matplotlib-users Mailing List for matplotlib

nappie74 wrote:
> Hi,
> I have to interpolated extra scattered data added in 2d map with grid
> regular data. 
> Is there any command to krigged or linear weighted interpolating this values
> with prexisting grid data and plotting the result using basemap.
> Thanks in advance.
> Pie.
> 
Pie: Sorry, no - that's beyond the capabilities of matplotlib and 
basemap. You might take a look at
http://sgems.sourceforge.net/
it does kriging, and appears to have a python interface.
-Jeff
-- 
Jeffrey S. Whitaker Phone : (303)497-6313
NOAA/OAR/CDC R/PSD1 FAX : (303)497-6449
325 Broadway Boulder, CO, USA 80305-3328

Jesper Larsen wrote:
> Hi Jeff,
>
> On Friday 06 July 2007 18:28, Jeff Whitaker wrote:
> 
>> Jesper: Hmm, I guess I never thought anyone would make a map that small. 
>> I tweaked some of the parameters to make it work better (svn revision
>> 3470). Here's the diff in case you just want to apply the patch manually:
>> 
>
> Thanks for the patch. And apparantly you were right until now;-) In any case I 
> would guess that at some point basemap would need to be changed.
>
> In shelf sea modelling we are now making setups with horizontal resolutions of 
> down to the order of hundreds of meters and global ocean models are not far 
> from this resolution either:
>
> https://www.navo.navy.mil/nipr_2006/modeling.html
>
> I am not entirely up to date with meteorological models but at least I know of 
> one limited area model that I use which has a resolution of 5 km.
>
> 
>> This will make drawing of meridians and parallels slower, however.
>> 
>
> What about making the resolution dependent on the size of the map if this is a 
> problem? I have a small method that I am using for creating nice contour 
> levels - although smarter methods definitely must exist. I have tried to 
> adapt it for producing what you need. If you decide to include something like 
> this please be aware that the Decimal(str(delta)) should probably be changed 
> (I don't think it will handle all cases well). Maybe it is faster simply to 
> increase the resolution as you have already done when it becomes necessary:
>
> def _getInterval(minval, maxval, ninter):
> """Returns list which resolves minval to maxval with at least ninter 
> intervals."""
> import decimal
> import numpy as npy
>
> # Calculate interval between increments
> delta = (maxval-minval)/ninter
> n = decimal.Decimal(str(delta)).adjusted()
> delta = 10**n
>
> # Round off minimum and maximum values
> xmin = minval/10**n
> xmax = maxval/10**n
> xmin = (xmin - xmin % 10)*10**n
> xmax = (xmax + xmax % 10)*10**n
>
> values = npy.arange(xmin, xmax+delta, delta)
> return values
> 
Jespers: I've found a way to do it that doesn't appear to slow things 
down significantly. Try the latest svn and please let me know how it goes.
-Jeff
-- 
Jeffrey S. Whitaker Phone : (303)497-6313
NOAA/OAR/CDC R/PSD1 FAX : (303)497-6449
325 Broadway Boulder, CO, USA 80305-3328

Hi Jeff,
On Friday 06 July 2007 18:28, Jeff Whitaker wrote:
> Jesper: Hmm, I guess I never thought anyone would make a map that small. 
> I tweaked some of the parameters to make it work better (svn revision
> 3470). Here's the diff in case you just want to apply the patch manually:
Thanks for the patch. And apparantly you were right until now;-) In any case I 
would guess that at some point basemap would need to be changed.
In shelf sea modelling we are now making setups with horizontal resolutions of 
down to the order of hundreds of meters and global ocean models are not far 
from this resolution either:
https://www.navo.navy.mil/nipr_2006/modeling.html
I am not entirely up to date with meteorological models but at least I know of 
one limited area model that I use which has a resolution of 5 km.
> This will make drawing of meridians and parallels slower, however.
What about making the resolution dependent on the size of the map if this is a 
problem? I have a small method that I am using for creating nice contour 
levels - although smarter methods definitely must exist. I have tried to 
adapt it for producing what you need. If you decide to include something like 
this please be aware that the Decimal(str(delta)) should probably be changed 
(I don't think it will handle all cases well). Maybe it is faster simply to 
increase the resolution as you have already done when it becomes necessary:
def _getInterval(minval, maxval, ninter):
 """Returns list which resolves minval to maxval with at least ninter 
intervals."""
 import decimal
 import numpy as npy
 # Calculate interval between increments
 delta = (maxval-minval)/ninter
 n = decimal.Decimal(str(delta)).adjusted()
 delta = 10**n
 # Round off minimum and maximum values
 xmin = minval/10**n
 xmax = maxval/10**n
 xmin = (xmin - xmin % 10)*10**n
 xmax = (xmax + xmax % 10)*10**n
 values = npy.arange(xmin, xmax+delta, delta)
 return values