ECMWF products for RA I and ACMAD

Dr Umberto Modigliani Head, User Support

2 November 2017

umberto.modigliani@ecmwf.int

October 29, 2014

The operational forecasting system High resolution deterministic forecast (HRES) :

– twice per day 9 km 137 levels, to 10 days ahead

Ensemble forecast (ENS):

– twice per day 51 members, 18 km 91 levels, to 15 days ahead – Monday/Thursday 00 UTC extended to 1 month ahead (Monthly Forecast, 18/36 km )

Ocean waves: twice per day

– HRES-WAM: 10 days ahead at 14 km (coupled)

– HRES Stand Alone Wave (SAW) model : 10 days ahead at 11 km – ENS-WAM: 15 days ahead at 28 km (coupled)

Seasonal forecast: once a month

– 51-members, ~80 km 91 levels, to 7 months ahead

– sub-set of 15 members is run for 13 months every quarter (30 years of hindcasts)

More details at:

http://www.ecmwf.int/en/forecasts/documentation-and-support

Slide 2 2

October 29, 2014

Seasonal forecasting System-5 and EUROSIP Seasonal System-5 - Key configuration choices for System-5 concern

– the resolution (Tco319 or TCo399, L91), same as Ext ENS

– the handling of the stratosphere,

– the use of stochastic physics,

– the sea-ice model configuration, and

– the initialization of the land surface.

There is the strategic question of how closely System-5 in should be aligned with the monthly forecast ensemble.

EUROSIP - It has been upgraded to latest version of Meteo-France seasonal forecasts, and after June-2016 Council decision JMA is an Associate Partner.

S5, 43r1 S4, 36r4

S5, 43r1 S4, 36r4 The figure shows preliminary results for NINO4 SST prediction (45

start dates,1993-3015, 5-member ensembles).

3

October 29, 2014

Drifters: MSL Pressure, 10m-Wind

PILOT/Profilers: Wind

Aircraft: Wind, Temperature, Humidity

SYNOP/METAR: MSLP, 2m-Rel.Hum SHIP: MSLP, 10m-Wind

Radiosonde balloons (TEMP): Wind, temperature and humidity

Note: We only use a limited number of the observed variables; especially over land. Primarily due to representativity issues.

Data assimilation: in-situ observations

4

October 29, 2014

Scatterometer ocean low-level winds: ASCAT

Imagers: SSMI, SSMIS, AMSR-E, TMI

Ozone GPS radio occultations

Sounders: NOAA AMSU-A/B, HIRS, AIRS, IASI, MHS

Geostationary+MODIS: IR and AMV

Data assimilation: satellite data

5

October 29, 2014

IFS Cycle 41r2 (8 March 2016) • Grid resolution is doubled to 9 km for HRES and 18 km for ENS, and tripled to 18 km for EDA

• ENS forecasts also improved by moving the resolution truncation from 10 to 15 days (to 36 km)

• Significant increase in forecast accuracy and computational efficiency

• The tracks and in particular the intensity of TCs are now more accurate especially for the ENS due to the increased resolution

CRPS, 41r1 CRPS, 41r2

6

More details at: https://software.ecmwf.int/wiki/display/FCST/Horizontal+resolution+increase

October 29, 2014

Resolution upgrade: octahedral reduced Gaussian grid

Sa!1279.16

Re-think the spectral wave number truncation to gridpoint number ratio (Wedi 2014) The cubic-octahedral grid (TCo1279) at ECMWF (Wedi et al 2015) A new grid for the IFS (Malardel et al., ECMWF Newsletter 146)

8

October 29, 2014

Cycle 43r1 includes changes to the IFS components

• Scaling of convective mass fluxes for high resolution and change to mass flux limiter

• Modified coupling coefficients to reduce diurnal cycle T2m errors

• Limitation on the ocean wave spectral steepness for high winds

• Introduction of a global fix for tendency perturbations in SPPT to improve humidity conservation

• Increase in the resolution of EDA variance calculation

• Implementation of a slant-path radiative transfer for all clear-sky sounder radiances

• Explicit handling of correlated observation error for hyperspectral sounders

• Weak constraint formulation in the stratosphere (model levels 1-44)

The figure shows the 2mT bias reduction in 43r1 at 00 UTC (DJF).

9

October 29, 2014

Cycle 43r1 includes changes to the IFS components • The new ocean (NEMO version 3.4) configuration includes a higher vertical (75 instead of the 42

layers used in operation) and horizontal (1/4 instead of 1 degree) resolutions

• Introduction of a dynamical sea-ice model (LIM 2.0)

• ORAS5 ocean reanalysis

The figures show the reduction in RMSE of ensemble-mean forecast of sea-ice at t+10d between 43R1 (sea ice fraction predicted with NEMO/LIM2) and 41R2 (sea ice fraction fixed).

10

11

• WMO designated ECMWF as a World Meteorological Centre in May 2017 • ECMWF forecast products support WMO members including severe weather (SWFDP) and research • More HRES & ENS parameters made available as WMO Additional data

• ECMWF participates in 29 WMO committees, expert teams, working groups, task teams etc (e.g. Many on data and representation and formats; GRIB, BUFR etc)

• ECMWF supports WMO research (TIGGE, TIGGE-LAM,

S2S databases & Polar Prediction Project)

Collaborations

Data accessible by RA I and ACMAD

• RA I has access to ECMWF products classified as WMO Essential, WMO Additional. Some specific products are ACMAD-only

• Original ACMAD agreement made by Council at its 72nd session in December 2009.

• Products are transmitted via EUMETCast Meteorological Data Distribution (MDD) RA I (Africa). The data is also available via the GTS and WIS.

• All products are provided twice daily from 00 and 12 UTC base times. • Product resolution is 0.5 degree latitude/longitude in GRIB2 format using

JPEG compression. • The present volumes are ~150 MB after compression per day • Old 1x1 GRIB1 data will be discontinued soon: Anyone still using it? More details at: http://www.ecmwf.int/en/forecasts/datasets/public-wmo-and-acmad-datasets

Slide 12

Data accessible based on HRES

For all WMO Members (global distribution and ACMAD)

Essential (public) products underlined Slide 13

Parameter Level Domain Steps

GH 500 G 0 to 240 by 24

T, u, v 850 G 0 to 240 by 24

MSL surface G 0 to 240 by 24

u, v 700/500/200 G 0 to 168 by 24

Relative humidity 850/700 G 0 to 168 by 24

Divergence, vorticity 700 T 0 to 144 by 24

Significant wave height, mean wave period, mean wave direction

surface G 0 to 168 by 24

Data accessible based on HRES For ACMAD only G= Global; T= Tropics between 35S and 35N; A=4 octants centred on Africa; TA = Tropics between 35S and 35N and from 90W to 90E

Slide 14

Parameter Level Domain Steps

MSL surface A 0 to 168 by 6

u, v 925 A 0 to 168 by 24

Divergence 925/200 TA 0 to 144 by 24

2m T, 10m u, 10m v, total precipitation

surface A 0 to 168 by 6

Significant wave height, mean wave period, mean wave direction

surface G 0 to 168 by 6

Skin temperature, Soil temperature, Surface pressure, Total column water vapour

Surface A 0 to 168 by 6

New data based on HRES For all WMO Members (global distribution and ACMAD)

0.5 x 0.5 resolution

Slide 15

Parameter Level Domain Steps

GH 850/250 G 0 to 72 by 6 72 to 144 by 12

T 500/250 G 0 to 72 by 6 72 to 144 by 12

u, v 925 G 0 to 72 by 6 72 to 144 by 12

Divergence, vorticity 925/250 G 0 to 72 by 6 72 to 144 by 12

2m T, 10m u, 10m v, total precipitation

surface G 0 to 72 by 6 72 to 144 by 12

Peak wave period and Mean zero-upcrossing

surface G 0 to 72 by 6 72 to 144 by 12

Data accessible based on ENS

For all WMO Members (global distribution and ACMAD)

Slide 16

Parameter Level Domain Steps

ENS Mean and Standard deviation

GH 500 G 0 to 240 by 24

T, WS 850 G 0 to 240 by 24

MSL surface G 0 to 240 by 24

Probabilities

Total precipitation > 10/20mm G 00-24, 12-36, …, 120-144

10m wind gusts > 15/25 m/s G 00-24, 12-36, …, 120-144

Significant wave height > 2/4/6/8 m G 12 to 144 by 12

New data based on ENS

For all WMO Members (global distribution and ACMAD)

Slide 17

Parameter Level Domain Steps

ENS Mean and Standard deviation

WS 250 G 0 to 240 by 24

Probabilities

Total precipitation > 1/5/25/50/100mm G 00-24, 12-36, …, 216-240

10m wind speed and 10 m wind gusts

> 10/15/25 m/s > 10 m/s

G 00-24, 12-36, …, 216-240

Temperature anomalies at 850

±1, ±1.5, ±2 std dev

G 12 to 240 by 12

Data accessible based on ENS

For all WMO Members (global distribution and ACMAD)

Slide 18

Name Description Format Domain

ENSgrams Site specific forecast up to 10 days

Graphical only

G

Extreme Forecast Indices 0 to 144 by 12 Graphical only

G

Tropical Cyclone tracks and SST

For all WMO Members (global distribution and ACMAD)

Essential (public) products underlined

Slide 19

Name Level Domain Steps

Tropical Cyclone Trajectory (including genesis)

BUFR G up to 10 days

SST surface G 1 to 6 month

SWFDP contribution

• ECMWF supports WMO SWFDP • ECMWF contributes to six SWFDP Regional Projects:

– Eastern Africa – Southern Africa – South Pacific – Central Asia – Bay of Bengal – South East Asia

• Participants in these projects can access selected ECMWF forecast charts for their project

Slide 20

Other free datasets

• Multi-model real-time (TIGGE) • Sub-seasonal to seasonal (S2S) • Atmospheric composition (CAMS) • Reanalysis (ERA-Interim, ERA5)

http://apps.ecmwf.int/datasets/

• Global Floods (GLOFAS – JRC) http://www.globalfloods.eu/

Slide 21

Free datasets

TIGGE: multi-model real-time forecasts http://apps.ecmwf.int/datasets/data/tigge (delay of 48 hours after initial time)

The TIGGE dataset consists of ensemble forecast data from ten global NWP centres, starting from October 2006, which has been made available for scientific research. TIGGE was established as a key component of THORPEX to accelerate the improvements in the accuracy of 1-day to 2-week high impact weather forecasts for the benefit of humanity. THORPEX programme finished at the end of 2014. TIGGE will continue for a further 5 years.

October 29, 2014

Free datasets S2S (WWRP/THORPEX/ WCRP joint research project) http://wwww.s2sprediction.net

Bridging the gap between weather and climate

http://apps.ecmwf.int/datasets/data/s2s

(delay of 3 weeks after initial time)

MISSION: to improve forecast skill and understanding on the subseasonal to seasonal timescale, and promote its uptake by operational centres and exploitation by the applications community.

Specific attention will be paid to the risk of extreme weather, including tropical cyclones, droughts, floods, heat waves and the waxing and waning of monsoon precipitation.

23

Reduced-fee licenses for WMO NMHS

• New NMHS non-commercial licence for ECMWF web products and ecCharts: 3,500 euros

• Existing NMHS non-commercial licence for all ECMWF data: 42,000 euros

• Can be used to fulfil national governmental obligations related to the protection of life and property and for Research Projects and Education

Slide 24

October 29, 2014

Distribution of ECMWF real-time products to NMHSs of WMO

25

Full Access to real-time products Access to graphical real-time products

=

October 29, 2014

Contributing to the EU Copernicus Programme

26

2019 • 2018 2020 2021 2025 Strategy

End of current Copernicus Delegation Agreement (CAMS & C3S)

ECMWF 4YP

Climate Data Store infrastructure and toolbox Completion of ERA5 (1979 - near real time) and preparation of ERA6 during period C3S multi-model seasonal forecasts, Including ECMWF, Meteo France, MetOffice, DWD, CMCC (+NCEP &

JMA)

Metrics for atmospheric composition forecasts endorsed by WMO/GAW Monitoring then assimilation of Sentinel-5p composition data

Ongoing work on new operational global flash flood prediction system GloFAS releases a global hydrological seasonal outlook system based on SEAS5 Operational delivery of fire danger indices from its multi-model ensemble prediction fire forecast

systems with verification activities subcontracted to Meteo France..

CEMS

CAMS

C3S

T255L60 – t+120h T511L60 – t+120h

CAMS: higher-resolution suite

More details at: https://atmosphere.copernicus.eu/implementation-ifs-cycle-43r3cams

October 29, 2014

Next generation of atmospheric reanalysis: ERA5 Reanalysis is worth repeating as all ingredients continue to evolve • model, data assimilation, observation reprocessing • with each new reanalysis, understanding of model/observations biases is improved

ERA5 takes on board 10 years of research and development in NWP model and data assimilation methods: • Uncertainty estimate (EDA) • surface analysis (SEKF) • improved bias correction • improved model physics • higher resolution,… Improved observations: 34 observation data records either reprocessed or updated

ERA5

ModEl: IFS (CY41R2, Mar 2016) obsERvAtions: Conventional & satellite AssiMilAtion: Ensemble 4D-VAR (EDA) REsolution: 31km (T639L137), hourly unCERtAintY EstiMAtE: 62km, 3-hourly PERiod: 1979-present

A dedicated ERA5-land component will be produced at a higher resolution.

28

Reanalyses: ERA5 ERA-Interim ERA5

Start of production August 2006 IFS Cy31r2

January 2015 IFS Cy40r3

Model input As in operations (inconsistent SST)

Appropriate for climate (CMIP5, HadISST.2, OSTIA)

Spatial resolution 79 km global 60 levels to 10 Pa

39 km global (or 31 km) 91 levels to 1 Pa (or 137 levels)

Output frequency 6-hourly Analyses Hourly

Time period 1979 - present 1979 - present (extension to ~1950)

Dissemination Monthly Monthly for ERA5; daily for ERA5T

Observations Mostly ERA-40, GTS Various reprocessed CDRs

Radiative transfer RTTOV7 RTTOV11

Analysis method 4D-Var 1D+4DVar rain

10-member ensemble 4D-Var (EDA) All-sky MW

Variational bias corrections

Satellite radiances

Also ozone, aircraft, surface pressure, radiosondes

Operated by Reanalysis team Forecast Department

October 29, 2014 30

The charts catalogue

October 29, 2014

Web services: faceted search Category • Conventional data • Satellite • …

Parameter • Radiances • Ozone • 2m temperature • …

Data Type • Microwave radiances • Infrared radiances • …

Instrument • ATMS • AORS • … • AVRHH

31

October 29, 2014

The dashboard Allows forecasters to create their own working environment

Allows the caching of products

32

October 29, 2014

ecCharts

Web based application to explore and visualize ECMWF data

● Easy and immediate access to

charts ● Native data resolution ● Interactive features (zoom, pan,

click, extract data information, …) ● Apply different colour styles ● Customisable parameters ● Operationally supported, highly

available service

October 29, 2014

The interactive charts …

34

We have increased number of clickable charts to display point forecasts

To discover more go on the Charts catalogue …

October 29, 2014

New products in ecCharts ecCharts products are updated on a regular basis

35

Cloud ceiling

M-climate

2m mean temperature computed over user selected time periods

October 29, 2014

Severe convection: CAPESHEAR

CAPEWindShearCAPESHEAR LL *][ 2

1=

• CAPE is a key ingredient of deep moist convection • Large vertical wind shear favors organized convection • Super cells occur where strong shear is combined with

large instability • CAPE values of less than 10 J/kg are filtered out to

emphasize convection rather than anomalous but insignificant CAPESHEAR

• Limitations: CIN not taken into account

36

October 29, 2014

Simulated Satellite Data SSD, by capturing detailed cloud and/or humidity features, show the high-resolution ECMWF forecasts as a weather satellite would see it.

Central wavelength

Description

6.30 (μm) Water vapour (~300hPa)

7.36 (μm) Water vapour (~500hPa)

10.79 (μm) Clouds (surface)

37

Support for ECMWF software

Available at http://software.ecmwf.int/

• Aim is to improve support for external users – Keep track of issues in a central place – Spread knowledge throughout the Centre

• Gribex discontinued • New ecCodes provides BUFR

decoder and has replaced GRIB API

October 29, 2014

MIR New Interpolation package

Flexible and maintainable design – Configuration driven – Plugin based (users can extend) – Share data-structures with future IFS dynamical core (ATLAS)

Any-to-Any Grid algorithm

– There is always a default algorithm

Kernel based on linear Interpolation Operators – Enabling caching of operators – Linear Algebra backend support for GPU’s & Accelerator cards (Intel Phi)

39

Data visualization: Metview

http://software.ecmwf.int/metview)

October 29, 2014

Partnership and Collaborations OpenIFS facilitates scientific research and teaching collaborations with NHMSs, academia and research institutions. Benefits to the Centre are manifold: it enhances ECMWF’s outreach and external collaborations and potentially leverages this substantial external effort to deliver IFS improvements.

Licensees: 55 currently, 44 in 2015, 21 in 2014.

41

https://software.ecmwf.int/wiki/display/OIFS/OpenIFS+Home

October 29, 2014

Outreach and training Training Catalogue • Computing • Meteorology • Software packages and applications

http://www.ecmwf.int/en/learning/training/training-catalogue Research annual seminar (Livestream)

11th to15th September 2017

Ensemble predictions: past, present and future

http://www.ecmwf.int/en/learning/workshops/annual-seminar-2017

Workshops A full list on the Research and technical topics http://www.ecmwf.int/en/learning/workshops-and-seminars

42

NEW

Training and Outreach: Courses and eLearning

43

Participants’ feedback on ECMWF training courses “Practical course …… we really like it!”

“Wunderbar! Remek! Traumhaft!”

“Every meteorologist should come to this course …”

eLearning modules expected to become available soon ecCodes for GRIB decoding; ecFlow; Convection part 1: Overview; Convection part 2: Mass flux approach and the IFS scheme; Introduction to Ensemble Forecasting; Introduction to MARS; Introduction to Metview; Introduction to parametrization; Learning how to cope with forecast jumpiness; Extended range forecasts; Single Column Model

Courses offered: advanced training in the field of numerical weather forecasting and in the use of ECMWF products and software eLearning modules are part of the blended courses Total attendance: 344 (79% from MS/Cs, 12% ECMWF)

EUROPEAN CENTRE FOR MEDIUM-RANGE WEATHER FORECASTS

https://www.ecmwf.int/en/en/learning/education-material/elearning-online-resources

October 29, 2014

Knowledge exchange

https://software.ecmwf.int/wiki/display/FCST/Forecast+User+Home

Forecast User Home

44

Mailing list to inform all WMO NMHS regularly used

Can also subscribe to forecast_changes@lists.ecmwf.int

October 29, 2014

Severe event catalogue

https://software.ecmwf.int/wiki/display/FCST/Severe+Event+Catalogue

45

October 29, 2014

BOND Programme – Provisional key dates

46

2017 2018 2019 2020

Data Centre design

complete

Data Centre handover to

ECMWF

HPC installation

Parallel run

HPC operational in

Bologna

Data Centre construction

starts

Nov ‘17 Q2 ‘18 Mid ‘19 End of ‘19 Q1 ‘20 Oct ‘20

Main contractor tender

Italian Senate ratified ECMWF agreement

October 29, 2014

ECMWF 2016-2025 strategy: overview

ECMWF’s purpose is to develop a capability for medium-range weather forecasting and to provide such weather forecasts to the Member and Co-operating States ECMWF is complementary to the National Meteorological Services and works with them in research, numerical weather predictions, supercomputing and training. Focus on high-impact weather, regime transitions and global-scale anomalies Integrated ensemble at high resolution at 5km by 2025 Earth-System model and analysis Scalable computation Environmental information services: Copernicus

47

October 29, 2014

ECMWF 2016-2025 strategy: goals for 2025

Forecast targets by 2025: Ensemble predictions of high impact weather up to two weeks ahead Seamless approach, aiming towards predictions of large scale patterns and regime transitions up to four weeks

ahead and global-scale anomalies up to a year ahead

Research goals by 2025: Research at frontiers of knowledge Ensemble-based analyses and predictions that raise the international bar for quality and operational reliability

reaching a 5 km horizontal resolution

Together - More collaboration: Partnering with universities and research institutes – OpenIFS Pooling expertise to improve scalability of data assimilation Continued support: Dedicated HPC, software, and data resources for Member States Advanced training

48