Check that ok_mleb=T does not change the results

[jgipsl]

Confirm that multi energy buget using one layer gives the same results as enerbil (standard version which is still kept in CAN).

To be done after #511

[jgipsl]

Added option MULTI_LAYER_CONTROL = 4 to do this test, see commit [6357]

[alanso]

Test were made with r6358.

Settings for the enerbil test:
MULTI_LAYER_CONTROL=4

Settings for mleb one layer case:
MULTI_LAYER_CONTROL=3

OK_GS_FEEDBACK=n
OK_HYDROL_ARCH=n
OK_IMPOSE_CAN_STRUCTURE=y
OK_MLEB=y

JNLVLE_OVER=0
JNLVLS=1
JNLVLS_CANOPY=1
JNLVLS_UNDER=0

The collapsed mleb scheme (i.e. run with 1 layer) could not reproduce the results obtained with enerbil. For longer runs, the largest deviation is seen for the sensible heat flux. The deviations are first observed in fluxsens and fluxlat during the first day of the simulations and then likely propagate to other variables.