VOL. 16, NO. 7
JOURNAL OF CLIMATE
1 APRIL 2003
The Seasonal Evolution of the Atmospheric Circulation over West Africa and
Equatorial Africa
SHARON E. NICHOLSON
AND
JEREMY P. GRIST
Department of Meteorology, The Florida State University, Tallahassee, Florida
(Manuscript received 8 December 2001, in final form 23 September 2002)
ABSTRACT
This paper examines the mean annual cycle of rainfall and general circulation features over West Africa and
central Africa for 1958–97. Rainfall is examined using a 1400-station archive compiled by the first author. Other
circulation features are examined using the NCEP–NCAR reanalysis dataset. Important features of the reanalysis
zonal wind field are shown to compare well with the seasonal evolution described by the radiosonde observations.
In addition to the well-known African easterly jet (AEJ) of the Northern Hemisphere, the seasonal evolution
of its Southern Hemisphere counterpart is also described. Thermal wind calculations show that although the
southern jet is weaker, its existence is also due to a local reversal of the surface temperature gradient. In the
upper troposphere, a strong semiannual cycle is shown in the 200-mb easterlies and a feature like the tropical
easterly jet (TEJ) is evident south of the equator in January and February. The paper describes the movement
of the rainbelt between central and West Africa. An asymmetry in the northward and southward migration of
the rainbelt is evident.
The paper discusses the influence that the jets may have on rainfall and possible feedback effects of rainfall
on the jets. Evidence suggests that the midtropospheric jets influence the development of the rainy season, but
also that the rainfall affects the surface temperature gradient and in turn the jets. In the Northern Hemisphere,
east of 208E, the axis of the TEJ is located so that it may promote convection by increasing upper-level divergence.
However, west of 108E and in the Southern Hemisphere, the location of the TEJ is consistent with the suggestion
that it is the equatorward outflow of convection that produces the TEJ.
1. Introduction
Sahelian West Africa is one area of the world that
has undergone a remarkable change in climate. Thirtyyear rainfall means have declined by 30%–40% between
1931–60 and 1968–97. The protracted dry spell since
the late 1960s has prompted study of the region’s climate, as has the Global Atmospheric Research Program
(GARP) Atlantic Tropical Experiment (GATE), which
took place in 1974. Since that time, numerous papers
have increased our understanding of rainfall variability
in the Sahel. These have shown, for example, that the
ENSO signal and tropical Atlantic sea surface temperatures (SSTs) have a strong enough correlation with
Sahelian rainfall to be useful predictors on seasonal
timescales (Semazzi et al. 1988; Ward 1994; Moron and
Ward 1998; Lamb 1978; Rowell et al. 1995). In addition,
systematic differences in the basic state between wet
and dry years have been noted with some consistency.
In wet (dry) years the midtropospheric African easterly
jet (AEJ) tends to be weaker (stronger) and more poleward (equatorward), while the 200-mb tropical easterly
Corresponding author address: Dr. Jeremy P. Grist, James Rennell
Division, Southampton Oceanography Centre, European Way, Southampton SO14 3ZH, United Kingdom.
E-mail: [email protected]
q 2003 American Meteorological Society
jet (TEJ) and the monsoonal flow tend to be stronger
(weaker) (Newell and Kidson 1984; Fontaine et al.
1995; Kanimitsu and Krishnamurti 1978; Grist and
Nicholson 2001). There is less agreement as to the role
that these features play in rainfall variability. Similarly,
differences in the intertropical convergence zone (ITCZ)
between wet and dry years have been noted, but there
is some disagreement as to whether variations in its
location or intensity are more important in determining
rainfall variability (Kraus 1977a,b; Ilesanmi 1971; Miles
and Folland 1974; Tanaka et al. 1975; Nicholson 1980;
Citeau et al. 1989; Ba et al. 1995). Thus, the fundamental reasons for the region’s unique rainfall variability have not been conclusively determined.
In contrast to the wealth of studies over West Africa,
there have been relatively few studies of equatorial Africa (CLIVAR 2000). Our present understanding of meteorological processes in the region is correspondingly
weak. Notable studies by Laing and Fritsch (1993) and
Mohr and Zipser (1996) have described the spatiotemporal distribution of mesoscale convective systems.
Apart from this, little is known about either the synoptic
or mesoscale systems that shape the rainy season or
about the factors governing interannual and interdecadal
variability. Recent Tropical Rainfall Measuring Mission
(TRMM) data have shown that the frequency of light-
1013
1014
JOURNAL OF CLIMATE
ning there is by far the greatest in the world. State-ofthe-art satellite algorithms overestimate rainfall in this
region by a factor of 2 or 3 (McCollum et al. 2000;
Nicholson 2000), suggesting a poor understanding of
rain processes in the region. In this context it is interesting to note that extensive biomass burning in this
region produces high smoke concentrations, a factor
shown by Rosenfeld (1999) to suppress convective rainfall. The region clearly merits further study as even basic
aspects of the general atmospheric circulation are poorly
understood.
Considering the gap in the quantity of research existing between West and central or equatorial Africa, it
is ironic that the regions exhibit a number of similar or
common circulation features. Thus, much stands to be
gained by studying the two regions in parallel. This
point has been missed, however, by most previous studies. An example is the midtropospheric easterly jet in
the Southern Hemisphere. A number of authors [beginning with Burpee (1972)] have acknowledged the existence of this jet, but neither its annual cycle nor its
structure has been described in detail. Likewise, its influence on regional meteorology has not been considered.
In this article we begin to address this deficit in the
literature by jointly examining the annual cycles of circulation features over West Africa and central Africa.
This allows us to better describe the basic state of the
atmosphere in these regions, something that is of fundamental importance to modeling efforts. Our interpretation of the results also offers some clarity to the meteorological processes at work, particularly in central
Africa. This can facilitate the interpretation of modeling
results and the design of forecast methods.
This article makes several new contributions to the
understanding of the meteorology of these regions. Our
work provides the first analysis of the annual cycles of
the midtropospheric easterly jets and the TEJ over Africa. Since past studies of these features have generally
been limited to the local rainy season, usually only the
boreal summer (and often only August) has been examined. Here we show, counter to previous assumptions, that the TEJ is neither restricted to the summer
months nor to the Northern Hemisphere and that the
AEJs exist during most months. We also demonstrate a
characteristic weakening of the northern AEJ in midsummer. By using the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis dataset that was recently made available, we are able to describe the important features of the wind and temperature fields in a
more spatially and temporally consistent manner than
has been done previously. This includes an analysis of
the associated wind shears. In recognizing the possible
shortcoming of the NCEP reanalysis data, we also use
radiosonde observations to confirm some of the results
based on NCEP. This combination allows us to relate
the evolution of various circulation features and thermal
VOLUME 16
FIG. 1. Location of the study regions West Africa (W) and
equatorial Africa (C).
fields to one another and to suggest a physical basis for
this evolution and the interrelationships. We also present
the first analysis of the seasonal evolution of the precipitation field in the context of the various dynamical
features.
The structure of this article is as follows. In section
2 an overview of the dynamical systems affecting West
Africa and central Africa is presented. Section 3 describes the data and methodology. Sections 4, 5, 6, and
7 describe the results of the analysis of zonal wind,
temperature, meridional wind, and the rainbelt, respectively. In section 8 there is some discussion about the
evolution of important dynamical features and their interrelationships. Section 9 presents the summary and
conclusions.
2. Overview of dynamical systems influencing the
study region
a. West Africa
The area described as West Africa is defined in Fig.
1. The rainfall field over West Africa is characterized
by a zone of maximum precipitation that migrates north
and south throughout the course of the year. This zone
lies to the south of the surface position of the ITCZ and
its seasonal excursion roughly parallels the seasonal excursion of the ITCZ. The rainbelt is the loci of disturbances that are dynamically linked to the midlevel AEJ.
Westward-propagating mesoscale disturbances, termed
cloud clusters, are the dominant convective system. Certain environmental factors, such as vertical wind shear,
1 APRIL 2003
NICHOLSON AND GRIST
buoyant energy, low-level jets, and latitude, determine
whether convection organizes into these long-lived systems (Laing and Fritsch 1993). The frequency of cloud
clusters and the amount of rainfall associated with them
is modulated by transient synoptic-scale African or easterly waves (Houze and Betts 1981; Thompson et al.
1979).
These waves originate as a consequence of a joint
baroclinic–barotropic instability associated with the vertical and horizontal shear of the midtropospheric AEJ.
The waves, as well as the cloud clusters, are generally
confined to a relatively narrow latitudinal zone near and
south of the jet (Laing and Fritsch 1993; Norquist et al.
1977; Albignat and Reed 1980; Burpee 1972). This zone
corresponds to the region between the axes of the AEJ
and the upper-tropospheric TEJ (Tourre 1981; Bounoua
1980). Their growth and development is influenced by
several factors, including the magnitude of the horizontal and vertical shear of the AEJ (e.g., Rennick 1981),
and latent heat release (Norquist et al. 1977).
Both observations and models have shown that characteristics of the AEJ (particularly horizontal and vertical shear) and the mean zonal flow influence the characteristics of these waves (Burpee 1974; Simmons 1977;
Mass 1979; Rennick 1976, 1981; Kwon 1989). Although the shear instability associated with the jet is
present throughout the rainy season, the waves appear
to contribute to the development of large-scale rainfall
systems only during late summer (Reeves et al. 1979;
Chen and Ogura 1982; Miller and Lindzen 1992). For
the waves to organize rainfall, they must produce moisture convergence and be reasonably close vertically to
the moist layer. Past studies have suggested that in the
Sahel, these prerequisites are present only in late summer. This fact may help to explain the contrast between
the rainfall regimes in June–July and August–September, noted in several studies, and the fact that mesoscale
convective complexes are much more frequent during
the latter period (Laing and Fritsch 1993), when wave
amplitude is larger (Duvel 1990; Grist 2002).
Squall lines, which may develop from isolated convection or from cloud clusters, are also important convective features. They likewise appear to be related to
wave activity (e.g., Reed et al. 1977; Payne and
McGarry 1977), but the association is inconsistent (e.g.,
Bolton 1981, 1984; Rowell and Milford 1993). The generation of squalls requires conditional instability, vertical wind shear in the lower troposphere, a shallow
moist layer with dry air aloft, and probably a minimum
intensity of the AEJ. Growth may be triggered by surface heating, topography or large-scale, low-level convergence, as is found just ahead of the wave trough
(Bolton 1981; Mansfield 1977; Rowell and Milford
1993).
Another dominant feature of the circulation during
the West African rainy season is the TEJ. During the
height of the rainy season it is of the order 20 m s 21 at
200 mb and 78N. Although it tends to be much stronger
1015
in wet years, there appear to be some difference of opinions as to its relationship with rainfall. Druyan and Hall
(1996) and Reiter (1969) suggest that the increased divergence of its exit region might promote convective
rainfall, while Thorncroft and Blackburn (1999) concluded that the TEJ was a response to the convective
rainfall.
It is evident that the dynamical features influencing
rainfall in the region exhibit strong seasonality. Consequently, a description of the annual cycle should be
helpful in understanding rainfall variability in the region.
b. Central Africa
In this article, we use the term central Africa to denote
the central equatorial Tropics of Africa, ranging from
approximately 58N to 108S and extending westward to
Cameroon and eastward to the Rift Valley highlands.
The geographical limits described here and indicated in
Fig. 1 are basically pragmatic. Areas somewhat farther
north (Central Africa Republic, Chad, and Sudan) and
west (Nigeria and other countries of the Guinea Coast)
are meteorologically controlled by the same processes
and features that influence Sahelian West Africa and
peak rainfall occurs in the boreal summer, when the
ITCZ moves to its far northern limits. In contrast, central
Africa experiences a strongly bimodal annual cycle, coincident with both the northward and southward passage
of the ITCZ. Peak rainfall tends to be in the two transition seasons. This zone may have many common features with eastern equatorial Africa, generally termed
East Africa. However, the latter has been comparatively
well studied and also contrasts with central Africa in
some key climatic features, such as proximity to the
Atlantic versus Indian Oceans, the occurrence of rainfall
in the boreal summer, and the influence of the Rift Valley
highlands and the barrier they provide between the two
regions.
Rainfall variability in the region shows some commonalities with that over East Africa. The variance spectra for rainfall throughout the region show very strong
peaks on the order of 5–6 yr and generally smaller peaks
at roughly 3.5 and 2.3 yr, the same peaks found in the
Southern Oscillation (Nicholson and Entekhabi 1986).
The variability shows strong links to both the Southern
Oscillation and to SST variability in the eastern Atlantic
(Nicholson and Entekhabi 1987). The variability is generally out of phase with that in the Sahel [i.e., dry (wet)
Sahel corresponds to wet (dry) Central Africa]. Although wave activity has not been identified, rainfall
tends to be organized into mesoscale convective systems
(MCSs) analogous to those in Sahelian West Africa
(Laing and Fritsch 1993; Mohr and Zipser 1996).
Very little is known about the synoptic situations and
atmospheric dynamics that control regional rainfall. A
few studies concentrating on southern Africa (e.g., Kumar 1978) have shown dynamical teleconnections be-
1016
JOURNAL OF CLIMATE
tween that region and central Africa, suggesting at least
some degree of midlatitude influence from the Southern
Hemisphere.
3. Data and methodology
The NCEP–NCAR reanalysis project dataset (Kalnay
et al. 1996) is utilized for this study. It consists of daily
and monthly gridded data (u, y , Z, etc.) at 2.58 3 2.58
horizontal resolution with 17 vertical levels. This is better spatial resolution than provided by Oort’s (1983)
global gridded analysis, which has previously been used
to describe the region. It is hoped that this increased
resolution will afford a more realistic description of the
regional circulation features. The dataset is currently
available from 1948 to 1997. However, when our analyses were carried out, the dataset extended back only
to 1958.
The NCEP reanalysis data are derived as a combination of observations and model results, based on the
NCEP global operational model. The observational data
consist primarily of global radiosonde data, Comprehensive Ocean–Atmosphere Data Set (COADS) surface
marine data, aircraft data, surface land synoptic data,
Special Sensor Microwave Imager (SSM/I) surface wind
speeds, satellite cloud drift winds, and satellite sounder
data. Details on the data and their quality control, the
model, data assimilation techniques, and limited validation are given in Kalnay et al. (1996).
Grist and Nicholson (2001) utilized the NCEP reanalysis data for a study that contrasted four wet years
(1958–61) and four dry years (1982–85) in the Sahel.
In view of the limited availability of observational data
in the analysis sector and of the discontinuity in the
NCEP data when satellite data were introduced in the
late 1970s, a limited comparison with actual pibal and
radiosonde data was made. Many features of the wind
field indicated by the NCEP data for wet and dry years
were confirmed by the actual wind data. These included
the annual cycle of intensity and seasonal excursion of
the jet core, and the generally weaker AEJ during wet
years. However, the reanalysis depicted anomalously
strong and deep westerlies in wet years. The magnitude
of this contrast could not be confirmed by the other
sources of wind data available. This study deals only
with the mean annual cycle; the ability of the reanalysis
to capture interdecadal variability should not be a concern. Because of this, and because there are some important upper-air observations in equatorial southern Africa early in the reanalysis period, NCEP data for the
period 1958–97 were used in the study.
The first step in this study is to examine the mean
zonal wind field for the years 1958–97. An analysis is
carried out for two sectors: 108W–108E and 108–308E.
The western sector includes both the continent and the
Atlantic Ocean to the south, while the eastern extends
only through the continent and includes equatorial Africa. By analyzing the two different sectors it was hoped
VOLUME 16
to gain insight into any longitudinal variation in variables that are attributable to the land–sea contrast in the
Southern Hemisphere. Characteristics of interest include
the general structure, the intensity and location of the
various jet streams, their seasonal development and associated horizontal and vertical wind shear.
In order to highlight possible problems in the reanalysis, key features of the wind field are compared to
radiosonde observations. The radiosonde observations
were dataset DS353.0 taken from NCAR data archive.
A summary of this is included in Fig. 2a. There was
sufficient data to enable the 700-mb latitudinal profile
for east and west sectors to be compared with the reanalysis for 1962–72. It is noted, however, that there is
very poor coverage south of the equator in the western
sector. This is important to remember when viewing
results from this area. For the 200-mb level, where fewer
observations were available, only the annual cycles for
key stations could be derived (not shown).
The same sections were used to produce zonal means
of the meridional wind and temperature fields for 1958–
97. The dynamical features are then related to the development of the rainy season via a comparison of their
annual cycles with the mean ‘‘structure’’ of the rainbelt
over West Africa. To do this, meridional cross sections
depicting rainfall as a function of month and latitude
are derived from a rainfall dataset (Fig. 2b) assembled
by Nicholson (1986) and recently updated (e.g., Nicholson et al. 2000).
4. The mean zonal wind field
Mean zonal winds for the two sectors are shown in
Figs. 3a,b. Three easterly jets are evident over West
Africa, an upper-tropospheric easterly jet (the TEJ)
around 200 mb and two midtropospheric jets near 650
mb. One of these is the well-known AEJ of the Northern
Hemisphere. The second is a Southern Hemisphere jet
and has not yet been described in detail. Here we use
the terms AEJ-N and AEJ-S to distinguish them. Their
annual cycles are also summarized in Fig. 4. Further
apparent are the two upper-tropospheric westerly jets of
the midlatitudes and a low-level westerly wind maximum near the equator.
In the western sector (Fig. 3a) the AEJ-N is discernable throughout the year, but during the Northern Hemisphere (NH) winter, when it is located around 08–58N,
it is relatively weak. It strengthens during the NH summer and also shifts to 108–128N. Core speeds range from
about 8 m s 21 in October–March to 12 m s 21 in June.
The altitude of the core also changes during the year.
It is at 600 mb in April–September, when the jet is
relatively strong. The core is closer to 700 mb during
the months of October–March, when the AEJ-N is relatively weak.
The AEJ-S is most clearly defined during September–
November. During this time its location is at 700 mb
between 58 and 108S. Mean core speeds reach 8 m s 21
1 APRIL 2003
NICHOLSON AND GRIST
FIG. 2. Distribution of (a) radiosonde and (b) rainfall stations used
in the study. (a) The western and eastern sectors used in the study
are indicated.
in October and November, but only 4–6 m s 21 in the
other months. Hence, it is weaker than the AEJ-N. During December–March there is a low-level (900 mb) easterly maximum at around 208S, but at other times of the
1017
year (April–July) this feature is not very distinct and
no semblance of the AEJ-S appears to be a discernible.
In the eastern analysis sector (Fig. 3b) the annual
cycle of these jets is similar, although some differences
are notable. The AEJ-N is somewhat weaker in this
sector, its core speeds barely reaching 10 m s 21 in May
and June, then again in September. The jet disappears
in November and its location in the NH winter ranges
from about 38 to 58N. Its core is near 600 mb in all
months, except October, when it lies near 650 mb.
The AEJ-S is more pronounced in the eastern sector.
It is best defined in August–November. Its position ranges from about 58S in August to 88S in November. Core
speeds range from about 6 m s 21 in August to 10 m s 21
in October. Some semblance of a midtropospheric jet is
apparent in December–March, but in April–July it is
replaced by a weak lower-level easterly maximum near
800 mb.
The horizontal shear at 600 mb was calculated as the
difference in zonal wind (Fig. 3a) between adjacent grid
points, which are spaced at 2.58 intervals. The variation
of horizontal shear with latitude and month is shown in
Fig. 5a. The seasonal migration of the area of maximum
shear is clearly associated with the migration of the AEJN. Cyclonic shear south of the jet reaches 10–12 3
10 26 s 21 in the eastern sector. The anticyclonic shear
north of the jet, where there is a rapid transition to
westerly flow, reaches 14 3 10 26 s 21 in April and May.
The vertical shear associated with the AEJ was calculated as the difference in the zonal wind velocities between 700 and 850 mb. The seasonal migration of the
maximum in the vertical shear is determined by the
seasonal strengthening and migration of the AEJ-N and
the low-level westerlies. It is at its most poleward in
August. The vertical shear beneath the jet core is around
5–6 3 10 24 m s 21 Pa 21 much of the year, but it appears
to peak at around 7 3 10 24 m s 21 Pa 21 in June. Both
the horizontal and vertical shears associated with AEJS are considerably weaker than the AEJ-N. This reflects
both the weaker midtropospheric jet and the weaker
westerlies in the Southern Hemisphere.
Figure 6 compares the radiosonde and reanalysis
mean annual cycle of zonal wind at 700 mb for both
sectors. In general, there is very good agreement between the two. The magnitude, location, and nature of
seasonal migration of the AEJ are very close. The good
agreement partly reflects the fact that the reanalysis has
been produced by the assimilation of radiosonde (and
other observations) into the NCEP model. The area of
weaker easterlies on the equatorward side of the AEJN is also evident in both datasets. It is noticeable that
there is increased variability in the radiosonde values at
this latitude. This may indicate some longitudinal variation in the wind field. Despite the fact that there are
fewer radiosondes in the SH, they still show the formation of the jet around August–September. The radiosondes show the AEJ-S weakening fairly dramatically
in November and continuing to move poleward. In the
FIG. 3. (a) Mean zonal wind speed (m s 21 ) for Jan–Dec, averaged between 108W and 108E for 1958–97. (b) Mean zonal wind speed (m s 21 ) for Jan–Dec, averaged between 108 and 308E
for 1958–97.
1018
JOURNAL OF CLIMATE
VOLUME 16
1 APRIL 2003
NICHOLSON AND GRIST
1019
FIG. 5. Annual cycle of zonal wind shear. (top) Horizontal shear
of the zonal wind at 600 mb in the sector from 108W–108E; units
are 10 26 s 21 and the solid contours are for negative values. (bottom)
Vertical shear of the zonal wind, averaged for the layer from 850 to
700 mb in the sector from 108W–108E; units are 10 24 m s 21 Pa 21
and the solid contours are for positive values.
FIG. 4. Mean intensity and location of jet core (monthly averages)
of the (a) AEJ-N and (b) AEJ-S for both sectors. The dashed lines
indicate standard deviations from the monthly mean.
eastern sector, the reanalysis depicts this evolution differently with the weakening occurring about two months
later than the radiosondes. According to both reanalysis
and radiosondes the AEJ-S has all but disappeared by
April.
The annual cycle of the magnitude of the TEJ features
a distinctive double maximum (Fig. 7). A northern summer maximum occurs in July, when the TEJ is most
northward. A second, weaker (around 14 m s 21 ) maximum occurs in February, when the TEJ is close to its
most southern latitude. As it moves north and crosses
the equator, its magnitude decreases to around 6 m s 21 .
At its most northerly point, the magnitude of the TEJ
is at around 19 m s 21 . The jet weakens as it moves back
south. According to Fig. 3b, the TEJ is not a distinct
feature in November, but reappears at 88S and 8 m s 21
in December. Through most of the year the TEJ is stronger in the eastern sector.
The apparent appearance of the TEJ in the Southern
Hemisphere is interesting. The existence of upper-level
easterly jets near (Davies and Sanson 1952) and just
south (Bond 1953) of the equator has been noted for
some time and they are considered part of the Walker
circulation. These Southern Hemisphere easterlies are
weaker and not considered part of the TEJ (Reiter 1969).
The TEJ is believed to owe its existence to the presence
of the Tibetan high and the large meridional temperature
gradient that forms between the Tibetan highlands and
1020
JOURNAL OF CLIMATE
VOLUME 16
FIG. 6. Zonal wind at 700 mb between 208S and 308N for (a) western sector and (b) eastern sector. Dashed lines
connect values at radiosonde stations and solid lines represent NCEP reanalysis. Values represent the means between
1962 and 1972, with the exception of St. Helena Island. For this station means for the years 1983–94 were used as a
proxy. Because of the paucity of data the radiosonde line is not extended south of the equator in the western sector
(observations from individual stations are marked as crosses).
the Indian Ocean during the northern summer (Koteswaram 1958). Thus, it is usually considered to be restricted to the NH and the boreal summer (e.g., Hastenrath 1988). However, the steady seasonal migration
of the easterly wind maximum in Fig. 7 suggests the
same process operates year-round. This and the importance of the longitudinal variation in the jets will be
discussed in section 8c.
The equatorial westerlies (Figs. 3a,b) are better developed in the western analysis sector than in the eastern, because they result primarily from the southeasterly
trades associated with the South Atlantic high, which
become westerly upon crossing the equator. This system
is less extensive in the continental sector farther east.
The equatorial westerlies are best developed in July–
September.
FIG. 7. Annual cycle of the 200-mb easterly wind (108–308E) maximum. (Top) intensity, (bottom) location. Dashed lines indicate standard deviations from the monthly mean.
5. Zonal mean temperature
Figure 8 shows a meridional cross section of zonally
averaged temperature in the eastern sector (Fig. 8a) and
the western sector (Fig. 8b) during each month of the
year. Considering the eastern sector, we note that the
temperature field is characterized by weak latitudinal
gradients in the mid troposphere and upper troposphere
from about 208S to 208N, and two temperature maxima
in the lower troposphere in subtropical latitudes. The
maximum in the NH migrates from 108N in NH winter,
when maximum temperatures reach 300 K, to 228N in
NH summer, when temperatures exceed 308 K. The
maximum in the SH is generally between to 158 and
208S. The maximum is most pronounced in October,
with maximum temperatures of over 300 K at 158S. It
is least pronounced between April and June when a
maximum of less than 300 K is spread over some 208
of longitude.
The NH maximum reflects the intense heating of the
Sahara and its southern margin, while its seasonal shift
is associated with the seasonal excursion of sub-Saharan
rainbelt (see section 7). The temperature gradient resulting from this low-level maximum (Fig. 9) reaches
a maximum of 10 K (1000 km) 21 in June, when the jet
is most intense and the vertical shear is strongest. Similarly, the maximum in the SH is indicative of the warming of the air above the semiarid Kalahari, prior to the
onset of rains.
There is a notable difference between the eastern and
western sectors in the SH, which appears to reflect the
land–sea contrast. The western region has a low-level
temperature maximum at around 108S. This is probably
FIG. 8. (a) Vertical cross sections of mean temperature (K) for Jan–Dec, averaged between 108 and 308E (eastern sector), 1958–97; (b) vertical cross sections of mean temperature (K) for
Jan–Dec, averaged between 108W and 108E (western sector), 1958–97.
1 APRIL 2003
NICHOLSON AND GRIST
1021
1022
JOURNAL OF CLIMATE
due to the advection of air warmed by the land to the
east. However, the fact that the temperature maximum
is elevated to around 850 mb, in August–December appears to indicate that the surface air is also modified by
the relatively cool SSTs.
It has been claimed that the AEJ-N forms as a result
of the reversal of temperature gradient. We investigated
this, and also whether it is the case for the AEJ-S, by
comparing the zonal wind shear with the thermal wind
(U t) between 600 and 925 mb. The zonal wind shear
was calculated as the difference in zonal wind between
600 and 925 mb. Values were calculated for all months,
for the eastern and western sectors. The monthly means
of the basic states from the western and eastern sectors
were used in the calculation. The thermal wind was
calculated from the following formula:
Ut 5
R dT 600 mb
ln
,
f dy 925 mb
where R is the gas constant 287 J deg 21 kg 21 , f is the
Coriolis parameter at the appropriate latitude, and T is
the mean zonally averaged temperature for the layer
between 600 and 925 mb.
If the temperature gradients are responsible for the
jets, a strong linear correlation between thermal wind
and observed wind shear near the jet location would be
expected. However, the values would not be comparable
close to the equator, where the Coriolis parameter goes
to 0. For this reason, the calculations were restricted to
the latitudes 8.758–13.758N and 8.758–13.758S. In addition, only winds of magnitude greater than 2 m s 21
were plotted. The average temperature for the layer
600–925 mb was estimated by a polynomial interpolation of the temperature profile to the midpoint (762.5
mb) of the layer.
The upper three panels of Fig. 10 shows the results
of the comparison of the wind shear and the thermal
wind for the western sector for 1) both the Northern
and Southern Hemispheres, 2) just the Northern Hemisphere, and 3) just the Southern Hemisphere. The lower
three panels show the same but for the eastern sector.
All graphs show strong correlations, with greater than
99% significance. Thus our calculations support the hypothesis that the vertical wind shear (i.e., the AEJ-N)
is caused by the temperature gradient. In addition, the
calculations further suggest that although the AEJ-S is
weaker, it likewise forms as a result of the local reversal
of the surface temperature gradient. In light of this it is
interesting to note that the slope in Fig. 10f for the
western sector in the Southern Hemisphere is much
smaller than the others. The western SH section is oceanic. Therefore, the difference in the thermal wind–wind
shear relationship may be because the meridional surface temperature gradient is weaker. Alternatively, it
may be because the paucity of upper-air observations
in the sector make the thermal wind relationship difficult
to characterize.
VOLUME 16
FIG. 9. Mean temperature gradient [K (1000 km) 21 ] at 850 mb,
averaged for 108–308E.
The annual cycle of the AEJ closely resembles that
of the maximum temperature gradient (MTG) at 850
mb. Figures 11 and 12 indicate that differences in the
surface temperature gradient between the NH and SH
are not just in magnitude. Their annual cycles also exhibit different characteristics. Figure 11 shows that in
the NH, the poleward migration of the MTG is nearly
coincident with its strengthening at the peak of the boreal summer. By contrast, shortly after the southern
MTG moves poleward the MTG weakens substantially.
We suggest that the differences in the annual cycles
are due to the varying influences of rainfall. Figure 12
shows the annual cycle of rainfall and 850 mb temperature at different latitudes. It is evident in both hemispheres that the rainfall moderates the annual cycle of
temperature. This is presumably by increasing the ratio
of latent to sensible heat transfer from the surface, although increased cloudiness may also be influential. In
the Southern Hemisphere there is more rainfall poleward
of the MTG and, hence, more cooling. As a consequence, the MTG (and, in turn, the AEJ-S) weakens.
However, poleward of the MTG in the NH (208N) there
is little rain and little moderation of temperature. Thus,
unlike the SH after the onset of the rains, the MTG
continues to increase until it is near its maximum poleward location.
6. Mean meridional winds and the ITCZ
Figure 13 shows the mean meridional wind component in the western sector. All 12 months are depicted.
In all months a quadrant pattern depicting a Hadley cell
circulation is evident. The reanalysis shows that the meridional circulation strengthens in the winter hemisphere, in agreement with previous observations (e.g.,
Lau 1984). However, this strengthening is less evident
for the Southern Hemisphere. Lindzen and Hou (1988)
show that this asymmetry in the circulation occurs when
thermal heating is displaced from the equator.
The southerly flow near the surface and in the lower
atmosphere reflects the southeasterly trade winds on
the equatorward flank of the South Atlantic high. It
generally extends to 750 or 800 mb and is evident in
all months throughout the Southern Hemisphere lat-
1 APRIL 2003
NICHOLSON AND GRIST
1023
FIG. 10. Monthly mean values of zonally averaged vertical wind shear (600–925 mb) against
thermal wind for 8.758N, 11.258N, 13.758N, 8.758S, 11.258S, and 13.758S: (a) AEJ-N and AEJS, eastern sector; (b) AEJ-N, eastern sector; (c) AEJ-S, eastern sector; (d) AEJ-N and AEJ-S,
western sector; (e) AEJ-N, western sector; and (f ) AEJ-S western sector.
itudes of the analysis sector. Its northernmost position
reflects the position of the ITCZ. The flow is strongest
close to the surface, with maximum speeds of generally 3 or 4 m s 21 . It is also strongest during the
months of April–August, when peak flow is just to
the south of the ITCZ. Using the surface transition
between southerly and northerly winds as an indicator
of ITCZ position, it is apparent that the ITCZ migrates
from around 98N in January and February to 208N in
July and August.
In the eastern sector (not shown), the meridional flow
is generally more weakly developed than in the western
FIG. 11. (top) Annual cycle of 850-mb MTG for Northern Hemisphere (108–308E), location in degrees latitude north (dashed lines)
and magnitude (solid lines). (lower) Same for southern 850-mb MTG,
only it is shown with a 6-month lag to aid comparison with Northern
Hemisphere and the location is in degrees latitude south. Units of
MTG are K (1000 km) 21 .
sector, described above. As with the equatorial westerlies, the explanation lies in the fact that this system
is best developed over the ocean, rather than over the
continental sector that dominates the eastern sector at
these latitudes.
7. The structure of the rainbelt
Figures 14a,b show rainfall as a function of latitude
and month for the areas 108W–108E and 108–308E, respectively. These figures are constructed from station
rainfall at all stations within the averaging sector. The
general characteristics of the tropical rainbelt shown in
the cross section are manifestations of the large-scale
dynamics governing rainfall. Two main features are evident in both cross sections: the latitudinal excursion of
the rainbelt during the course of the year and the asymmetry of the northward and southward portions of this
migration. Some other features, however, are representative of regional peculiarities. For example, the maximum near 58N in June in the western sector is a result
of the geographical configuration of the continent: at
this latitude, rainfall over the land area is dominated by
a small area of peninsulas and highlands in eastern Nigeria and coastal Cameroon that experience extremely
heavy rainfall between May and July. Likewise, the relatively low January and February rainfall from roughly
58 to 118S is related to coastal effects of the Benguela
Current and not large-scale dynamics.
1024
JOURNAL OF CLIMATE
VOLUME 16
sector, the southward excursion is rapidly accelerated
from September to October, abruptly ending the rainy
season in the central Sahel/Soudan zone. In the eastern
sector, the biggest shift occurs between October and
November, so that the eastern Soudan zone receives
significant October rainfall.
The rainbelt is considerably more ‘‘intense’’ (in terms
of rainfall per month) during the southward excursion.
In the western sector (Fig. 14a) rainfall is on the order
of 350–400 mm month 21 from September to November
in a large sector within the latitudinal belt 48–18N. Rainfall does not exceed 350 mm month 21 during the northward excursion. In the eastern sector (Fig. 14b) the area
with rainfall in excess of 200 mm month 21 is likewise
limited to the latter half of the year. As discussed in
section 8b this asymmetry may be related to the appearance of both midtropospheric jets during the southward excursion. The intensity suggested by the cross
sections in Figs. 14a,b is a function of the strength of
the rain-producing processes and the time that a location
is affected by the core of the rainbelt. Because of the
more rapid southward excursion, there is a shorter, more
intense rainy period during the southward excursion
than that suggested from the monthly totals at individual
stations. Over the tropical Americas, Horel et al. (1989)
noticed a similar asymmetry in the timing (but not the
intensity) of the march of the convective rains.
The classic interpretation of the seasonal excursion
of the rainbelt is that it represents the northward and
southward movements of the ITCZ. A comparison
with the ITCZ, as defined by wind convergence, is
discussed in the following section. A somewhat alternate explanation for the seasonal excursion of the
rainbelt and a hypothesis concerning its asymmetry
are also presented.
8. Discussion
FIG. 12. Mean annual cycle (1958–97) of 850-mb temperature
(dashed lines, right axis) and precipitation (solid lines, left axis) (108–
308E) for (a) 208, 158, 108, 58N, and (b) 208, 158, 108, 58S. Units are
8C for temperature and mm for rainfall.
The northward excursion occurs during the six
months from February to August, when the axis of the
rainbelt migrates from roughly 98 or 108S to 118N (88N
in the eastern sector). The southward excursion is more
rapid, with the axis migrating back to 108S by November, that is, within 4 months. This asymmetry was noted
earlier by Nicholson (1981) and is evident throughout
the east–west extent of the continent. In the western
To examine the relationship between the annual cycle
of the AEJ-N and the rainbelt, the mean locations of
both are plotted in Figs. 15a,b. These are compared as
well with the location of the strongest horizontal temperature gradient and with the ITCZ. The ITCZ location
is defined by the shift from northerly to southerly meridional flow.
In the western sector, the seasonal excursion of each
of these features is quite similar. The asymmetry of the
northward and southward excursion noted earlier for the
rainbelt is evident in all of these indices. The jet core
lies roughly 48–58 south of the surface convergence, and
the maximum temperature gradient lies in between. The
rainfall maximum is about 58 south of the AEJ in most
months (it is about 108 to the south for July–September).
In the eastern sector (Fig. 15b), the seasonal excursion
of each of these parameters is nearly the same as in the
western. Although the strong similarity of the seasonal
excursions of the wind and temperature parameters in
both sectors could be an artifact of the model used to
1 APRIL 2003
NICHOLSON AND GRIST
1025
FIG. 13. Mean meridional wind speed (m s 21 ) for Jan–Dec, averaged between 108W and 108E,
the western sector.
produce the reanalysis data, the rainfall data are completely independent. The major trends seen in the other
parameters, especially the asymmetry of the northward
and southward excursions and the small points of contrast between the eastern and western sectors, are also
apparent in the rainfall data. Thus, the geographical associations among these parameters appear to be real
climatological features of the atmosphere.
fails to account for the asymmetry of the rainbelt during
the northward and southward excursion, with the most
intense rainfall occurring from August to November, and
for the preferential occurrence of large and well-organized disturbances (MCCs) late in the rainy season (Duvel 1990; Laing and Fritsch 1993).
a. The ITCZ
It is interesting to note that the rainbelt is most intense
when both the AEJ-N and the AEJ-S are well developed.
During this time, late in the Sahelian rainy season, there
is a preferential occurrence of large convective events
and the rainbelt lies roughly between the two axes of
the jets (Figs. 14a,b). There is thus circumstantial evidence that the juxtaposition of the two jets plays a role
in determining the dynamics of the rainfall regime and
the development of the rainbelt during August–November. Although it should be acknowledged that the wind
shear associated with the AEJ-S is relatively weak, the
most intense mesoscale convective systems in Southern
Hemisphere Africa are on the equatorward side of this
jet (Laing and Fritsch 1993; Mohr and Zipser 1996) and
these large, organized disturbances are approximately
The parallel seasonal excursions of the ITCZ, AEJ,
and rainbelt suggest a fundamental relationship between
all of these parameters. Although Figs. 15a,b show parallels, in the mean, between the ITCZ (i.e., the surface
convergence) and the location of the rainbelt, there is
clear evidence that the two are decoupled on interannual
timescales (Grist and Nicholson 2001). The location of
the ITCZ shows little contrast between the wet and dry
year composites analyzed in that study, while the location of the core of the AEJ-N varies greatly, as does
the location of the rainbelt. Both are displaced equatorward during dry years over West Africa. Attributing
the position of the ITCZ as a control on rainfall also
b. African easterly jets
1026
JOURNAL OF CLIMATE
VOLUME 16
limited to the latitudes between the AEJ-N and the AEJS during the seasons when both jets are present (Fig.
16). Moreover, the MCSs appear to be quite rare in East
Africa, in the longitudes where the AEJ-S appears to
disappear. Consequently, this connection may be worthy
of further exploration.
It has generally been accepted (e.g., Krishnamurti and
Pasch 1982; Thorncroft and Blackburn 1999) that the
origin of the AEJ-N is basically the temperature gradient
induced by the contrast between the Sahara and the
humid Guinea coast region to the south. Our results
further confirm this and also demonstrate that the origin
of the AEJ-S is also related to the surface temperature
gradient. In this case, the temperature contrast is produced by semiarid regions of the Southern Hemisphere,
compared to the subhumid, perennially vegetated lands
to the north.
As suggested in section 5, this implies a feedback
between the rainfall regime and the jets. In the case of
the AEJ-N, the maximum temperature gradient (and jet
core) lies in proximity to the boundary between the dry
region north of the jet and the area near and south of
the core where rainfall is occurring. The effect of the
rainfall is to reduce surface temperatures (as a consequence of soil moisture and rapidly growing vegetation
cover). This, together with the northward progression
of the solar maximum, displaces the maximum temperature gradient (and the jet core) farther northward.
Analogous arguments were used by Webster (1983) to
explain the northward advance of the Indian monsoon.
In the case of the AEJ-S, there is greater rainfall on the
poleward side of the jet. Consequently, the poleward
surface temperature cools more and the feedback appears to work in the opposite sense.
There is considerable asymmetry in the northward
and southward excursion of the AEJ-N (and rainbelt).
This may be partially a consequence of the asymmetry
in the surface heating over West Africa. During the
northward excursion, which is prior to the Sahel/Soudan
rainy season, the region is dry and, hence, comparatively
warm. During the rainy season, the juxtaposition of the
summer heating of the Sahara and the relatively moist
and vegetated surface over the Sahel maintains the temperature maximum at high latitudes. In September, the
Sahara rapidly cools as the solar maximum moves to
the equator and the rainy season ends, rapidly drying
the Sahelian land surface. Consequentially, the latitude
of maximum temperature (and maximum temperature
gradient) shifts abruptly southward after September.
←
FIG. 14. Cross section of mean rainfall (mm month 21 ) as a function
of month and latitude, averaged for (a) western sector, 108W–108E
(183 stations are used to produce this cross section), and (b) 108–
308E (420 stations are used to produce this cross section). The location
of the AEJ-N and AEJ-S is indicated by the black solid lines. Because
of the lack of stations (Fig. 2b) the contours are not plotted south of
the equator in the western sector.
1 APRIL 2003
NICHOLSON AND GRIST
FIG. 15. Latitudinal location of the AEJ, center of gravity of the
rainbelt over Africa, MTG; and ITCZ: (a) western sector (108W–
108E) and (b) eastern sector (108–308E).
c. Tropical easterly jet
We further examine the relationship between the seasonal cycles of rainfall and the TEJ in Fig. 17. The
figure shows the precipitation and the zonal wind at 200
mb for August and February 1989. The zero contour of
the du/dy field near the easterly flow is included to show
the axis of the TEJ. The distinctive Asian and West
African–Atlantic branches of the TEJ are evident. Although this is just an example from one year, the features
of relevance are common to most years. In fact, the
August plot is very similar to Fig. 8 in Koteswaram
(1958). We identify three distinct regions (A, B, and C).
Region A, central Africa, is in the left exit region of
the main Asian branch of the TEJ. It is therefore likely
to be a region of enhanced upper-level divergence that
could promote convective activity, as suggested by Reiter (1969). This suggests that the variability of the TEJ
may influence rainfall variability in equatorial Africa.
This is not the situation in region B, West Africa (west
of 108E), where the main area of convective activity is
to the north of the jet axis. The location of convective
activity relative to the TEJ is consistent with the work
of Thorncroft and Blackburn (1999). They attribute the
formation of the TEJ to the easterly deflection of the
southward outflow of convection due to the Coriolis
1027
FIG. 16. Jan and Oct MCSs over Africa, based on their 85-GHz
ice scattering signatures from SSM/I data during sunset passes (from
Mohr and Zipser 1996). Data are for 1993 only. The approximate
mean location of the AEJ-S and AEJ-N cores has been added, showing
that most of the MCSs lie between the cores.
force. The implication is that in region B, variability in
the TEJ is a response to, and not a cause of, rainfall
variability. This is supported by differences maps of
wind field between West African wet and dry years.
These show that the largest differences in the TEJ are
west of 08E. For the TEJ to promote convective activity
in this region, it would have to be through enhancing
convection as the right entrance region of the West African–Atlantic branch of the TEJ, as suggested by Druyan (1998).
Region C, in the Southern Hemisphere, is not usually
considered part of the TEJ. However, it appears to be
the Southern Hemisphere version of the process Thorncroft and Blackburn (1999) described. Convection located around the southeastern coast of Africa creates an
upper-level outflow. The northward branch of the outflow is deflected westward by the Coriolis force. This
explains the seasonal migration of the 200-mb easterly
wind maximum (Fig. 7) and why it reaches a minimum
as it crosses the equator.
Summing up the TEJ discussion, it is noted that the
strong zonal flows near regions B and C are a response
to convective activity. The convective activity in region A, however, may be enhanced or promoted by the
upper-level divergence associated with the TEJ (Asian
branch) exit region. It is also noted that the features
of the 200-mb zonal wind described here and the mech-
1028
JOURNAL OF CLIMATE
FIG. 17. Contours of 200-mb u for 1989; the solid line is the zero contour of the du/dy field
(indicating the TEJ axis). Shading represents precipitation (mm day 21 ). Wind data are from
NCEP and rainfall data are from first authors archive: (a) Aug, and (b) Feb. Significance of
letters is explained in the text.
VOLUME 16
1 APRIL 2003
1029
NICHOLSON AND GRIST
anisms attributed to them are not covered in classic
descriptions of the TEJ (e.g., Koteswaram 1958; Hastenrath 1988).
9. Summary and conclusions
The annual cycle of the basic state over West Africa
and central Africa has been described in some detail.
The NCEP reanalysis of zonal wind compared well with
available radiosonde observations, adding confidence to
the climatologies produced. The increased resolution of
the dataset provided clear depiction of such features as
the AEJ-N, AEJ-S, ITCZ, TEJ, and the monsoonal flow.
The seasonal migration of these features is asymmetric
in that the northward excursion occurs much more slowly than the southward. This creates a similar asymmetry
in the seasonal cycle of rainfall, with the rainy season
taking on a very abrupt end. There is also an asymmetry
in the intensity of the rainbelt, which is substantially
more intense during the months when both the AEJ-N
and the AEJ-S are present (i.e., August–November). The
belt of intense rains is roughly bound by the cores or
the two jets during this period. The cores also roughly
define the zones where mesoscale convective systems
occur over equatorial Africa. This suggests that the AEJs
play an important role in the development of the rainy
season.
Observations and calculations support the claim that
the AEJ-N is formed from the local reversal of the surface temperature gradient. This study also shows that,
although weaker, the AEJ-S is formed in a similar manner. A comparison of the MTGs in the Northern and
Southern Hemispheres revealed that precipitation appears to moderate the MTG. It thus, in turn, moderates
the annual cycle of the jets. In the Northern Hemisphere,
rainfall moderates the surface temperature at the latitude
and south of the MTG but not over the practically rainless Sahara. The net effect is that the jet strengthens as
it is displaced north after the onset of the rains. In the
Southern Hemisphere the rainfall is more abundant poleward of the MTG. Consequently, shortly after the onset
of the rains and as the jet begins to move poleward, it
weakens to the point where it is not a very distinct
feature.
From the above discussion it is clear that there is
considerable feedback between the two African easterly
jets and the tropical rainbelt. Not only does the rainy
season influence the jets via the MTGs, but we suggest
that the jets also modulate the development of the rainy
season. The clearest example is that the AEJ-N controls
the development and organization of rain-bearing systems via the generation of waves. This may also be the
case with the AEJ-S, but that has not yet been established.
Our results similarly suggest that the tropical rainbelt
also has some influence on the tropical easterly jet at
200 mb. The Asian branch of the TEJ is forced remotely
by the Tibetan high. This may act as a mechanism of
rainfall variability over central Africa, which is in the
jet’s left exit region. Over West Africa, the TEJ seems
at least partly a response to the equatorward outflow of
convection just to its north. This process would appear
to explain the TEJs distinct semiannual cycle, its weakening over the equator, and its appearance in the Southern Hemisphere in January and February.
Acknowledgments. This study was supported by NSF
Grant ATM-9521761 (Climate Dynamics), with some
contributions from Florida State University for the support of Grist. The assistance of NCAR Scientific Computing Division in accessing and retrieving data from
their archives is acknowledged.
REFERENCES
Albignat, J. P., and R. J. Reed, 1980: The origin of African wave
disturbances during Phase III of GATE. Mon. Wea. Rev., 108,
1827–1839.
Ba, M. B., R. Frouin, and S. E. Nicholson, 1995: Satellite-derived
interannual variability of West African rainfall during 1983–88.
J. Appl. Meteor., 34, 411–431.
Bolton, D., 1981: Generation and propagation of African squall lines.
Ph.D. Thesis, Imperial College, London, 265 pp.
——, 1984: Generation and propagation of African squall lines.
Quart. J. Roy. Meteor. Soc., 110, 695–721.
Bond, H. G., 1953: Easterly jet streams over Darwin. Weather, 8,
252–253.
Bounoua, L., 1980: Relation entre champ de vent, genese et progations des lignes de grains ouest Africaines. Ph.D. thesis, Institute
Hydrometeorologique de Formation et de Recherches, 108 pp.
[Available from IHFR BP 7019, Seddikia, Oran, Algeria.]
Burpee, R. W., 1972: The origin and structure of easterly waves in
the lower troposphere. J. Atmos. Sci., 29, 77–90.
——, 1974: Characteristics of North Africa easterly waves during
the summers of 1968 and 1969. J. Atmos. Sci., 31, 1556–1570.
Chen, Y., and Y. Ogura, 1982: Modulation of convective activity by
large-scale flow patterns observed in GATE. J. Atmos. Sci., 39,
1260–1279.
Citeau, J., L. Finaud, J. P. Cammas, and H. Demarcq, 1989: Questions
relative to ITCZ migrations over tropical Atlantic ocean, sea
surface temperature and Senegal river runoff. Meteor. Atmos.
Phys., 41, 181–190.
CLIVAR, 2000: CLIVAR Africa Implementation Plan. WCRP Informal Report 5/2000, ICPO Publ. Ser. 35, Int. CLIVAR Project
Office, Southampton Oceanography Centre, Southampton, United Kingdom, 33 pp.
Davies, D. A., and H. W. Sansom, 1952: Easterly jets over East
Africa. Weather, 8, 343–345.
Druyan, L. M., 1998: The role of synoptic systems in the interannual
variability of Sahel Rainfall. Meteor. Atmos. Phys., 65, 55–75.
——, and T. M. Hall, 1996: The sensitivity of African wave disturbances to remote forcing. J. Appl. Meteor., 35, 1100–1110.
Duvel, J. P., 1990: Convection over tropical Africa and the Atlantic
Ocean during northern summer. Part II: Modulation by easterly
waves. Mon. Wea. Rev., 118, 1855–1868.
Fontaine, B., S. Janicot, and V. Moron, 1995: Rainfall anomaly patterns and wind field signals over West Africa in August (1958–
1989). J. Climate, 8, 1503–5101.
Grist, J. P., 2002: Easterly waves over Africa. Part I: The seasonal
cycle and contrasts between wet and dry years. Mon. Wea. Rev.,
130, 197–211.
——, and S. E. Nicholson, 2001: A study of the dynamic factors
influencing the variability of rainfall in the West African Sahel.
J. Climate, 14, 1337–1359.
1030
JOURNAL OF CLIMATE
Hastenrath, S., 1988: Climate and Circulation in the Tropics. D.
Reidel, 455 pp.
Horel, J. D., A. N. Hahmann, and J. E. Geisler, 1989: An investigation
of the annual cycle of convective activity over the tropical Americas. J. Climate, 2, 1388–1403.
Houze, R. A., Jr., and A. K. Betts, 1981: Convection in GATE. Rev.
Geophys. Space Phys., 19, 541–576.
Illesanmi, O. O., 1971: An empirical formulation of an ITD rainfall
model for the tropics: A case study in Nigeria. J. Appl. Meteor.,
10, 882–891.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.
Kanamitsu, M., and T. N. Krishnamurti, 1978: Northern summer tropical circulations during drought and normal rainfall months.
Mon. Wea. Rev., 106, 331–347.
Koteswaram, P., 1958: The easterly jet stream in the tropics. Tellus,
10, 43–57.
Kraus, E. B., 1977a: Subtropical droughts and cross-equatorial energy
transports. Mon. Wea. Rev., 105, 1009–1018.
——, 1977b: The seasonal excursion of the intertropical convergence
zone. Mon. Wea. Rev., 105, 1052–1055.
Krishnamurti, T. N., and R. J. Pasch, 1982: Large-scale mean state. The
GARP Atlantic Tropical Experiment (GATE) Monographs, GARP
Publ. Ser. 25, World Meteorological Organization, 183–218.
Kumar, S., 1978: Interaction of Upper Westerly Waves with Intertropical Convergence Zone and Their effect on the Weather over
Zambia During the Rainy Season. Government Printers, 36 pp.
Kwon, H. J., 1989: A reexamination of the genesis of African waves.
J. Atmos. Sci., 46, 3621–3631.
Laing, A., and J. M. Fritsch, 1993: Mesoscale convective complexes
in Africa. Mon. Wea. Rev., 121, 2254–2263.
Lamb, P. J., 1978: Large-scale tropical Atlantic surface circulation
patterns associated with subsaharan weather anomalies. Tellus,
30, 240–251.
Lau, N.-C., 1984: Circulation statistics based upon FGGE level IIIB analyses produced by GFDL. NOAA Data Rep. ERL GFDL5, 427 pp.
Lindzen, R. S., and A. Y. Hou, 1988: Hadley circulations for zonally
averaged heating centered off the equator. J. Atmos. Sci., 45,
2416–2427.
Mansfield, D. A., 1977: Squall lines observed in GATE. Quart. J.
Roy. Meteor. Soc., 103, 569–574.
Mass, C., 1979: A linear primitive equation model of African wave
disturbances. J. Atmos. Sci., 36, 2075–2092.
McCollum, J. R., A. Gruber, and M. B. Ba, 2000: Discrepancy between gauges and satellite estimates of rainfall in equatorial
Africa. J. Appl. Meteor., 39, 666–679.
Miles, M. K., and C. F. Folland, 1974: Changes in the latitude of the
climatic zones of the Northern Hemisphere. Nature, 252, 616.
Miller, R. L., and R. Lindzen, 1992: Organization of rainfall by an
unstable jet with an application to African waves. J. Atmos. Sci.,
49, 1523–1540.
Mohr, K. I., and E. J. Zipser, 1996: Mesoscale convective systems
defined by their 85-GHz ice scattering signature: Size and intensity comparison over tropical oceans and continents. Mon.
Wea. Rev., 124, 2417–2437.
Moran, V., and M. N. Ward, 1998: ENSO teleconnections with climate
variability in the European and African sectors. Weather, 53,
287–295.
Newell, R. E., and J. W. Kidson, 1984: African mean wind changes
between Sahelian wet and dry periods. J. Climatol., 4, 1–7.
Nicholson, S. E., 1980: The nature of rainfall fluctuations in subtropical West Africa. Mon. Wea. Rev., 108, 473–487.
——, 1981: Rainfall and atmospheric circulation during drought periods and wetter years in West Africa. Mon. Wea. Rev., 109,
2191–2208.
——, 1986: The spatial coherance of African rainfall anomalies: In-
VOLUME 16
terhemispheric teleconnections. J. Climate Appl. Meteor., 25,
1365–1381.
——, 2000: Validation of TRMM and other satellite estimates of
rainfall over selected regions of Africa. Preprints, 15th Conf. on
Hydrology, Long Beach, CA, Amer. Meteor. Soc., 363–366.
——, and D. Entekhabi, 1986: The quasi-periodic behavior of rainfall
variability in Africa and its relationship to the Southern Oscillation. Arch. Meteor. Geophys. Bioclim., 34A, 311–348.
——, and ——, 1987: Rainfall variability in equatorial and Southern
Africa: Relationships with sea-surface temperatures along the
southwestern coast of Africa. J. Climate Appl. Meteor., 26, 561–
578.
——, B. Some, and B. Kone, 2000: An analysis on recent rainfall
conditions in West Africa, including the rainy season of the 1997
ENSO year. J. Climate, 13, 2628–2640.
Norquist, D. C., E. E. Recker, and R. J. Reed, 1977: The energetics
of African waves disturbances as observed during phase III of
GATE. Mon. Wea. Rev., 105, 334–342.
Oort, A., 1983: Global atmospheric circulation characteristics: 1958–
73. NOAA Prof. Pap. 14, 180 pp (147 microfiches).
Payne, S., and M. McGarry, 1977: The relationship of satellite inferred convective activity to easterly waves over West Africa
and the adjacent ocean during the phase II of GATE. Mon. Wea.
Rev., 105, 413–420.
Reed, R., D. Norquist, and E. Recker, 1977: The structure and properties of African wave disturbances as observed during phase
III of GATE. Mon. Wea. Rev., 105, 413–420.
Reeves, R., C. Ropelewski, and M. Hudlow, 1979: Relationships
between large-scale motion and convective precipitation during
GATE. Mon. Wea. Rev., 107, 1154–1168.
Reiter, E., 1969: Tropospheric circulation and jet streams. Climate of
the Free Atmosphere, Vol. 4, D. F. Rex, Ed., World Survey of
Climatology, Elsevier, 85–203.
Rennick, M. A., 1976: The generation of African waves. J. Atmos.
Sci., 33, 1955–1969.
——, 1981: Some sensitivity experiments with an African wave model. J. Atmos. Sci., 38, 106–113.
Rosenfeld, D., 1999: TRMM observed first direct evidence of smoke
from forest fires inhibiting rainfall. Geophys. Res. Lett., 26,
3105–3108.
Rowell, D. P., and J. R. Milford, 1993: On the generation of African
squall lines. J. Climate, 6, 1181–1193.
——, C. K. Folland, K. Maskell, and M. N. Ward, 1995: Variability
of summer rainfall over tropical North Africa (1906–1992); observations and modelling. Quart. J. Roy. Meteor. Soc., 121, 669–
704.
Semazzi, F. H. M., V. Mehta, and Y.-C. Sud, 1988: An investigation
of the relationship between sub-Saharan rainfall and global sea
surface temperatures. Atmos.–Oceans, 26, 118–138.
Simmons, A. J., 1977: A note on the instability of the African easterly
jet. J. Atmos. Sci., 34, 1670–1674.
Tanaka, M., B. C. Weare, A. R. Navato, and R. E. Newell, 1975:
Recent rainfall patterns. Nature, 255, 201–203.
Thompson, R. M., Jr., S. W. Payne, E. E. Recker, and R. J. Reed,
1979: Structure and properties of synoptic-scale wave disturbances in the intertropical convergence zone of the eastern Atlantic. J. Atmos. Sci., 36, 53–72.
Thorncroft, C. D., and M. Blackburn, 1999: Maintenance of the African easterly jet. Quart. J. Roy. Meteor. Soc., 125, 763–786.
Tourre, Y. M., 1981: The squall line over West Africa and tropical
eastern Atlantic ocean during GATE. Ph.D. dissertation, University of Virginia 147 pp. (ISBN 82–08502).
Ward, M. N., 1994: Tropical North African rainfall and worldwide
monthly to multidecadal climate variations. Ph.D. thesis, University of Reading, United Kingdom, 313 pp.
Webster, P. J., 1983: Mechanisms of monsoon low-frequency variability: Surface hydrological effects. J. Atmos. Sci., 40, 2110–
2124.