Hi Jagadish,
By default regridding is by linear interpolation, which results in
"smoothing" (sort of).
Check out regridding using the "nearest" (g=@NRST) transformation:
http://ferret.pmel.noaa.gov/Ferret/documentation/users-guide/Grids-Regions/GRIDS#_VPINDEXENTRY_669
The "exact" (@XACT) regridding transformation is also a possibility
to suppress smoothing of any kind at all (leaving missing value
voids instead).
- Steve
===========
On 2/13/2014 10:08 AM, jagadish
karmacharya wrote:
Dear
all,
I
wonder if anyone had a chance to think about the problem
below, which I posted sometime back.
Any advice would be greatly appriciated or is it not possible
to carry out in ferret?
Thanks in advance,
Jagadish
On Friday, December 6, 2013 1:08
PM, jagadish karmacharya
<j_karmacharya@xxxxxxxxx> wrote:
Dear
all,
I want to mask a coarse resolution data
(upper left plot) outside a region of
interest. I can do that (upper right
plot) but the probelm is: because of
corase data resolution only portion of
desired area is covered (and also some
undesired area outside the boundary). I
can solved the problem of coverage by
first interpolating the coarse
resolution data to a high resolution
grid (lower left plot) but the problem
is: data is now smoothened.
What I want is: to keep the original
magnitude at each location but also be
able to mask out data outside region of
interest. Someone suggested this could
be done by resampling data but I have no
idea how to do that. I would appreciate
any suggestions.
Below is my script for generating those
plots
sh d
currently SET data sets:
1> GPCPv2.2.precip.JJAS.mean.nc
name title
I J K L
PRECIP Average Monthly Rate of Precipi
1:144 1:72 ... 1:19
2> NPL_adm0_box.nc (default)
name title
I J K L
MASK IF PT_IN_POLY(TEMP[L=1,D=3],LON
1:455 1:300 1:1 1:1
! originial data
set vi ul
shade/nolab
precip[d=1,l=1,x=79:89,y=25:31];go land
1,,1
! mask interpolated to the data grid of
2.5 deg resolution
set vi ur
def
grid/x=mask[d=2,gxy=precip[d=1]]/y=mask[d=2,gxy=precip[d=1]]
gxy1
let dummy1=0*x[gx=gxy1]*y[gy=gxy1]
let
N_NoTime1=reshape(mask[d=2,l=1,gxy=gxy1],dummy1);
let Nmask1= if N_NoTime1 then 1
let prm1=precip[d=1]*nmask1
shade/nolab/lev
prm1[l=1,x=79:89,y=25:31];go land 1,,1
! mask and data interpolated to a fine
grid of 0.1 deg resolution
set vi ll
def axis/x=78E:91E:0.1 xax; def
axis/y=25N:31N:0.1 yax; def
grid/x=xax/y=yax gxy2
let dummy2=0*x[gx=gxy2]*y[gy=gxy2]
let
N_NoTime2=reshape(mask[d=2,l=1,gxy=gxy2],dummy2);
let Nmask2= if N_NoTime2 then 1
let prm2=precip[d=1,gxy=nmask2]*nmask2
shade/nolab/lev
prm2[l=1,x=79:89,y=25:31];go land 1,,1
Best regards,
Jagadish
|