Here's an example. The behavior appears in older versions of Ferret as well.
yes? define axis/t="1-jan-2006:12":"5-jan-2006:12":86400/units=seconds/t0=1-jan-1970 time5
yes? define axis/y=0:20:5/units=degrees_north yax
yes? let tvar = L[gt=time5]
yes? let ytvar = tvar + sin(y[gy=yax]/30)
! A shade plot shows the shift
yes? set view upper
yes? shade/title="day axis in units of sec-since-1970" ytvar
! A line plot is correct
yes? set view lower
yes? plot/step=connected/color=red tvar
Now redefine the time axis with an origin closer to the data. The plot is correct. Using other units such as hours or minutes with the 1970 time origin also fixes things.
yes? define axis/t="1-jan-2006:12":"5-jan-2006:12":86400/units=seconds/t0=1-jan-2005 time5
yes? set view upper
yes? shade/title="day axis in units of sec-since-2005" ytvar
yes? set view lower
yes? plot/step=connected/color=red tvar
Reported by @AnsleyManke on 19 Feb 2014 23:28 UTC Reported in the Ferret Users List: http://www.pmel.noaa.gov/maillists/tmap/ferret_users/fu_2013/msg00733.html
Here's an example. The behavior appears in older versions of Ferret as well.
Now redefine the time axis with an origin closer to the data. The plot is correct. Using other units such as hours or minutes with the 1970 time origin also fixes things.
Migrated-From: http://dunkel.pmel.noaa.gov/trac/ferret/ticket/2143