© Crown copyright Page 1 Cloud-resolving simulations of the tropics and the tropical tropopause layer Glenn Shutts June 13 2006
© Crown copyright Page 1
Cloud-resolving simulations of the tropics and the tropical tropopause layer
Glenn Shutts June 13 2006
© Crown copyright Page 2
Convective mass flux – “pumping up the lens”
Homogeneous intrusion solution of Gill(1981) adapted for equatorial beta-plane i.e. f = y
cold
EQ N
Zero PV region embedded in background linear meridional PV variation
The large-scale perspective
cold
jet
© Crown copyright Page 3
Balanced geostrophic wind in and around the lens
NEQ
jetstream
1
2
3 13
8
© Crown copyright Page 4
‘big domain simulations of tropical convection
Need CRM domain size > 2000 km (c/)1/2 ~ 1000 km
Need horizontal gridlengths ~ 1 or 2 km
Uniform resolution too computationally demanding for long runs anisotropic grid
Explicit time stepping t ~ 5 sec; run length > 15 days !
3-phase cloud microphysics scheme
Smagorinsky-Lilly turbulence closure
© Crown copyright Page 5
Met Office Cloud-resolving mode (LEM) configurations
Acknowledgment: ECMWF for computer support/resources
Equatorial beta-plane with SST variation, imposed tropospheric cooling (1.5 K/day) and ‘Trade Wind forcing function’ to drive surface easterlies
1. 96 x 7680 x 30 km dx= 1km dy= 1 km Lat. range +/- 35 degs
2. 3840 x 7680 x 30 km dx= 2 km dy= 40 km Lat. range +/- 35 degs
3. 7680 x 7680 x 30 km. dx= 2km dy=10 km. Lat. range +/- 35 degs
4. 40000 x 5120 x 30 km dx=2.44 km dy= 40 km Lat. range +/- 23 degs
Domain dimensions and grids used:35 N
35 S1 2 3
4
Domain shapes
Grid-box shapes
© Crown copyright Page 6
Cooling profile used in simulations
-1.5 0K/day
20 km
15 km
10 km
30 km
© Crown copyright Page 7
Anisotropic grid run
dx= 1 km dy=40 km
Domain :
3840 km
7680 km
35 N
35 S
© Crown copyright Page 8
Hovmuller diagram of rainfall rate averaged from 10 S to 10 N
14 m/s
-13 m/s
© Crown copyright Page 9
U field in vertical sections along the Equator at 6-hourly intervals
Red is westerly
Blue is easterly
-25 < u < 28 m/s
© Crown copyright Page 10
Wave structure at the equator
Fourier decompose u,v and T in the x-direction at all height and at 15 minute intervals
plot the amplitude and phase for each zonal wavenumber as a function of time and height
e.g. u= A(z,t) cos[2kx/Lx + (z,t)] A(z,t) is the amplitude; (z,t) is the phase
© Crown copyright Page 11
Time-height plot of the amplitude of wavenumber 1 for u field (wavelength=3840 km)
0 (white) < amp(u) < 15 m/s (black)
© Crown copyright Page 12
Time-height plot of the phase of wavenumber 1for u field (wavelength= 3840 km)
Kelvin wave phase slope
Period ~ 69 hours0 360 degrees
Black grey white
© Crown copyright Page 13
Time-height plot of the amplitude of wavenumber 1 for v field (wavelength=3840 km)
0 (white) < amp(v) < 11 m/s (black)
© Crown copyright Page 14
Time-height plot of the phase of wavenumber 1 for v (wavelength= 3840 km)
Period ~ 33 hours
n=2 equatorially-trapped inertia-gravity wave
© Crown copyright Page 15
Phase of wavenumber 2 in potential temperature field (wavelength= 1920 km)
Boomerang-shaped phase lines
© Crown copyright Page 16
‘boomerang structure’ – Wheeler et al (2000)
T’ contours
© Crown copyright Page 17
Phase of wavenumber 5 in u field
n=1 equatorially-trapped inertia-gravity wave
© Crown copyright Page 18
Time-height section of u at a point on the equator
Time 0 15.4
days
Z (km)
© Crown copyright Page 19
u perturbation from radiosonde data taken during the ARM Nauru99 field experiment (Boehm and Verlinde,2000)
© Crown copyright Page 20
‘circum-equatorial ’ simulation of tropical atmosphere
horizontal domain extent = 40,000 km
gridlengths dx=2.44 km dy=40 km
Meridional extent: 23 S 23 N
11 day simulation from uniform easterly (5 m/s)
initial dry atmosphere
© Crown copyright Page 21
u in a N-S slice through the domain
23 S 23 NEQ
30 km
12 km
narrow jetstream due to meridional SST
© Crown copyright Page 22
U field in longitude/height section along equator at day 10
40,000 km
z
30 km
Red= 25 m/s Blue= -25 m/s
© Crown copyright Page 23
Potential temperature perturbation + ice mixing ratioalong equator on day 11
30 km
0 40,000 kmSuperimposed ice cloud
amplitude in stratosphere ~ 10 K
© Crown copyright Page 24
From Boehm and Verlinde (2000) time-height section of temp. perturbation and cirrus
© Crown copyright Page 25
Ice mixing ratio at z= 13.9 km
5120 km
8550 km
Sub-domain view
© Crown copyright Page 26
Ice mixing ratio at z= 16.8 km
© Crown copyright Page 27
Ice mixing ratio at z=17.6 km
© Crown copyright Page 28
Correlation of log(qi) and ’ at z= 13.9 km
log(q I )
’
x (longitude)0 40,000 km
© Crown copyright Page 29
Horizontally-averaged ice mixing ratio profile
30 km
20 km
10 km
0 4 x 10-5Ice mixing ratio (kg/kg)
© Crown copyright Page 30
Horizontally-averaged temperature profile
30 km
20 km
10 km
150 200 250 300
T (K)
Too cold !
© Crown copyright Page 31
Summary
convective mass flux terminating near 12 km drives ascent and adiabatic cooling in the TTL
- ‘inflating the lens’
‘Big domain’ simulations of tropical convection using anisotropic grids are a useful intermediate solution to the problem of insufficient computer power
squall lines are organized by Kelvin waves propagating eastward at 14 m/s.
Observed ‘boomerang-shaped’ wave systems are found in the simulations
Model cirrus tends to occur in cold phase of convectively-coupled wave system