The ESRL website will be unavailable for 24 hours starting Friday, March 6th at 5:00pm MT due to building maintenance.
 

8.1

Project MCAT (Mountain induced Clear Air Turbulence):

Background, Goals, Instrumentation, and Methodologies

Alfred J. Bedard Jr
National Oceanic and Atmospheric Administration
Environmental Technology Laboratory
R/E/ET4, 325 Broadway, Boulder, Colorado 80303-3328
email: Alfred.J.Bedard@noaa.gov

Peter Neilley
National Center for Atmospheric Research
Research Applications Program, Boulder, CO 80307-3000


1. INTRODUCTION

Project MCAT97, a Federal Aviation Administration (FAA) sponsored program designed to document aircraft hazards in the Colorado Springs area, resulted from National Transportation Safety Board recommendations following a United Boeing 737 accident at Colorado Springs, Colorado. Laboratories from the National Center for Atmospheric Research (Research Applications Program) and the National Oceanic and Atmospheric Administration (Forecast Systems Laboratory and the Environmental Technology Laboratory) combined to design and execute a field experiment at Colorado Springs to study and define mountain induced aircraft hazards in the region. We describe the history and the goals of the experiment. The instruments and platforms participating in the experiment included a Doppler lidar, 3 wind profilers, an instrumented aircraft, surface weather stations, a mobile CLASS, and an infrasonic observatory. We also review the methods used to execute the experiment, which included daily short term and long term weather briefings. The field study continued from February through early April 1997 and successfully measured a range of phenomena. The purpose of this paper is to provide an overview of the experiment, providing background for several other papers appearing in these proceedings.

2. BACKGROUND

Both the National Transportation Safety Board and the General Accounting Office (GAO 1993) have made recommendations concerning steps to reduce mountain flying risks. These risks are evidenced by the occurrence of more than ten major accidents and incidents during the last ten years in the vicinity of mountains. Further, the General Accounting Office found the general aviation accident rate was about 40% higher for mountainous western states than for all other continental states during FY 1992. In response, laboratories of the National Oceanic and Atmospheric Administration (NOAA) and the National Center for Atmospheric Research (NCAR) have combined resources to perform a field experiment addressing terrain-induced flight hazards. For the first time we were able to bring together state-of-the-art remote sensing and in situ instrumentation to study low-level aircraft turbulence hazards. We have obtained a dataset documenting many different mountain-generated flow phenomena that pose a threat to aviation. In particular, the Doppler lidar measurements during MCAT97 provide some of the clearest documentation of rotors, flow reversals, and mountain-wave-induced flow features. These data represent an important resource.

3. THE MCAT 97 EXPERIMENT: HISTORICAL PERSPECTIVE

The Mountain-Induced Aeronautical Hazards (MAH) program originated after the crash United Boeing 737 in Colorado Springs on 3 March 1991. The National Transportation Safety Board listed the cause of this crash as unknown and made the following two recommendations:

1. Develop and implement a meteorological program to observe, document, and analyze potential meteorological aircraft hazards in the area of Colorado Springs, Colorado, with a focus on the approach and departure paths of the Colorado Springs Municipal Airport.

2. Develop a broader meteorological aircraft hazard program to include other airports in or near mountainous terrain, based on the results obtained in the Colorado Springs area.

These recommendations initially resulted in FAA support to the participating laboratories to provide a training document for helping pilots to use visual clues during mountain flying. Subsequently, we created the handbook, Hazardous Mountain Winds and Their Visual Indicators, which has been published as an FAA Advisory Circular (Carney et al. 1996). In collaboration with the FAA, we designed a plan for the MAH (Mountain Aviation Hazards) program that addresses all aspects of mountain flying safety, with goals ranging from improving pilot training to creating hazard detection and warning systems.

The first phase of this program included the design and implementation of a pilot experiment to study potential low-level, mountain-induced hazards in the Colorado Springs area. A major field experiment performed from February through early April 1997 in the Colorado Springs, Colorado, area focused on defining terrain-induced, low-level hazards to aircraft. The field program brought to bear a variety of remote sensors, as well as surface in-situ and aircraft measurement systems. The experiment, funded by the FAA, was implemented by the collaborative efforts of laboratories within NOAA and NCAR. Specifically, the laboratories were the Research Applications Program (RAP) of NCAR, the Environmental Technology Laboratory (ETL) of NOAA, and the Forecast Systems Laboratory (FSL) of NOAA. These laboratories also contributed resources to enhance the observing campaign and help insure success. The field program created valuable resources that can be used in the following ways:

  • To increase our understanding and definition of atmospheric hazards.

  • To evaluate improved prediction techniques.

  • To compare with numerical simulations.

  • To develop detection and monitoring systems for low-level, terrain-induced aircraft hazards.

The contributing laboratories all have a long history in combining both analytical and field studies of broad aspects of mountain flows, particularly focusing on how terrain-modified flows affect flight. The combined talents and experiences of our laboratories constitute a unique resource to be applied to the challenge of improving air safety near complex terrain.

4. OVERVIEW OF THE GOALS OF THE COLORADO SPRINGS EXPERIMENT IN THE CONTEXT OF THE MAH PROGRAM

The primary objectives of the Colorado Springs experiment MCAT97 were as follows:

1. To obtain datasets from a variety of systems:

  • Sufficient for defining low-level, terrain-induced flight hazards in the Colorado Springs area.

  • Suitable for the development, testing, and tuning of turbulence detection algorithms for the terminal area. Such algorithms will be used to provide real-time mountain-induced hazard warnings to controllers and pilots. For this purpose, the terminal area is defined as the 3- to 5-nm-long, 1-nm-wide approach/departure corridors extending off both ends of runways 35R and 35L, plus the runways themselves.
2. To apply and test the utility of the synoptic climatology and windstorm damage surveys, as well as output from weather prediction models, to help guide the experiment and for improving turbulence predictability.

3. To evaluate remote sensor capabilities, scan strategies, and site details for mountain-flow hazard detection and definition.

These goals are integral to the MAH program and constitute an important first step toward increasing mountain flight safety. Concentrating on defining low-level flight hazards, the MCAT97 experiment will provide an initial hazard definition, data resources for use with detection algorithms, a test of the methodologies used to design the experiment, and a first evaluation of remote sensor strengths and weaknesses in complex environments.

The deliverables from project MCAT97 will include the following:

1. A report containing recommendations and addressing the following specific areas:

  • algorithms with a detection and display system adjusted for Colorado Springs

  • improved forecast methodologies

  • definition of the nature of aeronautical turbulence hazards in the Colorado Springs area.

This report will identify datasets created during the experiment and summarize key flow regimes for Colorado Springs. The datasets covered will include:

  • datasets valuable for simulations of aircraft response

  • experimental datasets valuable for comparisons with fine-scale numerical models and laboratory-scale models

  • datasets connecting atmospheric flow fields with pilot reports (PIREPS).
2. A demonstration of a working model of a Colorado Springs mountain aircraft hazard detection and display system. Materials and presentations of a working model of a Colorado Springs hazard detection system were given to the FAA during 1997.

5. Expected Hazards

A variety of potential hazards to aviation involving wind flows over and around mountains exist in many areas and around a number of airports (Carney et al. 1996). In planning for MCAT97, we had the goal of detection of four possible mountain-induced phenomena that may impact low-altitude aircraft operations in the Colorado Springs area:

1. Wind surge over a mountain ridge giving rise to a traveling horizontal-axis vortex.

2. Zones of horizontal shear and mesoscale eddies downstream from Pikes Peak giving rise to shear-generated turbulence and possible small-scale vertical-axis vortices.

3. Strong mountain wave with turbulence associated with rotors and intense vertical shear.

4. Mechanically generated and thermally enhanced turbulence generated by strong flow over the mountainous terrain southwest through northwest of the Colorado Springs airport.

Phenomenon 1 has been illustrated in the laboratory in water-tank simulations. These simulations show that a surge of wind over an obstacle will produce a wind-shear layer that will "roll up" into a horizontal-axis vortex and move downstream. An aircraft encounter with such a vortex could be dangerous. Vertical-axis vortices arising from phenomenon 2 are also proposed as a possible explanation for a number of the so-called "freak wind" incidents observed at the ground in the Colorado Springs area and are also a potential hazard to aircraft. Phenomenon 3 (strong mountain wave) is a well-known source of turbulence at all altitudes; prior to MCAT97, no attempt to observe and document the particular manifestation of strong mountain waves in the Colorado Springs area had been made. Phenomenon 4 was expected to be present any time low-level flow from the west quadrant was anticipated and, therefore, is considered to be a more common hazard for low-level aircraft operations in the vicinity of the Colorado Springs airport. All of these hazards occur within the context of a few fairly well-defined, broader-scale weather situations that bring strong winds aloft with a westerly component over the Colorado Springs area. These larger-scale weather patterns, described briefly below, will be discussed in detail by Brown et al. (1998).

  • Southwest chinook: Characterized by strong southwest flow aloft and gusty and sometimes quite strong southwest surface winds, particularly in the southern part of the Colorado Springs area.
  • Northwest chinook: Characterized by strong west to northwest winds aloft and strong surface winds in the northern part of the Colorado Springs area, associated with either warm-air advection (WAA) or cold-air advection (CAA) below 25,000 ft altitude.

Another weather situation can bring strong terrain-enhanced surface winds to the area, but is believed not to be a primary source of the mountain-induced hazardous phenomena (except possibly for item 4) that was the focus of MCAT97. This situation, referred to as "bora", is described as

  • Bora: Characterized by a sudden shift of the surface wind into the north or northeast, followed by falling temperatures and a period of strong winds from the north quadrant, and usually associated with passage of a rapidly southward-moving cold front down the High Plains and over the Palmer Divide north of Colorado Springs. Although these conditions produce strong shear and turbulence, we currently believe that the chinook conditions produce somewhat greater turbulence and thus pose a greater threat to aviation. This is a hypothesis we need to test.

Fourteen chinook cases were measured by the lidar during the experiment. These cases often had strong winds above 100 m AGL, but were not so strong at the surface. However, they still represented a threat to aviation. During MCAT97, there were relatively few southwest chinooks, no exceptionally strong northwest chinooks, and many boras. Overall, the occurrence of hazard-prone weather situations during MCAT97 is believed to be somewhat less than usual for the February through March period. arise. A brief overview of each of the sensor systems deployed for MCAT97 follows.

6. DESCRIPTION OF THE EQUIPMENT DEPLOYED

In this section, a brief description of the suite of field sensors deployed for the MCAT97 field program is presented. The set of sensors deployed consisted of a broad range of meteorological instruments, each of which provided a unique view of the meteorological environment associated with the set of expected hazards. Since there is considerable diversity in the hazards that could occur, and because there is little experience in collecting data on some of the expected phenomena (e.g., rotors), a spectrum of sensors was required to ensure a high degree of observational success. The sensors chosen have been proven in other field experiments to provide the type of measurements that would be needed to form a nearly complete picture of the evolution of the hazards and the environments and mechanisms by which they

6.1 Doppler wind profilers

Doppler wind profilers provide measurements of the three components of wind in a vertical column above the ground. In addition, profilers equipped with radio acoustic sounding systems (RASS) also provide measurements of the virtual temperature (moisture-adjusted temperature) above the site. The depth and vertical resolution of the wind and temperature measurements from a profiler are dependent upon the frequency of the emitted radiation. In the case of MCAT97, so-called "boundary-layer" profilers operating at 915 MHz were deployed. These profilers typically provide measurements up to a height of about 2 km at a vertical resolution of about 75 m. Traditionally, the temporal resolution of the measurements is between 15 and 30 min, although techniques under development show promise of extracting high-quality wind measurements from the profilers at frequencies of about one minute. Similar research has also shown promise of extracting direct measurements of turbulence from the profilers. The profilers deployed during MCAT97 recorded profiles every minute to allow high sample rate wind and turbulence metrics to be computed in post-analysis of the data.

The principal advantage of wind profilers is providing relatively low-cost and continuous measurements of the vertical structure of the wind vector above the ground. A disadvantage is only providing a one-dimensional view of the wind field as a function of time. Assumptions or other corroborating information must be used in order to determine a more complete three-dimensional picture of the wind fields. In addition, wind profiler measurements are sometimes contaminated by various sources of clutter, which often requires careful attention to data quality during analysis.

Three wind profilers were deployed during MCAT97. Two of the profilers were deployed upstream (west) of the Colorado Springs airport by about 10 km, one located west of the northern and one located west of the southern approach/departure corridors. The third profiler was deployed near center field and a few hundred meters east of runway 35R/27L. Figure 1 shows a geographic layout of the profiler sites along with the other instruments deployed. The western two profilers were designed to measure the wind flow in the lee of the mountains, but still upstream of the primary approach and departure corridors of the airport. The third profiler was designed to measure the conditions as close to the airport as practical. The profilers operated unattended and continuously throughout the program, except for occasional equipment repair and maintenance.

6.2 Doppler lidar

A Doppler lidar is similar to a traditional weather radar [e.g., a Weather Surveillance Radar 1988 Doppler (WSR-88D) Next Generation Weather Radar (NEXRAD) or a Terminal Doppler Weather Radar (TDWR)] in that it can remotely measure the radial component of wind in a three-dimensional volume surrounding the instrument. In both cases, the measurement is accomplished by sending out a pulse of radiation and then looking at the frequency shift of the returned signal. The degree of shift is directly related to the component of wind velocity toward or away from the lidar. The time between the transmitted pulse of radiation and the returned signal indicates the distance from the radar or lidar to the particular wind measurement. By scanning the direction of the transmitted radiation through various elevations above the horizon and various compass azimuths, a three-dimensional picture of the radial component of wind can be formed.

A lidar differs from a traditional weather radar in that it transmits a narrow beam of laser energy instead of a focused cone of radio frequency radiation. Therefore, a lidar provides a relatively narrow view of the wind and thus wind measurements with considerably greater spatial resolution. In addition, the use of a laser eliminates many of the clutter problems associated with imperfectly focused radiation beams inherent in traditional weather radars. The largest advantage of a lidar over a traditional weather radar in its application to MCAT97 arises from the fact that the primary reflectors of the laser's energy are dust rather than precipitation particles. This enables the lidar to provide wind measurements in the clear-air environments that characterize the low levels of the atmosphere during strong flows off the terrain. In these cases there often are insufficient meteorological reflectors for traditional weather radars to provide suitable velocity signals. Thus a lidar may be the only instrument that can reliably sample a complete volume of radial velocities during strong terrain-induced wind events. The primary disadvantages of a lidar are that it cannot operate unattended, is limited to operation in clear-air conditions, and only measures one component of the wind field. However, given that the lidar is the only practical device that can reliably sample the clear-air wind fields over a terminal volume, its deployment for MCAT97 was considered essential. Other techniques exist for inferring the other components of the wind field.

The lidar was collocated with the third Doppler wind profiler, just east of the midpoint of runway 35R/17L. The lidar operated at nearly all times with potential for developing significant terrain-induced flows. Four general types of scans were used by the lidar: (i) horizontal sweeps at various elevation angles; (ii) vertical sweeps at specified azimuths; (iii) narrow, rapid, horizontal sweeps across the approach and departure corridors extending from the ends of runway 35R/17L; and (iv) fixed azimuth/elevation stares. The choice of which scan strategy to use during a particular event was determined by the project scientists and lidar operators based on the other instruments currently operating and the general meteorological situation. A combination of the scanning strategies was utilized in most cases observed.

6.3 Instrumented aircraft

The KingAir research aircraft, instrumented by the University of Wyoming, flew approximately 40 h in support of MCAT97 objectives. It is a twin turboprop aircraft equipped with special meteorological equipment to measure winds, temperature, pressure, moisture, cloud microphysics, and other meteorological parameters at the high rate of 16 Hz (or about 5-m horizontal resolution at typical flight speeds). The aircraft was flown during those times when the conditions were particularly conducive for strong terrain-induced wind flows in the vicinity of Colorado Springs. The aircraft concentrated on sampling the meteorological environment along the approach and departure corridors of runways 35R/17L by performing repeated simulated approaches and departures out to 5-8 km from the runway ends. Returns for subsequent approaches and departures usually were done at constant altitude (typically, 8000 ft MSL) 5 km west of the airport. These return legs were designed to sample the environment above the airport. Approaches were performed down to the lowest elevation deemed safe by the pilot, which was typically about 50 ft above the runway. In a typical flight, a series of about ten approach/ departure patterns were flown. The aircraft was in close communication with the lidar operators during the flights, with corroborating rapid sector scans across the flight paths typically performed by the lidar during aircraft operations.

The primary motivation for deploying the research aircraft was that it was the only instrument that could provide in situ, objective measurements of the actual wind flow and associated hazards along the approach and departure corridors. In this sense, the aircraft measurements could be considered the "truth" to which observations from other sensors would be compared. The primary disadvantage was that the aircraft operation was expensive and thus was limited to operation during only a few select events throughout the course of the program. In addition, aircraft measurements provide only a one-dimensional set of observations along the flight path, making it difficult to assemble three- or four-dimensional views of small-scale coherent structures (e.g., rotors) encountered.

In addition to the standard meteorological measurements, the aircraft indirectly measured atmospheric turbulence in several ways. First, the state-of-the-aircraft measurements, including vertical acceleration, provided measurements of the response of the aircraft to the turbulence encountered. More objective methods of extracting state-of-the-atmosphere (e.g., aircraft-independent) turbulence statistics were applied during the postprocessing of the data. The resultant statistics are necessary to determine the response that any aircraft would have to the sampled turbulence fields.

The output of the turbulence postprocessing of the aircraft data was the cube root of the eddy dissipation rate (EDR). The eddy dissipation rate is one of the more commonly used measures of atmospheric turbulence. It is a state-of-the-atmosphere measurement, as opposed to a state-of-the-aircraft measurement. Its units in the meter-kilogram-second system are (m2 s-3)1/3, which will be referred to as just "MKS." For large, commercial transport category aircraft, EDR values in excess of about 0.25 MKS are usually coincident with the onset of moderate turbulence, while values in excess of about 0.50 MKS are associated with severe turbulence. However, these thresholds are only approximate and depend on many factors such as aircraft type, weight, airspeed, and subjective interpretation. The EDR was computed on a one-second basis in a variety of ways and using each of the orthogonal components of the wind. It remains a topic of current scientific research as to the best or most appropriate method to compute EDR. Since no generally accepted "best" method has emerged, a large number of different EDR metrics were computed.

6.4 Surface weather stations

Seven research-grade surface weather stations were deployed for MCAT97. Surface weather stations provide inexpensive, accurate, and continuous measurements of wind, temperature, moisture, pressure, and, in some cases, precipitation and radiation rates at relatively high frequency (typically between 2 and 10 s per sample). One weather station was deployed at each of the Doppler wind profiler sites, with the remaining stations deployed at (a) Pikes Peak (14,100 ft), (b) Cheyenne Mountain (~9000 ft), (c) the southern end of the Rampart Range (~8100 ft), and (d) 5.5 km southwest of the airport center field (~6000 ft) as shown in Figure 1. These weather stations provided valuable, continuous measurements of the meteorological conditions near the surface. The mountain-top stations provide measurements of the conditions aloft that are critical to understanding the environment under which various terrain-induced hazards form.

Figure 1. Geographic layout of the various sensors deployed during MCAT97.

In addition to the deployed weather stations, the data from a number of existing stations were collected. These stations included a seven-station network at the U.S. Air Force Academy (AFF), the Phase II Low-Level Wind Shear Alert System (LLWAS) network at the airport, the high-rate observation data from the Colorado Springs Airport Surface Observing System (ASOS) station, and a network of stations operated by the city of Colorado Springs. The quality and characteristics of the data collected from these existing weather stations varied considerably. The approximate locations of all of the weather station data collected during MCAT97 are shown in Fig. 1.

6.5 CLASS soundings

A mobile (van-based) CLASS system was deployed during MCAT97 and took 24 soundings during the course of the program. These soundings consisted of horizontal wind, temperature, and humidity as a function of pressure at 10-s increments (about 50-m vertical increments) from the ground until signal was lost (typically about 40,000 ft). The primary objective of the soundings was to sample the upstream, relatively less perturbed, tropospheric-deep environment associated with particular terrain-induced wind events. Numerous scientific investigations have shown that the character of mountain-induced flow heavily depends on the details of the ambient wind and temperature profiles upstream of the terrain. Therefore, these soundings were taken in order to provide a basis for understanding the observed downstream, terrain-induced flows from a general scientific perspective. Further, these soundings are essential in order to pursue fine-scale numerical modeling of the observed events.

6.6 PIREPS

Two types of PIREPS were collected during MCAT97. First, the standard suite of PIREPS transcribed and broadcast on the National Weather Service (NWS) real-time weather data transmissions was archived. In addition, an aviation-band radio receiver was set up on the airfield and tuned to the two active tower Air Traffic Control (ATC) frequencies for Colorado Springs. All voice traffic broadcast on these channels were recorded on the audio channel of a VCR, while the current time was recorded simultaneously on the video channel. In practice, few of the aircraft-originated broadcasts were picked up by the receiver, and thus PIREPS of significant weather in the terminal area were generally recorded by this mechanism only if the tower controller repeated the report. Recordings were made only during times of interest to the program.

6.7 Infrasonic observatory

The infrasonic observatory is a research instrument that has shown promise in detecting aviation-significant turbulence and other meteorological phenomena by listening to the inaudible, low-frequency noise generated from these events (Bedard, 1978). An infrasonic observatory was deployed during a portion of MCAT97 collocated with the Doppler lidar. Preliminary results and a review appear in these proceedings (Bedard and Craig, 1998).

7. GENERAL EXPERIMENT CONDUCT AND OPERATIONS

Operations of MCAT97 commenced on 3 February and continued through 5 April 1997. Most equipment was deployed and operational by the end of the first week. Since no strong wind events occurred during this first week, there was no loss of data significant to the program during this project spin-up time. The profilers and anemometers collected data continuously from deployment through the end of operations. The remaining equipment was operated as conditions dictated. A day-by-day general summary of MCAT97 operations will appear in a report to the FAA.

Operational decisions were made on a daily basis during a routine morning meeting of the principals involved in the program. The meeting typically involved an informal discussion of the program's most current operational event and other logistical issues, followed by a formal review of the current and forecasted relevant weather conditions. This formal review was followed by a discussion about the operation of the aircraft, lidar, and mobile sounding in the subsequent 24-48 hour period. Decisions to operate equipment were based upon several factors including the current and forecast conditions, the operational status of the equipment, and the availability of the necessary personnel to operate the equipment. In general, the lidar operated during all periods in which a reasonable chance of some significant terrain-induced wind flow was forecast, whereas the aircraft would more typically wait for favorable conditions to develop before deploying. The mobile sounding system typically launched balloons every 6 h during periods in which the aircraft was operating. Air traffic control communications were generally recorded during all periods that the lidar operated.

8. HAZARDS MEASURED

This section reviews hazard-producing conditions for the Colorado Springs area. The atmospheric hazards include:

  • rotors and flow reversals
  • mountain waves and severe downslope winds
  • terrain-modulated synoptic-scale hazards
  • low-level stochastic (random) turbulence
  • interplay between diurnal circulations and synoptic-scale flows
  • mesoscale eddies
  • bora flows.

The following subsections include a review of some of these hazards and an overview of results.

8.1 Rotors and flow reversals

Although considerable evidence exists for the presence of vortices near mountains (e.g., Bedard 1990), there has been a lack of direct measurements documenting not only the strength and size but also the conditions under which they are created. There are many potential physical processes that can create vortices, ranging from direct interaction of flow with terrain (in this case, shear layers can roll up into concentrated eddies) to larger-scale rotations and flow reversals associated with lee waves.

During MCAT97, the Doppler lidar proved to be invaluable for measuring and visualizing smaller-scale atmospheric features, and a preliminary assessment indicates that many of these features represent potential hazards to flight operations. Analyses to date have identified a number of situations involving rotors and flow reversals, including:

  • a low-level flow reversal associated with a northwest chinook
  • a larger-scale rotor beneath the crest of a lee wave
  • a small-scale rotor at low level associated with a northwest chinook
  • a small-scale rotor at upper levels associated with a northwest chinook
  • a small-scale flow reversal at quite low altitude associated with a southwest chinook.

We anticipate that subsequent analyses will identify many additional such events. If aircraft were to fly through these situations, relative airspeed changes greater than 40 m s-1 are possible. The aircraft response will depend upon the details of the interaction (especially heading, altitude, and the wind-speed strength). Oliver and Poulos (1998) describe a case study of a system that produced a flow reversal.

Real-time monitoring of the Doppler lidar display provided some indications of the existence of translating rotors, but more analysis is required to clearly establish their size and strength. It is clear that during northwest chinook conditions, coherent surges of flow moved from near mountain-top locations to the airport (a distance of about 20 km). The arrivals of surges at the airport were accompanied by increases in surface winds. At times, surges of higher-velocity flow would be coupled with a zero-velocity flow (a potential sign of rotation if the mean flow is subtracted). More analysis of these situations is needed.

The Doppler lidar measurements during MCAT97 have provided some of the clearest documentation of rotors and flow reversals in the vicinity of mountains. Hence, these data represent an important resource.

8.2 Low-level stochastic (random) turbulence

Stochastic turbulence refers to the seemingly random motion of the atmosphere. Turbulence is always present in the atmosphere, but fortunately a majority of the time it is weak enough not to significantly affect aircraft. It is distinguished from coherent terrain-induced hazards (e.g., waves and rotors) in that no simplified set of equations can explicitly describe its four-dimensional evolution. It is further differentiated from coherent phenomena in that it is not characterized by a predominant length scale, but rather spans a wide spectrum of scales typically from meters to kilometers. This wide span of scales means turbulence usually has a variety of impacts on an aircraft ranging from performance changes (e.g., airspeed fluctuations) to reduced controllability. As these impacts become strong enough, turbulence may become a significant hazard to safe flight.

Turbulence encounters by aircraft at sufficient altitude are usually not a threat to airframe integrity. The risk of crash due to turbulence at cruise altitudes is less since the altitude changes caused by strong turbulence are often relatively small. Nonetheless, turbulence encounters are one of the leading causes of passenger and crew injury. On the other hand, at low altitudes, strong turbulence can become a major factor in the ability of an aircraft to safely execute an approach or departure. This is particularly true for smaller, general aviation aircraft, which are more strongly affected by turbulence. Landing gear failure and runway excursions are some of the common accidents that are brought about (in part) by turbulence.

There are many mechanisms by which the normal, benign level of turbulence can evolve to become a hazard to aviation. In most cases, the mechanics of how the turbulence arises is not related to subsequent impacts that it may have on an aircraft. However, understanding these mechanisms, and, in particular, the environments that are conducive to them, is an important element in understanding, detecting, and forecasting mountain-induced flow hazards in general, because many of the turbulence generation mechanisms are directly or indirectly associated with terrain-induced flows. The more common of these mechanisms by which flow over terrain can give rise to stochastic turbulence are described below.

a. Mechanical turbulence. The generation of "mechanical" turbulence is the result of the flow being forced to follow complex trajectories as it flows over and around the terrain. The stresses placed on the flow in these circumstances are often unstable and quickly degenerate into turbulence that is carried downstream by the mean flow. The intensity of the turbulence generated by this mechanical process is more strongly dependent on the strength of the flow over the terrain and the character of the terrain, and less so on other factors. Severe and extreme mechanical turbulence has been scientifically documented at low levels in the lee of some mountains. The intensity of the turbulence generated in this fashion decreases downstream of the mountain as the turbulent kinetic energy is dissipated in the form of heat. Mechanical turbulence is generally found at and below mountain tops.

b. Mountain wave-induced turbulence. Mountain waves can generate strong turbulence near mountains through at least three processes. First, as the mountain waves amplify, the associated surface wind speeds increase, often dramatically, leading to an enhanced generation of mechanical turbulence near the surface. Second, the waves themselves can amplify to the point of breaking, a process that gives rise to strong turbulence in a manner analogous to ocean waves breaking on a beach. The third mechanism is associated with enhanced vertical wind shear along the wave that can lead to the amplification and subsequent breaking at smaller scales through a process known as Kelvin-Helmholtz instability. The latter two mechanisms are the most common cause of terrain-induced turbulence at high altitudes above mountains. Severe to extreme turbulence encounters at high altitudes are often reported during the course of the winter season when large-amplitude mountain waves are present. Strong low-level turbulence associated with mountain-wave-induced downslope windstorms may primarily be a consequence of enhanced mechanical turbulence, although more research, including analysis of the MCAT97 dataset, is needed to confirm this hypothesis.

Stochastic turbulence strong enough to impact aviation is often present when other strong mountain-induced flow structures are present. For example, it is probably uncommon to have terrain-induced rotors present without a noticeable degree of stochastic turbulence. However, the presence of turbulence does not imply the presence of the other structures. The presence of multiple phenomena (e.g., rotors plus turbulence) generally enhances the likelihood that any particular terrain-induced flow will be hazardous. For example, flight simulator studies have shown that the ability of a pilot to successfully recover from a strong low-level wind shear is greatly reduced when stochastic turbulence is added to the wind shear flow even though the contributing phenomena (wind shear or turbulence) were not a hazard individually. Results such as these imply that the onset of aviation hazardous conditions may not be determined solely by the individual phenomena, but must be considered in the broader environment in which they are commonly embedded.

8.3 Mesoscale eddies

Mesoscale eddies may be an important source of low-level flow phenomena hazardous to aircraft operations in the Colorado Springs area. We were able to confirm the existence of at least one characteristic eddy circulation during the field phase of MCAT97. This feature occurs in northwest chinook situations, probably most consistently in warm-air advection (WAA) northwest chinooks. Called the "Fort Carson eddy," it may be defined by the simultaneous occurrence of surface winds from the northwest quadrant at AFF or Colorado Springs of 15 knots or more and winds with an easterly component at Fort Carson. It is believed that this eddy is a vortex downstream of the Pikes Peak massif, and it is possible that when the Fort Carson eddy is observed that there may be an oppositely rotating eddy south of Fort Carson. During MCAT97, the Fort Carson eddy was observed on a few occasions during northwest chinooks, but was weak and not persistent. We noticed more cases of Fort Carson eddies in the few weeks just prior to the experiment. The Doppler lidar detected many small-scale intense vortices along the lateral shear associated with the Fort Carson eddy on several occasions. These situations will require careful analysis.

The existence of other terrain-anchored eddies in the Colorado Springs area will require more detailed analysis of surface data collected during the experiment.

8.4 Bora flows

Bora flows at Colorado Springs are characterized by sudden shifts of the surface wind into the north or northeast, typically accompanied by falling temperatures and a period of strong winds from the north. These surges of cold air are made more complex by their interactions with the Palmer Divide ridge just north of Colorado Springs. These surges are relatively shallow and are driven by density differences between the warmer ambient air and the colder bora air. Such gravity currents move rapidly toward regions of lower altitude like flowing water.

The initial flow tended associated with a bora event in the Colorado Springs area tended to move over the eastern region of the Palmer Divide ridge and approach the airport from the northeast, with the subsequent flows backing to the north as the cold air deepened and surged directly over the Palmer Divide to the north. One hazard is represented by the wind-shift line, which can result in sudden wind direction changes and wind vector changes in excess of 20 m s-1. Another hazardous aspect of these flows is the potential for strong vertical wind shears. This is especially true when the upper-level flows are associated with northwest or southwest chinooks. In such situations, the shallow, cold, underlying, northerly bora air will transition rapidly to the chinook flow, accompanied by large wind shear and turbulence. This will be important for takeoff and landing at Colorado Springs. Paradoxically, once the cold air following the leading edge of the bora is entered, flight can be smoother because the stable air tends to suppress turbulence. Another complicating factor is that at times cold air surges from both the north and west interact over the Colorado Springs area.

9. CONCLUDING REMARKS

The MCAT97 data sets represent an important resource for improving our understanding of mountain related aircraft hazards. Moreover, the operation of a variety of remote sensors can be evaluated and their value for detection and warning assessed. The presentations at this conference are only a first step to apply the results ( Brown et al., 1998, Bedard and Craig, 1998, Oliver and Poulos, 1998, Poulos and Oliver, 1998, and Levinson and Banta,1998).

10. ACKNOWLEDGMENTS

The successful completion of MCAT97 would not have been possible without generous support from numerous individuals and groups in the Colorado Springs area. A large number of individuals from NOAA and NCAR contributed in many ways to the experiment and future publications will acknowledge their efforts by name.

We are most grateful for the support of the Federal Aviation Administration, and particularly thank Dr. Aston McLaughlin.

11. REFERENCES

Bedard, A. J., Jr., 1978: Infrasound originating near mountainous regions in Colorado. J. Appl. Meteorol. 17, 1014-1022.

Bedard, A. J., Jr., 1990: A review of the evidence for strong, small-scale vortical flows during downslope windstorms. J. Wind Eng. Ind. Aerodyn. 36, 97-106.

Bedard, A.J., Jr., and R. Craig, 1998: Infrasonic detection of atmospheric turbulence in the vicinity of mountains. Proc. Eighth Conf. on Mountain Meteorology, 3-7 August 1998, Flagstaff, AZ (This Volume).

Brown, J.M., E.J.Szoke, and D. Levinson, 1998: Synoptic weather patterns associated with strong winds and low-level turbulence at Colorado Springs. Talk to be presented at the Eighth Conf. on Mountain Meteorology, 3-7 August 1998, Flagstaff, AZ.

Carney, T. Q., A. J. Bedard, Jr., J. M. Brown, J. McGinley, T. Lindholm, and M. J. Kraus, 1996: Hazardous Mountain Winds and Their Visual Indicators. Handbook. Dept. of Commerce, NOAA, Boulder, Colorado, 80 pp. [Recently republished as Federal Aviation Administration Advisory Circular AC00-57.]

General Accounting Office (GAO), 1993: Aviation Safety, FAA can better prepare general aviation pilots for mountain flying risks. Report GAO/RCED-94-15. U.S. General Accounting Office, Gaithersburg, Maryland.

Levinson, D., and R. M. Banta, 1998: Effects of Northwesterly downslope winds interacting with a Bora on low-level aircraft turbulence during MCAT97. Talk to be presented at the Eighth Conf. on Mountain Meteorology, 3-7 August 1998, Flagstaff, AZ.

Oliver, L.D., and G.S. Poulos, 1998: Frontal passage, mountain waves, and flow reversals in the vicinity of the Colorado Springs, CO airport. Proc. Eighth Conf. on Mountain Meteorology, 3-7 August 1998, Flagstaff, AZ (This Volume).

Poulos, G.S., and L.D. Oliver, 1998: Mountain wave/frontal dynamics in the lee of Pike's Peak. Proc. Eighth Conf. on Mountain Meteorology, 3-7 August 1998, Flagstaff, AZ (This Volume).