Survey (USGS) terrain;
  • Short- and longwave radiation parameterizations were updated every 1200 s;
  • The soil and vegetation parameterizations were active with vegetation determined from a 5-min vegetation-type dataset and a soil type of sandy clay;
  • Eddy diffusion was parameterized using standard anisotropic K theory;
  • Horizontal Smagorinsky constant: 0.5 (all grids);
  • Vertical Smagorinsky constant: 0.25 (all grids); and
  • Inverse eddy Prandtl number: 3.
The model was initialized with the analysis field from the Rapid Update Cycle (RUC) model run [available every 3 h at the National Centers for Environmental Prediction (NCEP)]. At the time of this study, NCEP RUC analyses contained 3D fields of the atmospheric variables at 60-km horizontal spacing. RAMS was run using four-dimensional data assimilation (4DDA) where, after initialization, model fields were nudged toward the RUC analyses on the outermost six grid points on grid 1 (inner grids were not subject to nudging). The model configuration allowed for horizontally inhomogeneous initialization while also ensuring that model evolution would be partly constrained by the analyzed atmospheric conditions that existed on 1 April 1997.

b. Limitations

This study, as an intercomparison between a numerical simulation with the specific setup described above, is inherently limited in its ability to discern those numerical components that contribute to model error relative to observations (or for that matter, aid in the characterization of observational error). Specifically, one can envision a future study focused on executing a wide array of sensitivity studies to ultimately discern which model components or combination of components were responsible for error. That study would require error and sensitivity analysis for a matrix of model configuration options that include initial conditions, boundary condition, grid spacing and nesting, parameterization choices, filtering options, and key or customized intraparameterization elements. Such an approach would provide considerably more guidance for the development of predictive tools for rotor-type flows, and aircraft safety applications. This study utilizes the RAMS numerical model as a tool for interpreting otherwise incomplete observations, while simultaneously providing detailed and unique insight into the ability of modern mesoscale numerical models to simulate such flows.

For the simulation described herein, although grid spacing was chosen such that rotor flows could be numerically resolved, it is clear that their detailed interaction with the underlying boundary and their subrotor structures (e.g., Doyle and Durran 2002) were unable to be resolved. Smaller horizontal resolution would improve the representation of smaller-scale orographic features, which in turn, would affect the simulated mountain wave system dynamics. This is one example of many sources of error in such a simulation. Eastman et al. (1998) and Weigel and Rotach (2004) have clearly shown that improved soil moisture representation improves complex terrain forecasts. Doyle et al. (2000) show that the choice of model, and therefore the dynamical core and parameterization options, impact the details and sometimes crucial mesoscale features for a given case study. In addition, for this simulation, as described above, the initialization and boundary conditions were generated from analyses of much coarser resolution than simulated and did not incorporate data from MCAT itself. Improved initial and boundary conditions would also be expected to impact the outcome of such simulations.

5. Observations and modeling results for 1 April 1997

a. Mesoscale setting

The weather maps shown in Fig. 4 characterize the conditions in Colorado for this case study. During the daytime hours of 31 March 1997, a leeside trough deepened over the eastern plains of Colorado. In response to this feature, surface stations showed strong southerly flow on the plains, to the east of the trough, and weaker southwesterly downslope flow along the foothills to the west. Meanwhile, a Pacific cold front moved eastward throughout the day, arriving in Utah by late afternoon (0000 UTC 1 April; Fig. 4a). By 0600 UTC (Fig. 4b), the cold front was progressing across Colorado, with the leeside trough still intact (Fig. 4b). At 1200 UTC the synoptic-scale cold front had crossed over the Rocky Mountains, and had propagated eastward, beyond the study area (Fig. 4c). West-to-southwest winds aloft were associated with an upper-level trough (Fig. 4d).

b. 915-MHz radar wind profiler observations

Vertical profiles of the horizontal wind from the three ESRL 915-MHz radar wind profilers are shown in Fig. 5. The profiles are hourly averages of the horizontal wind, providing a background context for discussion of the more detailed Doppler lidar measurements. The cold front passage, as determined by profiler data (marked with an arrow on each plot), was evident in the hourly data between 1100 and 1500 UTC, depending on the radar site. The hourly averaged profiler winds are plotted at the end of the hour over which the winds were averaged. Therefore, if the wind shift associated with the front passage is plotted at 1300 UTC, then the front passed the profiler sometime between 1200 and 1300 UTC. Since the lowest height of the profiler measurements was 140 m, there could be a time lag between the time the front passed at the surface, but was deep enough to be measured by the wind profiler. Therefore, if the front passed near the end of the hour, the wind shift may not appear until the next hour’s profile. Similarities among the profiler measurements included west-to-southwest flow above ∼4 km MSL, steady southerly to westerly winds throughout the profiles for much of the time before the front passage, and a switch to flow with an easterly component coinciding with the front passage. This easterly postfrontal flow deepened with time at all radar sites.

Differences among the three sites included easterly flow at WOC before the front passage (Fig. 5c) and much stronger prefrontal southwesterly or westerly winds at FCN compared with PAF and WOC. These differences can be explained by the lidar measurements. Lidar constant elevation angle scans revealed much horizontal variability in the winds near the WOC profiler site. These terrain-induced variable winds were confined to a small area near WOC and, from lidar vertical-slice scans (not shown), appeared to be rotors. Nearly horizontal lidar scans (not shown) indicated that the strongest winds reaching the plains had a narrow horizontal extent, with the maximum winds flowing over the FCN site, missing the PAF and WOC profilers. This explained the significant differences in wind speed below 3.5 km MSL between FCN and PAF, placed <20 km apart.

c. Doppler lidar east–west cross sections and model results

Between 0000 and 1000 UTC 1 April 1997, the lidar detected lee waves and associated flow reversals, or rotors. In the east–west Doppler lidar cross sections shown in Fig. 3, a small sampling of the variety of wave structures seen throughout this day are presented. The lidar was located at x = z = 0 km (to the left of the lower-left-hand corner of each plot). Positive radial velocity values (orange and red) indicate flow toward the lidar, and negative radial velocities (green) indicate flow away from the lidar. The Front Range is to the west, beyond the right edge of each plot, and the basic flow was right to left. In each lidar scan, a lee-wave crest is indicated by a “hump” in the ribbon of highest radial velocities (dark brown to red shades). Underneath each wave crest was a rotor, as indicated by green radial velocities.

A convenient way to assess the winds measured by the Doppler lidar throughout a period of several hours is to select a region, or window, of range–height plots at a given azimuth angle (e.g., 270° as in Fig. 3), extract a vertical profile of horizontal winds from this window, and then plot the profiles in a time–height format. Such time–height plots have proven to be an effective way to compare lidar scan data and model results (Darby et al. 2002). To obtain the horizontal winds parallel to the plane of the range–height scan, the lidar radial velocities are converted from polar to Cartesian coordinates and divided by the cosine of the lidar beam’s elevation angle. It is assumed that the cross-beam divergence is negligible and that w sin φ ≪ u sin φ, where w is vertical velocity and φ is the elevation angle. Figure 6a, middle plot, shows a contour plot of the horizontal winds derived using this technique on the lidar data shown in Fig. 3a. East–west cross sections were chosen for analysis because of the abundance of lee waves and rotors detected at this azimuth angle throughout the study period, and for easy comparison to model results.

After assessing all east–west cross sections, a window extending from 12 to 14 km west of the lidar was chosen for further analysis (annotated in Fig. 6a, middle plot, and its geographical location noted on Fig. 1 with light gray shading). This region generally included a portion of the lee wave and a rotor, when present, and almost always had good signal strength from the surface to at least 2 km AGL. After interpolation to a Cartesian grid, at each vertical level, the horizontal winds were averaged over the 12–14-km window for every east–west scan taken during the time period shown in Fig. 7.

Figure 7 presents a continuous and unique record of rotor activity to the west of the lidar from the initial observations through frontal passage and provides the background for a discussion of the lidar measurements interleaved with model simulation results. The three phases of the time period covered in the following discussion are indicated at the bottom of Fig. 7. There was the early evening lee-wave and rotor evolution phase when the features dissipated for a short period (0000–0300 UTC). At 0300 UTC, the period of lee-wave and rotor amplification began, as seen in the increased westerly flow (darkly shaded contours) and increase in depth of the rotors (lightly shaded contours) until ∼0900 UTC. The final phase of the discussion begins at 1000 UTC, between the time when the front propagated through PAF, and when it entered the lidar analysis window at 1030 UTC, as it neared the foothills. After the front passage, the lee wave no longer existed, and the easterly flow behind the front quickly increased in depth, dominating the low-level flow.

1) Early evening lee-wave and rotor evolution (0000–0300 UTC)

In the first 2 h plotted in Fig. 7 (0000–0200 UTC), lidar scans indicated strong leeside flow that descended in altitude over time (darkly shaded contours), overlying a rotor at the surface (lightly shaded contours) that decreased in depth coincident with the descent of the stronger westerly flow. Figure 3a shows a lee-wave crest and rotor representative of this period.

We can evaluate the model performance in reproducing the complex features of the lee-wave–rotor system by comparison of model results to the lidar measurements. In Fig. 6, in each pair of u-component plots, the top plot is a contour plot of the east–west component of the wind from an xz cross section extracted from the innermost model grid. The model cross section is 270 m north of the lidar, the xz cross section closest to the lidar, and it has been cropped to match the sampling region of the lidar. Contour plots of the lidar scans closest in time to the model plots are in the lower panel of each pair. Each pair in Fig. 6 will be discussed in conjunction with the description of the lidar time–height plot (Fig. 7). Below the u-component plots are plots of the modeled potential temperature, θ, from the same xz slice as the u component, but extending many kilometers westward beyond the range of the lidar measurements.

At 0115 UTC (Fig. 6a) the lidar measured westerly leeside flow of at least 12 m s−1. The curvature of the wave crest is seen between 10 and 20 km distance from the lidar. Underneath this wave crest was a flow reversal ∼1 km deep. The model results at this time (top plot, Fig. 6a) did not show the lee-wave structure (note absence of curvature in the westerly flow), although the westerly wind speed maximum of >10 m s−1 was close to the observed 12 m s−1. The model had easterly flow at the surface, but it extended eastward beyond the lidar, rather than being confined to the foothills region, as in the lidar data. The modeled θ contours indicated a weak vertical gradient as opposed to what was seen at later times.

During the time that surface pressure rose slightly at PAF (between 0130 and 0330 UTC; Fig. 8e), the lee-wave and rotor activity disappeared for a short time from the sampling window 12–14-km west of the lidar (Fig. 7). Range–height scans from this period (not shown) indicated that the wave structure significantly decreased during this time at all distances from the lidar, while weak flow reversals a few 100s of m deep remained within the 4–10-km range of the lidar, outside the sampling region for Fig. 7.

At the end of this period of pressure increase, an upper-level flow reversal (light shaded region above 3 km AGL; Figs. 6b and 7) and the extinction of the rotor to the west of the lidar were seen in the lidar data (note the zero contour line near the surface between x = 10–18 km in Fig. 6b). The comparison to the model simulation at 0330 UTC indicates that the upper-level flow reversal seen in the lidar data was simulated (Fig. 6b). However, the horizontal extent of the model’s upper-level flow reversal was not as large as the measured extent at this time and the modeled leeside westerlies were stronger than what was measured by the lidar. The model results also indicated a shallow rotor at the foothills, which was occurring just a short time in advance of the rotor measured by the lidar (as seen in Fig. 7). Given these discrepancies in timing and wind strength between the model results and the lidar measurements, the model did capture the primary features of the wind near this time. An increase in static stability is indicated in the modeled θ field.

2) Lee-wave amplification (0300–1000 UTC)

At 0300 UTC the enhanced leeside westerlies again appeared in the sampling window of the 270° range–height scans (Fig. 7), with the exception of a break in the appearance of the stronger westerlies and rotor coinciding with the brief rise in pressure at PAF at 0445 UTC (Fig. 8e). During the overall surface pressure decrease at PAF that began at 0330 UTC (Fig. 8e), the depth and strength of the leeside westerly flow and surface rotor was variable, but the overall trend, until the cold front reached the foothills at 1030 UTC, was an increase in the depth and strength of the rotor. Figure 3b shows the accelerated leeside westerly flow at its deepest, with a shallow flow reversal underneath. Lester and Fingerhut (1974) found from aircraft data that as lee waves intensify, the turbulent zone beneath the wave crests also intensifies, which, in turn, suggests that in our case study, turbulence increased between 0300 and 0900 UTC.

The increase in the intensity of the wave activity during this time was reflected in the evolution of the cross-barrier pressure gradient, in addition to the overall pressure decrease at the lidar site. Across the Pike’s Peak massif over a distance of 20 km at 1200 m AGL, the modeled Δp (not shown) was −50, −200, and +∼10 Pa for 0600, 1000, and 1300 UTC, respectively. The fourfold increase in cross-barrier pressure deficit between 0600 and 1000 UTC was consistent with the intensification of the wave activity, whereas the reduction and sign reversal of the postfrontal gradient was consistent with the cessation of wave activity (to be discussed later). We note also that during periods of weak rotor activity (0600 UTC), a relatively weak adverse pressure gradient was found under the lee wave in the simulation, and that during stronger rotor activity a stronger adverse pressure gradient developed such as at 1000 UTC. Both of these solutions were consistent with the findings of Doyle and Durran (2002).

Another important factor in the evolution of this lee-wave system was implied by the modeled changes in atmospheric stability with time. The simulation shows that by 0000 UTC 1 April, low stability existed in the lower troposphere due to boundary layer heating, as maximum temperatures during the afternoon of 31 March approached 20°C under clear skies on the plains. Through 0900 UTC 1 April 1997, the modeled atmospheric stability increased dramatically due to both nocturnal, radiatively driven cooling near the surface under clear skies overnight, and, later, due to cold air advection at low levels behind the cold front. According to model cross sections, atmospheric stability in the first 4 km above the elevation of the plains near Colorado Springs increased from <1 K km−1 at 0000 UTC to ∼4 K km−1 at 0900 UTC. After 0900 until 1400 UTC, the atmospheric stability in the same layer remained approximately constant. Denver soundings (not shown) indicated an increase in stability due to substantial overnight cooling between 0000 and 1200 UTC.

A model–lidar comparison for 0700 UTC is shown in Fig. 6c. The lidar data show a lee wave, including a shallow surface rotor west of the sampling window. The maximum wind speed in both the lidar plot and model plot were in good agreement at 16 m s−1, although the model’s horizontal extent of the high winds ranged farther eastward than measured. The model results also included a rotor near the foothills, but it extended into the higher terrain, which was not supported by the observations, a likely consequence of inadequate detail or representation of terrain, or physiographic features (such as zo, surface roughness length) in the model simulation combined with the slight distance between the lidar measurements and the model xz cross section. Modeled θ continued to indicate an increase in stability and increased wave activity upstream of the lidar.

Evolution of the wave structure was significant over the following hour. At 0800 UTC (Fig. 6d), west of the lidar there were two wave crests, as seen in Fig. 3c. Based on the lidar plot of wind speed in Fig. 6d we note that the wave structure evolved such that the secondary descent of stronger (8 m s−1) lee-wave flow in the lee of the massif nearly intersected the ground surface approximately 12 km west of the lidar. The contours seen in Fig. 6d explain the structure of the time–height series shown in Fig. 7, at 0800 UTC. The 12–14-km sampling window was between rotors, explaining the downward curvature of the westerly component contours seen at 0800 UTC in Fig. 7. The model results (top of Fig. 6d), indicated a single-crest lee wave with a surface rotor.

In the final lidar–model comparison seen in Fig. 6e (0915 UTC), the simulation results were encouraging. The two-crested lee-wave structure was seen in both the lidar and model results, with the modeled lee wave lower in altitude, stronger in speed, and differing in wavelength than what was measured. The modeled wave crests were west of the measured wave crests. None of these modeled differences is unexpected given the larger modeled wind speed and the dependence of the lee-wave structure on the horizontal component of the wind. The large horizontal extent of the underlying easterly flow as measured by the lidar was captured by the model, although the modeled easterly flow covered a larger east–west area. An elevated flow reversal, smaller in horizontal extent than observed, was found in the model results near x = 21 km and z = 4 km. This flow reversal was a further indication of the increasing wave intensity and the potential for nonlinear behavior (e.g., wave breaking). The deep rotor underlying the lee-wave flow of 14 m s−1 (Fig. 7) indicated that strong vertical wind shear was present at this time. The modeled increase in stability over the course of the prefrontal amplification of the lee waves was indicated by comparing the θ plot in Fig. 6a with that of Fig. 6e.

To quantitatively assess the lidar–RAMS comparisons shown in Fig. 6, a root-mean-square error (rmse) analysis was performed by gridding the Doppler lidar data to the RAMS x and z coordinates. Only data to the west of the lidar were used. The results for the entire window to the west of the lidar are shown in the second column of Table 1. The rmses ranged from 4.5 to 8.9 m s−1 for the entire window analyzed. Since the vertical resolution of the model was much greater in the lower 1 km of the domain (designed to contain the rotor circulation), we split the window of interest into two portions, and performed the analysis again for data below 1 km and above 1 km to see if the rmses improved where the model had increased vertical resolution. For data below 1 km, the results did improve for all times except 0115 UTC, with the 0700 and 0800 UTC comparisons showing significant improvement. The model’s relative success in simulating the rotor at 0700 and 0800 UTC led to the two lowest rmses of the analysis. If the modeled rotor was too shallow and the upper-level strong westerlies dipped too close to the surface, then the rmses were quite large (as at 0915 UTC).

The third column contains rmses for grid points above 1 km. Except for 0115 UTC, the rmses were larger for this window than for the entire window or the lowest 1 km only. This upper region had lower resolution and contained the accelerated leeside flow and upper-level flow reversals, features not necessarily modeled well.

3) Cold front passage (1000–1230 UTC)

The passage of the cold front at PAF was seen in the surface meteorological data (Fig. 8). At 0945 UTC the surface pressure began to rise (Fig. 8e), a drop in wind speed accompanied the wind shift (Figs. 8c,d), the temperature dropped [superimposed upon the nocturnal cooling (Fig. 8a)], and the mixing ratio increased (Fig. 8b). Note that the winds behind the front were from the northeast (Fig. 8c). Synoptically, the cold front had a northeast–southwest orientation, moving through Colorado from the northwest (Fig. 4b), but the front approached PAF first from the northeast, a result of the local terrain features, as discussed earlier. Fluctuations in the atmospheric variables (particularly in the moisture and wind direction time series) from before the front passage until sunrise (just after 1300 UTC) imply that the frontal passage at the surface was not a distinct, abrupt passage.

The dual-crested wave with two rotors and elevated flow reversal seen in Figs. 2c and 6e, indicate how complex the winds were shortly before the cold front passage. Shortly after the scan shown in Fig. 2c was completed, a significant change in the lidar-measured winds near the foothills occurred as the cold front progressed from PAF to the foothills. The lee-wave–rotor system abruptly shut down. The surface easterly flow behind the cold front (light shaded region at the surface after 1030 UTC; Fig. 7), quickly increased in depth and speed as the cold front propagated toward the foothills (as also seen in the profiler data). Previous studies have also shown a reduction or elimination of lee-wave activity with the passage of a cold front (Neiman et al. 2001; Darby et al. 1999, 2000).

Easterly flow above 1 km AGL, extending to at least 4 km AGL, was detected for a brief period, between the abrupt cessation of the lee-wave and rotor activity, but before the easterly flow at the surface, behind the cold front, reached the lidar analysis window (Fig. 7). This easterly flow aloft also appeared in the WOC profiler data (Fig. 5c) at 1000 UTC, but not at FCN or PAF (Figs. 5a,b), showing how localized this upper-level flow feature was. The cause of this flow feature is unknown.

With the pressure rise behind the cold front (Fig. 8e), the distinct pressure minimum to the lee of the Rocky Mountains no longer existed and the acceleration of the southwesterly winds to the surface subsided, consistent with the modeled reversal in cross-barrier pressure gradient. As cold air filtered into the region, potentially warmer air in the lee-wave flow was unable to penetrate to the surface. Lidar data show a well-established postfrontal northeasterly flow underlying the now weaker southwesterly flow aloft. Rotor activity at low levels was eliminated as the synoptic conditions associated with the cold frontal passage dominated the study region. The winds eventually shifted from northeasterly to easterly, upslope conditions developed, and snow began to fall by 1800 UTC.

Figure 9 (1300 UTC 1 April), represents model results well after the cold front passage, and shows that the wave flow subsided substantially in the model as in the lidar measurements. Both upstream and downstream stability were significantly lower than at 0915 UTC, thereby supporting lower-frequency wave activity. Associated with the cold frontal passage, an easterly flow of 3–6 m s−1 existed in a layer from the ground up to 1300 m, similar to that seen in the lidar data at 1230 UTC (Fig. 7).

6. Implications for aircraft

The WKA flight on 1 April 1997 occurred from 0353 to 0642 UTC, during the wave amplification period (Fig. 7). Flight tracks for this day are shown in Fig. 10. Simulated approaches began when the ferry portion of the flight ended at ∼0500 UTC 1 April 1997. The WKA approached the airport from the north, for the simulated approaches, descending as low as 15 m AGL (Bedard and Neilley 1998). During the simulated departure, the plane climbed in altitude, headed west, and eventually turned northward to fly along the Front Range before heading east again for the next simulated approach and departure.

The calculated turbulence intensities from the flight on 1 April 1997 are shown in Fig. 11. Turbulence during the simulated approaches and departures along the north–south runway mainly fell in the category of light turbulence, with a few instances of moderate turbulence (nonshaded times in Fig. 11). The measurements from the western traverses (shaded times in Fig. 11), which were about 10 km west of the airport at 550–765 m AGL, show many instances of moderate turbulence. During the sixth shaded period of the WKA measurements, marked by an arrow in Fig. 11, the lidar performed range–height scans at 210°, 240°, 270°, and 310° azimuth, slicing through the region of the WKA’s western traverses. The 270° range–height scan from this series is shown in Fig. 3b. The star symbol in Fig. 3b indicates the approximate region the King Air flew through during its western traverse and its encounter with moderate turbulence. At the time of this scan, the WKA was 10.7 km west of the lidar, and 4 s away from flying through the plane of the range–height scan. The cause of the moderate turbulence was clearly evident in the lidar measurements. The aircraft was skimming the edge of a rotor. Wind shear values near the region of the star were 1.09 × 10−2 s−1 for ±50 m in the vertical and 1.03 × 10−3 s−1 for ±337.5 m in the horizontal.

In lee-wave conditions, severe turbulence is likely to occur in the updraft side of the rotor, and moderate turbulence is likely to occur in the downdraft side (Lester and Fingerhut 1974). In this case, the King Air was closer to the downdraft side of the wave when it encountered moderate turbulence, consistent with these previous observations. Although this study is focused on the interaction of a lee-wave–rotor system with an incoming cold front, these rare coincident observations by lidar and aircraft confirm the value of Doppler lidar for aircraft safety applications. In combination with numerical forecasts of sufficient resolution, these two tools would provide a basis for aircraft guidance through predicted and observed instances of potential danger to aircraft.

7. Summary

Measurements from various types of instruments and platforms have been combined with a mesoscale model simulation in order to establish the life cycle of a lee-wave–rotor event in Colorado Springs, Colorado. The Doppler lidar measurements revealed lee waves and rotors along the lee of the Rocky Mountains, in the Colorado Springs area. Early in the study period a slight rise in pressure on the plains, as measured at PAF, coincided with a weakening of the mountain and rotor directly west of PAF, followed by several hours of falling pressure on the plains. The falling pressure coincided with the amplification of the lee wave west of PAF, along with the increase in depth of the accompanying rotor. A research aircraft obtaining turbulence measurements during the amplification period encountered moderate turbulence when skimming the edge of the rotor. After ∼7 h of amplification, the demise of the lee-wave–rotor system occurred with the passage of a cold front.

For ease of comparison with model results, lidar east–west cross sections were used for comparisons with RAMS. The modeled east–west component of the wind was compared to lidar measurements for five different times during the evolution of the lee waves and rotors. A qualitative look at the model results indicated that at times the model captured many of the features measured by the lidar, but at times the wind speeds or the spatial extent of certain features were over- or underpredicted. The upper-level flow reversal seen at 0330 UTC and the dual-wave structure seen at 0915 UTC were two examples of complex features captured fairly well by the model.

To quantify the comparisons, we calculated an rmse for the model simulations, based on gridded lidar data. The rmses were rather large, but there were improvements in the errors for four out of the five comparisons when we only considered the lowest 1 km of the model grid, which had much higher vertical resolution than the rest of the grid. There were times when the model adequately reproduced features of the lee-wave–rotor system, but the horizontal location or the depth of the modeled feature differed from the measured feature enough that the rmses were large and not necessarily a good indicator of the model’s success. The modeled pressure trends were consistent with the observed pressure trends.

The simulation and measurements shown in this paper illustrate that over a short time span shear can increase and flow direction can shift, creating difficult landing conditions at PAF or other nearby airports. The Pike’s Peak massif and associated terrain features (such as Cheyenne Mountain) were significant sources of flow variability depending on the evolving upstream conditions. The rapidly varying conditions during even this moderately strong lee-wave case present a significant challenge not only to forecasters, but also for airport personnel trying to anticipate the safest possible routes for incoming aircraft. In particular, traveling or amplifying lee waves can create moderate turbulence. Given that this was a case of moderate upstream flow (up to 16 m s−1), and that at 4 km MSL flows are sometimes double this speed, one may reasonably conclude that aircraft in this area (particularly near the mountains) would occasionally be subjected to more intense turbulence. This conclusion is supported by simulations of the flow conditions in this region associated with the 1991 aircraft accident near the Colorado Springs Airport. The model simulation of that event indicates the eastward propagation of rotors from the Pike’s Peak massif, and particularly Cheyenne Mountain (National Transportation Safety Board 2001; Doyle and Durran 2002).

Acknowledgments

The authors would like to acknowledge several people who made the experiment a success. We thank Dr. Al Bedard for obtaining the funding to bring the NOAA/ESRL Doppler lidar to Colorado Springs for MCAT. We thank Mr. Jim Howell, Dr. David Levinson, Dr. Robert Banta, and Mr. Hutch Johnson for their tireless efforts in the field to obtain the Doppler lidar data. The aircraft turbulence data were kindly provided by Dr. Peter Neilley. We thank Dr. Clark King and Ms. Catherine Russell for the installation and the data acquisition of the radar wind profiler data and the PAF tower data. The work of Ms. Darby was partially supported by the U.S. Department of Energy, under the auspices of the Environmental Meteorology Program of the Office of Biological and Environmental Research, through Interagency Agreements DE-AI03-99ER62842 and DE-AI02-04ER63860.

REFERENCES

  • Banta, R. M., , L. D. Olivier, , and D. H. Levinson, 1993: Evolution of the Monterey Bay sea-breeze layer as observed by pulsed Doppler lidar. J. Atmos. Sci., 50 , 39593982.

    • Search Google Scholar
    • Export Citation
  • Banta, R. M., and Coauthors, 1997: Nocturnal cleansing flows in a tributary valley. Atmos. Environ., 31 , 21472162.

  • Banta, R. M., , L. D. Olivier, , P. Kaufman, , D. H. Levinson, , and C. J. Zhu, 1999: Wind flow patterns in the Grand Canyon as revealed by Doppler lidar. J. Appl. Meteor., 38 , 10691083.

    • Search Google Scholar
    • Export Citation
  • Banta, R. M., , L. S. Darby, , J. D. Fast, , J. O. Pinto, , C. D. Whiteman, , W. J. Shaw, , and B. D. Orr, 2004: Nocturnal low-level jet in a mountain basin complex. Part I: Evolution and implications to other flow features. J. Appl. Meteor., 43 , 13481365.

    • Search Google Scholar
    • Export Citation
  • Bedard Jr., A. J., , and P. Neilley, 1998: Project MCAT (Mountain Induced Clear Air Turbulence): Background, goals, instrumentation and methodologies. Preprints, Eighth Conf. on Mountain Meteorology, Flagstaff, AZ, Amer. Meteor. Soc., 163–171.

  • Brinkmann, W. A. R., 1974: Strong downslope winds at Boulder, Colorado. Mon. Wea. Rev., 102 , 592602.

  • 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
  • Clark, T. L., , W. D. Hall, , R. M. Kerr, , D. Middleton, , L. Radke, , F. M. Ralph, , P. J. Neiman, , and D. Levinson, 2000: Origins of aircraft-damaging clear-air turbulence during the 9 December 1992 Colorado downslope windstorm: Numerical simulations and comparison with observations. J. Atmos. Sci., 57 , 11051131.

    • Search Google Scholar
    • Export Citation
  • 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
  • Darby, L. S., , A. Gohm, , L. B. Nance, , S. Gabersek, , R. M. Banta, , and S. Sandberg, 2000: Foehn flow in the Austrian Alps interrupted by a cold front passage, Part I. Preprints, Ninth Conf. on Mountain Meteorology, Aspen, CO, Amer. Meteor. Soc., 79–82.

  • Darby, L. S., , R. M. Banta, , and R. A. Pielke Sr., 2002: Comparisons between mesoscale model terrain sensitivity studies and Doppler lidar measurements of the sea breeze at Monterey Bay. Mon. Wea. Rev., 130 , 28132838.

    • Search Google Scholar
    • Export Citation
  • Doyle, J. D., , and D. R. Durran, 2002: The dynamics of mountain-wave induced rotors. J. Atmos. Sci., 59 , 186201.

  • Doyle, J. D., and Coauthors, 2000: An intercomparison of model-predicted wave breaking for the 11 January 1972 Boulder windstorm. Mon. Wea. Rev., 128 , 901914.

    • Search Google Scholar
    • Export Citation
  • Eastman, J. L., , R. A. Pielke, , and D. J. McDonald, 1998: Calibration of soil moisture for large-eddy simulations over the FIFE area. J. Atmos. Sci., 55 , 11311140.

    • Search Google Scholar
    • Export Citation
  • Fast, J. D., , and L. S. Darby, 2004: An evaluation of mesoscale model predictions of down-valley and canyon flows and their consequences using Doppler lidar measurements during VTMX 2000. J. Appl. Meteor., 43 , 420436.

    • Search Google Scholar
    • Export Citation
  • Fingerhut, W. A., , and P. F. Lester, 1973: Lower turbulent zones associated with mountain lee waves. Department of Meteorology Paper No. 30, San Jose State University, 123 pp.

  • Flamant, C., and Coauthors, 2002: Gap flow in an Alpine Valley during a shallow south foehn event: Observations, numerical simulations, and hydraulic analog. Quart. J. Roy. Meteor. Soc., 128 , 11711210.

    • Search Google Scholar
    • Export Citation
  • Förchtgott, J., 1949: Rotor streaming in the lee of mountain ridges. Bull. Meteor. Czech., 4 , 1416.

  • Förchtgott, J., 1969: Evidence for mountain-sized lee eddies. Weather, 24 , 255260.

  • Gerbier, N., , and M. Berenger, 1961: Experimental studies of lee waves in the French Alps. Quart. J. Roy. Meteor. Soc., 87 , 1323.

  • Hoinka, K. P., 1985: Observation of the airflow over the Alps during a foehn event. Quart. J. Roy. Meteor. Soc., 111 , 199224.

  • Holmboe, J., , and H. Klieforth, 1957: Investigation of mountain lee waves and the air flow over the Sierra Nevada. Final Report, Dept. of Meteorology, UCLA, Contract No. AF19(604)-728, 290 pp.

  • Kuettner, J., 1959: The rotor flow in the lee of mountains. Geophysical Research Division Res. Note 6, Air Force Cambridge Research Center, 20 pp.

  • Kuettner, J., 1968: Lee waves in the Colorado Rockies. Weatherwise, 21 , 180195.

  • Lester, P. F., 1994: Turbulence: A New Perspective for Pilots. Jeppesen Sanderson Training Products, 212 pp.

  • Lester, P. F., 1995: Aviation Weather. Jeppesen Sanderson Training Products, 335 pp.

  • Lester, P. F., , and W. A. Fingerhut, 1974: Lower turbulent zones associated with mountain lee waves. J. Appl. Meteor., 13 , 5461.

  • Lilly, D. K., 1978: A severe downslope windstorm and aircraft turbulence event induced by a mountain wave. J. Atmos. Sci., 35 , 5977.

  • Lilly, D. K., , and W. Toutenhoofd, 1969: The Colorado Lee Wave Program. . Clear Air Turbulence and Its Detection, Y. Pao and A. Goldburg, Eds., Plenum Press, 232–245.

    • Search Google Scholar
    • Export Citation
  • Lyra, G., 1943: Theorie der stationaeren Leewellenstromung in freier Atmosphaere. Z. Angew. Math. Mech, 23 , 128.

  • National Transportation Safety Board, 2001: Aircraft accident Report: Uncontrolled descent and collision with terrain, United Airlines flight 585, Boeing 737-200, N999UA, 4 miles south of Colorado Springs Municipal Airport, Colorado Springs, Colorado, March 3, 1991. NTSB Rep. AAR-01/01, NTIS Rep. PB2001-910401, 214 pp. [Available online at http://www.ntsb.gov/publictn/2001/AAR0101.pdf.].

  • Neff, W. D., 1994: Mesoscale air quality studies with meteorological remote sensing systems. Int. J. Remote Sens., 15 , 393426.

  • Neiman, P. J., 1988: Doppler lidar observations of a downslope wind storm. Mon. Wea. Rev., 116 , 22652275.

  • Neiman, P. J., , F. M. Ralph, , R. L. Weber, , T. Utall, , L. B. Nance, , and D. H. Levinson, 2001: Observations of nonclassical frontal propagation and frontally forced gravity waves adjacent to steep topography. Mon. Wea. Rev., 129 , 26332659.

    • Search Google Scholar
    • Export Citation
  • Nicholls, J. M., 1973: The airflow over the mountains: Research 1958–1972. WMO Tech. Note 127, World Meteorological Organization, Geneva, Switzerland, 73 pp.

  • Pielke, R. A., and Coauthors, 1992: A comprehensive meteorological modeling system—RAMS. Meteor. Atmos. Phys., 49 , 6991.

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

  • Poulos, G. S., 1996: The interaction of katabatic winds and mountain waves. Ph.D. dissertation, Colorado State University, 297 pp.

  • Poulos, G. S., , and J. E. Bossert, 1995: An observational and prognostic numerical investigation of complex terrain dispersion. J. Appl. Meteor., 34 , 650669.

    • Search Google Scholar
    • Export Citation
  • Poulos, G. S., , J. E. Bossert, , T. B. McKee, , and R. A. Pielke, 2000: The interaction of katabatic flow and mountain waves. Part I: Observations and idealized simulations. J. Atmos. Sci., 57 , 19191936.

    • Search Google Scholar
    • Export Citation
  • Queney, P., 1955: Rotor phenomena in the lee of mountains. Tellus, 7 , 367371.

  • Queney, P., , G. Corby, , N. Gerbier, , H. Koshmieder, , and J. Zierep, 1960: The airflow over the mountains. WMO Tech. Note 34, World Meteorological Organization, Geneva, Switzerland, 135 pp.

  • Scorer, R. S., 1955: The theory of airflow over mountains—IV. Separation of flow from the surface. Quart. J. Roy. Meteor. Soc., 81 , 340350.

    • Search Google Scholar
    • Export Citation
  • U.S. General Accounting Office, 1993: Aviation safety: FAA can better prepare general aviation pilots for mountain flying risks. Rep. RCED-94-15, 72 pp. [Available online at http://archive.gao.gov/t2pbat4/150580.pdf.].

  • Vosper, S., , and S. Mobbs, 2002: Numerical simulations of lee-wave rotors. Preprints, 10th Conf. on Mountain Meteorology, Park City, UT, Amer. Meteor. Soc., CD-ROM, P15.3.

  • Weigel, A. P., , and M. W. Rotach, 2004: Flow structure and turbulence characteristics of the daytime atmosphere in a steep and narrow Alpine valley. Quart. J. Roy. Meteor. Soc., 130 , 26052627.

    • Search Google Scholar
    • Export Citation
  • Whelan, R. F., 2000: Exploring the Monster. Mountain Lee Waves: The Aerial Elevator. Wind Canyon Books, 169 pp.

  • Worthington, R. M., 2002: Mountain waves launched by convective activity within the boundary layer above mountains. Bound.-Layer Meteor., 103 , 469491.

    • Search Google Scholar
    • Export Citation
  • Zhong, S., , and J. D. Fast, 2003: An evaluation of the MM5, RAMS, and Meso-Eta models at subkilometer resolution using field campaign data in the Salt Lake Valley. Mon. Wea. Rev., 131 , 13011322.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Map of experiment region, centered on PAF, and the Colorado Springs Airport runways (boldface lines to the left of PAF). Terrain contours are every 100 m from 1100 to 2300 m MSL and every 200 m between 2300 and 4300 m MSL. The Doppler lidar was stationed at PAF. The 915-MHz radar wind profilers were stationed at PAF, WOC, and FCN. The light gray dashed box indicates the approximate location of the RAMS grid 4. The small gray box 12–14-km west of the lidar indicates the region over which the vertical profiles of lidar-measured horizontal winds shown in Fig. 7 were averaged.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 2.
Fig. 2.

Topography of the RAMS model domain. (a) Topography for the outermost grid (grid 1). Terrain contours are every 200 m from 1200 to 3400 m MSL. (b) Topography for the innermost grid (grid 4). Terrain contours are every 100 m from 1800 to 4000 m MSL.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 3.
Fig. 3.

Doppler lidar range–height radial velocity scans pointing toward the west along the 270° radial. West is to the right. Tick marks are in 2-km increments. The color bar at the bottom of each plot represents radial velocities in m s−1. Black arrows indicate the westerly component flow (positive radial velocities). Red arrows indicate the easterly component flow (negative radial velocities). The lidar is located at x = z = 0 and times are UTC. (a) Lee wave and rotor at 0120 UTC. (b) Deep lee wave and rotor at 0610 UTC. The red star indicates approximate position of the research aircraft as it traversed along the foothills at this time. (c) Lee wave and rotors at 0938 UTC, shortly before the cold front passage.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 4.
Fig. 4.

Weather maps, with the standard symbols, obtained from NOAA’s National Climatic Data Center: (a) surface, 0000 UTC 1 Apr 1997, (b) surface, 0600 UTC 1 Apr 1997, (c) surface, 1200 UTC 1 Apr 1997, (d) 500 mb, 1200 UTC 1 Apr 1997.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 5.
Fig. 5.

The 915-MHz radar wind profiler measurements for 1 Apr 1997. Time (UTC) runs from right to left. The wind barbs point into the wind. Long barbs represent winds of 10 kt and short barbs represent winds of 5 kt. The arrow on each plot indicates the profile containing the wind shift associated with the cold front passage: (a) FCN profiler, (b) PAF profiler, and (c) WOC profiler.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 6.
Fig. 6.

(top) Comparison of the east–west wind component from an east–west cross section of RAMS grid 4 and (middle) the lidar east–west wind component obtained from an east–west range–height scan. The RAMS cross section has been cropped to match the lidar’s range. West is to the right. Note that lidar data within 2-km range of the lidar have been deleted because the elevation angle of the beam is too steep to calculate a horizontal wind in this region. Solid contours represent the westerly component flow and dashed contours represent the easterly component flow and contours are every 4 m s−1. The bold line is the zero contour. The two vertical lines on the lidar plot indicate the window from which vertical profiles of the horizontal wind were extracted for the time series shown in Fig. 7. Westerly wind speeds >16 m s−1 have heavy shading and the easterly component flow has light shading. Arrows serve as reminders as to which way the wind was flowing. (bottom) The modeled potential temperature (K) along the same x axis as the u-component plots, but with extended range to the west: (a) 0115, (b) 0330, (c) 0700, (d) 0800, and (e) 0915 UTC 1 Apr 1997.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 6.
Fig. 6.

(Continued)

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 6.
Fig. 6.

(Continued)

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 6.
Fig. 6.

(Continued)

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 6.
Fig. 6.

(Continued)

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 7.
Fig. 7.

Time–height series of vertical profiles of the horizontal wind between 0000 and 1230 UTC 1 Apr 1997. Solid contours represent westerly component flow. Dashed contours represent the easterly component flow. Westerly flow greater than 8 m s−1 has been shaded dark gray. All easterly component flow (rotors before 1000 UTC and postfrontal flow after 1000 UTC) has been lightly shaded. The zero contour is boldface. These profiles represent conditions 12–14 km west of the lidar, nearly adjacent to the foothills.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 8.
Fig. 8.

Surface meteorological measurements for 1 Apr 1997 at the PAF site: (a) temperature (°C; 2 m AGL), (b) mixing ratio (g kg−1; 2 m AGL), (c) wind direction (°; 10 m AGL), (d) wind speed (m s−1; 10 m AGL), and (e) pressure (mb; 1 m AGL). The vertical dashed line indicates the time of the front passage at PAF.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 9.
Fig. 9.

Model output from grid 4, 1300 UTC 1 Apr 1997. West is to the right. Vectors indicate horizontal wind (see legend for speeds) while contours represent potential temperature (K) in increments of 1 K.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 10.
Fig. 10.

Terrain map with WKA flight tracks overlayed. The time of the flight tracks is 0400–0642 UTC 1 Apr 1997.

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Fig. 11.
Fig. 11.

Time series of turbulence intensity on 1 Apr 1997 derived from the eddy dissipation rate calculated from wind sensors onboard the WKA. The categories of light, moderate, and severe apply to commercial aircraft. The shaded regions indicate when the WKA was executing western traverses. (Data courtesy of P. Neilley)

Citation: Monthly Weather Review 134, 10; 10.1175/MWR3208.1

Table 1.

The rmse of RAMS results compared with gridded Doppler lidar data. Only grid points to the west of the lidar were used. Refer to Fig. 6 to see the data used in the comparison. The first column indicates the time of the comparison. The second column is the rmse for the entire window west of the lidar. The last two columns separate the window at 1 km AGL, with the third column showing the results from below 1 km AGL and the fourth column showing the results from above 1 km AGL.

Table 1.
Save