R help - Jul 2006 - Normal score transform of spatial data

List:

I have 2 related questions:

(1) first I have x-y-z data, where x & y are the geographic locations of 
point values, z. I need to perform a normal score transform on the 
z-values and maintain their geographic location. So, how do I go from 
columns x-y-z to x-y-z-t (or x-y-t), where the t-values are the normal 
score transforms of the z-values? Can I use qnorm(ppoints(data)) to do 
this; how?

(2) I also need to do a reverse transform of a matrix of normal score 
values (v) to my original data value units; I have seen that I can use 
something like approx(t, x, v)$y. Again, the matrix represents a spatial 
grid. So, how do I maintain the same spatial ordering resulting from the 
reverse transform?

Thank you for your help.

Regards,
Tom


-- 
Thomas E Adams
National Weather Service
Ohio River Forecast Center
1901 South State Route 134
Wilmington, OH 45177

EMAIL:	thomas.adams at noaa.gov

VOICE:	937-383-0528
FAX:	937-383-0033

Hi Tom,

I do not know about R but you might take a look at the GSLIB codes. It has
executables for both tasks using a simplfied GEOEAS format. I recently did some
Normal Score Kriging and use a combination of GSLIB, GEOEAS, and R (for the
contouring of the resulting grid). BTW I used GEOEAS instead of one of the R
packages because of the way the distance in each lag is calculated. I have not
explored the matter but my impression is some codes/packages use the mid-point
of each lag interval and some use the average pair distance for the
interval...need to look at that sometime [Moral: use packages with caution--not
because of any inherent flaw but because of differences in approach. You need
to know what is being done.]

Regards,
Michael Grant


--- Thomas Adams <Thomas.Adams at noaa.gov> wrote:
> List:
> 
> I have 2 related questions:
> 
> (1) first I have x-y-z data, where x & y are the geographic locations
of
> point values, z. I need to perform a normal score transform on the 
> z-values and maintain their geographic location. So, how do I go from 
> columns x-y-z to x-y-z-t (or x-y-t), where the t-values are the normal 
> score transforms of the z-values? Can I use qnorm(ppoints(data)) to do 
> this; how?
> 
> (2) I also need to do a reverse transform of a matrix of normal score 
> values (v) to my original data value units; I have seen that I can use 
> something like approx(t, x, v)$y. Again, the matrix represents a spatial 
> grid. So, how do I maintain the same spatial ordering resulting from the 
> reverse transform?
> 
> Thank you for your help.
> 
> Regards,
> Tom
> 
> 
> -- 
> Thomas E Adams
> National Weather Service
> Ohio River Forecast Center
> 1901 South State Route 134
> Wilmington, OH 45177
> 
> EMAIL:	thomas.adams at noaa.gov
> 
> VOICE:	937-383-0528
> FAX:	937-383-0033
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>