Open potash opened 7 years ago
Could this be related to the fact that Lake Michigan (and the other Great Lakes) are not tagged as with natural=coastline
? Or are there clipping issues like this for all waterbodies in the metro extracts?
This also came in to Mapzen support (see Desk 1051). A user sent a Chicago coastline shapefile that was downloaded in October 2015 that has coast and land shapes that were as expected.
This is a problem with any extracts that include Lake Michigan. For example, Milwaukee returns an empty coastline and a large land polygon there, too.
I tested with Tampa and a few custom extracts along the coast, and they were as expected. It appears that something upstream is not happy with the Lake Michigan relation edits.
Often times when water goes missing it's newly moved to the OSM water layer that's a separate data download off Metro Extracts. I haven't confirmed in your case, but check for it there. The same thing has been happening in vector tiles along the Gulf coast (see Texas).
@nvkelso thanks for pointing out the water layer. However, in the case of Chicago the water layer appears empty:
$ ogrinfo chicago_illinois_water_coast.shp chicago_illinois_water_coast
INFO: Open of `chicago_illinois_water_coast.shp'
using driver `ESRI Shapefile' successful.
Layer name: chicago_illinois_water_coast
Geometry: Polygon
Feature Count: 0
Extent: (0.000000, 0.000000) - (0.000000, 0.000000)
Layer SRS WKT:
GEOGCS["GCS_WGS_1984",
DATUM["WGS_1984",
SPHEROID["WGS_84",6378137,298.257223563]],
PRIMEM["Greenwich",0],
UNIT["Degree",0.017453292519943295]]
FID: Integer64 (11.0)
In this case, to clarify, the Water > Coastlines shapefile is empty (0 features) and the Coastlines > Land shapefile has one feature in it that is a large bounding box, possibly the size of the whole lake.
These layers were populated in the past (per the 2015 example sent through support).
In the Chicago extract the Lake Michigan polygon (osm id 1205149) is clipped to something very small and incorrect.