While on OpenStreetMap, I noticed my area lacks a lot of area representing natural forests. I noticed another town not far from me does have forests mapped. I very much doubt that whoever did this actually did this manually, but instead imported the data from another source. So I wondered if there was a way to import features to OSM such as this. I think it may be possible but I have no idea how to do it. Any help would be appreciated. I attached a screenshot of what I’m talking about below.
Depends on the country which sources you can use to import land use.
I’m in Pennsylvania, USA
While there are a few imports of landcover I’m aware of in the US, this tends to be quite complex, especially guaranteeing that the data are of sufficient quality to be imported into OSM.
I noticed another town not far from me does have forests mapped. I very much doubt that whoever did this actually did this manually, but instead imported the data from another source.
I actually suspect that these woods you’re seeing are manually drawn. For instance, take one of the natural=wood polygons in your screenshot: Relation: 7784501 | OpenStreetMap. The changeset that initially added it reads to me like it was one mapper’s hobby to detail the landcover in their local area. Many of the other landcover areas around there have similar histories.
3 Likes
Indeed, many mappers including my past self make drawing landuse and natural areas a bit of a hobby. See also: the #forest-mapping channel in OSMUS slack!
Also, you may have noticed that much of Canada has had forests imported. So it definitely is possible. I recommend reaching out to the users doing that import for their process.
Also also, you might look into the NLCD, National Land Cover Database. It’s high enough resolution it may be importable after some processing. I’ve thought many times about doing this but ultimately decided that (IMO) large swaths of landcover polygons are actually out of scope of OSM. Also, its license may not be compatible with OSM, I’m not sure.
It is possible but it is a ton of effort, work, and organization. You have to find sources, verify their license is compatible, possibly get license approved for import if it’s compatible but happens to be named differently, transform the data, notify everyone, bear with Europeans who have done their imports 10 years ago and now think imports are bad, then add the data to OSM taking a lot of care to merge any existing data.
See Import/Guidelines - OpenStreetMap Wiki for an introduction to the process. See We need to talk about the Canvec elephant in the room - #11 by pnorman for a recent thread about Canadian forest imports.
Because of this, I wouldn’t recommend doing imports for an area smaller than a state. It would very likely be faster to manually draw landcover for a few counties (even starting from complete beginner as an OSM editor) than to try to get an import arranged and completed.
2 Likes
Oh, I completely agree with everything you said. I def didn’t mean to downplay the complexity of an import like that… In my experience imports tend to end up being much more complex than I initially guessed. Haha