Digital twin data assimilation is the use of data assimilation, which combines observations with short-range model forecasts to estimate a system state and its uncertainty, to keep a digital replica of part of the Earth system consistent with observations.12

Data assimilation is typically sequential. A previous model forecast is compared with newly received observations, the model state is updated to reflect them, and a new forecast is started.1 ECMWF describes the purpose as determining a best possible atmospheric state and its uncertainties, and notes that its analyses still carry uncertainties arising from the observations and the model.1

Destination Earth describes the creation of several digital replicas covering different aspects of the Earth system, based on simulations and observations.2 That description places observations and simulations in the same system, which is the setting in which assimilation applies.

Which assimilation scheme any Destination Earth twin uses is not stated by the retained sources, and the ECMWF page concerns numerical weather prediction rather than digital twins as such. The term should therefore be read as a combination of two separately documented ideas, and claims about a particular twin need that twin’s own documentation.12

Footnotes

  1. ECMWF, “Data assimilation”. https://www.ecmwf.int/en/research/data-assimilation ↩ ↩2 ↩3 ↩4

  2. Destination Earth, “Destination Earth (DestinE) home page”. https://destination-earth.eu/ ↩ ↩2 ↩3