Exploring ordination as a method for normalizing disparate datasets: implications for digital change detection

Researchers interested in digital change detection are well aware that the present plethora of earth observation satellites brings with it a mixed blessing. The evolution of satellite technology has evolved rapidly with regard to changes in the spatial and spectral resolution of acquired data. This has revealed numerous research possibilities, with associated unique challenges, that were previously impossible. Potential implications of this are immediately apparent when research questions require an effective transition from data to knowledge. Specifically, this research employs principal components analysis (PCA), a multivariate ordination technique, to perform multi-date change analysis. The application of PCA offers a method for standardization of disparate image types, a necessary step in any digital change-detection procedure. Change detection is now a common procedure in studies employing remotely sensed data to study differences between two dates. However, interest is gaining in its potential use as an analytical tool for identifying and monitoring environmental change in disparate data sets, acquired at multiple time intervals. The reasons for this are twofold: scientists and concerned members of the public are increasingly interested in understanding anthropogenic influences on the environment; and the archive of satellite data continues to grow permitting longer and more sophisticated analyses of environmental change. This research uses satellite imagery acquired of a rapidly urbanizing coastal city in Southern China. Four images obtained from three different satellite sensors (Landsat 5 TM, SPOT HVR, Landsat 7 EMS+) form a twelve-year chronology. PCA is used to help elucidate coastal urbanization by offering one method of normalizing differences in the spatial and spectral resolution of these data. This paper contributes to the development of methods for normalizing disparate datasets, an essential requirement for conducting meaningful investigations of environmental change.