LC1LC2 Cyclones in the Unified Model

1
LC1/LC2 Cyclones in the Unified Model

• Ian Boutle
• Thanks to Richard Forbes Victoria Sinclair

2
Talk Outline

• What are LC1 LC2
• Motivation for simulating them in the UM
• Problems to overcome and how we solved them
• Comparison of the new simulations with previous
  results
• What this new tool can give us

3
IGCM History

• Primitive Equation, Spectral model of Hoskins
  Simmons (1975)
• Simulates 1/6 of a hemisphere, then reflects this
  12 times to make the whole globe.
• Adds fastest growing normal mode at wavenumber 6
  to basic state, to simulate baroclinic instability

4
Basic States
LC1
LC2
Typical zonal mean state of Mid-latitude jet and
potential temperature distribution
Barotropic shear added, confined to the bottom
half of the troposphere
From Thorncroft et al (1993)
5
LC1 Cyclones

• Predominantly anticyclonic wave breaking
• Strong cold front

• Cyclones move north, anticyclones move south
• Decay after 10 days

6
LC2 Cyclones

• Predominantly cyclonic wave breaking
• Strong warm front

• Not much poleward movement
• Doesnt really decay

7
Motivation

• Much work done with these classic cyclone
  examples
• Recent focus on more complex systems effects of
  friction/moisture/radiation
• IGCM adapted by many users over time simple
  boundary layer scheme added, moisture radiation
  parameterisations
• But

8
Motivation 2

• UM has state of the art parameterisations for
  all these areas
• Complex boundary layer scheme of Lock et al
  (2000), with realistic treatment of boundary
  layer cloud and precipitation
• Complex cloud microphysics, radiation and
  convection schemes
• Much scope for investigating LC1/LC2 in a more
  real atmosphere

9
Problems

• UM doesnt do 1/6 of a hemisphere either
  global model or rectangular limited area
• Limited area normally used with forced boundaries
  (eg. NAE) or bi-cyclic boundaries for idealised
  studies
• Needed to implement East-West cyclic boundary
  conditions, with North South boundaries fixed
• Limits domain to be a Cartesian f-plane

10
Limited area domains
IGCM

• IGCM had 600 spherical geometry
• Easy to have jet centred on 45N with wrap around
• UM has square domain imposed on spherical
  geometry
• Cannot have wrap around at 45N

600
UM
45N
11
Problems 2

• UM solves full deep atmosphere equations not
  easy to get balanced initial conditions
  analytically
• Specify wind field, balance temperature with
  thermal wind balance, then pressure with
  hydrostatic balance
• Run model for an hour to allow adjustment to
  models own balance

12
Problems 3

• How do we calculate the fastest growing normal
  mode?
• Polvani Esler (2007) provide an alternative
  method perturb initial temperature field

13
Initial Conditions
LC1
LC2
14
Basic States
LC1
LC2
Typical zonal mean state of Mid-latitude jet and
potential temperature distribution
Barotropic shear added, confined to the bottom
half of the troposphere
From Thorncroft et al (1993)
15
LC1 Simulation

• MSLP
• Evolves in a similar fashion to IGCM
• Similar highs lows
• Dont move as far north/south
• Temperature
• Similar cold front domination, but perhaps more
  wrapped up
• Lags IGCM by 2 days in development

16
LC2 Simulation

• MSLP
• Evolves in a similar fashion to IGCM
• Very strong lows
• Does start to decay eventually though
• Temperature
• Similar warm front domination
• Lags IGCM by 2 days in development again

17
TEKE Comparison
2 day lag for LC2
4 day lag for LC1
18
Differences

• Less poleward movement of LC1
• More wrapped up LC1
• Qualitative 2 day lag behind IGCM
• Quantitative 4 day lag for LC1
• Quantitative 2 day lag for LC2
• LC2 does decay eventually

19
Why the difference?

• Balasubramanian Garner (1997) investigated the
  effects of spherical geometry on LC1 cyclone.

• LC1 cyclone initiated in IGCM-type model, full
  PV field inverted in Cartesian geometry and run
  in grid-point model.
• Found 2-day delay in peak of TEKE.

20
Attributed to smaller meridional movement of lows
in Cartesian geometry the flow zonalises much
quicker in spherical geometry (hence deviation
from the mean is less). This is a purely
geometric effect.
21
Their LC1, day 9
My LC1, day 11
Both at TEKE peak and lowest MSLP
22
Differences

• Less poleward movement of LC1
• - Geometric Effect
• More wrapped up LC1
• - Geometric Effect
• Qualitative 2 day lag behind IGCM
• Quantitative 4 day lag for LC1
• - Extra 2 days is Geometric Effect
• Quantitative 2 day lag for LC2
• LC2 does decay eventually

23
Upper Level Analysis

• UM is an f-plane, whereas IGCM has full Coriolis
  variation

PV on 335K, initial condition
Should continue increasing here
Approx is ok here
Should tend to zero here
24

• LC1, IGCM day 7
• Big streamer of high PV moved from Pole into
  midlatitudes
• Lots of PV to feed off reinforce structure

25

• LC2, IGCM day 7
• More wrapped up than LC1
• Able to move more high PV southwards

26

• LC2 ends with completely cut-off upper level PV
  anomaly
• The prevents the decay of the surface cyclone
• This isnt possible in UM, hence cyclone decays

Day 15
27
Differences

• Less poleward movement of LC1
• - Geometric Effect
• More wrapped up LC1
• - Geometric Effect
• Qualitative 2 day lag behind IGCM
• Quantitative 4 day lag for LC1
• - Extra 2 days is Geometric Effect
• Quantitative 2 day lag for LC2
• LC2 does decay eventually
• - Due to f-plane approximation

28
What does the model give extra?
Why is the boundary layer deep here? Air is
moving East-West only, should be the same
temperature as sea surface. no buoyancy to grow
boundary layer! Must be shear-induced turbulence
(wind shear is capable of mixing air turbulently
higher than would otherwise happen due to
buoyancy effects.)
LC1, Boundary layer depth
29
Moisture
Cloud structure around low Cloud and rain on
warm conveyor belt Some convective rain in high
pressure
Day 9
30
Conclusion

• LC1/LC2 represent 2 classic idealised cyclones,
  used in many studies
• Similar cyclone types now available in Idealised
  UM
• Clear why differences exist
• The cyclones look realistic
• Added functionality of the UM allows for new
  areas of research