Attribution of extreme climate events in coupled and uncoupled GCM experiments

National Centre for Atmospheric Science (NCAS)

University of Reading

Buwen Dong

Rowan Sutton

Len Shaffrey

Climate event attribution

There is still no consensus about the best methodology for climate event attribution (CEA).

A common approach uses AGCMs forced by prescribed SSTs with and without anthropogenic influences (e.g., Pall et al. 2011, Otto et al. 2012, Christidis et al. 2013, Christidis and Stott 2014 …).

AGCM’s lack explicit ocean-atmosphere coupling, so: -> Are the conclusions from such studies robust?

Pall et al. Nature, 2011

Role of ocean-atmosphere coupling in climate event attribution

Coupled experiments: PD (1994-2011), and EP (1964-1981) Uncoupled experiments: PD and EP (SSTs from coupled experiments).

MetUM-GOML: HadGEM3-A

(1.875° x 1.25°, 85 levels)

coupled to the Multi-Column

K-Profile Parameterization

ocean model (Klingaman et

al. 2011).

Key advantages: • Cheap: < 5% of the cost of the

atmosphere, allowing high (1

metre) ocean vertical

resolution. Small SST biases.

• Disadvantage : Lack of

interactive ocean dynamics.

We test robustness using a “perfect model” approach

Relatively large warming over North and tropical Atlantic, Indian and western tropical Pacific oceans in all seasons. Model simulates many features well.

Seasonal mean SST changes between PD and EP in DJF and JJA

Obs

Model

DJF JJA

Seasonal mean SAT changes between PD and EP in DJF and JJA

Obs

Coupled

Uncoupled

DJF JJA

large warming over NH continent in DJF and over Europe and Africa in JJA. Coupled model simulate these features. Not sensitive to air-sea coupling

Very similar spatially aggregated PDFs in both coupled and uncoupled simulations

Land fraction with very warm winters (>2Ϭ) increased by 15 times while it increased by 25 times for hot summers (> 2Ϭ).

Spatially aggregated PDF of land fraction of SAT anomalies in PD and EP for DJF and JJA

DJF JJA

Obs

Coupled

Uncoupled

Attribution conclusions for SAT changes derived from AGCM experiments generally robust - not sensitive to air-sea coupling.

Contrasting precipitation change over East Asia in JJA.

Seasonal mean precipitation changes between PD and EP for DJF and JJA

Contrasting precipitation features over Australia in DJF.

Coupled

Uncoupled

DJF JJA

Significant differences in several monsoon regions.

Contrasting precipitation features over East Asia and South China Sea in JJA

Seasonal mean circulation and precipitation changes between PD and EP over East Asia in JJA

Local changes in the coupled response bear a similarity to observed changes

Coupled

Uncoupled

Obs

Summary

Attribution conclusions for surface air temperature changes derived from AGCM experiments are generally robust and not sensitive to air-sea coupling (might be exceptions in some regions).

However, changes in precipitation and circulation in some regions show large sensitivity to air-sea coupling, notably in summer monsoons.

Coupled simulations show generally better agreement with observations.

Results demonstrate that the AGCM method has significant limitations and may lead to erroneous attribution conclusions in some regions and situations.

Contrasting precipitation

features over Australia summer monsoon in DJF in

the coupled and uncoupled

responses.

Seasonal mean precipitation changes between

PD and EP over Australia in DJF

Contrasting responses over the Maritime continent and the Coral Sea.

Obs

Uncoupled

Coupled