BB: Background. Digital Twin prototyping platform, supporting the integration of Earth-system modelling, Earth observation and new sensors (drones, buoys, etc.).

• Define uncertainty of models and predictions. Tricky 

BB: I wrote the concept note from the perspective of what an end user would like to see

• Goal: easy to prototype Digital Twins
• Whole cascade: Numerical modeling to data simulation 
• Applied to different scales: local, regional, and global.

BB: One of the applications would be Observing System Experiments

• How much an observation impact a simulation
• As a user I go to the platform, put a bunch of observations in, and test the predictions

KK: What do you expect to be implemented? Jupyter notebooks? or something that we develop?

• BB: We need to think about the interfaces. Notebooks could be but it's more about being able to run numerical models on different HPCs, accessing the data that I need, etc.
• BB: numerical model + data assimilation scheme + novel observations = what is the prediction? The goal would be to streamline the development 
• KK: HPC systems work in a very different way, policies, peer-reviews etc. they are more protected and make things tricky. HPC procurement is coming up. GRNET is trying to ease the processes for moving data etc.
• RO: Cross-HPC systems could be used, upscaling and downscaling for e.g. post-processing. We would need clearer requirements for the platform. 
• BB: harmonised access to computing resources (cloud, htc, hpc); added value here is integrate numerical models with innovative in-situ observations
• KK: Interoperability with Data Spaces / Lakes should be a bullet for sure
• MD: we might have use cases from Portugal
• MS: Re: about infrastructure perspective - is there a realistic chance to connect our infrastructure with DestinE infrastructure? (DestinE feels a closed society). For example, can we connect EODC to the DestinE data lake?
  • CC: EODC is part of the DestinE data lake along with CloudFerro
• BB: Bring your own data / data-bridge functionality; discussion ongoing at interTwin with ECMWF
  • RO: can we process anything via this bring your own data in our infras? 
  • CC: ECMWF is happy to stay in the loop, but already participating in other two proposals for the same call
• MS: In the call they didn't include CDSE, and VITO is there. We should plan to use their data or connect with them somehow.
• MS: Several Austrian universities (under ACONET) are organising a joint use case for this call.
• CC: From interTwin (University of Trento + CMCC) are asking whether we are preparing something together for this call.

TZ: ENES would be an additional use case.

TZ: Please consider the integration modes from the text. The intended integration can be performed at different levels, for example: (i) directly integrated in the existing Destination Earth simulation and data fusion system (i.e. full integration mode), (ii) integrated in a sequential workflow where new digital twins operate their own simulation and data fusion tasks interfacing with the existing Destination Earth digital twins (i.e. coupling mode) or (iii) as data post-processing applications without own Earth-system component simulation tasks (i.e. post-processing mode).

RO: Focus on a couple of domains/sub-domains and find partners that we would be interested in collaborating. We explain what we would like to do and ask for user stories.

TZ: Had suggested CB to contact JRC (Guido) about the use cases that are in the pipeline. I can also discuss with Guido. 

SLV: Benjamin shared with me this week the following. Relevant? https://publications.jrc.ec.europa.eu/repository/handle/JRC122456. But this an old publication (2020). There is a new one with 9 use cases. Charis Chatzikyriakou will find and share it.