11/5/09
2
Sahel Rainfall Interhemispheric SST Contrast
NH Cold
NH Warm
Strong relationship to SST changes
11/5/09
3
A West African Monsoon Index
The West African summer monsoon index (WASMI) is defined as an area-averaged seasonally (JJAS) at 850 hPa within the West African monsoon domain (5º-17.5ºN, 20ºW-40ºE) (Li and Zeng, 2002, 2003, 2005)
Correlation maps between the seasonal (June-September) rainfall and WASMI (1979-97). The shaded areas indicate significant at the 95% confidence level.
A West African Monsoon Index
The West African summer monsoon index (WASMI) is defined as an area-averaged seasonally (JJAS) at 850 hPa within the West African monsoon domain (5º-17.5ºN, 20ºW-40ºE) (Li and Zeng, 2002, 2003, 2005)
Sahel Rainfall Interhemispheric SST Contrast
11/5/09
4
Results from perfect SST “AMIP” ensembles (L. Goddard, IRI)
… but current seasonal SST forecasts are without skill!
SST along the equator
as observed (thick black)
in coupled ocean–atmosphere general circulation models (GCMs)
⇒ Strong biases in GCMs including a failure of reproducing cold tongue
Errors in AGCMs during spring
Possibility of improving coupled simulations
Richter and Xie (2008)
Atlantic SST bias in coupled climate models
11/5/09
6
summer ITCZ
spring ITCZ
ITCZ position and rainfall intensity affect densely
populated regions in West Africa
Sahel rainfall climatology
Guinea rainfall climatology
Atlantic ITCZ annual cycle
11/5/09
7
Atlantic ITCZ Variability
First EOF (23%) of the June-August rainfall from GPCP 1979-2001 (contours in mm/day). June-August SST anomaly (colors, in °C & white contours, every 0.2°) and surface wind anomaly (vector, in m/sec) are determined by regression on the time series of the rainfall EOF.
First EOF (33%) of the March-April rainfall from GPCP 1979-2001 (contours in mm/day). March-April SST anomaly (colors, in °C & white contours, every 0.2°) and surface wind anomaly (vector, in m/sec) are determined by regression on the time series of the rainfall EOF.
11/5/09
8
Zonal Mode - „Atlantic Nino“
Dominant mode during boreal summer
Increased precipitation (grey shaded, mm/day) during warm events
⇒ SST in the equatorial cold tongue important for regional climate prediction
Kushnir et al. (2006)
NTA
MEAN
Covariance
11/5/09
9
Point 1
• Tropical Atlantic Variability fundamentally involves the interaction between the ocean surface and the ITCZ�
• Its patterns are highly seasonal�
• In the spring the North-South SST gradient anomaly dominates causing rainfall fluctuation over Brazil�
• Are ocean dynamics involved?
11/5/09
10
NCAR Climate Community Model (CCM3)
couple to simple 75m ocean ML model
Standard T42 configuration
11/5/09
13
Point 2
• In the spring the North-South SST gradient anomaly develops well in a Mixed-Layer Ocean Model.�
• Ocean Dynamics are NOT crucial here. �
• However, the signals are possibly to strong in this model and one can assume that ocean mixing will damp the SST anomalies�
• How large is the upper ocean mixing?
Free falling microstructure profiler:
• two vertical velocity shear sensors
• high resolution temperature sensor
• acceleration sensor
• depth range: 0-500m
Quantify diapycnal mixing in the upwelling regions of the Tropical Atlantic Ocean
Marcus Dengler�(IFM-GEOMAR)
11/5/09
14
Heat Budget
Direct estimates of the diapycnal heat flux at the base of the mixed layer: up to 90 W/m² during cold tongue onset
Microstructure measurements with a loosely tethered profiler (M. Dengler, IFM-GEOMAR)
11/5/09
15
Tropical Atlantic Climate Experiment
Spring Summer First EOF of rainfall anomaly [mm/day] and regressed SST and surface wind
after Kushnir et al. (2002)
SST and circulation scheme
⇒ Potential for prediction of Rainfall over Africa and Brazil
after Schott et al. 2004
Motivation/Introduction
Equatorial Cold Tongue
Cold SSTs develops during boreal summer in the eastern equatorial Atlantic
Strong interannual variability and long term warming trend
ATL3 annual cycle: Max in April ~29°C Min in Aug. ~24°C
11/5/09
16
Why are the rain fall predictions over Africa so bad?
Dominant pattern of precipitation error associated with dominant pattern of SST prediction error based on persistent SST anomalies (Goddard & Mason ,Climate Dynamics, 2002)
FLAME – Mean (MJJ) 15m Temperatures
Cold tongue in the eastern equatorial Atlantic with box for cold tongue index
Tropical Atlantic Climate Experiment
FLAME – Model
• 1/12° horizontal resolution • 45 vertical levels (10m near the surface, 250m below 2000m) • simulation from 1990 to 2002
11/5/09
17
Tropical Atlantic Climate Experiment
Cold Tongue (MJJ) – Temperature Anomaly
Temperature Transport Anomaly (MJJ)
Point 4
• The zonal (Antlantic NINO) mode affects the West African Monsoon and thus rainfall in the boral fall season over Subsaharan Africa.
• Significant societal impacts have been demonstrated, for example drought mitigation, tropical disease mitigation..
• Ocean dynamics are at the heart of the cold SST development in the eastern tropical Atlantic.
• Current assimilation models perform poorly
11/5/09
18
TACE: Tropical Atlantic Climate Experiment TACE core period six years: 2006/7 - 2011/12
To advance understanding of coupled ocean-atmosphere processes and improve climate prediction for the Tropical Atlantic region
Specific goals are: a) To advance understanding of the key processes that control SST, interactions with the AMI (Atl. Marine ITCZ), and related climate predictability in the eastern tropical Atlantic. b) To contribute to the design of an enhanced sustained observing system for the tropical Atlantic region.
11/5/09
19
1) Determine oceanic processes important in regulating SST in the tropical Atlantic and associated atmospheric responses
2) Improve SST forecasts on seasonal to interannual time scales in the tropical Atlantic
3) Provide parameterizations and model improvements to global and regional prediction centers
4) Investigate response of tropical Atlantic region to global warming, including teleconnection patterns
Tropical Atlantic Climate Experiment
Cruise Track M68/2
Mooring array at 23°W BMBF Verbundvorhaben Nordatlantik
11/5/09
20
Tropical Atlantic Climate Experiment
3 Moorings at 23°W 0.75°N, 0°N, 0.75°S (including the PIRATA ADCP in the central mooring) and one at 21.5°W, 0°N
CTD-profiler at 23°W, 0°N 1000m - 3500m (John Toole, WHOI)
CTD/O2-profiler At 23°W, 5°N 120m – 1000m (new SFB)
Ocean Observing System
Interannual Variability
Seasonal dependence of interannual variability: Strongest during boreal summer (June/July)
11/5/09
21
Interannual Variability
Boreal summer cold tongue:
Warm events in 2002, 2006, 2007
Cold event in 2005
EUC and cold tongue variability
warm (cold) event in JJA 2002 (2005), with relaxed (intensified) winds in the west during boreal spring (MAM)
EUC embedded in shallower (deeper) thermocline at 23°W during boreal summer 2002 (2005)
Hormann and Brandt (2009)
Interannual Variability
Zonal Velocity Temp. Gradient Temperature