[ncl-talk] 4 sub-blocks of a 3D matrix in satellite data

[Yang, Xiaoxin]

Dear Dennis,
Thank you for your reply. It is said that each observation has an averaging kernel matrix. The grid_pressure_composite includes 34 levels, with index 0 to 16 referring to hdo and the remaining levels to h2o. Hdo and h2o vapor mixing ratio are retrieved simultaneously.
Each averaging kernel matrix has 4 sub-blocks correspond to the HDO/H2O ratio estimate , the H2o estimate and then the cross-terms.
The guide also explains that if the matrix is 32x32 then there are 4 components that are each 16x16 elements long.
Yet I still felt uncertain about the structure. Will this clarification help you in figuring out the structure of the sub block of the 3D matrix?


在 2016年10月11日，下午10:46，Dennis Shea <shea at ucar.edu<mailto:shea at ucar.edu>> 写道：

TES-Aura satellite data for HDO:

Not sure what to say.  We have not seen these files. Certainly, I have not read any documentation. Does the indexing in the documentation you read say it is 1- or 0-based?



On Tue, Oct 11, 2016 at 4:27 PM, Yang, Xiaoxin <yang.1143 at osu.edu<mailto:yang.1143 at osu.edu>> wrote:
Dear all,
I’m processing the TES-Aura satellite data for HDO, but could not figure out the structure of the 4 sub-blocks of a 3D matrix, i.e., averagingkernel(grid-targets*grid_pressure_composite*grid_pressure_composite).

As is stated in the guide, the averagingkernel consists of 4 sub-blocks of the matrix [0,0],[1,0],[0,1] and [1,1]. Subblock [0,0] ranges from indices ns to 16, and subblock[1,1] ranges from indices 17+ns to 33…

Should it be something like this?

ak=F->AveragingKernel(vid,:,:) ; vid means valid observation meeting certain criteria
ak at _FillValue = -999
;averaging kernel for HDO
Add = ak(:,ns:16,ns:16)

But the result of the corrected HDO using this averaging kernel does not seem right. Hope you would offer me more suggestions on this.
Thank you very much.
Best,
Xiaoxin


_______________________________________________
ncl-talk mailing list
ncl-talk at ucar.edu<mailto:ncl-talk at ucar.edu>
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20161012/740dedf2/attachment.html