Tropical Convection: A Product of Convergence

1
Tropical Convection A Product of Convergence
2
But What Drives Convergence?

• ONE THEORY CISK
• Conditional Instability of the Second Kind
• A Positive Feedback Mechanism . . .

3
(No Transcript)
4
CISK Convergence Driven by LH Release Aloft
Is this the Whole Story?
5
Other Process.

• Barotropic Instability
• Sea Surface Temperature Gradients
• (Lindzen and Nigam)
• All processes play a role to some extent

6
But how do they compare?

• General Circulation
• Conv driven by upper-level Div
• Local Circulation
• Conv driven by SST gradient

7
Basic Hypothesis

• -Momentum Balance of Hadley circulation aloft
  does not account for total low-level moisture
  Convergence
• -SST directly influence Convection apart from
  thermodynamic properties
• -Variation or Gradient in SST pattern important
  for Convection
• In Tropics
• Small Changes Large Influence

8
Environment of Tropical Ocean
9
Basic Approach/Methodology

• ATS capped at 700mb (height of inversion)
• Inversion decouples upper ATS from below
• No influence from LHR in cumulus towers (CISK)
• Convergence in lower layer driven by SST Gradient
• Pressure Gradient
• Well Mixed BL SST and gradients
  correlated in vertical
• Model Eddy (anomalous) surface flow
• Zonally averaged flow well represented by Hadley
  Circulation
• Compare model with observational data (FGGE) in
  order to determine relative importance of
  low-level forcing in eddy convergence

10
Model Development

• Vertical Temp structure of BL linear function of
  SST
• Flow in Boundary Layer Incompressible
• Given Temp Density Pressure via
  Hydrostatic Eq

11
Momentum Equations Balance of PGF, Coriolis,
Friction

• Zonal Component
• Coriolis PGF Turbulent
  Stress (friction)
• Meridional Component

12

• Compute Eddy SLP from Observed temperature using

13
Initial Results

14
Major Approximation/Error

• -Lindzen Nigam assume top of Boundary Layer
• (taken to be 700mb or 3km) is flat and does not
  vary
• in time
• -Convection occurs instantaneously
• -These simplifications are later revised in order
  to
• Get realistic flow pattern in the model
• (back-pressure effect)

15
Back-Pressure adjustment

• -In original model, BL (700mb sfc) is a rigid sfc
  that cant be modified
• -In reality, vertical motion above SFC LOW raises
  the top of the BL (700mb sfc) and this adiabatic
  expansion acts to cool the lower tropopause
  raises pressure Negative
  feedback
• -This cooling is eventually dampened by ample LHR
  
• But it takes time for convective clouds to
  develop (30mins)

16
2 Major New Variables Introduced

• Deviation of 700mb layer from flat 3km sfc
• Proportional to uptake of mass via
  convergence
• Proportional to cooling of tropopause
• If large cooling offsets warm SST
  Convergence suppressed
• Time Scale Cloud development time
• Represents adjustment time of ATS to reach
  steady state
• If small, LHR quickly compensates cooling from
  h
• Convergence excessive

17
Revised Equations in Model

• Allows for modulation of 700mb sfc with upward
  vertical motion
• variation in top of BL

18
Note new variables directly proportional to each
other

• time scale
  conv/div

19
New Solutions

• If tau30s looks like old model
  (excessive convergence)
• If tau3hrs Weak to no convergence (Big
  back-pressure)
• If tau30mins resembles flow from real
  data

20
Solution with tau30mins
21
Both Gradients Important

• Forcing from Meridional
• -Represents ITCZ better
• Forcing from Zonal
• -Represents SPCZ better

22
(No Transcript)
23
Criticisms/Notes

• Questionable parameterizations
• -3km can be considered too high for mean
  Boundary Layer
• -Time Adjustment of 30 mins chosen b/c it looks
  the nicest
• (No theoretical Justification)
• Poor Results for NH Winter
• -Boundary Layer is shallower
• -Greater influence from motions aloft
• Are Results repeatable
• -How does model compare against other reanalysis
• and data sets (future work)
• Conceptual Problem

24
Inherent Ambiguity What drives what?Low level
vs. Upper Level

• SST gradient Pressure gradient
  Low-level flow
• (Lindzen Nigam)
• Deep Convection/LHR Pressure gradient
  Low-level flow
• (Gill others)
• Different Forcing can yield similar results
• Each Mechanism only valid given assumptions made