ADECEES: Anomaly DEtection of CO2 Emissions via Ensemble Segmentation (Papers Track)
Andrianirina Rakotoharisoa (Imperial College London); Simone Cenci (University College London); Rossella Arcucci (Imperial College London)
Abstract
Latest studies show that we are not on track to limit global warming below 1.5°C compared to pre-industrial levels. Reaching Net Zero is an essential target to reduce global warming and requires accurate and global monitoring of global emissions. In this paper, we introduce our Anomaly DEtection of CO2 Emissions via Ensemble Segmentation (ADECEES) system for the identification of consequences of CO2 emissions on the atmosphere relying on partial diffusion and ensemble segmentation. We apply our system on a global XCO2 dataset and illustrate that it can be used both for the detection of point sources and the detection of variation of emissions.