From this interpretation station we watch video in conjuction with viewing the TM imagery. Through the use of GIS, GPS and the time code recorded on the video, we can tell the exact location of each video frame on the TM imagery. We can then write to a data file the position and the assigned vegetation alliance. It is this data set that we use in creating the decision rules for our models.