Jennifer Catto

1
Extra-tropical cyclones and Storm Tracks

• Jennifer Catto
• Supervisors Len Shaffrey and Kevin Hodges

2
Introduction

• Higher resolution atmosphere models should be
  able to represent structures of storms better
  e.g. fronts.
• Will compare HiGEM, HadGEM and ERA-40
• would like to look at ECMWF interim reanalysis in
  the future.
• Using Kevin Hodges tracking program and some
  case studies to do this.

3
Feature tracking

• Using Kevin Hodges Feature tracking method
  described in Hoskins and Hodges (2002)
• Fields filtered to T42 and background field -
  (wavenumber 5) removed
• Once tracks are found they are referenced back to
  full resolution for further analysis

• Example of storm tracks pressure tracks
  generated within 75W to 65W and 30N to 40N.

4
Vorticity (850hPa) Tracking Statistics I - Track
Density
(Cyclones per month per 5 spherical cap)

• Vorticity preferred for tracking as it picks up
  more small scale features and is not an
  extrapolated field.
• Both models capture main features of storm tracks
  including the Mediterranean track.
• Both too strong in some places especially HiGEM
  in the Pacific

5
Vorticity (850hPa) Tracking Statistics I - Track
Density Errors
(Cyclones per month per 5 spherical cap)

• Quite different error patterns between models in
  Atlantic
• Track density too strong near Greenland
• Big differences in Mid-East Pacific - Tracking of
  trailing cold fronts?

6
Vorticity (850hPa) Tracking Statistics II -
Genesis Density
(Cyclones per month per 5 spherical cap)

• Both models capture main features of genesis
  including Mediterranean cyclogenesis.
• Overestimated genesis in mid-Pacific in HiGEM
• Underestimated genesis in mid-Atlantic in both
  models.

7
Summary so far

• Tracking statistics show that both models do a
  reasonable job representing the storm tracks.
• Large differences between the models HiGEM
  generally produces more storms.
• Have looked at atmosphere only runs, SSTs,
  baroclinicity.
• Want to investigate differences further by
• looking at individual cyclones in the models
  case studies
• investigating the mechanisms producing these
  differences PDFs of cyclone variables

8
Case Studies - MSLP
HiGEM1.2
HadGEM1.2
Mean Sea Level Pressure (hPa)
9
Case Studies - vorticity
HiGEM1.2
HadGEM1.2
850hPa Vorticity (x10-5s-1)
10
Maximum intensity vorticity on vorticity tracks

• HiGEM storms a lot more intense than HadGEM.
• More high intensity storms in the Atlantic than
  Pacific

11
Maximum intensity pressure on vorticity tracks

• Smaller difference between the two models what
  we expect as vorticity is more sensitive to
  scale.
• Indicates that storms may be of smaller scale in
  HiGEM.

12
Growth Rates
Pressure tendency
Vorticity tendency

• Some cyclones of larger deepening rates in HiGEM
  than HadGEM
• Difference more obvious in vorticity tendency
  many more rapidly intensifying cyclones

13
Conclusions so far

• HiGEM produces more storms generally than HadGEM
  shown by track density and genesis density.
• There are more extreme high intensity storms in
  HiGEM than HadGEM.
• Storms in HiGEM are smaller scale and more
  coherent as can be seen in the case studies.
• It is possible that there is more secondary
  cyclogenesis in HiGEM than HadGEM.

14
Future Work

• Investigate mechanisms further
• Spatial analysis
• Latent heating - vertical velocity and
  precipitation distributions
• Deformation strain
• Greenland tip-jets.
• Atmosphere only model runs.
• ECMWF interim reanalysis
• Teleconnections.
• Role of coupling to the ocean.
• Climate change run (1 CO2 increase per year) of
  HiGEM1.2

15

• Different kinds of secondary cyclones from
  Parker (1998).

• An early model of a secondary cyclone forming
  from a wave on the cold front of a primary
  cyclone from Parker (1998) after Bjerknes and
  Solberg (1922)

16

• Surface analyses for 12 UTC 25-28 April 1992
  IOP 3 from FASTEX from Parker (1998) after
  Hewson (1993).

17
Vorticity Tracking Statistics IIGenesis Density
Errors
18
Pressure Tracking Statistics ITrack Density
(Cyclones per month per 5 spherical cap)

• Less cyclones overall as pressure tracking only
  picks up larger scales noticeable in
  Mediterranean especially.

19
Pressure Tracking Statistics ITrack Density Error
(Cyclones per month per 5 spherical cap)

• Similar error patterns to vorticity tracking
• Large difference in Mid-East Pacific so
  probably not tracking trailing cold fronts.

20
Pressure Tracking Statistics IIGenesis Density
(Cyclones per month per 5 spherical cap)

• Differences around Japan sea ice?
• More secondary cyclogenesis in HiGEM than HadGEM?

21
Pressure Tracking statistics IIGenesis Density
Error
(Cyclones per month per 5 spherical cap)
22
Case studies vertical velocity
700hPa vertical velocity (Pas-1)
23
Eady Growth Rate Maximum- A measure
of the baroclinicity
where
Mean growth and decay rate from ERA-40 from
850hPa vorticity
Eady Growth Rate for ERA-40 calculated at 700hPa
24
Eady Growth Rate 850hPa
HadGEM1.2-HiGEM1.2
25
Sea Surface Temperature
ERA-40
26
Sea Surface Temperature
27
Regions of interest
Atlantic
Pacific
Mid-Pacific
Mid-Atlantic
Gulf Stream
Kuroshio Current
28
Maximum intensity vertical velocity

• Values found taking the minimum within a 10 area
• HiGEM upward velocities stronger than HadGEM
• Values slightly higher in Pacific than Atlantic

29
Vertical Velocities - Sensitivity to sampling
method
30
Growth Rates Gulf Stream Kuroshio Current
Pressure tendency
Vorticity tendency

• Expect to see larger differences in smaller
  regions.
• Vorticity tendency much higher in HiGEM in both
  Gulf-Stream and Kuroshio current regions.
• Unusual differences in pressure tendency for
  Gulf-Stream.
• Slightly higher number of rapidly deepening
  cyclones in HiGEM in Kuroshio current region.

31
Secondary cyclones

• Secondary cyclones
• Develop in smaller local baroclinic regions (e.g.
  trailing cold fronts)
• Horizontal scales lt 1000km
• Can develop explosively

• Other factors influencing genesis and development
  of frontal cyclones
• Latent Heat Release
• Deformation strain and stress

32

• Surface pressure analysis for 18 GMT 21st Jan
  1995. Showing parent cyclone LA and a wave on
  its cold front, LU.

33
Regions of interest
Atlantic
Pacific
Mid-Pacific
Mid-Atlantic
Gulf Stream
Kuroshio Current
34
Growth Rates Mid-Atlantic Mid-Pacific
Pressure tendency
Vorticity tendency

• Much greater number of storms in the Mid-Pacific
  region in HiGEM than HadGEM seen previously in
  track density
• Same number of Mid-Atlantic storms in the models
  much higher vorticity tendency in HiGEM.