Skip to Main Content
A 1980 national land-cover classification (USGS LUDA aggregated to 1 km) was combined with 2000 satellite imagery (AVHRR NDVI time series) to derive land-cover change information for forest, urban, and agriculture categories in southeast Michigan, part of the USFS North Central Region. To derive useful land-cover change data using a heterogeneous dataset, we (a) interpreted a sample of NAPP photographs from 2000 following 1980 LUDA protocols and used this interpreted land-cover for calibrating a 2000 AVHRR ISODATA classification, and (b) used NRI land-cover statistical data to develop relationships between 1980 NRI and 1980 LUDA proportions for the three classes to develop expected proportions in each class for 2000 AVHRR. We further constrained this by using the NRI predicted proportions with an allocation procedure in IDRISI. Results showed that these approaches we adopted in order to produce comparable classifications improved on AVHRR ISODATA classifications alone.