Relationships between dominant plant species, fractional cover and Land Surface Temperature in a Mediterranean ecosystem

Abstract

The Hyperspectral Infrared Imager (HyspIRI) is a proposed satellite mission that combines a 60m spatial resolution Visible-Shortwave Infrared (VSWIR) imaging spectrometer and a 60m multispectral thermal infrared (TIR) scanner. HyspIRI would combine the established capability of a VSWIR sensor to discriminate plant species and estimate accurate cover fractions with improved Land Surface Temperatures (LST) retrieved from the TIR sensor. We evaluate potential synergies between Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) maps of dominant plant species and mixed species assemblages, fractional cover, and MODIS/ASTER Airborne Simulator (MASTER) LST utilizing multiple flight lines acquired in July 2011 in the Santa Barbara, California area. Species composition and green vegetation (GV), non-photosynthetic vegetation (NPV), impervious, and soil cover fractions were mapped using Multiple Endmember Spectral Mixture Analysis with a spectral library derived from 7.5m imagery. Temperature-Emissivity Separation (TES) was accomplished using the MASTER TES algorithm. Pixel-based accuracy exceeded 50% for 23 species and land cover classes and approached 75% based on pixel majority in reference polygons. An inverse relationship was observed between GV fractions and LST. This relationship varied by dominant plant species/vegetation class, generating unique LST–GV clusters. We hypothesize clustering is a product of environmental controls on species distributions, such as slope, aspect, and elevation as well as species-level differences in canopy structure, rooting depth, water use efficiency, and available soil moisture, suggesting that relationships between LST and plant species will vary seasonally. The potential of HyspIRI as a means of providing these seasonal relationships is discussed.