C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and...

45
Climate Change Seasonal Forecasts – Météo-France M-F: Sophie Marnoni-Lapierre, Lucas Grigis, Chrisan Viel, Ludovic Bouilloud and Raphaël Legrand + Alberto Troccoli (WEMC), Laurent Dubus (EDF), Linh Ho (WEMC), Yves- Marie Saint-Drenan (ARMINES). C3S Energy,webinar on seasonal climate and energy indicators

Transcript of C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and...

Page 1: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

Climate Change

Seasonal Forecasts – Météo-France

M-F: Sophie Martinoni-Lapierre, Lucas Grigis, Christian Viel, Ludovic Bouilloud and Raphaël Legrand+ Alberto Troccoli (WEMC), Laurent Dubus (EDF), Linh Ho (WEMC), Yves-Marie Saint-Drenan (ARMINES).

C 3 S E n e r g y , w e b i n a r o n s e a s o n a l c l i m a t e a n d e n e r g y i n d i c a t o r s

Page 2: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

OUTLINES

123

Basics on seasonal forecasts (5mins)

Seasonal operational stream (10mins)

4 Indicators assessment (15mins)

Seasonal energy indicator production (15mins)

5 Demonstration and conclusions (5mins)

Page 3: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts are climate data

Some important specificities of Long Range Forecast :

- forecasting of Climate Variability, not weather time-integration/statistics of climate variables, probability of scenarios

- predictability comes from the slow components of the Climate System (oceans, ground, sea ice...) through boundary conditions concept of “forcing”

Page 4: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Potentialities and limits in Europe

ENSO (El Niño – La Niña)

Stratosphere

SST*

Cryosphere

Main forcings on Atlantic and Europe at seasonal scale

Forcings: components of the climate system that evolve slowly and could influence atmospheric conditions

The predicability in Europe :✔ Is still modest (little added value vs climatology), mainly available on

large scale output and after time-integration/statistics✔ Can be very different from one place to another, from one year to

another (depending on the strengh of the forcings and on the type of climate)

The European climate is partly influenced by these forcings, but mainly chaotic.

(*SST: “Sea Surface Temperature”)

Page 5: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

The current landscape of operational Seasonal Forecast

A robust and well-organised operational production

http://www.wmo.int/pages/prog/wcp/wcasp/gpc/gpc.php http://www.ecmwf.int/en/forecasts/charts/seasonal/ http://iri.columbia.edu

… supported by active research projects

Page 6: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Copernicus Climate Change Service (C3S)

Agriculture Energie Transport Tourism Health Insurance Infrastructure Coastal areaWater Management

Figure and comments taken from https://climate.copernicus.eu/sites/default/files/2019-03/13.30_02_A%20reflection%20on%20climate%20services_Marta%20Bruno%20Soares.pdf

The role of climate services:● “Non linearity”: Facilitate

the access to huge amount of climate dataset (e.g. homogeneity and SI infra.)

● “Usability”: Apply user’s knowledge and expertise on climate data.

● “R&D”: Enhance applied science and services

Page 7: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

OUTLINES

123

Basics on seasonal forecasts (5mins)

Seasonal operational stream (10mins)

4 Indicators assessment (15mins)

Seasonal energy indicator production (15mins)

5 Demonstration and conclusions (5mins)

Page 8: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecast for the energy sector: C3S SIS Energy overview

C3S SIS ENERGY

HISTORICAL

SEASONAL

PROJECTIONS

Operationalservice

Page 9: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecast for the energy sector: C3S SIS Energy overview

C3S SIS ENERGY

Operationalservice

Data available online: open data policy

Page 10: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

The science

Large ensembles: each system provides at least 50 ensemble members for the real-time forecast and at least 25 ensemble members per year for re-forecasts Diversity : ECMWF, Météo-France, Met Office, CMCC, DWD

Accessibility of data and products (open data policy)

✔ Graphical and digital products available through the Climate Data Store https://climate.copernicus.eu/charts/c3s_seasonal/

✔ Monthly updates, published on each 13th, available for operational schedule

✔ Sectoral seasonal forecast products are being produced in Sectoral Information System Services

C3S’s ambitions on Seasonal Forecast

Page 11: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

CDSBias

Correction Energy Indicators Public VM

Energy Data

One-shot data retrieval:(soon from CDS API):1. ECMWF SEAS5 Hindcasts:

• Monthly from January 1993 to decembre 2016

• 1° horizontal resolution; gridded values.

• Six-hourly T2m and 10m-wind• Daily GHI and TP

2. ERA5 reanalysis (same parameters, period and resolution).

Real time correction on climate indicatorsforecasts:• Quantile mapping

scripts to mapshindcast/ forecastsaccording to ERA5 distribution within a ±15days climaticwindow

Processed data and output validated and

moved to public directory on VM for

transfer and inclusion into CDS

Near real time data retrieval:(soon from CDS API)ECMWF SEAS5 Forecasts:

• Monthly from January 2017• 1° horizontal resolution;

gridded values.• Six-hourly T2m and 10m-wind• Daily GHI and TP

Metadata necessary for energy indicatorscomputations:• Demand (Nuts0), solar (gridded) and

hydropower (Nuts0) models• Wind power curves for Onshore and offshore

wind production.

Seasonal Forecast

Energy Indicators indicatorshindcast/forecasts):[DMP compliant file naming and format]

• 1° gridded [Netcdf files]:SPV : daily averaged CFWON and WOF : 6-hourly releasedinstantaneous CF• Nuts0 aggregated [CSV files]:EDM : daily valueHRO and HREf : daily value

Climate Indicators

NUTS0Aggregation

Python scripts using Eurostat NUTS0 (country scale) shapefiles. Processing script run in no-hangupmode on VM, using Python with: iris, geopandas, cartopy, pandas, cartography, os, time.

Climate indicators forecast:[Netcdf format; DMP compliant file naming and format]• 1° horizontal resolution gridded values• 6hourly T2m and 10m-wind• Daily GHI and TP

AggregatedClimate Indicators

CSV format and DMP compliant file naming and format)

Skillassessment

« Health check »: File checking through basic control (size, missing characters, etc.) and data control through computation of biais, anomaly

coefficient correlation and ROC score.

Seasonal stream: workflow

Page 12: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts models used for the indicators

ECMWF : SEAS5✔ 51 forecasts and 25 hindcasts (re-forecasts) members✔ 51 members initialized on the 1st of the month

Météo-France : System 6✔ 51 forecasts and 25 hindcasts members✔ 51 members : 1 initialization on the 1st, 25 on the 25th and another 25

members on the 20th (lagged initialization)

End of summer

2019

Page 13: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts models used for the indicators

ECMWF : SEAS5✔ 51 forecasts and 25 hindcasts (re-forecasts) members✔ 51 members initialized on the 1st of the month

Météo-France : System 6✔ 51 forecasts and 25 hindcasts members✔ 51 members : 1 initialization on the 1st, 25 on the 25th and another 25

members on the 20th (lagged initialization)

MetOffice : GloSEA5✔ 2 members initialized each day (lagged initialization), ~62 members

by month✔ Rolling release model : both hindcasts and forecast are computed

“on-the-fly” every month

End of summer

2019

Autumn 2019

Page 14: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Forecasts production

Forecast production process Object oriented python3 ; Only standard and Free-libre and

open-source libraries ; 9 classes, one parameters and one

driver/launcher file ; Custom methods for quality

assesment, country aggregation, etc…

Forecast treatment is light and fast ! ~ 500 (250) Mio for both raw and bias adjusted hindcast file by

(cumulated) variable and by month + 40 (20) Mio terciles proba ; Less than 5 Mio for all CA csv’s ; Time of compilation ~ 45min for gridded variables (+ 1h for CA var).

Page 15: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

OUTLINES

123

Basics on seasonal forecasts (5mins)

Seasonal operational stream (10mins)

4 Indicators assessment (15mins)

Seasonal energy indicator production (15mins)

5 Demonstration and conclusions (5mins)

Page 16: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Indicators dedicated to demand and supply balance

Market

TSODSO

Production

Consumption

Indicators you need

Page 17: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Climate indicators

2m air temperature

TA-

Total Precipitations

TP-

Global Horizontal Irradiance

10m Wind Speed

WS-

● 24-hourly● Daily amounts● 1° gridded and NUTS0 aggregation

● six-hourly● Instant values● 1° gridded and NUTS0 aggregation

● 24-hourly● Daily amounts● 1° gridded and NUTS0 aggregation

● six-hourly● Instant values● 1° gridded and NUTS0 aggregation

Page 18: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Energy indicator: same model for all time frame

Electricity demand

On&offshorewind production

Photovoltaic production

Run-of-River &Reservoir

production

● Capacity factor of aggregated PV prod. installed in each pixel.● Computed from GHI and T2m following Saint-Drenan et al. (2018)● Gridded values

Saint-Drenan et al., 2018: https://doi.org/10.5194/asr-15-51-2018

● Production of widely used turbines: vestas V136/3450 (3.45MW) for onshore and a Vestas V164/8000 (8.MW) for offshore

● Computed from 10m wind raised up to 100m● Gridded values

● From T, GHI, 10m-WS, Calendar, estimation of the detrended load; trend estimated on 2010-2018.

● Stat. model (GAM models); reference ENTSO-E data.● NUTS0 aggregation

● Hydropower generation using lag times of total precipitation, temperature and snow depth.

● Stat. model (Random forest); reference ENTSO-E data.● NUTS0 aggregation

More info was given on C3S Webinar the 22 May 2019https://climate.copernicus.eu/webinars

Page 19: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts models used for the indicators

End

of

sum

mer

201

9A

utum

n 20

19

More informations e.g here:https://confluence.ecmwf.int/display/COPSRV/Description+of+the+C3S+seasonal+multi-system

● 51 forecasts and 25 hindcasts (re-forecasts) members. All members initialized on the 1st of the month

● Ocean analysis : OCEAN5 (5-members analysis)● Ocean-Atmosphere coupling: 1-hourly of Nemo v3.4 (0.25°

ORCA grid) and IFS v43r1 (~36km)

● 51 forecasts and 25 hindcasts members, 1 initialization on the 1st, 25 on the 25th and another 25 members on the 20th (lagged initialization)

● Ocean : NEMO v3.6 (0.7° ORCA grid)● Atmosphere : ARPEGE-climat v6.2.5 (0.5° ; ~50km) and SURFEX v8.1● 3-hourly with OASIS v3.0

● 2 members initialized each day (lagged initialization)● Rolling release model : both hindcasts and forecast are computed

“on-the-fly” every month● Ocean : NEMO (0.25° ORCA grid)● Atmosphere : UM Met Office Unified Model (0.83°x0.56° ; ~60km in mid

latitude) and JULES v3.0 (surface reanalisys) ● Ocean-atmosphere coupling : 3-hourly

Page 20: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal dataset

Variable Timescale Source

Seasonal 1 day 1 deg Country

Precipitation «TP-» Seasonal 1 day 1 deg Country

Seasonal 6 hours 1 deg Country

Seasonal 1 day 1 deg Country

ENERGY INDICATORS

Seasonal 1 day / Country

Seasonal 6 hours 1 deg Country

Seasonal 1 day 1 deg Country

Hydro Power «HRE» Seasonal 1 day / Country

Highest Temporal

Resolution

Highest Spatial

Resolution

Spatial Aggregation

CLIMATE INDICATORS

Temperature «TA-»EC, MF,

MOEC, MF,

MOWind (10 and 100m)

«WS-»EC, MF,

MOSolar Radiation at

surface «GHI»EC, MF,

MO

Electricity Demand « EDM »

EC, MF, MO

Wind Power «WON» and «WOF»

EC, MF, MO

Solar Power «SPV»EC, MF,

MOEC, MF,

MO

Tasks planned for “Fall 2019”

Climate Indicators

Energy Indicators

Page 21: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

OUTLINES

123

Basics on seasonal forecasts (5mins)

Seasonal operational stream (10mins)

4 Indicators assessment (15mins)

Seasonal energy indicator production (15mins)

5 Demonstration and conclusions (5mins)

Page 22: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts hindcasts and climatology period

Hindcasts : up-to-date SF model running on past period✔ Build a SF model climatology ;✔ Allow bias adjustment ;✔ Give access to statistics (anomalies and probabilities).

Reference data✔ Reference data is ECMWF ERA5 at 1° horizontal resolution ;✔ Climatology period : 1993-2016 (1993 is the beginning of satellite

altimetry data assimilation) ;✔ Same (1993-2016) 24 years hindcast period.

Page 23: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Bias adjustment: Quantile mapping

Bias adjustment via quantile mapping : simple and efficient✔ Ranking the raw values in their climatic distribution (centiles) ;✔ Taking the correspondent value in the ERA5 reference distribution ; ✔ Linear interpolation between the centiles ;✔ Linear extrapolation of the extreme values ;✔ The possible extreme forecast values are taken from ERA5 extremes

values in the same climatic window.

Page 24: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Assessment on climate variables: starting point

TA-+ 30 d + 180 dLocation:Berlin

ERA5 and raw and bias adjusted ECMWF SY05 Hindcasts distributions within a +/- 15d climatic window

Both the two distributions are similar, QM went ok!

Page 25: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: bias adjustment map examples

Page 26: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: bias adjustment map examples

Page 27: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: bias adjustment map examples

Page 28: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: bias adjustment map examples

North Sea

Budapest

Page 29: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: bias adjustment location examples

North Sea

Bias Corrected (BC) and Raw hindcasts

ECMWF SY05

Reduced bias on Sea, more inertial = more predictable

Mean Error

RMSE

Page 30: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: bias adjustment location examples

Budapest

BC and Raw hindcasts

ECMWF SY05

Higher bias onshore, more inerannual variability = less predictable

Mean Error

RMSE

Page 31: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: bias adjustment location examples

BC and Raw hindcasts ECMWF SY05

Higher bias offshore than onshore: higher mean wind speed = higher mean error

Mean Error

Mean Error

Page 32: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: skills

Skill scores will come along the forecasts outcome ✔ Computed on the hindcasts vs ERA5✔ Chosen scores are :

✔ Interannual correlation coefficient (ensemble mean skill)✔ ROC diagram (distribution skill)

Page 33: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: skills

Anomaly correlation coefficient Determine wether our forecast is correlated with observation.

MTFR

No surprise, poor skills on a step-by-step basis.

Page 34: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: skills

Anomaly correlation coefficient Determine wether our forecast is correlated with observation.

MTFR

High skills in the Tropics.

Poor skills at mid latitudes.

Scores are different from one season to the other.

Page 35: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: skills

Anomaly correlation coefficient Determine wether our forecast is correlated with observation.

ECMWF

Similar skills patterns

Page 36: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Hindcasts production

Hindcasts production process Object oriented python 5 classes, one parameters and one driver/launcher files. Custom methods for quality assesment, country aggregation,

etc…

Hindcast bias removal is the heavyest task 9 (4) Gio for both raw and bias adjusted hindcast file by

(cumulated) variable and by month ; Time of compilation ~12h per month per variable ; 24*25 members to country aggregate : 1 hour by month by

variable.

Fortunately, done once for all (except for UKMO GloSEA5…)Note that heavy, possible and done with 64 Gio of RAM…

Page 37: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

OUTLINES

123

Basics on seasonal forecasts (5mins)

Seasonal operational stream (10mins)

4 Indicators assessment (15mins)

Seasonal energy indicator production (15mins)

5 Demonstration and conclusions (5mins)

Page 38: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: climate output

Page 39: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal energy output

Photovoltaic productionmean anomaly

[init June, J.A.S. 2019]

Offshore wind productionmean anomaly

[init June, J.A.S. 2019]

Onshore wind productionmean anomaly

[init June, J.A.S. 2019]

Page 40: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Conclusions● C3S-SIS-Energy provide an operational service of production of climate and

energy indicator. Meteorology reference taken: ERA5.● All outputs are bias corrected with reference given by ERA5● 3 time periods covered in the project:

● Climate & Energy indicators addressing the demand and supply balance:

Electricity demand

Wind powerproduction

Hydroproduction

Daily, NUTS0

photovoltaicproduction

Daily, NUTS0 6h, 1° & NUTS0 6h, 1° & NUTS0

Past Real time Future

Historical(1973-2016)

Climate projection(up to 2100)

Seasonal forecasts(→ 7months, each month)

Seasonal:

Wh

ole

da

tas

et

fre

ely

av

ail

ab

le o

n

the

Cli

ma

te D

ata

Sto

re (

CD

S)

htt

ps:

//cd

s.cl

imat

e.co

per

nic

us.

eu/

We

bin

ar

22

/05

/19

We

bin

ar

05

/07

/19

We

bin

ar

12

/09

/19

Page 41: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Perspectives

Short term:a) Set operational the real-time service and make available all the data in the CDSb) Extend to the UKMO GloSEA5

c) Release the whole documentation, user guide and demonstrator.

Longer term depending on user’s needs for future project:

d) Possibility to extend to larger domain over the whole globe

e) Address other needs for the demand and supply balance as gaz consumption, maximum ampacity, other solar productions…

f) Test other statistical models to make the bias adjustment

g) More use cases and teaching tools

h) Others… ?

Page 42: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

Climate Change

WEBINAR ON SEASONAL FORECASTS

Sophie Martinoni-Lapierre, Lucas Grigis, Christian Viel, Ludovic Bouilloud, Raphaël Legrand (Météo-France), Alberto Troccoli (WEMC), Laurent Dubus (EDF), Linh Ho (WEMC), Yves-Marie Saint-Drenan (ARMINES).

C 3 S S I S E n e r g y

ENDNeed more details on [email protected] or [email protected]

Page 43: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Extra slides

ECMWF : SEAS5✔ Ocean analysis : OCEAN5 (5-members analysis)✔ Ocean-Atmosphere coupling: 1-hourly of Nemo v3.4 (0.25° ORCA

grid) and IFS v43r1 (~36km)

Météo-France : System 6✔ Ocean : NEMO v3.6 (0.7° ORCA grid)✔ Atmosphere : ARPEGE-climat v6.2.5 (0.5°) and SURFEX v8.1✔ 3-hourly with OASIS v3.0

UK Met Office : GloSEA 5✔ Ocean : NEMO (0.25° ORCA grid)✔ Atmosphere : UM Met Office Unified Model (0.83°x0.56°) and

JULES v3.0 (surface reanalisys) ✔ Ocean-atmosphere coupling : 3-hourly

Page 44: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: skills

Anomaly correlation coefficient Determine wether our forecast is correlated with observation.

MTFR

A large part of Europe exibits low skills.

Get better in Easter Mediterranée.

Page 45: C3S Energy,webinar on seasonal climate and energy indicators · CMCC, DWD Accessibility of data and products (open data policy) ... ORCA grid) and IFS v43r1 (~36km) 51 forecasts and

ClimateChange

Seasonal forecasts: skills

Anomaly correlation coefficient Determine wether our forecast is correlated with observation.

ECMWF

Often better offshore than onshore