Oceanic mixed layer heat budget in the Eastern Equatorial Atlantic using ARGO floats and PIRATA buoys M. Wade (1,2,3), G. Caniaux (1) and Y. du Penhoat (2) 1. CNRM/GAME (Météo-France / CNRS), Toulouse, France 2. IRD LEGOS, Toulouse, France 3. LPAOSF, Dakar, Sénégal

Introduction 1.SST anomalies in the Eastern Equatorial Atlantic (EEA) condition precipitation over the West African continent ( Vizy and Cook, 2002 ) 2.Interaction between the Atlantic Cold Tongue (ACT) and the West African Monsoon ( e.g. Okumura and Xie, 2004; Gu and Adler, 2004; Caniaux et al., 2011 ) 3.Need to understand the physical processes affecting SSTs in the EEA SAHEL ACT Equatorial Front Reynolds et al. (2007)’s SSTs – 15 June 2005

Introduction 1.Observation ( e.g. Merle et al., 1980; Foltz et al., 2003 ) and modeling studies ( Philander and Pacanowski, 1986; Yu et al., 2006; Peter et al., 2006 ) suggest that subsurface processes are at the origin of upper temperature decrease in the ACT 2.Further south SSTs are largely dominated by surface heat fluxes 3.Observations fail to close heat budgets within 150 Wm -2 in the EEA ( Foltz et al., 2003 ) We present an oceanic mixed layer heat budget combining in-situ observations from ARGO floats and PIRATA buoys and outputs from atmospheric NWP models.

ARGO Floats October 2007

Method Heat storage Hori. advec. Entrainment Solar h. flux Non-solar h. flux with Surface heat fluxes: ECMWF OSCAR data set (Helber et al., 2007) + Reynolds et al. (2007)’s SSTs Ohlmann (2003) and

Subdivision of the EEA 9 boxes reflecting the regional characteristics of the dynamics and thermodynamics Northern boxes ( warmer SSTs, Guinea Current, heavy precipitation ) ACT boxes ( SEC and EUC ) Southern boxes ( SECC, Angola dome ) De Boyer Montégut et al. (2004) MLD climatology

Spatial and Temporal Distributions of Profiles profiles over 3 years ( ) 2.The number of floats increased during the EGEE cruises 3.20 profiles/mon. in Jan and 100 profiles/mon. in Jan Lower number of floats (4-9) and profiles in boxes 2, 3, 6 than in the other boxes (11-18) 5.4 PIRATA buoys included EGEE1 EGEE2EGEE3 EGEE4EGEE5 EGEE Latitude vs. time Number vs. time

Comparison ARGO-Reynolds’ SSTs 1.Good agreement between ARGO and Reynolds’ SSTs after quality control 2.Important seasonal cycle in each box, weaker north (3°C- 4°C) than south of the equator (4°C-6°C) with north-south SST gradients was a year with anormally cold SSTs in the ACT (boxes 2, 5, 8) ARGO SSTs Reynolds et al. (2007)’s SSTs SSTs vs. time

Mean Annual MLD Heat Budget ( ) 1.Mean budgets are positive except in boxes 6 and 9 2.Surface heat fluxes dominate 3.A significant part of solar heat flux goes out of the ML 4.Mean horizontal advection stronger in boxes 1, 2, 8, 9 5.Weak contribution of entrainment 6.The residuals are strong, meaning that one or several terms are missing BOX 7 BOX 8 BOX 6 BOX 5 BOX 3 BOX 2 BOX 4BOX 1 BOX Horizontal advection Net heat flux Fsol*I(-h) Residual Entrainment Heat storage

Seasonal Cycle 1.Amplitude of cooling / warming: -100 to +50 Wm -2 2.Cooling from April to July, weaker in the ACT 3.Semi-annual signal 4.The Σ ( calculated terms) > storage term (except in box 9): the residual term has a negative contribution Heat storage Sum of the calculated terms

1.The seasonal cycle of heat storage is mainly due to the seasonal cycle of the surface heat fluxes 2.The surface heat fluxes have a >0 contribution, except in austral winter in boxes 6 and 9 3.Entrainment is insignificant throughout the year Seasonal Cycle Horizontal advection Net heat fluxF sol *I(-h) Entraînment

Seasonal Cycle of the Residuals 1.Negative residuals (reaching -120 Wm -2 ) 2.Weak residuals in the southernmost boxes 3.The residuals are due to errors and missing terms 4.Turbulence data collected during AMMA/EGEE ( Dengler et al., 2010 ) and other studies ( Rhein et al., 2010 ) compare well with the residuals (phase, amplitude, spatial distribution) 5.Vertical mixing is probably the leading missing term Heat Storage5 Wm -2 Horizontal Advection5 Wm -2 Entrainment18 Wm -2 Net surface heat flux17 Wm -2 Fsol I(-h)7 Wm -2 Residual27 Wm -2 Statistical errors

Conclusions 1.ARGO floats provide realistic ML budgets in the EEA at seasonal time scales 2.Cooling in April-July and warming the remainder of the year with a short cold season 3.The ML cooling rate in the ACT is weaker and shorter than elsewhere 4.Strong residuals suggest that vertical diffusion is the most important process during the ACT setup (shear between the SEC and the EUC in the ACT and between the Guinea Current and the Guinea Undercurrent in the northern boxes)