# Re: using linint2 for WRF interpolation

From: Dennis Shea <shea_at_nyahnyahspammersnyahnyah>
Date: Sat, 17 Jan 2009 12:54:30 -0700

fi = a->T ; (Time,bottom_top,south_north,west_east)
printVarSummary(fi) ; look at the variable
printMinMax(fi, True ) ; contributed.ncl

WRF XLAT/XLONG are (Time,south_north,west_east)
Generally, they do not change with time.
Hence, read then as 2D arrays

; These are [*][*] (two dimensional)
lat2d = a->XLAT(0,:,:)
lon2d = a->XLONG(0,:,:)
lat2d_at_units = "degrees_north" ; not required for interpolation
lon2d_at_units = "degrees_east"

Your 'target' (rectilinear) grid is lat[*],lon[*].
This could be from some other model
or you could manually create via

latS =-20 ; regular grid output
latN = 35
lonL =-35
lonR = 35

mlon = 50 ; however many points you want
lon = fspan (lonL, lonR, mlon)
lon!0= "lon"
lon_at_units = "degrees_east"

nlat = 40 ; however many points you want
lat = fspan (latS,latN, nlat)
lat!0= "lat"
lat_at_units = "degrees_north"

-----
v5.1.0 has a faster code but the interface is exactly the same

Erik Noble wrote:
> Thank you Dennis, for pointing me to the rcm2rgrid instructions.
> Although there are examples, I'm not clear from the examples on what
> to put for the "lat" and "lon" arguments. It says that they are 1D
> arrays that specify the lat and lon coordinates of the regular grid
> and that they should be monotonically increasing. My model output is
> regional ( latitudes 35W to 35E, longitudes 20S to 35N).

?latitudes with W and E ... no
?longitudes with N and S ... no

This is confusing

Would I put
> these values in for the "lat" and "lon" arguments respectively(
> {-35:35}",{-20:35})?
>
> function rcm2rgrid (
> lat2d [*][*] : numeric,
> lon2d [*][*] : numeric,
> fi : numeric,
> lat [*] : numeric,
> lon [*] : numeric,
> Option : numeric
> )
>
> return_val : numeric
> Thank you for your help.
> -Erik
> On Sat, Jan 17, 2009 at 1:13 PM, Dennis Shea <shea_at_ucar.edu> wrote:
>> http://www.ncl.ucar.edu/Document/Functions/Built-in/rcm2rgrid.shtml
>>
>> WRF data have multidimensional lat/lon arrays:
>> XLAT[*][*], XLONG[*][*]
>>
>> linint2 had coordinates prototypes as one dimensional:
>> lat[*], lon[*]
>>
>> The rcm2rgrid in version 5.1.0 is more efficient.
>>
>>
>>
>> Erik Noble wrote:
>>> Hi. I am trying to learn how to use NCL to interpolate WRF rain data
>>> to observed precipitation from Satellite (TRMM).
>>> Some say use the linint2 function to bilinearly interpolate.
>>> Some don't:
>>> http://www.ncl.ucar.edu/Support/talk_archives/2008/0484.html
>>>
>>> Do we have any examples of using linint2 with wrf data?
>>> -Erik
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>>
>>
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Sat Jan 17 2009 - 12:54:30 MST

This archive was generated by hypermail 2.2.0 : Wed Jan 21 2009 - 13:09:21 MST