Re: Averaging between arrays with missing values

Thanks for the input. I do see the difference between the two
approaches. As for the 3rd approach, that was where I was heading and
was hoping to avoid as the original data files were in .hdf format,
meaning I would have had to convert them within NCL to .nc and then
used NCO operators. Either way, problem solved. Thanks!

On Jun 4, 2008, at 12:56 PM, Dennis Shea wrote:

>
> NCL has numerous "dim_*" [dim_*_Wrap] functions.
> Generally speaking, they operate on the rightmost dimension ... for
> efficiency.
> This may require reshaping/reordering the arrays using the variables
> named dimensions
>
> Let's say MISR_ASO_2002 (time,lat,lon) [, 2004 and 2006]
> [1] if the times on each file are ntim1,ntim2,ntim3
> NTIM = ntim1 + ntim2 + ntim3
> x = new ( (/ NTIM,nlat,mlon/), typeof(MISR_ASO_2002),
> getFillValue(MISR_ASO_2002) )
>
> x(0:ntim1-1,:,:) = (/ MISR_ASO_2002 /)
> nStrt = ntim1
> nLast = ntim1+ntim2-1
> x(nStrt:nLast,:,:) = (/ MISR_ASO_2004 /)
> nStrt = nLast+1
> nLast = nLast+ ntim2-1
> x(nStrt:nLast,:,:) = (/ MISR_ASO_2004 /)
> x!0 = "time"
> x!1 = "lat"
> x!2 = "lon"
>
> MISR = dim_avg(x(lat|:,lat|:,lon|:))
>
> [2] This *may* give a different answer than [1] .... do u know
> why? :-)
> MISR = (dim_avg(MISR_ASO_2002(lat|:,lon|:,time|:)) \
> + dim_avg(MISR_ASO_2003(lat|:,lon|:,time|:)) \
> + dim_avg(MISR_ASO_2004(lat|:,lon|:,time|:)) )/3
>
> [3] or if they are on files and time is unlimites
>
> ncra MISR_ASO_2002.nc MISR_ASO_2003.nc MISR_ASO_2006.nc
> MISR_ASO_FOO.nc
> Michael Tosca wrote:
>> I have a set of arrays (all same month) that contain lat and lon
>> points. I wish to compute an average value for each lat/lon point
>> from all the arrays, however, some points (different in each array
>> dataset) contain missing values, so that when i just add the arrays
>> and divide by 'n' any lat/lon point that has at least ONE missing
>> value in one of the arrays is essentially "thrown out".
>>
>> My question is: how can i average over the arrays and simply
>> disregard a point when it has a missing value and use the remaining
>> points to compute the average. For example, if I have 10 arrays,
>> and one lat/lon point only has a vlue in 4 of the 10 arrays, I want
>> the average value of that point in the resulting dataset to be an
>> average of the 4 available data. The "avg" function doesn't work
>> because it averages over a single array, whereas I want to average
>> between several arrays. Here's a sample of the code I'm using to
>> average:
>>
>> MISR_ASO_ElNino = MISR_ASO_2002
>> MISR_ASO_ElNino = (MISR_ASO_2002 + MISR_ASO_2004 +
>> MISR_ASO_2006 ) / 3
>>
>> Thanks!
>> -Mike
>> _______________________________________________
>> ncl-talk mailing list
>> ncl-talk_at_ucar.edu
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
> --
> ======================================================
> Dennis J. Shea tel: 303-497-1361 |
> P.O. Box 3000 fax: 303-497-1333 |
> Climate Analysis Section |
> Climate & Global Dynamics Div. |
> National Center for Atmospheric Research |
> Boulder, CO 80307 |
> USA email: shea 'at' ucar.edu |
> ======================================================
>

_______________________________________________
ncl-talk mailing list
ncl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Wed Jun 04 2008 - 14:03:55 MDT