Recently I have been thinking of a possible project for grad school, and I've decided to try linking Ulleungdo and the Pacific Northwest through some sort of anthropogenic reforestation premise with a focus on biodiversity indexes. That's a long way off, but I decided to jump into GIS and create a map of different vegetation types. To do it, I used QGIS to create ndvi maps of Ulleungdo in both July and April. Using the July ndvi map, I was able to digitize an area for total vegetation cover, since everything is leafed out at that point. Then I used the April ndvi map to digitize evergreen areas (darkest values), then mixed areas (gray values). I'm using the blanket term "evergreen canopy" and not "coniferous cover" to include broadleaf evergreens such as Camellia japonica. After that I had some trouble with the difference tool in QGIS but GRASS came to my rescue (surprising since I had been fighting with it about rasters to no avail). By removing intersecting polygons (evergreen/mixed) from the overall vegetation polygon I was able to get deciduous vegetation. The result is a beautiful map of vegetation cover that I hope to supplement with more field research.
Yay for open source GIS!
