Impact of including 20m layer

Case Description SCAM Runs
CAM runs
Comments
L27
27 levels versus 26
Vanilla cam3.5 (cam3_5_36)
SCAM: L27 rainfall timeseries more intermitent due to ZM
Similar patterns to CAM runs occur with HK convection stronger and ZM weaker
CAM: L27 total rainfall stronger in deep tropics
ZM convection too weak; HK convection too strong; LS rain increases
ZM heating weaker at p>500mb stronger for P<500mb
HK heating weaker at p<500mb stronger for P>500mb
L31
31 levels versus 30
Vanilla cam3.5 (cam3_5_36)
SCAM
CAM
L27_no_hmax_inL27
27 levels versus 26
For L27 do not allow ZM convection to originate in the layer nearest the surface (cam3_5_36)
w/ L26
arm95
toga
w/ L27
arm95
toga
SCAM: Increases PRECZ by 29%; reduces PRECSH by 90%; PRECT remains constant (arm95)
Surface fluxes improved but DC timing is to strong an early in L27_nohmax_inL27 (arm95)
PREC changes to look almost identical to L26 (toga)
Heating/moistening distributions more similar to L26 and much smoother
Diurnal cycle timings are now too early
CAPE is reduced by 50%
CAM
L26_mix_parcel_pbl
L26
Continue to launch parcel from khmax, but assume the parcel is well
mixed in the PBL (constant theta, q) and then use that value at the
launch level (cam3_5_36)
w/ L26
arm95
arm97
toga
SCAM: In arm cases ther is stron buoyancy in the lowest part of deep PBLs which
makes sense given that theta is being mixed from higher up.

General Zhang-McFarlane Changes

Case Description SCAM Runs
CAM runs
Comments
L26_dilute_bug01
CAM: cam_conv_03
L26 fix for parcel entropy at launch level to include tpert correctly
Vanilla cam3.5 (cam3_5_36)
w/ L26
arm95
toga
SCAM:
SCAM: L26_parcel_nrg01
CAM: cam_conv_01
L26 with parcel KE control on convective top
Initial parcel energy = 10J/kg
Parcel PE->KE conversion efficiency = 0.2
Vanilla cam3.5 (cam3_5_36)
w/ L26
arm95
arm97
SCAM: Increases sensitivity to dryness just above the boundary layer, but simulation is much more noisy with strong stablization/destablization periods. SCAM: There was not much sensitivity in the mean simulation.
SCAM: L26_parcel_nrg02
CAM:
As L26_parcel_nrg01 but removing the tpert value from the parcel calculation as it is just too
large such that the buoyancy regularly overcomes any inhibition. Vanilla cam3.5 (cam3_5_36)
w/ L26
arm95
arm97
SCAM: Rdeuced excessive buoyancy in the PBL due to the TPERT contribution during daytime conditions. However, there was still a lot daytime max precipitation not observed.
SCAM: L26_parcel_nrg03
CAM:
As L26_parcel_nrg01, but now including the level below the LCL in the KEPAR assessment. In addition there is an additional dependency of PE->KE efficiency depending on whether there is +ve or -ve buoyancy.
Positive buoyancy - low efficiency = 0.2
Negative buoyancy - low efficiency = 0.8
Vanilla cam3.5 (cam3_5_36)
w/ L26
arm95
arm97
SCAM:
SCAM: L26_parcel_nrg03
CAM: cam_conv_02
As L26_parcel_nrg01, but now including the level below the LCL in the KEPAR assessment. In addition there is an additional dependency of PE->KE efficiency depending on whether there is +ve or -ve buoyancy.
Positive buoyancy - low efficiency = 0.2
Negative buoyancy - low efficiency = 0.8
Vanilla cam3.5 (cam3_5_36)
w/ L26
arm95
arm97
SCAM:
SCAM: L26_parcel_nrg04
CAM: cam_conv_j04 (jaguar)
Add the dynamic effects of the large scale velocity to the parcel energetics calculation Vanilla cam3.5 (cam3_5_36) (CAM - pini_ke = 10; pe2ke_eff = 0.2)
w/ L26
arm95
SCAM: The effects are trivial (<< 1 J/kg) for even the strongest ascent. No point in doing a full CAM simulation.
SCAM: L26_clear_cloudy_01
CAM:
Allow the entrainment rate to reflect the pre-existance of convective cloud by setting equal to 1/(cloud top height at previous timestep). Vanilla cam3.5 (cam3_5_36)
w/ L26
arm95
arm97
SCAM: Increases the sensitivity to the presence of cloud but in a too jerky manner such that it is either high entrainment or low entrainment from one timestep to the next. This is probably because there is no consideration of cloud fraction in the calculation. Maybe worth a CAM run.
SCAM: L26_clear_cloudy_02
CAM:
Allow the entrainment rate to reflect the pre-existance of total cloud and adjust the entrainment rate according to E = E(1-cld(i,k)). This will give reduced effective entrainment in response to increased cloudiness in a smoother fashion than L26_clear_cloudy_01. Definitely worth a CAM run. Vanilla cam3.5 (cam3_5_36)
w/ L26
arm95
arm97
toga
SCAM:

Organization runs

Case Description SCAM Runs
CAM runs
Comments
CAM4-alpha control
CAM: cam3_6_07_cntl
Control simulation with MG microphysics (cam3_6_07)
SCAM: CAM:
Mapes organization WITH advection of ORG
CAM: cam3_6_07_convorg00
Mapes organization WITH advection of ORG + MG microphysics (using surface u,v) (cam3_6_07)
SCAM: CAM:
Mapes organization WITH advection of ORG
CAM: cam3_6_07_convorg02
Mapes organization WITH advection of ORG + MG microphysics (using PBL top u,v) (cam3_6_07)
SCAM: CAM:
Mapes organization WITHOUT advection of ORG
CAM: cam3_6_07_convorg01
Mapes organization WITHOUT advection + MG microphysics (cam3_6_07)
SCAM: CAM:

Towards CAM4

Case Description SCAM Runs
CAM runs
Comments
CAM3_6 Control
CAM: cam3_6_07_cntl
Control simulation with MG microphysics (cam3_6_07)
SCAM: CAM:
CAM3_6 + ZM tau=1800
CAM3_6 with ZM convective timescale set to 1800s (cam3_6_07)
SCAM:Reduces TMQ by 1-2mm over land but has a minimal effect over tropical ocean. There is a transfer of rainfall away from shallow to deep. Not really worth a cam run at this stage.
CAM:
CAM3_6 +
CAM3_6 with efforts to restrict convection dependent on condensation within cores (cam3_6_07)
arm95
arm97
toga
gate
SCAM:
CAM:
CAM3_6 +
CAM3_6 with efforts to include OMEGA/Z_convection timescale. Tau is adjusted according to the depth of convection and an in-cloud verical velocity. The in-cloud vertical velocity is linked to a dynamic calculation in closure where there is an assumed initial parcel energy (pini_ke = 10J/kg )and an assumed efficiency of potential energy conversion for the parcel (pe2ke_eff=0.2). (cam3_6_07_var_zmtau01) SCAM:
CAM:
CAM3_6 +
As cam3_6_07_var_zmtau01, but the process begins at the convective initiation level (hmax) rather than the level of free convection. This then allows a true CIN effect to be considered (cam3_6_07_var_zmtau02)
cam3_6_07_var_zmtau02 - Incorrect capping of pnrg<0
w/ Obs.
w/ cam3_6_07_cntl cam3_6_07_var_zmtau02 - Correct capping of pnrg<0; tau_wmean turned off
w/ Obs.
w/ cam3_6_07_cntl cam3_6_07_var_zmtau02 - Correct capping of pnrg<0; tau_wmean turned on
w/ Obs.
w/ cam3_6_07_cntl
SCAM:
CAM:
CAM3_6 +
CAM3_6 with Minghua's convective height restriction so ZM has to top out at preesure <650mb (cam3_6_07_zmtop_gt650mb) SCAM: Very minimal effects on all IOP simulations so it may work well for subtropical regions
CAM: Some effects on the sub-tropical periphery, but otherwise quite minimal.

CAM4 tuning

Case Description SCAM Runs
CAM runs
Comments
CAM4 dev control Control simulation of cam4 development code (camdev07_cam3_6_15) versus cam3_6 (cam3_6_07)
- Includes lots of diagnostics
SCAM:
CAM:
CAM4 dev 01 - Stronger Downdrafts Downdraft strength as % of updraft strength
Control (cam4_dev00, alfa=0.1) = 10% ; Test (cam4_dev01, alfa=0.5) = 50%
SCAM:Increases TMQ by ~2 mm
CAM:
CAM4 dev 02 - Weaker Downdrafts Downdraft strength as % of updraft strength
Control (cam4_dev00, alfa=0.1) = 10% ; Test (cam4_dev02, alfa=0.01) = 1%
SCAM: Reduces TMQ by ~1mm
CAM: Reduces PREH2O by ~0.3mm
CAM4 dev 03 - More liquid in closure clouds
Increase maximum allowed liquid in clouds from 1g/kg to 2g/kg
just for closure purposes
yrs 1 w/ camdev07_cam3_6_15
SCAM:
CAM: Worth a run.
CAM4 dev 04 - Colder freezing in closure clouds
Decrease temperature at which in-convection condensate freezes from 0C to -10C
just for closure purposes
yrs 1 w/ camdev07_cam3_6_15
SCAM: Minimal effects over land: Decreases average buoyancy, Increase PREH2O by at least 1mm.
Over ocean: PREH2O increases
CAM: Not worth a run.
CAM4 dev 05 - Add trigger of mean humidity in convecting column
Add the additional constraint for convection whereby the mean relative humidity of the convecting column has to be greater than 50%
yrs 1 w/ camdev07_cam3_6_15
SCAM: Over ocean no effects (RH>50 always)
Over land: PREH2O and PRECT strangely increases due to less frequent, but stronger events (good). T,q averages better in dry environment.
CAM: Worth a run.
CAM4 dev 06 - Convective parcel temperature excess. Change the temperature excess used for the initial convective parcel (CAM3.5 = 0.5 K in agreement with Tiedke). Change to 2 K.
yrs 1 w/ camdev07_cam3_6_15
SCAM: Over land: Average buoyancy less
CAM:
CAM4 dev 07 - Add convective gustiness due to TOTAL precipitation. From the Redelsperger et al. (2000, J. Clim.) paper add the functional from of the relationship between precipitation and convective gustiness from the TOGA-COARE. Then add the gustiness to the lowest model level wind (ubot,vbot) for surface flux calculations. SCAM: Increases lhflx markedly in TOGA and GATE.
CAM: Improves distribution of precipitation in the western Indian Ocean and in the Western Pacific. But overall precipitation is still too high. LHFLX is about right in the tropics but still way too high in the sub-tropics.
CAM4 dev 07a - Add convective gustiness due to convective rainfall only From the Redelsperger et al. (2000, J. Clim.) paper add the functional from of the relationship between convective rainfall and convective gustiness from the TOGA-COARE. Then add the gustiness to the lowest model level wind (ubot,vbot) for surface flux calculations. SCAM: .
CAM:
CAM4 dev 08 - Add convective gustiness due to cloud base convective updrafts. From the Redelsperger et al. (2000, J. Clim.) paper add the functional from of the relationship between convective updrafts and convective gustiness from the TOGA-COARE. The add the gustiness to the surface wind for surface flux calculations.
yrs 1 w/ camdev07_cam3_6_15
SCAM:
CAM:
CAM4 dev 09 - Add convective gustiness due to convective downdrafts. From the Redelsperger et al. (2000, J. Clim.) paper add the functional from of the relationship between convective downdrafts and convective gustiness from the TOGA-COARE. The add the gustiness to the surface wind for surface flux calculations.
yrs 1 w/ camdev07_cam3_6_15
SCAM:
CAM:
CAM4 dev 10a - Increase dilution Test senstivity to deep convection dilution. Increase from 1/km to 2/km
yrs 1 w/ camdev07_cam3_6_15
SCAM:
CAM:
CAM4 dev 10b - Decrease dilution Test senstivity to deep convection dilution. Increase from 1/km to 0.5/km
yrs 1 w/ camdev07_cam3_6_15
SCAM:
CAM:
CAM4 dev 11 - Remove minimum moiste static energy constraint for detrainment
yrs 1 w/ camdev07_cam3_6_15
SCAM:
CAM:

CLAM Deep Convection Tuning

Case Description SCAM Runs
CAM runs
Comments
RBN00 camdev07_cam3_6_23
Control experiment similar to u37b with extra tendency output
SCAM:
CAM:
RBN01 camdev07_cam3_6_23
Detrainment of the updraft is able to begin below the minimum saturated moist static energy level.
yrs 1 w/ Obs.
SCAM:
CAM:
RBN02 camdev07_cam3_6_23
Mutiply tpert by a factor of 2 for the purposes of the convective parcel temperature.
arm95
arm97
toga
gate
SCAM:
CAM:
RBN03 camdev07_cam3_6_23
Include convective gustiness of Redelsperger et al. (2000) based on total convective rainfall
arm95
arm97
toga
gate
SCAM:
CAM:
RBN04 camdev07_cam3_6_23
SCAM:
CAM:
RBN05 camdev07_cam3_6_23
SCAM:
CAM: