Characterising Vegetation Structural and Functional Differences Across Australian Ecosystems From a Network of Terrestrial Laser Scanning Survey Sites and Airborne and Satellite Image Archives

Tuesday, 15 December 2015
Poster Hall (Moscone South)
Stuart R Phinn1, John Armston2, Peter Scarth2, Kasper Johansen1, Michael Schaefer3, Lola Suarez4, Mariela Soto-Berelov4, Jasmine Muir1, William Woodgate4, Simon Jones4 and Andre Alexander Held5, (1)University of Queensland, St Lucia, Australia, (2)Department of Science Information Technology Innovation and the Arts, Remote Sensing Centre, Brisbane, Australia, (3)CSIRO Marine and Atmospheric Research, Canberra, Australia, (4)RMIT, Melbourne, Australia, (5)CSIRO Canberra, Canberra, ACT, Australia
Vegetation structural information is critical for environmental monitoring, management and compliance assessment. In this context we refer to vegetation structural properties as vertical, horizontal and volumetric dimensions, including: canopy height; amount and distribution of vegetation by height; foliage projective cover (FPC); leaf area index (LAI); and above ground biomass. Our aim was to determine if there were significant differences between vegetation structural properties across 11 ecosystem types in Australia as measured by terrestrial laser scanner (TLS) structure metrics. The ecosystems sampled included: mesophyll vineforest, wet-dry tropical savannah, mallee woodland, subtropical eucalypt forest, mulga woodland/grassland, wet eucalypt forest, dry eucalypt forest, tall and wet eucalypt forest, and desert grassland/shrublands. Canopy height, plant area–height profiles and LAI were calculated from consistently processed TLS data using Australia’s Terrestrial Ecosystem Research Network’s (TERN) Supersites by the TERN AusCover remote sensing field teams from 2012-2015. The Supersites were sampled using standardised field protocols within a core set of 1 ha plots as part of a 5 km x 5 km uniform area using a RIEGL-VZ400 waveform recording TLS. Four to seven scans were completed per plot, with one centre point and then at 25 m away from the centre point along transect lines at 0o, 60o and 240o.

Individual foliage profiles were sensitive to spatial variation in the distribution of plant materials. Significant differences were visible between each of the vegetation communities assessed when aggregated to plot and ecosystem type scales. Several of the communities exhibited simple profiles with either grass and shrubs (e.g. desert grassland) or grass and trees (e.g. mallee woodland). Others had multiple vegetation forms at different heights, contributing to the profile (e.g. wet eucalypt forest). The TLS data provide significantly more detail about the relative vertical and horizontal distribution of plant materials. TLS data are providing a step change in satellite image based vegetation mapping, and refining our knowledge of vegetation structure and its phenological variability. Open access plot scale TLS measurements are available through the TERN Auscover data portal.