Land cover change monitoring within the east central Louisiana study site: A case for large area surveys with LANDSAT multispectral scanner data
Results established for four digital procedures developed for characterizing the radiometric changes between multidate LANDSAT spectral data sets into meaningful measures of land cover/use dynamics are documented. Each technique's performance was contrasted against digitized land use change maps, which were produced from contemporaneous, retrospective aerophoto coverage, in a cell by cell comparison over a one half by one degree area in east central Louisiana as a standard for comparison. The four techniques identify from 10.5 to 13.0% loss in area of forestland in a five year period; however, they differ more by how accurately this amount of change is distributed, the need for ancillary ground truth, and amount of usable information that is extractable. All require some method of digitally co-registering the two data sets. All are capable of providing tabular statistics as well as map products. Two are capable of detecting changes and identifying their locations. The other two, in addition to this, provide information to qualify land cover conditions at each end of the study interval.