Michigan Forest Land

David Martin started a great conversation on the mailing list regarding importing the Michigan State Forest Compartments Shapefile from http://gis-michigan.opendata.arcgis.com/datasets/dfe0bcec31184b57b9f0d96bc02d6548_1.

Kevin Kenny surveyed the metadata and posted the results here: https://lists.openstreetmap.org/pipermail/talk-us/2019-March/019315.html

Max Erickson has raised concerns about whether or not the compartments file is the correct way to go. I’ll add that an “MDNR Parcels” shapefile can be found here: http://gis-midnr.opendata.arcgis.com/datasets/1c93c259111241ddb75d308c5d3f2704_3?geometry=-85.022%2C42.793%2C-84.774%2C42.837

The compartment file is divided into a lot of very small areas. They can be coalesced into ones that have ‘UNIT_NAME’ or ‘MANAGEMENT’ in common, but need some minor fixes to the topology (which appear to be just roundoff error).

The ones that are encoded as ‘state park’ or ‘wildlife’ have names that are what a user might expect. The ones that are encoded ‘state forest’ have two name-like fields, ‘UNIT_NAME’ and ‘MANAGEMENT’. I’ve tried slicing them both ways, and you can find the resulting files at https://kbk.is-a-geek.net/tmp/mi_sf.zip They are mostly consistent topologically, but have some minor tweaks (e.g., breaking up long linestrings) before they can be conflated. I’d appreciate if the locals could give it a look-see.

The parcels file has better names - presumably ‘Sault Ste. Marie Forest Management Unit’ is what we want to have replace part of the former ‘Lake Superior State Forest’? It’s even more fragmented though. The ‘compartments’ file is split into areas of about three sections (1920 acres, 7.77 km²) while the ‘parcels’ file is mostly in quarter-quarter-sections (40 acres, 16.2 ha). It has a lot more fields, but they’re Public Land Survey System data that we won’t want to import. (“northeast ¼ of northeast ¼ of Mackinac County, township 43N range 6W section 9” is not very useful to us!) At first glance, it appears that if I aggregate it by ‘Unit_Name’, I’ll get nearly the same result that I got from the compartments.

I don’t have enough local knowledge to do the conflation or to decide what the best aggregation strategy might be. I can surely help with the programming, or with the import proposal - I’ve done a couple of similar ones in New York.

It sounds like there is more in the data source than just state forests. Is the plan to focus on just the state forests? I have interest if it goes beyond that as I have spent time on boundaries of state parks.

Yes, state parks and state wildlife management areas are in the data set. I was focusing on state forests because their data appears to be organized differently. I’ll try to prepare state forest and state wildlife area extracts for you over the weekend.

The ZIP file at https://kbk.is-a-geek.net/tmp/mi_sf.zip is updated - there are now four .osm files in it: the two for state forests (one organized by ‘management’ and the other by ‘unit name’); one for state parks; and one for state game areas.

DO NOT UPLOAD - the tagging is sketchy and there are known to be validation errors. But it’s a start so that people can review what’s in there, verify that the geometries appear more or less as expected, and discuss what other tagging can be added mechanically.