Hong Kong is building a new airport at Chek Lap Kok (CLK) located on partly reclaimed land at the base of Lantau Island, a rugged mountain with a maximum elevation of nearly 1000 m. (see Fig. 1 ). Roughly (US) $20 billion is being spent on construction of the airport and associated infrastructure projects. Due to the close proximity of the New Airport to Lantau Island, the New Airport will be affected by significant terrain-induced windshear and turbulence when certain meteorological conditions exist. In order to enhance flight safety and operational efficiency at CLK during such conditions, the Operational Windshear Warning System (OWWS) program was created. Weather Information Technologies, Inc. (WITI) together with the National Center for Atmospheric Research (NCAR), the University of Wyoming (UW) and the Hong Kong University of Science and Technology (HKUST), under the sponsorship of the Hong Kong Observatory, participated in the OWWS program. The program was managed by B. Donaldson (WITI), R. Wagoner and W. Mahoney (NCAR) and J. Chen (HKUST). The major objective of the program was to develop and implement a Windshear and Turbulence Warning System (WTWS) for the new airport. On 17 July 1997, after 44-months of research and development, the WTWS system was accepted by the Hong Kong Government. This $16 million project included basic and applied research on wind flow over Hong Kong's rugged terrain, a scientific field study, warning system concept and feasibility studies, system design, construction, implementation and training.
The WTWS provides real-time hazardous weather information to air traffic controllers and pilots in order to enhance flight safety in the terminal area and predictions of hazardous weather to air traffic managers and aviation forecasters for supporting strategic decision making. The WTWS is unique in that it is the first system, worldwide, to provide real-time alerts of terrain-induced turbulence and convective and terrain-induced windshear. The WTWS also provides terrain-induced turbulence prediction, airport surface wind prediction, and numerical weather prediction guidance.
The WTWS generates graphical and text output designed for easy interpretation by pilots, air traffic controllers, air traffic managers, and aviation forecasters. The system interfaces with other airport systems allowing its products to reach a broad user community including airport authority staff and airline offices.
Prior to the development of the WTWS, several basic studies were conducted by Hong Kong researchers aimed at gaining insight into the meteorological conditions near CLK. These studies included analyses of routine weather observations, special observing programs and modeling studies. Analyzed variables included windspeed and direction, temperature, clouds, visibility, general rainfall, thunderstorms and fog. Methods used to conduct the studies included investigative flights by light aircraft, water tank experiments and wind tunnel experiments. These and other studies indicated that significant turbulence and windshear occur at CLK during specific meteorological conditions. This result eventually led to a decision by the Hong Kong Government to create the OWWS program, which was designed to investigate the detailed flow environment near CLK and, based on the scientific results, build and implement an operational windshear and turbulence warning/forecasting system.
One of the first major tasks of NCAR, primarily the Research Applications Program (RAP) and the Mesoscale and Microscale Meteorology (MMM) divisions, was to perform a detailed meteorological review of historical data and perform analyses to better understand the flow conditions near CLK. This component of the research program was primarily conducted by B. Foote, P. Neilley, T. Keller,T. Clark, H.M. Hsu, and C. Wade. The First Meteorological Report included: a) a review of scientific theory on flow around complex terrain; b) an analysis and identification of conditions which could cause Terrain-Induced Windshear and Turbulence (TIWT) near CLK; c) numerical experiments aimed at gaining additional insight on conditions that produce TIWT; and d) an estimate of the timing and location of significant TIWT at the new airport.
Results of the scientific review indicated that the primary parameters important for determining the nature of the flow include windspeed and direction, stability, Froude and Richardson numbers, and the presence of critical levels. These parameters were further studied by performing small scale modeling simulations using the Clark model. The preliminary model results indicated that both mechanical and gravity-wave processes contribute to the windshear and turbulence conditions and confirmed that the intensity and location of the turbulence is sensitive to windspeed and direction, stability and critical-level heights.
An assessment was made to identify the phenomena that would likely affect aircraft operations at CLK. Crosswinds, longitudinal windshear, large wind changes, turbulence, updrafts, and downdrafts were identified as the principal aircraft-affecting phenomena that would be expected during TIWT episodes. At that point in the study it was unclear which condition, if any, would dominate.
Knowledge gained from the first meteorological study was used, along with other information, to design a scientific field experiment to understand the fine-scale flow in the vicinity of CLK. The scientific field program began in March 1994 and concluded in September 1995. The basic objectives of the program were: a) to quantify the frequency and severity of turbulence and windshear at CLK; b) to more clearly define the meteorological conditions under which significant TIWT occurs; c) to validate the airflow predictions made by the small-scale model; d) to determine the elements necessary to develop the WTWS; and e) to collect verification data for the WTWS.
The major observational platforms operating in Hong Kong included the NCAR King Air research aircraft, a scanning Doppler lidar, an integrated sounding system, a wind profiler, and a network of surface weather stations. Digital flight data from Cathay Pacific Airways' 747-400s operating into Hong Kong were also collected.
The scientific field study was performed by scientists at RAP (Neilley, D. Blanchard, Cornman and Keller), MMM (Clark, J. Coen, and Hsu) and UW (A. Rodi). The study indicated that moderate to severe terrain-induced turbulence occurs in the vicinity of CLK and that terrain-induced turbulence is found more frequently than terrain-induced windshear. The terrain-induced turbulence is found almost exclusively in a well-defined region in the wake of the terrain surrounding CLK. Ambient windspeed is a first-order parameter in determining the magnitude of the terrain-induced turbulence, whereas ambient wind direction governs location. Other parameters (stability, critical-level) affect the flow response but are secondary. Both mechanical and gravity-wave processes appear to be important in the dynamics and the net response of both processes appears to be quasilinear and similar. There is no indication of resonance or unusual responses. For the largest transport category aircraft (e.g., Boeing 747) about 1700 hours of moderate and 20 hours of severe or stronger turbulence can be expected at the new airport each year with higher frequencies for smaller aircraft. Significant TIWT will occur in episodes that are typically several days in duration and separated by several weeks. The wealth of scientific information provided by the field study were directly applicable to the WTWS design.
The results of the scientific studies coupled with feedback obtained from air traffic controllers, airlines, pilots, and aviation meteorologists were used to develop an operational concept for the WTWS. The WTWS concepts and user needs were established by NCAR/RAP (C. Biter, Mahoney, T. Lindholm, Neilley and Cornman) over a two-year period culminating with a demonstration of a prototype WTWS in October 1995. The major factors and issues raised by the users for consideration in the design of the WTWS included:
Each of these items were carefully considered in the design of the WTWS. The WTWS was designed by staff at RAP (G. Wiener, Mahoney, D. Albo, Neilley, Cornman, C. Morse, L. Carson, M. Dixon, D. Fletcher, M. Limber, K. Goodrich, G. Cunning), and MMM (S. Low-Nam, D. Gill, and W. Kuo), and by staff at HKUST (Chen, A. Lau, D.Yeung, A. Kwok and J. Ho).
The WTWS products were designed to enhance the safety, capacity, and efficiency of operations at Chek Lap Kok by providing: 1) pilots with concise alerts for convective and terrain-induced windshear and turbulence conditions, 2) air traffic managers and supervisors with information to aid in effective decision making during adverse weather, and 3) aviation forecasters with high-resolution real-time meteorological data and forecast guidance for the Hong Kong region.
To enhance system flexibility, the WTWS was designed to run on standard, commercially-available hardware platforms running a UNIX operating system. The system software is modular and is written in the C and Fortran programming languages. The software was designed to conform to international standards for open systems allowing the WTWS software to run on vendor-neutral platforms.
The WTWS integrates data from various sensors and sources including anemometers, Doppler weather radar, Doppler wind profilers, numerical weather prediction models, and global weather observations (surface, aloft, aircraft, satellite, ships, etc.). A simplified schematic diagram of the system is shown in Fig. 2. The WTWS imports and exports data directly to a number of external systems including the Meteorological Processing System (METPS), Terminal Doppler Weather Radar (TDWR), Aerodrome Meteorological Observing System (AMOS), Automated Weather Stations (AWS), Royal Observatory Main Computer, wind profilers, and airport master clock. Indirectly, the WTWS receives data from the World Area Forecasting System (WAFS) and the Global Telecommunications System (GTS).
The WTWS contains algorithms which produce turbulence and windshear products based on sensor data. Many of the algorithms utilize an analysis technique often referred to as "fuzzy logic." This technique makes use of disparate data types and keeps important information throughout the decision process maximizing algorithm performance. Sensor inputs include AWS, TDWR, AMOS, and wind profilers and each of these inputs are used by specific windshear and turbulence detection algorithms. The weather product algorithms output gridded information which is then integrated using decision algorithms. By employing such a scheme, a new turbulence or windshear algorithm can be added to the system by tailoring its output to the standardized gridded format.
The integration decision algorithms produce a set of grids which are then read by turbulence and windshear alert generation algorithms. The alert generation algorithms are responsible for determining hazard level and whether hazardous weather is impacting the runway corridor. The alert status of the runway corridor in the form of alphanumeric alerts, as well as windshear event shapes, are forwarded to the alert integration algorithm. The alert integration algorithm is then responsible for integrating the alerts generated by TDWR as well as those generated by WTWS and forwarding the integrated alerts to the graphic and alphanumeric alarm displays.
A version of the Penn State/NCAR Mesoscale Model version 5 (MM5) is used to provide the prediction of the atmospheric condition around the new airport. The MM5 is a state of the art non-hydrostatic limited-area model, which has a long history of development primarily for mesoscale atmospheric phenomena . The MM5 modeling system was adapted for the purpose of providing real-time, short range prediction of the mesoscale atmospheric conditions conducive to terrain-induced windshear and turbulence.
The operational MM5 system is comprised of: 1) a pre-processing component which ingests, decodes, and quality-checks incoming data and prepares an analysis for the assimilation cycle; 2) the numerical model which provides meteorological forecasts at regular intervals; and 3) a post-processing component which performs diagnostics and distributes the model forecast products.
The modeling system operates on a high-speed workstation platform. The system ingests the stan-dard World Meteorological Organization (WMO) message format data-feed and analyzes the irregularly spaced data to a uniform MM5 grid in 3-dimensional space. The lateral boundary conditions are provided by an external global model, making the information available in the internationally recognized and accepted WAFS format.
Although the MM5 produces weather forecasting guidance for the entire Hong Kong region, airport specific aviation forecasts are generated by using a post-processing algorithm. This algorithm, which is similar to Model Output Statistics (MOS) techniques commonly used at major forecast centers, provides 12-hr forecasts with 30-min resolution of wind and turbulence at CLK.
The primary WTWS product suite consists of terrain-induced turbulence detection, terrain-induced windshear detection, convective microburst/windshear detection, gust front detection, precipitation intensity, storm motion, terrain-induced turbulence prediction, airport surface wind prediction, and mesoscale numerical weather prediction guidance. The graphical and text formats are designed for easy interpretation by pilots, air traffic controllers, air traffic managers, and aviation forecasters. The WTWS utilizes four different display types: a graphical situation display, an alphanumeric alarm display, and two system operation and monitoring displays. The graphical display illustrated in Fig. 3 is one of the interfaces to deliver hazardous weather warning information and other meteorological products to the users. The display shows the horizontal profile of various hazardous weather areas, vertical wind profiles near the approach and departure corridors, as well as text warning messages. The current and forecasted situation can be examined. Movieloop views of the recent product history are possible. The meteorological situation is displayed in several user-selectable ranges and variable detail levels. Critical products and important situation changes can be visually highlighted on the display and/or announced by audible signals.
The alphanumeric alarm display is designed to alert controllers to time-critical weather hazards, and to provide textual warnings for communication to pilots. Alerts are given as "microburst," "windshear," or "turbulence," with associated intensity and location. For "windshear" and "microburst" alerts, intensity is given as headwind "loss" or "gain" in knots, and for turbulence, the intensity is given as "moderate" or "severe." The intensity is the maximum expected along the alert corridor, and the alert location is where the event is first expected to be encountered. Event locations for windshear alerts are given as 1, 2, or 3 nautical miles (nm) on approach or departure or on the runway. Event locations for turbulence alerts are given as departure or approach.