Opened 23 months ago

Closed 23 months ago

Last modified 8 months ago

#2921 closed help (answered)

How can I output the dew-point temperature in JULES?

Reported by: mtodt Owned by: pmcguire
Component: JULES Keywords: dew point, temperature, dew-point temperature, JULES
Cc: Platform: JASMIN
UM Version:


How can I output the dew-point temperature in JULES?

Change History (6)

comment:1 Changed 23 months ago by pmcguire

By inspecting the JULES 4.9 FORTRAN code, it appears that the dew-point temperature is currently only calculated and written to the (annual) dump files when the l_fire namelist variable is set to be TRUE. When l_fire is set to be TRUE, this in turn sets the internal l_metstats to be TRUE.

In the branch off of GL6 (branch from Pier Luigi Vidale) that we have been using (i.e. Rose Cylc suite u-bb422) , the l_fire namelist variable is set to FALSE.

If l_metstats is set to TRUE, then the metstats_timestep() routine is called each timestep, which keeps a running total and each day at local midnight, it calculates dewp_ave_00h.

In order to calculate the dewpt at each timestep (i.e., every 3 or 6 hours), it calls:
CALL dewpnt(msi(:)%spec_hum,msi(:)%press,msi(:)%temp,land_pts,dewpt_ext(:))
in order to compute dewpt_ext (at 1.5 meters altitude in the air) from the specific humidity, pressure, and temperature.

The dewpnt() code is called by the elevate() subroutine every timestep, irrespective of setting the l_fire flag to TRUE. But it doesn't store the tdew variable; this variable is a local variable; and it isn't written to the output files.

If you want the dew-point temperature written to the output NETCDF file every timestep (i.e, every 3 or 6 hours), one would need to modify the FORTRAN code to do so. Another option is to write the 1.5-meter prognostic variables (specific humidity, pressure, and temperature) to disk for every time step, and then compute a diagnostic variable for dew-point temperature afterwards for each time step from the NETCDF files.

Does this help? I can look into it further if you want.

comment:2 Changed 23 months ago by mtodt

Dew-point temperature doesn't seem to be saved, as it is only a local variable in the source code. It might be easier to calculate it during post-processing, especially since surface pressure and temperature and humidity at 1.5m are already included in my output list.


Addendum: Seems like you've been quicker than me.

Last edited 23 months ago by mtodt (previous) (diff)

comment:3 Changed 23 months ago by pmcguire

  • Status changed from new to accepted

comment:4 Changed 23 months ago by pmcguire

It looks like we reached the same conclusions.
I will close the ticket now.

comment:5 Changed 23 months ago by pmcguire

  • Resolution set to answered
  • Status changed from accepted to closed

comment:6 Changed 8 months ago by pmcguire

  • Cc mtodt removed
  • Reporter changed from pmcguire to mtodt
Note: See TracTickets for help on using tickets.