Re: lats and lons
- From: Walter Roberson <roberson@xxxxxxxxxxxx>
- Date: Sun, 17 Aug 2008 19:13:39 GMT
jenya polyakova wrote:
could you please suggest how would I then (without using
griddata) approach my problem? I tried several other ways
but with no success. Thanks.
The reason you have had no success so far is that you have
not clearly defined what you are trying to do.
As I wrote in one of your other threads:
Latitudes and longitudes are not enough information to convertif a significant angular latitude range is being used.
a planar grid. Latitudes and longitudes give you two coordinates
in a spherical coordinate system, but you need the third coordinate
(radius) in order to be able to use the information. Without
the third coordinate, and knowledge of the projection system in use
(e.g., Mercator, Robinson, Lambert) you cannot project onto a
planar grid. Assuming a constant radius is *not* a good approximation
And sure enough, you -are- using a significant angular latitude range.
You are not just trying to create a map of a few square miles, where
the spherical distortion can be ignored.
You have not clearly defined your coordinate system, whether it is
geographic or geocentric, and you have not taken into account the
oblateness of the Earth. On the other hand, according to the constants at
http://topex.ucsd.edu/geodynamics/14gravity1_2.pdf
and a bit of calculation, the maximum geographic vs geocentric
latitude distance works out to be roughly 12 seconds, which is
not significant on the coarse grid you are using. Likewise, the
oblateness is roughly one part in 300; as your grid is only 71 units
for latitude, I suspect you won't care if a point very near a grid
boundary trickles over to the next element.
Mostly, though, you have not defined your projection system and
any accompanying altitude data (or the reference point(s) that
the altitudes are defined relative to.) Saying that you want "a grid"
is not sufficient. *Are* you assuming a cylindrical projection? And do
you care about the significant distortions that is going to introduce
towards the north of your graph? Or to phrase it a different way:
this graph that you are producing, what are it's important characteristics?
Do the horizontal and vertical need to be on the same scale so that you
can do measurements in arbitrary directions? Do the horizontal and
vertical each need to be on linear scales so that a distance measured
in a fixed direction near the top of the graph would be comparable to
a distance measured in the same direction near the bottom of the graph?
Is it important to be able to measure angles off of the graph?
Or is this whole thing just about creating a vague web-page icon of
the Northern Hemisphere, where correctness in any mathematical sense
is not important as long as it looks more or less okay artistically?
--
Q = quotation(rand);
if isempty(Q); error('Quotation server filesystem problems')
else sprintf('%s',Q), end
.
- Follow-Ups:
- Re: lats and lons
- From: Rodney Thomson
- Re: lats and lons
- References:
- lats and lons
- From: jenya polyakova
- Re: lats and lons
- From: Kelly Kearney
- Re: lats and lons
- From: jenya polyakova
- Re: lats and lons
- From: Walter Roberson
- Re: lats and lons
- From: jenya polyakova
- lats and lons
- Prev by Date: Re: solutions of a equation
- Next by Date: Re: Data Input to Matrix and Average Function
- Previous by thread: Re: lats and lons
- Next by thread: Re: lats and lons
- Index(es):
Relevant Pages
|
Loading