-
Notifications
You must be signed in to change notification settings - Fork 0
ACCESS_CM2_AMIPconfigTests
As part of the development process a new recommended CABLE configuration for ACCESS should be defined. CABLE has many different parametrization options, such that not every new switch can be tested independently in AMIP simulations.
AMIP simulations to test #176 (https://trac.nci.org.au/trac/cable/ticket/176) in a default/fixed format provide a chance to test what we feel 'default' should be. Please add your thoughts below on what should be tested or included in a default AMIP test).
soil_struc=default
fwsoil=Haverd2013
gs_switch=medlyn
litter=.true.
rev_corr=.true.
l_revised_coupling=.true. !if possible
soil_struc=default
gw_model=true
or_evap=true
gs_switch=medlyn
fwsoil=Haverd2013
litter=.true.
rev_corr=.true.
l_revised_coupling=.true. !if possible
*This is what is in the Coupled Run at present
cable_user%diag_soil_resp='ON '
cable_user%fwsoil_switch='standard'
cable_user%leaf_respiration='OFF'
cable_user%run_diag_level='BASIC'
redistrb=.false.
satuparam=0.8
wiltparam=0.5
cable_user%cable_runtime_coupled=.TRUE.
cable_user%l_new_reduce_soilevp=.FALSE.
cable_user%l_new_roughness_soil=.FALSE.
cable_user%l_new_runoff_speed=.FALSE.
cable_user%ssnow_potev=''
l_casacnp=.false.
icycle=0
Noting that this is just for the physical climate and attempting to minimise the science changes within CABLE between this default and ACCESS1.3
soil_struc=default - i.e. soilsnow
fwsoil=Haverd2013 - Ticket #93 improved drought response and link between transpiration and photosynthesis
gs_switch=default
rev_corr=.true. - Ticket #139 addressing completeness of the correction terms in ACCESS
l_revised_coupling=.true. - Ticket #132 correcting the meteorological forcing used on the implicit call to CABLE in ACCESS
cable_user%ssnow_potev=''
redistrb=.false.
The choice on gw_switch (i.e. medlyn or another) may come down to whether the dry asymptotic limit is critical.
Use of cable_user%ssnow_potev='P-M' is better science-wise, but is incorrectly coded.
gs_switch = medlyn - #56 data derived functional form for the stomatal conductance
gw_model - #97 - noting that we may have to consider inland water bodies in more detail
litter and/or Or_model - impacting soil evaporation
redistr = .true. - hydraulic redistribution
We should also follow up on
l_runoff_reduction
l_reduce_soil_evap - in partnership with Or_model and litter
fix_access_roots
Same switches as detailed by Mark (mrd561).
Aside from general metrics (e.g. climatologies), also suggest:
-
Rainfall variability at seasonal and interannual scales (e.g. coefficient of variation)
-
Partitioning between sensible and latent heat at seasonal scales. Noting limited obs but known to be poorly simulated by default (offline) CABLE
Initially, I have some confusion about the 'default configuration' mentioned all over the page. Jhan's default configuration is using the same ones as in ACCESS 1.3; we still need this run to compare with old results as a sanity check as we still have not solved the present-day run problem. Ian's default should better be named as 'CM2 basic configuration' to compare with the old default and other test configurations to finally decide on a 'CM2 default configuration'. Mark's default should better be named as 'Ticket#176 test configuration' which only differ from Ian's by one configuration - gs_switch. If you have enough side evidence to accept Haverd2013 into your basic configuration, it seems that the Medlyn gs_switch would fit the criteria as well. How about adopting 'Ticket#176 test configuration' as the 'CM2 basic configuration'? Then Ian's first alternate configuration is the same as Mark's which would be the 'Test1 configuration'; and Ian's second alternate configuration would be the 'Test2 configuration'. So, 4 simulations are needed up to this moment.
Note to Ian - Ticket #97 is Mark's ticket and should not be confused with hydraulic redistribution (name of an existing subroutine) which is Longhui Li's work even though Mark might be doing something similar.
Simple fixes to known problems of too much spring evap and decoupled photosynthesis/transpiration are resolved with fwsoil=Haverd2013 and litter=.true. If we run out of time for more complex configurations, my recommendation is to at least turn these two on.
soil_struc=default
fwsoil=Haverd2013
litter=.true.