the sensitivity of cloud dynamics and boundary layer ... · experiments comparing generated cloud...
TRANSCRIPT
The sensitivity of cloud dynamics andboundary layer structure to 3-D
radiative heating effects*
I3RC WorkshopWorkshop
William O’Hirok
University of California, Santa Barbara
*a radiative narrative
Water vapor
0
Z
Surf ↓ 3D > ICATOA ↑ 3D < ICAABS 3D > ICA
SZA = 0
3D ICA(horizontal transport)
Water vapor
0
Z
Surf ↓ 3D < ICATOA ↑ 3D > ICAABS 3D > ICA
SZA = 45
3D ICA
ICA3D
SZA = 60°
0 0.1 0.2K/hour
12
km
0
Shortwave heating rate profile0 km 100
12
km
0
MMCR derived cloud field
3D shortwave radiative effectstend to cancel on the scale ofGCMs
ICA
3D
3D-ICA
MMCR
Shortwave heating rate (K/hour)
0 km 100
12
km
0
model resolution (km)
Heating Rate (3D –ICA)
+ 20 %
60 %
- 20 %
0°
60°
Is 3D radiation relevant to climate modeling?(superparameterization)
All 3D shortwave radiation studies have been static
What is the impact on cloud structure and the resultingradiative field?
3DRT CloudResolving Model (WRF-3DRT)
Weather and Forecasting Model
(WRF)
RRTM_SW16 Band
k-distribution
Monte Carlo5 % column accuracy
moisture, temperature,
pressure, and aerosol fields
Toa upwelling, surface downwelling
irradiances, atmospheric heating rates
WRF 3DRT
non-hydrostatic 2D – 400/512 columns x 80 layers 250 m layer and column resolution
240 minute simulation 3 second time step
5 minute radiation time step shortwave - 3D*
longwave -1D RRTM microphysics - water, ice, graupel, snow, rain (Lin et
al.) surface layer physics - Monin-Obukov
turbulence – TKE scheme*ocean/soil surface layers - thermal diffusion
open/periodic boundaries initial temperature perturbation
* modified for air-surface exchange
MONTECARLO
TAUOPT. PROP
COLUMNSTORAGE
band loop
READ ATM,CLOUD, AER
PROFILES&COMPUTE
COLUMN OPTICS.
DISORT
RRTM
ICA out* 3D out
cloud field
RRTM MCWRAPPER
column loop
ERRORyes*For experiments ICAperformed by Monte Carlo
Case 2 Case 3 Case 4
F↑ F↓
MC – DISORT MC – DISORT MC – DISORT
Wm-2 Wm-2 Wm-2
Laye
r #La
yer #
Two not so well thought out experiments comparing generated cloud fields using *3D and ICA radiation schemes
A. Deep convection over water - periodic boundary conditions fordynamics and radiation.
B Deep convection over land - open x boundary for dynamics, periodicboundary for radiation.
*3D radiation on 2D field
periodic boundary 35 min
periodic boundary 55 min
periodic boundary 110 min
periodic boundary 195 min
periodic boundary 220 min
.25 .40
15
0
0 - 80 min 0 - 240 min3DICA
3DICA
domain average condensate(liquid, ice, snow, graupel,rain)
cloud fraction (liquid, ice)
heig
ht (k
m)
15
0
heig
ht (k
m)
.0 .0
Periodic boundary
g kg-1 fraction
Albedo (3D/ICA)
Transmission (3D/ICA)
Atm. Heating (3D/ICA)
Time 13:00 17:00 13:00 17:00
Domain Average
0.8
1.2
0.4
1.6
0.4
1.6
1.01.0 1.0
13:00 17:00
accumulated precipitation
latent Heat
atmospheric absorption
vertical wind variance
2.0
0.0
5.0
0.01.0
0.0
2.0
1.0
1.0
1.2
0.0
0.8
Domain average ratios (3D/ICA)
simulation time (m) 0 24060 120
ICA ICA
3D 3D
13:402.4
2.0
0.0
0.8
0.4
1.6
1.2
0.2
0.1
2.4
2.0
0.0
0.8
0.4
1.6
1.2
0.2
0.1
8.0
2.0
0.5
4.0
1.0
0.
8.0
2.0
0.5
4.0
1.0
0.
cloud condensate (g/kg) SW heating rate (K/hour)
return time window for water vapor
simulation time (m) 0 24060
20
0120
heig
ht (k
m) qv(t) - qv(0) at 0 km
0 km 100 0 km 100
cloud condensate cloud condensate
SW heating rate SW heating rate
vertical velocity vertical velocity
2.4 K/hour1.2 K/hour
-20 m/s
+20 m/s
+11 m/s
-19 m/s
6 g/kg4.1 g/kg
ICA 3D14:40
open boundary 120 min
open boundary 240 min
domain average accumulated precipitation
simulation time (m) 0 24060 120 0 128
12050
6025
0 0
km
mm
mm
open boundary
0 - 240 min3DICANR
1.0 .80
15
0
0 - 240 min3DICA
domain average condensate(liquid, ice, snow, graupel,rain)
cloud fraction (liquid, ice)
heig
ht (k
m)
15
0
heig
ht (k
m)
.0 .0
open boundary
g kg-1 fraction
NR
0 - 240 min3DICANR
0 - 240 min3DICANR
domain average precipitation
50
0250 500 1000 2000 4000
model resolution (m)
5 2 RT min timestep – little difference
open boundary
mm
2d underestimates 3D radiative effects
2d overestimates turbulence
Periodic boundary conditions may enhance modeled chaotic process
Open boundary conditions lose energy
What to do?
x
y
compute-0-1
RRTMMCWRAP
compute-0-2
RRTMMCWRAP
compute-0-3
RRTMMCWRAP
compute-0-4
RRTMMCWRAP
compute-0-5
RRTMMCWRAP
compute-0-…
RRTMMCWRAP
MC
compute-0-1
RRTMMCWRAP
compute-0-2
RRTMMCWRAP
compute-0-3
RRTMMCWRAP
compute-0-4
RRTMMCWRAP
compute-0-5
RRTMMCWRAP
compute-0-…
RRTMMCWRAP
WRF in
WRFHeating rates
multiband optical properties
column group atmospheric properties
all column single band optical properties
all column single band fluxes
Multi-spatial/spectral cluster computing (ROCKS based, psuedo MPI)
Spatial integrator
Errorcheck
Spectralintegrator
photon count message
LITERAL MELT DOWN
Hence, a two field(!) conclusion 3D radiation …