MapInfo area calculation method

[Davidll]

Dear list,

We have an issue with the area being reported by MapInfo.

For geographical objects the area reported by MapInfo (using the
Area() function), differs from the area being reported by other
systems eg. ARC, Oracle Spatial, Microstation. These other systems all
report the same area.

At the latitude we are looking at (about 22 degrees south) the
difference is a factor of 0.9974 (the area reported by the other
systems is smaller by this factor than the MapInfo area). We are using
MapInfo data in a Lat/Long GDA94 coordinate system.

Some previous discussion on this topic on an Oracle Spatial forum
elicited this response (from Oracle):

"Oracle's result is smaller than the MapInfo result by a factor of .
997. This is precisely the ratio expected if MapInfo is returning
spherical area using the authalic sphere for this GRS80 ellipsoid.
This authalic sphere has radius 6371007 meters. The Oracle result is
an ellisoidal area, corrected for meridional and prime normal radii of
curvature at the average location of the polygon. At -22 degrees
latitude, this ellipsoidal correction is precisely a factor of 0.997."

I'm just wondering if anyone else has looked into this issue and
whether or not you have come up with similar results. Is there a way
to get MapInfo to report ellipsoidal area?

Thanks in advance for any reponses.

Best regards,
David Llewellyn

[Bill Thoen]

Davidll wrote:
> For geographical objects the area reported by MapInfo (using the
> Area() function), differs from the area being reported by other
> systems eg. ARC, Oracle Spatial, Microstation. These other systems all
> report the same area.
>

Take the cartesian area instead of the standard, spherical area, and I
think you'll find that you'll get a match with the other software.

[Davidll]

Thanks for that Bill,

The areas do seem to compare. However, I first have to convert the lat/
long table to a projected table (MGA94) in order to use
CartesianArea().

Is this just a difference between cartesian and spherical area? Is the
'ellipsoidal area', as it is described by Oracle Spatial, really just
a cartesian area for projected data?

Cheers
Davidll

[Bill Thoen]

Davidll wrote:
> Thanks for that Bill,
>
> The areas do seem to compare. However, I first have to convert the lat/
> long table to a projected table (MGA94) in order to use
> CartesianArea().
>
> Is this just a difference between cartesian and spherical area? Is the
> 'ellipsoidal area', as it is described by Oracle Spatial, really just
> a cartesian area for projected data?
>

I don't know what Oracle uses, but I'm pretty sure MapInfo uses an
adaptation of Girard's Theorem
(http://math.rice.edu/~pcmi/sphere/gos4.html) to calculate areas on a
sphere (not an ellipsoid). And for Cartesiean areas I think they just
assume a flat surface and calculate areas as described in any
computational geometry reference (tessellating a polygon into trapezoids
or triangles and summing those areas).