Algorithm for the Retrieval of Columnar Water Vapour from Hyperspectral Remotely Sensed Data

A new algorithm for the retrieval of columnar water vapour content is presented. The proposed procedure computes the area of the H2O absorption feature centred around 940 nm in order to assess its integrated columnar abundance as well as its density at ground. The procedure utilises the HITRAN 2000 database as source of H2O cross-section spectra. Experimental results were derived using radiometrically calibrated hyperspectral images collected by the AVIRIS sensor over Cuprite mining district (Nevada, USA). Numerical simulations based on the MODTRAN 4 were also employed in order to investigate the algorithm performance. An additional empirical H2O retrieval procedure was tested using data gathered by the VIRS-200 imaging spectrometer.