EGU23-6122, updated on 27 Jun 2023
https://doi.org/10.5194/egusphere-egu23-6122
EGU General Assembly 2023
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.

Destination Earth On-Demand Extremes Digital Twin

Roger Randriamampianina1 and the On-Demand Extremes Digital Twin Team*
Roger Randriamampianina and the On-Demand Extremes Digital Twin Team
  • 1Norwegian Meteorological Institute, Development Centre for weather forecasting, Norway (rogerr@met.no)
  • *A full list of authors appears at the end of the abstract

Destination Earth (DE) On-Demand Extremes Digital Twin (DT) (DE_330_MF) offers configurable digital twin engines for forecasting environmental extremes at a sub-km scale. DE_330_MF aims at providing an on-demand workflow with co-design of high resolution predictions about extreme weather events combined with decision making support for impact sectors including hydrology, air quality and energy meteorology, with the use of a physics-based and data driven model system and computationally-intensive data-flow organised on the EuroHPC high performance computing platform. 

The DE_330_MF core software system is developed in four parts. Firstly, code adaptation needed for running on hybrid CPU and accelerator architectures will be done. Here, our most computationally intensive parts will be optimised for running on accelerators at the end of the first phase. Secondly, our modelling system is being adapted for running on a sub-km grid scale in an efficient on-demand setup. During phase 1 we primarily use operationally and pre-operationally mature components. Tailor-made modules targeted on impact modelling (e.g. wind energy) will be implemented as well. Thirdly, we work on post-processing of the model output, so that this can be distilled into a limited number of variables essential for end-users including uncertainty quantification for these variables. The approaches will be tailor-made for the specific targeted types of extreme events. Triggers for detecting extreme events are being developed as part of the post-processing work. Fourthly, and very importantly, we are designing and developing workflow management tools and scripting for the core software. Input and output data are also managed, with a particular focus on the exploitation of using high-density observations in high-resolution weather forecasting and verification. Overall we aim at running a hectometric scale On-Demand Extremes DT with a typical resolution of 750 or 500 m, exploring a finer resolution of up to 200 m for special applications where it is meaningful. In the first phase, capability demonstration is the focus by examining a range of carefully selected high-impact cases for the three above-mentioned impact areas, as well as other applications such as agriculture. The goal here is to explore the added values with co-designing workflow combining the on-demand, event- or user-driven solutions with weather forecasting on the one hand, and impact sectors on the other hand, in an integrated value chain for decision-making support. In this effort, containerised solutions, including one hydrological and two air quality models, will run next (one way coupling) to the on-demand extremes DT.

This work is funded by the EU under agreement DE_330_MF between ECMWF and Météo-France. The on-demand capability proposed by the Météo-France led international partnership is a key component of the weather-induced extremes digital twin, which ECMWF will deliver in the first phase of Destination Earth, launched by the EC.

On-Demand Extremes Digital Twin Team:

All experts involved in the Destination Earth On-Demand Extremes Digital Twin (DE_330_MF)

How to cite: Randriamampianina, R. and the On-Demand Extremes Digital Twin Team: Destination Earth On-Demand Extremes Digital Twin, EGU General Assembly 2023, Vienna, Austria, 23–28 Apr 2023, EGU23-6122, https://doi.org/10.5194/egusphere-egu23-6122, 2023.