ECMWF, as one of the Entrusted Entities for the Destination Earth (DestinE) initiative, invites Tenders for the purpose of demonstrating that Machine Learning/ Deep Learning (ML/DL) based methodologies can augment DestinE datasets and products from the Weather-induced and geophysical extremes Digital Twins (DT).

Under this Tender, the DT concept shall be enhanced with data-driven models tethered to high resolution simulations to provide synthetic model data. Specifically, machine learning models will generate physically-realistic synthetic ensemble members and provide physically-consistent interpolation in time.

The successful Tenderer will produce new high-impact ML/DL methods and software solutions that can work close to the data generation location of the DTs of DestinE and eventually become integrated in the DTE. The output shall be a well documented software package, covering both training and inference data, trained ML/DL models for achieving the above tasks, and an assessment of product quality with respect to the chosen reference output.

Other activities are further detailed in Volume II of the ITT, including further developments the Tenderer may propose.

ECMWF intends to award a single contract for a maximum period of 24 months, which is expected to commence in September or October 2023.

Tenderers should note that ECMWF will also consider the Total Cost of Ownership beyond the contract period and ultimately make an award decision based on the best interests of the Centre taking value for money into account.