Observations of Tropical Atlantic Variability

Lecturer: Jim Hurrell

Summary by: Olga Zolina and Sungsu Park

1. Introduction

The seasonal cycle is the largest atmosphere-ocean signal in the tropical Atlantic. For this reason, the annual mean climate and the seasonal variability about that mean will be reviewed. Interannual and longer-term variations are not negligible, however, and at least in part can be interpreted as modulations to the average annual cycle. Much of the early interest in tropical Atlantic variability (TAV) was motivated by the enormous social and economic impacts it has on the local populations in parts of South America and Africa. Some of these early studies will be reviewed, in particular those that link variations in Sahel and Nordeste rainfall to anomalous tropical sea surface temperature (SST) patterns.

Another aspect of the tropical Atlantic is that, unlike the tropical Pacific where interannual variability is dominated by the El Nio Southern Oscillation (ENSO) phenomenon, multiple competing influences of comparable importance affect the tropical Atlantic. Anomalies in tropical Atlantic SST are primarily driven by changes in surface winds, which can be forced either locally or remotely. Local forcing, for instance, arises from changes in the position and intensity of the Inter-Tropical Convergence Zone (ITCZ), while remote forcing comes from variability over the extratropical Atlantic or from ENSO. Also, superimposed on the mean seasonal cycle are two modes of coupled atmosphere-ocean variability. The first is characterized by a north-south interhemispheric gradient in SST, with associated changes in trade winds, that exerts considerable influence on the regional climate. These fluctuations display markedly large power on time scales of 8-16 years. The second is an equatorial mode of variability similar to ENSO. Although weak relative to the Pacific variability, the Atlantic equatorial SST anomalies affect regional rainfall, such as the Gulf of Guinea. All of these aspects of TAV will be reviewed. Later talks will focus more on the primary mechanisms of variability.

2. Annual mean climate

Over the eastern Pacific and the Atlantic Oceans, the ITCZ looks much like our conceptual model: a narrow, well-defined band of high clouds, convection and rainfall. Its mean position is north of the equator, between about 5-8N. Over the tropical continents, the ITCZ is much broader. Minima in rainfall are associated with descending motions on the eastern flanks of the large subtropical anticyclones in both hemispheres. These anticyclones produce the tropical easterlies, which are steady and strong over the Atlantic. The southern trades cross the Equator and converge with the northern trades near the mean latitude of the ITCZ. Over the eastern tropical Atlantic, the winds are more meridional and rotate clockwise toward the African continent in association with the monsoon. This wind direction favors coastal upwelling over most of the African coast south of the Equator. Upwelling also occurs along the Equator in the central Atlantic, where the southeasterly winds are favorable for Ekman divergence, and over the subtropical African coast in the Northern Hemisphere. The warmest SSTs exceed 27C and are mostly located north of the Equator. The average wind pattern results in a deeper thermocline in the western tropical Atlantic than in the east: the 20C isotherm deepens from 50 m depth in the east to about 150 m in the west.

3. Seasonal cycle

The ITCZ migrates seasonally. It is farthest south in boreal spring (March-May), just reaching into the Southern Hemisphere, and farthest north between September and November, near 10N. Over the tropical continents, rainfall maxima more closely follow the annual march of the sun, which produces a meridionally broad pattern in the annual mean.

The seasonal migration of the oceanic rainfall generally follows the seasonal cycle in SST. Near 30W, for instance, the warmest SSTs are near 5S during March and April, but move northward to 5-10N by September. Pronounced seasonal changes in the surface wind stress are also associated with the seasonal movement of the ITCZ. At the Equator near 30W, for example, weak and variable winds prevail from January through May but, as the ITCZ moves northward, the southeasterly trades pick up and the SSTs cool down. This is even clearer if one examines the seasonal evolution of equatorial SST and surface wind as a function of longitude. Here, the seasonal cycle of wind is strongest over the western Atlantic as just described. Over the eastern Atlantic, the strongest southerlies occur when the ITCZ is farthest north (September-November) and the African monsoon is well developed. Associated with the intensification of the easterly wind stress in May, a cold tongue develops and reaches its peak amplitude in August.

The seasonal changes in the wind, particularly over the western tropical Atlantic, induce a response in the equatorial wave-guide that changes the slope of the thermocline and the sea surface. The slope is flat from January through June, relative to the deep thermocline in the west and the shallow thermocline in the east during the second half of the year.

The response of the current systems in the tropical Atlantic to the strong seasonal cycle in surface winds includes a number of striking phenomena. One of these is the North Brazil Current (NBC). During November-April, when the southeasterly trades are weak and the ITCZ is displaced to the south, the NBC flows continuously along the South American coast and helps to transport heat northward. However, when the southeast trades strengthen in May, the NBC abruptly veers offshore between 5 and 10N and the feeds the North Equatorial Counter Current (NECC), which exhibits its own remarkable seasonal cycle.

These are just a few examples of the strong seasonal cycle in the atmosphere/ocean system over the tropical Atlantic. The salient point is that any interannual and longer-term variability is superimposed (and somewhat dwarfed) by these seasonal changes, which is opposite the case in the tropical Pacific. Yet, the interannual and decadal variations have strong climate impacts, which we will now review.

4. Rainfall variability over tropical North Africa and Northeast Brazil

Variability of summer rainfall over the tropical continents could be due to internal atmospheric variability, land surface forcing, or oceanic forcing. The former is a relatively small effect in the tropics. Land surface forcing can be important, especially in individual years. We will focus, however, on oceanic forcing.

The Sahel

ropical North Africa encompasses a variety of rainfall regimes. On the southern flanks of the Sahara lies the Sahel, where up to 90% of the annual mean rainfall occurs during July, August and September. Immediately to the south is the region known as the Soudan, and on its southwest flank is the Guinea Coast. During July-September, the ITCZ is at its northernmost location and the Guinea Coast experiences a mini-dry season. In fact, summer rainfall over the Guinea Coast is negatively correlated (-0.4) with rainfall over the Sahel.

The temporal character of Sahel rainfall is well known. The long-running drought from the mid 1960s until recent summers is perhaps the most striking example of multi-decadal variability evident in the instrumental database. Variations in Sahel rainfall are associated with distinct patterns in SST. In particular, drier-than-average years are typically associated with cooler-than-average tropical Atlantic SSTs north of the Equator, with warmer-than-average SSTs south of the Equator. SST anomalies of opposite sign are observed during wet years. Moreover, observational and model studies have indicated that rainfall anomalies over the Sahel are associated with a pattern of global SST anomalies. This pattern comes out as a leading EOF of global SSTs, and it is characterized by anomalies of opposite sign over the two hemispheres. The associated principal component time series of this so-called interhemispheric contrast mode correlates with summer rainfall over the Sahel at about 0.6. Recent research has also suggested that ENSO has both an indirect and a direct influence on Sahel rainfall. The former occurs via an influence on tropical Atlantic SSTs, while the latter may involve anomalous stationary equatorial waves that interact to enhance subsidence over tropical North Africa during warm event years.

The Nordeste

As for the Sahel, the relationship between rainfall anomalies over the northeast region of Brazil and tropical Atlantic SST points to a strong role for oceanic forcing, with an out-of-phase relationship between hemispheres that acts to modify the position of the ITCZ. Wetter-than-average rainy seasons (February through June) are typically associated with warm (cold) SST anomalies south (north) of the Equator, consistent with a southward shifted ITCZ. The opposite is true of dry periods. Nordeste rainfall anomalies are dominated by variability on interannual time scales, with less evidence for multi-decadal variations than for the Sahel.

ENSO is known to have a large influence on Nordeste rainfall. During warm events, drier-than-average conditions prevail. At the same time, SSTs over the tropical North Atlantic are warmer-than-average. Because tropical Pacific SST variations are significantly correlated with tropical North Atlantic SST variations, it is difficult to unambiguously infer which ocean is directly linked to rainfall variations over Northeast Brazil.

5. ENSO influence on the tropical Atlantic

Other lecturers will talk in more detail about the influence of ENSO on the tropical Atlantic and, in particular, will try to quantify this influence through model experiments. The main results from observational studies can be summarized as follows. During boreal spring under El Nio conditions, anomalies include weak northeasterly trade winds, warm SSTs and low sea level pressure (SLP) over the tropical North Atlantic, a northward shift of the ITCZ, and a decrease in rainfall over the Nordeste. The largest changes follow the tropical Pacific SST anomalies by 3-6 months. Secondary (and less robust) features include weak cold SST anomalies, increased SLP, and slightly enhanced southeasterly trades over the tropical South Atlantic.

6. The cross-equatorial SST gradient

Many analyses of tropical Atlantic SST variability have identified the presence of a "dipole" structure, especially in relation to rainfall variability over the Nordeste and Sahel regions as discussed in section 4. The dipole pattern is characterized by SST anomalies of opposite sign on either side of the Equator. It shows up as one of the leading two EOFs of tropical Atlantic SSTs, or from joint analyses of SST and surface winds, and it exhibits considerable variability on time scales of 8-16 years. However, in such analyses, the leading EOFs are not well separated, so the possibility of degeneracy and effective mixing of modes exists.

Indeed, rotated EOF analyses of tropical Atlantic SSTs yield discernibly different results from their unrotated counterparts. In particular, the results suggest that most of the variance in SST north and south of the ITCZ is not correlated at zero lag. These results are supported by simple one-point correlation maps, which show little evidence of a dipole structure in the original SST data. In fact, tropical North and South Atlantic SST anomalies of opposite sign and of absolute magnitude 0.2C or larger occur no more frequently (15% occurrence rate) than expected by chance. This does not mean, however, that significant fluctuations in the cross-equatorial SST gradient do not occur, nor that such fluctuations are unimportant. Analyses of historical instrumental SST records show that significant meridional gradients, in the absence of dipoles, occur in nearly 50% of all months over the tropical Atlantic.

7. Atlantic ENSO

There is some evidence of a coupled atmosphere-ocean mode of variability in the equatorial Atlantic that is dynamically akin to ENSO. As in the Pacific, eastern equatorial Atlantic SSTs are dominated by the seasonal cycle. Here, SSTs are warmest during northern spring, when the equatorial winds are weak and the thermocline is deep. As the year progresses, the southeasterly trade winds strengthen, the thermocline depth decreases and the SSTs cool. Superimposed on this annual cycle are SST anomalies that can exceed 0.5C, especially during boreal summer. Warm equatorial Atlantic SST anomalies are associated with relaxed easterlies and a southward displacement of the ITCZ, and the overall spatial patterns bear a strong resemblance to those over the tropical Pacific associated with ENSO. However, the smaller amplitude of the fluctuations and the smaller percentage of variance explained by the equatorial Atlantic mode relative to ENSO reinforces the view that the tropical Atlantic is affected much more than the tropical Pacific by remotely forced variability or by other, local modes that are dynamically distinct from the equatorial variations.