Transcript Effect of topographical resolution on cirrus clouds using a high

Effect of topographical resolution
on cirrus clouds
using a high-resolution GCM
(Seiki et al., JGR, under review)
Tatsuya Seiki1,
Chihiro Kodama1, Akira T. Noda1, Masaki Satoh1,2,
Tempei Hashino3, Yuichiro Hagihara4, Hajime Okamoto3
1.JAMSTEC, 2. The University of Tokyo, 3. Kyushu University, 4. JAXA
The 4th International Workshop on Nonhydrostatic Models (NHM2016)
November 30 to December 2, 2016
Cirrus
Cirrus is a category of
“thin” and “high” clouds
Cirrus is broadly distributed,
hence has strong CRF
 27% for longwave
(Chen et al., 2001)
Cirrus image from wikipedia
However,
GCMs have large bias in cirrus
(IPCC-AR5; Li et al., 2013)
Structures in cirrus (according to Sassen et al., 2007)
 Each cell has structures smaller than 1km
 Several cells are embedded in meso-scale systems less than 200km
 GCMs (dx~100km) have not been capable of capturing these structures
 GCRMs (dx~10km) was used for capturing meso-scale system <THIS STUDY>
Objectives
Resolving effect of orographic gravity wave on cirrus
Typical
Gravity Wave from
Durran (1986)
Homogeneously
Nucleated
Ice Number Concentration [m-3]
Vertical velocity w, trigger of cirrus
• amplitude [ 10 cm s-1 ~ 100 cm s-1 ]
• Horizontal
M
Mwave length [10 km ~ 1000km]
wˆ  h , k
Cirrus Layer
• Ice number concentration (NI) strongly depends on w in cirrus layers
• Joos et al. [2008; 2010] suggested that
NI increases over mountain using a gravity wave parameterization
 Does an increase in horizontal resolution improves simulated NI ?
Numerical Settings (resolution)
NICAM was used (NICAM = Non-hydrostatic ICosahedral Atmospheric Model)
Cloud Microphysics: Double-Moment Bulk (Seiki et al., 2014; 2015)
Horizontal resolution was different between grid (Δx) and topography (Δx*factor)
１）gtopo30 is averaged within a control volume
with averaging radius of r0
r0 
Agrid   f r 0
２）Smoothing by hyper diffusion
ztopo
t

r0


   K H   2 ztopo ,
 ztopo
1
4
K H  x 
16
 g max




p fac
.
Smoothed Topography (NICAM Default)
NOT-Smoothed Topography
G8
dx = 28km
fr0 = 1
G8T
dx = 28km
fr0 = 3
G9
dx = 14km
fr0 = 1
G9T
dx = 14km
fr0 = 3
G10
dx = 7km
fr0 = 1
G10T
dx = 7km
fr0 = 6
Resolution of topography
dx = 7km
dx = 28km
Topographical effect on w and NI
SMOOTHED Topo
NOT-SMOOTHED Topo
Cloud Optical Depth
Cloud Optical Depth
Evaluation of NI (compared to CALIPSO+CloudSat)
Global
IWC [kg m-3]
Cloud Optical Depth
NI [m-3]
Rockies
Global
Andes
Maritime Continent
NI [m-3]
NI [m-3]
NI [m-3]
• Global NI does not improves by changing resolution and topography.
( IWCs from all experiments are in good agreement with obs. )
• NI increases by 50-100 % over the mountain, however, the sensitivity differs by regions.
Summary
topographical effect on Global NI
What’s the response of cirrus to the change in topography ?
As topography becomes fine,
1. Cirrus is strongly perturbed by stronger w
2. Cirrus becomes isolated because of finer horizontal wavelength
Orographic cirrus does not always develop/broaden
NI does not significantly increases in terms of global mean
Topographical data has significant impact on NI when dx ≥ 28 km (G8T)