Characterization of intra-annual reflectance properties of land cover classes in southeastern South Dakota using Landsat TM and ETM+ data

Landsat-7 and Landsat-5 have orbits that are offset from each other by 8 days. During the time that the sensors on both satellites are operational, there is an opportunity for conducting analyses that incorporate multiple intra-annual high spatial resolution data sets for characterizing the Earth's land surface. In the current study, nine Landsat thematic mapper (TM) and enhanced thematic mapper plus (ETM+) data sets, covering the same path and row on different dates, were acquired during a 1-year time interval for a region in southeastern South Dakota and analyzed. Scenes were normalized using pseudoinvariant objects, and digital data from a series of test sites were extracted from the imagery and converted to surface reflectance. Sunphotometer data acquired on site were used to atmospherically correct the data. Ground observations that were made throughout the growing season by a large group of volunteers were used to help interpret spectroradiometric patterns and trends. Normalized images were found to be very effective in portraying the seasonal patterns of reflectance change that occurred throughout the region. Many of the radiometric patterns related to plant growth and development, but some also related to different background properties. The different kinds of land cover in the region were spectrally and radiometrically characterized and were found to have different seasonal patterns of reflectance. The degree to which the land cover classes could be separated spectrally and radiometrically, however, depended on the time of year during which the data sets were acquired, and no single data set appeared to be adequate for separating all types of land cover. This has practical implications for classification studies because known patterns of seasonal reflectance properties for the different types of land cover within a region will facilitate selection of the most appropriate data sets for producing land cover classifications.