Here's another approach to do what I think you want (picture attached):
! plot your time series define axis/t=1-jan-1990:1-jan-2000:5/units=days t1990 plot SIN(T[gt=t1990]/100)
! overlay a line through 0.0 from 15-sep-1990 to 15-jul-1998 plot/over/vs/line/nolab {`T[gt=t1990,t=15-sep-1990]`,`T[gt=t1990,t=15-jul-1998]`},{0,0}Using PLOT/vs you need only make sure that the X coordinates are in the same time encoding as the base time series plot.
Hope this helps - steve
==============================================
Gary Strand wrote:
I'd like to add a simple, constant value straight line to a plot of a time
series of several quantities. For example, a time series of several model runs'
average sea surface temperatures, and I want to include the mean Levitus
value. The hitch is that not all the model run time series are of the same
length, and, IIRC, it's not easy to specify the begin and end of a time axis.What I've done up until now is something like:
plot/nolabel sst1,sst2,sst3,sst4,sst5
let lev = if sst1 eq 1.e10 then 0 else 18.39
plot/over/line=1 levBut that's not real clean if sst1 is much shorter in time than the other four
runs.Any ideas?
PS - It would be nice to have "user" coordinates to "/nouser" transforms, both
directions, readily callable by FERRET, like GrADS does, so one can use geo-
graphic coordinates and the "inches" coordinates on the plotting page
more-or-less interchangably.
--
Gary Strand Climate Change Research Section, NCAR (303) 497-1336
strandwg@ucar.edu http://www.cgd.ucar.edu/ccr/strandwg
--
Steve Hankin
NOAA/PMEL, 7600 Sand Point Way NE, Seattle, WA 98115-0070
ph. (206) 526-6080 -- FAX (206) 526-6744