RF02 SCM Intercomparison

1
RF02 SCM Intercomparison

• Coordinators
• Matt Wyant and Chris Bretherton, UW
• Results submitted to date by
• Andreas Chlond, MPI-Hamburg
• Hitoru Kitagawa, JMA
• Cara-Lyn Lappen, CSU
• Vince Larson, UW-Milwaukee
• Adrian Lock, UKMO
• Stephan de Roode, KNMI

2
Participating SCMs
Name SCM Turbulence Cld. Frac. Microphys.
Austin CCCMa4?
Chlond ECHAM4-5 moist TKE we pdf Sundquist
Kitagawa JMA 1st-order K RH-pdf Sundquist
Lappen CAM3 CAM3UW Nonlocal, sfc-based. K-profile, explicit-we RH/stab RH Autoconv./coll., N 65 cm-3
Larson 2GPDF-HOC From pdf Khair.-Kogan w. joint pdf
Lock UKMO Nonlocal, explicit-we RH-pdf Autoconv./coll., N 100 cm-3
Menon GISS SCM Dry adjustment RH/stab Autoconv./coll. (del Genio)
Roode RACMO EC CY23R4 K-profile, explicit-we Tiedtke Sundquist, PµLWC
3
Case description

• Identical to LES case except suggested
  sensitivity studies
• Vertical resolution
• LR Operational Dz, Dt.
• HR Dz 10 m, Dt 5 s
• Precipitation (P) vs. no precipitation (NP)
• Cu convection allowed (C) vs. no Cu (NC)
• Most SCMs dont allow aerosol, CCN, or droplet
  number to be specified.
• Interest in relation of drizzle to LWP as well as
  their evolution.
• Results are preliminary and have known omissions,
  glitches.

4
LR-P-C (Default) Initialization

• Mainly fine.
• JMA loses cloud fast.
• UKMO drizzles a lot.
• CAM doesnt have ug.
• RACMO dry above PBL.

5
LR-P-C Evolution

• LWPs 100-150 g m-2 except for JMA, RACMO.
• All models but JMA hold onto cloud.
• High-LWP models range from 0-1 mm d-1 drizzle.

6
Surface drizzle vs. LWP

• Diverse sensitivities.
• Microphysical parameterizations or droplet size
  assumptions?

7
Cloud-base drizzle vs. LWP

• Max(drizzle flux profile)
• Isolates production (vs. evap.)

8
High-resolution (HR-P-C) results

• Results broadly similar to LR.
• JMA holds onto cloud better.
• CAM and CAM-UW have higher LHF/LWP/drizzle.

9
HR surface drizzle vs. LWP

• Same diversity as LR.

10
Precip vs. no-precip sensitivity studies

• In drizzly models (except JMA), LWP increased
  substantially by drizzle suppression.

11
Summary

• SCMs display a wide variety of drizzle-LWP
  sensitivities, scattering on both sides of
  observations.
• In some SCMs, drizzle is substantially reducing
  LWP.
• Most SCMs could not specify 65 cm-3 cloud
  droplet concentration.
• Overall, the case specifications seem effective.
  Specified surface fluxes and interactive
  radiative cooling profiles would have been easier
  for SCMs.