[Talk-us-massachusetts] Town boundaries redux - was Re: Talk-us-massachusetts Digest,

[Greg Troxel]

Bill Ricker <bill.n1vux at gmail.com> writes:

> Those ATLASes are good reading.
> They are the Bible for the circumambulate the bounds exercise for each
> town/city.
> I've downloaded a bunch of them.
>
> Warning on data: the latitude longitude and state X-Y posits listed in
> those were referenced to an antique geode and datum and can NOT be directly
> used. AFAIK.

This is a really good question, and a friend and I dug into it in 2000
(before OSM existed) and I just looked at my notes.

Bottom line up front: I believe that someone has transformed the atlas
coordinates into NAD83 for the towns survey points layer, even if it's
been lost track of how and who.  There is or has been a "state
geodesist" which I think was in MassDOT* that I speculate would have
helped.

*That seems odd at first glance, but it's really MassDOT that is
concerned with cm-level positioning (and MassDOT runs MaCORS).


The basic version of the story about town corner coordinates is that the
coordinates in early legislation and I believe in the Atlases are in
something called the "New England Datum", which was adopted in 1879/1880
by the USCGS as their first datum.  As triangulation networks extended,
they were added on (but without a readjustment, and hence coordinates in
MA remaining fixed) and was renamed to the United States Standard Datum
of 1901 and then the North American Datum (of 1913) when Canada and
Mexico joined.  Then, there was a general readjustment in 1927, but most
coordinates didn't change much.  Of course the adjustment of NAD83 was a
big change around here.

So, it's mostly fair to just treat the old coordinates as USSD, and
somewhat less fair as NAD27.

My friend wrote to NGS, and heard back.  I don't want to forward private
mail, but if you are a US geodesy nerd and have read a lot of NGS
publications you would instantly recognize the person writing back as
among the most qualified to answer any question like this.

The NGS person looked up records and found stations sort of near Stow
(~20 and ~30 miles) for which there were both New England Datum and
NAD83 coordinates, and said that based on those the transform from the
New England Datum to NAD 83 (1996) is:

Latitude = -0.201 seconds
Longitude = - 1.770 seconds

which is very very roughly (assuming 111 km for a degree)
 
 -6.2m in latitude
-36.9m in longtitude

Taking 42.5 -71.5 as my standard point for these sorts of conversions (a
Western Middlesex centric view!), and asking NCAT for NAD27 to
NAD83(1986), I get a shift of that's similar.  (Do this yourself at
https://www.ngs.noaa.gov/NCAT/ to get a nicer-to-read output.)

Input Coordinate	Output Coordinate	Total Change + Uncertainty
Latitude
N42° 30′ 00.00000″
N423000.00000
42.5000000000
Longitude
E288° 30′ 0.00000″
W0713000.00000
-71.5000000000
Ellipsoid Height (m)
Not given
Orthometric Height (m)
Not given
Reference Frame
NAD27
Geopotential Datum
Not given
	
Latitude
N42° 30′ 00.33483″
N423000.33483
42.5000930095
Longitude
E288° 30′ 1.76680″
W0712958.23320
-71.4995092223
Ellipsoid Height (m)
Not given
Orthometric Height (m)
Not given
Reference Frame
NAD83(1986)
Geopotential Datum
Not given
	
Latitude
0.33483″ ±0.001826″
(10.332 m ±0.0563 m)*
Longitude
1.76680″ ±0.002502″
(40.341 m ±0.0571 m)*
Ellipsoid Height
Not given
Orthometric Height
Not given

Then, I asked for the same thing in USSD, which is closer to the values
from NGS in latitude, and there the signs make sense as values to add to
north latitude and west longitude, which makes sense in the historical
context of NAD83 and 2000.  (Now, we are in an ITRF world where
longitude is often positive east.)

Input Coordinate	Output Coordinate	Total Change + Uncertainty
Latitude
N42° 30′ 00.00000″
N423000.00000
42.5000000000
Longitude
E288° 30′ 0.00000″
W0713000.00000
-71.5000000000
Ellipsoid Height (m)
Not given
Orthometric Height (m)
Not given
Reference Frame
USSD
Geopotential Datum
Not given
	
Latitude
N42° 29′ 59.79866″
N422959.79866
42.4999440729
Longitude
E288° 30′ 1.75341″
W0712958.24659
-71.4995129427
Ellipsoid Height (m)
Not given
Orthometric Height (m)
Not given
Reference Frame
NAD83(1986)
Geopotential Datum
Not given
	
Latitude
-0.20134″ ±0.003060″
(-6.213 m ±0.0944 m)*
Longitude
1.75341″ ±0.009674″
(40.036 m ±0.2209 m)*
Ellipsoid Height
Not given
Orthometric Height
Not given

So, I think the best thing that can reasonably be done with the old
coords is to treat them as USSD and to use NCAT.

>  (If Paul @ MassGIS or anyone else has proj4 or similar definitions of the
> old geode / datum /  coordinate systems in the Atlases, that would be
> wonderful to facilitate direct use. I tried to reverse engineer based on
> X-Y being statehouse cupola centered and it didn't work well IIRC. If we
> had a transform(s), it might be practical to georeference the plates as
> layers for our editors or Grass/QGis??)

I would say use NCAT and go for it, and please share!  As I understand
it NCAT has done all the hard work and the best you are likely to do is
reinvent it with some errors.  In particular there is a lot of thinking
about whether to transform e.g. USSD to NAD83 by doing USSD->NAD27 and
NAD27->NAD83 or to do it 1-hop.

> Mass DOT has a public GIS website that has town corners as a layer (in
> addition to USGS and Mass geodetic disks/monuments) with modern GEODE/DATUM
> posits. The atlas will explain more but this gives the modern GPS posit if
> you want to find it. IDK the license status of that data wrto OSM, but
> should be usable for QA/QC.

Earlier versions were clearly usable, and while there is this somewhat
odd CC-BY license tags on the MassGIS website every time I or anyone
else has asked they grant permission.  Looking at the entire history
over the entire time OSM has existed, it is entirely clear that MassGIS
has been and is completely ok with us using their data.

I also discussed with the Secretary of State's office and it seems that
they don't believe public records generated by the state can be subject
to copyright (as in the state could not assert copyright), but it also
seems this is slightly a grey area.  (This is similar to some other
states actual laws that the state can't assert copyright, but here it's
a bit indirect and doesn't have a court precedent.)