CDO

Try the CDO operator remapnn for this talk:

cdo remapnn,t1.nc sU10.nc sU10_on_t1_grid.nc

this doea the conversion . But there is something very strange i observed :

when i use cdo to read information regarding the transformed variable i get good results that is mean meadin,max,min

cdo info sfix.nc
-1 : Date Time Level Gridsize Miss : Minimum Mean Maximum : Parameter ID
1 : 2000-01-01 00:00:00 10 7200 0 : -8.2600 -0.10724 5.9000 : -1

however when i use ncl to extract grid information i get value in 2000. i dont have clue what went wrong, apparently it seems like that values are not correctly represented at grids

(0,0,0,0) -21035
(0,0,0,1) -21035
(0,0,0,2) -21035
(0,0,0,3) -21035
(0,0,0,4) -21035
(0,0,0,5) -21035
(0,0,0,6) -20992

is there any way to solve the problem, i am still struggling with it.

without the file, it's hard to guess

hi @Ralf Thehos Mueller please fingf the files in attachment

your input is a short variable, i.e. 16bit integer. CDO used double internally, thats why the info operator shows reasonable results.

with ncl, you get the plain integer values. Use the copy operator together with the -b option, e.g.

 cdo -b F32 copy <in> <out>

btw: ncview can handle it

Thanks a lot, this explains it all , also i came across this solution, when i type printVarSummary i got the following in NCL

scale_factor : 0.01
add_offset : 207.65
units : m/s
actual_range : ( -32.79, 37.59 )
unpacked_valid_range : ( -120, 120 )
valid_range : ( -32765, -8765 )

which explains :

here are “scale_factor", and "add_offset”. Once you apply those to your numbers, you are fine. Some software packages apply these automatically once you open the variable, and sometimes (particularly when looking into the file) you still need to apply them.

However i liked your solution more, thanks a lot Ralf.