1. Tropical Convection and MJO
Shuguang Wang

2. Super cloud cluster on Nov 24, 2011
Convection is coupled with a range of dynamic equatorial waves and madden julian oscillations. This is illustrated in the these OLR.
Which I happened to be part of it, more cloud processes. Which can be more specified shown in WK diagram.
24 Nov 2011, Meteo7 IR
Tropical convection organize into large scale

3. Tropical convection f = 0 f = 0.5x10-4 s-1 f = 1x10-4 s-1
Liu and Montcrieff 2004 3

4. Large scale influence of convection
ECMWF Training Course
02 May 2000
Large scale influence of convection
Consider dry static energy:
Qr: radiation
c: condensation
e: evaporation
Taking spatial average over a large area, thermodynamic equation becomes :
Eddy transport
Large-scale vertical motion
Heat effects of convection
Moist Processes

5. Convection: large scale effects
ECMWF Training Course
02 May 2000
Convection: large scale effects
Thermodynamic eqn
Apparent heat source
Moisture eqn
Apparent moisture sink
Apparent momentum source
Momentum eqn
These are “observables”
Moist Processes

6. Yanai et al., 1973, JAS ECMWF Training Course 02 May 2000
Moist Processes

7. Sounding array in tropical field programs
Radiosondes are released at the “polygon” network every 3-6 hours
Flux of humidity is calculated using sounding observations
Based on the measurement of all the sources and sinks, the budget of moisture and energy can be constructed. Large scale vertical motion can also be derived.
Large scale vertical motion
Horizontal advection of moisture and temperature
Sea surface temperature (SST)
Surface wind speed

8. Sounding array observations in tropical field campaign
GATE (Atlantic)
TOGA-COARE
(Western Pacific)
DYNAMO
(Indian ocean)
Sounding array: Continuous sampling of atmospheric states over several months
Products: areal average of budget terms for thermodynamic variables

9. Sounding array observation: Column modeling
Precipitation fits well Obs, but other important fields may not
Temperature, moisture
Radiative fluxes
Reflectivity, or microphysics
~64 km
22 km
Dx = 1 km
TOGA: Rain

10. Column modeling with parameterized large scale dynamics
Standard method
Alternative modeling strategy
Large scale vertical motion
Sea surface temperature (SST)
Relaxation to large scale winds
Horizontal advection of moisture
WTG or wave coupling (WPG)
Sea surface temperature (SST)
Relaxation to large scale winds
Free troposphere temperature
Radiation
Goal: Simulate cloud statistics
Goal: Simulate rainfall

11. Two methods: WTG and Gravity-wave coupling
C R M
relax Θ
climatology
C R M
Use theta’, schematic
: timescale
k: wavenumber
ε: momentum damping at1 days
Both:
LS vertical motion
temperature variations

12. Parameterizing large scale dynamics for TOGA-COARE
Surface fluxes
WTG
Damped-wave
Surface fluxes drive the responses at the given wavenumber or time scale; In fact, surface flux underestimated around day 80, while E at day 90 is similar to those at the first major MJO event, around day

13. Parameterizing large-scale vertical velocity for TOGA-COARE
WTG
Damped-wave
Now we see model results with large scale vertical velocity parameterized with two methods: WTG and Gravity-Coupling.
Both methods capture the gradual ramp up during the first event, but overestimate rain during the 2nd event.
Given the complexity of MJO, I contend that these results are considered to be quite good.
- Rain is not specified by the large-scale forcing. Expect much less agreement.
- Good agreement in rainfall for the 1st MJO event, less agreement for the 2nd MJO event

14. Parameterized large scale vertical velocity
obs
Potential Temperature: model-obs
WTG
WTG
In WTG, vertical velocity and temperature perturbations well correlated. But not the case in Gravity-wave-coupling. This suggests the WTG is more sensitive temperature in upper troposphere.
Gravity-Wave
Gravity-Wave

15. Madden-Julian Oscillation (MJO)
(Madden and Julian 1971, 1972)
30 to 90 day tropical weather cycle
No satisfying theory to explain all the features
Difficult for both global climate/operational
models and regional models

16. Dynamical structure
Rui and Wang (1990).

17. Tropical convection organize into large scale Super cloud cluster on Nov 24, 2011
Convection is coupled with a range of dynamic equatorial waves and madden julian oscillations. This is illustrated in the these OLR.
Which I happened to be part of it, more cloud processes. Which can be more specified shown in WK diagram.
24 Nov 2011, Meteo7 IR
This super cloud cluster is Madden-Julian Oscillation (MJO) at its initialization phase
MJO: a global weather phenomenon
Influence rainfall over all the tropics and subtropics

18. Tracking the November 2011 cloud cluster
Outgoing Longwave Radiation 2.5S-2.5N
Madden Julian Oscillation
Kelvin wave
Equatorial Rossby Wave
Convection is coupled
with a range of large scale
dynamic processes:
Madden Julian Oscillation
(5 m/s eastward)
Kelvin waves
(13 m/s eastward)
Equatorial Rossby waves
(Westward propagation)
Convection is coupled with a range of dynamic equatorial waves and madden julian oscillations. This is illustrated in the these OLR.
Interaction between convection and large scale processes is the key.
CAWCB/Bureau of Meteorology

19. 200 hPa Velocity Potential Anomalies (10N-10S)
Gottschalck et al 2013
Red dashed: eastward propagation of upper-level divergence

20. DYNAMO: RMM index
OLR, 850/200 hPa zonal wind (Wheeler and Hendon 2004)

21. DYNAMO: RMM index
Longitude
Longitude

22. DYNAMO: setup of large domain WRF simulations
Continuous free run for > 2 months

23. DYNAMO: large domain cloud-permitting WRF simulations
Longitude

24. DYNAMO: budget of moist static energy
M = cpT+gZ+LQv
Oct 1

25. DYNAMO: Surface fluxes and radiative fluxes

26. Concluding remark
Large scale effects of tropical convection (Q1, Q2) are observable using the sounding network
Column modeling offers insights to both physics (cloud microphysics, radiation, etc.) and dynamics (parameterized large scale dynamic)
Understanding coupling between convection and large scales is the key for tropical dynamics
Convection is crucial for MJO, and responsible for poor simulation and prediction of MJO events in numerical models
High resolution WRF modeling shows some promising results for precipitation