Euroforecaster 2017 Newsletter

The European
Forecaster
Newsletter of the WGCEF N° 23

September 2018
C ontents
3 Introduction
Minutes of the 23rd Annual Meeting of the Working Group on Co-operation

4
Between European Forecasters (WGCEF)
Sting Jets and other processes leading to high wind gusts:

10
wind-storms “Zeus” and “Joachim” compared
16 Forecasting Freezing Rain in the UK – March 1st and 2nd 2018
24 The Extreme Wildfire, 17-19 July 2017 in Split
30 Changing the Way we Warn for Weather
Storm naming: the First Season of Naming by the South-west Group:

33
Spain-Portugal-France
38 Can we forecast the sudden dust storms impacting Israel's southernmost city?
45 The 31st Nordic Meteorological Meeting
46 Representatives of the WGCEF
Cover:
Ana was the first storm named by the Southwest
Group (Spain, Portugal, France) during winter
2017-2018. It affected three countries with great
impacts.
Crédit Météo-France
Printed by Meteo France

Editors Stephanie Jameson and Will Lang, Met Office
Layout Kirsi Hindstrom- Basic Weather Services
Published by Météo-France
COM/CGN/PPN - Trappes
I ntroduction
Dear Readers and Colleagues,
It’s a great pleasure to introduce the 23rd edition of our newsletter ‘The European Forecaster’. The publica-
tion is only possible due to the great work and generosity of Meteo-France, thus we want to express our
warmest gratitude to Mr. Bernard Roulet and his colleagues. We kindly thank all the authors for submitting
articles, particularly as they all work in operational forecasting roles and thus have only limited time for
writing an article. Many thanks go to Mrs. Stephanie Jameson, Dr. Will Lang, Mrs. Evelyn Cusack and Ms.
Laura Paterson for reviewing the incoming articles. Last, but not least, we want to thank Dr. Andre-Charles
Letestu for updating our WGCEF website www.euroforecaster.org regularly.
During recent months, the weather has been quite exciting. A strong zonal flow during late autumn and
early winter brought many storms to western and north-western Europe (e.g. storms Caroline, Dylan and
Carmen) and to south-western Europe (storms Ana and Bruno) as well. Some storms also affected central
Europe, like storms Eleanor or David. Most of these weather events were associated with strong wind gusts,
but some were associated with high precipitation amounts too - especially in the higher parts of the Alps,
where new snow records were observed. By the end of January, the weather regime changed; during
February and March cold continental air from the north-east influenced many parts of Europe. Whereas in
central and eastern parts of Europe these air masses were often dry, over the United Kingdom and Ireland,
blizzard-like snowfall conditions were observed (referred to as the ‘Beast from the East’ by the media).
During April the winter conditions turned rapidly into summer ones. Very warm air masses from the south,
and later on from the east, covered central and eastern Europe and even some parts of Scandinavia and the
UK. This air was unstable, especially over central Europe, and therefore many thunderstorms occurred. As a
result of the weak flow, these thunderstorms moved very slowly or even stayed over the same region for
many hours, therefore accumulating more than 100mm of rain within a short period. This led to many flash
floods and landslides in those areas.
As forecasters we are taking a leading role in supporting people, decision-makers and civil protection agen-
cies in being prepared for these severe weather events - to minimise, or even avoid, damage to humans and
infrastructure. To keep, or even improve, the high quality of our forecasts, cooperation and scientific
exchange between the different NMSs is essential, which is one important goal of our working group.
Finally, we want to send many thanks to Mrs. Julianna Drozdzynska, Mr. Piotr Manczak and their colleagues
who organised and hosted our excellent 2017 meeting in Warsaw.
We hope you find this newsletter enjoyable and informative.
Best regards,
Christian Csekits and Jos Diepeveen

Chairpersons WGCEF
The European Forecaster 3

Minutes of the 23rd Annual Meeting
of the Working Group on Co-operation Between
European Forecasters (WGCEF)
Thursday 12th – Friday 13th October 2017
Warsaw, Poland
OGRODNIK, Michal - IMGW/Hydrometeorolo-

gical Service of Poland
PATKAI, Zsolt - OMSZ/Hungary
RALIENE, Vida - Hydrometeorological Service
of Lithuania
REY, Jaime - AEMET/Spain
ROULET, Bernard - Meteo France
SANDEV, Marjan - Czech
Hydrometeorological Institute
SKELBAEK, Michael - DMI/Denmark
List of Participants (in alphabetical order): VANHAMEL, Thomas - RMI/Belgium
Participants via videoconference:
BLAAUBOER, Dick - Eumetnet Mr. Attilio DI DIODATO (Italy), Mr. Panos GIANNOPOU-
CSEKITS, Christian - ZAMG/Austria LOS (Greece); Mr. Andre-Charles LETESTU
(Switzerland); Mr Mats Johansson, SHMI (Sweden).
CUSACK, Evelyn - MET EIREANN/Ireland
Meeting Venue:
DIEPEVEEN, Jos - KNMI/The Netherlands
Institute of Meteorology and Water Management –
DOUBLET, Karen Helen - Met Norway National Research Institute
DROZDZYNSKA, Julianna - MGW/ ul. Podleśna 61
Hydrometeorological Service of Poland 01-673 Warszawa
HAUSEN, Robert - DWD/Germany Thursday, 12th of October
HEWSON, Tim - ECMWF

HLADNIK, Veronika - Slovenian Environmental Introduction
Agency
IMGW Deputy Director General Piotr Kowalczak
JAMESON, Stephanie - UK Met Office welcomed the WG to Warsaw and introduced the
JONASDOTTIR, Elín Björk - Icelandic Met Office participants to the IMGW facilities.
KALIN, Lovro - DHMZ/Meteorological and As WG Chair, Christian thanked Piotr and made his
Hydrometeorological Service of Croatia introductory remarks. Christian and Jos then
welcomed the new members to the group: Jaime
KRUMINA, Laura - Latvian Environment,
Rey from Spain, Laura Krumina from Latvia, Elin
Geology and Meteorology Centre (LEGMC) Bjork Jonasdottir from Iceland and Stephanie
LAINE, Mikko - FMI/Finland Jameson from the UK.
LEITAO, Paula - PMA/Portugal
In an overview of group activities, Christian noted
MANCZAK, Piotr - MGW/Hydrometeorological that WGCEF delegates were Co-Chairs in the
Service of Poland Forecasters’ Session of the EMS Annual Conference,
4 The European Forecaster

which took place in Dublin from 4 th to 8 th of • At the beginning of October, KNMI launched new app.
September 2017 (Evelyn Cusack and Christian • Development of own visualisation and production
Csekits). Furthermore the WGCEF Chair is member of system, Geoweb.
the PSC, the programme and science committee of
the EMS. Karen Helen Doublet /Norwegian Met
Service
WGCEF was also heavily involved in the Eumetnet • Several episodes of extreme precipitation
Task Team on Improved Services and Impact Based events/flooding during the last year, and there is a
Warnings (Laura Paterson, Jos Diepeveen and project to update extreme weather warning plans.
Christian Csekits). WGCEF has also participated in • NWP cooperation between SMHI and METNorway will
the Eumetnet Drafting Team on forecasting (Jos include FMI this year. Ensemble system MEPS 2.5 km,
Diepeveen and Christian Csekits). 10 members. Cooperation on a HPC-solution as well
• New trajectory models for drifting objects. Flexible
The meeting agenda was agreed and actions from
solution, different modules: drifting oil slick, persons
the last meeting were agreed.
in water, different objects. General input data: current,
waves and wind
Members’ Updates • Developing finemesh models in coastal areas and in
fjords:
In this round table session, participants were invit- - grid size 800 m, wave model and ocean
ed to introduce themselves and give a short update model.
on new developments within their NMS. - grid size 2.5 km atmospheric model.
Cristian Csekits, ZAMG/Austria Robert Hausen DWD/Germany

• During nights there is now only one technical shift • Change in DWD-law regarding providing free data
(instead of two; operational since summer 2017) • The “Sinfony“project is developing an integrated
• New ZAMG-App for fire brigade, police and ambu- forecast system
lance • Add-on DWD-App: warnings and reports for flood-
• INCA now uses forecast data from AROME (instead ings, storm tides and avalanches
of ALARO) from 2018/19 on
• New operational forecasting and emergency room (7 Tim Hewson, ECMWF
instead of 4 positions) ready in November 2017 • Agreement reached that ECMWF computing facilities
– notably the next -Supercomputer – will be re-located
Evelyn Cusack, Met Eireann to Bologna in Italy
• New Assistant Director: Dr Sarah O'Reilly • Two new model cycles were introduced since the
8 new meteorologists are training at KNMI: last WGCEF meeting. New seasonal forecast system
October/November/December. KNMI trainers coming (System 5) will become operational in November.
to Dublin in January and February. • For the recent devastating hurricanes in the North
• European Met Society in Dublin was a great Atlantic ECMWF forecasts, though not perfect, outper-
success.....the largest ever over 850 (110 from formed forecasts from other centres (including special
Ireland). Met Eireann had a follow-on 1-day seminar in high resolution LAMs).
October 2018 which energized all the staff. We gave
15-minute presentations from EMS. Veronika Hladnik, Slovenian
• We continue to automate remote rainfall stations in Environmental Agency
mountains etc and 60 are being tested. • Implementation of Aladin-SI model version CY40
• Facebook account Arso Vreme
Jos Diepeveen KNMI/Netherlands • Rejuvenation of the forecasting team Facebook
• There is ambition to develop to an Early Warning account Arso Vreme.
Centre
Devastating hurricane Irma hits the SSS Islands in the Stephanie Jameson, UK Met Office
Caribbean, especially St. Maarten • Weather challenges over the past year:
• 2017 is the 200 year anniversary of the birth of Buys - Storm Naming reaches ‘Ewan’ for the
Ballot, founder KNMI in 1854 (1817-1890) 2016/2017 season. Particularly notable impacts
• New Dual-pol radar in Herwijnen and update BULL associated with Storm Angus (flooding) and Storm
supercomputer Doris (winds).

- Aviation impacts due to dense fog over • Ongoing refurbishment of forecasting office
Christmas period. • Regional Maritime Centre established in Split.
• UM Model Upgrade:
Laura Krumina, Latvian Environment,
- GM: Horizontal resolution increases
Geology and Meteorology Centre (LEGMC)
- UK High Resolution (UKV): Introduction of hourly
• Continued renovation of the meteorological and
cycling, change from 3 to 4 Dimensional Analysis (4D-
hydrological network – reduction of manual observa-
Var), higher resolution lateral boundaries.
tions, but real time data flow and web cameras –
- Ensemble: Increase in number of ensemble
completion is planned by the end of 2018;
members, increased horizontal resolution.
• Development of new automatic verification system
• Other Business Development: for general forecasts; verification of forecasts on daily
- Transform & Efficiency Programme, increase in basis – fast feedback on models and forecasters
Media & Aviation Consultancy Services, developing performance;
involvement with the UN Ops & Crisis Centre. • Revision of hydrological and meteorological warn-
ings criteria; closer cooperation with civil protection
Mats Johansson, SMHI Sweden authority and municipalities. The first steps to new
• Due to new products (road condition forecasts) there warnings dissemination and verification system;
has been one more night shift for my group since • Development of the personnel competency mainte-
October (2 total). SMHI also have 2-3 night shifts in nance and assessment system for the Forecasting and
Stockholm working mostly with aviation forecasts. Climate Department (3 years cycle); competency
• Upgrading radar stations, 6 of 12 are done. The plan assessment for general and marine forecasters at the
is that all of them will be modified by late 2018. end of 2017;
• The work to upgrade observation stations is ongoing • More activities in the NORDMET with an emphasis
and will continue also next year. on the work in the NAMCON consortium, developing
• Running a ten member high resolution EPS (Arome) closer cooperation with Estonian Environment
together with Norway and Finland. Finland joined Agency. Priorities - field of aviation forecasting
earlier this year. (common SWC, etc.) and IT solutions;
• Impact based warning project is running and Phase • Analysis of past and future climate for Latvia, activi-
I will be done in July next year. Parameters that are of ties were performed in cooperation with FMI – basis
most interest are wind, snow, flooding and high sea for adaptation plans;
levels. • Development of hydraulic model application, flood
• From the coming ice season we will work together risk maps for flood risk
with FMI with the ice-chart for the Baltic Sea • areas and hydrological simulation and forecasting
• Aviation forecasting group has been working togeth- system. Activities are done in cooperation with SYKE
er with DMI for a couple of years. The group has also (FINLAND);
started cooperation with FMI. • Use of SENTINEL satellite images to detect ice cover
/ ice jams in inland waters, flooded areas.
Elin Bjork JÓNASDÓTTIR, Icelandic Met
Office Mikko Laine, FMI Finland
• Harmonie is run in two separate instances, one by • A joint Nordic (Norway, Sweden, Finland) model
IMO the other in collaboration with DMI (IGA) MEPS has replaced our own Harmonie model.
• New satellite images, including SCAT • New weather warning system, now some of our
• Visual Weather and Moving weather updates pend- warnings are given up to 5 days ahead.
ing. Free choice of area for warnings.
• CAP warning system for severe weather. • Meteorologist are co-operating with
research/developers/sales –personnel.
Lovro Kalin, DHMZ Croatia • New services/products: meteorologists were on site
• There are around 15 forecasters (2 apprentices) on Nordic World Ski Championships in Lahti.
There has been an increase of daily work – and
increasing pressure on staff/schedules (including Paula Leitao, IPMA/Portugal
health issues) • New lightning detector network is operational
• 2017 was a tough season (cold, hot, drought, fires) • Pre operational severe weather warning, which
• VisWe is used in operational forecasts for visualisa- congregates different phenomena resulting from
tion and forecast production. convection

• Forest fire risk based on impacts is operational Zsolt Patkai, OMSZ/Hungary
• Improving connections with other national authori- • A new Aviation Meteorology portal has been
ties and knowledge in order to Impact Based launched (aviation.met.hu). Widespread weather
Forecasts and Warnings information available: Bulletins, Observations,
• Low level area forecast (GAMET – over mainland) Regional Area Forecast for VFR flights, Tons of fore-
• Aerodrome warnings cast maps based on AROME (e.g. thermals, wave
• Contact with UK and Spain in case of significant forecast, wind, cross-sections, convective indices).
weather over the border. • Staff number is stable, but people are coming and
• On the 17 June a forest fire caused 64 casualties going, only last year 5 forecaster colleagues decided
• Later this year to leave. The main reason: salaries are very low, they
- Contact with Morocco in case of significant have not been raised since 2008.
weather over the border. • This year WMO released a new International Cloud
- Pre operational AUTOMETAR at small aerodromes Atlas. Linked to this, our Met. Service also released a
during night time (Pico, Graciosa e Corvo – Azores) book called Cloud Atlas (Cloud Atlas - OMSZ)
- Pre operational severe weather warning by
counties Vida Raliene, Hydrometeorological
• Planning for near future
Service of Lithuania
- New radar for Madeira
• New supercomputer SGI ICE X Replacing Hirlam 4km
- New lightning detector network at Madeira and
by Harmonie 2,5 km.
Azores
• TRAINING COURSE “BALTIC+“ on 14–16 March in
- New meteorological visualization and process-
Vilnius, involving 19 Forecasters and 9 Teachers from
ing system
7 countries and EUMETSAT. The information about this
Andre-Charles Letestu, Meteoswiss event was published in the “EUMETRAIN training
• Simplification of warnings bulletin“ in July's issue.
AutoMETARS during the night at Geneva’s airport. • The presentation "Baltic+ 2017 course on
New apps: type of precipitation in radar animation. Convection – Collaborative Effort for High Quality
New forecasters, employed as AMFe later becoming Convection Training" was made in ESSC-2017.
AMFm.
Jaime Rey, AEMET/Spain
Piotr Manczak, IMGW/ • Since June Harmonie-Arome cycle 40 is fully opera-
Hydrometeorological Service of Poland tional in AEMET´s Bull supercomputer. There are often
• Works on nowcasting tool SEiNO are in progress. problems in convective situations
Particular modules of the SEiNO system are responsi- • Warnings issuance in CAP format to Meteoalarm
ble for different calculations. For example: GRS – esti- and Civil Protection authorities is in test phase.
mation of precipitation field based on rain gauges, • Ninjo licenses expired in January and were not
radar and satellite data; SCENE – detection of convec- renewed. It didn´t fulfill our expectations (e.g. only
tion and precipitation nowcasting using extrapolation one product was generated through Ninjo for external
vectors; SNOF – Fourier analysis is used to particular users)
hydrometeors (similarly to STEPS models in other
countries); ENSEMBLE and PROB modules of this Bernard Roulet, Meteo France
nowcasting system are tested. • Models
• Works on forecast simulator have begun this year. - Global model ARPEGE : test of new scheme for
It's going to be a training tool which will allow users to convection was not yet conclusive (so not included in
work operationally on archive data. Thera are works in future version)
progress on interface and database. This simulator - Ensemble High Resolution Model AROME :
will be used in the future to examine candidates for training plan for forecasters and test of new algo-
the job, to raise and periodically check forecasters' rithm to select coupling models in ARPEGE ensem-
qualifications. ble.
• In 2015 and 2016 we experienced staff redundan- - OVERSEAS AROME : improvements for tropical
cies and automation of about 40% of the synoptic phenomena including tropical cyclones
stations. This year, however, a slightly opposite trend - AROME-NWC : test of new algorithm to merge
is visible. Human observations were restored in sever- forecast precipitation and extrapolated precipitation
al synoptic stations, only during the day as yet. by radar.

• Progress due to new contracts and appropriations
• Vigilance • 3 new Doppler radars - new Harmonie and
- Surveys of public and authorities to identify the Ensemble model
needs and to build future evolutions
- Working group on specifications for a smaller Thomas Vanhamel, RMI/Belgium
scale vigilance and more impact based warnings • New products and developments:
- New model ALARO13 operational with resolu-
• Cloud Sourcing Applications tion of 1.3 km
- Public observation applications for smartphone - Bellavista (BELLS lightning detection)
- Test of connected cars in partnership with - Open data: growing catalogue (climate data,
Continental ALARO-4 output, lidar) with GIS-application
Marjan Sandev, Czech • General news concerning the weather office

Hydrometeorological Institute - (Ongoing) organisational changes: automatiza-
• The basic changes in the Integrated Warning Service tion of some products & from 3 -> 2 forecasters on
System at Czech Hydrometeorological Institute (in duty during the day: test phase
near future) are as follows: - Road condition forecasting (preliminary devel-
• Alert Editor as a module of the Visual Weather opment phase)
system
• Future performance of warning web page (from the Newsletter, Website and Social Media
beginning of 2018?) The WG passed their great thanks and respect to the
• Project on prevention of security risks caused by reviewers of the last newsletter, and also for the print-
extreme meteorological phenomena - their specifica- ing of the beautiful magazine, kindly provided by
tion and innovation of forecasting and warning Meteo France. We are very grateful that Meteo France
is willing to print the letter also in the near future!
Andre Charles gave an update about the website

(www.euroforecaster.org) . Presentations will be put
on this site in the secured area.
Jos offers all participants to connect via Facebook to

the (closed) group of WGCEF, you can also encourage
collegues to join: send a message to Jos Diepeveen,
via FB/Email!
EUMETNET Update
Dick updated the WG on current and planned devel-
opments in EUMETNET
systems with respect to climate change. The project is
financed by the Ministry of Foreign Affairs, and its Future
main goals are: The next phase, 2019-2023, of EUMETNET
- Analysis of impacts caused by extreme meteo- programmes is being prepared. Drafting teams in the
rological phenomena area of Observations and Forecasting are drafting the
- Definitions and Database of Extremes requirements for the various programmes in these
- Impact analysis and draft criteria for issuing areas. The main theme of the Forecasting Programme
impact-based alerts in the next phase will be “impact based forecasting
- Evaluation of the success of predictions of and warning”, which is currently a main subject on
dangerous hydrometeorological phenomena the agenda of many NMHSs. The subject is related
- Estimation of the evolution of frequency of with several activities within the programme like prob-
dangerous phenomena with regards to climate ability forecasting, nowcasting, impact assessment,
change verification of IbW and education&training.
Activities and projects will be integrated in a smaller
Michael Skelbaek, DMI/Denmark number of programmes which will focus on:
• ‘Black November’ – 20% got fired or left voluntarily - EMMA/Meteoalarm

- Forecaster support Members’ Presentations
- NWP cooperation The majority of the second day of the meeting was
- Education&Training devoted to presentations by members on a variety of
EUMETNET is striving for more flexibility. This means subjects. The contents of these presentations can be
projects with flexible start and duration, flexible found on the WGCEF website, www.euroforecaster.org .
number of participating Members. New activities may Our IMGW hosts also arranged a visit to the Forecast
arise during the next programme phase and fit in the Office and the Remote Sensing Department.
existing programmes.
Today Before the close of the meeting, there was discussion
All projects in the forecasting domain, EMMA, EMMA- around the location and theme of the next meeting.
H, Eumetcal, C-SRNWP, SRNWP EPS II, ASIST are well
underway, details can be found in the presentation on Alyssa Razy kindly offered to host the autumn 2018
the website. meeting on behalf of the Meteorological Service of
The yearly Heads of Forecasting meeting in Warsaw in Israel. WG subgroups were also formed to propose
May 2017 was focused on planning the new phase of possible topics for discussion. It was also proposed
projects, especially “Impact based warnings and fore- that the next meeting should be 2 days in length,
casting”. There were special presentations on the ESFS possibly involving a social event and/or sight seeing.
projects, where interested NMHSs may cooperate in a
storm forecasting activity in cooperation with ESSL. Christian and Jos then again complimented our IMGW
Also a very interesting presentation was given on the hosts and organisers Piotr and Julianna before thank-
HIWeather project from WMO/WWRP by Brian Golding. ing the members for their participation and closing
The importance of including other disciplines like econ- the 2018 meeting.
omy, sociology, communication in the production
chain of impact forecasting was discussed. Jos Diepeveen 20th October 2017
Related non-EUMETNET projects:
- Aristotle, started February 2016 ending January
2018: http://aristotle.ingv.it/ . Seven individual
NMHSs and EUMETNET are joining, a project financed
by and in support of the ERCC in Brussels. Currently in
a semi-operational pilot phase. Much related with
current impact based forecasting activities;
- ANYWHERE (EnhANcing emergencY manage-
ment and response to extreme WeatHER and climate
Events, H2020-DRS-1-2015-700099) ) is funded with-
in EU’s Horizon 2020 research and innovation
programme. The principal objective of ANYWHERE is
to enable society as a whole and the main civil protec-
tion agencies to respond more rapidly than today to
extreme climate and weather events, and to better
cope with the high social, environmental and
economic impacts related to these extremes.
Friday 13th October
EUMETNET Task Team on Storm-Naming

Update and Discussion
Evelyn gave an update on the progress and plans of
the storm-naming TT, including the aspirations for
further harmonisation of the European schemes and
possible integration of some of the existing regional
schemes, notably in between the southwestern and
western schemes. A lively debate ensued, with the
intention of passing recommendations to the TT meet-
ing in Spring 2018.

Sting Jets and other processes leading to high
wind gusts:
wind-storms “Zeus” and “Joachim” compared
Bruno Gillet-Chaulet bruno.gillet-chaulet@meteo.fr
“It is very difficult to predict,
especially the future”
Introduction Niels Bohr
On Monday, 6th March 2017, a major wind-storm, Both cyclones were the result of a “ B a r o c l i n i c
named ‘ Z eus’ by the media, hit France. This low D e v e l o p m e n t ” . They both had similar moving
system brought absolute record wind gusts. It was speeds (about 80 km/h) and pressure drop rates
one of the most violent that the country had known (above 1 hPa per hour), from 1005 to 990 hPa in 12
for years. In Brittany (“Finistère”), the level of the hours in the case of “Zeus”, and from 995 to 965
wind speed was unexpected and surprised lots of hPa in 24 hours in the case of “ J o a c h i m ” .
forecasters on shift at that time. As a consequence, “Explosive cyclone development” is the phrase
this event gave us the opportunity to better under- sometimes used to describe the development of
stand the mechanisms that are at the source of these systems. Nevertheless, “Zeus” was smaller in
highest gusts and in particular, to discuss the size than “Joachim” (which suggests a speeding-up
concept of the “Sting Jet”, raised in association with of overall processes involved in this case). An
this type of hazard more and more often in the last analytical approach ( Santurette and Joly, 2002 )
few years. This article describes the wind-storms highlights the differences between both (FFigure 1)
“Zeus” (6/03/2017) and “Joachim” (16/12/2011) when the systems had reached maturity.
and e x p lo r e s t he d if f er e nt p r oc e sse s l e a d in g t o “Joachim” is typical of the first stage (I) of develop-
high wind gusts. ment of the conceptual cyclone model of S hapiro
and Keyser (1990) (FFigure 2, a). The cold and warm
fronts are well structured and are of “Ana” type.
Characterisation of the wind-storms They join on the inside of the low. This is a
“Zeus” and “Joachim”: “Norwegian” configuration. Note the two branches
of the upper jet stream (brown arrows). The cyclone
Indicators, given by the Climatology Division of is located on the left exit of the upstream jet streak,
Météo France, allow us to characterize wind-storms and on the right entrance of the downstream jet
and compare them with each other (Tempêtes en streak. A dynamic potential vorticity (PV) anomaly
France métropolitaine, Météo France, 2017). On an (black line) precedes the surface low. The isobars
historical level, the previous wind-storm most are quite regularly spaced. Note that the model
comparable to “Zeus” was “Joachim”. The compari- “analysis” (brown lines) is close to the observa-
son of these systems is very instructive. The table tions.
below (Table 1) provides a summary.
“Zeus” “Joachim”
Maximum gusts at maturity 190 km/h, Ouessant-Stiff 212 km/h, Puy de Dôme (1415m)
of the systems phase I 193 km/h, Camaret (absolute record) 139 km/h, Clermont-Ferrand
06-07 H, 6/03/2017, Western France 06-08 H, 16/12/2011, Centre France
Other significant gusts phase II 144 km/h, Chambéry-Aix

14-15 H, 6/03/2017, Eastern France
Other significant gusts phase III 184 km/h, Dramont

174 km/h, St Cézaire sur Siagne (absolute record)
00 H, 7/03/2017, South-eastern France
Affected Surface (% of the Country) 32% 31%
“Lamb (1991)” Index 595 618
“Storm Severity Index” (SSI) 54 50

Table 1: Characteristics of the wind-storms.

“Zeus” is typical of the last sta ge (IV) of develop- K i n e t i c E n e r g y ” ( T K E ) . The formula (with
m e n t ( S h a p i r o - K e y s e r m o d e l ) . The warm front ARPEGE/AROME -Météo France-) is close to :
partially encircles the low-pressure core (back-bent W gu s t s = W me a n + α * √ T K E , α depending on the
warm front). The cold front is of “Kata” type. The model. These relevant fields allow us to view the
cyclone is located on the left exit of a single structures of the gusts (FFigure 3, top).
upstream jet streak. A “PV” anomaly overhangs the In the case of “J oa chim”, the highest gusts were
surface low (black dashed line). The isobars (black essentially connected to the cold front and were
lines) are tightened on the “equatorial side” of the linked to the presence of a strong low level jet
l o w (manual “analysis”). Note that the model 8 0 kts) at an altitude of 1500m (vertical cross
(8
“analysis” (brown lines) is not low enough section C-D, bottom). This is the warm conveyor belt
compared to the observations. Nevertheless the (WCB) (FFigure 2, b). In the case of “Zeus”, the high-
main characteristics of the system seem to be est gusts were located on the equatorial side of the
properly modelled. The distinction between the surface low with a peak very close to the centre of
stages of extra-tropical cyclone development the cyclone, next to the back-bent warm front where
(Shapiro-Keyser model) is essential because of the the isobars were tightest. These gusts were linked
consequences on “C C o nv e y o r B e l t s ” (explained to the presence of a strong low level jet (9 90 kts) at
below). Moreover, it is important to notice that, an altitude of 1000m (vertical cross section A-B,
when mature, “J o a c hi m” never goes beyond the bottom). This speed may have been underestimated
first stage, but “Zeus” comes rapidly, if not directly, by the model in accordance with remark 1. Note
to the last stage (but without undergoing the interim how the strongest “isotachs” extend toward the
stages). surface with a peak of TKE. These features refer to
the concepts of cold conveyor belt (CCB) and sting
jet (SJ) (FFigure 2, b and c) discussed below.
Characterisation of the gusts:
Already note that the geometric configuration of the
For a few years, the wind gusts predicted by numeri- gusts has consequences on their predictability. For
cal models, have used the parameter “Turbulence a system like “Joa chi m”, a small forecast error in
Figure 1: Graphical view (ANASYG) of the wind-storms.
Figure 2: Conceptual view of an extra-tropical cyclone. Based on Shapiro and Keyser (1990) and Schultz and Browning (2017).

Figure 3: Gusts (threshold 100km/h, shading) and vertical cross-sections associated with the wind-storms.
position and intensity of the low would lead to a The water vapour image (top left) shows an axis of
small inaccuracy of wind estimation (almost the strong subsidence overhanging the equatorial side
same regions would be affected with more or less of the surface low (marker +) in front of the upper
the same level of wind speed). For a system like cloud head (marker C) of the system. This dry and
“Zeus”, a similar small forecast error could lead to subsiding air mass moves along the north side of
missing the level of the highest gusts and the areas the upper jet streak and is accompanied by a
affected. There is no doubt that issuing accurate dynamic PV anomaly (marker A) and the upper cold
warnings is easier in the first case than in the front (marker B).
second... The E-View RGB composite image (top right) clearly
shows the trace of the thick warm front, on the
“polar side” of the surface low (marker +). The wind-
The Sting Jet in question: ing of low cloud, trace of the back-bent warm front
(marker D) in front of the upper cloud head (marker
It has been known for a long time that the most C), on the equatorial side of the surface low
damaging winds of an extra-tropical cyclone are distinctly appears. Subsidence, previously
often located on the equatorial side of the low mentioned, prevents mid and high cloud develop-
(Schultz and Browning, 2017). Some metaphorical ment in this region. Finally, note the signature of the
expressions used to describe this phenomenon are kata-cold front near the surface (marker E) and the
“poisonous tail” or “the sting at the end of the tail”. upper levels (marker B).
The concept of the sting jet was introduced by The radar images (bottom) show continuous moder-
Browning in 2004 (Browning, 2004). The author ate rain associated with the warm front. The ring-
associates the notion with a slantwise jet, coming shaped echoes mark the surface low. At 06 H, note
from the mid and upper levels and descending the occurring of a narrow hook-shaped band of
toward the surface, leading to damaging gusts precipitation (marker F) eponymous of a sting jet.
(Figure 2, c). Some characteristics of a sting jet can Indeed, this observation is concomitant with the
be found with “Zeus”. Satellite and radar pictures highest wind gusts. Afterwards, the cyclone fills as
display special features (Figure 4). it drifts eastwards over land. As a consequence, and

Figure 4: Satellite and radar images of wind-storm “Zeus” at maturity of the system.
because of increasing friction, gusts tend to identified (FFigure 3). It is probably the signature of
decrease and precipitation echoes become unstruc- the sting jet, but we should point out that verifica-
tured. tion would require the computation of backward
X-Y from figure 3) along
The vertical cross-section (X Lagrangian trajectories, in order to know the origin
the axis of the highest gusts (FFigure 5) shows the of air parcels in the region of maximum wind
vertical velocity (blue curves). A slantwise subsi- (Coronel, 2016).
dence tends to overhang the low level jet previously
Figure 5: Vertical cross-section, wind-storm “Zeus”.

Other processes leading to high and, on the other hand, the ones with strengthen-
ing winds near the ana-cold front (““ Joa chim”), in
wind gusts: conjunction with the “warm conveyor belt” (WCB).
Significant wind gusts can also occur outside the Moreover, convective gusts and down-slope wind-
maturity period of a cyclone. storms (mountain waves) can also occur. Figure 7
During phase II (TTable 1), “Zeus” was weakening. gives a summary of these occurrences.
The outstanding value measured in Chambéry-Aix
was caused by a stormy episode (FFigure 6, a). The It is important to mention that the use of the term sting
convective aspect explains the level reached. Gusts jet has become quite ambiguous because it refers both
associated with thunderstorms can take several to damaging wind gusts (generally) over a small area in
forms: down-bursts (micro-bursts, macro-bursts) or a short period of time and a slantwise circulation due to
even tornadoes. These are generally isolated, so as a subsidence from mid and upper levels towards low
result, are particularly hard to forecast. To issue levels (this last definition being selected in this paper,
accurate warnings is troublesome. Fi g ur e 5). Yet, as a recent study ( Coronel , 2016)
During phase III (table1), “Zeus” was moving away suggests: “The strongest wind gusts do not seem to be
towards Mediterranean regions. On the back of the connected directly to the sting-jet air streams (in most
system, a dry subsiding air stream fell into place cases)”. Indeed, it is the combination of both a cold
over the Southern Alps. Mountain waves produced conveyor belt and a sting jet (increasing momentum
the “down-slope wind-storm” phenomenon with within the boundary layer) that leads to the highest
record gusts (FFigure 6, b). Windward higher-pressure gusts. Moreover, concerning the location of the sting
and leeward lower-pressure are well represented by jet, it seems that the sting jet takes place above the
tightened isobars along the relief. Consequently cold conveyor belt, rather than beyond it. This study
Potential Temperatures (green curves) dive on the says: “This differs from the 1987 storm studied by
lee side. It has to be said that (global and limited Browning (2004) [...], for which the sting jet extends
area) models have proved successful in predicting beyond the cold-conveyor-belt jet”. In that respect,
this type of event, even if the exact intensity remains diagram (FFigure 2, c) could be updated and the term
difficult to assess. Accurate warnings are generally “Hooked Jet” could be used to name this occurrence. In
issued. any event, the expression “Jet d’Occlusion” (“Occlusion
Figure 6: Gusts during phases II and III of wind-storm “Zeus”.
Conclusion: Jet”) which is often used in French to describe that type

of hazard should not be used here (as there is no ques-
The examination of the wind-storms “Z eus ” and tion of an occluded front in this case). Let us remember
“ J o a ch im ” has provided an analysis of the main that the tightening of the isobars on the equatorial side
processes leading to high wind gusts in European of a cyclone at the last stage (Shapiro-Keyser Model)
countries. Indeed, these systems constitute the next to the back-bent warm front is a potential marker
base form of two families of extra-tropical cyclones: of this type of event.
on the one hand, the ones with strengthening
winds on the equatorial side of the low (““Zeus”), In an operational context, forecasters have to be
because of the simultaneous presence of the “cold attentive to all processes leading to high wind
conveyor belt” (CCB) and probably a sting jet (SJ), gusts. To be able to identify these processes as

Figure 7: Configuration of high wind gusts connected to extra-tropical cyclones. Surface isobars and fronts. Brown line symbolises
the relief.
soon as possible, allows forecasters to better Shapiro, M. A., and D. Keyser, 1990: Fronts, jet
manage their predictability and to better qualify streams and the tropopause. Extra-tropical
uncertainty. In that sense, Ensemble Prediction Cyclones, The Erik Palmén Memorial Volume, C. W.
Systems (EPS), at large and small scales, are really Newton and E. O. Holopainen, Eds., Amer. Meteor.
helpful. Finally, forecasters of the 21st Century expe- Soc., 167-191.
rience a new challenge when facing the risk of an Schultz, D.M., K.A. Browning, 2017: What is a sting
extreme weather event: among a set of numerous jet ?, Weather, March 2017, Vol. 72, N°3.
forecasts, there could be one forecast member, at
least, which simulates the “right solution”. The
purpose is to obtain the greatest possible benefit
from numerical models whilst avoiding false
alarms.
Note:
All times are UTC.
Reference:
Browning, K. A., 2004: The sting at the end of the
tail: Damaging winds associated with extra-tropical
cyclones, Q. J. R. Meteorol. Soc., 130, pp. 375–399.
Coronel B., Ricard D., Rivière G., Arbogast P. 2016:

Cold-conveyor-belt jet, sting jet and slantwise circu-
lations in idealized simulations of extra-tropical
cyclones. Q. J. Royal. Meteorol. Soc., 142, 1781-
1796.
Lamb H.H., Frydendahl K., 1991: Historic storms of

the North Sea, British Isles and Northwest Europe.
Cambridge, Cambridge University Press, 203 p.
Météo France, 2017, Tempêtes en France métropoli-

taine, “tempetes.meteo.fr”.
Santurette, P., A. Joly, 2002: ANASYG/PRESYG,
Météo-France’s new graphical summary of the
synoptic situation, Meteorol. Appl. 9, p. 129-154.

Forecasting Freezing Rain in the UK –
March 1st and 2nd 2018
Stephanie Jameson, Met Office, Exeter, United Kingdom.
stephanie.jameson@metoffice.gov.uk
Introduction brought a bitterly cold polar continental air mass

(with origins in Finland and north-east Russia) to our
Late February and early March 2018 saw some shores. The air mass was unstable both to sea
remarkable weather conditions across the United surface temperatures and daytime temperatures,
Kingdom, as an omega block over Europe led to an with the strong easterly flow helping the westward
easterly flow and a polar continental air mass becom-
propagation of snow showers inland and low level
ing established across the UK. The ‘Beast from the
convergence lines developing over north-eastern
East’ (as the UK press dubbed the spell of bitterly cold
Britain acting as a focus for more continuous snow
wintry weather) brought significant snowfall, remark-
showers through the day.
able low temperatures and severe travel disruption. In
fact, it was the most significant spell of snow and low
Meanwhile, the deep area of low pressure located to
temperatures for the UK overall since 2010 - a year
the north of Spain on the 12 UTC analysis chart
which stands out in the record books for having the
(named Storm Emma by the Portuguese Institute for
coldest December in the UK Series (one of the stan-
dard datasets used by the Met Office to contextualise Sea and Atmosphere, IPMA) continued to track
weather events, and which dates back to 1910). In northwards, still in a deepening phase as it crossed
particular, the night of March 1 st and morning of the Bay of Biscay, before gradually starting to slow
March 2nd 2018 saw a significant freezing rain event and fill as it neared the south coast of Devon. Warm
that affected south-western parts of the UK and which advection ahead of Storm Emma led to quite a
will be the focus of this article. complex frontal analysis in the vicinity of the UK, with
an upper warm front indicating the leading edge of
the milder air overriding the extremely cold polar
Synoptic Situation: continental air at the surface. This led to an area of
precipitation blossoming ahead of the main surface
Despite being the first day of Meteorological Spring, front, which fell initially as snow, but turned increas-
March 1st was in fact bitterly cold. The surface analy- ingly to freezing rain, as warm advection continued to
sis chart in Figure 1 shows how high pressure posi- occur at mid-levels. By 00 UTC on Friday 2nd March,
tioned over Scandinavia and low pressure in the Bay the leading warm front associated with Storm Emma
of Biscay led to a strong, deeply unstable easterly had crossed northern France and the English
flow becoming established across the UK. This Channel, pushing north across southern coastal
Figure 1. Surface analysis charts, for Thursday 1st March and Friday 2nd March 2018. Surface analysis chart valid at 12 UTC on Thursday 1st
March 2018 (left). Surface analysis chart, valid at 00 UTC on Friday 2nd March 2018 (centre). Surface analysis chart, valid at 12 UTC on Friday 2nd
March 2018 (right).

counties of England. As Storm Emma moved slowly these are then grouped into one of 8 general patterns
north-westward towards the coast of Cornwall, veer- (by looking at the correlation between the 30 regimes
ing flow on the north-east flank of the system contin- in terms of their MSLP anomalies). This allows the
ued to import milder (or less cold) air from the south, user to take a probabilistic view of which weather
allowing a gradual thaw to take place through late regime is most likely through the forecast period.
Friday and into Saturday 3rd March.
Figure 2 shows the output from Decider, based on
the GloSea5 00 UTC run on Wednesday 7th February.
This indicated an unusually high forecast probability
Mid to Long-Range Forecast for Regime 1 to occur later through the month. As can
be seen in figure 2, Regime 1 refers to a generally
Prior to the event in question, a prolonged cold spell blocked pattern with little Atlantic mobility. The
had been well forecast several weeks in advance. For image depicts an idealised version of what Regime 1
example, the Met Office product Decider had high- looks like in terms of PMSL and temperature biases
lighted a significant cold spell as early as 7 February relative to average. This shows a positive PMSL
2018, some three weeks beforehand. Decider is a anomaly to the north of the UK and a negative PMSL
medium-to-long range forecasting tool which aids anomaly to the south of the UK, resulting in a blocked
meteorologists by categorising ensemble members pattern and leading to an easterly flow being set up,
by matching their synoptic pattern to a range of signalling the potential for a significant cold spell for
idealised synoptic patterns. It is updated several the country.
times a day using output from a variety of ensemble
model forecasting systems, including MOGREPS-G By the following week, it was possible to refine the
(Met Office), Glosea5 (Met Office), ECMWF medium medium-range forecast further, with a strong signal
range (ECMWF) and GEFS (NCEP) output. Ensemble for Regime 27 to take hold during the last week of
members are assigned to the closest matching February. Figure 3 shows the Decider output based
weather regime out of a set of 30 possible predefined on the ECMWF 00 UTC run on Tuesday 13th February,
weather regimes across Europe. To help reduce the highlighting high forecast confidence in Regime 27
typical spread in output at longer range timescales, taking hold. An idealised version of how Regime 27
Figure 2. Long range output from Decider, the Met Office 5 week weather regime forecast (8 regimes), based on the GloSea5 00 UTC run on
Wednesday 7th February 2018.

Figure 3. Medium-range output from Decider, based on the ECMWF 00 UTC run on Tuesday 13th February 2018.
looks in terms of PMSL and temperature biases rela- The general weather pattern based on these regime
tive to average is shown in Figure 3. This once again forecasts were highly confident for both the long
depicts an area of high pressure to the north of the range and medium range forecasts, though there was
UK, with an area of low pressure to the south, indicat- still uncertainty in the detail, with small changes in
ing a strong easterly flow with a negative temperature the exact position of the anticyclone leading to varia-
anomaly becoming dominant during the end of the tion in the surface flow and the sensible weather at
month. any given location, as a result.
Figure 4. MOGREPS-G, Met Office’s global ensemble forecasting system, postage stamps of MSLP valid at 12 UTC Friday 2nd March, using
the 06 UTC run on Sunday 25th February.

By Sunday 25th February, there was very high confi-
dence in the synoptic pattern, with deeply unstable,
cold air and unusually low 1000-500hPa thickness
spreading across the country. This meant most areas
were expected to see snow at some point, and signif-
icant accumulations were expected in places. There
was also some signal for potential blizzard condi-
tions to develop through Thursday 1st and Friday 2nd
March as Storm Emma was forecast to track north-
wards towards the UK, though there was uncertainty
in the timing of this feature, as different model runs
gave differing solutions.
Figure 4 shows the ensemble output from

MOGREPS-G, the Met Office’s global ensemble fore-
casting system, using the 06 UTC run on Sunday 25th
February, valid at 12 UTC Friday 2nd March. At the time
of writing, MOGREPS-G is comprised of 18 ensemble
members (1 control and 17 perturbed members) with
a vertical resolution of 70 levels (with a model top of Figure 5. UKV snow depths for 24 hours up to 09 UTC on Friday 2nd
80 km) and horizontal resolution of approximately March 2018, using the 09 UTC run on Thursday 1st March 2018.
20km at mid-latitudes. At this stage, there was a
strong signal for the low pressure system and associ-
ated fronts to affect the southern half of the UK, with
many members suggesting the possibility of a signifi- However, the approach of the higher wet bulb poten-
cant blizzard affecting southern and central parts of tial temperature air mass associated with Storm
the UK through Thursday and into Friday. Emma, was expected to have an impact on the type of
precipitation falling. Not only was snow across this
region expected to become ‘wetter’, but the risk of
Short-Range Forecast freezing rain became evident. Figure 6 shows high
resolution model output from the UKV model showing
Other model output nearer the time in the lead up to ‘precipitation type’ forecast for 00 UTC and 06 UTC
Friday 2nd March indicated some substantial snowfall respectively, on Friday 2nd March. Here, swathes of
accumulations, as well as a significant freezing rain red indicate areas where precipitation is expected to
signal across southern parts of the UK. Figure 5 shows fall as freezing rain or sleet. This was a particularly
the UKV models snow depths in the 24 hours up to strong signal in the NWP forecast, most notable for
09 UTC on Friday 2nd March, using the 09 UTC run on the size of the area covered, which by UK standards
Thursday 1st March 2018. This model output was was large.
considered a good broad guide by the duty Chief
Meteorologist on the day, but some spatial errors In general, freezing rain events are infrequent across
were anticipated across northern areas due to the the UK and when they do occur are usually fairly
exact location of snow showers and the convergence localised and transient in nature, so an event on this
bands that were forecast to develop. What is more, scale had the potential to be very disruptive, as well
the model diagnostics assume a snow to liquid ratio as climatologically quite rare.
of 10:1, but due to the extremely cold, low dewpoint
characteristics of the air mass, a ratio of 12:1 or 15:1 Figure 7 shows output from the 03 UTC run on 1st
was considered more appropriate. However, this March of MOGREPS-UK showing the forecast risk of
adjustment was expected to apply less and less freezing rain at 00 UTC and 06 UTC Friday 2nd March
across southern and south-western areas through the 2018, on the left and right respectively, with the area
course of Thursday evening, as warm advection and shaded red indicating where more than 80% of
increasing dewpoints would lead to a wetter type of ensemble members are forecasting freezing rain to
snow falling here. Some 5-10 cm were expected wide- occur. This area highlighted the far north-east of
ly across parts of Devon, Somerset and east Wales, Cornwall and parts of Devon, including the higher
with areas of 10-20 cm and possibly 30-50cm over ground of Dartmoor and Exmoor, as being most at
higher ground, through the course of the evening. risk.

Figure 6. Deterministic, post-processed model output from the UKV model, showing precipitation type at 00 UTC Friday 2nd March 2018 (left)
and at 06 UTC Friday 2nd March 2018 (right). Output from the 15 UTC run of the UKV from Thursday 1st March 2018.
Figure 7. Post-processed output from MOGREPS-UK, the Met Office’s regional ensemble forecasting system, showing risk of freezing rain,
valid at 00 UTC and 06 UTC Friday 2nd March 2018, using the 03 UTC run on Thursday 1st March 2018.
Upon further investigation, and by looking at the began to occlude out, with the effect of potentially
forecast tephigram profiles, as shown in Figure 8, it allowing freezing rain to turn back to sleet or wet
was possible to see a clearly defined, classic ‘warm snow. At the same time, the slight deepening of the
nose’ developing as warm advection began to take boundary layer shown by the 09 UTC profile, may
place aloft. The elevated melting layer evident on also mean that freezing rain or sleet falling from the
the profiles would allow any frozen precipitation to melting layer may have time to re-freeze as it fell,
change phase back to liquid before falling through reaching the ground as ice pellets instead. Although
the sub-zero surface layer and hence become subtle, these nuances in the forecast were signifi-
supercooled - posing a severe icing threat on cant in terms of impacts, as supercooled rain freez-
contact with any sub-zero surface. Interestingly, by ing on contact on the ground would give a greater
09 UTC, the elevated melting layer was becoming ice risk and be considered more hazardous than
less pronounced, perhaps as the milder air aloft that of frozen precipitation falling.

Figure 8. Tephigram profiles from the 00 UTC run on Thursday 1st March of the deterministic Global Model, valid for 15 UTC Thursday
1st March (left), 00 UTC Friday 2nd March (centre) and 09 UTC Friday 2nd March 2018 (right).
Warnings Issued As forecast confidence increased, further evidence was

gathered to indicate that there was a greater likelihood
The Met Office operate an impacts-focussed National of some significant impacts occurring. There was
Severe Weather Warning System (NSWWS). In light of concern that roads would become blocked due to drift-
the high confidence in the general forecast, the ing snow, trees and infrastructure would be damaged
Guidance Team Meteorologists who are responsible and interruptions to power supplies and other utilities
for the public weather service warnings were able to may occur. As a result, an amber warning (see figure 9)
issue a yellow warning for the evening of Thursday 1st was issued at 12:16 local on Tuesday 27th February, this
March some 4 days in advance. The warning level is time falling into the medium likelihood of high impacts.
assessed using a ‘likelihood’ versus ‘impact’ matrix, Finally, by the morning of Thursday 1st March, all the
and initially the yellow warning was issued with the evidence available pointed to the potential for a severe
low likelihood of having a high impact. Figure 9 shows weather event over parts of Devon, Somerset and south
the area covered by the yellow warning that was Wales, posing significant threat to life and limb and
issued at 12:54 local on Sunday 25th February. This culminating in the issue of a red warning, which fell in
warning was updated at 04:59 local on Thursday 1st the category of high likelihood of high impact. This
March, to increase the areal coverage into north Wales warning (see figure 9) was issued at 08:13 local on the
and across southern coastal counties of England. morning of Thursday 1st March 2018.
Figure 9. National Severe Weather Warning Service warnings issued for south-western parts of the UK for Thursday 1st March. The yellow
warning (left) was issued at 12:54 on Sunday 25th February for the south-west UK on Thursday 1st March, the amber warning (centre) was
issued at 12:16 on Tuesday 27th February for the south-west UK on Thursday 1st March and the red warning (right) was issued on Thursday 1st
March for the south-west UK later that same day.

Impacts and Historical Context There were many significant impacts across south-
western parts of the UK due to the weather through
Figure 10 shows a cross-section of the solid precipita- the 48 hour period from 1st March to 2nd March 2018,
tion that fell through Thursday 1st March into Friday with numerous requests for military support across
2 nd March, captured at Uffculme in East Devon, Devon, Cornwall, Avon and Somerset.
30 miles north-east of Exeter, whilst Figure 11 shows There was disruption on the road network across the
photographs taken of ice glazed vegetation and vehi- region. For example, one of the main routes in east
cles following the freezing rain event on Friday Devon - the A380 - was closed, so that the authorities
2nd March. could focus resources on keeping the main coastal
Figure 10. Cross-section of solid precipitation that fell at Uffculme, Devon, during Thursday 1st March and Friday 2nd March 2018.
routes open, whilst another main trunk road, the A38

– which runs over high ground on the eastern side of
Dartmoor - was closed following a number of road
traffic accidents. Similarly, the A303, a main road
through parts of East Devon and Somerset, came to a
standstill later in the day, with drivers stranded there
for 12 hours. Overnight into Friday 2nd March, many
roads across southern parts became treacherous after
freezing rain fell across the region, with reports of
glazed ice in Hampshire, Dorset and Devon.
There was also an impact on the train lines, with

National Rail reporting mass disruption across the
south-west, whilst a South Western Railway train to
Weymouth in Dorset, broke down during the evening,
leaving passengers trapped onboard for 11 hours.
Meanwhile, over 5000 schools were closed across
England, as frozen pipes caused issues with water
Figure 11. Images of ice glazed vegetation and vehicles taken supplies in places. It was reported that there were
following the freezing rain event on March 2nd 2018 in Exeter, Devon. 10 weather-related deaths from across the country.

Figure 12. Monthly averages (days) of snow
falling and lying for 1971-2000
at St. Mawgan, near the coast of Cornwall (altitude
103 metres), Long Ashton, near Bristol (altitude 51
metres) and Yeovilton, Somerset (altitude 20
metres).
To give this some historical context, it is interesting to Acknowledgements and References

note that settling snow does not occur frequently in
the south-west of the UK. In fact, lowland areas do not Prior, J. & Kendon, M. (2010): The UK winter of
record any lying snow in roughly one out of every 3 2009/2010 compared with severe winters of the last
years. In coastal areas of Devon and Cornwall, the 100 years. Weather, January 2011, Vol. 66, No.1.
average number of days of lying snow (defined as a
day where more than 50% of the ground is covered at Neal, R., Fereday, D., Crocker, R. & Comer, R. E. (2016):
0900 at the observation site) is typically less than 3 A flexible approach to defining weather patterns and
per year. This figure increases to around five to ten their application in weather forecasting over Europe.
days per year for inland areas, whilst the highest Meteorol. Appl. 23: 389-400.
ground in the region – that of Dartmoor and Exmoor – https://www.metoffice.gov.uk/research/news/2016/
have more than 20 days per year on average. Figure 12 new-weather-patterns-for-uk-and-europe
shows a table of monthly averages of snow falling and Symons, L. & Perry, A. (1997): Predicting road hazards
lying at a variety of different sites at different altitudes. caused by rain, freezing rain and wet surfaces and the
role of weather radar. Meteorol. Appl. 4, 17-21.
Meanwhile, true freezing rain events – where super-
cooled rain droplets fall, as opposed to ordinary rain Brooks, C. E. P. & Douglas, C. K. M. (1956): Glazed
falling onto frozen surfaces – are climatologically rare frost of January 1940. Meteorological Office
in the UK. In the past century, only a handful of notable Geophysical Memoirs, No. 98, M0584f.
freezing rain events have occurred. In particular, one Snow and Low temperatures February to March 2018,
event occurred across mid-Wales, the Midlands and 13 March 2018. Met Office:
Norfolk in January 1996, when an intense area of high https://www.metoffice.gov.uk/climate/uk/interesting
pressure over western Russia fed polar continental air /february2018-snow
across the UK, and a series of weather fronts began to
Information and data, including climate summaries
push in from the south. A similar synoptic situation led
and statistics, is compiled by the Met Office National
to an event that occurred across southern England and
Climate Information Centre https://www.metoffice.
south Wales during the last two days of December
gov.uk/climate.
1995, where sections of the M4 motorway were closed
for almost two days when a series of road traffic acci- Information, including synoptic charts and forecast
dents occurred as a result of icing. It was thought that data, as well as information on the NSWWS can also
traffic accidents were actually fewer than might have be found online at the Met Office Digital Library and
been due to the fact that the event fell on a day where Archive:
many people were on holiday between Christmas and https://digital.nmla.metoffice.gov.uk/archive/
New Year. Widespread freezing rain also occurred on
the 20th to 21st January 1966, as well as on Christmas With special thanks to Stephen Moseley, Met Office, for
Eve in 1968 when Wales and the west Midlands were providing access to archived post-processed model
badly affected. During the period of the 17th to the 19th data, Robert Neal, Met Office, for his permission to use
March 1969, there was another freezing rain event, graphics showing ouput from Decider, and Keith
this time across the eastern Pennines, whilst a more Fenwick, Met Office, for providing profile analysis and
notable and widespread glazed frost occurred in the cross-section of solid precipitation in Uffculme.
January 1940. Additional thanks to Met Office colleagues Dan Suri,
In comparison, it is clear that the event of this March, was Laura Paterson, Will Lang and Jason Kelly for their
remarkable and relatively speaking, climatologically rare. advice and input.

The Extreme Wildfire,
17-19 July 2017 in Split (Croatia)
Tanja Trošić Lesar, Meteorological and Hydrological service, Zagreb, Croatia
e-mail: trosic@cirus.dhz.hr
Marija Mokorić, Meteorological and hydrological Service, Zagreb, Croatia
e-mail: mokoric@cirus.dhz.hr
Introduction
The summer months of 2017 were characterized by
dry periods and extremely high air temperatures. In
the Adriatic coastal area, such weather conditions
were most pronounced across the southern part of
the Croatian region of Dalmatia, There were 4 expres-
sive heat waves in Dalmatia, and the most prolonged
with the highest air temperatures began at the begin-
ning of the third week of July and ended at the end of
the first week of August. Thus, last summer, in
Croatia, as in other areas of the Mediterranean, there
was an increase in the number of forest fires. Some of
them affected large areas of the Adriatic coast, and
the most prominent was recorded near Split, the
second largest city in Croatia and the centre of
Dalmatia.
The wildfire in the Split area that occurred from 17 to Figure 1. The area affected by the fire in Split (top) and NASA
19 July 2017, according to the National protection and satelite imagery of the Split area on 17 July 2017 (bottom) (source:
HUKM)
rescue directorate fire department, began on July 17,
2017 at 0042h in Central European Time at Tugare,
about 10 km east of Split. On 17 July, the fire quickly synoptic situation on 17 July 2017, when the fire was
spread from the east to north-east over the slopes of the most severe, as well as on 18 and 19 July, when it
Mt. Mosor, as a result of a strong Bura wind and the was placed under control and finally localized.
local orography. Several houses, business premises
and olive groves were burned, and the fire also
reached the astronomical observatory at Mosor.
Weather conditions
Furthermore, the fire approached the Žrnovica missile
launch facility, which was defended with a great effort from 17 to 19 July 2017
by the firefighters and soldiers in attendance. The
As mentioned, the summer months of last year were
town of Split was endangered by the Bura from the
dry and extremely warm. Therefore, the hazard class
northeast. The air force could not participate in the
derived from the meteorological hazard index for
firefight until 1500h on July 17, because of the strong
Bura. Three Canadair aircraft were involved in the fire the ‘fire emergence and spread’ by the Canadian
extinguishing, as winds began weakening. The fire method was mostly ‘large’ or ‘very large’. According
was placed under control on 18 July at about 0700h, to data from the Split-Marjan meteorological
localized on 19 July at about 1100h and was station, the highest danger class was recorded in
completely extinguished on 25 July 2017. Around the Split area on 25 June, and continued until 19
4300 ha of vegetation were burnt down in the fire and July 2017, when the fire was localized. A high
the fire front was occasionally 40-km long, making it danger class indicates the dryness of the canopy or
one of the biggest fires in Croatia. Fortunately there the presence of dried vegetation, which is affected
were no human casualties. by high air temperature and lack of rain. Wind and
This paper describes the course of the fire and the low relative humidity are also contributing factors.

The meteorological index component, according to have also been used, for example, by Carvalho et
the Canadian method, which shows how the fire al. (2008) in Portugal, Tian et al. (2011) in China
can spread is the Initial Spread Index (ISI). The and also by Vu eti and Vu eti (2011) and Kozari
Canadian forest fire weather system components and Mokori (2012) in Croatia. The details of the
Figure 2. The mean and maximum wind velocities (a), mean wind direction (b), as well as the air temperatures at 2 m height (° C) and rela-
tive humidity (%) at the Split-Marjan station from 16 to 19 July 2017.

Canadian forest fire weather system can be found in caused a deep Bura which blew on 17 July 2017. In the
Van Wagner and Pickett (1985). The ISI recorded at middle of 17 July, the anticyclonic ridge over north-
the Split-Marjan station at 1200 UTC on 16 July west Europe spread over the western Balkans, and
2017, before the outbreak of fire, was very high cyclonic conditions moved over south-east Europe.
(24). On 17 July at 1200 UTC, when the fire was The jet flow was still visible at 300 hPa, from Sicily to
uncontrollably spreading, the ISI value was 20. the Ionian Sea. The Bura then weakened, and along
Then, on 18 July, the ISI decreased to 14. Rapid fire the coast and the open sea, it turned to the north-
spread is associated with ISI values of more than west. Then, during the night and in the morning of 18
13, whilst values of more than 18 are associated July most of Croatia, including the Mid-Adriatic, was
with fire which is rapidly expanding. The estimated influenced by an anticyclonic ridge. With the shifting of
the upper level trough axis to the east of Croatian terri-
fire spread speed is 18.3m per minute when the fire
tory and the strengthening of the anticyclonic branch
occurs in the treetops and when multiple fire fronts
from the north-west of Europe on 18 July, the air pres-
appear. The high values of the initial spread index
sure gradients decreased and the Bura weakened. The
are due to the high wind speeds which were in the
South Adriatic was affected by the forward side of the
Adriatic region on 16 July 2017.
upper ridge, and the rearward side of an upper trough,
which was then centred over the Ionian Sea, having
During the Bura on 17 July 2017 the relative humidity
tracked slowly east. The jet stream had moved away
was between 20 and 30%, and when the Bura weak-
from the Mediterranean area. Then, on 19 July, the
ened on 18 July, the relative humidity was higher than
anticyclonic ridge weakened leading to an elevated
35%. Furthermore, according to the measurements pressure field over Croatia, and from the middle of the
from the Split-Marjan automatic station (see Figure 2) day equalized pressure field reduced to mean sea
on 17 July, the Bura had weakened and was moderate level. At upper levels, the ridge in geopotential height
to strong, whilst in the middle of the day the wind was spread across the Balkans, whilst the upper low
briefly turned to the east. On 18 July, the Bura weak- was still over the area of Asia Minor.
ened, such that in the morning the wind was mostly
weak with an east to south-east direction. At midday
and early in the afternoon, a moderate south-westerly
wind was blowing. Then the wind was mostly weak,
with the characteristics of a coastal circulation. Sharp
changes in wind direction have been observed in the
fire and made the behavior of the fire unpredictable,
and the situation extremely dangerous.
Synoptic analysis
from 17 to 19 July 2017
The surface analysis based on the synoptic maps
(from DWD and the reanalysis from ECMWF) show that
on 17 July anticyclonic conditions extended from
north-western Europe towards Croatia (see Figures 3
and 4). At the same time, over south-east Europe,
there was an area of low pressure, whose centre was
over Asia Minor. Therefore, there was a large surface
pressure gradient over the Adriatic. In the upper
atmospheric layers (isobaric surfaces 850, 500 and
300 hPa) on 17 July at 0000 UTC, central and south-
western parts of Europe were under the influence of an
upper level low. Along with this, the separation of the
upper level cyclone with a cold air core over the south-
eastern part of Europe is visible, so that the upper
level cyclone is placed over the southern Adriatic. The
jet stream stretched from eastern Europe across the Figure 3. Synoptic situation on 17 July 2017 at 0000 UTC (a) and on
central Adriatic and southern Italy, and turned toward 18 July 2017 at 1200 UTC (b) (source: German Meteorological Service,
the Ionian Sea at 300 hPa. Such synoptic conditions DWD)

Figure 4. Distribution of geopotential height (dam), wind speed and direction (shades of colour) at 850 hPa (left) and 300 hPa which are
filled marking the wind speed above 30 m s-1 (right), on 17 July 2017 at 0000 UTC with shades of green (a and b), and on 18 July 2017
at 1200 UTC (c and d) (according to the ECMWF analysis).
Figure 5. Wind speed and wind direction at 10 m height (shaded in color) and pressure reduced to mean sea level (right) to ALADIN/ALARO
HR analytical fields at 4 km (left), and wind speed and direction at 10 m height (shaded in colors) to analytical fields ALADIN/ALARO DADA
model, dynamic adaptation at 2 km (right), 17 July 2017 at 0000 UTC (a and b) and 18 July 2017 at 1200 UTC (c and d).

Regarding the change in speed and direction of the
wind on 17 and 18 July, the wind field was analyzed in
more detail from the reanalysis of the mesoscale
numerical prognostic model ALADIN/ALARO Croatia. At
0000 UTC on 17 July, the strongest pressure gradient
was in the northern part of the Mid-Adriatic. On 18 July,
the wind was light along most of the Adriatic coast,
and even in the Split area. Trošić and Trošić (2010)
have also shown the strong influence the local orogra-
phy and synoptic situation has on wind rotation during
summer Bura, at its onset in the Northern Adriatic.
However, at midday and through the early afternoon on

18 July in the Split area, there was occasionally either a
moderate south-westerly or moderate north-westerly
wind at 850 hPa (see Figure 4c). The circulation at 850
hPa over south-eastern Europe was possibly causing
the north-western "synoptic wind" with a wind speed of
up to 6 m s-1 over the Split area that then supported the
development of the south-western sea breeze, with
wind speeds of up to 8 m s-1 (see Figures 5c and 5d).
The vertical wind profile from the pseudotemp of the

ALADIN/ALARO mesoscale numerical forecast model
shows wind shear on 18 July for the Split-Marjan
station at 1200 UTC (Figure 6d). Analysis of the verti-
cal atmospheric structure was shown to be important
by Byram (1954) for the examples of the big fires on
the Croatian islands Kornati and Brač, by Vučetić et al.
(2007) and Mifka and Vučetić (2012), respectively.
The turbulent kinetic energy model forecast for midday

and the early afternoon on 18 July for the area around
Split also shows wind shear in the near-surface level of
the atmosphere, with instability present in the dry air.
Figure 7 shows the ALADIN/ALARO model of turbulent
kinetic energy at 4 km on 18 July at 1500 UTC. Along
with the equalized air pressure reduced to mean sea
level over Croatia, and with the weakening and further
shift of the upper level cyclone to the east on 19 July,
the wind weakened and there was no turbulence,
although there was still instability in the dry air. On
19 July, the wildfire was localized.
Conclusion
The behaviour of the front of the fire on 17 July was
dominated by a strong or very strong Bura due to the
influence of the local orography, such that it was called
"a fire carried by the wind". The next day, on 18 July,
when the main synoptic systems moved, the Bura
weakened to moderate and turned south-westerly, after
Figure 6. Pseudotemp from the mesoscale numerical model
which the fire was mostly affected by the coastal circu-
ALADIN/ALARO HR analysis for the Split-Marjan station, on 17 July
at 0000 UTC (a) and 1200 UTC (b), and 18 July at 0000 UTC (c) lation. Therefore, it can be concluded that the behav-
and 1200 UTC (d). iour of the fire was conditioned not only by the wind

Figure 7. ALADIN/ALARO model turbulent kinetic energy forecast at 4km on 18 July at 1500 UTC, model run from 17 July at 0000 UTC.
speed, but also by the combination of wind shear (wind Tian, X, McRae, DJ, Jin, J, Shu, L, Zhao, F, Wang, M
direction and wind speed change by height), i.e. turbu- (2011) Wildfires and the Canadian Forest Fire Weather
lence in the near-surface layer of the atmosphere, insta- Index system for the Daxing'anling region of China.
bility in the dry air, and the thermodynamic processes International Journal of Wildland Fire 20, 963-973.
that were caused by the extreme fire itself. The meteo-
rological parameter pointing to the possible occurrence Trošić T, Trošić Ž (2010) Numerical Simulation and
of wind shear and the influence of orography was the Observational Analysis of the Bora of Pag’s Ribs.
prognostic field of the turbulent kinetic energy of the Boundary-layer meteorology 134, 353-366.
numerical mesoscale model of ALADIN/ALARO Croatia.
Van Wagner CE, Pickett TL (1985) Equations and
FORTRANProgram for the Canadian Forest Fire
References WeatherIndex System. Canadian Forestry Service,
Government of Canada, Forestry Technical Report, 33,
Byram GM (1954) Atmospheric conditions related to 18 pp.
blow up fires. A Publication of the National Wildfire
Coordinating Group, PMS 815-NFES 2239, 29 pp. Vučetić V, Ivatek-Šahdan S, Tudor M,Kraljević L,
Ivančan-Picek B, Strelec Mahović N (2007) Weather
Carvalho, A, Flannigan, MD, Logan, K, Miranda, AI, Analysis during the Kornat Fire on 30 August 2007.
Borrego, C (2008) Fire activity in Portugal and its rela- Croatian Meteorological Journal 42, 41-65.
tionship to weather and the Canadian Fire Weather
Index System. Vučetić M, Vučetić V (2011) Fire risk analysis during
International Journal of Wildland Fire 17, 328-338. the Kornat fire on 30 august 2007. Vatrogastvo i
upravljanje požarima 1, 12-25.
Mokorić, M., Kozarić T (2012): Kvarner fire 23rd and
24 July 2012 year –weather analysis, Vatrogastvo i
upravljanje požarima 2, 53-66.

Changing the Way we Warn for Weather
Elín Björk Jónasdóttir,
Group leader for weather services, IMO
Iceland’s weather is in general rather violent, at least rologists, as societal impact has not been the focus
compared to most of Europe. In general, it is windy, point of forecasting until now. Single message warn-
relatively cold and often wet. This is not new to those ings are also likelier to get a reaction from the public
of us who live here, but to the two million tourists that as well as decrease confusion or mixed messages
travel to Iceland every year, weather and the effect it from different sources. During the time the system
can have on holiday plans often comes as a surprise. was in development it was decided among the stake-
holders that IMO will never issue a red (very high like-
Familiar with the weather or not, many Icelanders are lihood, very high impact) warning without the consent
dependent on the weather. Fishermen, drivers, the of the Civil Protection Authorities. This relieves some
aviation industry, the tourism industry and farmers all of the pressure from the forecasters among other
depend heavily upon weather forecasts and warnings. things.
In the past decade or so, with widespread sources of
weather information it has become clear that a single
message warning for weather and the effect it can First winter experiences
have is the best way to communicate the impact of
weather and the necessary precautions people need The warning system was launched on 1. November
to make. It is also safe to say that the literacy of 2017 with a small media campaign and introduction
weather has decreased among the population and for the public. Introducing the system turned out to be
many have difficulty making decisions given a general made easier than anticipated by the weather itself as
forecast. With all this in mind the Icelandic the first deep low-pressure area of the winter was fast
Meteorological Office developed a new impact-based approaching. Before noon on the first day the first
colour coded warning system based on the Common warnings were out. During the next 48 hours all fore-
Alerting Protocol (CAP) recommended by WMO. casts converged on a deep cyclone close to Greenland
with a very sharp front passing Iceland from south-
The system was developed by the IMO IT- west to northeast.
Development team in cooperation with the Division
for Forecasting and Warnings. The system is based on Considering time of year and week it became clear
CAP 1.2. We issue warnings for up to 5 days, for rain, that upgrading the warning to high impact was neces-
wind, blizzard, snow and lightning. Colours are based sary as the traffic towards the capital area is often
on EMMA tradition, yellow, amber and red and for the quite dense on Sunday evenings. Other impacts
first version the warnings go out for pre-established considered were delays to take-off and landing at KEF
forecasting areas, 10 in total. All warnings are issued Airport, the major international airport in Iceland. In
in Icelandic and English simultaneously and the warn-
ings are open via XML repository available for anyone
to consume and visualise. Warnings are visualized on
the IMO’s webpage as well as in the IMO app. The
impact based system replaced the older system
which was strictly based on climatological thresholds.
One of the major improvements made in preparation

of launching the new system was that Icelandic Civil
Protection Authorities as well as other stakeholders
such as the Road Administration and Aviation author-
ities now join the team of forecasters and natural
hazard sepcialists at IMO in a daily meeting held in
the monitoring room at IMO. It was also clear from the
start that the impact factor would have to be estimat-
ed by Civil Protection personnel along with the meteo- Figure 1: The forecasting areas of Iceland

collaboration with Civil Protection and the Road During a meeting with all stakeholders on Friday 3rd
administration the upgrade to high impact – high like- November it was briefly discussed what would need
lihood was made on Friday, with 48 hours lead time to to happen for this weather event to escalate to a red
the actual event. This led to an amber warning for the warning. Consensus was reached among the partici-
southwest and west of Iceland and yellow warnings pants that since this was not during rush hour, the
for the rest of the forecast areas. The main reason for timing would almost completely exclude the possibili-
escalating to amber so far in advance was the fact ty of a red warning, with one exception. If lightning
that the impact would happen during a weekend and were to strike the high voltage lines from the
it was important to get the information out sooner Reykjanes peninsula towards the capital, taking out
rather than later so that travellers would know to power for thousands of homes for an extended period
change their plans for the returning trip towards the we might reconsider. Lightning in Iceland is not exact-
capital area. ly rare, but not common either.
Figure 2 ECMWF forecast +118h Valid Sunday 5th Nov. 2017 at 18Z
Figure 5 Visualisation of CAP warnings on IMO's website 5th

November 2017
As the weather approached it became clear that the

Figure 3 ECMWF forecast +66h Valid Sunday 5th Nov. 2017 at 18Z warnings were justified, maximum 10 min. wind-
Valid Sunday 5th Nov. 2017 at 18Z speeds exceeded 20 m/s in all areas of the country
and wind gusts exceeded 35 m/s widely. Search and
rescue teams were called out to control damage at
construction sites, help stranded travellers on roads.
While the inner edge of the front was passing, light-
ning did strike the Reykjanes peninsula and a high
voltage power line. Power outages were extensive and
included the main radar for the airport, parts of the
airport and the towns close to it as well as parts of the
capital area. All in all around 40.000 people are
thought to have been out of power for a while during
the evening. Power was restored just as the front
passed, the southwestern most point at round 23Z in
the evening. We did not go to red, and as it turned out
Figure 4 ECMWF forecast +18h valid Sunday, Nov. 5 01 0 2017 at 18Z no warnings went to red the entire winter.

Figure 6 High resolution Harmonie-Arome forecast for 10 m winds
valid 18Z on 5. Nov. 2017
Yellow fatigue and unexpected

consequences
The warning system was well regarded and accepted Some complaints were communicated to IMO, mostly
by the public and stakeholders. The change from a regarding the complexity of the system, lack of docu-
strict climatological threshold system to a more flexi- mentation or other things that mostly fixed them-
ble impact-based system was challenging for every- selves as everybody got more used to the new
one, not least the forecasters. Forecasters do not system. An unexpected complaint was that some
issue press releases regarding especially dangerous Icelanders found the use of the colour amber objec-
weather or jumps in forecasts anymore, but the addi- tionable. Amber or orange, is in Icelandic “appel-
tional work of issuing warnings for all of the forecast- sínugulur” – which, granted, is a rather difficult word
ing areas added quite heavily to the work load for the to say. It has however been in the Icelandic language
on-duty forecasters. The first storm provided a good for some 150 years, or since oranges first appeared in
example, but as the winter wore on a certain misun- letters and books. Some complaints are always
derstanding became obvious. The public seems to expected with dramatic changes to a system that has
only rely on the colours and do not read the whole been the same for years, but this one certainly was
warning text, which can often be specific, or look at not and just shows that one can never be completely
where in the matrix of impact and likelihood the warn- prepared.
ing actually lies. In addition, yellow warnings started
to miss their mark when weather was such that IMO Next steps in development of the system will be to
issued yellow warnings for wind, snow or blizzard for include free-drawn warning areas and thus limiting
days – indicating that without proper caution and the area under warning. Some minor adjustments will
preparation the weather could be hazardous, so on also be done on the forecasters interface as well as
days 5-10 of these stretches of bad weather the the production of a handbook with well worded warn-
Search and Rescue teams were deployed more often ings but IMO still counts the launch of this new
to help travellers who had taken a chance on moun- system a success.
tain passages etc.
It was also brought to our attention that school

districts, fire departments and others were using the
system as guidelines to when to close schools, ask
parents to accompany children or issue reduced
services of some sort. Most often amber warnings
were used as guidelines, which in some cases is too
late. Cases can be made for schools asking parents to
accompany young children to school much earlier
than during an amber warning, and sometimes during
amber warnings, it is not necessary to limit the
services of home health care professionals or other
services.

Storm naming: the First Season of Naming by
the South-west Group: Spain-Portugal-France
Paula Leitao (IPMA- Instituto Português do Mar e da Atmosfera), Bernard Roulet
(Meteo France), Jaime Rey (AEMET)
Introduction gust associated with the passage of a storm or low

pressure (regional winds such as the Mistral are not
Following the success of storm naming by MetEireann taken into account).
and the Met Office in 2016-2017 (cf WGCEF Task Team • Area: Atlantic Ocean and western
on Storm Naming in Europe ; Authors Cusack, Mediterranean Sea.
Paterson, Lang, Csekits, WGCEF newsletter N°22), • The NMS who first issues an orange/red warn-
three other countries in the south-western part of ing (or Vigilance) names the storm and informs the
Europe - Spain, Portugal and France - decided to join three others.
the process of storm naming, following recommenda- • A named Storm keeps the same name its whole
tions of the Task Team. life.
System specifications were discussed by partners The coordination with the North-west Group is crucial
during the year 2017 via mail exchanges, then web- to ensure that the same storm gets a single name. If a
conferences during Autumn 2017 allowed the Group storm is named by one group and moves to the other
to finalise the process of storm naming for the south- group, it keeps the same name. In the rare cases
west. The month of November was then used as a where a barotropic storm is a post-tropical storm, it
blank test, before the system entered its operational keeps the name given by NHC Miami preceded by ‘Ex’.
phase in December 2017, for the winter season of Notice that this procedure was used by Met Eireann in
2017-2018. October 2017 with Ex-Hurricane Ophelia.
Emails are favoured for exchanging information

Principle of storm naming between the National Meteorological Centres, but
direct phone calls between operational forecasters
The following principles of naming were applied : are possible if needed. The following form is used by
• A depression will be named only if at least an the first NMC who name a storm, to inform the NMCs
orange warning is expected for wind speed or wind partners.
STORM NAMING COORDINATION FORM (SW GROUP)

PROPOSED NAME FORECASTED LOW CENTRE LOCATION Latitude
(at first orange/red wind warning (in decimals)
onset time) Longitude
(in decimals)
FIRST WIND DATE (YYYYMMDD) FIRST WIND WARNING DATE (YYYYMMDD)

WARNING ISSUING TIME (HH:MM CET) ONSET TIME (HH:MM CET)
MAXIMUM WIND Orange OTHER ORANGE/RED Rain Coastal event

WARNING LEVEL Red WARNINGS ISSUED Snow Others
REMARKS
WARNINGS Météo-France: http://vigilance.meteofrance.com/

WEBSITES IPMA: https://www.ipma.pt/en/
AEMET: http://www.aemet.es/en/eltiempo/prediccion/avisos
Meteoalarm: http://www.meteoalarm.eu/index.php?lang=en_UK
Naming list winter season 2017-2018 of the SW group

Ana-Bruno-Carmen-David-Emma-Felix-Gisele-Hugo-Irene-Jose-Katia-Leo-Marina-Nuno-Olivia-Pierre-Rosa-
Samuel-Telma-Vasco-Wiam

everywhere, and up to 160 mm in 24 hours in many
places. The main impacts were the closure of two big
sea ports, Tarifa and Algeciras, and the large number
of incidents, 260 in Galicia alone. The concept of storm
naming had a warm welcome in Spain, becoming front
page news in all important media. On the downside,
flight cancellations in Europe were attributed to Ana in
some media, even before it was formed.
Portugal
A red warning was issued for wind and an orange
warning was issued for heavy precipitation, snow and
coastal events. There were significant damages caused
by the wind, with 144 km/h gusts reported, and red
warning levels of precipitation were observed. On
Madeira Island, orange level winds were observed
when the winds veered northerly on the passage of the
cold front, with wind gusts exceeding 90 km/h. After a
long draught period, naming this low helped commu-
nicate to the media and public the important change in
the weather pattern and the strong effects it caused.
Case study of the first named storm

‘Ana’
‘Ana’ was named by AEMET on the 8th of December
2017 with a forecast position of a deep low pressure
(964 hPa) in the Bay of Biscay, south-west of Brittany,
on the 11th of December 2017 00UTC. This storm
affected the three countries of the South-west Group.
France
An orange warning was issued, not only for winds, but
also for other parameters connected to the storm:
snow, coastal events, heavy precipitation and
avalanches (in the Alps). Having a named storm
meant all these parameters could be connected to the Other storms of the season
storm, making the Vigilance more understandable for Storm Bruno 2017/12/26
the public.
Storm Bruno mostly affected France with gusts up to
120 km/h in Brittany (orange warning) and north of
Spain, with gust up to 110 km/h (orange warning). In
Spain Portugal an orange warning was issued for wind and
A "Special Warning" mentioning Ana was issued on coastal events, with gusts rising up to 115 km/h on
the 8th of December along with more than 30 orange the passage of the cold front associated with this low.
warnings, but on the 11th of December the situation
ended with half of Spain under orange warnings and
Storm Carmen 2018/01/01
with 16 red warnings (3 for precipitation). Wind gusts An orange warning was issued in France. The cold
were over 120km/h, and not only in mountain or front associated with the storm was very unstable – a
coastal stations. In Galicia, precipitation was very wind power generator was destroyed by the localised
persistent, with more than 80 mm in 12 hours almost tornadoes which occurred.

at 00UTC, with a central pressure of 965 hPa – requir-
ing an orange warning for the wind on the passage of
the cold front. On the 27th February, this frontal system
crossed mainland Portugal, with significant damage to
trees caused by freezing rain (a very rare phenomenon
in Portugal) and disruption to car traffic caused by
snow fall. Later, Emma moved eastward, filling to
979hPa over the Atlantic to the west of Portugal on the
1st March, then moving northward towards the United
Kingdom and Ireland. As a result of storm Emma and
its associated frontal systems during this period,
orange warnings were kept in force over mainland
Portugal, whilst a tornado formed over the sea and
entered the town of Faro on the 28th February. A red
warning for wind and coastal events and an orange
warning for rain was kept in force for the Madeira
Islands where 140 km/h wind gusts were observed at
the airport, with a significant impact to air traffic.
In Portugal, an orange warning was issued for coastal

events and yellow level (74 km/h) wind gusts were
observed at the extreme north of the coastal area of the
country.
In Spain an orange warning was issued for the mouth

of the river Ebro (90 km/h).
In France, when the warm front associated with the
Storm David 2018/01/17
low moved northward on 28th February, it came up
This low crossed England, but was not named by the against the very cold air advected across the country
Met Office as the impact of the storm was not expect- in the days before by a strong Scandinavian high
ed to reach orange warning thresholds there. pressure. The result was heavy snowfall over France,
especially in the Languedoc region: more than 40cm
The low was named by Meteo-France due to an orange of heavy and wet snow fell near the sea and on the
Vigilance being required for the extreme north of town of Montpellier. A red warning for snow was
France. The storm deepened quickly in the North Sea, issued, but major disruption occurred on the A9
resulting in the greatest impacts being seen in Belgium motorway and in the town of Montpellier.
(orange warning), and the Netherlands (red warning).
David´s path was too high in latitude to affect Spain
and Portugal.
Storm Eleanor 2018/01/02

This storm was named by Met Eireann, but also had a
great impact in France (bringing strong winds, heavy
rainfall and leading to coastal destruction in the
Channel). The name was widely used by media.
Storm Emma 2018/02/26

This was a long lived storm with strong impacts, named
by the Azores Regional Department of IPMA. It had a
forecast position at 41N 36W on the 26th February 2018

Emma also had a great impact in Spain, with strong gusts of up to 120 km/h were observed at the airport,
gust winds, up to 110 km/h, and heavy rainfall (140 causing significant impacts.
mm in 24 hours in Grazalema) in the south-west. As in
France, the collision of the two air masses triggered an In mainland Portugal, 100 km/h wind gusts were
abnormal episode of widespread snow, covering the observed across the coastal region, with a red warning
northern half of the country, even affecting cities at for coastal events and heavy precipitation being issued.
sea level such as Bilbao (5 cm).
In Spain, red warnings were issued for wind gusts in
Storm Felix 2018/03/09 the Cordillera Cantábrica and for coastal events in the
west of Galicia.
A vast complex low persisted over the Atlantic but a
new deep low with a minimum pressure forecast of Storm Gisele 2018/03/14
979 hPa at 47N 24W on the 9th March at 00UTC, was
named as Felix by IPMA, due to an orange warning in AEMET named Gisele, a new deep low pressure
force for the Madeira Islands. Wind gusts of up to 150 system forecast with a low centre of 968 hPa at 48N
km/h were observed across mountain areas and 15W on the 14th March at 06UTC (orange warnings for

wind and coastal events), a few minutes before the Conclusion
Azores Regional Department of IPMA took action and
issued an orange warning for wind and coastal events The storm naming system by the South-west Group
for the 13th March. Wind gusts of up to 100 km/h were was quite successful in raising the awareness of citi-
observed at Graciosa Island airport. zens about meteorological hazards related to storms
When Gisele moved eastward, an orange warning for and encouraging people to follow the recommenda-
wind and coastal events was issued in mainland tions of the authorities.
Portugal. A wind-gust of 110 km/h was observed and
several meso-vortices formed, causing at least one Coordination between NMSs of both naming groups
tornado, on the passage of a very unstable cold front, involved in storm naming was quite good, but it is
on the 14th March. important to work on a more efficient system, that
does not rely on exchanging e-mails.
Storm Hugo 2018/03/23
AEMET named Hugo, a new deep low pressure fore- The existence of a pre-defined list of names known by
cast with a low centre of 971 hPa at 48N 09W on the the media and the public allows the public to guess
23th March. A "Special Warning" was issued regarding which low is going to be named next but may produce
the explosive cyclogenesis of Hugo. Red warnings ‘fake news’, for instance, when some private compa-
were issued for coastal events on the north-west coast nies try to pre-empt the naming of a storm before the
and widespread orange wind warnings were issued, NMS does so. At the same time there is additional
with gusts of up to 120 km/h expected. complexity for the public in understanding whether
the named low will affect their country or not. The
An orange warning for coastal events and yellow warn- solution may be in making use of graphical and text
ing for wind and snow was issued for mainland tools and through the use of common communicating
Portugal. An orange warning level for wind was systems (eg. websites, social media) to reach the
observed with wind gusts over 100 km/h at towns in public in the most efficient way.
the mountainous areas of central mainland Portugal.
Further work and coordination are necessary before
Storm Irene 2018/04/16 reaching a single naming system in Europe.
The Azores Regional Department of IPMA named
Irene, a new deep low pressure system, forecast with
low pressure centre of 952 hPa at 49N 29W on the
16th April, due to an orange warning for wind and
coastal events being issued. Wind gusts of up to
110 km/h were observed. This low then moved north-
ward over the Atlantic.

Can we forecast the sudden dust storms
impacting Israel's southernmost city?
Anna Pavlov, Nachum Yudovitz, Alissa Razy, Elyakom Vadislavsky, Alon
Shtivelman, Harel Muskatel, Amit Savir (Israel Meteorological Service)
Eilat (Fig. 1) is the southernmost city in the State of This article discusses the cases in which these clouds
Israel, and it lies on the shores of the Red Sea. The caused severe dust storms in the city resulting in the
city is characterized by very high temperatures in the closure of the airport, and the possible ways to
summer, and a pleasant and comfortable climate predict this phenomenon.
during winter, which attracts millions of tourists a
year, making it one of Israel's major tourist cities.
The meteorological background
Eilat is known for its desert climate, which combines a
small annual precipitation with low humidity. The rain As described above, in order for haze or sandstorms
regime in Eilat is characterized by a low frequency of to develop in the Eilat area due to this synoptic
rain events, but some of the rain events that do occur scenario, clouds of significant vertical extent must
in the city are high-impact events, causing damage move in from a large sandy region. However, not all
and disruption due to Cb clouds accompanied by clouds coming from a dust-source region cause sand-
strong winds (microburst), flash floods and hail, and storms. There must be sufficiently strong vertical
reduced visibility (sometimes to the point of dust motion to cause a microburst, alongside relatively
storms). weak synoptic winds which fail to neutralize the gust
front. This allows a cloud of dust to be advected to the
This relatively rare weather in Eilat is usually caused city. The source region of dust advected to Eilat can be
by southern weather patterns, and mainly by the one seen in Fig. 2 (largely the Sinai Peninsula).
known as a Sharav Low – a warm-core low pressure There are two main problems facing the forecaster in
system that travels along the coast of North Africa predicting dust or sandstorms in Eilat:
from the area of Morocco, advecting warm and dry 1. A small number of measuring stations in the
desert air. Sometimes its warm front is accompanied Sinai region providing real-time information about the
by highly-developed medium clouds that may create formation of sandstorms in the area
thunderstorms in the Eilat area. 2. The rapid development of the phenomenon
Occasionally dust clouds will also form under these

storm clouds, and quickly reduce visibility. This Analysis of events
phenomenon is variously reported by weather
observers as haze, dust/dust storms or Two instances of a sudden, rapid decrease in visibility
sand/sandstorms. Hence the terms dust, haze and were selected. One took place on the afternoon of
sand are used interchangeably here. 18.05.2017, when the visibility dropped to 1500
Figure 1: Eilat is the southern-

most city in the State of Israel

evidenced by a difference of 33 to 35 degrees
between 850 hPa to 500 hPa), along with a zone of
lifting air around Eilat during the time of the Cb forma-
tion as seen in the vertical velocity at 700 hPa (Fig 5).
In such events, the convectively active clouds tend to

result in microbursts, leading to sudden extreme
changes in wind speed and direction at the surface.
Analysis of the satellite images from these events
indicates the development of a line of clouds with
significant vertical extent passing over sandy areas
and approaching Eilat. In the images from 18.05.17,
the cloud line moves from Sinai toward the city (Fig
6a), whereas in the image from 01.03.17 (Fig 6b),
these clouds are developing rapidly north of Eilat.
Figure 2: The source region in the Sinai Peninsula and southern Forecasting tools and limitations
Israel of dust advected
The tools available to the forecaster to assist in
meters at the airport and to 500 meters in the city itself. predicting this sudden, disruptive phenomenon can
In the second event, which took place on 01.03.2017, be divided into two main groups: tools that enable the
the visibility at the airport dropped to 5 kilometers, and prediction of the event up to a few days in advance,
for a short period down to 900 meters. and tools for real-time prediction (nowcasting).
In both events a warm low tracked from Egypt towards In advance of the event, the forecaster will make use
Israel (Fig. 3). As the warm sector moved over Israel, of global models to track the expected path of a warm
advecting hot air into the lower levels of the atmos- or relatively southern low that is accompanied by an
phere, an upper-level trough approached. The resul- upper-level trough and highly unstable conditions.
tant south-southwesterly flow transported a lot of The seasonal climatology of such events is also taken
moisture from the tropics (Fig 4). This synoptic system into account – the advection of heavy dust and haze
caused highly unstable atmospheric conditions (as is far more likely in late winter and spring than in
Figure 3 (a): The sea-level pressure chart for 18.05.2017 at 00Z; (b): The sea-level; c): The sea-level pressure chart from 01.03.2017 at 00Z;
(d): The sea-level pressure chart from 01.03.2017 at 12Z

autumn (and never in summer). The forecaster will
also make use of dust models, checking for indica-
tions of increased dust load along the expected track
of the low. Some of these dust model products are
shown in Fig. 7 for the 18.05.17 dust event: Fig. 7a is
an optical depth chart, and Fig. 7b shows cross-
sections of dust density, illustrating the vertical distri-
bution of the dust. Most of the dust was seen west of
Israel, and was expected to reach Eilat later than it did
as a result of synoptic changes, rather than from a
microburst. In addition, the use of water vapor and
"Airmass" satellite images provides a way to identify
tropical plumes of moisture extending northward to
the eastern Mediterranean, which is essential for acti-
vating the warm front.
Due to these tools' limitations in predicting local-

scale phenomena, and the fact that the region where
the warm lows form is data-sparse, the accurate fore-
casting of such events largely relies on real-time
monitoring via remote sensing, and on regional
models and other tools developed by the Israel
Meteorological Service.
One of the regional models used by the IMS is the

COSMO model at 2.8 km resolution, to predict local
phenomena. The high-resolution model can provide
more important indications of thunderstorm develop-
ment and associated gust fronts. One clear example
is from an event on 01.05.2018. The ground-level
wind field indicated a possible microburst (Fig. 8a),
Figure 4 (a): The "Airmass" MSG satellite imagery showing a
while the rain field (Fig. 8b) showed unusually large
tropical air mass accompanied by a large amount of high and
medium clouds over the eastern Mediterranean. accumulations forecast in the same area. It should be
(b): The highly developed clouds as seen in the 700 hPa relative noted that in our two haze events of interest, there
Figure 5: The ECMWF charts of vertical velocity at 700 hPa, indicating the zone of lifted air NW of Eilat during the time of the Cb formation.
The strongest updrafts, characterized by large negative omega (ω) values, can be seen at 12Z, whereas in the subsequent charts the values
have decreased rapidly, indicating the decay of the Cb cells.

Figure 6 (a): Clouds of significant vertical extent pass over the sandy areas and approach the city of Eilat (18.05.2017)
Figure 6 (b): Clouds of significant vertical extent pass over the sandy areas north of Eilat, approaching the city (01.03.2017)
Figure 7 (a): Optical depth from the CAMS model,

showing the mass of dust
approaching Israel from North Africa. In the red circle,
possible signs of the dust developing in Sinai.
(b): East-west cross-sections of dust density along
Israel

Figure 8 (a): The surface wind chart of the COSMO model from 01.05.2018, indicating a possible downburst.
(b): Heavy precipitation accumulation is forecast in the same area.
Figure 9 (a): Radar images from 18.05.2018 (left) and 01.03.2018 (right).
n red circles the areas of the CB clouds.
was no clear indication of a microburst in the model's more than 250 kilometers from the area of interest.
low-level wind field. Therefore, only extremely tall clouds will be detected,
while the rest of the precipitating clouds will not be
One of the main tools used in real time is the rain radar. picked up at all, or will be detected as weak echoes.
The IMS' radar is located on the roof of the building, (Additional radars that are located farther south do not

Figure 10: An image from the lightning detector owned and operated by
the Israel Electric Corporation (IEC).
Figure 11: The sharp drop in visibility (18.5.2017) as seen in the visibility sensor of Eilat airport.
operate continuously.) It is impossible to conclude In the May event, lightning was first detected at 11:20Z
solely from the weak echoes in the radar images from in western Sinai, while slightly after 12:00Z the visibili-
the May event (Fig. 9) that there are Cb clouds with high ty in Eilat was already dramatically impaired due to the
tops. Only cross-referencing the information with other sandstorm created by the significant Cb's gust front
tools, including satellite imagery, makes it possible to (Fig 11).
correctly assess the intensity of the clouds. In particular
the "Severe Storms" satellite image indicates the An additional remote-sensing tool is the "Dust" satel-
cloud's significant vertical extent, and intensity. lite imagery. It can be used to track both larger dust
Another tool that complements the radar is the light- events from significant storms, and the development
ning detector (Fig. 10), which is owned and operated of smaller, localized sandstorms. Unfortunately, in
by the Israel Electric Corporation (IEC). Its primary limi- many of these events the same cloud structures that
tation is that it mainly detects cloud-to-ground light- produce the microbursts responsible for the heavy
ning, while the Cb clouds in these events tend to dust events, also obscure the dust that is kicked up
develop from medium clouds (Altocumulus beneath them. For instance in the May event (Fig. 12),
Castellanus), and therefore at first the majority of the significant Cb clouds (indicated by the high cloud
lightning is inter-cloud and goes largely undetected. tops in red) hid the dust event below them.

Figure 12: The "Dust" satellite image,
18.5.2017 13Z. Indications of dust hiding below
the large amount of clouds (circled
Summary
While Israel's proximity to deserts with fine dust parti- sparse areas, and remote sensing tools do not always
cles leaves it vulnerable to advective haze events, these provide clear indications of the developing phenome-
generally do not reach the city of Eilat due to the local non, thus posing a challenge to forecasters. Since these
wind regime. However heavy dust events do occasional- events are poorly forecast by global models, it has been
ly impact the city, as on 18.05.2017 and 01.03.2017, necessary to develop and use in real time mesoscale
due to microbursts and strong gust fronts in the dust and microscale forecast products to aid the forecaster,
source regions. These events develop quickly in data- although certain limitations remain.

The 31st Nordic Meteorological Meeting
The 31st Nordic Meteorological Meeting was held in Reading, Pererik Åberg from Swedish Television, and
Reykjavík 18th to 20th June 2018. Øyvind Byrkjedal from Kjeller Vindteknikk. Talks from
participants ranged widely, from model developement
More than 60 meteorologists attended the three-day to education, observation networks and remote sens-
meeting which included nine sessions of talks and an ing as well as impact based warnings and communi-
afternoon trip to the west of Iceland with a conference cation with the public and stakeholders.
dinner held in the town of Borgarnes. The meeting
was organized by the Icelandic Meteorological Society
in cooperation with The University of Iceland and the
Icelandic Meteorological Office. The Nordic
Meteorological Meeting is held every two years, and
every tenth year in Iceland.
Topics of the meeting were varied and included topics

such as development of numerical weather prediction
models and Nordic operational cooperation, verifica-
tion methods and comparison of results, probability
forecasting, meteorological and climate information
for the energy sector, communication of forecast and
meteorological information, remote sensing systems
and new data sources and Arctic studies such as
observations of climate change and climate model-
ling. Invited speakers were Patrick Samuelsson from
SMHI, Mikko Laine from FMI, Andrea Vajda from FMI, Figure Most of the participants of NMM31 along with family
Ewan O’Connor from FMI and The University of members joining the excursion, taken at Snorralaug hot spring
in Reykholt, west Iceland.
Photo: Hallvard Mørk Tvete

Euroforecaster 2017 Newsletter

Uploaded by

Copyright:

Available Formats

Euroforecaster 2017 Newsletter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Euroforecaster 2017 Newsletter

Uploaded by

Copyright:

Available Formats

The European

Newsletter of the WGCEF N° 23

Minutes of the 23rd Annual Meeting of the Working Group on Co-operation

Sting Jets and other processes leading to high wind gusts:

16 Forecasting Freezing Rain in the UK – March 1st and 2nd 2018

24 The Extreme Wildfire, 17-19 July 2017 in Split

30 Changing the Way we Warn for Weather

Storm naming: the First Season of Naming by the South-west Group:

45 The 31st Nordic Meteorological Meeting

46 Representatives of the WGCEF

Printed by Meteo France

Dear Readers and Colleagues,

We hope you find this newsletter enjoyable and informative.

Christian Csekits and Jos Diepeveen

The European Forecaster 3

OGRODNIK, Michal - IMGW/Hydrometeorolo-

HAUSEN, Robert - DWD/Germany Thursday, 12th of October

HEWSON, Tim - ECMWF

4 The European Forecaster

Cristian Csekits, ZAMG/Austria Robert Hausen DWD/Germany

The European Forecaster 5

6 The European Forecaster

The European Forecaster 7

Marjan Sandev, Czech • General news concerning the weather office

Andre Charles gave an update about the website

Jos offers all participants to connect via Facebook to

8 The European Forecaster

EUMETNET Task Team on Storm-Naming

The European Forecaster 9

Other significant gusts phase II 144 km/h, Chambéry-Aix

Other significant gusts phase III 184 km/h, Dramont

Affected Surface (% of the Country) 32% 31%

“Lamb (1991)” Index 595 618

“Storm Severity Index” (SSI) 54 50

10 The European Forecaster

Figure 1: Graphical view (ANASYG) of the wind-storms.

The European Forecaster 11

12 The European Forecaster

Figure 5: Vertical cross-section, wind-storm “Zeus”.

The European Forecaster 13

Figure 6: Gusts during phases II and III of wind-storm “Zeus”.

Conclusion: Jet”) which is often used in French to describe that type

14 The European Forecaster

Coronel B., Ricard D., Rivière G., Arbogast P. 2016:

Lamb H.H., Frydendahl K., 1991: Historic storms of

Météo France, 2017, Tempêtes en France métropoli-

The European Forecaster 15

Introduction brought a bitterly cold polar continental air mass

16 The European Forecaster

The European Forecaster 17

18 The European Forecaster

Figure 4 shows the ensemble output from

The European Forecaster 19

20 The European Forecaster

Warnings Issued As forecast confidence increased, further evidence was

The European Forecaster 21

routes open, whilst another main trunk road, the A38

There was also an impact on the train lines, with

22 The European Forecaster