MODIS has surface reflectance product too from where we can produce land cover output after supervised classification. From my knowledge of concern Landsat, Sentinel etc. only provide surface reflectance products. Later land cover is derived by us. But why does MODIS provide land cover product (derived from surface reflectance) by itself? Is it because their spectral resolution (250m, 500m, 1000m) is lower than others (10m, 30m) and so a land cover output derived by us will have less accuracy?