Re: regridding problem axis listings

To: Bob Smith <Bob.Smith@xxxxxxxx>
Subject: Re: regridding problem axis listings
From: Ansley Manke <ansley.b.manke@xxxxxxxx>
Date: Wed, 09 Jun 2004 11:22:19 -0700
Cc: oar.pmel.ferret_users@xxxxxxxx, Bill Stern <Bill.Stern@xxxxxxxx>, Matthew Harrison <Matthew.Harrison@xxxxxxxx>
In-reply-to: <Pine.LNX.4.44.0406090904470.29592-100000@bgs>
References: <Pine.LNX.4.44.0406090904470.29592-100000@bgs>
Sender: owner-ferret_users@xxxxxxxxxxxxx
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax)

Hi Bob,
Thanks for supplying more information. The missing first value would
correctly show up for a linear interpolation (the default regridding method).
That method uses the coordinates from the source axis, TIME, interpolated
to the points on the destination axis, T2. There is no way to interpolate
16-AUG backwards to 14-AUG since there is no point on axis TIME
before 16-AUG.

The @NRST regridding transformation uses the same logic. It chooses the point
on the source axis nearest to the point on the destination axis, requiring that there
be points on the source axis on both sides of the destination point.. Your first
coordinate on the source axis 16-AUG and so the way this is implemented Ferret
doesn't have a way to go backward to 14-AUG. I agree that one would expect
Ferret to match the point if the grid cells overlap. I'll look into what it would take
to change this.

The @AVE transformation will give you an August value, since it uses weighting
from partial grid cells.

Ansley Manke

Bob Smith wrote:

Regridding over years
Fellow ferreters, I'm having a problem regridding from our orignial time
coordinate to a new one. The original data represents a series of six month aug-jan
runs from 1991-2001. The new time coordinate extends from aug 15 1991 to jul 14 2001
When we do the transformation the first august point is dropped.
yes? sho/l=1:12 ax t2
name axis # pts start end
T2 TIME 121 r 14-AUG-1991 00:00 13-AUG-2001 12:00
T0 = 15-JAN-1901
CALENDAR = JULIAN
Axis span (to cell edges) = 121

L T TBOX TBOXLO TSTEP (MONTHS(JULIAN))
1> 14-AUG-1991 00:00:00 1 29-JUL-1991 18:45:00 1086.916
2> 13-SEP-1991 10:30:00 1 29-AUG-1991 05:15:00 1087.916
3> 13-OCT-1991 21:00:00 1 28-SEP-1991 15:45:00 1088.916
4> 13-NOV-1991 07:30:00 1 29-OCT-1991 02:15:00 1089.916
5> 13-DEC-1991 18:00:00 1 28-NOV-1991 12:45:00 1090.916
6> 13-JAN-1992 04:30:00 1 28-DEC-1991 23:15:00 1091.916
7> 12-FEB-1992 15:00:00 1 28-JAN-1992 09:45:00 1092.916
8> 14-MAR-1992 01:30:00 1 27-FEB-1992 20:15:00 1093.916
9> 13-APR-1992 12:00:00 1 29-MAR-1992 06:45:00 1094.916
10> 13-MAY-1992 22:30:00 1 28-APR-1992 17:15:00 1095.916
11> 13-JUN-1992 09:00:00 1 29-MAY-1992 03:45:00 1096.916
12> 13-JUL-1992 19:30:00 1 28-JUN-1992 14:15:00 1097.916
yes? sho/l=1:12 ax time
name axis # pts start end
TIME TIME 60 i 16-AUG-1991 12:00 16-JAN-2001 12:00
T0 = 01-JAN-1991 00:00:00
CALENDAR = JULIAN
Axis span (to cell edges) = 3471.75

L T TBOX TBOXLO TSTEP (days)
1> 16-AUG-1991 12:00:00 30.5 01-AUG-1991 06:00:00 227.5
2> 16-SEP-1991 00:00:00 30.5 31-AUG-1991 18:00:00 258
3> 16-OCT-1991 12:00:00 30.5 01-OCT-1991 06:00:00 288.5
4> 16-NOV-1991 00:00:00 30.5 31-OCT-1991 18:00:00 319
5> 16-DEC-1991 12:00:00 30.75 01-DEC-1991 06:00:00 349.5
6> 16-JAN-1992 12:00:00 122 01-JAN-1992 00:00:00 380.5
7> 16-AUG-1992 12:00:00 121.75 02-MAY-1992 00:00:00 593.5
8> 16-SEP-1992 00:00:00 30.5 31-AUG-1992 18:00:00 624
9> 16-OCT-1992 12:00:00 30.5 01-OCT-1992 06:00:00 654.5
10> 16-NOV-1992 00:00:00 30.5 31-OCT-1992 18:00:00 685
11> 16-DEC-1992 12:00:00 30.75 01-DEC-1992 06:00:00 715.5
12> 16-JAN-1993 12:00:00 121.5 01-JAN-1993 00:00:00 746.5
yes? list /x=120E/y=5N ($varb)_ens_mn[l=1:12],($varb)_ens_mn2[l=1:12]
!-> list /x=120E/y=5N precip_ens_mn[l=1:12],precip_ens_mn2[l=1:12]
WARNING: Listed variables have ambiguous coordinates on axes: T
LONGITUDE: 118.8E
LATITUDE: 5N
Column 1: PRECIP_ENS_MN[T=01-AUG-1991 06:00:02-MAY-1993 12:00] is PRECIP[D=1991aug2001jan.precip]/NUM_ENS*86400
Column 2: PRECIP_ENS_MN2[T=29-JUL-1991 18:45:29-JUL-1992 00:45] is PRECIP_ENS_MN[GT=T2@NRST]
PRECIP_EPRECIP_E L / 1: 5.84 ....
L / 2: 4.04 4.04
L / 3: 10.22 10.22
L / 4: 1.49 1.49
L / 5: 6.88 6.88
L / 6: 13.63 13.63
L / 7: 8.69 13.63
L / 8: 8.00 13.63
L / 9: 9.81 13.63
L / 10: 9.17 8.69
L / 11: 4.83 8.69
L / 12: 5.16 8.69

Any thoughts on why the first point is being dropped. Any help would be most appreciative.

References:
- regridding problem axis listings
  - From: Bob Smith

Previous by thread: regridding problem axis listings
Next by thread: Continuous colorbar

[Thread Prev][Thread Next][Index]

Dept of Commerce / NOAA / OAR / PMEL / TMAP