| Popis: |
Since the launch of the first earth resources monitoring satellite, remote sensing imagery has been used to provide information on the progress of urbanization, and land cover and land use change. The launch of the first SPOT satellite marked a significant improvement in spatial and spectral resolution for discriminating individual targets and increased the potential to acquire more information regarding land cover and land use. This study aims to investigate the capability of using SPOT digital imagery for monitoring land use change in the urbanised catchment of Tuggeranong Creek in the Australia Capital Territory, and assess its effects on catchment hydrology. SPOT multispectral and panchromatic imagery was acquired over the study area for January 1987 and September 1990. This imagery was digitally processed and analysed using microBRIAN (MB) V3.01 software to derive information on land cover and land use within the catchment. Multi-temporal imagery was co-registered to a base map with sub-output pixel accuracy. In order to improve spatial resolution, the multispectral imagery was merged with panchromatic imagery acquired on the same day using HIS and HPF techniques. The HPF technique retained more integrity of the original multispectral data than did the HIS technique. Both HPF merged and unmerged (original) image sets were used to assess the possibility of using higher spatial resolution imagery in subsequent classification and change detection analysis. On the basis of statistical calculation, non-vegetation classification results were found to be consistent between merged and un-merged imagery, but not consistent for vegetation classes. The inconsistency was found to be the result of seasonal differences in phenology and sun angle. However more small sub-pixel sized features such as houses and lawns were identified using merged imagery. Regression differencing and post classification comparisons were performed on both merged and unmerged image sets to detect temporal changes which had occurred between both image dates. As expected, merged imagery led to more sub-pixel sized examples of change being highlighted using both the HPF and HIS techniques. However, errors associated with multi-temporal image registration, compounded by classification errors arising viI from seasonal differences, meant that the reliability of all identified incidences of change could not be validated. Nevertheless, post classification change detection was found to be the most useful approach for identifying the nature of change from one type of land use to another. The results of classification and change detection techniques were used to diagnose likely changes in catchment hydrology attributable to changes in land use. Preliminary hydrologic analyses found that catchment yield is more sensitive to changes in land use than runoff volume or peak flood discharge. This study confirms that SPOT imagery can be used for mapping and monitoring land use change in urban areas. SPOT imagery was found to be suitable for providing information on land use and land cover changes and assessing the likely hydrologic consequences of such change. The use of imagery from anniversary dates would further improve the reliability of hydrologic assessments based on remote sensing of land use change. |
