Table 1. Criteria of land use/cover types used in selecting the training data for creating the classification models.

All training data were collected using high resolution imagery in Google Earth.

Land use/cover types	Criteria
Forest	Areas of natural or semi natural tall woody vegetation cover (≥5 m) generally with more moderate canopy (≥80% or more cells in the training polygons are trees that means ≥80% forest canopy in a pixel).
Shrub land	Natural or semi natural shrub vegetation generally with a more opened canopy and short woody vegetation (<5 m), where 80% or more cells in the training polygons were shrubs.
Agriculture/Pasture	Areas cover with annual or biannual crops and/or fallow lands mixed with grasslands Often with clear signs of cultivation (i.e., presence of linear till rows).
Plantation	Stand of trees or shrubs cultivated in rows for commercial purposes. This class includes eucalyptus, coffee, fruit trees palm and coconut trees.
Urban/infrastructure	Area with human-made structures including Residential, commercial, industrial, transportation, roads, and mixed urban where more than 95% of cells in training polygons were infrastructure.
Barren land	Areas of exposed soil and barren area with less than 20% vegetation cover
Water	River, open water, lakes, ponds, reservoirs.