Selecting the spatial resolution of satellite sensors required for global monitoring of land transformations

Abstract

The spatial resolution of the next generation of sensors for the global monitoring of vegetation is assessed with particular reference to the proposed Moderate Resolution Imaging Spectrometer (MODIS). The main innovative use of such instruments will lie in their ability to monitor land transformations at global and continental scales. Reliable monitoring is shown to rely on the success with which the changes in the phenomena being analysed can be separated from other temporal changes. Depending on the type of spatial change being monitored, sensor properties such as accuracy of registration, resolution and radiometric sensitivity are shown to have greatest importance. An empirical investigation of the required spatial resolution is based on eight Landsat multispectral scanner system images of the normalized difference vegetation index degraded to candidate resolutions between 250 m and 4000 m. Pairs of images from different dates were registered and different images were then generated. Spatial analysis by Fourier and scale variance analyses indicate that resolutions finer than I km are highly desirable for change detection. A resolution of 250 m will probably generate an impractically high quantity of data on a global basis if all the proposed spectral bands are included. A sensor with a resolution of 500 m is recommended as providing the best compromise between detail of changes detected and the size of the resultant data volume but several other options are also suggested, including one involving one or two finer resolution bands to assist multitemporal registration.