Estimation of grassland CO2 exchange rates using hyperspectral remote sensing techniques

Although the link between CO2 exchange and spectral reflectance is well established in laboratory conditions, limited research has been conducted in the field. To determine the applications of remote sensing in the estimation of the northern mixed grasslands as a carbon sink, the primary objective of this study was to evaluate several narrow-band vegetation indices and band depth analysis in the prediction of leaf CO2 exchange rates in a northern mixed grass prairie ecosystem. Spectral reflectance and CO2 exchange measurements were collected from 13 sites located in Grasslands National Park, Saskatchewan, Canada. Pearson's correlation found a significant relationship between the CO2 exchange rates and the Photochemical Reflectance Index (PRI). Linear regression showed that the PRI explained 46% of the variance seen in the leaf CO2 exchange rates. This is somewhat lower than previous experiments conducted in laboratory conditions; however, the current study was conducted in field conditions, where there are a number of different species in the field of view and background effects from soil, litter and dead materials, which are negligible in laboratory conditions.