Document Sample
TREND 12 Powered By Docstoc

Working Group on Training, the Environment and New Developments in
                     Aeronautical Meteorology

                       (Former ATEAM)

                    NEWSLETTER No. 12

                            February 2001



               Summary and selected review of papers presented at the
               Ninth American Meteorological Society Conference on
               Aviation, Range and Aerospace Meteorology, held in
               Orlando, Florida, 11-15 September 2000 and at the
               General Assembly of the European Geophysical Society
               Meeting, held in Nice, France, in April 2000.
                                                                          TREND Newsletter No. 12


      The Commission for Aeronautical Meteorology (CAeM) at its eight session in 1986, noting
the rapid developments in the application of modern forecast techniques and methodologies,
and considering the potential benefit of the application of advanced techniques and
methodologies to aeronautical meteorological services and the development of advanced
techniques for the quality control of meteorological data and forecasts, decided to establish a
Working Group on Advanced Techniques Applied to Aeronautical Meteorology (ATEAM). The
Commission re-established the Working Group at both its ninth and tenth sessions in 1990 and
1994. The eleventh session of CAeM held in Geneva from 2 to 11 March 1999, considering the
potential benefit of the application of advanced techniques and methodologies to aeronautical
meteorological services listed above as well as the need for CAeM to take a leading role in
considering the implication of Agenda 21 in the field of aeronautical meteorology, decided to
establish a Working Group on Training, the Environment and New Developments (TREND) in
Aeronautical Meteorology to replace the ATEAM Working Group. The TREND is currently
composed of Dr H. Pümpel (Austria), Chairman, Ms S.S.Y. Lau (Hong Kong, China), Mr M.
Obayashi (Japan), Mr V. Ahago (Kenya), Mr G. Ross (UK) and Dr R.A. Petersen (USA).

     CAeM-XI gave the following terms of reference to the TREND Working Group:

      (a)          To foster training of meteorological personnel in aeronautical meteorology
                   aimed at upgrading aeronautical meteorological practices and to act as a focal
                   point for the Commission regarding training;

      (b)          To keep under review the impact of aviation on the environment, particularly in
                   relation to meteorological conditions influencing the impact of aviation in the
                   terminal area and the impact of aircraft emissions on the environment in the
                   en-route phase of flight;

      (c)          To encourage airlines to contribute to the global database of information
                   enabling an ongoing assessment of the impact of aviation on the environment;

      (d)          To promote research and development on the forecasting of meteorological
                   phenomena of importance to aircraft operations, in particular the forecasting
                   and warning of en-route meteorological hazards;

      (e)          To review and report on research and development in techniques and
                   technologies related to aeronautical forecasting in particular with reference to
                   very-short-range terminal forecasts;

      (f)          To prepare and review guidance material on advanced techniques relevant to
                   aeronautical forecasting;

      (g)          To review and report on procedures for monitoring and verifying aeronautical

                                             Page i
                                                                        TREND Newsletter No. 12

      (h)         To advise on the introduction, and promote the exchange through technology
                  transfer, of modern forecasting techniques and technologies relevant to
                  aeronautical forecasting;

      (i)         To maintain close liaison with the Commission for Atmospheric Sciences,
                  particularly with respect to aeronautical meteorological research and
                  development requirements;

      (j)         To liaise with ICAO and aeronautical user organizations in relation to impacts
                  of aviation on the environment;

     In line with these terms of reference, TREND will continue the publication of the TREND
(formally ATEAM) Newsletter to report on training, research and development in aeronautical
meteorology and the impact of aviation on the environment.

                                           Page ii
                                                                        TREND Newsletter No. 12


One of the major tasks of the Commission for Aeronautical Meteorology (CAeM) Working Group on Advanced
Techniques Applied to Aeronautical Meteorology (ATEAM) since its establishment in 1986, has been the
development and publication of a series of Newsletters providing information on research and development as well
as on advanced techniques and technologies related to aeronautical meteorology.

At the Eleventh session of CAeM held in March 1999, the TREND Working Group was installed to continue the
successful work of the ATEAM Working Group, with a new focus on environmental matters which have also
become a main concern in aviation meteorology.

Eleven Newsletters have been produced and distributed since 1986. The Eleventh Newsletter published in January
2000 was devoted to the summary of selected papers and posters presented at the American Meteorological
Society (AMS) Eighth Conference on Aviation Range and Aerospace Meteorology, held in Dallas, Texas, USA,
from 10 to 15 January 1999.

In pursuing the tradition of ATEAM, members of TREND attended the General Assembly of the European
Geophysical Society Meeting held in Nice, France in April 2000 and the 9th Conference on Aviation, Range and
Aerospace Meteorology held by the AMS in Orlando, Florida, in September 2000.

In all, 124 scientific and technical papers were presented at the AMS conference. The papers were assembled in 9
broad subject areas. Subject areas included programme overview, aviation accidents and case studies, aviation
operations and support, aviation icing, forecasting and evaluation/verification, turbulence and wind shear, sensors
and systems, advances in weather radar support for local storm research and aviation and thunderstorm impacts.
Furthermore, 23 papers presented at the General Assembly of the European Geophysical Society Meeting were
considered suitable for this issue of the Newsletter.

Conference papers and poster sessions thought to be the most relevant to the Aeronautical Meteorology
Programme were reviewed and summarized by Dr Neil Gordon (President of CAeM) and Dr Herbert Pümpel
(Chairman of TREND). The material was subsequently edited and published by the WMO Secretariat.

The purpose of this review is to encourage a wider audience that has not had the opportunity to attend this
conference to have a look at the contents and merits of papers presented, and to make an informed choice
on papers relevant to their field of work, and/or containing relevant information for their routine duties.
Provided comments are necessarily subjective and are not intended to pass judgement on individual papers,
but to serve as an at-a-glance pointer to interesting information contained in these papers.

For further in-depth reading of the original articles and the full contact address of authors please access the AMS
Journals Online Information at o r write to the following address:

        American Meteorological Society
        45 Beacon Street
        Massachusetts 02108-3693

As you indicated in your positive feedback regarding previous ATEAM Newsletters, I trust that you will find the
content of this Twelfth TREND Newsletter stimulating and informative.

                                                     Page iii
                                                                                              TREND Newsletter No. 12


Foreword .............................................................................................................................. iii

                                                                 PART I

Ninth American Meteorological Society Conference on Aviation, Range and Aerospace
Meteorology, Orlando, Florida, 11-15 September 2000

Session 1:                     Programme Overview.......................................................................... 1

Session 2:                     Aviation Accidents and Case Studies.................................................. 1

Session 3:                     Aviation Operations Support ............................................................... 2

Session 4:                     Aviation Icing ....................................................................................... 6

Session 5:                     Forecasting and Evaluation/Verification .............................................. 14

Session 7:                     Turbulence and Windshear ................................................................. 24

Session 8:                     Sensors and Systems ......................................................................... 29

Joint Session:                 9th Aviation Conference and 20th Severe Local Storm Conference..... 36

Poster Session:                ............................................................................................................. 39

                                                                 PART II

General Assembly of the European Geophysical Society Meeting
Nice, France, April 2000....................................................................................................... 42

                                                                 Page iv
                   PART I

9th American Meteorological Society Conference
on Aviation, Range and Aerospace Meteorology

               Orlando, Florida
            11-15 September 2000
                                                                      TREND Newsletter No. 12

                              Session 1: Programme Overview

                     Paper 1.2: Global Challenges and Opportunities in
                                 Aeronautical Meteorology

     Neil D. Gordon, President of the Commission for Aeronautical Meteorology (CAeM)
                   and Director of MetService, Wellington, New Zealand

Worldwide, aviation meteorological services face major challenges in meeting increasingly
demanding requirements from their customers. However, scientific and technological advances,
as well as new ways of organizing delivery of services, also provide major opportunities for better
meeting those customer needs.

The Aeronautical Meteorology Programme of the World Meteorological Organization (WMO)
plays a pivotal role in the global co-ordination and advancement of meteorological services for
aviation. It is carried out under the technical responsibility of the Commission for Aeronautical
Meteorology (CAeM), one of eight WMO Technical Commissions. This talk will cover the role of
CAeM, and its plans and dreams for the improvement of services.

Comment: The main reason for including this reference is to note that the opportunity was taken
at this conference to “fly the WMO flag” and promote the work done by CAeM, including TREND.

                      Session 2: Aviation Accidents and Case Studies

                   Paper 2.3: Case Study Verification of RUC/MAPS Fog
                                 And Visibility Forecasts

           Tatiana G. Smirnova, CIRES/Univ. of Colorado and NOAA/FSL, Boulder
                          Stanley G. Benjamin, and John M. Brown

The operational Rapid Update Cycle (RUC) provides hourly-updated short-range forecasts of
atmospheric and surface conditions as part of its ongoing hourly assimilation cycle, including
evolution of soil moisture and temperature, and also snow temperature and snow depth if snow

Recently performed validations of hydrological cycle components, such as precipitation,
evapotranspiration and soil moisture, as well as of soil and skin temperature demonstrate the
model's capability to represent surface processes with a good degree of realism. This is a
prerequisite for successful application of RUC products to predict conditions that can adversely
impact terminal operations, such as fog and low ceilings.

Late last year, with the availability of the new computer at the National Centers for Environmental
Prediction, additional aviation impact variables became routinely available as part of RUC output.
Among these is a diagnosed horizontal surface visibility product based on a modified version of
the Stoelinga-Warner algorithm applied to native-grid RUC output.

In this paper examples of verification of these short-range visibility forecasts by the RUC and its
developmental counterpart at FSL, the Mesoscale Analysis and Prediction System (MAPS) are
described. For selected cases having widespread fog as indicated by satellite imagery and
METAR reports, areas of poor visibility as forecast by the RUC (or MAPS) will be compared with

                                                    Page 1
                                                                           TREND Newsletter No. 12

available observations and discuss performance. Improvements anticipated with the major RUC
upgrade later this year are discussed.

Comment: It is clear that the future lies in model-based guidance for terminal ceilings and
visibilities. Numerical models do not (yet) provide explicit predictions of these variables, so some
kind of parameterization is required. The work described here is of interest, though somewhat
handicapped by being applied to the RUC model running at 40 km resolution, which is barely
sufficient to model the detailed mesoscale features which give rise to regions of low visibility

                            Session 3: Aviation Operations Support

                      Paper 3.4: An Interactive Gridded Aviation Weather
                             Database: Results of a Pilot Project

                  Richard Verret, MSC, Dorval, Canada, Marie-France Turcotte
                      Vanh Souvanlasy, Michel Baltazar, and Mario Ouellet

The aviation industry is heavily dependent on accurate and timely information on current and
forecast weather conditions for flight planning and safety. The demand to have access to
weather information quickly and in an intuitive and easily understandable fashion is ever
increasing. Advancements in computer sciences and in numerical modeling of the atmosphere
have brought forward the capability of responding to the demand. Numerical weather prediction
models can ingest an ever-increasing amount of data from various sources and produce high-
quality gridded forecasts in relatively short periods of time. Computers powerful enough are
available at relatively low costs to process impressive amounts of information produced by the
numerical models, and coming from other sources, and generate quickly, user tailored
information products in graphical formats. Computer applications are available to build user-
friendly interfaces and make the information available on networks. This opens a whole era of
new aviation weather products thus allowing a quick and intuitive understanding of actual and
forecast aviation weather conditions.

In that context, the Canadian Meteorological Centre, in collaboration with Nav Canada, has
developed an interactive aviation weather database (AWeD) to be the core component of an
aviation weather display system intended to be used as a briefing-aid tool. The domain of the
aviation database covers all of Canada, adjacent waters as well as a significant portion of United
States. The gridded data is available at a 24 km horizontal resolution, at every 1000 feet in the
vertical from mean sea level up to 40 000 feet and at a 3-hour time resolution from zero to 48
hour projection times. The database is updated twice per day (00 and 12 UTC) in real time. In its
current state, the aviation weather database includes: temperatures, winds, icing, turbulence,
cloud fraction, relative humidity, vertical velocity at all flight levels. It also includes: tropopause
pressure and temperature, freezing level, total cloud cover, instantaneous precipitation rate at the
surface and station pressure. Real-time observation data, METAR, SPECI, FA, SIGMET,
AIRMET and PIREP, are also incorporated in the database.

The database gridded aviation-impact variables can be interactively queried on Internet through a
user-friendly interface which allows users to enter flight parameters, such as departure and
arrival points, alternate airports, check points along the planned route, estimated elapse time of
the flight and flight level. Series of meteorological products, all tailored to each particular flight, in
plan view and/or vertical cross-section along the route can then be requested. Approximately,
eighteen pilots have been selected to test the system in real time and in a real operational
environment and to provide feedback. Accesses to the system were monitored and statistics
gathered. The results of the project show that the gridded database is a useful tool for pre-flight

                                                       Page 2
                                                                       TREND Newsletter No. 12

planning. Most available products were judged as good as or better than the conventional
products. The main conclusion of the project is that the interactive gridded database is mature
enough to be implemented for operational usage.

Comment: A very interesting approach of providing completely automated products, which can
provide valuable information for briefing pilots. It is intended to be complementary to the official
information and briefings, but does raise questions about how to deal with any inconsistencies
which could arise between such numerical guidance and official products, and about pilot
education to interpret the information. The system was to go operational in late 2001.

                     Paper 3.5: Recent Enhancement and Plans for the
                           Aviation Digital Data Services (ADDS)

                             Lynn A. Sherretz, NOAA/FSL, Boulder
                             Greg Thompson, and Philip Kennedy

This presentation demonstrated (via the Internet) recent enhancements to the Aviation Digital
Data Service (ADDS) and described plans for future work. A key recent enhancement is the
ADDS Flight Path Tool (FPT) that enables users to generate interactively for user-specified flight
routes time-sequenced vertical and horizontal graphical cross-sections of forecasts of key
atmospheric variables (e.g., wind, icing, and turbulence) and observations (e.g., PIREPs and
METARs). The challenge in developing the FPT was to enable ADDS to generate the cross-
sections very quickly from gridded data generated by the Rapid Update Cycle and algorithms.

Recent enhancements include: improving the utility of the JAVA-based PIREP and METAR
viewers; improving the performance (i.e., ability to respond to user requests in a timely manner)
of the METAR viewer by incorporating progressive disclosure; expanding the domain of ADDS to
include the Northern Hemisphere; enabling ADDS to display NEXRAD observations at frequent
intervals; and transferring ADDS to the NWS Aviation Weather Center (AWC). The last
enhancement includes enabling ADDS to access all its data from AWC and developing software
that monitors the "health" of ADDS and alerts AWC support staff as necessary.

An important activity for the coming year entails implementing ADDS in an Automated Flight
Service Station (AFSS) and conducting a study of the utility of using ADDS as a interactive
component of flight briefings. Planned enhancements to ADDS include enabling ADDS to:
monitor observations and forecasts for user-specified conditions and alerting users when those
conditions are met; and display automated turbulence reports (pending airline approval), VORs,
and Airways.

ADDS is being developed by the NCAR Research Applications Program (RAP), NOAA Forecast
Systems Laboratory (FSL), and NWS Aviation Weather Center (AWC) under the auspices of the
Federal Aviation Administration (FAA) Product Development Team for the Aviation Gridded
Forecast System (AGFS). Funding is provided by the FAA Aviation Weather Research (AWR)

Comment: The site at is highly recommended. The ADDS is the
primary data dissemination for products based on the Aviation Gridded Forecast System, which
has been under development for a decade.

                                                    Page 3
                                                                         TREND Newsletter No. 12

                    Paper 3.6 Creating an Aviation “Centre of Expertise"

                            Steve Ricketts, MSC, Edmonton, Canada

In late 1997, Environment Canada's Prairie and Northern Region decided to re-organize its
weather service program along a "centre of expertise" model. All aviation-forecast services were
consolidated in Edmonton, along with forecasting responsibility for the Canadian Arctic. The
Prairie Aviation and Arctic Weather Centre (PAAWC) produces aviation forecasts for the three
prairie provinces, the Northwest Territories, and Nunavut (just over 50% of Canada by area), and
it looks after 1/3 of the aerodrome forecasts (TAFs).

This "new" centre and structure has created both opportunities and challenges. It has allowed a
focus on meeting the needs of a large client, NAV CANADA, which is a private, not-for-profit
corporation responsible for the Air Navigation System (ANS) in Canada, including aviation
weather services. Goals of the PAAWC include improving the quality of the weather services
and to work on establishing better working relationship with NAV CANADA and other aviation
users. It has implemented several ideas along these lines: restructured its operations, developed
innovative tools to monitor the weather better and streamline forecast production, increased
client contact, and done aviation-specific technique development.

Recently, significant effort has gone into the new graphical area forecast (GFA) product and
investigating ways to improve the quality (both accuracy and utility) of its aviation forecasts. This
last item is important and a big challenge, considering the tremendous spatial and temporal
variability in the weather parameters that are most critical to aviation operations (e.g., ceiling,
visibility, wind, icing, and turbulence), and the varied needs of the aviation community. The
PAAWC also has started using performance measurement data and feedback from clients to
improve its services.

This talk will cover some of the work that has been done in the PAAWC, and review the benefits
(and drawbacks) of the "centre of excellence" approach, as seen through the PAAWC's eyes.

Comment: A very interesting presentation by Merv Jamieson, which usefully highlighted some of
the management issues associated with rapid change and the need to be much more customer

         Paper 3.7: TAF Tools: Development of Objective TAF Guidance at CMC

              Pierre Bourgouin, Canadian Meteorological Centre, Dorval, Canada
                   Richard Verret, Laurence Wilson, and Jacques Montpetit

Terminal aviation forecasts (TAFs) are site-specific forecasts that are currently prepared every 6h
manually, using guidance from the operational numerical weather prediction (NWP) models and
available observations. TAFs include forecast information on ceiling, visibility, weather,
obstructions to visibility and wind. It is believed that gains in forecast production efficiency can be
realised by automating as much of the production process as possible, leaving the final control of
the forecast contents with the operational forecaster. Statistical methods are used because they
are cheap compared to other solutions. There are three major components of the project, one
for very-short range forecasts, one for the short range and finally, one that blends the two
techniques together and possibly incorporate other available information.

Numerical weather prediction models have difficulties forecasting precise weather elements for a
specific site as needed for a TAF. Persistence, especially conditional climatology, is in fact very

                                                      Page 4
                                                                     TREND Newsletter No. 12

difficult to beat during the first few hours. It has been shown that a system based on
observations is superior to persistence climatology and to NWP-based statistical systems. To
take advantage of these results, we are developing a very short-term forecasting technique
based solely on current available observations. About 40 years of hourly observations are used
to develop forecast equations relating observations at a time to observations at a later time
To+dT. The equations are developed using a Multiple Discriminant Analysis (MDA) technique.
MDA has recently been shown to give superior forecasts to CART for cloud amount. Most of the
work so far has been devoted to the construction of a large database consisting mainly of hourly
observations but some results will be shown.

Observation-based systems may provide the best possible forecast at very-short ranges but their
skills degrade rapidly in time. It was decided to develop a perfect-prog system to forecast the
different elements required to write a TAF. Reanalyses from the National Centre for
Environmental Prediction are used to derive site-specific predictors such as temperature,
vorticity, moisture advection, stability indices, etc. The predictors are paired with observations
which have been processed to be representative of a time-step of 3h. Equations will also be
developed using a MDA technique. The presentation describes the technique design and results
to date.

The third component of the project, which will blend the two techniques together, and possibly
incorporate other available information, is still in its early development stage.

Comment: A very comprehensive approach to blending “traditional” pure observations-based
statistical forecasting with model-based information. It is still early days on how well it will

     Paper 3.16: CGen: Enabling AWC Forecasters to Generate Convective
                            SIGMETS via AWIPS

                     Dennis M. Rodgers, NOAA/OAR/FSL, Boulder, USA
                                Greg Pratt, and Jim Frimel

CGen, (Convective SIGMET Generator) is an extension of the AWIPS D2D display system. This
tool provides the means to graphically define advisory areas, lines, or points, and outlook areas,
and to create Convective SIGMET (WST) text using a graphical user interface. CGen is
composed of graphic and text tools designed to increase the efficiency of WST generation. Since
CGen is an extension of D2D, the full display capabilities of D2D may be utilized in the WST
preparation process.

The WST preparation process will fall into at least one of three modes for each of three regions
covering the Continental U.S.: a null forecast; a modification of the previous hour's issuance; or
one or more new advisories. Each of these modes is supported by CGen.

Advisories may be created, manipulated, or deleted using the mouse pointer in the main pane of
D2D. Advisory anchor point locations are identified with respect to aviation navigation beacon
(VOR) locations. Anchor point and affected states information is automatically inserted into the
WST message. Individual WST advisory text is created in the graphic user interface (GUI) on
D2D, then inserted into East, Central, and West WST bulletins, assembled into the WST
message, and displayed in the CGen Text Display window on the AWIPS text workstation. In the
CGen text window, the forecaster may inspect the message, perform final editing or insert free-
form text, and issue the message. Scheduled, Special, and Correction issuances are possible
with CGen.

                                                   Page 5
                                                                      TREND Newsletter No. 12

CGen is an end-to-end tool. The previous hour's WST advisories can be displayed as editable
graphics, with associated text, for editing of the geographic extent and/or text. Bulletin header
and footer, issue time, valid time, and forecaster identification are updated in the text

Gains in forecaster productivity over current WST preparation methodology and forecaster
acceptance of CGen will be discussed.

Comment: Traditional workstation developments have focussed on better and faster ways of
presenting increasing volumes of information for the forecaster. This paper is about using
workstations to enhance forecaster productivity, in preparation of products. Although this
development may not be operationally implemented as described, the concepts behind it are well
worth a look.

                                   Session 4: Aviation Icing

 Paper 4.2: Freezing Drizzle and Supercooled Large Droplet (SLD) Formation in Stably
    Stratified Layer Clouds: Results from Detailed Microphysical Simulations and
                              Implications for Aircraft Icing

                  Roy M. Rasmussen, NCAR, Boulder, USA, Istvan Geresdi,
                      Greg Thompson, Kevin Manning, and Eli Karplus

In this paper an evaluation was made of the role of: 1) low Cloud Condensation Nuclei (CCN)
conditions, and 2) preferred radiative cooling of large cloud drops as compared to small cloud
drops near cloud top, on cloud droplet broadening and subsequent drizzle formation in stably
stratified layer clouds. The evaluation is performed by simulating cloud formation over a two-
dimensional idealized mountain using a detailed microphysical model implemented into the
NCAR/Penn State MM5 mesoscale model. The height and width of the two-dimensional
mountain was designed to produce an updraft pattern with extent and magnitude similar to
observed freezing drizzle cases. In addition, an evaluation was also made of two different
methods of ice initiation on freezing drizzle formation, and the role of cloud depth and cloud top
temperature on the formation of freezing drizzle

Comment: A thorough investigation into the relative importance of airmass characteristics such
as maritime vs. Continental, availability of cloud condensation nuclei, and presence of ice nuclei.
Somewhat surprising that in this setup, the roles of radiative cooling, shear and shallow
embedded convection do not appear to be dominant.

               Paper 4.3: Environments Associated with Large Droplets,
            Small Droplets, Mixed-Phase Icing, and Glaciated Conditions Aloft

                Ben C. Bernstein NCAR, Boulder, USA. and Frank McDonough

During the winters of 1997 and 1998, the NASA-Glenn Twin Otter research aircraft sampled a
variety of icing and non-icing environments over the Great Lakes region. Conditions encountered
include freezing rain, freezing drizzle, small/cloud droplets, ice crystals and mixtures thereof.
Occurrences of these phenomena are objectively compared with observational data from GOES-
8 satellite, METARs, and stability derived from aircraft temperature profiles.

GOES-8 infrared satellite indicated that small-droplet and "non-classical" large-droplet icing
occurred most often occurred when cloud top temperatures (CTTs) were warmer than -17C,

                                                    Page 6
                                                                       TREND Newsletter No. 12

while mixed-phase icing occurred within a cooler subset of CTTs. A significant percentage of
large- and small-droplet icing occurred beneath melting zones in "classical" environments, when
CTTs were much colder. A GOES-8 based icing algorithm is shown to perform quite well for all
but the "classical" icing encounters.

Surface observations of cloud coverage and certain precipitation types provide good information
about icing and SLD encounters aloft, while other observations provide much less. In particular,
surface reports of freezing drizzle, freezing rain, and ice pellets are excellent indicators of SLD
environments in the lower atmosphere. Results differ significantly between non-classical and
classical situations, and with location relative to melting zones. Aircraft temperature profiles
reveal that the stability in non-classical SLD environments varies greatly. Non-classical SLD was
observed with in-cloud temperature structures that ranged from isothermal to moist adiabatic.
SLD formed in relatively low stability tended to be isolated to situations where relatively low
concentrations of cloud condensation nuclei are expected. Classical SLD primarily occurred in
very stable environments beneath the melting zone.

Comment: The good correlation between surface observations of freezing drizzle and SLD aloft
underlines the importance of reliable surface obs of this phenomenon including automated sites!
In case of low warm cloud tops and no overlying high cloud, a combination of several spectral
channels on geostationary satellites does reasonably well during daytime (no VIS available at
night!), bit in the classical case of melting and refreezing below dense multilayer cloud, satellite
data are not a great help.

                      Paper 4.4: Aircraft Icing Detection using S-Band
                            Polarization Radar Measurements

                       S. M. Ellis NCAR, Boulder, USA, J. Vivekanandan
                           E. A. Brandes, J. L. Stith, and R. J. Keeler

The combination of supercooled liquid water content (SLWC) > 0.2 g/m3, cloud temperature
approximately -10 C, and median volume diameter (MVD) > 30 micron causes significant aircraft
icing. The amount of SLWC is highly variable and depends on moisture content and temperature
profile. In regions where temperatures are below 0° C, super-cooled liquid water (SLW) is
generated through transport of liquid droplets from warmer temperatures (updrafts), or the
condensation of vapor into cloud droplets. These super-cooled droplets are subject to contact
freezing, or riming, when they come into contact with a solid body such as hailstones or aircraft.
Three parameters, namely liquid water content, MVD and temperature, are used for
characterizing the severity of icing conditions. The presence of SLW is difficult to detect, yet
constitutes a significant threat to aviation safety. The ability to detect and track SLW remotely
from the ground would result in greatly increased air safety near airports. A dual-wavelength (Ka-
and X-band) radar technique can be used for detecting the amount of liquid water and droplet
size in cloud. However, the technique and equipment are complicated and expensive to deploy
and operate. Therefore, alternate methods of remote SLW detection are under investigation.
Recent studies have shown the utility of polarimetric radars to distinguish hydrometeor particle
types. In this paper, we investigate the detection of SLW using ground based S-band polarization
radar measurements. The NCAR S-Pol radar (a 10 cm wavelength dual polarimetric ground
based radar) was operated during the Mesoscale Alpine Program (MAP) in coordination with
several aircraft equipped with microphysical instrumentation (including 2-D probes, and the
Rosemount icing probe). Based on flight reports and icing detection instruments, we have
identified several time periods that the planes experienced icing conditions within regions of S-
Pol observations. The radar observations, as well as the particle identification output, are
compared to the 2-D probe measurements in regions the aircraft observed super-cooled drops.
In this paper we present preliminary results of the study to determine if polarimetric radars are
capable of detecting regions of aircraft icing hazards.

                                                    Page 7
                                                                        TREND Newsletter No. 12

Comment: Dual or slant polarization is used to discern between precipitation particles that are
spherical (typically small droplets, small graupel) and those that are plated or oblique due to
either wind resistance (large drops) or crystalline form (ice, graupel). Aircraft in-situ
microphysical probes confirm the assumptions to some extent, abut the fuzzy logic algorithm
needed to determine particle types (up to 17 different ones are attempted to define) will probably
require a fair amount of tuning to different climates and precipitation regimes. Compared to
specialized radars of very short wavelength, the S-band SPOL Radar has sufficient range to be
considered operationally useful in real-time detection of icing conditions.

A great effort is spent on a proof-of-concept work to identify critical values of SLW in clouds.
Using polarimetric radar in different regions. Problems remain at small drop/crystal sizes, and
more work on classification algorithms will be needed before practical results can be expected.

              Paper 4.5: Application of a Mixed-Phase Microphysics Scheme
                                  To Predict Aircraft Icing

                            Gregory Thompson, NCAR, Boulder, USA
                            Ben C. Bernstein, and R. M. Rasmussen

The most sophisticated microphysics option within the PSU/NCAR research mesoscale
numerical model (MM5) has undergone extensive testing and modification. The bulk
microphysical parameterization allows for the prediction of mixing ratios of cloud water, rain,
cloud ice, snow, and graupel as well as the number concentration of ice. The impetus for this
work is improving forecasts of aircraft icing under sponsorship by the Federal Aviation
Administration. Such improvement implies correct forecasts of supercooled liquid water which, in
turn, implies proper handling of ice initiation and growth. The main focus of this study is
numerical simulations of clouds and precipitation as they relate to aircraft icing and, of particular
importance, freezing drizzle/rain (or supercooled large drops, SLD). Many tests were conducted
in two-dimension using idealized flow over a barrier before continuing to fully three-dimensional
simulations involving case studies of freezing drizzle/rain events. While precipitation amount and
type reaching the ground is important, research aircraft data were also examined and compared
against the model simulations of supercooled liquid water. Observational data suggests that
many freezing drizzle/rain cases are non-classical in the sense that collision/ coalescence is
responsible for cloud droplet growth to drizzle/rain sizes. On the other hand, the classical
mechanism consists of ice initiation, growth to snow, and subsequent melting through a warm
layer (temperature greater than 0C) before passing back into a sub-freezing layer and
supercooling to freezing drizzle/rain. In this paper, both classical and non-classical freezing
drizzle/ rain cases are investigated using numerical simulations.

Comment: This very thorough investigation of the MM5’s capability to detect icing potential
based on its fairly advanced microphysics shows that the forecast system as a whole needs to
perform well in order to achieve this ambitious goa:. A misplacement of the precipitation area,
under-or overestimation of intensity (changing the melting layer through draining heat used in
melting) and excellent vertical resolution (depth of the melting layer needs to be sufficient in order
to melt the ice particles) will suffice to spoil the game for even a perfect microphysics package!
To see the potential for such detection, however, is a heartening thought.

                                                     Page 8
                                                                         TREND Newsletter No. 12

                     Paper 4.7: Information Requirements for Improved
                                   In-Flight Icing Decisions

       Laurence N. Vigeant-Langlois, MIT, Cambridge, USA and R. John Hansman, Jr.

The influence of potential information regarding in-flight icing conditions on pilot decision-making
was investigated to define functional requirements of both forecasting and remote-sensing
systems. Potential icing information and display features were manipulated in a web-based
experiment which involved 230 pilots. Display features included: a graphical plan view depiction
of icing severity, vertical view depiction, single and multiple icing severity levels as well as sensor
range, were varied in a part-task simulation experiment. Using information from each display,
pilots were presented with a set of four flight scenarios and probed on their routing decisions and
comfort level with those decisions. The experiment also included a subjective display preference
evaluation. Results show that all of the displays improved pilot decision-making over existing text-
based icing information. Displays, which included vertical depiction of icing conditions, were
found to support improved decision-making over those displays which only presented horizontal
depiction. Sensor range was not found to be a strong factor in the experiment; however the
minimum range tested was 25 nautical miles, which may be in excess of current technical
capabilities. The depiction of the severity of icing conditions was not found to be as important as
accurate information on the location of icing conditions.

Comment: The importance of meteorological information being understood by the pilots is
beautifully demonstrated in this paper. It is somewhat sobering to see that the presentation of a
forecast product about a serious condition such as icing weighs more than the actual content.
Three-dimensional graphics indicating both vertical and horizontal distribution of icing threat are
obviously what pilots want most, and this is a challenge to the science to produce a forecast
product that deserves the enhanced trust instilled by a better presentation.

         Paper 4.8: Simulations and Observations Implicating Mesoscale Gravity
                      Waves in Producing an Environment which is
                               Conducive to Aircraft Icing

              Michael L. Kaplan, North Carolina State University, Raleigh, USA
        A. J. Riordan, Y.-L. Lin, A. W. Huffman, K. M. Lux, and Kenneth T. Waight, III

An in-depth analysis of the dynamical and microphysical processes responsible for the ATR72
plane crash near Roselawn, Indiana (USA) is presented. Observed surface and upper air data,
satellite data, and numerical simulations using MASS version 5.12 at 44, 20, and 10 km
resolutions are used to assess the atmospheric conditions present at the time of the crash.
Preliminary results from the 10 km simulation indicate that high icing potential existed at the
altitude of the crash and near the time of the accident. The mesoscale numerical simulations
suggest that a ducted, large-amplitude, hydrostatic, propagating, internal gravity wave was the
key organizing mechanism for an icing environment. This wave is diagnosed from model cross-
sections which indicate a region of descent around 400mb coupled with a region of ascent
around 650mb propagating through the area surrounding Roselawn, Indiana from west-
southwest to east-northeast. Such a wave is likely to have produced three key signals necessary
for icing: 1) descent aloft resulting in the removal of frozen hydrometeors 2) ascent within the
lower-middle troposphere resulting in production of substantial cloud water, some of which is
supercooled, and 3) relatively strong vertical wind shear near the wave duct enhancing turbulent
mixing aiding the collision-coalescence droplet growth process. The gravity waves evident in the
simulation, will be verified against satellite, radar, wind profiler, and surface microbarogram

                                                      Page 9
                                                                        TREND Newsletter No. 12

Comment: In a very sobering analysis of the Roselawn accident the importance of mesoscale
dynamical features in the atmosphere on intensification of SLD conditions is demonstrated.
Whether gravity waves, convergence lines of rain bands are the decisive factors, enhanced
vertical velocity is the key to a tip in the balance between ice nuclei depletion and supply of fresh
supercooled water for droplet formation. Similar to the article above, we are lead to understand
that small-scale features in the forecast play a decisive role in the generation of hazardous icing

                     Paper 4.9: Percent Power Increase – A Simple Way
                                 To Quantify an Icing Hazard

    Donald W. McCann, NOAA/NWS/NCEP/Aviation Weather Center, Kansas City, USA
                             and Phillip R. Kennedy

Engineering studies on the icing hazard to aircraft have identified three meteorological variables
that influence how supercooled water accumulates as ice on aircraft surfaces: (1) cloud liquid
water content determines the amount of supercooled liquid water available for an aircraft to
intercept; (2) effective droplet diameter determines the percentage of available cloud liquid water
that impacts an airfoil as well as where ice accumulates; (3) air temperature determines the
location where the water freezes.

Icing forecasts using these three variables can be computed with numerical model output.
Weather forecasters must then communicate the threat to pilots in a manner that effectively
details how it will affect aircraft performance. Current reporting and forecast terminology
identifies the icing hazard as trace, light, moderate, or severe. Although these terms are based
on a degradation to aircraft performance, they have been criticized as being vague and

Ice accumulation increases airfoil drag. This increased drag must be countered by increased
thrust if the aircraft is to maintain speed and altitude. It stands to reason then that every aircraft
has a quantitative icing dial onboard - its engine power gauge (SHP, RPM, manifold pressure,
etc.). The increase required in engine power is a direct measurement of the icing hazard. If
pilots were to report such an increase as a percentage, and meteorologists were to forecast icing
hazards in the same manner, then much of the vagueness and subjectivity of the present
reporting system could be reduced.

In addition to being a simple and potentially effective means for the pilot to report icing, the
Percent Power Increase (PPI) can also be estimated from the model based icing forecast
mentioned above. PPI estimates are based on empirical/physical relationships between the three
meteorological variables and aircraft airfoil performance derived from wind tunnel data. PPI icing
forecasts are experimentally produced and published on the Aviation Weather Center

Comment: A down-to earth, common-sense approach to reporting icing intensity in PIREPs.
Although voice transmitted PIREPs are not expected to remain the communication of choice for
the next decade, power increase could easily be taken from an automated sensor and included in
the ADS messages or in the AMDAR code, particularly when these codes are changed to a
binary message in BUFR format.

                                                    Page 10
                                                                       TREND Newsletter No. 12

                Paper 4.10: Regional Icing Algorithm Performance Analysis

                             Tressa L. Kane, NCAR, Boulder, USA
                            Barbara G. Brown, and Ben C. Bernstein

Verification of in-flight icing forecasts previously has involved evaluations on either the national
scale or in the areas around large cities. However, forecast performance within specific regions
also is an important aspect of the evaluation of icing forecasts and diagnoses. In particular,
regional differences in performance may indicate changes or improvements that are needed to
accommodate the varying characteristics of the forecasts and icing situations. Moreover, these
evaluations may identify particular strengths and weaknesses in the forecasts. Performance of
the Integrated Icing Diagnostic Algorithm (IIDA) and icing AIRMET, the operational icing
forecasts issued by the National Weather Service's Aviation Weather Center, are evaluated over
15 regions across the continental United States. These regions were defined to cover the
AIRMET region while creating areas that are relatively homogeneous in terms of both climate and
topography. Naturally, the occurrence of icing varies among these regions. Moreover, the
frequency of pilot reports (PIREP) of icing is dependent both on the climate and on aircraft flight
patterns. The regional measures of overforecasting, the percent of total area and volume
covered by the forecast are normalized using the maximum expected icing extent, based on
numerical-model-based temperature and humidity analyses. Finally, a variety of IIDA thresholds
are evaluated to determine the extent to which the calibration of IIDA varies from region to region.
Results indicate that the performance of both IIDA and the AIRMET varies a great deal among
the regions.

Comment: Verification of icing algorithm or indeed forecast performance is just as tedious as it is
important for further development work on one hand and customer confidence on the other. The
strong dependence of verification results on the availability of PIREPs both positive and negative
make alternative measures such as area or volume afflicted by warnings necessary to determine
factors such as False Alarm Rate. The inclusion of climatological probability of icing conditions is
used to normalize this measure. Some progress can be seen in this respect, but the general
availability of objective verification data from sensors that are automatically transmitted will be a
prerequisite for fair and objective measures of quality in the long run.

                     Paper 4.11: Mixed-Phase In-flight Icing Conditions

                Marcia K. Politovich, NCAR, Boulder, USA, and James T. Riley

Recent investigations with modern instrumentation suggest that mixed-phase icing conditions are
more frequent and widespread than had previously been generally realized. A conservative
estimate is that an aircraft may be in mixed-phase clouds as much as 20% of the time that it is
operating in icing conditions. Substantially higher figures, 50-80%, may be appropriate in some
geographic areas.

Characterization of ice crystal and droplet size distributions in these regions is limited and are
needed to address questions of aircraft safety and remote sensing, and numerical and wind-
tunnel simulations. The data available from research flights do not suggest that there is any
difference in performance effects caused by structural icing resulting from flight in mixed-phase
cloud as compared to purely liquid cloud. However, these data cover a restricted range of
conditions and do not appear sufficient for generalization. If there is a difference in the nature of
icing in mixed-phase regions, such as the accretion rate, the texture and location of ice, this
needs to be accounted for in numerical accretion and performance models and in wind tunnels
used for icing research. Some methods have been used to create ice crystals in wind tunnels,
but generally natural conditions are simulated only with limited fidelity. Furthermore, it is

                                                    Page 11
                                                                       TREND Newsletter No. 12

sometimes difficult to assess the degree of fidelity necessary for the investigation of various
safety-related questions. The prevalence of mixed phase clouds poses a challenge to remote
sensing methods for icing detection. Radar reflectivity alone cannot clearly discriminate between
ice crystals and water drops, even if temperature is known. In general, ice crystals are typically
much larger than water drops, so that for short-wavelength cloud radars, non-Rayleigh scattering
effects can be a problem. Polarimetric techniques which enable identification of crystal habit or
detection of drizzle are in initial stages of testing for icing; their utility for general-use icing
detection is not known at this time. Knowledge of typical ice crystal habits, sizes, and
concentrations would help in development of techniques to deal with mixed-phase conditions.
These and other remote sensing problems are covered in the paper. Data obtained in mixed-
phase conditions, presented in a FAA-sponsored workshop in late 1998, are summarized. In
addition, calculations using representative ice and droplet concentrations and size distributions
are presented to show the temperatures and updrafts at which we can expect mixed phase
clouds to persist. To date, these calculations have been oversimplified with unrealistic ice crystal
concentrations and sizes. New calculations confirm that the conditions are indeed expected

Comment: The over-simplified concept of aircraft icing occurring mostly in purely supercooled
water clouds is ripe for the dustbin of science. It is being demonstrated that many processes
commonly found in the atmosphere (orographic uplifting, convergence lines, embedded
convection) are liable to maintain a continued presence of both supercooled droplets and ice
crystals in mixed clouds, and such conditions may be the rule rather than the exception. In-flight
measurements, however, will be needed to determine whether the high-risk supercooled large
droplet (SLD) environment is also commonly found in such mixed phase clouds.

               Paper 4.12: A Research Aircraft Verification of the Integrated
                             Icing Diagnostic Algorithm (IIDA)

                  Ben C. Bernstein, NCAR, Boulder, USA, Frank McDonough
                         Marcia K. Politovich, and Barbara G. Brown

The Integrated Icing Diagnostic Algorithm (IIDA) was designed to determine the potential for icing
and supercooled large droplets (SLD) to exist at locations across the contiguous United States
(CONUS) and southern Canada. It is currently employed by several regional airlines for dispatch
information regarding icing. Past verification exercises using icing pilot reports (PIREP) have
shown that IIDA is a relatively efficient algorithm, when compared to other automated icing
algorithms. While PIREP allow for evaluation of an algorithm's ability to identify general icing
conditions, they are difficult to use for verification of no-icing, SLD, and no-SLD
forecasts/diagnoses. Negative icing reports are relatively uncommon, and the assumption of "no-
icing" based on a lack of positive icing PIREP is not reliable. Verification of SLD/no-SLD aloft is
extremely difficult since PIREP do not contain information regarding droplet size, except for the
rare mention of FZDZ or FZRA in the weather or Comments fields.

During the winter of 1997-98, the NASA-Glenn Twin Otter research aircraft completed more than
50 hours of flight in a variety of conditions over the Great Lakes region. Numerous encounters
with FZDZ and FZRA are complemented by a large amount of data for conventional icing and no-
icing situations. The aircraft was well equipped for sampling and identifying these conditions,
and data collected by it provides reliable information for accurately determining the 3-D locations
of icing, no-icing, SLD, and no-SLD. In this paper, Twin Otter data are directly compared with
IIDA output to examine the algorithm's ability to identify SLD, no-SLD, and no-icing. This paper
provides the first absolute verification of diagnoses or forecasts SLD conditions aloft. The
technique applied is fairly strict, as only IIDA output for the adjacent horizontal grid points that
were within 1000 feet of the aircraft altitude. For each 3-D grid point, IIDA provides estimates of

                                                   Page 12
                                                                       TREND Newsletter No. 12

the potential for icing and SLD. This scaled, rather than binary, ability of the algorithm allows to
discriminate between higher and lower likelihood of icing and SLD. Results are very encouraging.
IIDA indicated high or moderate potentials for icing and SLD for most of the Twin Otter icing and
SLD encounters, respectively. IIDA also demonstrated capability for differentiating between "yes"
and "no" icing and SLD situations.

Comment: Avoiding the by now well-known problems of using PIREP for icing verification
studies, use of a research aircraft permits a fairer and much more detailed study in 3 dimensions.
The results show a very acceptable performance of the tested IIDA algorithm in icing conditions,
but we may be forgiven to assume that given the cost of operating a research aircraft, the sample
cases were tilted in favour of probable icing conditions, reproducing to a lesser extent the
difficulty of verifying false alarms.

         Paper 4.13: An Aircraft Flight Test Program in Natural Icing Conditions:
            Part 1 – Forecasting for the Desired Meteorological Conditions

       Wayne R. Sand, Aviation Weather Consulting, Boulder, USA, and Cleon J. Biter

An aircraft flight test program was conducted recently in natural icing conditions to test new
aircraft deicing boot operational procedures. Locating icing with a minimum amount of flight time
is critical to the success of such a program since flight test is extremely costly in terms of the
equipment and personnel required to properly execute such operations. Further, the tests often
require a certain type, or types, of icing to meet the demands of the test or certification flights.
Based on extensive flight experience and scientific research, the authors have developed
methodology to forecast and direct aircraft into specific types of icing conditions. An
understanding of the meteorological conditions which produce icing and the type of icing to be
expected in each condition is essential. The technique uses information readily available on the
Internet and telephone communications with the aircraft to direct them into the icing conditions in
real-time. Both maximum intermittent and maximum continuous conditions were successfully
located and penetrated using these techniques, while the total flight-test hours were minimized.
The flight test program consisted of 13 days with 36 icing encounters. The program was
completed ahead of schedule with all goals being realized. This paper discusses the specific
techniques used to forecast the desired icing conditions and to direct the flights into those

Comment: A refreshingly down-to earth approach to practical icing forecasts based on products
freely available on the Internet shows that even with limited resources out in the field a pretty
good job can be done in forecasting not only the presence, but also intensity and type of icing.
Makes excellent reading for the operational forecaster willing to improve.

         Paper 4.14: An Aircraft Flight Test Program in Natural Icing Conditions:
           Part 2 – Correspondence between the Forecast Icing Conditions,
                The Actual Icing Conditions and the Type and Amount of
                               Ice Accreted on the Aircraft

            Ralph Sorrells, Mitsubishi Heavy Industries America Inc., Dallas, USA,
                              Wayne R. Sand and Cleon J. Biter

“Part 1 – Forecasting for the Desired Meteorological Conditions” discussed the methodology
used to forecast and direct aircraft into desired natural icing conditions. This paper uses case
studies from the flight test program to illustrate the correspondence between the forecast icing
conditions, actual icing conditions and the airframe ice that collected on the test aircraft during
the icing flights. The test aircraft was instrumented with four high-quality video systems to record
the accretion of ice on various parts of the airframe as well as readings from the aircraft

                                                    Page 13
                                                                        TREND Newsletter No. 12

instrument panel. The videos show how airframe ice accretes differently under different
meteorological conditions. In addition, a data system recorded the status of an onboard ice
detector plus aircraft and engine parameters required for drag calculations. Using these data,
comparisons are made between meteorological conditions and the nature of the ice accreted on
the aircraft.

Comment: As for section 4.13.

              Paper 4.16: Integrated Icing Diagnostic Algorithm Assessment
                                   At Regional Airlines

                             Danny L. Sims, FAA, Atlantic City, USA,
                                 C. B. Fidalgo, and T. C. Carty

Aircraft icing is a major aviation safety hazard and has been linked to numerous aircraft
accidents. In an attempt to reduce the hazards to aviation, the Federal Aviation Administration
(FAA) Aviation Weather Research Program (AWRP) has sponsored research and development
activities aimed at providing timely and accurate icing detection and forecasts. Utilizing AWRP
funding, the Research Applications Program at the National Center for Atmospheric Research
(NCAR/RAP) has combined a number of different techniques into an Integrated Icing Diagnostic
Algorithm (IIDA) that makes use of the strengths of individual algorithms while minimizing their
respective weaknesses. The IIDA uses gridded data from the Rapid Update Cycle model;
GOES-8 satellite imagery; NEXRAD mosaics; and National Weather Service surface
observations to produce three-dimensional grids of Icing Potential, Supercooled Large Drop
(SLD) Potential, and Icing Type. In order to obtain operational feedback on IIDA, the FAA
William J. Hughes Technical Center (WJHTC) Communication/Navigation/Surveillance
Engineering and Test Division, Weather Branch (ACT-320) conducted an assessment of its use
by regional airlines. Dispatchers at Air Wisconsin and Atlantic Coast Airlines participated in the
assessment during the 1998-1999 winter season. The assessment focused on the utility and
perceived benefit of IIDA, along with suggested enhancements specific to dispatcher use.
Feedback was collected via an Internet-based user log, observations, interviews, and an Internet-
based questionnaire. Results of the assessment indicated IIDA was useful to airline dispatch
operations, and that dispatchers felt it was accurate in the identification of in-flight icing

Comment: A nice and reasonably promising attempt to repackage icing information from
different data sources into one meaningful, 3-dimensional parameter that is easily interpreted.
Results from user feedback indicate the expected: Although the Algorithm necessarily will not
provide much new information compared to its component factors, it is popular with users and
instills more confidence than individual parameters.

                     Session 5: Forecasting and Evaluation/Verification

                 Paper 5.1: Forecasts Aids to Lessen the Impact of Marine
                      Stratus on San Francisco International Airport

       F. Wesley Wilson, MIT Lincoln Laboratory, Lexington, USA, and David A. Clark

San Francisco International Airport (SFIA) is unable to use independent parallel approaches to its
closely-spaced parallel runways when marine stratus is present in the approach. Delay programs
are imposed to regulate the flow of traffic to match the true arrival capacity of the airport. Failure
to forecast accurately the times of onset and dissipation of stratus in the approach results in

                                                    Page 14
                                                                            TREND Newsletter No. 12

costly airborne holding and diversions, or in wasted capacity, as the traffic management planners
fail to match the arrival rate to the actual airport capacity. Four forecast algorithms have been
developed, which predict the time of dissipation of summer Marine Stratus at SFIA. These
algorithms are based on both dynamical and statistical analysis. In addition to operational data,
these algorithms use information from special sensors that have been installed for this project:
SODAR to measure the height of the inversion base, pyranometers to measure the intensity of
the solar radiation at the surface, and time series of 10m winds, temperature, and humidity. The
COBEL column model bases its forecast on explicit analysis of the evolution of the boundary
layer, with special attention to radiation and cloud water. The COBEL initialization relies heavily
on the special sensors. The other algorithms are based on statistical analysis. Each utilizes
different data features. The Local Statistical Forecast Model (LSFM) relies primarily on trends in
the project sensor data. The Regional Statistical Forecast Model (RSFM) relies on the standard
regional surface hourly observations. The Satellite Statistical Forecast Model (SSFM) relies on
trends in the regional GOES visible data. A consensus forecast algorithm is under development,
which will integrate the forecasts from the individual algorithms. These automated forecasts are
intended for use as guidance by operational forecasters. Performance statistics indicate that
each algorithm rivals the skill of the operational forecasts. An operational assessment is
planned, in which the forecasters will evaluate these models and their value in the preparation of
the operational forecasts.

Comment: A clearly defined problem with a limited number of independent variables really cries
out for an algorithm-based forecast guidance. Pure local column approaches, however, often
suffer from the lack of advective influence, which can be quite prominent in a land-sea-boundary
situation. Thus, the inclusion of a regional data set with all tendencies promises to give extra
value to the algorithms. Watch this space for the promised integration of all algorithms into a
guidance system.

         Paper 5.3: Steps to Improve Ceiling and Visibility Forecasts for Aviation

           James J. Gurka, NOAA/NESDIS, Suitland, USA, and Frederick R. Mosher

Delays to the U.S. air carrier system cost billions of dollars every year. The majority of these
delays are weather related, mostly due to thunderstorms and low ceilings and visibilities.
Recognizing that the impact of weather on the air carrier system can be mitigated by improved
terminal forecasts, the NWS has put a high priority on improving its service to the aviation
community. Among the goals documented in "Vision 2005": the National Weather Service
Strategic Plan for Weather, Water and Climate Services, 2000 - 2005, are: "ensure local airport
warnings for established criteria have a probability of detection of at least .80 and a false alarm
rate of .40 or less (2005)"; and "reduce the false alarm rate by 50 percent and increase the
probability of detection by 50 percent for critical ceiling (200 feet) and visibility (1/4 mile) forecasts
as contained in aviation terminal forecasts (2005)". While much effort has been given to
improving forecasts of thunderstorms, the forecasting of low ceilings and visibilities has
historically received relatively little attention. This is reflected in the verification scores in terminal
forecasts which have remained essentially unchanged over the last two decades. This paper will
focus on the steps necessary for the NWS to achieve its ambitious goals of improved ceiling and
visibility forecasts, including:

(1)     improved understanding of low cloud and fog processes;
(2)     improved mesoscale numerical models for low cloud forecasting;
(3)     improved the utilization of satellite data;
(4)     development of a system of algorithms similar to the "System for Convection Analysis
        and Nowcasting (SCAN) that will integrate all information required for forecasting ceilings
        and visibility;
(5)     improved use of climatological data in the aviation forecast process;

                                                       Page 15
                                                                        TREND Newsletter No. 12

(6)    initiation of an intensive program to educate all NWS field meteorologists in the
       forecasting of low clouds and fog;
(7)    provision of more focus on aviation forecasting in the NWS offices and the offer of
       incentives for improved performance.

Comment: The technological progress in low visibility operations for individual aircraft seems to
have been a major factor in the neglect of improving visibility and ceiling forecasts over the last
two decades. Whereas the improvements in NWP have been reflected very clearly in upper wind
and temperature forecasts, and high resolution regional models have had a positive impact on
the prediction of en route significant weather, the highly localized nature of low visibility /ceiling
conditions has prevented similar progress. The dramatic impact of low visibility/ceiling conditions
on the acceptance rate of hub airports is now generating a renewed interest in this field. The
article could be seen as a step towards recognizing the problem, but unfortunately also highlights
the lack of a detailed step-by-step strategy to date.

                   Paper 5.4: A Fuzzy Logic System for the Analysis and
                         Prediction of Cloud Ceiling and Visibility

                  A. Bruce Carmichael, NCAR, Boulder, USA, Kevin R. Petty,
                   Gerry M. Wiener, Melissa A. Petty, and Martha N. Limber

Low ceiling and poor visibility largely contribute to aviation safety and efficiency related problems.
In an effort to provide accurate and timely diagnoses of these atmospheric conditions, a method
is presented for the analysis and short-term forecast (0-6 hr) of cloud ceiling and visibility. Fuzzy
logic is used to develop an integrated algorithm that combines in situ, satellite, climatology and
numerical model data. The RUC (Rapid Update Cycle) and COAMPS (Coupled Ocean-
Atmospheric Mesoscale Prediction System) models serve as the primary numerical models in the
investigation, with model simulated, state of the atmosphere variables, being translated into cloud
ceiling and visibility fields. Operationally, forecasters currently examine data subjectively to reach
conclusions about the future state of ceiling and visibility, whereas the algorithm developed
herein uses these data to objectively analyze and forecast ceiling and visibility by means of
physically based parameters. Continued algorithm development focuses on implementing other
data sources that may prove beneficial in reducing ceiling and visibility forecast errors.

Comment: The use of fuzzy logic as so often is an indicator for the complexity of the problem at
hand. Whereas forecasters find themselves often in a situation where several sources of data
provide partly conflicting information and become cynics in the process, fuzzy logic attempts to let
statistics do the hard choice of what to believe. An interesting attempt.

                 Paper 5.5: Evaluation of Ceiling and Visibility Prediction:
               Preliminary Results over California Using the Navy’s Coupled
                Ocean/Atmosphere Mesocale Prediction System (COAMPS)

                     Daniel A Geiszler, SAIC, Monterey, USA, John Cook,
                   Paul Tag, W. Thompson, Richard Bankert, and J. Schmidt

A ceiling and visibility (C&V) algorithm developed at the National Center for Atmospheric
Research (NCAR) is examined during a six-month period using the Coupled Ocean Atmosphere
Modeling Prediction System (COAMPS). Consecutive 12-hour simulations were run using a triply
nested (81, 27, 9km) Lambert conic conformal projection over the California coast beginning in
July 1999 and ending 31 December 1999. During this time period, a qualitative comparison was

                                                    Page 16
                                                                      TREND Newsletter No. 12

made between visible satellite imagery and the ceiling height forecasts of the highest resolution
nest at the 6th (18Z) and 12th (00Z) forecast hours. A quantitative comparison was made with 3
hourly METAR at Monterey (KMRY), San Francisco (KSFO), Los Angeles (KLAX), San Luis
Obispo (KSBP), and Bakersfield (KBFL). The temperature, dewpoint, wind speed, wind direction,
ceiling height, and visibility of each METAR were compared with the nearest grid point in the 9
km grid. COAMPS and the C&V algorithms were found to compare favorably with the visible
satellite imagery. In many cases, the structure of the coastal stratus was well represented along
the coast, and the periodicity of the cloud coverage was well captured by the model. In contrast,
statistical comparisons with the individual stations suggest that the model and algorithms
performed very poorly. Occurrences of low ceiling and low visibility are underpredicted by the
algorithm at KMRY, KSBP, and KLAX and overpredicted at KSFO. An examination of the grid
points around these stations, however, shows that in over 50% of the cases in which the model
"misses" predicting low ceiling heights, it misses by only a couple grid points (< 20 km). Other
possible explanations for the inaccurate forecasts include a lack of aerosol information in the
C&V algorithms, terrain misrepresentation within the model where the coastal mountain ranges
are smoothed beyond the shoreline, and interpretation of the Automated Surface Observing
System (ASOS) ceiling measurements.

Comment: The stark difference between a large-scale, qualitative assessment of marine stratus
presence and the detailed analysis of conditions at individual sites is a lesson in verification
ground truths: In aviation meteorology it is simply not enough to predict marine stratus, but
whether it will cover coastal airports or not is the make-or-break question.

           Paper 5.6: AWIPS ERA Real Time TAF, FWC, and Lamp Ceiling and
                    Visibility Verification Programme, at WFO, Tulsa

             James M. Frederick, NOAA/NWS, Tulsa, USA, and Steve A. Amburn

An overview of the Terminal Aerodrome Forecast (TAF) verification program that was developed
at the National Weather Service Weather Forecast Office in Tulsa was presented. TAF
prevailing forecast data and the associated FWC and LAMP ceiling and visibility forecasts are
gathered for each routinely issued TAF by software on AWIPS (Advanced Weather Interactive
Processing System). The verification is done using the FWC/LAMP ceiling and visibility
categories. The data are updated every six hours and are available to be viewed by forecasters.
Numerous statistical verification are available, including frequency distribution, bias, absolute
errors, probability of detection, false alarm rate and critical success index. The data can be
sorted by cycle time or six-hour forecast periods, and can be viewed in tabular or graphical form.
The graphical user interface, which allows users to easily follow FWC and LAMP trends and
gives forecasters immediate feedback on their TAF, was presented.

Comment: A beautiful example of a small office developing a sensible, simple-to-use method to
verify its TAFs. Although not all problems were solved at once, it is encouraging to see that local
initiatives can contribute a lot to quality awareness. The advantage of doing verification locally
under self-developed rules is that acceptance of the method by staff is virtually guaranteed and a
positive influence on quality thus much more likely than from any imposed system.

                  Paper 5.7: TAF Verification: Performance Measurement
                                 Or Quality Improvement

                Kent A. Johnson, MSC, Kellwood, Canada, and Uwe Gramann

For decades, aviation aerodrome forecast (TAF) verification systems have been used to provide

a quantitative representation of forecast quality. These numbers were often confusing and

                                                   Page 17
                                                                        TREND Newsletter No. 12

found, given the nature of weather patterns, to be highly variable in both space and time.
Although verification scores can provide an indication of forecast quality, efforts to use such
results as an impetus for TAF improvement were limited. In order for TAF verification scores to
be used in a quality improvement mode, several conditions must be met. First, aviation
meteorologists must have a rudimentary understanding of the objective scoring system. Second,
near real-time statistics must be available to operational aviation forecasters, in order to assess
past performance and consider cases where substantial improvement could be made. A PC-
based verification system has been developed with two distinct components. An overall
database allows for generation of statistical measures such as False Alarm Ratio (FAR),
Frequency of Hits (FOH) and Probability of Detection (POD). A training database provides
forecasters the opportunity to write several TAFs for the same situation and study the differences
in verification scores. Both components of the verification system were described and
demonstrated. Finally, it is suggested that, through effective use, within an operational setting, a
simple verification system can effectively improve TAF quality.

Comment: Similar to section 5.7 above, another refreshingly simple and straightforward method.

           Paper 5.8: An Observation-Based, Statistical System for Short-Term
            Probabilistic Forecasts of Aviation Sensitive Weather Parameters

     Joby L. Hilliker, Penn State University, University Park, USA, and J. Michael Fritsch

Accurate and timely short-term (< 6 h) forecasts of high-impact aviation weather parameters,
such as low ceiling, poor visibility, and thunderstorms are a critical component to airline
operations. Adverse weather conditions not only plague the efficiency of commercial operations,
but are also a safety concern, particularly for general aviation pilots and passengers. A review of
weather-related Accident Briefs indicates that the vast majority of fatal accidents (over 80%) is
caused by low ceiling, poor visibility, and thunderstorms. Moreover, the number of aircraft delays
has increased markedly – up over 60% in September 1999 from a year earlier. Clearly, improved
weather guidance would be of great value to the aviation industry, especially if the guidance
provided a measure of the uncertainty associated with the forecasts. Traffic-flow-management
personnel recognize the inherent uncertainty when predicting weather, and attempt to account for
it through careful cost-benefit decision-making – all in an effort to minimize the airlines’ operating
costs. The uncertainties incorporated into cost-benefit analyses can best be captured, then, via
reliable probabilistic forecast guidance. This suggests that short-term, observations-based,
statistical forecast techniques that yield a quantitative measure of uncertainty would be most
useful to the aviation industry, especially if such techniques would provide sharp, timely, and
reliable probabilistic forecasts of high-impact aviation weather parameters. A framework for an
automated system that provides probabilistic guidance for aviation weather parameters is
presented. Preliminary results from early versions of the automated system are also presented.
The system utilizes NEXRAD Information Dissemination Service (NIDS) radar data, surface
mesonet data, and Velocity Azimuth Display (VAD) wind information as input to a stepwise
regression procedure. Output consists of probabilistic categorical forecasts for lead times of 6,
12, 18, 30, and 60 min. The system is capable of providing updates every 6 min., thereby
accounting for rapid changes in local conditions. Forecasts are made for a selected airport and
for an array of grid points centered on the selected airport. The array covers a domain
comparable to the approach-control zone around the selected airport (i.e., the area within a
radius of about 60 km. Of the airport). Thus, the system is designed to provide probabilistic
categorical forecasts of aviation-sensitive weather parameters at the airport as well as the spatial
distribution of probabilities of certain weather categories (such as the presence of thunderstorms)
in the region surrounding the airport. In this manner, air-traffic controllers can obtain high-
resolution, rapidly updated guidance for improving air-traffic flow. It is expected that this
observations-based system will provide an advancement in two aspects of aviation:

                                                    Page 18
                                                                       TREND Newsletter No. 12

1) reliable, ultra-short-term probabilistic forecasts will ameliorate air-traffic congestion during
   adverse conditions, thereby reducing airline costs and passenger complaints, and
2) short-term guidance on dangerous weather elements in the near vicinity of airports can
   diminish the risk of General Aviation accidents.

Comment: A thorough analysis of the problem at hand is followed by an outline of how it could
be done. The proof of the pudding will be the eating.

        Paper 5.11: Short Term Forecasting of Snowbands Using Doppler Radar
                         Observations and Cloud-Scale Model

                         Mei Xu, NCAR, Boulder, USA, Juanzhen Sun,
                           N. Andrew Crook, and Roy Rasmussen

The accuracy of extrapolation technique for forecasting snowfall decreases rapidly after 30
minutes, due to the effect of storm evolution. In order to extend the predictability of phenomena
such as snowbands and freezing rain events, it is necessary to use a numerical model that
simulates storm evolution. In this study we use a cloud-scale model to forecast a winter storm
which produced heavy snow in the New York City area on December 10, 1997. Well-defined
snowband structures were the dominating feature of the storm. To initialize the model, Doppler
radar observations are used. Radar reflectivity and radial velocity are assimilated into the model
using the adjoint technique. The assimilation system, which includes a cloud-scale model and its
adjoints, determines the 3D wind, thermodynamical, and microphysical fields of the storm by
minimizing the difference between radar observed variables and their model predictions.

Preliminary results show that using two volume scans of radar observations separated by 6
minutes, the retrieval is able to recover the band structure of the storm. The wind fields are
reasonable and fit relatively well to the observed radial velocity. One-hour forecasts of this event
have been performed using the retrieved fields as initial conditions. Results show that the model
simulates the motion of the snowband reasonably well. Work is currently being done to improve
the microphysical representation in the model to allow a better quantitative snowfall forecast. We
will also use the output from a mesoscale model to provide boundary conditions for the cloud-
scale model so that the forecast can be extended beyond one hour.

Comment: One of the first real attempts at ingesting Weather Radar Data in a truly mesoscale
model. The complexity of using the adjoint technique, storm-scale microphysics representation
and the feedback on the dynamics of the storm movement will probably require more work, but
appear to hold considerable promise.

                  Paper 5.12: FAA Terminal Convective Weather Forecast
                                 Algorithm Assessment

                   K. E. Theriault, MIT, Lexington, USA, Marilyn M. Wolfson,
                           Barbara E. Forman, Robert G. Hallowell,
                         Michael P. Moore, and Richard J. Johnson, Jr.

Predicting convective weather is extremely important to aviation, since thunderstorms are the
largest cause of national airspace delay. Accurate 1-hour convective weather forecasts meet
critical terminal traffic planning needs of the TRACON and ARTCC supervisors and traffic
managers, as well as airline dispatchers and pilots. The Terminal Convective Weather Forecast
(TCWF) product has been developed by MIT Lincoln Laboratory as part of the FAA Aviation

Weather Research Convective Weather Product Development Team (PDT). Lincoln began by
consulting with air traffic personnel and commercial airline dispatchers to determine the needs of

                                                   Page 19
                                                                       TREND Newsletter No. 12

aviation users. Users indicated that convective weather, particularly line storms, caused the most
consistent problems for managing air traffic. The "Growth and Decay Storm Tracker" developed
by Wolfson et al. (1999) allows generation up to 1-hr forecasts of large scale, organized
precipitation features with good, and operationally very useful accuracy.

The TCWF has been tested in Dallas since 1998, in Orlando since 1999, and in New York since
FY2000 began. These have been informal demonstrations, with the FAA Technical Center
assessing utility to the users, and with MIT LL modifying the system based on user feedback and
performance analyses. The TCWF is now ready for a formal Assessment. Ultimately, the FAA
Technical Center will judge whether TCWF is suitable for inclusion in the FAA's operational
Integrated Terminal Weather System, which has an unmet requirement for 30+ min forecasts of
convective weather. The Memphis International Airport has been selected for the TCWF
Assessment, since it has not been exposed to the forecast product during prior demonstrations.
Operations were said to begin in early spring, 2000. Operational feedback would be assessed by
the FAA Technical Center and MCR Corp. that would perform a quantitative benefits
assessment. This paper describes the current TCWF algorithm, gives examples of operational
impact of terminal forecasts, and analyzes the technical performance of TCWF during the
Assessment in Memphis.

Comments: An extrapolation-based, spatially refined algorithm delivers useful results for steadily
moving, larger convective or frontally organized precipitation bands. All the usual limitations
(growth, decay, reformation, etc) apply.

                  Paper 5.13: FAA Terminal Convective Weather Forecast
                                    Benefits Analysis

                      Jim S. Sunderlin, MCR Federal, McLean, USA, and
                           Gary Paull, MCR Federal, Bedford, USA

Thunderstorms are a major cause of disruptions to air traffic at airports causing inefficiencies for
both airlines and passengers. Improvements in convective weather forecasting provide traffic
managers with the ability to better anticipate thunderstorm impacts on airspace availability, thus,
reducing ground delays, airborne delays, diversions, and cancellations. The Terminal
Convective Weather Forecast (TCWF) product has been developed by MIT Lincoln Laboratory
as part of the FAA Aviation Weather Research Program. TCWF is intended to improve
convective weather forecasts, resulting in efficiency benefits. Currently, the product is being
demonstrated at airports in Orlando, Dallas Fort-Worth, Memphis and New York. MCR has
conducted interviews with Tracon and ARTCC Traffic Management Unit (TMU) personnel for
both the Orlando and Dallas Fort-Worth airports. These discussions address specific
experiences using TCWF during thunderstorm activity and provide the basis for assessing
benefits. Careful understanding and structuring of these efficiency benefits is necessary to
develop an accurate framework for quantifying the savings achieved by using the product. A
thorough analysis of validated inputs from multiple experts provides credible estimates of the
benefits. It is imperative that the estimated savings be compared to historical flight disruption
data associated with thunderstorms to provide a context for proper interpretation.

This report describes the primary scenarios in which TCWF has assisted traffic managers in
addressing thunderstorm impacts, benefit estimates based on specific efficiencies identified by
the users, and a comparison of the results to historical flight disruption data.

                                                   Page 20
                                                                      TREND Newsletter No. 12

Comment: See Section 5.12 above.

                   Paper 5.14: Improvement of Terminal Area Forecasts

     C. Pan, Donna Tucker and David A. Braaten, University of Kansas, Lawrence, USA,
                   Peter A. Browning, NOAA/NWS, Pleasant Hill, USA,
                            Israel Jirak, and David Beusterien

An effort to improve the quality of terminal area forecasts (TAFs) at the Kansas City International
Airport (MCI) is undertaken by the University of Kansas and the NWS forecast office in Pleasant
Hill, Missouri. In consultation with forecasters at Pleasant Hill, a list of critical weather
parameters in making TAFs has been determined. A multilinear regression technique is then
applied to produce 1,3 and 6 hour forecasts of visibility and ceiling under instrument flight rule
(IFR) and low instrument flight rule (LIFR) conditions. A fuzzy logic model technique using
internal functions is also being applied to the same dataset. Forecast results from these two
techniques are compared. TAFs for MCI issued by NWS forecasters have been verified for the
period January 1995 through December 1999. This period represents observations exclusively
from the Automated Surface Observing System (ASOS). Statistical measures of forecast
performance from this period will be used as a benchmark for measuring the skill of the new
objective TAF forecast system. The ultimate goal of this research is to improve the flight safety
with the improved TAFs.

Comment: Yet another group of hopeful developers running into the usual problems of MLR in
Ceiling and Visibility forecasts: small datasets combined with a low correlation between upper-air
predictors and the predictands make life not very easy for them.

             Paper 5.15: Wind and Temperature Verification Statistics for the
                Operational Terminal Area PBL Prediction System at the
                         Dallas-Fort Worth International Airport

         J. J. Charney, North Carolina State University, Raleigh, USA, M. L. Kaplan,
              Y.-L. Lin, K. T. Waight, K. D. Pfeiffer, J. A. Thurman, and C. M. Hill

This paper addresses the multi-month statistical accuracy of the Terminal Area PBL Prediction
System (TAPPS). TAPPS is being run operationally in support of the Aircraft Vortex Spacing
System (AVOSS) demonstration at the Dallas-Fort Worth (DFW) International Airport under the
auspices of the NASA-Langley Research Center. AVOSS is an element of NCI’s Terminal Area
Productivity Program (TAP) that is designed to incorporate wake vortex observations, wake
vortex decay and transport algorithms, the observed and predicted weather state, and system
integration to determine the minimum safe spacing for departing and arriving aircraft in order to
avoid unsafe encounters with aircraft wake vortices. A concept demonstration of the system is
planned for Dallas-Ft. Worth (DFW) airport in 2000. The motivation behind the development of
this system is the need for increased airport capacity while maintaining the present level of
safety. Means of safely increasing airport capacity is a critical issue as the number of flights to
and from U.S. airports increases substantially in the near future.

The centerpiece of TAPPS is a mesoscale numerical weather prediction model that is run twice
daily over a grid mesh centered on the DFW International Airport. TAPPS (version #1) is
integrated over a matrix of 60X60X56 grid points and essentially represents Version 5.13 of the
Mesoscale Atmospheric Simulation System (MASS). This mesoscale numerical weather

                                                   Page 21
                                                                         TREND Newsletter No. 12

 prediction model employs the Blackadar PBL parameterization scheme, Kuo-MESO convective
parameterization scheme, a comprehensive soil moisture hydrology and land use
parameterization scheme, as well as very detailed vertical resolution within the lower part of the
PBL. The highest resolution version of the model utilizes a 12 km horizontal fine grid mesh
nested within a 24 km horizontal coarse grid mesh. The coarse grid mesh is initialized from NWS
ETA analyses data sets and integrated for 24 hours of real-time on a DEC-ALPHA 600
workstation. The fine grid mesh is then integrated for 21 hours, starting 3 hours after the
observation time, and using the coarse grid mesh results as lateral boundary conditions.
Routinely available products include 15 m vertical resolution soundings of the cross-runway wind
component, along-runway wind component, virtual potential temperature, eddy dissipation rate,
turbulent kinetic energy, and wind variance up to ~1005 m elevation. Also available are time-
height sections of these same fields based on 30-minute interval simulated model results.

Observations from the DFW PBL profiler, RASS, and MIT Lincoln Laboratory AWAS profiles are
compared to the TAPPS #1 simulations for a multi-month time period. Error statistics, including
the mean absolute error, RMS error, and bias of the wind velocity components and temperatures
are discussed in terms of their vertical structure as well as their sensitivity to model initialization
time, to the diurnal cycle, and to different synoptic weather regimes. Comparisons are also made
against a smaller sample of verification statistics from the second generation of the operational
TAPPS (version #2), which includes improved initial soil moisture, improved initial winds, and ~6
km horizontal resolution simulations.

Comment: A laboured approach to a problem of increasing significance: Wake vortex decay
forecasting in order to reduce approach spacing done with a complex forecasting suite. It will be
interesting to see whether significant improvements over the existing information from operational
models can be achieved.

          Paper 5.16: An Evaluation of Using Lightning Data to Improve Aviation
                 Oceanic Convective Forecasting for the Gulf of Mexico

                Alan Nierow, FAA, Washington, USA, and Robert C. Showalter

The National Weather Service (NWS) Aviation Weather Center (AWC) developed a product that
overlays "long-range" lightning data onto Infra-Red (IR) satellite imagery. These lightning data
are an extension of the National Lightning Detection Network (NLDN) developed and operated by
Global Atmospherics, Inc. (GAI), and the Canadian Lightning Detection Network (CLDN) owned
by Environment Canada. The Federal Aviation Administration (FAA) and the NWS promoted an
evaluation of this product under the National Lightning Contract between the NWS and GAI, with
participation from government (NWS, FAA, and Department of Defense (DoD)) as well as non-
Government users (airline meteorologists and dispatchers) from April 1999 to January 2000.

Given that satellite imagery alone cannot always detect the presence of convection activity
underneath the clouds, the primary purpose of the evaluation was to determine if merging it with
'long-range' lightning data would aid aviation forecasters/dispatchers in identifying regions of
oceanic convection. The results of the evaluation indicated that CWSU meteorologists and other
Air Route Traffic Control Center (ARTCC) personnel, as well as airline dispatchers, found this
product beneficial in monitoring convective storm development over the Gulf of Mexico (GOM)
and the Caribbean. AWC meteorologists have used this lightning product for some time now and
found it quite useful in issuing International Significant Meteorological Information (SIGMETs) for
convection, and in routing aircraft away from hazardous thunderstorms over oceanic regions.

The evaluation indicated that there was a large difference in long-range NLDN/CLDN
thunderstorm detection efficiency between day and night due to diurnal influences on the

                                                     Page 22
                                                                       TREND Newsletter No. 12

ionosphere. For convective forecasting, thunderstorm detection is most important/critical for
flight planning and route deviation. This 'long-range' lightning product had an excellent
thunderstorm detection efficiency of close to 100% within 1000 km of CONUS (day/night), with
only a slight decrease to 70-80% within 3500 km (night) and 2000 km (day) of CONUS. The FAA
is endorsing a multifaceted program centered on the requirement to improve air traffic control
services in the offshore airspace of the Houston ARTCC and the U. S. controlled Oceanic Flight
Information Region (FIR) in the GOM region. Initial requirements regarding weather will address
shortfalls in oceanic convective observations and forecasting. Air traffic in the GOM region will
increase significantly in the next decade. To efficiently deal with the increase in capacity, it is
likely that the horizontal separation normally used between aircraft flying over oceanic regions will
be reduced. Accommodating this increase in GOM growth and capacity would dictate
improvements in the availability of current and forecast weather data as well.

The Oceanic Convective product has the potential to play a significant role in supporting
increased airspace capacity in the GOM region without degrading safety. Furthermore, it could
reduce the incidences of non-coordinated deviations due to weather through greater situational
awareness. This product will advance collaborative decision making processes, especially over
oceanic regions where surveillance and communications are not optimal. An assessment of this
Oceanic Lightning product over the GOM should clearly establish if the value of this product is
sufficient to:
1) develop a need for routine access to this product for the GOM region, and
2) expand the network to enhance data availability over the Caribbean, Central America, and
    South America with improved detection and location accuracy.

Comment: Radar networking is never easy or cheap, but fails totally offshore. Lightning
detection is simple, cheap and if done correctly can go a long way towards achieving reasonably
accurate identification and localization of active thunderstorms. Its limits are likely to be the use
for extrapolation and forecasting.

         Paper 5.17: Verification of Icing and Turbulence Forecasts: Why Some
               Verification Statistics Can’t Be Computed Using PIREPS

                   Greg S. Young, Barbara G. Brown, NCAR, Boulder, USA

Pilot reports (PIREPs) are routinely used to verify forecasts of icing and turbulence conditions,
including both operational and experimental forecasts. It is well known that PIREPs have
numerous characteristics that make them a difficult data source to use for verification. Among
these characteristics are the subjective nature of the reports and their spatial and temporal

The non-systematic nature of PIREPs creates the greatest difficulty for verification; because
observations are not consistently available at the same time and location, the sample of reports
does not provide a representative sample of the forecast grid. Impacts of this characteristic have
been taken into consideration in most recent studies of the quality of icing and turbulence
forecasts. In particular, most studies have limited the statistics computed, and have not
computed the False Alarm Ratio (FAR) and other statistics that require stratifying the data by the
forecast type. However, questions frequently are raised regarding why it is inappropriate to
compute these statistics using PIREPs. The purpose of this paper is to clearly demonstrate why
FAR and other statistics should not be computed when PIREPs are used as verifying
observations for forecasts across a grid or large domain. In addition, some examples of the
impacts of computing these statistics are presented.

The primary analytical result of this study is a simple demonstration that FAR is strongly related
to the relative frequencies of Yes and No PIREPs. Thus, when either the number of Yes or No

                                                    Page 23
                                                                          TREND Newsletter No. 12

PIREPs is changed, the FAR also changes. In contrast, other statistics (based on stratifying by
the observation type), such as PODy and PODn, change very little when the numbers of PIREPs
change. This effect is demonstrated for turbulence by supplementing the No observations with
AVAR observations and for icing by supplementing the no-icing PIREPs with PIREPs reporting
“clear above.” In both cases, the FAR estimate increases dramatically. These results are
supplemented by a simple simulation study, in which PIREPs are randomly eliminated from a
verification analysis. Results of these simulations suggest that when the PIREPs to be removed
are randomly selected from both the “YES” and “NO” PIREP subsets, there is little effect on any
of the verification statistics. However, when the eliminated PIREPs are selected from either the
“YES” or “NO” subsamples only, the FAR and other statistics are affected. Other statistics that
are impacted include the Bias and Critical Success Index, and the Heidke and Gilbert skill

The results of this study firmly demonstrate the limitations of verification statistics computed using
PIREPs. In addition, the study emphasizes the value that could be attained through systematic
collection of pilot reports.

Comment: Not too surprising, but certainly illustrating the fundamental problems of verifying
forecasts for events that are rare both in space and time, that are influencing decisions on the
availability of proof or otherwise (Pilots are less likely to fly into regions of predicted severe icing
or turbulence if they can). The nature of the phenomena to be verified influences not only their
predictability, but also their likelihood of detection. Automated reports such as AMDAR will
improve some aspects of the problem (reporting becomes more regular), but will not change
others (pilots still unhappy to fly into affected regions).

                             Session 7: Turbulence and Windshear

   Paper 7.1: Aircraft Encounters with Mountain Wave Induced Clear Air Turbulence:
         Hindcasts and Operational Forecast Using an Improved Global Model

               Stephen D. Eckermann, NRL, Washington, USA, Dave Broutman,
                                 and Julio T. Bacmeister

The first stage of a major upgrade to an operational mountain wave forecasting model, which we
refer to as MWFM was recently completed. The MWFM uses global forecast data to assess flow
patterns over the Earth's major topographical features, in order to locate regions where mountain
waves are generated. The new model (MWFM 2.0) uses numerical ray-tracing algorithms to
predict the subsequent radiation of three-dimensional nonhydrostatic mountain wave patterns
from the parent orography. While a major extension of the old, the new code is nonetheless
numerically efficient, allowing to forecast regional or even global mountain wave distributions
soon after global forecast data come on-line.

MWFM forecasts of turbulence due to mountain wave breaking are discussed. A brief
description of the old and new calculation methods is provided. The model was used previously
in research campaigns using NASA's ER-2 research aircraft. This aircraft (a modified U-2)
cruises at ~60,000-70,000 feet and, due to weight constraints, does not have a wing spur, which
makes it structurally vulnerable to turbulence. Though much of the weather-related turbulence
found lower down is absent at these stratospheric cruise altitudes, in-flight ER-2 data and MWFM
forecasts and hindcasts were used to demonstrate that breaking mountain waves are major
source of hazardous turbulence for the NASA's ER-2 and DC-8 research aircraft were deployed
from the Arena Arctica in Kiruna, Sweden, during the SAGE III Ozone Loss and Validation
Experiment (SOLVE), which took place during the winter of 1999-2000. Kiruna lies downstream
of large mountains that run the length of the Norwegian coast. Furthermore, research flights

                                                      Page 24
                                                                       TREND Newsletter No. 12

were also possible over southern Norway, Scotland, the Urals, Iceland, Greenland, Novaya
Zemlya, Franz-Josef Land and Spitzbergen, all regions with significant orography and hence
mountain wave turbulence potential. These facts, along with the hostile remote Arctic
environment, meant that mountain wave-induced turbulence was a critical safety for ER-2 flight
planning. Accordingly, we ran the new MWFM model in an operational forecasting configuration
throughout the SOLVE mission, and ran tailored forecasts in the field during the ER-2
deployments. Some selected results from the MWFM-SOLVE forecasting effort are presented.

Though less mission-critical, MWFM turbulence forecasts were also provided for the DC-8 at
altitudes ~200-250 hPa (~35,000 feet). he potential utility of these lower-altitude MWFM
turbulence forecasts for commercial aviation were investigated. Two NTSB reports of airliner
encounters with unforecasted severe turbulence were focused on, one near Yakutat, Alaska, the
other near Bishop, California, both of which resulted in crew injuries. MWFM hindcasts produce
zones of intense mountain wave-induced turbulence near the reported locations and altitudes of
these incidents.

Comment: Experience from field experiments (MAP) clearly shows a potential for high-resolution
non-hydrostatic models to come up with something resembling TKE. The big problem appears to
be a tendency of always producing these results regardless whether anything happens at all.

            Paper 7.3: Preliminary Results of the NCAR Integrated Turbulence
                      Forecasting Algorithm (IFTA) to Forecast CAT

               R. Sharman, NCAR, Boulder, USA, B. G. Brown, and S. Dettling

Accurate automated one to twelve hour forecasts of turbulence in a three-dimensional airspace
are needed by the commercial and general aviation community, but so far, these forecasts have
not generally provided acceptably high detection rates and at the same time acceptably low false
alarm rates, especially in clear air at upper levels. This is due in part to lack of observations at
the (relatively small) atmospheric scales that affect aircraft motion (microscale), and to
uncertainties in our knowledge about the relationship between the meteorologically observable
(relatively large) scale and the microscale. Until we have a better understanding of the linkage
between the large scale and the microscale, we are forced to use semi-empirical techniques to
diagnose the large scales for potential turbulence. It is shown that the discrimination capability of
these techniques by themselves is in general poor, but a combination performs better. This
paper describes an automated turbulence forecasting procedure (the Integrated Turbulence
Forecasting Algorithm, ITFA) whereby several turbulence diagnostics are fit to available
observations to produce the forecast. Intense verification exercises have been performed over
two winter seasons in which probabilities of yes and no detections were determined by
comparisons to observations in the form of pilot reports. The sparseness and qualitative nature
of this data produces some unavoidable uncertainty in the verification results, however,
preliminary results indicate the automated algorithms our competitive with turbulence forecasts
produced by knowledgeable human forecasters.

Comment: As for fuzzy-logic systems, there is the lingering doubt that combining a number of
algorithms all suffering from basic uncertainties in the models they are applied to (mostly due to a
lack of vertical resolution with respect to the scale of the turbulence-producing shear zones) will
be limited in scope, and may benefit mostly from statistical effects (hedging, elimination of biases

             Paper 7.4: The Turbulence Algorithm Intercomparison Exercise:

                                                    Page 25
                                                                      TREND Newsletter No. 12

                                Statistical Verification Results

            Barbara G. Brown, NCAR, Boulder, USA, Jennifer Luppens Mahoney,
             Judy Henderson, Tressa L. Kane, Randy Bullock, and Joan E. Hart

During the winter of 1998-99, forecasts of clear-air turbulence produced by a large number of
turbulence indices were compared. The 14 algorithms considered in the study include a number
of algorithms that have been available for many years, as well as algorithms that are newly under
development. The algorithm forecasts were based on output of the RUC-2 numerical weather
prediction model for the period 21 December 1998 to 31 March 1999. Forecasts issued at 1200,
1500, and 1800 UTC, with 3-, 6-, and 9-hr lead times were included in the study. Turbulence
AIRMETs, the operational turbulence forecast product that is issued by the NWS’s Aviation
Weather Center (AWC), also were included in the evaluation. The evaluation was limited to the
continental United States and to altitudes above 20,000 ft.

The forecasts were verified using “YES” and “NO” turbulence observations from pilot reports
(PIREPs), as well as “NO” observations based on automated vertical accelerometer (AVAR) data
that were obtained from a number of aircraft. The algorithms were evaluated as “YES/NO”
turbulence forecasts by applying a threshold to convert the output of each algorithm to a “YES” or
“NO” value. A variety of thresholds were applied to each algorithm. The verification analyses
were primarily based on the algorithms’ ability to discriminate between “YES” and “NO”
observations, as well as the extent of their coverage.

The study was comprised of two components. First, the algorithms were evaluated in near-real-
time by the Real-Time Verification System (RTVS) of the NOAA Forecast Systems Laboratory
(FSL),     with    results   displayed     on      the  World-Wide        Web      (http://www- Second, the verification results were re-
evaluated in depth in post-analysis, using a post-analysis verification system at the National
Center for Atmospheric Research (NCAR), with additional thresholds applied to each algorithm to
provide a thorough depiction of algorithm quality.

Results of the intercomparison suggest that some algorithms perform somewhat better than
others. In particular, these algorithms have somewhat larger values of the True Skill Statistic for
comparable thresholds, and they have a slightly larger overall discrimination skill statistic.
However, the best algorithms have very similar performance characteristics. In some (but not all)
cases the algorithm performance is slightly better than the performance of the AIRMETs. Results
of the study also suggest that further algorithm development is needed before newer algorithms
will show large improvements over some of the older algorithms. Moreover, algorithms like
Integrated Turbulence Forecasting Algorithm (ITFA) may benefit by not including some
algorithms that have relatively little forecasting skill.

A second algorithm intercomparison is in progress for winter 1999-2000. Available results of this
second evaluation will also be reported and compared to the winter 1998-1999 results.

Comment: This paper serves to support the comments on section 7.3.

                 Paper 7.5: Integrated Turbulence Forecasting Algorithm
                                Meteorological Evaluation

                                                   Page 26
                                                                    TREND Newsletter No. 12


    Victor S. Passetti, Basic Commerce and Industries, Inc. and FAA, Atlantic City, USA,
                         D. L. Simms, T. C. Carty, and J. A. Weinrich

Encounters with unexpected, hazardous, clear-air turbulence (CAT) pose significant safety risks
to the aviation community. To mitigate these safety hazards, the Federal Aviation Administration
(FAA) Aviation Weather Research Program (AWRP) has sponsored research and development
activities aimed at improving the detection and forecasting of CAT. Utilizing AWRP funding, the
Research Applications Program at the National Center for Atmospheric Research (NCAR/RAP)
has developed the Integrated Turbulence Forecasting Algorithm (ITFA). ITFA is composed of
numerous CAT indices that are calculated using the forecasted fields of the Rapid Update Cycle
(RUC) and assigned weighting factors based on turbulence observations obtained from pilot
reports (PIREPs) and aircraft vertical accelerometer measurements.

The output and weighting factors are integrated, with the resulting final output displayed as
graphical forecasts of CAT.

In order to gain feedback about the operational effectiveness of the ITFA, the FAA William J.
Hughes Technical Center (WJHTC) Communication/Navigation/Surveillance Engineering and
Test Division, Weather Branch (ACT-320), performed an event-driven meteorological evaluation
of the ITFA during the period 1 January to 30 April 2000. The evaluation involved using
Significant Meteorological Information advisories (SIGMETs) and PIREPs to identify regions of
significant CAT, and then correlating these turbulent regions with the output produced by the
ITFA before, during, and after the CAT events. Results of the evaluation identified the added
value of the ITFA products to the aviation forecasting community and provided NCAR valuable
insight for future development of the algorithm.

Comment: See Section 7.3.

        Paper 7.6: Obervational and Numerical Simulation-Derived Factors that
                   Characterize Turbulence Accident Environments

              Michael L. Kaplan, North Carolina State University, Raleigh, USA,
                   Y.-L. Lin, A. J. Riordan, K. M. Lux, and A. W. Huffman

Dual approaches are employed to determine the synoptic, meso-alpha, and meso-beta scale
dynamical factors that organize an environment conducive to severe turbulence that affects
commercial aircraft. The first approach involves performing a 44-case-study turbulence
categorization analysis. 2.5 degree 6-hourly data sets of NCEP Reanalyses including wind,
temperature, relative humidity and height on mandatory pressure surfaces along with high
frequency GOES infra-red and visible satellite imagery are employed to perform diagnostic
analyses of 44 case studies where turbulent conditions lead to accidents on commercial aircraft.
The analyses indicated that a synoptic and meso-alpha scale environment predisposed to severe
turbulence is favored by proximity to the following phenomena: 1) the curved right entrance
region of a jet streak, 2) convection, 3) low relative vorticity, 4) upward vertical motion, 5)
horizontal cold advection, 6) strong vertical wind speed and directional shear, and 7) shallow
layers of low Richardson number. In particular, accidents are favored near convection and low
relative vorticity.

The second approach involves developing meso-beta scale numerical simulations of the
precursor environment that organizes severe turbulence. The Mesoscale Atmospheric

                                                   Page 27
                                                                       TREND Newsletter No. 12

Simulation System (MASS) numerical weather prediction model is utilized to perform 12 km
nested-grid simulations during two turbulence events that led to accidents: one in proximity to
deep convection and a second which is devoid of deep convection. Both accidents occurred in a
region of low relative vorticity within the curved right entrance region of a jet streak.

In an effort to refine the meso-beta scale precursor conditions that result in turbulence, both of
these simulations are evaluated for the wind, temperature, and moisture conditions that are
conducive to inertial instability, conditional symmetric instability, and shearing instability. These
instability signals indicate the potential for internal gravity wave generation mechanisms as a
precursor to extreme turbulence.

Comment: Dynamical meteorologists rejoice: A process-oriented rather than statistical approach
to turbulence generation. The role of mesoscale precursor processes is analyzed in a coherent
way and point to gravity waves as potential causes for severe turbulence events. Not that the
findings are particularly new, but at last they seem to point to a convincing method to derive
turbulence potential from dynamical processes that may be easier to identify than very-small
scale shear zones in numerical models.

                            Paper 7.7: A Windshear Hazard Index

    Fred H. Proctor, NASA/LRC, Hampton, USA, David A. Hinton, and Roland L. Bowles

In context of aviation science, windshear refers to a change in wind speed or direction
experienced by an airplane over a particular distance or length of time. An aircraft exposed to a
hazardous level of windshear may suffer a critical loss of airspeed and altitude, thus endangering
its ability to remain airborne. In order to characterize the hazard of windshear to aircraft, a
nondimensional index was developed based on the fundamentals of flight mechanics and our
understanding of windshear phenomena. This paper reviews the development and application of
the Bowles F-factor, which is now used by onboard sensors for the detection of hazardous
windshear. It was developed and tested during NASA/FAA's windshear program and is now
required for FAA certification of onboard radar windshear detection systems.

Reviewed in this paper are: 1) definition of windshear and description of atmospheric phenomena
that may cause hazardous windshear, 2) derivation and discussion of the F-factor, 3)
development of the F-factor hazard threshold, and 4) its testing during field deployments.

Comment: It was about time to define properly what windshear is and when it is dangerous to

              Paper 7.11: Modeling of Atmospheric Effects on Wake Vortices

           Robert E. Robins, Northwest Research Associates, Inc., Bellevue, USA,
                                   and Donald P Delisi

When instrument flight rules (IFR) are required due to poor visibility, aircraft are required to
remain a minimum distance behind the preceding aircraft, with the result that airport capacity can
be adversely affected. In certain atmospheric conditions (e.g., strong cross winds), this minimum
distance is greater than it needs to be for safe operations. If it could reliably be determined when
minimum following distances can be reduced, then airport capacity during IFR could be
increased. An approach was developed to modeling the evolution of aircraft wake vortices which
can be used to predict the length of time that wake vortices remain a hazard to following aircraft.
In this approach, atmospheric conditions are represented by vertical profiles of potential

                                                    Page 28
                                                                     TREND Newsletter No. 12

temperature, cross wind, and turbulence. Either eddy dissipation rate (EDR) or turbulence kinetic
energy (TKE) can be used to characterize the turbulence. The wake vortices are characterized by
their initial position, lateral separation, and initial descent speed. The model has been used by
the NASA Langley Research Center as the predictor component in a prototype system (called
AVOSS for Aircraft VOrtex Spacing System) that could be used by air traffic control to determine
when it is acceptable to reduce aircraft spacing.

In this paper, the modeling approach is described and examples shown of how various
atmospheric conditions can be expected to affect the evolution of aircraft wake vortices.

Comment: This paper is here to represent a fast-growing body of work on wake vortex
modelling. The increasing load on hub airports is forcing a rethink on spacing of approaching
aircraft, with ensuing consequences on acceptance capability and hence profitability. More funds
are expected to be invested in this field in the near future.

                              Session 8: Sensors and Systems

                   Paper 8.7: An Initial RUC Cloud Analysis Assimilating
                                      GOES Cloud-Top

                Dongsoo Kim, NOAA/FSL, Boulder, USA, and S. G. Benjamin

An initial cloud analysis for the RUC has been developed and refined since early 1999 and is
planned for implementation later in 2000. This analysis performs both cloud clearing and cloud
building, and uses the background hydrometeor (5 types) mixing ratio fields forecast explicitly in
the RUC using the MM5 mixed-phase cloud microphysics. This technique is described.
Assimilation of the NESDIS GOES sounder-based cloud-top pressure data has been tested in a
real-time parallel version of the RUC run at FSL. Positive effects on forecasts of cloud tops and
precipitation have been demonstrated. In order to enhance the resolution and coverage of the
sounder-based cloud product, a similar product based on the higher-resolution GOES imager
data has been developed. The two products are compared for relative strengths and
weaknesses regarding hourly assimilation into the RUC. Initial performance is presented at the

Comment: The overall concept of RUC as a system with a high update rate of data ingest has
been crying out for an objective way of using the wealth of information on the hydrological cycles
hidden in satellite data, which are one of very few data sources that are ubiquitous and have a
very high update rate. Progress both in variational analysis schemes and microphysics were the
prerequisite for notable progress in this field. Watch this space for more exciting developments.

            Paper 8.8: Measurement of Hazardous Winter Storm Phenomena
                          at the Portland International Airport

                 Bradley A. Crowe, MIT Lincoln Laboratory, Lexington, USA,
                            James E. Evans, and David W. Miller

Wind shear and lightning strikes to aircraft are classically associated with summer convective
weather. However, a recent study of West Coast airports found that severe wind shear and
lightning strikes can occur during winter storms in the Portland Oregon terminal area. These
unexpected phenomena arise from the local topography (especially surface cold air outflows from

the Columbia River gorge) and, the weakly electrified nature of the fall/winter/early spring
convective storms in the Portland area.

                                                   Page 29
                                                                     TREND Newsletter No. 12

This paper describes initial results of measurements with a transportable pencil beam C-band
Doppler weather radar at the Portland Oregon International airport (PDX). The radar will be used
to collect data on winter and early spring storms and wind shear events. It is believed that the
vertical wind shear events, experienced by aircraft arriving and departing PDX, could have similar
headwind changes as summer convective wind shear phenomena (e.g., > 30 knot losses or
gains in airspeed).

Both PPI and RHI Doppler radar scans were utilized so as to determine the vertical shear of the
wind shear events and to characterize the 3D structure. A second, but equally important,
objective of this study is to determine whether the baseline ASR-9 Weather System Processor
(WSP), which is planned for PDX, will adequately detect the vertical wind shear. The C-band
Doppler radar data will be used to create synthetic ASR-9 reflectivity and Doppler radial velocity
data. This synthesized ASR-9 data will be analyzed to evaluate the ability of the ASR-9 to detect
the vertical wind shear. The paper characterizes key meteorological features of the wind shear
and lightning storms observed and then discuss the viability of various options (e.g., an
unmodified ASR-9 WSP, a modified WSP, and a WSP augmented with data from other sensors
such as vertical profilers) for providing timely warning on hazardous phenomena.

Comment: Although at first glance a paper with a very local significance, it has been included
here for the practical merits of using available (and hence free) information from existing ATS
radar systems for the detection of severe weather events. For many countries and locations
where the high cost of a dedicated TDWR-type system would be unavailable for cost reasons,
this example could be a very useful approach.

                 Paper 8.9: Retrieval of Cloud Microphysics during the Mt
                       Washington Icing Sensors Project (MWISP)

Charles C. Ryerson, U. S. Army Corps of Engineers Cold Regions Research and Engineering
            Laboratory, Hanover, USA, George G. Koenig, and Forrest R. Scott

MWISP, the Mt. Washington Icing Sensors Project, was conducted in April, 1999, to assess the
ability of remote sensing devices to measure the microphysics of inflight icing conditions,
principally liquid water content, drop size and the presence of drizzle drops. An additional goal
was to characterize the icing environment in complex terrain in a mid-latitude, early spring
environment. Both goals required continuous, high quality in situ and remote measurements of
cloud microphysical conditions, an opportunity afforded by a mountain summit observatory, and
the close proximity of remote sensing devices near the base of the mountain. This was
accomplished through use of a wide variety of instrumentation located at both the Mt.
Washington summit (1916 m) and near the mountain base (811 m) including Particle Measuring
Systems (PMS) probes, Rosemount ice detectors, an icing radiosonde system and rotating

CRREL operated three PMS probes at the summit, an FSSP (2-47 mm), a 2D cloud probe (0-800
mm) and a 2D precipitation probe (0-6400 mm). The three instruments were located on a mast
extending about 2-m above the Mt. Washington Observatory 12-m high concrete observation
tower to provide maximum exposure to winds. The instruments were operated during the full
project, the entire month of April, and acquired approximately 150 hours of measurements.

Though PMS probes are typically exposed to harsh icing conditions when carried on aircraft, the
mountain summit presented several unique challenges to acquiring high quality measurements.
Airflow over an aircraft wing is generally consistent because angle of attack varies over only a
few degrees. On Mt. Washington, wind direction changes over time, both because of

                                                  Page 30
                                                                          TREND Newsletter No. 12

geostrophic changes but also as a result of turbulence over time periods of seconds to a few
minutes. Though probes can be rotated into the wind during large-scale direction changes, they
cannot track rapid fluctuations. Icing was a significant problem, and additional operational
obstacles included vibration-loosened circuit boards, signal noise produced by commercial radio-
frequency interference, and required periodic maintenance necessitating removal of probes in
environments with wind speeds of 20 ms-1 or more.

Because of the ferocity of the operating environment, the FSSP probe calibration drifted
considerably during the project. In addition, an older FSSP did not provide activity levels or total
strobes to the data acquisition system for making corrections to the data. Comparisons of the
PMS probe parameters with liquid water contents retrieved from rotating multicylinders and a
Rosemount ice detector are providing verification of the PMS probe measurements. In addition,
comparisons of summit measurements are being made with those of a radiosonde that measures
supercooled liquid water content, and measurements made by NASA’s Twin Otter research
aircraft during six flights over the summit. Results of these comparisons are presented with
discussions of reasons for physical consistency or inconsistency of the measurements. Such
instrument intercomparisons are generally not available in most field programs, thus affording the
participants of MWISP an opportunity for more reliable cloud microphysical measurements,
despite the operating difficulties presented by the Mount Washington environment.

Comment: A refreshingly practical report on an alternative to expensive research flight testing of
sensors: Mountain observatories have a significantly lower operating cost than research flight,
and thus promise a much larger number of test hours with a multitude of synoptic and
microphysical situations to be looked at.

               Paper 8.10: Verified Detection of Supercooled Large Droplets
                         with Dual-Polarized, Millimeter-Wave Radar

              Roger F. Reinking, NOAA/ETL, Boulder, USA, Sergey Y. Matrosov,
                Charles C. Ryerson, Robert A. Kropfli, and Bruce W. Bartram

Supercooled large cloud droplets (SLDs), those of 50-500 micron diameter, have been previously
identified as creating a particularly severe aircraft icing hazard, due to their ability to penetrate the
slip stream of an aircraft wing and freeze on the surface as rough ice. Several polarization states
have been evaluated with NOAA/ETL's Ka-band (8.66 mm wavelength) radar to distinguish
hazardous SLDs from the various types of cloud ice particles that commonly occur but do not
cause icing. Most recent tests focused on depolarization ratio, DR, obtained from transmitting a
45 -slant, quasi-linear polarization state and receiving the corresponding co- and cross-
polarization signals. Scattering calculations have demonstrated that this polarization state has a
significant advantage over the standard horizontal linear state because of the increased power
received in the cross-polarized received signal, which is considerably weaker than that received
co-polar channel, but required to differentiate particle type. Another of several advantages is that
the slant polarization state is comparatively insensitive to variations in settling of orientation of ice
particles, which causes uncertainty is horizontal DR measurements. Current and previous tests
have isolated warm large droplets from ice particles, and that in itself demonstrated the capability
of this kind of dual-polarization radar to differentiate between them. However, supercooled large
droplets are rare events and are not often present during short field experiments. Now, under
FAA sponsorship, verification of SLD detection with the radar has been found in data taken
during the 1999 Mt. Washington Icing Sensors Project (MWISP). The radar DR signature of the
large water droplets is independent of their temperature. However, their detection in the radar
DR signature, with size and supercooling confirmed with in-situ observations, provided
reassurance that millimeter-radar identification of actual icing hazards can

be accomplished. The expected radar signature for water droplets was observed in a sub-
freezing cloud while 150 m mean diameter droplets (SLDs) were observed with CRREL's 2DGC

                                                      Page 31
                                                                       TREND Newsletter No. 12

PMS cloud particle probe at Mt. Washington Observatory (MWO). Other supporting data
included visual observation of fog at the summit, substantial icing measured with the MWO
Rosemount icing probe, substantial cloud liquid water measured with ETL's microwave
radiometer, and sub-freezing temperatures verified with an NCAR CLASS sounding. The case is
presented and further supported by measurements that show how the various individual types of
ice particles were well-differentiated among themselves and from SLDs.

Comment: See Section 8.9 above.

           Paper 8.11: A Technique for Improving the Detection of Drizzle and
           Freezing Drizzle on ASOS Automated Weather Observing Stations

                            Charles G. Wade, NCAR, Boulder, USA

Drizzle, and in particular freezing drizzle, is a form of precipitation that is not well observed in
Automated Surface Observing System (ASOS) stations. It is typically reported as rain, snow, or
"unknown" precipitation (UP), unless the observation is supplemented by a human observer.
Under certain conditions (such as when combined with warm cloud-top temperatures aloft)
drizzle may be an important indicator of in-flight icing conditions near the station. Improving the
detection of drizzle at automated weather stations, therefore, may result in better forecasts of in-
flight icing conditions near these stations.

The reason generally given for the inability of ASOS stations to report drizzle is that the water
droplets are simply too small to be observed by the ASOS precipitation identifier: LEDWI (Light-
Emitting Diode Weather Identifier). Drizzle is defined as water droplets smaller than 0.5 mm in
diameter, while the LEDWI sensor was designed to observe somewhat larger precipitation
particles (> 1 mm diameters). An examination of the raw output data from LEDWI sensors at
ASOS stations shows, however, that there is a distinct signal that appears in these data when
drizzle is occurring that can be used to detect the presence of the drizzle. The purpose of this
paper is to describe the output data that are produced by the LEDWI sensor, and to show a
graphical technique for displaying these data that identifies drizzle as distinct from rain or snow.
By combining these results with other data measured at ASOS stations (ceiling)

Comment: Freezing drizzle as one of the most dangerous indicators of aircraft icing can be well
identified in a many cases by a dense surface observing system using adequate sensors.
Whether the ASOS or similar systems by other manufacturers are employed is of minor
importance provided they are capable of resolving the small drizzle-size droplets.

           Paper 8.12: Evaluation of Snow Forecasts Provided by the Weather
                 Support to Deicing Decision Making (WSDDM) System

                     Steven V. Vasiloff, NOAA/NSSL, Salt Lake City, USA,
                        Roy Rasmussen, Mike Dixon, and Frank Hage

The National Center for Atmospheric Research Weather Support to Deicing Decision Making
(WSDDM) system uses real-time snow water equivalent (SWE) data from snow gauges to
calibrate radar data. The calibration is done through adjustment of the coefficient in the radar Z-
S equation which relates radar reflectivity to snowfall rate (in liquid equivalent). The main output
from WSDDM is a 30 min prediction of SWE at a particular gauge site.

The WSDDM system has shown success in the eastern U.S., primarily in New York City and

                                                   Page 32
                                                                       TREND Newsletter No. 12

Chicago. This paper reports on an evaluation of the system for New York and for the more
mountainous terrain of the Salt Lake City area. The WSDDM system was installed at the Salt
Lake City National Weather Service forecast office in 1996. Unlike previous field tests, the Salt
Lake system ingests wideband level-II data from the WSR-88D as well as SWE data from 8 snow
gauges. Initial performance results were encouraging and showed large fluctuations in the Z-S
coefficient with time.

Recent activities have entailed modification of which elevation angles are used in the reflectivity
analysis. Data from 20+ winter storms in northern Utah are being used to evaluate the system for
the different gauge sites, including the site at the airport. These data have been analyzed using
a fixed Z=75S**2 and will be used as a benchmark for the WSDDM system.

Important considerations for this study include the large distance between the WSR-88D's 0.5
degree beam and the airport (~5000 ft) and the shallow nature of Utah's winter storms.
Considerations at other sites include beam blockage and varying altitude and range of the

Comment: This paper again was included for drawing the attention to a wide-spread problem in
mountainous terrain: Radar stations positioned at higher altitudes to avoid beam blocking by
orography are suffering from the failure to detect seeder-feeder processes in winter storms,
where much of the snow is generated at heights below the minimum beam elevation.

                Paper 8.14: Multi-Frequency and Polarization Radar-Based
                               Detection of Liquid Droplets

        J. Vivekanandan, NCAR, Boulder, USA, Guifu Zhang, and Marcia K. Politovich

Preliminary results of dual-wavelength radar data analysis and model calculations show the
potential for detecting liquid water content and droplet size. However, the dual-wavelength
technique is susceptible to non-Rayleigh scattering effects, mixed-phase precipitation, and
sampling volume differences between two radars. Only a limited verification of the technique is
presented in the previous studies by comparing liquid water path along the radar beam with the
corresponding microwave radiometer-based liquid path retrieval. In the absence of in situ aircraft
measurements, it is difficult to verify dual-wavelength-based liquid water content estimate and
droplet size. This limitation can be partially eliminated in the case of tri-frequency (X, Ka, and W-
band) measurements because two independent liquid water estimates can be derived using X
and Ka pair and Ka and W-band pair observations.

In this paper, tri-frequency radar measurements collected during the Mount Washington Icing
Sensing Project (MWISP) were analyzed. The retrieved liquid water contents and droplet sizes
were compared. The estimated liquid water path is compared with radiometer observation.

The attenuation corrected dual-wavelength ratio at X/Ka and Ka/W were used for estimating
particle size. The MWISP data analysis suggest the dual-wavelength technique is limited by a
number of radar performance factors such as, (a) sensitivity of X- and W-band radar systems, (b)
fluctuation in reflectivity measurements, and (c) sampling volume differences between radars.
The paper suggests preliminary specification for an optimal radar system for liquid water content
and droplet size estimations.

Comment: This paper is included for the benefit of those that thought dual polarization would
answer all questions on icing.

                Paper 8.16: The Detection of Convective Turbulence Using
                                Airborne Doppler Radars

                                                    Page 33
                                                                       TREND Newsletter No. 12


     Larry B. Cornman, NCAR, Boulder, USA, John K. Williams, and Robert K. Goodrich

Encounters with convectively-induced turbulence can result in severe upsets to commercial
aircraft. The 1998 turbulence encounter near Japan which resulted in numerous injuries and one
death is a very tangible example. Many commercial aircraft are being fitted with airborne
windshear radars which have the potential to detect significant convective turbulence. As part of
NASA's Aviation Safety Program (AvSP), a joint industry-research community effort has been
initiated to address this problem. This endeavor encompasses basic research into the
phenomena, numerical modeling, the development of robust detection and quality control
algorithms, end-to-end simulation and flight tests. This paper describes these varied efforts as
well as promising results.

Comment: Given the inherent problems of predicting turbulence generated by very localized
convection-induced gravity waves, this outlook to the future of in-flight detection systems is highly

          Paper 8.21: ITWS and ITWS/LLWAS-NE Runway Alert Performance at
                            Dallas-Fort Worth and Orlando

                  Mark A. Isaminger, MIT Lincoln Laboratory, Lexington, USA,
                            Bradley A. Crowe, and Erik A. Proseus

The Integrated Terminal Weather System (ITWS) provides runway-orientated wind shear and
microburst alerts in order to enhance the safety of flight operations at major U.S. airports. The
alerts are reported as either losses or gains of airspeed, representing performance decreasing or
performance increasing wind shears. The performance of ITWS as a stand-alone system has
been thoroughly documented in previous research. During the 1994 ITWS Demonstration and
Validation (DemVal) testing, the probability of detection (POD) and probability of false alarm
(PFA) at Memphis (MEM) and Orlando (MCO) for all loss events were >90 and <5 percent,

The Low-Level Windshear Alert System - Network Expansion (LLWAS-NE) also generates
runway alerts in the same format as ITWS. LLWAS-NE is not subject to viewing angle problems
such as that experienced from single-Doppler radar. However, false alarms caused by LLWAS-
NE sensor failures at some TDWR sites have reduced user confidence in the system. At those
ITWS sites with an LLWAS-NE, the alerts derived from TDWR data are integrated with LLWAS-
NE alerts in order to improve the alerting performance. The ITWS integration algorithm is
identical to the Terminal Doppler Weather Radar (TDWR) version with the exception of a few
adaptable parameter changes. The ITWS/LLWAS-NE parameters were modified slightly to take
advantage of the ITWS and TDWR algorithm performance differences. For instance, the ITWS
microburst algorithm tends to underestimate the actual wind shear loss of most events, so the
minimum velocity threshold to confirm a weak event was reduced slightly to account for this

In this paper, the performance of ITWS and the ITWS/LLWAS-NE integration algorithm at the
MCO and Dallas-Ft. Worth (DFW) demonstration sites was discussed. This performance
assessment is considered unique since the radar and anemometer data were combined to create
the runway truth. In addition, the Dallas Love (DAL) TDWR surface scan data was incorporated
every five minutes to provide an additional source of truth with a different viewing angle. The
focus of the research reported in this paper is to identify the shortcomings of both systems in
order to recommend modifications that will improve the integration performance. Thus, the radar

                                                      Page 34
                                                                       TREND Newsletter No. 12

data for each missed event and false alert was analyzed to ascertain the type and frequency of
the failure modes. Recommendations are also made for an improved integration algorithm.

Comment: A very detailed paper on a national system included here for the sobering insight into
the long, hard way to reduce false alarms, improve detection rates and make the system practical
to use. Again, a warning to those who believe in quick, off-the-shelf solutions to complex
atmospheric problems.

              Paper 8.23: A Comparison of GOES-8 Imagery with Cloud-Top
                           Penetrations by a Research Aircraft
               Frank McDonough, NCAR, Boulder, USA, and Ben C. Bernstein

Various studies have shown that it is possible to extract a variety of information about the
locations, temperature, and microphysical properties of cloud tops from GOES-8 imager data.
The visible channel can provide information about the water content, concentration of cloud
particles, and cloud depth. The short-wave IR channel has shown promise in detecting both
cloud phase and particle sizes, while the long wave IR channel is useful in determining cloud top
temperatures. Since the GOES-8 imagery has high temporal and spatial resolution, this cloud
information can be quite valuable for monitoring clouds that may cause aircraft icing. In recent
years, several algorithms have used GOES-8 measurements to identify supercooled liquid water
at cloud tops.

During the winters of 1997 and 1998, the NASA-Glenn Twin Otter research aircraft sampled a
variety of subfreezing clouds over the Great Lakes region. In this study, microphysical probe and
state-parameter data from Twin Otter cloud top penetrations are examined and broken into
several categories depending on cloud top temperature, cloud phase, liquid water content,
particle concentration, and particle size. The aircraft and GOES-8 data are objectively compared
to test for cloud top temperature accuracy and the ability of certain channels to differentiate
between various microphysical properties at cloud top. Results provide new information
regarding the use of GOES-8 data for in-flight icing detection.

Comment: An illustration of the difference between a proof-of-concept study and practical
implementation. Unfortunately, it still appears that satellites only scratch the surface of the cloud
icing problem!

              Paper 8.24: Estimation of Instrument Cloud Base Conditions at
                    Night Using GOES and Surface Temperature Data

                      Gary P. Ellrod, NOAA/NESDIS, Camp Springs, USA

Low ceilings and visibilities caused by fog and stratus clouds are one of the most significant
weather-related factors in aircraft accidents and operational flight delays. Detection of low clouds
or fog at night can be accomplished using an image derived from the temperature difference
between the short wave (3.9 micron) and long wave (10.7 micron) infrared (IR) imager channels
on the Geostationary Operational Environmental Satellites (GOES), sometimes referred to as the
“fog product”. One weakness of the product, however, is that it cannot easily distinguish between
clouds that cause very low ceilings and visibilities at the surface, versus higher-based stratus,
stratocumulus, and altostratus that do not represent significant hazards to aviation. Empirical
rules have been developed to help forecasters make such assessments from the satellite
imagery, but they require some experience for successful diagnosis of fog without the use of

surface data. A more quantitative product is desired that would provide an estimate of the
likelihood of low ceilings or visibilities that can result in hazardous flight conditions. At the

                                                    Page 35
                                                                           TREND Newsletter No. 12

National Weather Service’s Pacific Coastal Forecast Systems workshop held in July, 1998 in
Silver Spring, Maryland, there was a proposal to produce an experimental, enhanced GOES fog
product that would provide probabilities of certain critical ceilings and visibilities for aviation users.

Data from the summer of 1997 has been collected and analyzed, and a good correlation was
found between the GOES longwave infrared minus surface temperature difference, and
instrument cloud base conditions (<1000 ft). An experimental prototype product has been
developed that can be automatically generated, and will soon be available on an Internet Web
site. Examples of the product were shown, along with some preliminary verification, and a
discussion of known product weaknesses. Based on preliminary results, this appears to be a
promising approach to providing aviation forecasters with an additional satellite-based guidance

Comment: Very much as Section 8.23 above.

                                      Joint Session
             9th Aviation Conference and 20th Severe Local Storm Conference

        Paper J-2.1: Evaluation of the NCAR Thunderstorm Auto-Nowcast System

                            Cynthia K. Mueller, NCAR, Boulder, USA,
                           Tom Saxen, Rita Roberts, and James Wilson

The National Center for Atmospheric Research (NCAR) Auto-Nowcast (AN) system provides 30
and 60 min time and place specific forecasts of thunderstorm initiation, growth and dissipation on
a local-scale (~200 x 200 km grid with 1 km resolution). This expert system uses fuzzy logic
methodology and incorporates forecast predictors derived from years of field research, numerical
modeling studies, and theory. The system makes use of operational data sets including radar,
satellite, surface stations, lightning, profilers and radiosondes. Also included are diagnostic and
forecast information from numerical models that are run in real time. Development of the AN
system is sponsored by the; Federal Aviation Administration (FAA) Weather Research Program
as part of the Convective Weather PDT, Army Test and Evaluation Command (ATEC), National
Weather Service (NWS) as part of the System for Convection Analysis and Nowcasting (SCAN),
and National Science Foundation under the U.S. Weather Research Program.

In this paper, AN forecast from two operational sites, the NWS Weather Forecast Office (WFO)
at Sterling, Virginia, and the Forecast Office at White Sands Missile Range are statistically
compared to extrapolation forecast. Summary statistics for the 30 and 60 min forecast show that
the AN system has increased skill in terms of Probability of Detection (POD) and Critical Success
Index (CSI). Although the False Alarm Ratio (FAR) varies, generally it is similar for both
extrapolation and AN systems. As expected the differences between extrapolation and AN
forecast CSI scores were relatively small when examined over long time periods. The small
differences are partially due to large multi-cellular systems where an extrapolation forecast works
very well. These large systems tend to dominate the statistics because of their area coverage.
However, from an user-standpoint accurate forecast of change, e.g. storm initiation, growth and
dissipation, are extremely important.

To better understand the forecast capabilities various weather situations are reviewed. They
include; initiation of squall line at location of boundary collision, initiation along quasi-stationary
boundary, initiation and growth associated with a quasi-stationary boundary on an active storm
day, initiation of storms as a boundary moves into a field of cumulus clouds, storm growth at the
location of storms and boundary intersections and storm-storm mergers, extrapolation and

                                                       Page 36
                                                                       TREND Newsletter No. 12

formation of bow-echo, initiation and growth of severe thunderstorms and dissipation of a line of
thunderstorms. These case examples are encouraging because they show the AN ability to
improve over extrapolation forecast. The improvement is primarily associated with the AN ability
to correctly forecast new storm development.

Comment: A tall order for an automated nowcasting system is the prediction of newly developing
cells, decaying, cell splitting and reorganizing of squall lines. Extrapolation, however dear to the
hearts of many developers, will not do. Ingestion of other data types (low level winds, moisture,
etc.) will be instrumental in the success of nowcasting systems that improve on linear

             Paper J-2.3: Global Thunderstorm Guidance Forecasts from the
                       AVN Model from the WWSTORM Algorithm

     Donald W. McCann, NOAA/NWS/NCEP/Aviation Weather Center, Kansas City,USA

A large organization of professionals helps domestic aviation operations in the United States
avoid hazardous weather. As international operations have expanded, so has increased the
need for similar support for global hazardous weather avoidance. Two factors complicate
international aviation weather forecasting. These are, first, many flights are over data sparse
oceans, so there is less knowledge of the state of the atmospheric conditions that cause
hazardous aviation weather, second, international flights are typically much longer than domestic
flights which increases the necessary forecast lead time of hazardous weather areas for effective
flight planning.

Thunderstorms are the primary hazard at international flight altitudes. Forecasting thunderstorms
in general is rather simple conceptually namely; find the areas in which potentially unstable air
parcels can be lifted to their Level of Free Convection. Developed for domestic thunderstorm
forecasting guidance, the experimental VVSTORM algorithm examines numerical model output
for favorable conditions for thunderstorms using this simple recipe. Since VVSTORM was
designed to run on any numerical model resolution through its parameterization of parcel lifting,
the Aviation Weather Center began experimenting with VVSTORM on the global AVN numerical

This paper shows comparisons of AVN-VVSTORM output with 1) VVSTORM output from the
RUC2 and ETA models for mid-latitude convection over the United States, and 2) satellite
imagery for six-hour AVN forecasts over tropical oceans. Subjective experience with AVN-
VVSTORM has been very good in the mid-latitudes and good in the tropics. The only apparent
problem area is along the equator where the parameterized parcel lifting appears not to force
enough convection.

Comment: The elimination of Regional Area Forecast Centers and their regional SIGWX
information will create new demand for the global production of automated convection guidance.
Here is an approach being applied at the Kansas Aviation Weather Center.

                                                   Page 37
                                                                     TREND Newsletter No. 12

     Paper J-2.4: Use of a New Thunderstorm Potential Index for 12-hour Forecasts
                             Using Mesoscale Model Data

                      David I. Knapp, U.S. Army Research Laboratory,
                    White Sands Missile Range, USA,. and Gordon Brooks

Severe weather forecasters at the Air Force Weather Agency produce graphical forecasts of
thunderstorm areas with expected severity three times each day for dissemination to base
weather stations and other DoD customers. These Military Weather Advisory (MWA) forecasts
depict aerial coverage of storms expected during the valid forecast period, plus predictions of
storm intensity and time periods in each outlined area. Use of a new Thunderstorm Potential
Index (TPI) provides forecasters with a “first guess” mesoscale model-based tool depicting areas
of general and severe thunderstorms predicted to occur during the MWA valid period. TPI charts
are intended to give forecasters an idea of where appropriate MWA thunderstorm areas should
be forecast for the upcoming period. The TPI was originally developed using a “perfect prog”
approach with the goal of accurately predicting percent probabilities of local thunderstorm
occurrence within a 100km radius of a RAOB location during the 12 hours after RAOB time. 1200
UTC RAOB data from 13 selected locations across the continental United States were collected
from March to October of 1990 and 1993. The 13 locations represented diverse climatic regions
with a mix of sites from coastal regions of the East Coast and Southeast to the Northern-,
Central-, and Southern Plains and Rocky Mountains. Stability indices were calculated for all
RAOBs collected and correlated with thunderstorm reports (from surface observations and
lightning data) within the 100km radius of the RAOB sites. Three indices (K, LI, and SWEAT)
were found to be statistically significant in determining correct yes/no forecasts of thunderstorm
occurrence during the forecast periods.

Using multivariate discriminant analysis (using stepwise procedures) to determine the significant
indices, a forecast regression equation was derived which produces percent probability of
thunderstorm occurrence at a RAOB location. It was found that a threshold value of TPI > 46%
maximized the Probability of Detection and minimized the False Alarm Rate. Verification at
seven independent locations was completed using RAOB data from April to September of 1996.
Comparisons with NGM MOS TSV06 forecasts valid at each verification location produced very
favorable results. With these positive results from the independent verification data set, the TPI
was applied to stability indices derived from MM5 output. Contours of TPI=47% have been found
to closely match MWA contours for general thunderstorm areas produced by forecasters without
knowledge of TPI output. Additionally, TPI threshold values in excess of 70% have been found to
correlate with MWA areas of severe thunderstorms. Comparisons of the TPI forecasts and MWA
areas over the United States and Europe will be presented showing how the automated contours
match up to forecaster-produced MWA thunderstorm areas. Verification results will show
performance comparisons of the TPI and MWA using lightning data and severe reports during
each forecast period.

Comment: So there’s room for yet another thunderstorm probability index, included here for
educational purposes as its development is nicely described.

    Paper J-2.5: Ensemble Cloud Model Applications to Thunderstorm Forecasting

      Kimberly L. Elmore, NOAA/NSSL and CIMMS/Univ. of Oklahoma, Norman, USA,
                       David J. Stensrud, and Kenneth C. Crawford

A cloud model ensemble forecasting approach is developed to create forecasts which describe

                                                    Page 38
                                                                         TREND Newsletter No. 12

the range and distribution of thunderstorm lifetimes that may be expected to occur on a particular
day. Such forecasts are crucial for both anticipating severe weather and ensuring the smooth
flow of air traffic at busy, hub airports. Storm lifetime is an important characteristic to examine
because long-lasting storms tend to produce more significant weather, and have a greater impact
on air traffic, than do storms with brief lifetimes.

Eighteen days distributed over two warm seasons are examined. Soundings valid at 1800 UTC,
2100 UTC and 0000 UTC, provided by the 0300 UTC run of the operational Mesoeta model from
the National Centers for Environmental Prediction, are used to provide initial conditions for the
cloud model ensemble. These soundings are shown to represent a likely range of atmospheric
states around the location of interest. A minimum threshold value for maximum vertical velocity
within the cloud model domain is used to estimate storm lifetime. Forecast storm lifetimes are
verified against observed storm lifetimes, as derived from the Storm Cell Identification and
Tracking algorithm applied to WSR-88D radar data from the National Weather Service (NWS).

When kernel density estimates are applied to the pooled data set consisting of all 18 days, a
vertical velocity threshold of 8 m/s results in a forecast pdf of storm lifetime which is closest to the
observed pdf. Model results from all 18 days also reveal that the storm lifetime resulting from a
given input sounding cannot be determined by analyzing the bulk sounding parameters, such as
convective available potential energy, bulk Richardson number (BRN), BRN shear, or storm
relative helicity. Standard 2 x 2 contingency statistics reveal that, under certain conditions, the
ensemble model displays some skill locating where convection is most likely to occur.
Contingency statistics also show that when storm lifetimes of at least 60 min are used as a proxy
for severe weather, the ensemble shows considerable skill at identifying days that are likely to
produce severe weather. Because the ensemble model appears to have skill in predicting the
range and distribution of storm lifetimes on a daily basis, the forecast pdf of storm lifetime is used
directly to create probabilistic forecasts of storm lifetime, given the current age of a storm. Such
a product could furnish useful information to Air Traffic controllers by providing guidance about
how soon a storm is likely to affect (or cease to affect) air traffic at a specific location. Similarly,
this product could provide NWS forecasters with guidance about how likely it is that a particular
cell will affect a given community.

Comment: Another educational paper on how-to-do.

                                          Poster Session

           Poster 1.2: The Use of NM5 in an Aviation Weather Forecast System

                      James F. Bresch NCAR, Boulder, USA, J. G. Powers,
                             K. W. Manning, and J. G. Michalakes

A key component of the Advanced Operational Aviation Weather System (AOAWS) being
developed by the National Center for Atmospheric Research (NCAR) for Civil Aeronautics
Administration (CAA) is the MM5 model. The Pennsylvania State University / NCAR mesoscale
model Version 5 (MM5, Dudhia 1993; Grell et al. 1994) is a nonhydrostatic, primitive equation
mesoscale weather forecast model in wide use around the world. The flexibility of the MM5
system, along with the availability of inexpensive yet fast computers permits users to produce
their own mesoscale forecasts of the kind (or better then those) previously only produced by
operational centers. Having local control of a mesoscale model allows users to produce tailored
forecast products.

In this paper, a description of the AOAWS MM5 system is given, including details of the domain
configuration, physical parameterizations, and initial data. Examples of standard forecast

                                                      Page 39
                                                                       TREND Newsletter No. 12

 products as well as aviation-specific forecast products will be shown, along with case studies of
weather-related aircraft incidents in order to demonstrate the usefulness of such a mesoscale
forecast system.

Comment: This is just one example of the use of the MM5 model in support of operational
aviation forecasting. MM5 is triple-nested down to an inner domain of 15 km resolution. The
CAA works closely with NCAR, and will be collaborating on trials of 3D-VAR data assimilation as
part of the MM5 system. MM5 will be useful for a while yet, as the “next generation” community
mesoscale model in the United States – WRF (Weather Research and Forecasting model) – is
still in the early stages of development.

           Poster 1.8: Graphical Area Forecast (GFA) Breaking the Text Barrier
                                  in the New Millennium

                     Daniel Chretien, MSC, Dorval, Canada. and M. Crowe

In April 2000, Environment Canada introduced a graphical area forecast (GFA) service, replacing
the antiquated alphanumeric area forecast bulletins (FAs) that have been used, with little change,
since the 1940s. Domestic and international aviation users now have access to a new graphical
representation of aviation weather parameters over Canadian airspace, making it easier to get an
understanding for the weather affecting their plans.

Forecasters at several aviation weather centres across Canada (Kelowna, Edmonton, Toronto,
and Gander) use a software program (Edigraf) to prepare graphical depictions of the weather
over their area of responsibility, including clouds and weather, icing and turbulence, and freezing
level information. These regional depictions are combined by forecasters in Montréal into a large
national depiction, from which seven separate "domains" are automatically carved. The final
charts are sent across communication networks, are stored on a web site for easy access by
aviation users, and are FAXed to other users.

This is a first in the world, and it has been warmly received by the aviation community. It has
been an interesting process to overcome the challenges posed by getting forecasters to work
graphically, to coordinate across multiple weather centres, and working with the users to fine-
tune the product.

Comment: A leading-edge example of what will be an increasing trend of delivering information
to pilots in pictorial form, rather then the traditional (and sometimes complex to decipher) text

                   Poster 1.12: Current Work of the Aviation Applications
                           Research Group at the UK Met. Office

                  S. James, UK Met Office, Bracknell, Berks. United Kingdom,
                        C. Bysouth, T. Scott, D. J. Hoad, and R. Lunnon

This paper gives a brief overview of the current work of the group, in particular covering progress
since the eighth conference on Aviation, Range, and Aerospace Meteorology in 1999.

The topics covered will include: 1. Monitoring of forecasts of upper air wind and clear air
turbulence. 2. Development of techniques for identifying volcanic ash at aircraft cruising levels
using satellite imagery. 3. Wake vortex studies. 4. Activities to facilitate the safe introduction of

                                                    Page 40
                                                                      TREND Newsletter No. 12

reduced aircraft vertical separation minima in Europe in 2002. 5. Facilities for providing
climatological data to airlines for flight planning. 6. Developments in the prediction of lightning
strikes to helicopters. 7. Provision of meteorological information to support the design of
geostationary stratospheric airships. 8. Developments in the prediction of optimum routes for

Comment: AMS conferences can naturally be dominated by North American presentations,
concerns and systems. This was a good example of the variety of work being carried on
elsewhere in the world, with a somewhat different focus.

                                                   Page 41
                          TREND Newsletter No. 12

            PART II

General Assembly of the European
  Geophysical Society Meeting

          Nice, France
           April 2000

                Page 42
                                                                       TREND Newsletter No. 12

"The following part of the TREND Newsletter submitted by Dr Herbert Pümpel (Chairman of
TREND) contains 23 selected, reviewed and summarized abstracts prepared for the General
Assembly of the European Geophysical Society Meeting held at Nice, France, in April 2000. Dr
Pümpel also kindly commented on these abstracts. Abstracts were taken from the sessions on
"Ocean and Atmosphere", where one particular session was dedicated to aeronautical
meteorology problems. Other abstracts were taken from other sessions dealing with
topographically modified winds, in particular, the results of the recent Mesoscale Alpine
Programme (MAP). A large number of the papers presented on Aeronautical Meteorology had
focussed on technical aspects of wake vortex predictions therefore only a few were selected to
represent this highly specialized field of research. In future, plans are to report regularly on this
new series of aeronautical meteorology conferences in TREND Newsletters."

     The Future of Aviation Meteorology as seen by the WMO Fifth Long Term Plan

                                            H. Pümpel

Aviation Meteorology has undergone the most radical changes of all operational meteorological
institutions over the last two decades, and more changes are foreseen for the future. A dramatic
shift in the user requirements, the availability of global forecast products, the change from
manual products explained in personal briefings to de delivery of semi-automated products by
digital networks to pilots and dispatchers are only the visible aspect of these changes. The
Commission for Aeronautical Meteorology of WMO has instituted working groups with the aim to
support and guide individual Members during this highly critical process by fostering activities as
laid out in the Fifth Long Term Plan. These include:

-   Organization of Special Training events in aeronautical meteorology
-   Encouragement of Cost-Recovery principles
-   Implementation of the World Area Forecast System (WAFS)
-   Improvements and increased cost-effectiveness in meteorological observations
-   Improvements in forecasts for the terminal area with special emphasis on the first
    three hours
-   Support for a global system for collection of in-flight aircraft observations
-   Improvements to the forecasts of in-flight weather hazards to aviation, including
    turbulence, icing, volcanic ash, and tropical cyclones
-   Increased understanding and awareness of the impact of aviation on the

The presentation will highlight developments that are already under way, and outline the direction
those that are now anticipated for the next few years.

Comment: A short introduction to the Aeronautical Meteorology Programme of the WMO Fifth
Long Term Plan, explaining briefly the plans and intentions behind the plan.

                                                    Page 43
                                                                         TREND Newsletter No. 12

                        The Influence of Thunderstorms on Air Traffic
                         at the Airport of Frankfurt/Main (Germany)

                                     M. Sasse and T. Hauf
              Institute of Meteorology and Climatology of the University Hannover
            (+49) +511 762 4418

Safety and operating efficiency of air traffic can be significantly influenced by weather.
Consequences can be delays, diversions, cancellations and reduced scheduling integrity. In a
pilot study, the impact of thunderstorms on landing flight operations at Frankfurt/Main Airport
(Germany) has been investigated for selected days with thunderstorms in the years 1997 and
1998. As the airport Frankfurt is working at even more than 100% of its nominal capacity, delays
can be observed daily. To identify thunderstorm as the unambiguous cause of delays, days
without thunderstorm were used for reference and the difference in delay time was determined.
The delay itself was derived from the differences between the actual flight time and the estimated
flight time corresponding to the flight plan. The data was provided by the German ATM
organization (DFS). As a result of the study, a distinct increase of delay minutes by a factor 5-10
in 1997 and 1-4 in 1998 by thunderstorms was observed. Thunderstorms typically cause about
1000 delay minutes for a sum of 100 arriving planes within three hours of assumed impact time.
Since only the delay of arriving airplanes has been investigated, more studies are required to
estimate the total amount of delay and the total amount of arising costs the delay produces.

Comment: Somewhat later than at some large US airports, European hubs are beginning to
experience the knock-on effects of severe weather on schedules. The quantification of the
problem is a first step towards addressing it.

                     SIGMA: System of Real-Time Identification of Icing

                               J.-M. Carriere (1), C. Le Bot (1)
                (1) Météo-France; / fax:+33-561078209

Icing is one of the most important in-flight meteorological risks for aircraft, which pilots must try to
avoid. In order to distinguish the areas of potential icing risk, and particularly to determine them
to the nearest potential, an action has been undertaken at Météo-France to develop a system
able to better forecast the risk areas for aeronautical users, while limiting unnecessary forecast
icing risk areas. SIGMA is a system of real time identification for use by aeronautical weather
forecasters. It uses the operational observed and forecast data, like infra-red calibrated image
from the MeteosatV satellite, the weather radars composite images from the ARAMIS network,
the data from the numerical weather prediction models ALADIN or ARPEGE (particularly ground
temperature and icing index risk forecasts). The combination of information issued from these
parameters allows the assessment of icing risk potential, each information being complementary
to the others. From the satellite pictures, the cloudy areas are filtered, especially those with a
cloud top temperature in the interval [-15°C, 0°C]. The information calculated by the icing index
is added on these areas. The index overviews the points.

Comment: The application of concepts developed and tried in the US to European infrastructure
(models, satellite information) is beginning to bear fruit. The absence yet of a European funding
body equivalent to the US FAA is sorely felt.

                                                      Page 44
                                                                      TREND Newsletter No. 12

                Automatic Detection of Volcanic Ash from Satellite Images

                   T. R. Scott, S. Watkin, R.W. Lunnon and I.J. McNair
        UK Meteorological Office, London Road, Bracknell, Berkshire RG12 2SZ, UK

Volcanic ash clouds are a significant hazard to aviation, as an encounter can cause structural
damage to an aircraft and possible engine shutdown. Volcanic eruptions are routinely reported
to airlines through Volcanic Ash Advisory Centres, but these centres are often unaware of
eruptions that have occurred in isolated areas. A system for the remote detection of volcanic ash
is therefore under trial at the UK Met Office. This system is based on infrared Meteosat images
and numerical weather prediction data, and offers half-hourly monitoring of volcanoes over a
wide field of view which includes Europe, Africa and the Atlantic. The system is primarily
automated but detections require manual verification by forecasters because the single channel
system cannot distinguish between ash and water clouds. In the longer term the system will be
upgraded to exploit multiple channels from Meteosat Second Generation and provide automatic
distinction between the two cloud types. Development of a detection algorithm for this upgraded
system is already underway.

Comment: Given the great importance of timely warnings for Volcanic ash clouds, the
development of multi-spectral methods in detection is very promising.

                   In Situ-Measurements for Severe Convective Storms

                             O. Serrano (1), M.A. Marco (1)
      (1) Departamento de Experimentación en vuelo. INTA. Spain.

During the summer of 1997 a flight test campaign with an instrumented meteorological aircraft
was carried out in the Ebro Valley (Spain). Aircraft flight through severe storms containing hail
was conducted by INTA aircraft, collecting microphysics data. Temporal and spatial cloud
parameters have been measured during several horizontal traverses starting at cloud top and
finishing at cloud base. Meteorological C-band radar on ground was used as aircraft support.
The data has been obtained during a National Hail Suppression Programme. Information
provided by these flights permitted to make progress in acquiring knowledge of severe storms
characteristics in the Northeast of Spain overflown by one of the most widely used air corridors in

Comment: A new group entering the field of in-situ cloud physics experiments with a lot of
enthusiasm and amazing stamina. Using a mature Skyvan to penetrate severe storms must be
termed as dedication to science. It will be interesting to see results for non-Great-Plains-storms
for a change.

                      Windline for Parallel Runway Operations at SFO

           E. A. Spitzer (1), R. P. Rudis (1), J. N. Hallock (1) and G. C. Greene (2)
                      (1) Volpe National Transportation Systems Center
                              (2) Federal Aviation Administration.
               [+1] 617-494-3623

Wake turbulence poses a potential problem for side-by-side approaches to close-spaced parallel
runways (229 m apart at SFO), particularly when the ceiling is too low to permit visual operations.

The wake encounter risk depends upon the crosswind and the longitudinal displacement of the
aircraft pair. The SFO windline tracks landing wakes generated over the first 650 m of the

                                                   Page 45
                                                                       TREND Newsletter No. 12

runway, a region where few prior wake measurements have been made.

Comment: Selected as a representative of a very large number of papers on wake vortex
detection and prediction. It outlines rather nicely the rationale for this body of research.

                 ADWICE – Advanced Diagnosis and Warning System for
                             Aircraft Icing Environments

                   A. Tafferner (1), T. Hauf, C. Leifeld (2) and T. Hafner (3)
          (1) DLR-Institut für Physik der Atmosphaere, D-82234 Wessling, Germany,
       (2) Universitaet f. Meteorologie und Klimatologie, D-30419 Hannover, Germany,
                 (3) Deutscher Wetterdienst, D-63067 Offenbach, Germany.

ADWICE is being developed since 1998 in co-operation of DLR, the German Weather Service
(DWD) and the University of Hanover (IMUK). In principle, diagnosing and forecasting aircraft
icing would require information about the liquid water content in clouds. As nowadays routine
observations cannot provide such information. ADWICE merges forecast model data with
satellite, SYNOP and radar data in order to identify the icing risk region. Firstly, algorithms which
represent different icing scenarios use forecast fields of temperature, humidity and stability to
provide a first guess icing volume which generally overestimates the potential risk region. This
first guess field in then subjected to the APOLLO cloud detection scheme developed at DLR. It is
truncated horizontally in cloudless areas and from above in regions where the observed cloud top
temperature from APOLLO is warmer than in the forecast. After 2001 METEOSAT Second
Generation data will be used for cloud correction. Under certain conditions, radar and SYNOP
data allow correction of the icing volume from below. Finally, pilot reports will be used as an
additional data source and also to verify the icing forecasts. It is intended that after a testing
phase in spring 2000 ADWICE will be used operationally by DWD.

Comment: Somewhat similar to the French approach, using a different model and further
information from observations, it also expects significant improvements in discerning power of
the algorithm from multi-spectral methods possible with the advent of the next generation of
geostationary satellites, in this case Meteosat.

            Research for Improving In-Cloud Icing Forecast System in Canada

         André Tremblay (1), Anna Glazer (1), André Méthot (2), Paul Vaillancourt (1),
                         Stewart Cober (1) and George A. Isaac (1)
          (1) Cloud Physics Research Division, (2) Canadian Meteorological Centre
                 Atmospheric Environment Service, Dorval, Québec, Canada

Freezing precipitation and in-cloud icing are significant threats for aircraft operation and
represent a serious public safety problem. Icing on airframes reduces the ascending force of
aircraft and causes severe risks to flight operations. Because accurate predictions for these
natural hazards are of primary importance, many operational forecast algorithms need to be
improved. For example, statistically tuned icing algorithms use non-physical relationships, which
results in overforecasting icing threats. On the other hand, the Canadian operational scheme
proposed by Tremblay et al., 1996 (WAF, 11) is physically based but only provides yes/no
forecasts for icing events and doesn’t give any information for icing intensity.
Actual freezing precipitation forecast techniques only address the classical melting ice
mechanism (presence of a warm layer aloft) although the non-classical formation of freezing

                                                    Page 46
                                                                       TREND Newsletter No. 12

drizzle (without a warm layer) is commonly observed. In an attempt to improve icing and freezing
precipitation forecasts, a mixed-phase cloud scheme has been developed (Tremblay et al.;
Tellus, 18 A, 1996; MWR, 1999). This scheme is computationally fast, easy to implement and
has an operational potential. The scheme has been extensively tested by simulating virtually all
ice storms during the 96-97 winter season. The scheme has also been run in a quasi-operational
mode for the whole 97-98 winter season to provide forecasts and guidance for research aircraft
flights during the third Canadian Freezing Drizzle Experiment (Isaac et al.; BAMS, 1999). It was
also used for the Canadian Convair research aircraft missions over the Arctic Sea during the
Canadian phase of the FIRE.ACE experiment (April 1998). Finally the scheme was implemented
in Canadian Global Environmental Multiscale model to support the Alliance Icing Research Study
(AIRS) experiment held in Ottawa Canada (Dec/1999 – Jan/2000). A description of the scheme
is first presented. Examples of forecasts for the above-mentioned field projects are shown.
Comparisons with aircraft measurements are discussed.

Comment: The thorough and professional analysis of icing conditions in classical and non-
classical cases reflects both the long research tradition and the scale of the problem in
Canada, where icing conditions can be encountered virtually all year.

                     An Interactive Gridded Aviation Weather Database

                  M. F. Turcotte, R. Verret, V. Souvanlasy and M. Baltazar.
               Canadian Meteorological Center,

Canadian Meteorological Center in collaboration with Nav Canada. The ultimate goals of the
system are to help reducing time delays and improving efficiency of briefings at flight service
stations (FSS). Advancements in computer sciences and numerical modeling have brought
forward the capability of giving access to users to interactive, high-resolution aviation products.
This opens a whole area of new aviation weather products thus allowing a quick and intuitive
understanding of the actual and forecast aviation weather conditions. The gridded database,
which is the core component of the aviation weather system, includes aviation impact variables
such as turbulence, icing, freezing level, clouds, tropopause height and standard meteorological
fields. The database is generated from the Canadian Global Environmental multi scale model in
regional configuration on a 24 km horizontal resolution grid (soon to be 16 km), at 41 levels from
surface to FL400 and at every 3 hours from zero to 48-h projection time. The database is
updated twice a day at 00 and 12 UTC in real time. Once users have entered flight parameters,
such as departure and arrival points, alternate, estimated elapsed time, flight level and departure
time, meteorological products can be requested, all tailored to each particular flight, in plan view
and/or vertical cross section along the route and downloaded to the users. This interactive
system has been delivered in fall of 1999 for a beta test with twenty designated Canadian pilots
and personal of two FSS.

Comment: The size of the country, its large number of small airfields compared to the population
explains the need for an automated, easy-to-use information system for pilots based on NWP
output. The user-friendly design of the system should guarantee acceptance by pilots, hence its
common use and positive impact on flight safety.

              In flight Icing Forecast at the Canadian Meteorological Center

                                 M. F. Turcotte, R. Verret,
               Canadian Meteorological Center,

Amongst the Aviation Impact Variable (AIV), icing deserves special attention because of the
hazards it may pose to safety of passengers and crew members and also because of its impact

                                                   Page 47
                                                                       TREND Newsletter No. 12

on the efficiency of flight operations with the additional costs associated. A new icing algorithm
has been implemented at the Canadian Meteorological Centre (CMC) since the arrival of the
Canadian Global Environmental multi scale model (GEM) in regional configuration with a 24 km
horizontal resolution. This new algorithm is based on cloud liquid water content as forecast by
the driving model. This algorithm uses direct model output of the driving model to diagnose the
available supercooled liquid water content available for icing, based on the parameterization of
the mesoscale cloud physics and the precipitation formation processes. Liquid water content,
vertical motion and temperature are analyzed to determined whether or not there is supercooled
liquid water and hence icing or no icing. The supercooled liquid water algorithm has been found
to be superior to the Appleman scheme which is based solely on temperature, vertical motion
and dew point depression and is known to overforecast icing area and to misrepresent the
occurrence of icing as a function of temperature.

Comment: The backbone component of the system described above is the algorithm extracting
the essential information from model fields including liquid water. As in many such schemes,
success will critically depend not only on the microphysics getting the SLW content right, but also
the intensity and location of the synoptic structures leading to these conditions!

         Towards the Improvement of In-Cloud Icing and Freezing Precipitation
                   Forecasts: Comparison of Four Cloud Schemes

                         Anna Glazer (1) and André Tremblay (1)
                           (1) Cloud Physics Research Division
                 Atmospheric Environment Service, Dorval, Québec, Canada

The increasing complexity of numerical weather prediction models imposes increasing demands
on their microphysical components. The ability to forecast precipitation amounts and types, as
well as clouds and their phase, is becoming a necessity. In particular, forecasts of supercooled
liquid water are essential to predict in-cloud icing and freezing precipitation. Four explicit cloud
condensation schemes, including the operational scheme at the Canadian Meteorological Centre
(CMC), were analyzed and evaluated. For the 1996-1997 winter season, 16 typical cases were
selected. Forecasts of the Canadian Mesoscale Compressible Community model, at a 35-km
resolution, were evaluated for a domain covering the continental US and Canada. Clouds
forecasts were evaluated by comparing model-predicted with satellite-derived cloud top pressure.
Precipitation was verified by a direct comparison with surface observations. The skill in
predicting temperature and humidity was assessed by a comparison with radiosonde data. All
schemes underpredict the cloud cover and particularly the upper and mid level clouds. The low-
level clouds, however, are overpredicted by all schemes. The operational CMC scheme
overforecasts precipitation accumulation during the considered period. The remaining schemes
have a comparable skill and are better than the operational scheme. Comparisons with
radiosonde data have shown that temperature is well predicted and weakly depends on the
choice of the cloud physics scheme. The model develops a dry bias with time in the middle
atmosphere for all cloud schemes. Two of the four schemes explicitly predict the supercooled
liquid water. One of these, the mixed-phase cloud scheme (Tremblay et al.; Tellus, 18 A, 1996;
MWR, 1999) better reproduces typical cloud structures. Its computational cost is equivalent to
the actual operational scheme and it shows a superior forecast skill for winter storms. The
scheme has a potential to infer quantitative prediction of icing type and intensity. It was
successfully applied to predict freezing precipitation in winter storms.

                                                   Page 48
                                                                     TREND Newsletter No. 12

         Evaluation of Two-Cloud Scheme’s Ability at Forecasting Aircraft Icing

        Paul A. Vaillancourt, André Tremblay, Stewart G. Cober and George A. Isaac
                         Cloud Physics Research Division, Canada

The formation and lifetime of supercooled water (SLW) in clouds is important to study because of
two impacts on human activity. First, in-cloud icing causes serious loss of aircraft performance.
Second, SLW is directly related to the formation of freezing precipitation. SLW in clouds is very
common as shown by the numerous PIREP's reporting icing and detailed observations made in a
cloud physics research context. A useful forecast of SLW is one that would provide reliable
information on the extent and amounts of SLW. Furthermore, information on the size distribution
of the liquid drops is also crucial to help the aviation community in assessing the danger posed
by the icing conditions. At least one case of aircraft crash resulted from icing associated with
supercooled large drops or freezing drizzle. A research project was recently set out to develop a
cloud scheme capable of forecasting SLW and providing useful information on the size
distribution of liquid drops. As a first step, direct comparisons between observations gathered
during 11 research flights (Canadian Freezing Drizzle Experiment I), and “virtual flights” through
the model data of these simulated cases were made. It was found that the MC2 dynamical
model and a one equation mixed-phase cloud scheme (Tremblay et al. 1996, Tellus, 18A) gave
sufficiently good forecasts of the location and amounts of SLW to justify efforts to incorporate
more detailed microphysics into the cloud scheme. As a second step, whether a detailed six
equation cloud scheme (Reisner et al. 1996, QJRMS, 124) can provide more accurate forecasts
of SLW in a large scale (operational) context was being studied. A test whether the new
information provided, e.g. presence of large drops, is reliable was conducted. Conclusions would
be based on the comparison of over 30 research flights (CFDE I and III) with the simulated virtual
flights. Based on these conclusions the simple cloud scheme would be improved. Results and
conclusions from these model-data comparisons were presented at the conference. Where
humidity and temperature are favorable to icing conditions radar echoes are added to improve
this forecast. The radar echoes indicate the presence of water in suspension or precipitating, in
liquid or solid form without distinction. The picture from SIGMA is available every quarter of an
hour. From pilots reports collected during the winter 1997-1998, SIGMA and the French icing
index were evaluated and their efficiency compared to the standard aeronautical operational alert
messages called SIGMETs.

Comment: As above.

             Providing Air Traffic Control with Convective Weather Products

                               JM. Carrière (1), F. Autonès (1)
                                      (1) Météo-France
            : +33-561078209

To anticipate aircraft diversions, localizing and nowcasting thunderstorms is of utmost
importance. Furthermore, in order not to overload air traffic controllers, they must be provided
with an easy-to-use, already interpreted and formatted-as-usual information. A project is
presented which tries to address these needs. Information from the French weather radar
network are merged with those from the lightning detection network of Météo-France to identifiy

thunderstorms. The movement of convective cells is then diagnosed thanks to an algorithm
based on radar images time series. This allows nowcasting of the convective cells by simple
extrapolation, allowing for curved trajectories. This information in presented on a customized
terminal for air traffic controllers using the same geographical projection and information as the
one used with aircraft surveillance radars. Experiments were carried out in 1998 with both the

                                                  Page 49
                                                                        TREND Newsletter No. 12

en-route air traffic control center in Paris-Orly and the approach of Paris-Roissy. The correlation
between convective cells and aircraft diversions was studied. Based on the experience gained
during the 1998 experiments, the ergonomy of the product could be improved and it was then
used by air traffic controllers in real-time in 1999 in Paris-Orly and Paris-Roissy and presented to
other control centres.

Comment: A first step, using extrapolation over homogeneous terrain only (very sensible to
avoid the mountainous regions, where linear extrapolation in time of convective events is bound
to fail in many cases!!!). An important step towards recognition of the usefulness of convection
forecasts in ATM in Europe.

           Sensitivity of a Wake Vortex Model to Meteorological Input Variables

                              Michael Frech and Frank Holzaepfel
             DLR, Institute of Atmospheric Physics, D-82230 Wessling, Germany.

A large eddy simulation (LES) of the decay of wake vortices in the convective boundary (CBL) is
used to investigate the predictive capabilities of a simple wake vortex transport and decay model
(VORTEX) based on Corjon and Poinsot (1996). The sensitivity of VORTEX forecasts on the
quality of meteorological input parameters is studied, since only limited meteorological
background information is available operationally. Three simple methods were tested to account
for the structure of the CBL. Area-averaged CBL profiles computed from LES and simple model
assumptions on the structure of the CBL using surface data are used as input. A simple third
approach uses meteorological LES data at z=16 m which is a height representative for tower
measurements. For the trajectories, the first and second approach yields similar results. The
predicted residence corridors are erroneous for some time steps. The errors are quantified. The
third approach yields conservative results, which is an important result from an operational point
of view. For the CBL case, this study shows that surface measurements would be sufficient to
predict WV trajectories along the glide path with acceptable accuracy. Recently published
parameterizations of the turbulent decay of wake vortices were also tested. A weak sensitivity of
the predicted corridor on the type of parameterization was found. The precise knowledge of the
meteorological background conditions is more important.

Comment: A good example for the very engineering-like approach to wake vortices and their

                   Global Triangulation of Intense Lightning Discharges

                                          M. Fuellekrug
                Institut für Geophysik, Feldbergstr. 47, D-60323 Frankfurt/Main.
         +49 (69)798-23280

A global network of three electromagnetic measurement instruments is used to simultaneously
record time series of globally observable Extremely-Low-Frequency (ELF) magnetic field
disturbances which propagate with little attenuation around the globe within the Earth-ionosphere
cavity. The triangulation of individual lightning ashes results in a picture of the temporal evolution
of intense lightning discharge occurrences on the planetary scale during April 1998. The
lightning ash charge moments are calculated with the short pulse approximation of the normal
mode expansion. The majority of the triangulated lightning discharges exhibit charge moments
with a potential to excite transient optical emissions in the mesosphere, known as mesospheric
sprites, and ¢ 5-20 % may account for air breakdown at sprite altitudes in ¢ 50-70 km height.

                                                    Page 50
                                                                      TREND Newsletter No. 12

Radars can’t be everywhere, so lightning discharge triangulation is a cheap and efficient way to
fill in the gaps.

            Verification of Short Range Thunderstorm Forecasts Using Radar
                  Data to Assess their Benefit to the Aviation Community

                                         D. J. Hoad
                  The Meteorological Office, Bracknell, England, RG12 2SZ.
             [+44] (0)1344 856099

On the afternoon of 20th July 1998 severe thunderstorms occurred over the south-east of
England. The airspace sector over that region had to be closed for safety reasons at fairly short
notice. This caused severe disruption to air traffic. If severe thunderstorms can be forecast early
enough, Air Traffic Management (ATM) authorities can, in future, take action to maintain safety
levels, e.g. by re-routing aircraft to avoid the worst affected areas. The aim of this study was to
use radar data to verify short-range thunderstorm forecasts produced by the UK Met. Office’s
GANDOLF (Generating Advanced Nowcasts for the Deployment of Operational Land-based
Flood forecasts) system. The method of verifying the forecasts is intended to replicate the way in
which the forecasts could be used in an aviation context. Specifically, both the forecast data and
the radar data used to perform the verification are processed to calculate conditions under which
an airspace sector might be closed. The output from this study consists of forecast accuracy
statistics, which can be used by ATM authorities to determine the value of the forecasts in the
context of their own operations.

Comment: Again very similar to the French approach, this time focussing on the proof of concept
of gaining air traffic efficiency by using appropriate convection forecasts. European ATM units
obviously need to be convinced of what their US colleagues accept as a home truth.

                 Radar Windprofilers in the Austrian Aviation Metservice

                             A. Lanzinger (1) and W. Langhans
              (1) Austro Control, MET Innsbruck, (2) Austro Control, FT Vienna

Austro Control, the company responsible for maintaining air traffic safety in Austria, operates
three 1290 MHz radar windprofilers at Vienna, Innsbruck and Salzburg airports, respectively. The
paper presents the most important operational applications for detection of hazards mainly for
landing and departing aircraft. Relevant situations are, for instance, strong wind shear associated
with low level inversions, also with a stratus cloud layer (icing), and Foehn winds with highly
turbulent conditions in Innsbruck. Also, limitations of windprofilers for the diagnosis of other
hazards are discussed.

Comment: Topographically caused wind shear and turbulence are a serious problem for a
airports in a mountainous country, hence the investment in advanced technology for early

    Fine Scale Weather Modelling for Turbulence Prediction at Norwegian Airports

                                             I. Lie
        Norwegian Meterological Institute, P.O.Box 43 Blindern, N-0313 Oslo, Norway

Many small airports in Norway are located in deep valleys or other difficult locations where
terrain-induced turbulence is a major problem for the landing conditions. In order to give a short

                                                   Page 51
                                                                       TREND Newsletter No. 12

term prediction of turbulence conditions a model system consisting of three parts was
constructed: the operational HIRLAM model running at 10km horizontal resolution and covering
Scandinavia and the Norwegian sea, provides the large scale forcing. Hirlam provides initial and
boundary data to the Canadian MC2 nonhydrostatic model running at 1km horizontal resolution
and covering an area (60-80 km wide) around the airport. The MC2 model produces detailed
meteorological fields as well as quantities like turbulent kinetic energy. A special very fine scale
turbulence model with variable grid size down to 50 m, is then used for detailed predictions
around selected terrain features. The model system is described in more detail in the paper and
results are shown from the MC2 model for the Varnes airport near Trondheim. Results for the
Spitsbergen airport were shown and the MC2 model was also run for the case of the aircraft
accident the 29.August 1996.

Comment: A fascinating study stretching high-resolution modeling to the limits, giving results
that tally well with in-flight observations.

                        A Physical Approach to Estimate Wind Gusts

                                         O. Brasseur
        Institut d’Astronomie et de Géophysique G. Lemaître, Université catholique de
               Louvain, Chemin du Cyclotron 2, 1348 Louvain-la-Neuve, Belgium
                / Fax: +32-10-474722

The determination of wind gusts appears important for operational weather forecast as well as for
studies related to extreme climatic events. In this framework, a new wind gust estimate (denoted
WGE) method has been developed considering the properties of the boundary layer. More
specifically, it is based on the following assumption: surface gusts are explained by the deflection
of air parcels flowing in the boundary layer. The deflection process is attributed to the presence
of large turbulent eddies. Consequently the WGE method namely requires as input the turbulent
kinetic energy field computed by atmospheric models. In addition to the estimate of gusts, this
technique includes the specification of a validity interval around the estimate. At first tested on
extreme situations of explosive cyclogenesis, the WGE method has been able to produce
accurate estimates with a typical error of 5 ms-1. This technique has been also successfully
applied for longer periods (from January to March 1990) in order to check its behavior for a
variety of situations and to assess the statistics of gusts. The validity interval has proved its
usefulness with a reliability rate of about 80% for the prediction of daily gusts.

Comment: small test sample; the reliability rate of 80% is perhaps not as good as it sounds
given the relatively large error estimate of 5 ms -1.

                      Physical Initialization of the LM with Radar Data

        G. Haase (1), P. Gross (1), C. Köpken (2), C. Simmer (1) and W. Wergen (2)
        (1) Meteorological Institute, University of Bonn, Auf dem Hügel 20, D-53121
       Bonn, (2) German Weather Service (DWD), P.O. 100465, D-63004 Offenbach.

The assimilation of precipitation in numerical weather prediction models is problematic due to the
strong interconnections of wind and humidity fields. Inconsistent initial fields cause spin-up
problems of the hydrological cycle. These problems mostly vanish several hours past
initialization when dynamics and thermodynamics are in balance. A physical initialization scheme
for the Lokal-Modell (LM) of the German Weather Service is currently under development. This
scheme uses as input data operational measured 2-d radar reflectivities and synoptical

                                                   Page 52
                                                                        TREND Newsletter No. 12

observations. The analyzed profiles of vertical wind, temperature, specific water vapor and cloud
water content are modified within the LM, so that the observed precipitation is produced by the
model. The cloud base and top heights are taken from the synoptical observations. As boundary
conditions it is assumed that the precipitation flux at the cloud top is zero and at the cloud base it
is derived from the radar composite images. Inside the cloud the precipitation flux decreases
linearly with height. The vertical wind induced by precipitation is calculated with a simplified
precipitation mechanism using the vertical gradients of the partial water vapor density and the
precipitation flux. Inside the cloud the air is saturated. The physical initialization scheme is
explained as well as present preliminary results for a case study. These are compared with a LM
run using a latent heat nudging technique.

Comment: How do you get cloud tops from synoptic observations? This scheme probably needs
very good QC on the radar reflectivity data to give satisfactory results.

                    Numerical Simulations of Mountain Wave Generation
                               and Breakdown During MAP

                             J. D. Doyle, Naval Research Laboratory

A three-dimensional, non-hydrostatic numerical model, COAMPS, is used to investigate the
dynamics and characteristics of mountain wave generation and breakdown resulting from flow
over the complex topography of the Alps during the Mesoscale Alpine Programme (MAP). The
high-resolution model simulations are examined in detail with research aircraft flight level and
lidar backscatter data. The results suggest that wave breakdown occurs most frequently in low-
levels and may include energy transport to higher frequency through nonlinear processes.
Boundary layer effects are shown to be important in the wave launching process

Comment: Emphasis of the paper not quite as in the title and abstract, however very good
overview of mountain wave related activities during MAP and first results. There were nice cases
of mountain waves, but considerably less than expected wave breaking was observed.

            Experimental Investigation of a Shallow Foehn Event during MAP
                   Experiment Using a Ground Based Doppler Lidar

             P. Drobinski (1), A.M. Dabas (2), P.H. Flamant (1), A. Delaval (1), M.
           Aupierre (1), P. Delville (1), B. Romand (1), C. Boitel (1), J.M. Donier (2)
                                          and C. Loth (1)
        (1) Laboratoire de Meteorologie Dynamique (IPSL/CNRS), Palaiseau, France,
                         (2) CNRM, Meteo-France, Toulouse, France.

Strong, gusty downslope winds are observed in many mountainous regions of the world. They
are an outstanding feature of the weather during fall and spring in the northern Alps, where they
are called foehn, and are linked with an increase of temperature and a decrease of humidity and
cloudiness. Such meteorological events have been documented in the Rhine Valley during the
field experiment of the Mesoscale Alpine Programme (MAP) from mid-September to mid-
November 1999. The boundary layer dynamics in the Rhine Valley was documented by means
of lidars, surface meteorological stations, soundings and airborne in-situ measurements. This
paper presents an analysis of the local features of a south shallow foehn event during IOP 5 (2
October 1999), as observed by a ground-based Doppler lidar, with respect to the large-scale and
mesoscale observations. The transition between a no-foehn event and a foehn event is
investigated. The Doppler lidar recorded in detail the whole event lifetime including wind gusts,

                                                     Page 53
                                                                        TREND Newsletter No. 12

flow reversal. Such structures have not previously been observed with comparable detail by
conventional in-situ instruments.

Comment: Nice detailed case study. Comparison with “conventional” observations often gives
confidence in data, in some cases (of disagreement) however reminds you how careful you have
to be with measurements (“Wer mißt mißt Mist”).

               High-Resolution Numerical Simulation of Alpine PV Banners:
                                First Results from MAP

                Michael Sprenger (1), Christoph Schär (1) and Robert Benoit (2)
            (1) Swiss Federal Institute of Technology, GI-ETH, Zurich, Switzerland,
               (2) Recherche en Prévision Numérique, Environnement Canada.

Atmospheric flow past isolated mountains and associated flow-splitting processes may lead to
the formation of lee-vortices and shear-lines downstream. These features are associated with
anomalies of potential vorticity (PV). In the case of complex topography such as the Alps, the
wake structure may be characterized by numerous elongated filaments of PV (PV banners) which
extend downstream from the topography. In this study high-resolution numerical simulations of
Alpine PV banners are presented, including inter-comparison against observational data
collected in the MAP field phase during IOP4 on October 1, 1999. The numerical simulations are
conducted with the Canadian MC2 model at a horizontal resolution of 3 km and with 50 levels in
the vertical. Consideration will be given to both forecast and hindcast simulations. The former
were driven by the DWD/SMI forecasting chain and were conducted in real-time during MAP.
The latter are driven by operational ECMWF analysis. The simulation results and their evaluation
against observations will be utilized to judge the suitability of the modelling strategy and to assess
the underlying dynamics of Alpine PV banners.

Comment: PV banners are an interesting phenomenon in so far as they were seen in numerical
model output fields first and then sought (and apparently found) in nature.

                                                    Page 54

Shared By: