Discriminating California plant species across seasons using airborne VSWIR and TIR imagery

<p><b>Presented at the 2015 Annual AAG Meeting</b></p><p>Accurate knowledge of plant species seasonal and inter-annual distributions are required for many research and management agendas that track ecosystem health. Airborne imaging spectroscopy data have been successfully used to map plant species, but often only in a single season due to data availability. NASA’s proposed Hyperspectral Infrared Imager (HyspIRI) space-borne mission would capture imagery every 16 days using a visible near infrared/shortwave infrared (VSWIR) imaging spectrometer and every 5 days using thermal infrared (TIR) multi-spectral imager. These data will provide the opportunity to discriminate species over a much broader temporal scale and to incorporate thermal information. Here we evaluate: 1) the potential for seasonal discrimination among species; 2) the accuracy of species classifications while plants undergo water stress due to severe drought; and 3) the capability of VSWIR and/or TIR spectra to discriminate plant species. Simulated HyspIRI imagery was acquired in the spring, summer, and fall of 2013 and 2014 spanning from Santa Barbara to Bakersfield, CA with the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) and the MODIS/ASTER Airborne Simulator (MASTER) instruments. Single date and multi-date spectral libraries were created from these images: AVIRIS (224 bands from 0.4 - 2.5 μm), MASTER (8 bands from 7.5 – 12 μm), and AVIRIS + MASTER. We used canonical discriminant analysis (CDA) as a dimension reduction technique and then classified plant species using linear discriminant analysis (LDA). Our results show that compared to single date libraries, the multi-temporal spectral libraries yield higher classification accuracies they better capture spectral variation present in the imagery.