Hi Dear Ferreters,
Maybe you would never meet this problem, but it has occupied me for
several days.
I have a model output including temperature and salinity. I want to get
rid of the model long-term trend. So I use regresst.jnl to do it, but the
resulting data shift away from their should-be position. Please see the
attached gif file.
The original data:
1> time_mean.001997.01.01.dta.nc
name title I J K L
TEMP Potential temperature 1:182 1:86 1:32 1:52
SALT Salinity 1:182 1:86 1:32 1:52
This is the script I am using:
set region/x=100:295/y=-71:71/k=1:25/l=1:52
def sym lyear=1:52@ave
def sym tyear=1945:1996
def sym llyear=1:52
go regresst
!
! For temperature
!
let t2=temp-temp[l=($lyear)]
!
! Get rid of trend
!
let q = t2[l=($llyear)]
let p = t[gt=time,l=($llyear)]
let ttt = t2[l=($llyear)]-qhat[l=($llyear)]
!
! Save data to file
!
let ($tname1)=ttt
save/file=($filename) ($tname1)
!
! For salinity
!
let t2=salt-salt[l=($lyear)]
let ($tname5)=ttt
save/file=($filename)/app ($tname5)
Then I get the following data:
1> ./dtrend_E1_CONTROL.nc (default)
name title I J K L
DTEMP TTT 1:98 1:72 1:25 1:52
DSALT TTT 1:98 1:72 1:25 1:52
The figure at upper left of regt.gif is temperature anomaly for the first
layer. It is correct. The upper right figure is for the second layer and
showing southwestward shift of the horizontal position. So do the lower
figures. I try to figure out what is the problem in my script but I can
not get it. Can anyone give me a hint?
By the way, I am using FERRET ver.5.22 for SGI.
Thank you very much.
Haijun
===============================================
Haijun Yang
Department of Atmospheric and Oceanic Sciences
University of Wisconsin-Madison
1225 West Dayton St.
Madison, WI 53706
Phone: 608-262-1956(office)
===============================================
Attachment:
regt.gif
Description: GIF image
