Thanks for the question. Models can (and probably should) be different, given the different contexts and time windows the cropland masks are required. In case several model are tested in the exploratory period, it would be appreciated if they are documented in the final report (along with respective performances).