I'm plotting the monthly variation on the diurnal scale, Please have a look at the attachment.
currently SET data sets:
temp_h 6-hourly temperature 1:1 1:1 ... 1:1460
temp_m monthly temperature 1:1 1:1 ... 1:12
By default, I have a monthly time axis. Here, I tried to define a 2nd-time axis (for an hourly scale) using an auxiliary coordinate regridding;
If I understand correctly, you want to regrid your 6-hourly temperature timesearies temp_h(t) into a two dimensional variable with its horizontal axis spans a day and vertical axis representing days.
To do so, the RESHAPE function
is what you need. Define a fake "x" axis, which actually represents hours within a day:
let lmax = `temp_h,return=lmax`
define axis/x=0:24:6/edges/units=hours axis_hour ! data defined at 03:00, 09:00, 15:00, 21:00.
define axis/t=0:`lmax/6`:1/units=days axis_day
let grid_var = x[gx=axis_hour] + t[gt=axis_day]
let temp2d = RESHAPE(temp_h, grid_var)
shade temp2d
The code isn't tested. Likely to include error.
[ This solution is less than perfect because the periodicity of the day isn't exploited. I mean, before 03:00 of "today", there is the data at the 21:00 of "yesterday" and after the 21:00 of today, there is the 03:00 of "tomorrow". The fake "x" axis above doesn't represent this continuity. Perhaps you can use three different "x" axes, one spanning [-3, 15], one [3, 21], and the other [9, 27], and merge the three regridded fields . . . I don't know . . . ]
Ryo