The Indo-Australian monsoon and its relationship to ENSO in the CMIP5 models & reanalyses

N.C. Jourdain1*, A. Sen Gupta1, Y. Li1, A. S. Taschetto1, C. Ummenhofer2, A. F. Moise3, K. Ashok4

1- CCRC-UNSW/ARC-CoE-CSS, Sydney, Australia ; 2- Woods Hole Oceanographic Institution, MA, USA ; 3- CAWCR-BOM, Melbourne, Australia ; 4- CCCR-IITM, Pune, India

* n.jourdain@unsw.edu.au

Abstract Datasets and indices

Indian monsoon - ENSO

ReferencesJourdain, N.C., Sen Gupta, A., Taschetto, A.S., Ummenhofer, C., Moise, A.F., Ashok, K. (2013). The Indo-Australian monsoon and its relationship to ENSO and IOD in reanalysis data and CMIP3/CMIP5 simulations, Climate Dynamics, in press.

Li, Y., Jourdain, N. C., Taschetto, A. S., Ummenhofer, C. C., Ashok, K. and Sen Gupta, A. (2012), Evaluation of monsoon seasonality and the tropospheric biennial oscillation transitions in the CMIP models, Geophys. Res. Lett., 39, L20713

k

A large spread exists in both Indian and Australian ave-rage monsoon rainfall and in their interannual varia-tions diagnosed from various observational and reanaly-sis products. While the multi model mean monsoon rain-fall from 59 models taking part in CMIP3/5 fall within the observational uncertainty, considerable model spread exists.

Considering all North-Australia rainfall, most models reproduce the observed Australian monsoon-El Niño Southern Oscillation (ENSO) teleconnection, with the strength of the relationship dependent on the strength of the simulated ENSO. However, over the Maritime Continent, the simulated monsoon-ENSO connection is generally weaker than observed, depending on the abi-lity of each model to realistically reproduce the ENSO signature in the Warm Pool region.

Acronym I Institute Spatial Start Resol ReferencesD coverage date

CMAP α UCAR/ NCAR/ CISL/ DSS Global 1979 2.5o Xie and Arkin, 1997GPCP β NOAA/ OAR/ ESRL PSD, Global 1979 2.5o Adler et al., 2003GPCC γ Boulder, CO, USA Global land 1901 0.5o Rudolf et al., 2011AWAP δ BOM, Australia Austr. land 1900 0.25o Jones et al., 2009APHRODITE δ ERTDF, Japan S-E Asia land 1951 0.25o Yatagai et al., 2012TRMM-3B42 v6 NASA/ GIES/ DISC, 50oS -50oN 1998 0.25o Adler et al., 2000TRMM-3B43 v6 ζ USA 50oS -50oN 1998 0.25o Adler et al., 2000HadISST η Met O�ce, Global 1870 1.0o Rayner et al., 2003HadSST2 θ Hadley Centre, UK Global 1850 5.0o Rayner et al., 2006NCEP-NCAR I λ NOAA/ OAR/ ESRL PSD, Global 1948 2.5o Kalnay et al., 1996NCEP-DOE II µ Boulder, CO, Global 1979 2.5o Kalnay et al., 1996NCEP-CFSR π USA Global 1979 0.5o Saha et al., 2010ERA-40 ρ ECMWF, UK Global 1957 2.5o Dee et al., 2011ERAinterim τ ECMWF, UK Global 1979 0.7o Dee et al., 2011JRA-25 ψ JMA/ CRIEPI, Japan Global 1979 2.5o Onogi et al., 2007MERRA σ NASA Global 1979 0.5o Rienecker et al., 2011

40N

20N

0

20S

40S40E 80E 120E 160E 160W 120W 80W

NINO-34DMI a DMI b

AMAR

ISAS

LAUS

LIND

CMIP3 simulations from 24 models: 20c3m scenario (1850-2000)CMIP5 simulations from 35 models: historical scenario (1850-2006)→ at least 3 ensemble members for more than half of the models

Indian and Asian monsoons JJASAustralian and Maritime Continent DJFM

Australian monsoon - ENSO

J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND

−0.5−0.4−0.3−0.2−0.1

00.10.20.30.4

year

0

Jourdain et al., Clim. Dyn. 2013, in press.

J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND−0.4

−0.2

0

0.2

0.4

0.6

0.8

year

0

lag correlationJune-July NINO34 vs lagged NINO34

NINO34 leads LIND leads

lag correl.JJAS LINDvs NINO34

CMIP3 (up-lower quartiles)CMIP5 (up-lower quartiles)APHRODITE-HadISST

Monsoon - ENSO over the Maritime Continent Austral rainfall in CMIP5

J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND

−0.4

−0.2

0

0.2

0.4

0.6

J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND−0.6

−0.4

−0.2

0

0.2

0.4

0.6

J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND

−0.8

−0.6

−0.4

−0.2

0

0.2

0.4

NINO34 leadsAMARleads

AMAR vs lagged NINO34 (partial: DMI removed)

year

0

DJFM

NINO34 leads

DJFM Papuan rainfall vs lagged NINO34 (partial: DMI removed)

year

0

NINO34 leads

DJFM E-C-Indonesian rainfall vs lagged NINO34 (partial: DMI removed)

year

0

CMIP5CMIP3OBS. / REA.GPCC/HadISST

J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND

−0.8

−0.6

−0.4

−0.2

0

0.2

0.4

0.6

0.8

NINO34 leads

DJFM meridian wind stress North of Papua (gray area) vs lagged NINO34

year

0

LAUSE-C-Indonesia

Papua

100E 120E 140E

Equ.

10S

IPSL−CM5A−MR 1.14IPSL−CM5A−LR 1.03

inmcm3−0 1.69ingv−echam4 0.99

ipsl−cm4 0.53MIROC−ESM 1.83bccr−bcm2−0 0.61

ukmo−hadgem1 1.32cnrm−cm3 0.76

GISS−E2−H 1.01GFDL−ESM2G 1.03

giss−model−e−h 0.84GFDL−CM3 0.34

ncar−ccsm3−0 0.60MRI−CGCM3 0.41MPI−ESM−LR 0.98

giss−aom 1.93cccma−cgcm3−1−t63 0.73

cccma−cgcm3−1 0.73mpi−echam5 0.56

miroc3−2−medres 1.67HadCM3 1.15

CSIRO−Mk3−6−0 1.77MPI−ESM−MR 0.73ukmo−hadcm3 1.08

CMCC−CM 0.42giss−model−e−r 0.68

gfdl−cm2−1 0.50miub−echo−g 0.27

mri−cgcm2−3−2a 0.79ncar−pcm1 0.52

csiro−mk3−0 0.94CCSM4 0.11

miroc3−2−hires 1.40HadGEM2−CC 0.31

inmcm4 1.24iap−fgoals1−0−g 1.81

EC−EARTH 0.72FGOALS−s2 0.45

HadGEM2−ES 0.38GISS−E2−R 0.23

ERA40 0.10HadGEM2−AO 0.46

CanESM2 0.96AWAP−HadISST 0.00GPCC−HadISST 0.00

MIROC5 0.68NCEP−NCAR−I 0.08CESM1−CAM5 0.40FGOALS−g2 0.36bcc−csm1−1 0.39

GFDL−ESM2M 0.27NorESM1−ME 0.23

IPSL−CM5B−LR 0.25CNRM−CM5 0.23ACCESS1−3 0.64

FIO−ESM 0.36gfdl−cm2−0 0.59ACCESS1−0 0.23

CESM1−FASTCHEM 0.12CESM1−WACCM 0.19

NorESM1−M 0.17csiro−mk3−5 1.09

NCEP−DOE−II 0.09ERAinterim 0.13

NCEP−CFSR 0.22MERRA 0.14JRA25 0.11

GPCP−HadISST 0.00CMAP−HadISST 0.00

120E 160E 160W 120W 80W

−0.81

−0.56

−0.31

−0.060.06

0.31

0.56

0.81

1.06

1.31

1.56

1.81

2.06

2.31

2.56

2.81

3.06

3.31

(a) correl. AMARDJFM vs NINO34 (b) Equatorial Pac. SST anom.

(b)

−0.95

−0.85

−0.75

−0.65

−0.55

−0.45

−0.35

−0.25

−0.15

−0.05

0.05

0.15

0.25

0.35

0.45

0.55

0.65

0.75

0.85

0.95(a)

year

−2

year

−1

year

0

0.570.570.450.370.300.270.260.240.210.200.170.120.050.020.01

−0.02−0.05−0.08−0.13−0.14−0.14−0.17−0.18−0.19−0.21−0.21−0.24−0.24−0.25−0.25−0.27−0.27−0.29−0.33−0.36−0.36−0.37−0.37−0.38−0.38−0.40−0.42−0.43−0.44−0.45−0.47−0.52−0.53−0.53−0.56−0.56−0.57−0.58−0.59−0.59−0.60−0.60−0.61−0.64−0.65−0.66−0.67−0.71−0.73−0.74−0.77−0.77−0.78−0.82−0.84

M J S D M J S D M J S D M J S D

Surface warming of the equatorial Paci-fic associated with El Niño events drives an intensification of the Walker circula-tion. As this circulation has its sinking branch located near the Australian region, El Niño events tend to weaken average deep convection, i.e. rainfall, over Australia. Correlation between LAUS (Australian monsoon) and simul-taneous NINO34 indeed reaches -0.45 in AWAP/HadISST.

It appears that the ENSO-monsoon rela-tionship is weaker in the 7 reanalyses than in any of the observational products. NCEP-NCAR-I, NCEP-CFSR and MERRA are quite close to the observations, but the other pro-ducts show significantly weaker correlations than observed, in particular NCEP-DOE-II (r=0.21, vs. r=0.45 in the observations). Because of the complex mixture between modelling and data assimilation, it is diffi-cult to identify the origin of these differences.J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND J FMAMJ J ASOND

−0.5

−0.4

−0.3

−0.2

−0.1

0

0.1

0.2

0.3 LAUS vs lagged NINO34

year

0

LAUSleads

DJFM

lag correl.

- Based on the multi-model means, the CMIP3 and CMIP5 simulations generally reproduce the lagged ENSOmonsoon relationship, even if the instantaneous anti-correlation is weaker than in the observations (r = -0.28 in CMIP3, and -0.33 in CMIP5).

- 3/4 of the 24 CMIP models that are in the range ±0.1 from the observed correlation are from the CMIP5 ensemble.- The models that exhibit a short persistence of NINO34 auto-correlation are the models producing the weakest ENSO-monsoon rela-

tionship (linked to too many aborted El Niño or La Niña events? Guilyardi et al. 2003). - The strength of the ENSO-monsoon relationship also depends upon the amplitude of ENSO itself, Considering the CMIP models only,

we find that 35 % of the intermodel variance in DJFM LAUS-NINO34 correlation is explained by the amplitude of ENSO.

Based on the multi-model mean, the LIND-NINO34 correlation at zero lag is significantly underestimated by the CMIP5 models and even more so by the CMIP3 models.

The CMIP lag correlation has a different shape

than the observed lag correlation. In the observa-

tions, NINO34 anomalies in June–July are rela-

ted to a developing ENSO event rather than to a

terminating event (see observed correlation

between NINO34 in June–July and NINO34 in

December prior to the monsoon is only 0.15, whe-

reas the correlation between NINO34 in June-

July and NINO34 in December following the

monsoon is as high as 0.80. This asymmetry is

much less marked in the CMIP simulations,

Strong monsoon-NINO34 anti-correlation when considering Austra-lia and the Maritime Continent together (r~-0.84 with CMAP and GPCP), cf results from OLR by McBride et al. (2003). However, there is no corrrelation over most of Indonesia. In fact, the strong anti-correlation is largely related to rainfall over the ocean. The CMIP models are not able to capture this strong correlation. None of them is as much correlated to NINO34 as in the observations.

partial correlation

Part of the discrepancy is related to a significant positive relationship between ENSO and rain-fall over Papua in most of the CMIP5 models that does not exist in the observations, i.e. there is more rainfall over Papua during El Niño events in the CMIP models.

CMIP3 (up-lower quartiles)CMIP5 (up-lower quartiles)AWAP-HadISSTOBS/REA.

The monsoon southward flow anomaly related to El Nino is well repro-duced in the CMIP models. However, to the north of Papua, SSTs areanomalously warm in most of the models where the observed El Nino response is a strong cooling, which leads to an unrealistic moisture convergence over the Papuan region in the models and a positive rain-fall anomaly. Although the SST response to ENSO is generally impro-ved in the CMIP5 models, the rainfall-ENSO relationship in Papua is not better in CMIP5 than in CMIP3, because stronger SST bias in CMIP3 is partly compensated by stronger bias of the meridian circula-tion anomaly. In view of the correspondence between the ranking of correlation coefficients and SST response to ENSO in the Equatorial Western Pacific, it appears that the SST bias not only affects Papua, but the entire Maritime Continent.

Regional climate modelling

10N

0

10S

20S

70E 90E 110E 130E 150E 170E

18

16

14

12

10

8

6

4

2

0A

CCES

S1-0

CanE

SM2

CCSM

4FG

OA

LS-s

2

FIO

-ESM

Had

GEM

2-A

O

DJFM rainfall (TRMM-3B43) + ERAinterim winds historical CMIP5

WRF-NEMO

Tropical belt con�guration developed at LOCEAN-IPSL

- ¾ degree resolution- 2-way nesting ¼ degree

North & South boundary conditions :- ERAinterim- NEMO global ¼ (forced by ERAinterim)

2-way nesting

Now running on vayu (NCI)