CASA: time averaged vs instantaneous pools and fluxes in output?

[SeanBryan51]

Whilst looking into the behaviour of the CABLE MPI driver during a CASA-CNP run, I noticed the MPI master driver is currently hard coded to output instantaneous pools and fluxes when writing the NetCDF CASA output:

CABLE/src/offline/cable_mpimaster.F90

Lines 840 to 841 in 7b87f9c

	CALL WRITE_CASA_OUTPUT_NC (veg, casamet, casapool, casabal, casaflux, &
	CASAONLY, ctime, &

But in the serial driver, it is hard coded to output time averaged pools and fluxes via sum_casapool and sum_casaflux:

CABLE/src/offline/cable_serial.F90

Lines 655 to 657 in 7b87f9c

	CALL WRITE_CASA_OUTPUT_NC (veg, casamet, sum_casapool, casabal, sum_casaflux, &
	CASAONLY, ctime, ( ktau.EQ.kend .AND. YYYY .EQ. &
	cable_user%YearEnd.AND. RRRR .EQ.NRRRR ) )

Is there any preference for one over the other? Perhaps this behaviour should be configurable at runtime via a namelist option.

[har917]

@SeanBryan51 Typically we would analyse and report using averaged values - however when assessing mass conservation we need to compare the averaged fluxes to the difference in carbon (nitrogen, phosphorous) across the time step. That difference should be based on the instantaneous pools.

All that to say that there are legitimate reasons why we would want to be able to do both - possibly even from the within the same run - and, importantly, it's not that we would want to output instantaneous values of all the CNP variables, we would want a mix (so averaged-fluxes and instantaneous pools).

So flexibility would be good.