Multi-spectral imaging of vegetation for detecting CO_2 leaking from underground

摘要

Practical geologic CO_2 sequestration will require long-term monitoring for detection of possible leakage back into the atmosphere. One potential monitoring method is multi-spectral imaging of vegetation reflectance to detect leakage through CO_2-induced plant stress. A multi-spectral imaging system was used to simultaneously record green, red, and near-infrared (NIR) images with a real-time reflectance calibration from a 3-m tall platform, viewing vegetation near shallow subsurface CO_2 releases during summers 2007 and 2008 at the Zero Emissions Research and Technology field site in Bozeman, Montana. Regression analysis of the band reflectances and the Normalized Difference Vegetation Index with time shows significant correlation with distance from the CO_2 well, indicating the viability of this method to monitor for CO_2 leakage. The 2007 data show rapid plant vigorrndegradation at high CO_2 levels next to the well and slight nourishment at lower, but above-background CO_2 concentrations. Results from the second year also show that the stress response of vegetation is strongly linked to the CO_2 sink-source relationship and vegetation density. The data also show short-term effects of rain and hail. The real-time calibrated imaging system successfully obtained data in an autonomous mode during all sky and daytime illumination conditions.