Grid convergence

[Olly Betts]

On Mon, Feb 22, 2016 at 05:34:59PM +0100, Martin Sluka wrote:
> 22. 2. 2016 v 11:44, Olly Betts <olly at survex.com>:
> > Survex's approach also allows specifying declinations for different
> > parts of the survey, which is good if you are putting in data for caves
> > over a wide enough area that the declination varies significantly.

> Therion has the identical feature, each survey may have it’s manually
> defined declination.

That's not what I'm describing (though you can do that with Survex
too - it's the only approach Survex supported until 1.2.21).

But if you specify the declination for each survey, you would have to
(a) look it up for each one and (b) do this separately for every
sub-survey which was done on a significantly different date (because the
declination changes with time).

Also each IGRF model is predictive for dates after it was published, and
upgrading to the next model will change such declinations. Probably not
by much, but it's another reason against adding them in by hand.

Say you wanted to view all the caves from a country at once - the
declination can vary significantly across a country, and you'd almost
certainly be dealing with a lot of surveys over many years.  Specifying
declinations by hand is a lot of work, and calculating the declination
based on survey data at a single point can add significant errors.

With Survex you can specify a position to calculate declination at for
each area with caves, or for each cave.  Then you have the declinations
calculated automatically based on survey date and a representative
position - best of both worlds.

If you're wondering why the declination isn't calculated based on the
actual position of the station the measurement was taken at, it's a
chicken and egg problem - you don't know the position of a station
until after you need to know the declination (unless it's a fixed
point).

You could pick a point for the declination, then solve to get positions
for all the stations, then repeat using those positions to calculate
the declinations, and iterate like this until the answers (hopefully)
converge.  But that's significantly more work, and the difference it
makes is not going to be meaningful - the IGRF model is only a model,
and you're also going to soon get past the point where you're well below
the accuracy you can get from instruments.

Cheers,
    Olly