Description of the upgrade

The operational release of AIFS ENS v2 will introduce scientific and architectural changes to the AIFS ensemble model.

New and revised output variables include:

•  The introduction of a new wave component, including 11 wave variables, marking ECMWF's first operational data-driven wave forecasts.
• The addition of a new snow variable to the existing land component.
• The addition of 2 variables already present in AIFS Single (vsw and cp), to harmonise the two models.
• An improved representation of the stratosphere, with the addition of pressure level fields at 10hPa.
• Changes to the priority  of real-time files transmitted via ECMWF's dissemination service (ECPDS).

Architectural changes include:

• Replacing the afCRPS loss function used in AIFS ENS v1 with a multi-scale loss. See https://arxiv.org/pdf/2506.10868.
• Imposing the same variable bounding as is used by the AIFS Single model, improving physical consistency.
• Revised graph features, utilising more edges in the decoder.
• An improved training regime, using 1 years of additional training data compared to AIFS ENS v1.

This upgrade also includes a technical fix:

• In AIFS ENS v1.0, some parameters at step 0 have incorrect encoding of the ecCodes key productDefinitionTemplateNumber, which is used to describe the type of data. The correct encoding will be applied to all AIFS ENS output products with this upgrade
  
  
  • See https://codes.ecmwf.int/grib/format/grib2/ctables/4/0/ for a description of these template numbers.

Description of the model

The upgrade of AIFS ENS to version 2 introduces changes to the training regime and model architecture. 

The full training procedure for AIFS ENS v2 consists of:

• Pre-training, which remains the same as for AIFS ENS v1
  • Pre-training is performed on ERA5 data covering the years 1979–2022 over 300,000 training steps.
• Fine-tuning, which has been updated for AIFS ENS v.2 
  • Fine-tuning is performed on  years of additional data compared to AIFS ENS v1.0. In particular, fine-tuning is performed on ECMWF operational analysis data and available IFS 50r1 esuite analysis data, covering the years 2018-2024 over 7,900 training steps.
    
  • As in previous versions of AIFS ENS, IFS fields are interpolated from their native O1280 resolution (approximately 0.1°) using MARS default interpolation tools down to N320 (approximately 0.25°) for fine-tuning and initialisation of the model during inference. 

For both phases of training, a cosine learning rate schedule and a batch size of 16 is used. 

Further model changes include introducing a multi-scale loss and removing reference field truncation, imposing the same variable bounds used by the AIFS Single model to improve physical consistency, and revising graph features by using more edges in the decoder together with new edge features.

AIFS ENS v2 is implemented using the Anemoi framework. Anemoi consists of open source components, enabling the user community to run AIFS themselves.

News

Timeline of the implementation 

Datasets affected

This upgrade impacts:

• Set X - AIFS ENS
• Set XIII - AIFS ENS wave (new)

New wave stream

AIFS ENS v2 introduces a new stream for wave products. This new stream will be configured as follows:

• MARS and dissemination indicator: stream=waef (see Keyconfigurationvalues )
• Dissemination schedule: products will be disseminated according to the existing IFS ENS-WAM dissemination schedule
• Dissemination stream (ECPDS): Dissemination products will be grouped under XX-AIFS-WAEF on ECPDS, where XX is the base time (00, 06, 12, 18)
• Dissemination file name: products will be named according to the ECMWF file naming convention by default. For example

abc_a1_aifs-ens_ai_waef _pf_20250725T000000Z_20250808T180000Z_354h

• Dissemination priority: products will be disseminated according to the existing IFS HRES-WAM priority (priority=50)

Changes to dissemination priority

The priority of real-time AIFS files transmitted via ECMWF's dissemination service (ECPDS), will be changed on 12 May 2026.

Resolution

AIFS ENS v2 introduces a new pressure level to the stratospheric component.

Meteorological content 

Initial conditions

• AIFS ENS v2 uses the operational IFS control initial condition, regridded to an N320 grid.

Observations

• No observations are used to train AIFS ENS v2.

Meteorological impact

Coming soon!

Evaluation 

Coming soon!

Known issues

Coming soon!

Key configuration values

*Levels (hPa):  10 (new) , 50, 100, 150, 200, 250, 300, 400, 500, 600, 700, 850, 925, 1000

Technical content

GRIB encoding

The GRIB model process identification numbers for AIFS ENS v2 will be as follows:

grib_get -p generatingProcessIdentifier <file.grib2>

Software 

To handle the data of AIFS ENS v2 we recommend use of the ECMWF software package ecCodes 2.46.0 

On the ATOS HPC, this corresponds to  ecmwf-toolbox/2026.02.0.0

Availability of AIFS ENS v2 test data

Test data in MARS 

AIFS Single v2 test data is available from MARS with E-suite experiment version (expver) 105 (MARS keywords EXPVER=105, CLASS=AI): 

• Stream=enfo
• Stream=waef (new)

Users should only review data from 14 April 06z run onwards. Data prior to this date may not be representative.

Test data in dissemination

AIFS ENS v2 test data   is now available through the test dissemination system.

Users with access to ECPDS and the Products Requirements Editor (PREd) can login to the test systems here: 

• Test ECPDS:    https://xdiss-monitor.ecmwf.int/
• Test PREd (TPREd):    https://products.ecmwf.int/esuite/requirements/

To receive the test products, please ensure your firewalls are open to the relevant ECPDS Data Movers:

INTERNET:

• • 136.156.192.0/26
  • 136.156.193.0/26

RMDCN:

• • 136.156.196.0/26
  • 136.156.197.0/26

The AIFS ENS v2 test products are available as version number 105 (file names including '*0105' as expver) and are intended to be generated shortly behind near real-time. 

Please note that following changes are automatically applied to dissemination requirements in the test system :

• Changed expver=0001 to expver=105 (for Release Candidate Phase only)

Users can trigger the transmission of test products in the usual manner.

To receive AIFS ENS v2 test data automatically (regularly), please contact Data Support via   the Support Portal . 

If you don't have access to the ECPDS system or (T)PREd, or require any assistance with AIFS ENS v2 test dissemination products, please contact us via the Support Portal .

Test data on ECMWF's Open Data platform

AIFS Single v2 data is freely available via ECMWF's open data platform.

Test data, including new parameters introduced with this upgrade, can be accessed at the following link:  https://data.ecmwf.int/forecasts/testdata/

Test products on ecCharts 

Meteograms on Open Charts

Resources

Webinars

<iframe width="390" height="230" src="https://www.youtube.com/embed/XQawZQish84?si=XNISBWQbAe0tOKRI" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

<iframe width="390" height="230" src="https://www.youtube.com/embed/RMN1O0nLbbE?si=EEwQTYGNchqWzANO" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

References

• Let it snow! How machine learning will forecast snow in the AIFS, Raoult, N., Pinnington, E., Arduini, G., Rudiger, C., De Rosnay, P., Chantry M.

• Representing ocean wind waves in ECMWF's AIFS, Hahner, S., Bidlot, J., Kousal, J., Zampieri, L., Lessig, C., Chantry, M.
• Simon Lang, Martin Leutbecher, Pedro Maciel, “A multi-scale loss formulation for learning a probabilistic model with proper score optimisation” (2025), arXiv:2506.10868.