Interactive comment on “Patterns of land-cover transitions from satellite imagery of the Brazilian Amazon” by Finn Müller-Hansen et al

Some minor issues that could be amended at the discretion of the authors: 1) The segmentation procedure is an issue. While in terms of management quality monitoring looking at administrative divisions such as municipalities might appear most relevant, the authors nevertheless mention, that this leads to certain limitations, caused e.g. by the area of municipalities being too small for providing reliable estimates of transition probabilities. For a better detection of regional trends, maybe not in the context of particular municipalities, why not use some standard segmentation procedures, e.g. Voronoi diagrams? This could also simplify the normalization procedures.