Mapping the Impact of CO2 Sequestration Using NDVI Time-series from Multi-sensor Optical Satellite Data

Geologic Carbon Sequestration (GCS) has been proposed global wide as one “clean coal” technology for mitigating the impacts of global warming. It uses underground reservoirs to store liquefied CO2 that captured from air and found is an effective way for reducing greenhouse gas emissions. However, there is a concern of CO2 leakage from public, and they demand100% assurance that the storage site is safe. A reliable and cost-effective system for the site safety monitoring is thus on the demand. In coop with the general knowledge of top soil contaminating with CO2, to focus on vegetation stress above/around CO2 sequestration sites can reflect the condition of soil composition. This vegetation stress can be detected indirectly by using the Normalized Difference Vegetation Index (NDVI) from satellite image data. In this thesis, this innovative change detection technique is called “GEOES NDVI”.The GEOS NDVI techniques were developed in order to utilize NDVI time series images. It is a combination of the strengths of spike detection and NDVI differencing to detect possible CO2 leakage in sequestration sites in two scales level. It merges these techniques synergistically, allowing each to compensate for the deficiencies inherent in the other. The general strategy is to first confirm hotspot pixels using an NDVI difference image. Confirmed hotspots are then used to derive coarse, regional-level NDVI difference thresholds, followed by more restrictive local thresholds. On a very small local scale, the high resolution Landsat image time series can detect vegetation cover change related to plant stress. The techniques have been tested in three engineered carbon sequestration sites: Otway, Australia, and Ordos and Liulin, China. With its ability to verify vegetation changes using multi-temporal images from multi-satellites and differing resolutions the GEOS NDVI is an innovative technique for monitoring CO2 sequestration sites in both short and long term