Combination of “top-down” and “bottom-up” approaches increases the reliability of estimations;

Fully remote, doesn’t require any on-site operations;

High spatial (up to 30 meters) resolution;

High temporal (up to 1 week) resolution;

Global coverage.

Top-down approach

Bottom-up approach

Multispectral
satellite imagery

Local map, 30-meter resolution

Global map,
5-10 km resolution

AI COMBINED APPROACH

Machine learning algorithm

Ground stations

Ground stations

Global atmospheric transport model

Greenhouse gas observing satellites

Gathering multispectral imagery and gas concentrations from the satellites.
Feature extraction and preparation for data assimilation.

Estimating global carbon fluxes using data assimilation with transport model.

Restoring hi-res carbon fluxes with the use of extracted features.

Сollecting data

Modelling emissions

Aggregating data and preparing the personal account.

Providing estimations