All the fences there came from an Amherst GIS vector dataset. I may have added one or two near houses that were visible in the photos. I also deleted a bunch of obvious misidentified fence segments in that vector set (sometimes bits of fence turned out to be cars in the aerial photos). So its likely some automated algorithm was used for the fences. Sometimes you can see fences in lidar in open areas, just like power lines, but they weren't visible in this data. In this case, the fences are visible in the orthophoto
Same deal with the buildings - the ones on the east half of the map were from Amherst GIS. They are a little ragged. I added the ones on the west side manually from lidar and photos.
I can't tell the difference between earth bank and cliff in the lidar. I make a guess as to which is which based on context and sometimes simply to make it clearer (the brown earth bank symbol can get lost in the contours). Its up to the fieldchecker to make the final decision. They should all really be "generic steep ground thingy" symbol. The one thing that is correct is the facing direction - that you can tell from the lidar, not only from the contours, but the signature in the filtered bare earth is bright-over-dark. Sometimes cliffs don't face directly down the fall line.