NCAR's demonstration system merges NTDA output from multiple radars to form a 3–D turbulence map, revealing regions of moderate and severe turbulence within clouds and thunderstorms.

166 FAA, NWS and DoD NEXRADs are currently in operation, providing Doppler wind and wind variability measurements as well as the more familiar reflectivity. The NTDA is a software upgrade that uses these data to detect in–cloud turbulence before aircraft encounter it.

Making use of the wind variability data provided by Doppler weather radars, RAL scientists have developed and tested the NCAR Turbulence Detection Algorithm (NTDA), designed for use on the nation's network of NEXRAD radars. The NTDA utilizes NEXRAD Level II data – the reflectivity, radial velocity, and spectrum width – to perform data quality control and produce atmospheric turbulence intensity (eddy dissipation rate, EDR) measurements of "in–cloud" turbulence. The expert system combines multiple modules for quality control and turbulence estimation using a "fuzzy logic" methodology to produce a final EDR and associated quality control index, or confidence. By providing direct detection of turbulence, NTDA provides an important addition to radar reflectivity as an indication of in–cloud aviation hazards. NTDA development has been funded by the FAA's Aviation Weather Research Program.

Example of text–based graphic used in uplink demonstration.

The first version of the algorithm, NTDA–1, was delivered to the NWS Radar Operations Center and deployed on all NEXRADs in 2008. The next version, NTDA–2, accommodates recent NEXRAD changes and is targeted for deployment in the 2012–2013 timeframe. A prototype version, running in real–time at NCAR, processes Level II data from 133 NEXRADs around the U.S. and produces a 3–D mosaic of in–cloud EDR updated every 5 min. In addition to possible direct use by pilots, airline dispatchers and air traffic control meteorologists, the NTDA data are being incorporated into a new Graphical Turbulence Guidance Nowcast (GTGN®) product. Given the highly transient and spatially variable nature of in–cloud turbulence, maximum benefit of the NTDA product can be achieved if pilots can receive this information in the cockpit in nearly real time. This capability has been demonstrated in collaboration with United Airlines: character graphics depictions of in–cloud turbulence 100 miles ahead and 40 miles to either side of the planned route were uplinked to selected United Airlines flights via ACARS. Pilots indicated that this information could be very useful for making tactical decisions near storms, suggesting that this product could improve airline safety while reducing delays due to unneeded deviations.

Coverage provided in a 24 hour period of in situ EDR measurements (colored code in units of m 2/3 s -1) from UAL mostly 757–200 aircraft (top panel), with one–minute routine reporting and DAL 737–800 aircraft (bottom panel) with event–based EDR reporting plus 15–minute routine reporting.

For both research and operational purposes, there are currently an insufficient number of reliable, accurate, and timely measurements of atmospheric turbulence locations and intensities. Turbulence "measurements" in the form of pilot reports (PIREPs), although useful, do not comprise a satisfactory turbulence detection system. NCAR, under FAA sponsorship, has developed in-situ algorithms for inferring and reporting turbulence encountered by commercial transport aircraft. The algorithms compute an eddy dissipation rate, or EDR (m 2/3 s –1), from available on–board flight parameters. EDR is truly a "state–of–the–atmosphere" turbulence metric; it is widely used in the research community as a measure of turbulence intensity, and has been adopted as the ICAO standard for aviation turbulence. About 100 757 aircraft from United Airlines are presently equipped with software which downlinks EDR reports recorded at one–minute intervals (in cruise) In order to reduce communication costs, an improved algorithm has been developed that downlinks turbulence encounter information immediately if the EDR is above a certain threshold, and still provides routine reporting but at less frequent intervals. This improved algorithm has been recently implemented on about 70 Delta Air Lines 737–800s. The algorithm will also soon be deployed on DAL’s 767 fleet.  As the airline industry becomes more aware of the benefits of this program, it is anticipated that some form of in situ EDR observation systems will ultimately operate on the majority of commercial aircraft flying both continental U.S. (CONUS) and international routes. Although EDR is the preferred prediction metric for forecasting purposes, some users may also require the aircraft acceleration loads associated with an atmospheric EDR measurement. This is easily provided by an aircraft/configuration dependent linear mapping factor.

Remote sensing

The NEXRAD Turbulence Detection Algorithm is described in the Radar section. The feasibility of using other remote sensing techniques to detect turbulence for tactical avoidance is also being studied. These include airborne forward–looking Doppler radar, lidar, GPS scintillation, and forward–looking infrared interferometer methods.

Over the last several years the FAA has funded NCAR and others to develop a turbulence nowcast and forecast system for application over the continental U.S. The forecast system, named GTG for "Graphical Turbulence Guidance," provides contours of turbulence potential based on NWP model forecasts out to 12 hours lead time. The GTG system is part of the NCEP operational suite. An example as it currently appears on ADDS is shown below.

The GTG procedure uses numerical weather prediction model forecasts to compute a number of turbulence diagnostics which are then weighted and combined. The relative weights for the combination are dynamically optimized for best agreement with the most recent available turbulence observations (in situ EDR data and pilot reports). This procedure allows the algorithm to minimize forecast errors due to uncertainties in individual diagnostic performance and thresholds. Rigorous statistical verification exercises have been performed in which probabilities of yes and no detections were determined by comparing turbulence forecasts to PIREPs and in situ EDR data. These statistics have made it possible to compare performance of the individual diagnostics, as well as test various diagnostic thresholding and weighting strategies. The overall forecast performance using the weighted diagnostics provides superior skill to the use of individual diagnostics.

The GTG is a constantly evolving product, with specific mountain wave turbulence, clear air turbulence and in the future also convectively–induced turbulence diagnostics (CIT) as well as probabilistic forecasts. Furthermore, a global GTG forecast product has been developed that has been shared with NCEP and the UKMet Office as part of the ICAO World Area Forecast System (WAFS) efforts to be operational in 2020.

Muñoz-Esparza, D. and R. Sharman, 2018: An Improved Algorithm for Low-Level Turbulence Forecasting. J. Appl. Meteor. Climatol., 57, 1249–1263, https://doi.org/10.1175/JAMC-D-17-0337.1

Sharman, R., and J. Pearson, 2017: Prediction of energy dissipation rates for aviation turbulence. Part I: Forecasting nonconvective turbulence. J. Appl. Meteor. Climatol., 56, 317–337, https://doi.org/10.1175/JAMC-D-16-0205.1.

Sharman, R., and T. Lane, 2016: Aviation Turbulence: Processes, Detection, Prediction. Springer, 523 pp.

Sharman, R., C. Tebaldi, G. Wiener, and J. Wolff, 2006: An integrated approach to mid- and upper-level turbulence forecasting. Wea. Forecasting, 21, 268–287, https://doi.org/10.1175/WAF924.1.

DCIT EDR output at FL370 for 00 UTC 5/14/09; (Bottom) blowup image of the upper Midwest showing overlaid 1–hr in situ EDR tracks validating the DCIT diagnosis. (Note that there are temporal offsets between the DCIT analysis time and some of the EDR measurements, so the correlation is not exact.)

While GTG forecasts turbulence based on NWP model data, limitations of the model accuracy, resolution and timeliness limit GTG's ability to adequately diagnose convectively–induced turbulence (CIT). The NEXRAD Turbulence Detection Algorithm (NTDA; see Radar section) partially addresses this deficiency by measuring in–cloud turbulence. However, there may often be regions of CIT above or adjacent to storms that are in clear air or otherwise not detectable by Doppler weather radar. To address this gap, the Diagnose CIT (DCIT) algorithm is being developed. Since DCIT is primarily concerned with resolving areas of turbulence close to but not in convective clouds, traditional PIREPs cannot be used as "truth" in evaluation and tuning because of their inherent location and timing uncertainty. Therefore in situ EDR data are required to tune and verify the DCIT algorithm.

DCIT is being developed using statistical learning and analysis techniques that provide estimates of the importance of candidate predictor variables and produces an empirical predictive model. An initial version of DCIT based on the RAP model, NEXRAD, and geostationary satellite data has been developed and is running operationally at NCAR.

Map of Taiwan

Advancements in science and technology, particularly computer and networking capacity, have allowed many civil aeronautics authorities to expand and modernize their aeronautical weather capabilities and services with the intention of providing their end users with more timely, accurate, and ready to use weather information. The primary objective of developing and implementing an advanced capability is to provide high-resolution four-dimensional (space and time) aviation weather products to meteorologists, pilots, air traffic controllers, and airline dispatchers and station operators in order to enhance flight safety and aviation system capacity.

Since 1998, the National Center for Atmospheric Research (NCAR) has been partnered with the Taiwan Civil Aeronautics Administration (CAA) to develop an advanced aviation weather system called the Advanced Operational Aviation Weather System (AOAWS) for Taiwan aviation system users. NCAR's local Taiwan partner on the project includes the Institute for Information Industry (III) (1998- 2008) and InfoExplorer Ltd. (2009 to 2010), and International Integrated Systems, Inc. (2011 to present). The main objective of the AOAWS is to enhance operational safety, capacity and efficiency in the terminal and Taipei Flight Information Region (FIR). The AOAWS product suite is designed to aid both tactical and strategic decision making for the direct users of the system, specifically pilots, controllers, traffic managers, and forecasters at the Taipei Aeronautical Meteorological Center (TAMC) supporting the formulation of aviation weather products.

Illustration of the major components of the AOAWS. The WMDS is the Web-based Multi-dimensional Display System, an integrated display including all relevant aeronautical weather information. A high-resolution numerical weather model is at the heart of the system.

The AOAWS system includes Low-Level Windshear Alert Systems (LLWAS) at Taiwan Taoyuan International Airport (TTY) and Sungshan Airport, advanced aviation weather display systems at the Taipei Aeronautical Meteorological Center (TAMC), Taipei Flight Information Service (FIS), Sungshan Weather Station, TTY Weather Station and FIS, Kaohsiung Weather Station and FIS, and the Taipei Area Control Center (TACC). The AOAWS also includes a World Wide Web system that allows airlines, pilots and other aviation system users to have remote access to AOAWS products.

The AOAWS operational concept has been developed by coupling user requirements with current scientific and engineering capabilities. Extensive experience gained by NCAR during development, deployment and operation of the Terminal Doppler Weather Radar (TDWR), Low-Level Windshear Alert System (LLWAS) and other advanced aviation weather systems for theFederal Aviation Administration (FAA) in the U.S. and the Windshear and Turbulence Warning System (WTWS) for Hong Kong provides a solid framework for an AOAWS operational concept.

AOAWS System Overview

Advanced aeronautical weather systems, such as the AOAWS, typically include a high-resolution weather model, model data post-processing system for calculating aviation impact variables such as icing and turbulence, display systems that allow users to interactively view weather products both in plan view and along flight routes, and servers to disseminate the information to end users in remote locations. A conceptual illustration of the AOAWS is shown below.

User Needs

The best starting point for developing an operational concept is to review the issues and needs raised by the future users of the AOAWS. User needs were developed during meetings held in Taipei in 1996, 1997 and 1999. Organizations represented in the meetings included the Taiwan Civil Aeronautics Administration (CAA), Far Eastern Transport (FAT), China Airlines, EVA Air, Air Force Meteorological Center, Central Weather Bureau (CWB), TransAsia, Hwa Hsin Airlines, Formosa Airlines and U-Land Airlines.

Chun-Ming Jou (So-So), visiting engineer from the CAA in Taipei,  is working on the MDS display at the AOAWS lab in Boulder.

Major issues considered in the development of AOAWS concepts:

1. Convective windshear (including microburst) detection in the terminal areas has higher priority than other advanced aviation weather products.
2. Terrain-induced windshear and turbulence in the terminal area has the second highest priority.
3. Thunderstorm hazards in the terminal area are important followed by thunderstorm hazards in the domestic enroute region and lastly in the FIR region. Thunderstorm hazards include windshear, turbulence, lightning, and to a lesser extent, icing.
4. Knowledge of airport surface winds and ceiling and visibility are necessary for safe operations. Better forecasts of those conditions are highly desired.
5. For flight planning, accurate information (current and forecast) of winds and temperature aloft are required.
6. The users agreed that better weather detection and forecasts accuracy for all aviation weather information is desired.

With this in mind, a number of considerations for AOAWS development were reviewed and accepted. AOAWS design considerations include:

1. The windshear systems – Low–Level Windshear Alert Systems (LLWAS) and the Windshear Processor (WSP) – should have a high Probability of Detection (POD) and low False Alarm Rate (FAR).
2. If more than one sensing system is used for hazards detection, the resulting alert information should be based on an integrated approach.
3. Terminal alert information should be concise to keep the controller workload low.
4. The AOAWS should use standard terminology for describing weather phenomena. For example, products such as flight categories should use standard definitions for IFR, VFR, etc., and standard terms should be used for turbulence and windshear. Event intensities should also be given using standard terminology.
5. Alert conditions should be reserved for operationally significant (safety critical) events.
6. Crosswind shear may be an operational problem at some airports, particularly at CKS and Kaohsiung.
7. Vertical windshear is thought to be a common occurrence at CKS. More research is needed to understand this phenomenon and determine whether it is an aviation hazard.

The windshear alert generation strategy should be consistent with FAA and other established systems. For example:

1. Alerts should be provided, where possible, out to 3 nm on approach and departure.
2. Alert update rate should be approximately 60 seconds or faster.
3. The alert corridor around the runways should be 1/2 nautical mile wide on either side of the centerline to account for windshear event movement.
4. Windshear alerts should be provided as gains or losses in wind speed (knots).
5. The minimum windshear alert threshold should be 15 knots.
6. The minimum microburst threshold should be 30 knots (windspeed loss).

Ching-Huei Hsu (Jeff ), a visiting engineer from the CAA in Taipei,  working on the AOAWS MDS display at the Boulder lab.

Terminal Windshear Systems

From an operational perspective, the AOAWS windshear detection component provides two major functions:

1. To improve safety by providing tactical decision making information to pilots for windshear events.
2. To provide strategic meteorologists at the weather stations for use in decision-making for optimizing terminal efficiency and capacity.

Because airspace is limited, especially during bad weather, it will be important to utilize all available AOAWS information in order to optimize operations.

Maximizing the utility of the AOAWS windshear components is a function of how the system is ultimately used in an operational environment and this is governed by its capabilities and performance.

Convective Windshear and Microburst Products

Microbursts are well recognized as an aviation hazard. Until the U.S. TDWR and Enhanced LLWAS systems were developed and implemented, no explicit microburst alerts were available. Onboard reactive windshear devices now provide windshear alerts to help pilots recognize that they have entered a windshear event, but they are reactive and do not provide an alert that can help pilots avoid windshear before they enter the event. Forward-looking systems are now just coming into operation and these will provide alerts of a minute or two ahead of entering windshear events. The ground-based systems (TDWR and LLWAS), which define microbursts as windshears with peak-to-peak windspeed differences of > 30 kts over distances < 2.5 nm, have microburst probabilities of detection (PODS) > 90% and false alarm rates (FARs) < 10%. Due to the high performance, the U.S. airlines have developed the policy that aircraft operations (approaches or departures) must cease if a "microburst alert" is given by these systems. When the TDWR or LLWAS systems generate a "windshear-with-loss alert", which is defined as a peak-to-peak windspeed difference of 15-29 kts over a distance of < 2.5 nm, it is up to the pilot to determine whether to continue operating.

Convective windshear events and microbursts typically last about ten minutes. This short duration is due to a combination of the true life cycle of the events coupled with the fact that they are usually moving with the precipitation cells. At times when the precipitation cells are moving more slowly (< 20 kts), the windshear events may impact the airport operations for a longer period. The avoidance of the hazard is an obvious safety benefit; however, every delayed departure and missed approach will disrupt the traffic flow and hence, impact capacity (if airspace is limited). The geographical information provided to the air traffic supervisors and managers on the GSD provides an opportunity to see the "bigger picture" and minimize disruptions.

An example of an AOAWS web system homepage

Web System Format And Style

The AOAWS Web System will be styled after a Web system that was developed by NCAR and NOAA for the NWS Aviation Weather Center (AWC) in Kansas City, Missouri. The complete AWC Web System can be viewed here: www.aviationweather.gov

The web system developed for AWC was first implemented in 1996. Since that time it has evolved from a gridded data server to an excellent resource for accessing world wide aviation weather information. The AWC system was designed for non-meteorologists (pilots, airline dispatchers, etc.) and its interface is simple and intuitive.

The AOAWS Web system will follow the general style of the AWC system for the following reasons:

1. The AWC system has been developed with extensive user input
2. The AWC system is very successful
3. The aviation community will have continuity between the AWC system and the AOAWS system lessening the time necessary to become comfortable with its use
4. NCAR has experience in developing the AWC system and can incorporate many of the features and functions of the AWC system into the AOAWS Web system

After the completion of the AOAWS Web system, a Chinese language interface will be created in the appropriate places. The Help function will also be converted to Chinese to make it easier for Chinese speaking users to interpret the products.

Partners

• The Institute for Information Industry (III)
• Civil Aeronautics Administration (CAA)
• National Center for Atmospheric Research (NCAR)
• InfoExplorer Ltd.

What is R?

Introduction to R

R is a language and environment for statistical computing and graphics. It is a GNU project which is similar to the S language and environment which was developed at Bell Laboratories (formerly AT&T, now Lucent Technologies) by John Chambers and colleagues. R can be considered as a different implementation of S. There are some important differences, but much code written for S runs unaltered under R.

R provides a wide variety of statistical (linear and nonlinear modelling, classical statistical tests, time-series analysis, classification, clustering, ...) and graphical techniques, and is highly extensible. The S language is often the vehicle of choice for research in statistical methodology, and R provides an Open Source route to participation in that activity.

One of R's strengths is the ease with which well-designed publication-quality plots can be produced, including mathematical symbols and formulae where needed. Great care has been taken over the defaults for the minor design choices in graphics, but the user retains full control.

R is available as Free Software under the terms of the Free Software Foundation's GNU General Public License in source code form. It compiles and runs on a wide variety of UNIX platforms and similar systems (including FreeBSD and Linux), Windows and MacOS.

R software packages [LINK: https://www.r-project.org/]

• R verify [LINK: http://www.inside-r.org/packages/cran/verification/docs/verify]
• R SpatialVx [LINK: http://www.inside-r.org/packages/cran/SpatialVx/docs/SpatialVx]
• R extRemes [LINK: http://www.inside-r.org/packages/cran/extRemes/docs/extRemes]

VLAS represents the Doppler lidar variant of VDRAS.  VLAS provides very high-resolution wind information at the neighborhood scale and has been used to study atmospheric transport and diffusion in urban environments.