[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Gridding options

Ben Tupper <btupper@bigelow.org> writes:

> Craig Markwardt wrote:
> >
> > I don't exactly understand what your data is like.  It sounds like you
> > have 0.5 m x 15000 m resolution, ie. extremely well sampled along one
> > axis and poorly sampled along another.  If that's the case, then the
> > following description may need to be modified.
> You have the right idea.  The ship traveled along a long (mostly) straight
> path.  Every 10-20km the vessel stops and drops the CTD overboard, sampling
> every 0.5 m  over a total depth of 50m - 200m.

Okay now I understand.  So in this case X would be the distance along
the cruise path, and Y would be the depth from the surface. 

> I do see what you are describing.  This is quite similar (in
> methodology) to the iterative gridding process used by a built in
> function GRID in PV-Wave (which I am not using.)
> How are NRX and NRY, for the response function, determined?

The more appropriate question is probably, how broad should the
gaussian be in X and Y?  This depends on how much smoothing you want
to acomplish, and the new sampling.  For example, if your original
sampling was 10-20 km, then the interpolated image might have ~2 km
resolution.  With minimal smoothing, the gaussian sigma would be
around 15 km (ie, comparable to your sampling).  The response function
should have around +/- 2 sigmas = +/- 30 km, which is about 30 pixels.


Craig B. Markwardt, Ph.D.         EMAIL:    craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response