Classifying multiyear agricultural land use data from Mato Grosso using time-series MODIS vegetation index data

► We use 5 years of agricultural field site data to examine crop type separability. ► “Soy-Cover” and “Soy-Commercial” emerged as practical class groupings. ► NDVI with date-of-acquisition interpolation emerged as the optimal MODIS dataset. ► Under optimal conditions, 5-year classification accuracies exceeded 80%. ► Year-to-year variations in crop phenology highlight the need for multiyear studies.