Development of satellite vegetation indices to assess grassland curing across Australia and New Zealand

Development of satellite vegetation indices to assess grassland curing across Australia and New Zealand

Martin, D 2009, Development of satellite vegetation indices to assess grassland curing across Australia and New Zealand, PhD Thesis, School of Mathematical and Geospatial Sciences, RMIT University.

Document type:	Thesis
Collection:	Theses

Attached Files
Name			Description	MIMEType		Size	Downloads
Martin.pdf			Thesis		application/pdf	15.07MB	82

Title	Development of satellite vegetation indices to assess grassland curing across Australia and New Zealand
Author(s)	Martin, D
Year	2009
Abstract	Worldwide, satellite observations are used to assess the fire danger levels of grassland regions using inputs such as meteorological data, topographical data and fuel moisture content, which correlates closely to curing (senescence). This research presents a case study from Australia that assesses the utility of Earth observing satellites to predict grassland curing. Using the normalised difference vegetation index (NDVI), the National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) is utilised across the globe to monitor fire fuel potential. Since the 1980s, this satellite sensor has been used operationally in Australia to produce a satellite-curing index, which was initially developed and validated from the combination of AVHRR and in situ curing observations. These in situ observations, however, were only collected from a small sample of improved pastures and may not be appropriate for other grassland types. In this research, Earth Observation System (EOS) MODerate resolution Imaging Spectroradiometer (MODIS) satellite data, collected from 2005 to 2008, were analysed and compared to in situ curing measurements at twenty-one field sites (of improved pastures, native and mixed grasslands) from different regions of Australia. Derived from the first seven MODIS bands, over thirty vegetation indices were investigated for curing prediction. While no vegetation index was found to outperform NDVI, this index was able to predict curing with an absolute error of 11.84% (on a scale of 0 to 100%) for different grass types and in different bioclimatic regions.
Degree	PhD Thesis
Institution	RMIT University
School, Department or Centre	School of Mathematical and Geospatial Sciences
Keyword(s)	AVHRR MODIS grassland curing NDVI vegetation indices spectral reflectance

Access Statistics:	118 Abstract Views, 82 File Downloads - Detailed Statistics
Created:	Fri, 26 Nov 2010, 10:23:53 EST