Past ENSO : observations and simulations with the IPSL model for the Holocene and the last Millennium P. Braconnot, Laboratoire des Sciences du climat et de l’envrionnement, Gif-sur-Yvette, France With input from different collegues
Jan 03, 2016
Past ENSO : observations and simulations with the IPSL model for
the Holocene and the last Millennium
P. Braconnot, Laboratoire des Sciences du climat et de l’envrionnement,
Gif-sur-Yvette, France
With input from different collegues
High resolution proxy records to infer ENSO variability
From T. Correge, G. Cabioche, C. E. Lazareth, M. CarréS. Tudhope, M. Helliot
Massive corals
ENVIRONMENTALRECORDS FROMMASSIVE CORALS:
Annual banding in coral cores provides ~monthly resolution
over lifespan of the coral
ENVIRONMENTALRECORDS FROMMASSIVE CORALS:
Annual banding in coral cores provides ~monthly resolution
over lifespan of the coral
Chemical analyses of massive corals: A window on tropical climate
S. Thudope
T. Correge
~8000 yr old porites coral
Tracers in corals: Sr/Ca
8,8
8,85
8,9
8,95
9
9,05
9,1
9,15
9,2 21
22
23
24
25
26
27
28
1993 1994 1995 1996 1997 1998 1999 2000
Sr/
Ca
(m
mo
le/m
ole
)
SS
T (°C
)
Year
21
22
23
24
25
26
27
288,8 8,85 8,9 8,95 9 9,05 9,1 9,15 9,2
y = 169,26 - 16,092x R= 0,95994
SS
T (
°C)
Sr/Ca (mmole/mole)
Calibration of Sr/Ca vs SST in a live coral from New Caledonia
Sr/Ca appears to be a robust SST tracer ("proxy")
from Correge
(Tudhope et al., unpublished)
-4 -3 -2 -1 0 1 2 3 4sea surface tem perature anomaly (degrees C)
NINO3.4 regionJarvis
-4 -3 -2 -1 0 1 2 3 4sea surface tem perature anomaly (degrees C)
NINO3.4 regionJarvis
r2 = 0.62 for monthly data
Jarvis coral 18O and NINO3.4 SST
Marine bivalves
ENSO strength from corals and bivalves
0.00
0.05
0.10
EN SO'strength '
(s.d.,‰ 1 8 O )
130 ± 2 ka85kaM O D ER N 2-3 ka
6.5ka 38 - 42 ka
112ka
118-128ka
H95
-64*
DT
91-7
Dur
-2
M93
TB
FC
Lain
g F
C2
H95
-16
H96
-36*
H96
-18
H96
-6
H96
-27
H97
-10*
H95
-14*
H94
-2*
H95
-58*
H95
-18
H95
-25*
H95
-6
AG E (Tudhope et al, 2001)
From Living and fossil corals
Comparison with bivalvesCabioche and Correge
0 20 40 60 80 100 120 140 160
20
0
-20
-40
-60
-80
-100
-120
-140
sea-level curve from deep sea core data (Shackleton, 1987)
thousand yea rs be fo re presen t
sea-
leve
l, m
‘GLACIAL’‘INTER-
GLACIAL’‘INTER-
GLACIAL’‘GLACIAL’
tim e intervals from which multi-decadal coral cores have been collected and analysed
Paleo-ENSO from mollusk shells
• Ex East Pacific
R=0.74
TC
alla
o
Annual Niño1+2 index
50-70% reduction of ENSO activity
M. Carré
Multi proxy reconstruction for the early to mid Holocene
• Need to combine different proxy to reconstruct mean-state, seasonality and variability
• Variability reduced in Early to mid-Holocene• Link with the insolation forcing through Bjerknes feedback (Clement
et al. 1999, 2000, etc…) or monsoon activity (Liu et al. 2003) • La Niña type mean state ?
From literature : Braconnot et al. subm
ENSO in PMIP 6ka simulations
Almost all simulations show reduced ENSO The increased monsoon counteracts the development of ENSO Case where the change in large scale dynamics dominate the signal different from future or LGM where thermodynamics seems more important.
Zheng et al 2008
Niño 3
ENSO reduction underestimated? Pb model? Pb proxy interpretation?
Mid and Early HoloceneResponse to insolation forcing and
fresh water fluxes
Results from Yihu Luan’s PhDCollaboration S. Brewer, W. ZhengBraconnot et al. Subm.
Early vs mid-Holocene?
• Differences in the response depending on precession?
Insolation 6ka-0ka Insolation 9.5ka-0ka
Latit
ude
Latit
ude
> 20 W/m2 > 25 W/m2
- 6 kyr
Présent
- 9.5 kyr
- 6 kyr
Présent
- 9.5 kyr
From Berger 1978
Sum
mer
sol
stic
e
Sum
mer
sol
stic
e
Simulations with IPSLCM4 OAGCM
Oceans26%
2.3 Pg y-1
Oceans26%
Oceans26%
2.3 Pg y-1
LAND ATMOSPHERE OCEAN & SEA I CE
Oceans26%
2.3 Pg y-1
Oceans26%
Oceans26%
2.3 Pg y-1
LAND ATMOSPHERE OCEAN & SEA I CE
water & energy carbon cycle
Land29%
2.6 Pg y-1
Land29%Land29%
2.6 Pg y-1
CO2
carbon cycle
Land29%
2.6 Pg y-1
Land29%Land29%
2.6 Pg y-1
CO2
Aerosols & chemistry
chemistry
Aerosols & chemistry
chemistry
•Ocean-atmosphere-sea-ice-land surface•Same version as in AR4 and PMIP2•Resolution atm : 96x72xL19•Ocean : 2° (refined at eq) xL31
Simulation CTRL 6ka 9.5ka
Length (yrs) 1000 650 300
Test to insolation forcing only
Marzin and Braconnot 2009
Role of fresh water in the North Atlantic?
• Fresh water flux on present day mean state enhances ENSO in a multimodel ensemble simulations (Timmermann et al. 2007)
• Test for 9.5ka and 6ka using IPSL-CM4 and a crude representation of interactive ice-sheet melting with climate
Following Swingedouw et al. 2006
Simulation 6kawF 9kawF
Length (yrs) 800 280
Test to insolation forcing and fresh water forcing
Ice-sheet Fresh water interactive with climate, uniformly distributed over the ocean > 40°N in Atlantic and Arctic
Change in DJF interannual variability
SST : standard deviation
0k
6k
9.5k
Insolation Insolation +wF•Variability larger in model than in observations
•Variability increases from 9.5k to 0k
•Fresh water flux in NA enhances variability
Change in seasonality
Niño 3 box : annual mean cycle
Wind and precipitation changes in July-August
mm/d
Insolation
Fresh water
Braconnot et al. Subm. Luan et al. In prep
ENSO composites Typical evolution in Niño3 box
Rms computed over this periodSST isolinesPr : color
Months Yr previous Yr after
Ex: role of EH insolation
Niño
Niña
Normal= seasonal cycle
mm/d
Relative changes induced by changes in seasonality, El-Niño and La Niña on SST
Braconnot et al. Subm.
The last millenium: ENSO response to volcanic forcing
From M. Khodri
IPSL-CM4 ModelIPSL-CM4 Model
Calendar time
Gao et al. 2008Khodri et al. In prep
Influence of external (volcanic and solar) forcing on SST? Influence of external (volcanic and solar) forcing on SST?
Warm Medieval Period Warm Medieval Period (≈ 950-1250 Year A.D.)(≈ 950-1250 Year A.D.)
simulated cooling in Niño 3 in respone to volcanic forcing
Persistant cooling in data (multi-year La Niña or La Niña like state?
Transition to LIA Transition to LIA (≈1400-1700?)(≈1400-1700?)
Higher Low Frequency (decadal) Variability
Modulation of El Niño events?
Volcanic ForcingVolcanic Forcing
Mann et al 2009
SST (ENSO) response to Volcanic ForcingSST (ENSO) response to Volcanic Forcing
SS
T A
no
ma
ly (
°C)
SS
T A
no
ma
ly (
°C)
Niño 4 IndexNiño 4 IndexControl Run
Millennium Run
Volcanic Forcing (W/m2)
SS
T A
no
ma
ly (
°C)
SS
T A
no
ma
ly (
°C)
Niño 3 IndexNiño 3 IndexControl Run
Millennium Run
Volcanic Forcing (W/m2)
Selected Eruption (max >2.5 W/m2)Selected Eruption (max >2.5 W/m2)
Random SST with same variance () as in CTRL
Simulated SSTSST > for at least 9 months
Conclusion
• Insolation and fresh water flux damp the SST seasonal cycle in the east Pacific (link with large scale dynamics)
• Increase ENSO magnitude from 9.5 ka to 0k in response to insolation
• Early and mid-Holocene ENSO development is damped in Autumn
• Fresh water flux enhances ENSO magnitude • Volcanic eruptions favors increased decadal variability at
the MWP/LIA transition in last millennium simulation with insolation and volcanic forcing
• Colder SST (more La Niña? ) after volcanic eruption? What about the hypothesis of El-Niño 1 year after?
• Need to better understand relative impact of seasonality and variability on relationship between proxy records and ENSO variability to produce reliable model-data comparisons.