Coupled and Uncoupled Model Simulation of the - PowerPoint Presentation

Slide1

Coupled and Uncoupled Model Simulation of the

Global ENSO-TC Teleconnection

Ray Bell With thanks to Kevin Hodges, Pier Luigi Vidale, Jane Strachan and Malcolm Roberts

@

RayBell_MetSlide2

Introduction

Motivation

It

is important to evaluate the ability of GCMs to simulate

realistic ENSO associated TC

teleconnections for seasonal forecasting and before predictions are made for TCs

and climate change using GCMs (Mori et al. 2013). Lack of studies

on simulated

ENSO-TC teleconnection worldwide

.

Research Objectives

What are the advantages and shortcomings of the simulated ENSO-TC teleconnection?

Are the key mechanisms well represented in the simulation?Slide3

Previous work

Observations

The influence ENSO on TC activity has long been observed (NATL, Gray 1984; WPAC, Chan 1985; AUS, Nicholls 1975)

Reanalysis

NCEP (‘50-’05)

GPI.

Camargo (2007)Slide4

Previous work

GCMs

First research by Wu and Lau (1992) at resolution R15.

Understanding simulation of the ENSO-TC teleconnection in the NATL Shaman and Maloney, 2011: CMIP3 GCMs coarse resolution (~2

ox 2

o)Impacts of

ENSO on

Caribbean vertical wind shear were the most poorly simulated

Murakami and Wang, 2010: AGCM 20 km

C

aptured

the broad

tropical cyclone

response to ENSO in the North AtlanticRecent AOGCM students: Iizuka and Matsuura (2008; 2009) ; Kim et al (2013); Wang et al (2013)Slide5

Idealised GCM simulation

HiGEM

UK’s

High-Resolution

Global Environmental Model (

Shaffrey

et al

, 2009

)

HiGAM

:

AMIP (atmospheric model forced with observed SST and sea ice) 1979-2002

(

Strachan

et al

, 2013

)

HiGAM with HiGEM SST

: atmosphere component forced with coupled SST

1.25

o

x0.83

o

, ∆x50N = 90 km

1/3

o ocean model

HiGEM 1.1 Present-day integration

150 yrs

N144

High-resolution climate models:

Improved representation of

ENSO (Guilyardi et al.

2009)

Improved mean-state TC climatology (Strachan et al. 2013)

Improved ENSO-teleconnections (Dawson et al. 2012) Slide6

Tracking Algorithm (TRACK)

A 20 year

time-slice

of GCM simulated tropical storms

Bengstton

et al

(2007);

Strachan

et al

(2013)

1

)

Locate and track all

centres

of high relative

vorticity

 35000/yr

2

)

Apply a 2-day filter to the tracks

 8000 storms / yr

3) Analyse vertical structure of storm for evidence of warm-core (tropical storm structure)  120 storms / yr

Genesis through to LysisSlide7

ENSO Composites

Niño-3.4 normalised SST anomaly for DJF >1

El Niño

; DJF <1

La Niña

NH TCs prior to event (correlation of 0.9 with ASO

SSTa

)

IBTrACS (1979-2010); ERA-Interim (1979-2010)

7

El Nino years

(82-83, 86-87, 91-92, 94-95, 97-98, 02-03, 09-10)

6

La Nina years

(84-85, 88-89, 98-99, 99-00, 07-08, 10-11)HiGEM (150 years): 31

El Ninos; 25 La

Ninas

HiGAM AMIP (1979-2002) :

6

El

Ninos

;

4 La NinasSimulation of HiGEM’s ENSO is in Shaffrey et al (2009)Slide8

HiGEM

: ENSO simulation

Shaffrey

(2009)Slide9

Present-day TC Climatology

See Strachan et al (2013): assessment of HiGAM interannual variability and

intraseasonal

variability.

Bell et al (2013): evaluation of HiGEM and response to climate change

HiGEM

:

Good distribution of TCs

Too many in SH (strong SPCZ)

Too few in

NAtL

(cool SST and strong VWS)

Lack of recurvate in WNP (cool SST) Slide10

ENSO-TC Location

Match well

Over pronounce

variabilitySlide11

ENSO-TC Location

Match well

Over

pronounce

variabilitySlide12

ENSO-TC Frequency

Greater variability in ERA-Interim vs IBT

Small var. of NATL TCs

Small var. and wrong sign of NATL TCsSlide13

ENSO-SST

SSTs extend

too far

west.

Lack of

warming near

Peru coast.

(see

Guilyardi

e

t al, 2009Slide14

ENSO-

ppt

Stronger

signal

No

teleconnection

In CaribbeanSlide15

ENSO-Walker circulation

Convection

shifts

Westward as

SSTs extend

f

urther west

Lack of

upper-level

westerliesSlide16

ENSO-vertical wind shear

Over

Pronounce in

WPAC

Lack of VWS

dipole

Di-pole

partly explain

See-saw of

Activity

NATL-

EPAC (

aiyyer

&

Thorncroft

,

2006) Slide17

ENSO-relative vorticity

Good

s

patial

pattern

Weaker.

No change

around

C.AmericaSlide18

ENSO-upper level circulation

Divergence

not

constrained

To CPAC

Causes

inc.

TC

In NATL during

El Niño Slide19

Thermodynamic vs. Dynamic influences:

North Atlantic

Niño-3.4 SST

VWS

Vor

850

-

ω

500

RH

700

Cool trop SST

b

ias in

HiGEM

TC

Count

Too high

m

ean-state

VWS

RH is well

c

aptured in

HiGAM.

VWS is

most Important(Carmargo, 2007)

JASO275-340oN, 10-20o

NAverageSlide20

Thermodynamic vs. Dynamic influences:

Western

North Pacific

VWS

VWS is not important

Vor

850

-

ω

500

RH

700

Large mean ascent

c

auses

t

oo many

TCs

TC

Count

Vor

. explains

ENSO var.

(

Carmargo

,

2007) Slide21

Discussion

Errors in atmospheric teleconnections stem from: mean-state SST biases; spatial pattern of ENSO associated SST.

ENSO-TC teleconnection is good west of PAC in HiGEM

Larger variability of ENSO-TC teleconnection in HiGAM without presence of air-sea feedbacks: Larger source of heat and stronger circulation

El Niño SST

s

hifted

west

Walker circulation

Shifted westward

Upper-tropospheric westerlies do not reach into NATL

No change in VWS

No change in TCs (HiGEM)Slide22

Conclusion

HiGEM captures the observed ENSO-TC teleconnection in the Pacific and Indian Oceans. HiGEM does not capture the response in the NATL

NATL

: mean-state VWS is most important factor. Once VWS is low enough other parameters can influence (seen in HiGAM)

WNP

: mean state -

ω

500

biases and low level

vor

. most important for

ENSO-TC

Bell et al

(

2014)

Simulation

of the global ENSO-Tropical cyclone teleconnection by a

high resolution

coupled general circulation model,

JCLIM

(in review)Bell et al (2013): TCs and climate change including TC intensity projections from coupled and uncoupled simulations, JCLIM Slide23

Conclusion

ENSO-TC

teleconnection in a 150-year present-day high-resolution AOGCM. Limitations of atmospheric component

-

AMIP experiment.

HiGEM captures the observed ENSO-TC teleconnection in the Pacific and Indian Oceans. HiGEM does not capture the response in the NATL

NATL

: mean-state VWS is most important factor. Once VWS is low enough other parameters can influence (seen in HiGAM)

WNP: mean state -

ω

500

biases and low level

vor

. most important for

ENSO-TC

Bell, R. J., Strachan, J., Hodges, K. I., Vidale, P. L. and Roberts, M. (

2014)

Simulation of the global ENSO-Tropical cyclone teleconnection by a high

resolution coupled general circulation model,

Journal of Climate

(in review)