State Boundary Incorrect

Hello,
I want to apologize if this is something that’s already known to the community; I’m new to OpenStreetMap. I have been correcting names and geographies in my area recently and found that the Pennsylvania/West Virginia state line is about 350-400 feet north of the actual border. Is that correctable?

Thanks

According to a note on the border boundary, the National Geodedic Survey finds the east-west portion of the Mason-Dixon line precisely at latitude 39.7222206 . I’m not an expert, but I believe that this would define the PA-WV boundary. According to that measure, the western edge of PA-WV is pretty close; within 1 Meter. And the eastern edge of PA-WV would actually move further north by about 130 Meters.

Newer and better data is available for most state boundaries, but since it’s so advanced and laborious, not many have been updated. Consider this an advanced operation. Feel free to open a “Note” in OSM on the map, and someone will eventually look at it. But if you want to get into the details, here are some things to watch for:

In some cases, roads have been incorrectly bound to administrative boundaries. When changing an administrative boundary, be sure that the roads remain connected and follow the view in Bing imagery or GPS tracks.
The PA-WV boundaries have been optimized so that a single line defines county and state lines. This should not cause any problems but is something to be aware of.

I read a bit more about the Mason-Dixon line - the line may not be a straight line, but is marked by crownstones. In some cases, the latest TIGER data is based on precise survey data, but may vary from county to county.

Here’s a screenshot showing the issue located here. With the MapQuest layer selected, you can see both boundaries. The correct location is labeled Pennsylvania - West Virginia in the MapQuest view. The gray line above it is where the boundary will display in the standard or humanitarian views. I’m not sure I’m up for advanced editing - I’ll stick with editing in the browser for right now - so I’ll open a note on it for folks to check it out.

Thanks for the clarification. It’s possible that MapQuest might not be using the most accurate boundary data. I hope the real experts will either make corrections or give more of an explanation to your note.

The eastern boundary of WV at the junction with MD and PA has been identified as one of the survey monuments, but even that differs from the MapQuest boundary by a number of meters.

The monument node: http://www.openstreetmap.org/node/75390572/history#map=19/39.72100/-79.47670

Unfortunately, the last editor of the monument node doesn’t cite the source: Published coordinates? GPS / single waypoint? GPS / Multiple waypoints on multiple days and times of day? GPS in the winter after leaves are off the trees? The area of that monument has treecover which would affect the accuracy.

I suspect that some other monument points are needed along the Mason-Dixon line to further improve accuracy.

I drive over the border in several places near here multiple times a day. MapQuest does have this one correct.

Trivia: The last monument placed by Mason and Dixon was about 30 miles shy of the end of the southern border of Pennsylvania. It was only meant to survey the border with Maryland. We used to have family reunions at the park where it’s located.

It seems that most borders are things that the average citizen mapper can’t fully check. Perhaps the center line of a river as it was 150 years ago before a series of floods changed it. But, if I am reading things correctly, the southern border of Pennsylvania is an exception and is something that the citizen mapper can check and correct.

From http://en.wikipedia.org/wiki/Mason-Dixon_Line it seems that however it was originally laid out, the border is a series of straight line segments between the “stones” (every mile) and “crownstones” (every 5 miles). And apparently these are still there and still the official boundary. So with a GPS, especially one with waypoint averaging to get within an acceptable error, one could get the WGS84 lat/lon for each stone. Then using the editor of choice add nodes to the boundary way for each stone and set their location per the captured waypoint.

Perhaps these folks have something available to download and make correcting the border easier. I’m still a novice, so I’m not sure.