Single.Earth's Digital Twin represents a hierarchical physical system state. It is updated in near real-time and/or sporadically utilizing input from big data, including ground-based human observations, sensor-based observations (spaceborne or ground-based), models and human decisions. Such a system state is defined by models representing current land use, land cover, and ecosystem parameters (such as carbon removal and storage and biodiversity status) that are applied to the digital twin at different temporal and spatial scales. Human decisions are used to update the structure of the model(s) associated with the digital twin.
Figure below: Conceptual workflow of the constantly updated Digital Twin.
Our digital twin is designed to guarantee complete transparency, be highly reproducible, and be constantly updated and improved. The transparency and reproducibility of our methods are made possible by storing all the parameters and processes that make up the digital twin in an open-access GitHub repository and creating a cryptographically verifiable reference to these resources on the blockchain.
The constant updates and improvements are made possible thanks to big data repositories that are periodically updated, and to the endless effort of our scientific and technology teams who constantly experiment and develop new solutions in collaboration with partners.
Bi-weekly, we monitor the intactness of the forest canopy cover of the lands onboarded to the Single.Earth platform with our in-house monitoring tool that utilizes satellite data (Sentinel-1) and machine learning models. Single.Earth uses annually updated global forest loss datasets from Landsat time-series imagery to monitor forest cover. These datasets, referred to as the Hansen map, are high-resolution maps that cover all global land except Antarctica and several arctic islands. These maps generate the estimation of tree cover extent, loss (a stand-replacement disturbance that includes natural (e.g., fire or storm damage, and diebacks due to disease) and anthropogenic (e.g., mechanical removal) disturbances that are observable in Landsat imagery) and gain (a non-forest to forest change, which includes observable gains in tree cover after harvest and other disturbances) at a spatial resolution equal to a Landsat pixel of 30 m.
Curtis, P. G., Slay, C. M., Harris, N. L., Tyukavina, A., & Hansen, M. C. (2018). Classifying drivers of global forest loss. Science, 361(6407), 1108-1111.
Hansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. A., Tyukavina, A.,
... Townshend, J. R. G. (2013). High-Resolution Global Maps of 21st-Century Forest Cover Change. Science, 342(6160), 850-853. https://doi.org/10.1126/science.1244693