Science Briefs

How Does the Atlantic Cold Tongue Affect West African Rains?

Map of June 2013 Aqua-MODIS sea surface temperature data in the equatorial Atlantic

Figure 1. A map of sea surface temperatures in the equatorial Atlantic observed in June 2013 by the Moderate Resolution Imaging Spectroradiometer on NASA's Aqua satellite shows a darker "tongue" of cooler water extending westward from the African coast across the Gulf of Guinea. (Data courtesy: NASA Earth Observations)

The Atlantic Cold Tongue (ACT) is a large pool of locally cooler ocean surface water that develops each spring just south of the Equator in the Eastern Atlantic Ocean. It reflects the upwelling of deep cold waters by the action of southeasterly Trade Winds. Contrasts in temperature between the ACT and the surrounding ocean typically intensify during May and June.

Previous research suggests that the ACT may affect the onset of summer rains farther north, in West Africa's Sahel region. The intensification of the cold tongue increases the temperature contrast between the Gulf of Guinea and the Sahel, which in turn accelerates onshore, south-to-north near-surface winds. According to one theory, atmospheric moisture then surges into West Africa advancing the seasonal rains northward into the Sahel.

Computer simulations using a limited area (regional) climate model were recently analyzed to test this hypothesis. In these simulations, ocean surface temperatures are specified as an important data input at the model atmosphere's interface with the ocean. Five individual reference simulations used observed ocean surface temperatures during each of five seasons, April-October, 1998-2002. To illustrate the structure of the cold tongue, the left side of Fig. 2 shows the 5-year mean ocean surface temperatures for June 26-30, where the ACT is centered about the 298 Kelvin (25°Celsius) isotherm. These reference simulations successfully replicated many features of the seasonal climate evolution as the cold tongue develops and intensifies. In particular, the northward advance of seasonal rainfall using these ocean surface temperatures was similar to the observed timing.

Contour plot of temperatures with ACTContour plot of temperatures without ACT

Figure 2. At left: Ocean surface temperatures (K) for June 26-30, 1998-2002, for the reference simulations that included the ACT. The 298K isotherm is at the center of the cold tongue. At right: Ocean surface temperatures for the same time period for simulations without the ACT. Compared to the control, the ocean is some 1-4K warmer where the ACT has been replaced.

A second set of simulations was run in parallel, but using warmer ocean surface temperatures that eliminated the cold tongue. The right side of Fig. 2 shows the corresponding mean of the five sets of ocean surface temperatures for June 26-30, prescribed for five simulation experiments without the ACT.

Comparison of the two sets of simulations suggests that the absence of the ACT does not delay the northward migration of precipitation. The unambiguous impact of replacing the ACT with warmer ocean surface temperatures is to increase rainfall rates downstream over West Africa throughout the summer, but the timing of major rainfall events was not affected. Higher rainfall rates owe to additional moisture advanced toward West Africa in the experiments. The added moisture, in turn, results from higher humidity due to excessive evaporation above the experiments' warmer ocean surface temperatures.

Time plots of rainfall over West Africa.

Figure 3. Five-day mean (pentad) precipitation rates (mm/day) during June-September 2001 over West Africa (10°W-10°E) by latitude. Pentad 1 is June 1-5 and pentad 24 is Sept. 24-28. Top: Observed by TRMM satellite; Middle: Simulated with ACT; Bottom: Simulated without ACT. Note the initiation of precipitation maxima along 10°N beginning with pentad 5 (June 21-25) and the waning of precipitation at 5°N by pentad 10 (July 16-20).

The results are illustrated in Fig. 3, which shows how rainfall maxima change latitude over West Africa during the summer of 2001. Pentad (5-day mean) rainfall rates plotted on these charts are averaged over 10°W-10°E at each latitude from 0° to 20°N. The timing of latitudinal shifts of rainfall maxima for the reference simulation (Fig. 3, middle) match satellite observations of precipitation rates (Fig. 3, top) quite well, while the experiment without the ACT (Fig. 3, bottom) produces the same timing of latitudinal shifts, but with some enhancement of precipitation maxima.

Note that global winds, temperatures and humidity used as background data for the limited area simulations were the same in the experiment as in the reference simulations. Results imply that the signals from these global data lead to the particular sequence of precipitation events observed each year, irrespective of the presence of the ACT. It is possible, however, that the same global conditions that cause an early appearance of the ACT during some years could also cause early advances of seasonal rainfall to the Sahel, making the ACT a corollary consequence rather than a cause of Sahel precipitation timing.


Druyan, L.M., and M. Fulakeza, 2014: The impact of the Atlantic cold tongue on West African monsoon onset in regional model simulations for 1998-2002. Intl. J. Climatol., 35, no. 2, 275-287, doi:10.1002/joc.3980.


Please address all questions about this research to Dr. Leonard Druyan.