Show simple item record

dc.contributor.authorRoderick, Michael L.
dc.contributor.supervisorDr Richard Smith
dc.contributor.supervisorProf. Graham Lodwick

The monitoring of continental and global scale net primary production remains a major focus of satellite-based remote sensing. Potential benefits which follow are diverse and include contributions to, and improved scientific understanding of, ecological systems, rangeland management, famine warning, agricultural commodity trading, and the study of global climate change.A NOAA-AVHRR data set containing monthly observations of green vegetation cover over a ten year period was acquired and analysed, to extract information on seasonal conditions. The data were supplied as a vegetation index, commonly known as the Normalised Difference Vegetation Index (NDVI), with a spatial resolution of approximately five km. The data set was acquired from three different satellites, and calibration problems were known to exist. A new technique was developed to estimate, and subsequently remove, the calibration bias present in the data.Monthly rainfall measurements were used as surrogate ground truth to validate the NDVI data. For regions of native vegetation, linear models relating NDVI to previous rainfall were derived, using transfer function techniques in common use in systems engineering. The models demonstrate that, in mid-latitude regions, the NDVI is a linear function of rainfall recorded over the preceding seven or eight months.Annual summaries of the image data were developed to highlight the amount and timing of plant growth. Three fundamental questions were posed as an aid to the development of the summary technique: where, when and how much? These summaries highlight the extraordinary spatial and temporal variations in plant growth, and hence rainfall, over much of Western Australia each year.Standard analysis techniques used in time series analysis, such as classical decomposition, were successfully applied to the analysis of NDVI time series. These techniques highlighted structural differences in the image data, due to land use, climatic factors and vegetation type.Overall, the results of the research undertaken in this study, using NOAA-AVHRR data in Western Australia, demonstrate that vegetation indices acquired from satellite platforms can be used to monitor continental scale seasonal conditions in an effective manner. As a consequence of these results, further research using this type of data is proposed in rangeland management and climate change modelling.

dc.publisherCurtin University
dc.subjectvegetation monitoring
dc.subjectseasonal monitoring
dc.subjectvegetation indices
dc.titleSatellite derived vegetation indices for monitoring seasonal vegetation conditions in Western Australia
curtin.thesisTypeTraditional thesis
curtin.departmentSchool of Surveying and Land Information
curtin.accessStatusOpen access

Files in this item


This item appears in the following Collection(s)

Show simple item record