Transcript The RAPID ocean observation array at 26.5°N in the

The RAPID ocean observation array at
26.5°N in the HadCM3 model
Leon Hermanson,
Rowan Sutton, Keith Haines,
Doug Smith, Joël Hirschi
Content
•
•
•
•
•
Introduction to RAPID-WATCH and VALOR
The RAPID observation array
Getting an MOC from the observations
Creating an analogous array in HadCM3
Can models reproduce MOC without RAPID?
(but including other observations, eg. ARGO)
• Seasonal cycles in the model and
observations
• Assimilating the RAPID observations
Introduction to RAPID-WATCH
• The RAPID climate change programme ran from 2001—
2008
– Aim was to improve our ability to quantify the probability and
magnitude of future rapid change in climate, with a focus on the
role of the Atlantic Ocean's Thermohaline Circulation
• RAPID-WATCH follows on from RAPID and will run until
2014
– Aim is to continue monitoring the MOC using the moored array in
the North Atlantic and also exploit the data being collected
– Observations will be used to:
• determine and interpret recent changes in the Atlantic MOC
• assess the risk of rapid climate change due to changes in the MOC
• investigate the potential for predicting the MOC and its impacts on climate.
Value of the RAPID array (VALOR)
•
Aim is to assess the value of the RAPID array observations for
predictions of the Atlantic MOC and its impact on climate
•
Will provide information for an assessment of the RAPID array late
2011
•
The project will explore a range of issues concerning the design of
a potential MOC prediction system
•
There are four main activities:
1.
Ocean synthesis
2.
Hindcast experiments
3.
Ensemble design
4.
Pseudo-observations
The RAPID array
• Located at approximately 26°N
• Between Florida and Morocco
• Reasons for choosing this
latitude:
– Gulf Stream is mainly found in
Florida Straits and has been
measured there since 1981
– High correlation between
MOC and ocean heat
transport
– Steep continental slope in the
West
– Strong, steady trade winds
• Downside:
– Low decadal signal to noise ratio
Meridional transports at 26°N in the
Atlantic
Gulf Stream
MOC
Transport (Sv)
1 Sv =
106 m3 s-1
Ekman
Upper
mid-ocean
Gulf Stream transport
Ekman transport
Upper mid-ocean transport
Meridional Overturning Circulation
The RAPID array and HadCM3
26.25°N
W1 W2
Resolution: 1¼° x 1¼°
MARW
MARE
E
Meridional velocity at 26.25 °N
76.25W
76.25W
Meridional velocity at 28.8°N
76.25W
76.25W
Comparison of MOC components
Monthly means
Δ RAPID array
- - 26.25°N
28.8°N
HadCM3 control
Jan
July
Wind stress
NOC1.1
Using all obs except RAPID
Average
correlation
0.3 (0.4)
DPS2 PPE
Assimilation
Monthly means
0.6 (0.6)
0.9 (0.8)
-0.3 (0.0)
Component correlations
Correlation
Model RAPID
Gulf Stream with
Upper mid-ocean
-0.75
-0.15
Gulf Stream with
MOC minus Ekman
0.40
0.55
Upper mid-ocean with
MOC minus Ekman
0.30
0.75
Seasonal cycle of MOC at 26°N
Seasonal cycle of upper mid-ocean
Seasonal cycle of Gulf Stream
and Ekman
Gulf Stream transport
Ekman transport
Assimilating RAPID observations
• Four years of twice-daily temperature and salinity at five
merged profile locations
– Too high frequency
– Need to split up some profiles
– Differences in mean states
• Given the large amount of ARGO and other observations
2004—2008, will RAPID have an impact?
– Eastern boundary poor in current assimilation (without RAPID)
– In NEMO deep observations project onto MOC
– Alternative methods
Salinity
1951-2006
HadCM3
ABW
Levitus
05
Potential Temp
1951-2006
Correlations with transport
1000m
Lag -5
months
Scientific summary
• The way the observed MOC is calculated is particular to
the unusual bathymetry at 26°N
• As most of Florida is missing in HadCM3, the
components of the MOC are difficult to reproduce
• There are no compelling reasons to move the model
array to 28.8°N
• Assimilating all observations except RAPID creates an
unrealistic seasonal cycle
• RAPID observations may have a large impact at eastern
boundary and at deep levels
• RAPID transports could possibly be assimilated through
their covariance with temperature and salinity