The Annual Cycle of Precipitation

over the Tropical Atlantic, South America, and

Africa.Mutual Influences of Land and Ocean.

Michela Biasutti

David S. Battisti and Edward S. Sarachik Co-chairs

Seattle, July 7, 2003

1

OUTLINE

• Motivation

• Ocean-Land relationship in Tropical Atlantic Variability (TAV)

• Similarity between TAV and Annual Cycle (AC)

• Method

• Models

• Experimental design

• Results

• Local Control of Land Precipitation (uncoupled)

• Local Control of Ocean Precipitation (uncoupled)

• Remote Control of Land Precipitation (uncoupled)

• Remote Control of Ocean Precipitation (uncoupled & coupled)

• Conclusions

• Implications2

Motivation: SST & LAND PRECIPITATION

Correlation of precip and SST

60E 90E 120E 150E 180 150W 120W 90W 60W 30W 0 30E 60E 30S

15S

EQ.

15N

30N

45N

60NMitchell/Kaplan 1950:2000 (JJASO Sahel Rainfall, JJAS SST)

60E 90E 120E 150E 180 150W 120W 90W 60W 30W 0 30E 60E 30S

15S

EQ.

15N

30N

45N

60NMitchell/Kaplan 1950:2000 (FMAM FQI, FMAM SST)

−0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8

• Fact: Global SST is corre-lated with precipitation inSahel and Northeast Brazil.

• Working Hypothesis: Trop-ical Atlantic SST influencesland precipitation.

• Open Question: What arethe mechanisms?

• Is the SST influence di-rect?

• Is it mediated throughchanges in oceanic precip-itation?

• Is there a feedback: canland precip affect SST?

3

Motivation: AC & TAV ANALOGUE

• Atmospheric processes bring the SST signal inland ⇒ the same fastatmospheric dynamics must be involved in both the AC and the TAV.

• The gross features of AC and TAV in the deep tropics look alike

4

Motivation: AC & TAV ANALOGUE

meridional modecross-equatorial SST gradient

cross-equatorial wind towards warm SSTmeridional shift of ITCZ towards warm SST

5

Method: THE MODELS

• Uncoupled Community Climate Model Version 3 (CCM3) T42resolution.

⇒ The precipitation over Africa and South America is overestimated.

⇒ Tropical land is too cold (especially the Sahara).

• Coupled CCM3 coupled to a Slab Ocean Model (SOM) in the tropicalAtlantic only (prescribed SST elsewhere).

⇒ SST responds to radiative and turbulent heat fluxes.

⇒ No ocean dynamics (Ocean Heat Transport Convergence isparameterized with a flux adjustment, the Q-flux).

⇒ Climatology of CCM3+SOM is (by construction) nearly identical tothe climatology of CCM3.

6

Method:THE IDEA BEHIND THE EXPERIMENTS

We want to identify:

• what features of the AC over land and ocean are locally forced?

• what features of the AC over land and ocean are remotely forced?

• the mechanisms of mutual influences between land and ocean regions

IDEA: we suppress the AC of forcings over land and ocean regionsseparately ⇒ we can separate out local and remote responses.

7

Method: THE FORCINGS• Insolation over land.

• Insolation over ocean.

• Q-flux.

• Elevated Condensational Heating(Qcond) in selected areas

} or SST

8

Method: EXPERIMENTAL DESIGN

• Fixed insolation over land, AC of Ocean Forcings (Qflux+Insolation orSST)

⇓AC over LAND is a REMOTE response to AC over oceanAC over OCEAN is a LOCAL response

• Fixed Ocean Forcings, AC of insolation over land⇓

AC over LAND is a LOCAL responseAC over OCEAN is a REMOTE response to AC over land

9

RESULTS

A. LOCAL LAND

• AC of insolation;March SST⇒ AC of land precip

B. LOCAL OCEAN

• AC of SST;March Insolation⇒ AC of ocean precip

C. REMOTE LAND

• AC of SST;March Insolation⇒ AC of land precip

D. REMOTE OCEAN

• AC of land insolation;March SST⇒ AC of ocean

10

Results: LOCAL LAND

90W 60W 30W 0 30E

30S

15S

EQ.

15N

30N

CTL − 1st EOF of PRECIP (56%)

J F M A M J J A S O N D−1.5

−1

−0.5

0

0.5

1

1.5

21st PC

CTL:North-South displacement over land

and ocean

Extrema: February/August

Local Land:North-South displacement over land

Extrema: December/June

Gulf of Guinea is way off!

11

RESULTS

A. LOCAL LAND

• AC of insolation;March SST⇒ AC of land precip

B. LOCAL OCEAN

• AC of SST;March Insolation⇒ AC of ocean precip

C. REMOTE LAND

• AC of SST;March Insolation⇒ AC of land precip

D. REMOTE OCEAN

• AC of land insolation;March SST⇒ AC of ocean

12

Results: LOCAL OCEAN

90W 60W 30W 0 30E

30S

15S

EQ.

15N

30N

CTL − 1st EOF of PRECIP (56%)

J F M A M J J A S O N D−1.5

−1

−0.5

0

0.5

1

1.5

21st PC

CTL:North-South displacement over land

and ocean

Extrema: February/August

Local Ocean:North-South displacement over ocean

Extrema: March/August

Local control of ITCZ position

Non-local control of ITCZ intensity

What’s the role of elevated condensational heating (Qcond) in generatingsurface winds and convergence?

13

Results: LOCAL OCEAN: the role of Qcond.

Moisture Convergence (P-E) &Surface Wind Response to SSTchanges.

14

Results: LOCAL OCEAN: the role of Qcond.

Moisture Convergence (P-E) &Surface Wind Response to SSTchanges.

Moisture Convergence (P-E) &Surface Wind Response to Qcondchanges in the ITCZ. The ITCZ

Qcond drives a circulation that sustains

the original SST-induced displacement

of the ITCZ.

15

RESULTS

A. LOCAL LAND

• AC of insolation;March SST⇒ AC of land precip

B. LOCAL OCEAN

• AC of SST;March Insolation⇒ AC of ocean precip

C. REMOTE LAND

• AC of SST;March Insolation⇒ AC of land precip

D. REMOTE OCEAN

• AC of land insolation;March SST⇒ AC of ocean

16

Results: REMOTE LAND.

Precipitation response to SSTchanges.• Northeast Brazil /Guiana

• Equatorial Africa & Guinea /Sahel

17

Results: REMOTE LAND: the role of Qcond.

Precipitation response to SSTchanges.• Northeast Brazil /Guiana

• Equatorial Africa & Guinea /Sahel

Precipitation response to Qcond

changes in the ITCZ.• YES: South America

• NO: Africa

18

Results: REMOTE LAND: the role of Tsfc

300

320

340

360West Africa

____ Sept. θe _ _ _ Sept. SLP

____ March θe _ _ _ March SLP

1005

1015

1025

30S 15S EQ 15N 30N1005

1015

1025

300

320

340

360Central Africa

____ Sept. θe _ _ _ Sept. SLP

____ March θe _ _ _ March SLP

1005

1015

1025

30S 15S EQ 15N 30N1005

1015

1025

SST⇓

Land Tsfc

⇓θe and SLP

⇓Pmax position.

19

Results: REMOTE LAND: Inland Advection of SST.

+

• SST ⇒ ITCZ

• ITCZ ⇒ wind in Africa + wind

and SAT in Near East

• mean westerlies ⇒ Arabian SAT

• mean easterlies + wind anomalies

⇒ Sahara SAT︸ ︷︷ ︸

• African SLP ⇒ southerly wind

• ⇒ moisture convergence into the

Sahel 20

RESULTS

A. LOCAL LAND

• AC of insolation;March SST⇒ AC of land precip

B. LOCAL OCEAN

• AC of SST;March Insolation⇒ AC of ocean precip

C. REMOTE LAND

• AC of SST;March Insolation⇒ AC of land precip

D. REMOTE OCEAN

• AC of land insolation;March SST⇒ AC of ocean

21

Results: REMOTE OCEAN: ITCZ Intensity Change

Uncoupled response: Land climateforces rainfall intensity anomalies inthe Atlantic ITCZ.

• Does the ITCZ respond tochanges in land surface temper-ature?

• Does it respond to changes inland precipitation (Qcond)?

22

Results: REMOTE OCEAN: Response to Land Qcond

The response to imposed steady forcing in elevated condensational heatingover Africa and South America.

Uncoupled response• ITCZ intensity responds to remote

Qcond anomalies.

• Anomalies are co-located with the

mean ITCZ.

• Surface wind anomalies over ocean.

23

Results: REMOTE OCEAN: Response to Land Qcond

The response to imposed steady forcing in elevated condensational heating(Qcond) over Africa and South America.

Uncoupled response• ITCZ intensity responds to remote

Qcond anomalies.

• Anomalies are co-located with the

mean ITCZ.

• Surface wind anomalies over ocean.

Coupled response• wind anomalies ⇒ evaporation

anomalies

• latent heat flux anomalies ⇒ SST

anomalies

• anomalous SST gradient ⇒ the

ITCZ shift

• wind/evaporation/SST/ITCZ feed-

back.

24

COUPLED RESULTS

D. REMOTE OCEAN (coupled!)

CTL

• AC of Insolation over Land.

• AC of Insolation over Ocean.

• AC of Q-flux.

Fixed Insolation over Land

• March Insolation over Land.

• AC of Insolation over Ocean.

• AC of Q-flux.

︸ ︷︷ ︸⇓

Effect of AC of Insolation over Land

25

Results: REMOTE OCEAN: Central Atlantic response toAC of Insolation over Land

15S10S 5S EQ 5N10N15N

PRECIPITATIONControl

PRECIPITATIONOceanic Forcings Only

J M M J S N J M M J S N

15S10S 5S EQ 5N10N15N

SST & V

J M M J S N J M M J S N

SST & V

suppressed AC over land ⇒ suppressed

meridional annual march of ITCZ.

Concomitant changes in SST.

What forces the SST & ITCZ anomalies?

• Land surface temperature?

• Land precipitation?

26

Results: REMOTE OCEAN: Effect of Land Temperature andPrecipitation

Compare the control simulation to the simulation with fixed insolation overland and the simulation with fixed elevated condensational heating over

Africa and South America (Qcond)

15S10S 5S EQ 5N10N15N

PREC′

Land Insolation Effect

PREC′

Land Qcond

Effect

PREC′

Land Temperature Effect

J M M J S N J M M J S N

15S10S 5S EQ 5N10N15N

SST′ & V′

J M M J S N J M M J S N

SST′ & V′

J M M J S N J M M J S N

SST′ & V′

Land Qcond andland Tsfc havelarge and oppositeeffects.

Land Tsfc domi-nant.

27

How?

28

Results: REMOTE OCEAN: Development of ITCZ Anomalies inResponse to Land Forcing.

J M M J S N J M M J S N−15

−10

−5

0

5

T SA

HA

RA

Effect ofInsolation over Land

J M M J S N J M M J S N

−1

0

1

2

UN

TA

J M M J S N J M M J S N

−0.9

−0.6

−0.3

SS

T NTA

J M M J S N J M M J S N

−1

−0.5

0

0.5

∇ S

ST

J M M J S N J M M J S N

−6

−3

0

ITC

Z

December cold “anomalies” in theSahara ⇒ thermal high

⇓stronger NTA Trades ⇒ stronger

evaporation⇓

colder NTA+

feedbacks @ equator⇓

equatorial SST gradient⇓

ITCZ shift

29

Equatorial Feedbacks

30

CONCLUSIONS:Mutual Influences of Land and Ocean

31

Conclusions: REMOTE INFLUENCE OF SST ON LAND

The AC of SST influences the AC of land precipitation

in coastal areas: SST ⇒ ITCZ ⇒ Northeast Brazil and Guianaprecipitation.

in the Sahel: SST ⇒ Sahara Tsfc ⇒ SLP gradient ⇒ Sahel precipitation.

32

Conclusions: REMOTE INFLUENCE OF LAND ONOCEAN

Land climate influences the ITCZ intensity and position.

intensity: Land Qcond ⇒ free tropospheric temperature ⇒ stability overthe ocean ⇒ ITCZ intensity.

position: directly and indirectly forced wind ⇒ latent heat loss ⇒ SST ⇒ITCZ position.• Sahara Tsfc ⇒

north tropical At-lantic Trades.

• Land Qcond ⇒ equa-torial wind.

• Land Qcond ⇒ ITCZintensity ⇒ surfacewind at the edge ofthe ITCZ.

Changes in surface windtrigger coupled feedbacksamong wind, SST, andITCZ.

33

IMPLICATIONS:for models, TAV, and paleo

34

Implications: MODELING THE ANNUAL CYCLE

• CCM3 biases: reduce the overestimate of precipitation over land (by

changing albedo?) ?⇒ solve the underestimation of precipitation in theITCZ.

• CGCM biases: correct precipitation over equatorial coastal areas andSahel ⇔ correct SST.

• CGCM biases: correct march of Atlantic ITCZ ⇔ correct AC oftemperature in the Sahara and of precipitation in Africa and SouthAmerica.

35

Implications: TROPICAL ATLANTIC VARIABILITY &CLIMATE CHANGE

• SST influences land precipitation at the annual timescale in the sameway it does at the interannual timescale ⇒ the AC is indeed a usefulanalogue for TAV.

• AC of land influences AC of ocean ?⇒ continental variability influencesmaritime variability.

• Change in African climate (e.g. due to deforestation) ?⇒ NortheastBrazil (via ITCZ).

Remaining question: How does the response time of the ocean modify theresponse to continental forcing at the interannual timescale?

36

Implications: PALEO CLIMATE STUDIESe.g. the Green Sahara

• Simulation of precipitation at edge of Sahara ⇔ correct basinwideAtlantic SST boundary conditions

• Simulation of precipitation in Sahara ⇔ correct simulation of soilalbedo (soil moisture and vegetation).

• The greening of the Sahara should be visible in paleo records of theITCZ position (e.g. from the Cariaco basin). Is it?

Caveat: what’s the role of ocean dynamics?

37

THANXto parents, teachers, friends, and numerous beings who have given guidance

and support along the path.

David Battisti. Ed Sarachik. Dennis Hartmann. Mike Wallace. ChrisBretherton. Eric Steig. Class of ’96. Dan Vimont. Martin Widmann. ChrisThompson. John Chiang. Todd Mitchell. Kevin Rennert. Justin Wettstein.

Camille Li. Everybody past and present in JISAO. Fiamma Straneo.Rosalba Ciampi. Chris Flanagan. Eric Dee. Pokeys and FNBers. MaryLouGamba. Housemates. Martin Wallace. Mom & Dad and sisters and friends

back home. David, Lynn, Eric, Adrian, Nathaniel

...and more!

38