The West African Monsoon Dynamics. Part II: The ‘‘Preonset ... · monsoon, deﬁning the...

1 NOVEMBER 2003 3407S U L T A N A N D J A N I C O T

q 2003 American Meteorological Society

The West African Monsoon Dynamics. Part II: The ‘‘Preonset’’ and ‘‘Onset’’ of theSummer Monsoon

BENJAMIN SULTAN AND SERGE JANICOT

LMD/IPSL, CNRS, Ecole Polytechnique, Palaiseau, France

(Manuscript received 24 October 2002, in final form 21 May 2003)

ABSTRACT

The arrival of the summer monsoon over West Africa has been documented by using daily gridded rainfalldata and NCEP–NCAR reanalyses during the period 1968–90, and OLR data over the period 1979–90. Twosteps have been characterized through a composite approach: the preonset and the onset of the summer monsoon.

The preonset stage corresponds to the arrival in the intertropical front (ITF) at 158N, that is, the confluenceline between moist southwesterly monsoon winds and dry northeasterly Harmattan, bringing sufficient moisturefor isolated convective systems to develop in the Sudano–Sahelian zone while the intertropical convergencezone (ITCZ) is centered at 58N. The mean date for the preonset occurrence is 14 May and its standard deviationis 9.5 days during the period 1968–90. This leads to a first clear increase of the positive rainfall slope corre-sponding to the beginning of the rainy season over this Sudano–Sahelian area.

The onset stage of the summer monsoon over West Africa is linked to an abrupt latitudinal shift of the ITCZfrom a quasi-stationary location at 58N in May–June to another quasi-stationary location at 108N in July–August.The mean date for the onset occurrence is 24 June and its standard deviation is 8 days during the period 1968–90. This leads to a second increase of the positive rainfall slope over the Sudano–Sahelian zone signing thenorthernmost location of the ITCZ and the beginning of the monsoon season. This abrupt shift occurs mostlybetween 108W and 58E, where a meridional land–sea contrast exists, and it is characterized by a temporaryrainfall and convection decrease over West Africa. Preonset dates, onset dates, and summer rainfall amount overthe Sahel are uncorrelated during the period 1968–90.

The atmospheric dynamics associated with the abrupt ITCZ shift has been investigated. Between the preonsetand the onset stages, the heat low dynamics associated with the ITF controls the circulation in the low andmidlevels. Its meridional circulation intensity is the highest at the beginning of the monsoon onset. This canlead to 1) increased convective inhibition in the ITCZ through intrusion of dry and subsiding air from the north,and 2) increased potential instability through a greater inland moisture advection and a higher monsoon depthinduced by a stronger cyclonic circulation in the low levels, through higher vertical wind shear due to westerlymonsoon wind and midlevel African easterly jet (AEJ) increases, through enhancement of the instability characterof the AEJ, and through increased shortwave radiation received at the surface. During the monsoon onset, oncethe rainfall minimum occurred due to the convective inhibition, the accumulated potential instability breaks theconvective inhibition, the inertial instability of the monsoon circulation is released, and the associated regional-scale circulation increases, leading to the abrupt shift of the ITCZ. Then the ITCZ moves north up to 108N,where thermodynamical conditions are favorable.

It is suggested by the authors that the abrupt shift of the ITCZ, initiated by the amplification of the heat lowdynamics, could be due to an interaction with the northern orography of the Atlas–Ahaggar Mountains. Subsidenceover and north of this orography, due to both the northern branches of the heat low and of the northern Hadley-type cell, contributes to enhance the high geopotentials north of these mountains and the associated northeasterlywinds. This leads to the development of a leeward trough that reinforces the heat low dynamics, maintainingan active convective ITCZ through enhanced moist air advection from the ocean, increasing the northern Hadleycirculation, which reinforces the high geopotentials and the interaction with the orography through a positivefeedback. The fact that an abrupt shift of the ITCZ is only observed on the western part of West Africa mayresult from the enhancement of moisture advection, which comes from the west and has a stronger impact westof the Greenwich meridian.

The northwest–southeast orientation of the Atlas–Ahaggar crest can induce the interaction with the heat low,first in the east where the mountains are nearer to the ITF than in the west, and second in the west. Anotherconsequence of the possible orography-induced interaction with the atmospheric circulation is that the inducedleeward trough, increasing the cyclonic vorticity in the heat low, may stimulate moisture convergence in theoceanic ITCZ near the western coast of West Africa.

Corresponding author address: Dr. Serge Janicot, Laboratoire deMeteorologie Dynamique, Ecole Polytechnique, 91128 PalaiseauCedex, France.E-mail: [email protected]

1. IntroductionRainfall over West Africa is controlled by the ad-

vection of moisture from the Gulf of Guinea in the lowlevels of the atmosphere. Following the seasonal ex-

3408 VOLUME 16J O U R N A L O F C L I M A T E

cursion of the sun, the monsoon develops over this partof the African continent during the northern spring andsummer, bringing the intertropical convergence zone(ITCZ) and the associated rainfall maxima to theirnorthernmost location in August (Hastenrath 1995).This is the time for the rainy season in the Sahel. Ag-riculture in the Sudano–Sahelian zone is heavily de-pendent on the seasonal characteristics of rainfall, thatis, onset, length and termination of the rainfall season,seasonal rainfall amount, and intraseasonal rainfall dis-tribution during the rainy season. The onset of ‘‘useful’’rains, that is the first rains sufficient to ensure enoughmoisture in the soil at the time of the planting and notfollowed by prolonged dry spells that could prevent thesurvival of seedlings after sowing, is certainly the majorpoint for agriculturists. However, there is not a uniquedefinition for the date of the onset of the rainy season.It can be defined from traditional to semiempirical andscientific techniques, depending on local meteorologicalconditions, and leading for any particular year to a highdispersion of these dates for the same location (Ati etal. 2002).

The objective of this study is twofold. First, at a re-gional scale, we will characterize two main steps in theseasonal evolution of the monsoon over West Africa,that is, 1) what we call the ‘‘preonset’’ of the summermonsoon, defining the beginning of the rainy seasonover the Sudano–Sahelian zone based on the northwardmigration of the northern limit of the southwesterlywinds of the monsoon [called the intertropical front(ITF)], 2) the real ‘‘onset’’ of the summer monsooncharacterized by an abrupt northward shift of the ITCZfrom 58 to 108N (Sultan and Janicot 2000; Le Barbe etal. 2002), and leading to major changes in the atmo-spheric circulation over West Africa. Second, we willinvestigate in detail the atmospheric circulation changesassociated with the summer monsoon onset and we willpropose a mechanism to explain the occurrence of theabrupt shift of the ITCZ. Due to the availability of thedifferent datasets, this study is based on the period1968–90, a dry one over West Africa compared to thelong-term mean (Hastenrath 1995).

2. Datasets

Independent datasets have been used to investigatethe preonset and onset of the West African monsoon.Two of them describe rainfall and convection variability,two others are the U.S. and European reanalyses. Theydocument the period 1968–90 or the subperiod 1979–90. We will show that all the results obtained with thesedifferent datasets are strongly consistent. As they comefrom independent sources (rain gauge amounts, satellitemeasurements, atmospheric variables from radiosound-ings, pibals, etc.), it gives a high level of confidence toour results.

a. The IRD daily rainfall

Daily rainfall amount at stations located on the WestAfrican domain 38–208N, 188W–258E have been com-piled by the Institut de Recherche pour le Developpe-ment (IRD), the Agence pour la Securite de la Navi-gation Aerienne en Afrique et a Madagascar (ASECNA,and the Comite Interafricain d’Etudes Hydrauliques(CIEH). These data are available for the period 1968–90, including more than 1300 stations from 1968 to1980, and between 700 and 860 for the period 1981–90. These daily values were interpolated on the NationalCenters for Environmental Prediction–National Centerfor Atmospheric Research (NCEP–NCAR) 2.58 3 2.58grid, by assigning each station daily value to the nearestgrid point and averaging all the values related to eachgrid point. They were also interpolated in time, relatedto NCEP–NCAR daily wind fields since daily rainfallamounts were measured between 0600 LST of the dayand 0600 LST of the following day. We applied a timelag of 12 h between the average time of the NCEP–NCAR daily values (0900 UTC) and an approximatedaverage time of ‘‘daily’’ precipitation over the West Af-rican continent [2100 LST; Duvel (1989) indicates amaximum of high cloud coverage over land between1800 and 0000 LST. Sow (1997) points out a maximumof half-hourly precipitation over the Senegal between1700 LST and the end of the night, depending on thestations]. The greatest density of stations is located be-tween the latitudes 58–158N. Data on latitude 17.58Ncan also be taken into account since 30 to 45 stationsare available.

b. The NOAA/OLR dataset

Since 1974, launching polar orbital National Oceanicand Atmospheric Administration (NOAA) TelevisionInfrared Observation Satellite (TIROS) satellites hasmade it possible to establish a quasi-complete series oftwice-daily measures of outgoing longwave radiation(OLR), at the top of the atmosphere and at a resolutionof 2.58 latitude–longitude (Gruber and Krueger 1974).The interpolated OLR dataset (Liebmann and Smith1996) provided by the Climate Diagnostics Center hasbeen used here. In tropical areas, deep convection andrainfall can be estimated through low OLR values. Localhours of the measures varied during the period 1979–90 between 0230 and 0730 in the morning and between1430 and 1930 in the afternoon. Since the deep con-vection over West Africa has a strong diurnal cycle, thesample of daily OLR based on two values separated by12 h is enough to get a daily average. Moreover thisdataset has already been widely used for tropical studies.This dataset has been used over the period 1979–90 tovalidate and extend the results obtained with the IRDrainfall dataset, but also during the period 1979–2002to address a particular issue (see Fig. 5).


c. The NCEP–NCAR reanalyses

NCEP–NCAR have completed the first version of areanalysis project with a current version of the Medium-Range Forecast (MRF) model (Kalnay et al. 1996). Thisdataset consists in a reanalysis of the global observa-tional network of meteorological variables (wind, tem-perature, geopotential height, humidity on pressure lev-els, surface variables, and flux variables like precipi-tation rate) with a ‘‘frozen’’ state-of-the-art analysis andforecast system at a triangular spectral truncation of T62to perform data assimilation throughout the period 1948to present. This enables us to circumvent the problemsof previous numerical weather prediction analyses dueto changes in techniques, models, and data assimilation.Data are reported on a 2.58 3 2.58 grid every 6 h (0000,0600, 1200, and 1800 UTC), on 17 pressure levels from1000 to 10 hPa, which are good resolutions for studyingsynoptic weather systems. We used data covering theperiod 1 June–30 September, from 1968 to 1990, withone value per day by averaging the four outputs of eachday. An improved version of this reanalysis project isnow available (Kanamitsu et al. 2002) but it was notused here because we need to work on the longest avail-able dataset and this one only begins in 1979.

d. The ECMWF Reanalyses (ERA-15)

The European Centre for Medium-Range WeatherForecasts (ECMWF) completed a first reanalysis project(ERA-15), which used a frozen version of their analysis-forecast system, at a triangular spectral truncation ofT106 with 31 levels in the vertical, to perform dataassimilation using data from 1979 to 1993 (Gibson etal. 1997). Compared to NCEP–NCAR reanalyses, thereare 17 pressure levels from 1000 to 10 hPa, with anadditional level at 775 hPa and a missing one at 20 hPa.According to our objectives, daily data have been in-terpolated on the 2.58 3 2.58 NCEP–NCAR grid. TheERA-15 dataset has been used in comparison to theNCEP–NCAR dataset in order to evaluate the uncer-tainty of the wind field produced by the U.S. reanalyses.Previous studies using these two reanalysis datasets al-ready demonstrated their accuracy over West Africa fordescribing both interannual and synoptic timescale var-iability, if we consider the period after 1968 (Diedhiouet al. 1999; Poccard et al. 2000; Janicot et al. 2001).The comparisons done here also showed high consis-tency between these two reanalysis datasets, so onlyresults obtained from the NCEP–NCAR reanalysis,which covers the longest period 1968–90, will be shownin this paper.

3. The detection of the preonset and the onset ofthe summer monsoon

a. The mean rainfall fields

Figure 1 sums up the main features of the mean me-ridional excursion of the axis of maximum rainfall as-

sociated with the ITCZ, with the corresponding 925-hPa wind field superimposed. The black line over WestAfrica represents the zero isoline of the zonal compo-nent so as to delineate the domain of the monsoon winds(i.e., westerly component) and detect the location of theITF in the north. During April and the first half of May(Fig. 1a), the ITCZ is centered between the equator andthe southern coast of West Africa, extending over boththe Guinea gulf and the Guinean coast region south of108N. The rainfall field associated with the northern partof the ITCZ is captured by the IRD rain gauge datameasurements over the land and corresponds to the firstpart of the first rainy season over the Guinea coast re-gion. The isoline of 4 mm day21 is located around 88N.The corresponding wind field at 925 hPa is characterizedover West Africa by the confluence along the ITF ofthe southwesterly monsoon winds and the northeasterlydry winds called Harmattan. The monsoon winds arecontrolled by the pressure gradient between the lowpressures of the heat low centered along the ITF andthe oceanic high pressures of the Santa Helena anti-cyclone. Harmattan winds are controlled by the pressuregradient between the heat low and the Libyan andAzores anticyclones. The area delineated by the zeroisoline of the zonal wind component, chosen to char-acterize the monsoon winds, is limited at this stage ofthe seasonal cycle, south of 158N (the latitude of theITF). During the second half of May and in June (Fig.1b), the ITCZ remains at a quasi-stable location around58N. Convection increases over land with rainfall max-ima centered around the main orographic highs of Westand central Africa and the isoline 4 mm day21 locatedbetween 108 and 128N. This is the second part of thefirst rainy season over the Guinea coast region. The ITFis now positioned between 158 and 208N with its north-ernmost latitude between 08 and 58E. During the firsthalf of July (Fig. 1c), the ITCZ has already shifted tothe north, reaching a second quasi-stable state around108N, while rainfall maxima are still present at 58Nalong the Guinean coast. This transition stage is thenmarked by rainfall minima between 68 and 88N, in par-ticular over the Ivory Coast along 58W characterizinga climatological feature of the rainfall field on this areacalled the ‘‘V Baoule.’’1 At the same time, the area ofwesterly winds go on its extension over land and nowover the tropical Atlantic between 58 and 188N, the ITFreaching 208N at its northernmost latitude. Finally, dur-ing the second half of July and into August (Fig. 1d),the ITCZ remains along 108N with enhanced precipi-tation and the isoline of 4 mm day21 reaching 158N. Arainfall minimum is now evident along the Guineancoast region, associated to the so-called little dry season.The monsoon season is fully developed, due to the high-est pressures in the southern tropical Atlantic and the

1 This comes from a ‘‘V’’ shape of the annual rainfall isolines dueto a southward extension of low values along 58W.


FIG. 2. Mean 1968–90 rainfall time series (mm day21) computedon the grid points located between 108W and 108E, along 158N. Theline represents the time series filtered to remove variability lowerthan 10 days. The two arrows localize the breaks in the positiverainfall slope.

FIG. 1. Mean 1968–90 daily rainfall fields (mm day21) averagedrespectively (a) from 1 Apr to 15 May, (b) from 15 May to 30 Jun,(c) from 1 to 15 Jul, and (d) from 16 Jul to 31 Aug. Rainfall valuesare from individual rain gauges gridded on a 2.58 mesh. Only rainfallvalues greater than 4 mm day21 are colored. White areas are due tomissing data. The NCEP–NCAR 925-hPa wind field is expressed invectors and its scale (m s21) is displayed below. The black line rep-

←

resents the zero isoline of the zonal wind component to delineate thedomain of the monsoon wind. Thin curves display mean sea levelpressure isolines (expressed in hPa minus 1000).

northernmost location of the heat low over land. Thisis when the westerly wind area has its largest extension.

Figure 2 depicts the corresponding daily rainfall timeseries over the Sudano–Sahelian region, by averagingthe rainfall values for the grid points located between108W and 108E along 158N, that is over the longitudeband where the meridional land–sea contrast exists. Thefirst-order zonal symmetry of the mean rainfall fieldsand of the African monsoon system enables us to workon longitude-averaged variables. Figure 2 clearly showsthe progressive rainfall increase between April and Au-gust. It is however possible to detect two different stepsin this mean seasonal cycle, a first one around mid-Maywith a first increase of the positive rainfall slope, anda second one around late June with a second accelerationof the seasonal cycle leading to the rainfall maxima ofAugust. In the following section we will show that thefirst step is connected to the ITF crossing 158N, and tothe arrival of the monsoon winds advecting moist aironto the Sahelian latitudes. We will then consider thisstep as the preonset of the summer monsoon because itcorresponds to the beginning of the rainy season on theSudano–Sahelian region, a rainfall regime with a sig-nificant amount at the regional scale while the ITCZ isstill located at 58N. The rainfall increase of the Sudano–Sahelian zone then results from the intensification ofthe first rainy season over the Guinean coast region andthe northward extension of the rainbelt associated withthe ITCZ. We will also demonstrate that the second step,around late June, can be clearly associated with theabrupt northward shift of the ITCZ from 58 to 108N andwith major changes in the atmospheric circulation overWest Africa signing at this time the development of themeteorological summer monsoon system over West Af-rica. This is the reason why we will call this secondstep the summer monsoon onset.


FIG. 3. (a) Mean composite time series of daily rainfall (mm day21)averaged over 108W–108E along 158N, averaged over the period1968–90 by using as the reference date the date for each year whenthe (108W–108E) averaged zonal wind component at 158N equalszero going from negative to positive values. Bars represent the rawvalues and the curve the low-filtered values to remove rainfall var-iability lower than 10 days. Values are presented from t0 2 60 daysto t0 plus 170 days. (b) Same as (a) but for the unfiltered zonal windcomponent (black line) and the unfiltered wind modulus (dotted line),expressed in m s21. In both (a) and (b) the vertical line localizes thedate t0 (the mean date over the period 1968–90 is 14 May).

In the following two sections, we have applied a com-posite method to highlight these two steps. The idea fora better detection of the monsoon preonset (as well asthe onset) is that, as this step occurrence can have ahigh dispersion in time, a classic average of any me-teorological signal (as in Fig. 2 for instance) could betoo strongly smoothed due to this dispersion. So ourapproach has been to define the step occurrence for eachyear Y, called t0(Y), and then to compute the averageof any variable by taking the different t0(Y) as the samereference date t0 before performing the average. Wehope to enhance the rainfall and convection signals, aswell as any meteorological signal that could be tightlylinked to this step.

b. The composite signal of the summer monsoonpreonset

The meteorological signals associated to the preonsetstage of the summer monsoon are rather weak at a re-gional scale because the start of the rains over the Su-dano–Sahelian zone is seldomly abrupt and is usuallypreceded by a succession of isolated showers of uncer-tain intensity with intermittent dry periods of varyingduration (Ati et al. 2002). We based our approach onthe idea that local convection will begin to be initiatedmore or less regularly when a sufficient amount of mois-ture advected by the southwesterly monsoon winds willbe present over the Sahel. At the regional scale consid-ered here, the latitude of the ITF, represented by thenorthern boundary of the 925-hPa zonal wind zero iso-line, could be a good indicator of that. For each year Ybetween 1968 and 1990 we then define the date t0(Y)when the 925-hPa zonal wind component averaged over108W–108E equals zero at 158N, going from negativeto positive values. The zonal wind time series have beenfiltered to remove high-frequency fluctuations and weonly consider the evolution of a smooth seasonal cycle.This prevents us from defining too early a date due toa short wind fluctuation giving positive values and com-ing back rapidly to negative ones, meaning that the ITFis not yet maintaining north of 158N. We then computedthe averaged time series for the period 1968–90 by usingeach t0(Y) as the same reference date. We performedthe same averaging procedure for the wind modulus at158 and the rainfall at 158N.

Figure 3 shows the corresponding composite time se-ries computed between t0 2 60 days to t0 1 170 days.The average date t0 is the 14 May and the standarddeviation on the period 1968–90 is 9.5 days. Due to thedefinition of the method, we see in Fig. 3b that the datet0 corresponds to the reversal of sign of the zonal windcomponent at 925 hPa, going from negative values dueto the Harmattan occurrence to positive values due tothe monsoon winds occurrence, in association with thenorthward progression of the ITF. It also correspondsto a minimum of the wind modulus (Fig. 3b), and to amaximum of both the relative vorticity and the wind

convergence (not shown). The corresponding rainfalltime series (Fig. 3a) now depicts a clear break at t0 witha strong increase of the positive rainfall slope, meaningthat the arrival of the ITF at 158N can be considered ameaningful signal for the beginning of the rainy seasonover the Sudano–Sahelian zone.

c. The composite signal of the summer monsoononset

A similar approach has been used to highlight themeteorological signal associated with the summer mon-soon onset corresponding to the abrupt northward tran-sition of the ITCZ. This transition is highlighted whencomputing time–latitude diagrams of daily rainfall val-ues averaged over the longitudes 108W–108E. The zonaldistribution of rainfall over West Africa (Fig. 1) enablesus to use such diagrams without any significant loss ofinformation. Figure 4a depicts such an example for theyear 1978. This rapid shift of the ITCZ is clearly iden-tifiable on this type of diagram. It leads to a sharp re-versal of the meridional rainfall gradient over West Af-


FIG. 4. (a) Time–lat diagram from 1 Mar to 30 Nov 1978 of daily rainfall (mm day21), averaged over108W–108E and filtered to remove variability lower than 10 days. Values greater than 5 mm day21 are colored.(b) Time sections of diagram (a) at 58 (green curve), 108 (black curve), and 158N (red curve). In (a) and (b),the vertical line localizes the date selected for the ITCZ shift (17 Jun). (c) Composite time–latitude diagramof daily rainfall (mm day21) averaged over 108W–108E, filtered to remove rainfall variability lower than 10days, and averaged over the period 1968–90 by using as the reference date the shift date of the ITCZ foreach year. Values are presented from t0 (the shift date) 290 days to t0 1 140 days. Values greater than 5 mmday21 are colored. (d) Time sections of diagram (c) at 58 (green curve), 108 (black curve), and 158N (redcurve). In (c) and (d), the vertical line localizes the date of the ITCZ shift at t0 (the mean date over the period1968–90 is 24 Jun).

rica, which is also evident in Fig. 4b by the crossing ofthe rainfall time series at 58 and 108N.

As for the detection of the monsoon preonset, a quasi-objective method can be built up to define a date forthe ITCZ latitudinal shift for each year between 1968and 1990. An empirical orthogonal function (EOF) anal-ysis (Richman 1986) has been performed on time–lat-itude diagrams of daily rainfall values averaged over108W–108E, for each year from 1 March to 30 Novem-ber. Most of the rainfall variance decomposed by theEOF analysis is explained by the two first components.The first one (about 91% of the variance in 1968–90)is highly correlated with the rainfall time series at 108N(correlation of 0.9 in 1968–90) and the second one(about 9% of the variance in 1968–90) is highly cor-related with the rainfall time series at 58N (correlationof 0.75 in 1968–90). The rainfall indexes (i.e., the rain-fall time series) at 108 and at 58N can then be used tosum up rainfall variability over West Africa and to definea date for the ITCZ shift. For few years, especially thedry ones like 1983 or 1984, we must have consideredthe rainfall index at 7.58 instead of 108N because theaxis of maximum rainfall is located at a more southernlocation in summer. The time series of the rainfall in-

dexes at 58 and 108N for 1978 are shown in Fig. 4b. Arainfall maximum occurs during May–June when theITCZ is located at 58N. The abrupt shift of the ITCZfrom 58 to 108N can be defined by simultaneously, adecrease of the 58N rainfall index and an increase ofthe positive slope of the 108N rainfall index. Becauseof the lag between these two moments, an uncertaintyof a few days remains. So we look for an increase ofthe positive slope of a similar rainfall index at 158Nduring this time to specify an only date. For 1978 thedate of 17 June has been selected (see vertical line inFig. 4b).

This method has been used to define a date of theITCZ shift for each year from 1968 to 1990. The meandate t0 found for this shift over the period 1968–90 is24 June and the standard deviation is 8.0 days. Corre-lations computed between the preonset dates and theonset dates (r 5 0.01), as well as with the summerrainfall amount over the Sahel (r 5 20.16 for the preon-set dates, and r 5 20.21 for the onset dates) are notsignificant over the period 1968–90, indicating that in-dependent mechanisms control these different variables.To provide a more precise statistical distribution of theseITCZ shift dates, we have extended this analysis over


FIG. 5. Histogram of the ITCZ shift dates over the period 1968–2002 (bars; left scale) defined from the NOAA/OLR dataset, and thecumulative distribution function (line; right scale).

the period 1968–2002 by using the NOAA/OLR data.We first made sure that we can apply a similar approachto define these dates from OLR indexes at 58 and 108Nby verifying that we find dates similar or very close tothose found with the IRD rainfall indexes during thecommon period 1979–90, then we defined the ITCZshift dates from the OLR values over the period 1991–2002. Figure 5 shows the histogram of the ITCZ shiftdates over the period 1968–2002 as well as the corre-sponding cumulative distribution function. Over this pe-riod, the averaged date is the same (24 June) and thestandard deviation is very similar (7.0 days). This dis-tribution can be used as the first step of a probabilisticforecast of the summer monsoon onset date over WestAfrica, based on a 35-yr dataset specific to a dry periodcompared to the long-term mean (Hastenrath 1995). LeBarbe et al. (2002) showed that the mean date of theITCZ shift is not significantly different from the pre-vious wet period 1950–70. So our probability distri-bution function may also be valuable for a differentclimate state, but as the approach of Le Barbe et al. israther different from our approach, this point should beinvestigated more precisely; however, this is out of thescope of this paper. Hereafter, all the computations havebeen based on the period 1968–90 where the ITCZ shiftdates have been defined from the IRD rainfall indexes,because we get a more precise signal from daily rainfallamounts measured at the station network than fromtwice-daily OLR measurements. We checked howeverthat the conclusions infered from the results presentedbelow over the period 1968–90 are confirmed over theperiod 1968–2002.

Figure 4c shows the composite of the mean 108W–108E daily rainfall values averaged over the period 1968–90 by using the ITCZ shift date for each year as therespective reference date. This figure points out latitude–time rainfall variations between t0 (the shift date) minus90 days and t0 plus 140 days. Figure 4d shows the cor-

responding rainfall indexes at 58, 108, and 158N. Therainfall maximum at 58N, also evident at 108 and 158N,occurs about 10 days before t0. Then rainfall decreasesslightly, but over West Africa it only decreases to a rel-ative minimum value at the beginning of the ITCZ shift.At t0, the shift is detected by a new positive slope of therainfall index at 108 and at 158N. The ITCZ reaches thelatitude 108N about 10–20 days after t0 where rainfallincreases until mid-August. The withdrawal of the ITCZis quite sharp too, except that the June rainfall maximumat 58N has no counterpart in late summer when the mon-soon retreats equatorward and we do not observe anyrainfall minimum between the summer monsoon seasonand the second rainy season along the Guinea coast. Thisalso offers a different view from the progression of thenorthern limit of the ITCZ (see the 1–4 mm day21 iso-lines), which has a more gradual latitudinal variation dur-ing the onset than during the withdrawal.

Figure 6 shows similar composite time–latitude dia-grams but for OLR values and for meridional cross sec-tions at particular longitudes, from 208W to 108E. UsingOLR values enables us to test the previous results onan independent dataset and to consider a wider regionwithout any spatial gap. At 08 (Fig. 6c) and at 108W(Fig. 6b), the OLR patterns are very consistent with therainfall (Fig. 4c). Again we see the shift of the ITCZaround t0, from the latitude 58N at the time of the firstrainy season over the Guinea coast region to the latitude108N at the time of the summer monsoon season overthe Sudano–Sahelian region. This shift, as seen throughthe OLR values, is again concomitant with a temporarydecrease of convection over all the latitudes of WestAfrica at this longitude. Similar results have been ob-tained at 58E (not shown). At 108E (Fig. 6d), the OLRpattern is a bit different. At this longitude as well asfor more eastward longitudes crossing central Africa,the convection decrease at t0 is still evident but the ITCZshift does not occur anymore. The ITCZ progresses reg-ularly to the north before t0 and stays centered around68N after t0 while its northern boundary extends north-ward during the first half of summer. Finally at 208Wover the ocean, the northward excursion of the ITCZ iseven more regular with no latitudinal shift and no clearconvection decrease at t0. These results suggest that theITCZ shift leading to the abrupt summer monsoon onsetis working well on the longitudes where a land–seacontrast exists between 108W and 58E, that is where theconcept of the monsoon system, as a large-scale cross-equatorial atmospheric circulation from an oceanic basinon one side and a land area on the other side, is wellestablished. Outside this longitudinal band, the ITCZloses this particular behavior to show a rather progres-sive meridional excursion.

Although observed rainfall data are used in this work,the consistency of the seasonal cycle of precipitations inthe NCEP–NCAR and ECMWF reanalyses has been ex-amined to estimate the accuracy of the associated tem-poral fluctuations of the reanalysis winds (Sultan 2002;


FIG. 6. Composite time–lat diagrams of daily OLR (W m22), filtered to remove variability lower than 10 days, and averaged over theperiod 1979–90 by using as the reference date the ITCZ shift date for each year. Values are presented from t0 (the shift date) 90 days to t0

1 140 days. Values lower than 240 W m22 are shaded. (a) 208W; (b) 108W; (c) 08; (d) 108E. In (a)–(d) the vertical line localizes the dateof the ITCZ shift at t0 (the mean date over the period 1968–90 is 24 Jun).

not shown). It appears that both reanalyses produce avery good timing of the ITCZ shift, beginning at the datet0 defined from the IRD rainfall indexes; however, if theERA rainfall field exhibits a very clear abrupt shift, it isless marked in the NCEP–NCAR rainfall field becauseafter the monsoon onset, precipitation increases over theSahel band but it remains rather high along the Guineacoast and does not very clearly show the little dry seasonoccurrence in this area (not shown). We know that pre-cipitation is derived solely from the model fields forcedby the data assimilation (classified as the C-type variable)whereas the winds are strongly influenced by observeddata (classified as the A-type variable, the most reliableclass; Kalnay et al. 1996). So despite the slight discrep-ancy of the NCEP–NCAR rainfall fields, and also becausewe checked the consistency between the NCEP–NCARand ERA data, the following results based on the windfields can be considered to be reliable.

4. The atmospheric dynamics of the monsoon onset

In the previous section we have shown that it is nec-essary and possible to significantly differentiate what wecall the summer monsoon preonset (i.e., the beginning

of the rainy season over the Sudano–Sahelian zone char-acterized by the arrival of the ITF at 158N while theITCZ is still centered at 58N) and what we call the sum-mer monsoon onset (i.e., the ITCZ shift from 58 to 108N).The ITCZ shift appears as a main scientific issue for theunderstanding of the West African monsoon dynamicsbut it is also a key point for agriculturists. Ati et al.(2002), for instance, compared five methods to determinethe onset of the growing season in northern Nigeria. Theydetermined the best method to be a ‘‘hybrid’’ methoddefined to minimize the risk of false starts without short-ening the growing season beyond the necessary lengthfor the crop varieties selected. For an ensemble of eightstations located between 11.58 and 138N and using datacollected between 1961 and 1991 depending on the sta-tions, this optimal method provides 22 June as the meandate of the growing season, that is very close to our meandate for the monsoon onset (24 June). So it is necessaryto document this onset and to investigate some more ofthe possibly involved mechanisms.

a. A meridional viewWe have seen that it is possible to consider at a first

order the West African monsoon as a zonally symmetric


atmospheric system. So in this section, to complementthe results shown on rainfall and OLR (Figs. 4 and 6),we describe some of the main features of the monsoonseasonal cycle and its onset by using latitude cross sec-tions averaged over the longitude band 108W–108E. Fig-ure 7 depicts time–latitude diagrams for the relative vor-ticity at 925 hPa (Fig. 7a), the vertical velocity at 850hPa (Fig. 7b), the vertical velocity averaged on 700–500 hPa (Fig. 7c), the meridional wind component at600 hPa (Fig. 7d) and at 925 hPa (Fig. 7e), the zonalwind component at 600 hPa (Fig. 7f) and at 925 hPa(Fig. 7g), the vertical shear of the zonal wind between925 and 600 hPa (Fig. 7h), the top of the monsoon layer(Fig. 7i), and the precipitable water height (Fig. 7j) (allof these diagrams are expressed in the referential t0 290/t0 1 140, where t0 represents in average the 24 Junefor the monsoon onset).

The 925-hPa relative vorticity diagram (Fig. 7a) en-ables us to follow the meridional excursion of the ITFand of the heat low, expressed by positive maxima. Fromt0 2 90 to t0 2 10, the ITF moves gradually to thenorth, passing at 158N around t0 2 40, that is at thetime of the monsoon preonset as we previously saw (t0

2 40 is about the mean date defined for the preonset—14 May). From t0 2 10, the northward movement ofthe ITF accelerates, passing at 17.58N at t0 and goingnorthward until about t0 1 20 when it reaches about218N and stays during most of the summer with itsgreatest values before retreating southward. South of108N, relative vorticity is negative with minima locatedbetween the equator and 38N, which signs the anticy-clonic curvature of the winds crossing the equator. To-mas and Webster (1997) showed that negative absolutevorticity during northern summer just north of the equa-tor over the Guinea Gulf is a sign of inertial instabilityof the low-level atmospheric circulation. Both the high-est negative values of relative and absolute vorticityoccur at t0, suggesting that the summer monsoon onsetis characterized by a maximum of inertial instabilityover the Guinea gulf, which is released when the mon-soon begins to intensify.

Figure 7b depicts similar time–latitude diagram forthe 850-hPa vertical velocity. It again enables us to fol-low the meridional excursion of the ITF and the asso-ciated heat low, represented by the axis of dry convec-tion maximum in the low levels. Its seasonal migrationis very similar, upward motion being associated withcyclonic vorticity. The greatest values are presentaround t0 2 80, that is approximately at the beginningof April, then decrease regularly as the ITF moves tothe north. However a temporary enhancement of dryconvection occurs between t0 2 10 and t0 (it is thegreatest at t0) when the ITF is located at 17.58N andwhen deep convection in the ITCZ, still located at 58N,decreases (Figs. 4 and 6). Figure 7c (vertical velocityaveraged on 700–500 hPa) and Fig. 7d (meridional windcomponent at 600 hPa) describe the upper part of thetransverse circulation associated to the heat low. South

of the maximum of the upward 850-hPa vertical veloc-ity, which is located at 17.58N at t0, both the 600-hPanortherly wind and 700–500-hPa downward velocityhave the greatest values around 12.58N at t0, followinga seasonal cycle similar to the one of 850-hPa verticalvelocity. In the lower layers, the 925-hPa meridionalwind (Fig. 7e) also has its greatest values at t0 between58 and 108N. Then the whole meridional transverse cir-culation associated to the heat low has its greatest in-tensity at t0. Figure 7c shows that the subsiding branchof the low-level circulation, located at 12.58N at t0,clearly separates the dry convection area in the northfrom the lower part of the deep convection associatedto the ITCZ in the south.

The seasonal march of the ITF is also closely linkedwith the evolution of the African easterly jet (AEJ) de-scribed by the zonal wind component at 600 hPa, thepressure level of the jet core (Fig. 7f). This jet is locatedsouth of the ITF, between the dry convection area ofthe heat low and the deep convection of the ITCZ. Thehighest speed of the jet is observed close to t0 at 108N,concomitant with the enhancement of the low-level me-ridional circulation associated with the heat low and thenorthward acceleration of the relative vorticity maxi-mum axis. This is consistent with the work of Thorncroftand Blackburn (1997) who showed that the AEJ is most-ly controlled by the direct meridional circulation of theheat low, the northerly returning flow in the midlevelsinducing an easterly acceleration due to the planetaryvorticity advection. At 925 hPa (Fig. 7g), the westerlyzonal wind is increasing in the latitude band 58–158Nfrom about t0 2 30, with a stronger acceleration at 108Nfrom t0 2 10 (shown by the time section at this latitude),characterizing the monsoon enhancement. This leads toan increase of the vertical shear of the zonal wind be-tween 925 and 600 hPa, which is the greatest at t0 along108N (Fig. 7h).

This monsoon enhancement can be well characterizedthrough the estimation of the top of the monsoon layer(Fig. 7i). This estimation has been defined followingLamb (1983) by the computation of the pressure levelswhere the zonal wind component equals zero, goingfrom westerly downward to easterly upward. The areasshaded in Fig. 7i show the lowest pressure values, thatis, the greatest depth of the monsoon layer. We mustnot consider the shaded areas in the upper corners,which correspond to the northern part of the anticycloniccell controlling westerly winds at these latitudes and atthese periods of the year. From t0 2 90 to t0 2 10, thetop of the monsoon layer is located at a pressure levelaround 870 hPa and moves northward from 58 to 108N.Then, the monsoon depth grows drastically and extendsboth northward and southward, during summer reaching840 hPa north of 158N and 760 hPa at 58N. The monsoononset is then characterized by an abrupt and large me-ridional and vertical development of the westerly windlayer as the ITCZ moves from 58 to 108N. This evolutioncan be seen also in the precipitable water height for


FIG. 7. (a) Composite time–lat diagrams of NCEP–NCAR daily relative vorticity (10 26 s21), averaged over 108W–108E, filtered to removevariability lower than 10 days, and averaged over the period 1968–90 by using the ITCZ shift date as the reference date for each year.


←

Values are presented from t0 (the shift date) 2 90 days to t0 1 140 days. (b) Same as (a) but for the vertical velocity at 850 hPa (unit in100 Pa s21). Negative values represent upward velocity. (c) Same as (a) but for the vertical velocity averaged on 700–500 hPa (unit in 100Pa s21). Negative values represent upward velocity. (d) Same as (a) but for the meridional wind at 600 hPa (m s 21). Negative values representnortherly wind. (e) Same as (a) but for the meridional wind at 925 hPa (m s21). Negative values represent northerly wind. (f ) Same as (a)but for the zonal wind at 600 hPa (m s21). Negative values represent easterly wind. (g) Same as (a) but for the zonal wind at 925 hPa (ms21). Negative values represent easterly wind. (h) Same as (a) but for the vertical shear of zonal wind between 925 and 600 hPa (m s 21).(i) Same as (a) but for the top of the monsoon layer expressed in hPa (see explanation in the text). (j) Same as (a) but the precipitable waterheight (Kg m22). In (a)–(j) the vertical line localizes the date of the ITCZ shift at t0 (the mean date over the period 1968–90 is 24 Jun).

which the northward progression abruptly accelerates att0, bringing, for instance, the isoline 40 kg m22 previ-ously stationary around 98N since t0 2 60, to 158N att0 1 30 (Fig. 7j).

The development of the monsoon over West Africais also seen in other features of the atmospheric cir-culation: 1) the setup of the tropical easterly jet (TEJ)at 200 hPa, which attains its greatest speed in Augustwhen convection and rainfall are the highest, leading toan enhancement of the anticyclonic circulation in theupper levels over West Africa (not shown); 2) the abruptshift of the zonal winds at 600 hPa between 208 and308N (see Fig. 6f), which signs a similar meridionalshift of the axis of the high geopotentials at this pressurelevel (see discussion on Fig. 14c further in the text).

A synthesis of these results is provided through thecomputation of the mean composite pressure–latitudecross sections of the 3D wind averaged on 108W–108Eat t0 2 10, t0, and t0 1 20 (Fig. 8). They depict themeridional and zonal atmospheric circulation associatedto the West African monsoon system on the longitudeband where land–sea contrast exists and where the ITCZabrupt shift occurs. The three steps, t0 2 10, t0, and t0

1 20, represent, respectively, the beginning of the tem-porary convection decrease, the convection minimumassociated with the beginning of the ITCZ abrupt shift(summer monsoon onset), and the time when the sum-mer monsoon regime first establishes.

Figure 8a depicts the meridional cross section of theatmospheric circulation at t0 2 10 before the summermonsoon onset. We do not get a traditional view of themeridional Hadley-type circulations. They are in factlargely disturbed by the predominancy of the heat lowtransverse circulation between the surface and 500 hPa.This circulation is well developed with a dry convectionarea between 158 and 208N, with one subsidence areaaround 258N and another one in the midtropospherebetween 108 and 158N that merges with the northernsubsiding branch of a quite reduced northern Hadley-type cell rejected in the upper part of the troposphereby the ascendance area of the heat low. The northernHadley-type subsiding branch is in fact split into twodifferent branches in the midtroposphere apparently dueto the corresponding diffluent upper branches of the heatlow. The subsiding area located between 108 and 158Nin the midtroposphere appears to interact with the deepconvection of the ITCZ around 58N, which constitutesthe common upward branch of the two Hadley-type

cells, by advecting air masses downward and then weak-ening the upward vertical velocities in the midtropo-sphere. The heat low transverse circulation not onlydisturbs the northern Hadley-type circulation but alsodistorts the meridional circulation of the southern Had-ley-type cell by controlling the northern extension ofthe low-level southerly monsoon winds up to 17.58N.South of the equator, the latitudinaly extended subsidingarea in the whole troposphere signs the southern branchof the southern Hadley-type circulation. The heat lowcirculation also controls the zonal circulation of the low-er half of the troposphere (see isolines in Fig. 8). Thedry convection at the latitude of the ITF is associatedwith the cyclonic circulation of the low pressure centerin the lower levels, controlling the westerly componentof the monsoon winds, and with the anticyclonic cir-culation around 600 hPa at the top of the heat low con-trolling the speed of the AEJ. So the heat low appearsto be a key component of the monsoon system overWest Africa before its onset.

The step t0 is the most important one because it signsthe summer monsoon onset as the beginning of a drasticamplification of the monsoon system that will be settledabout 20 days later. One of the most striking points atstep t0 is the occurrence of a minimum in the convectiveactivity over West Africa, just before the abrupt north-ward shift of the ITCZ (Fig. 8b). Between t0 2 10 andt0, the monsoon system has not moved northward butsome part of it has intensified: the transverse circulationof the heat low is the strongest at t0, associated bothwith an increase of the dry convection at 17.58N andof the subsiding branch between 108 and 158N. Thisintensification leads also to a stronger zonal circulationwith higher westerly wind speed in the monsoon layerand a bit higher AEJ speed in the midlayer, resulting toa maximum of the vertical zonal wind shear between925 and 600 hPa (Fig. 7h). We also observe a slightweakening of the deep convection in the ITCZ still lo-cated at 58N, which is consistent with the minimum ofrainfall values and maximum of OLR values. The north-ern Hadley-type cell is still reduced with its northernsubsiding branch split into two parts because of its lo-cation above the heat low. The heat low enhancementat t0 may be an explanation for the temporary convectionminimum observed over West Africa at this step: theassociated transverse circulation could stimulate intru-sion of dry air from the northern mid- and upper layersof the troposphere into the moist deep convection lead-


FIG. 8. Composite pressure–lat cross sections of the NCEP–NCAR 3D atmospheric circulation averaged over 108W–108E, filtered to removevariability lower than 10 days, and averaged over the period 1968–90 by using as the reference date the ITCZ shift date for each year: (a)at t0 2 10; (b) at t0; (c) at t0 1 20. The meridional-vertical circulation is represented by streamlines, shaded areas depict vertical velocity(unit in 100 Pa s21; negative values for upward motion), contours indicate zonal wind velocity (m s 21; negative values for easterly wind).

ing to some convective inhibition in the ITCZ; as at thesame time this transverse circulation helps to enhancewesterly moisture advection in the monsoon layer andto increase the vertical zonal wind shear, both factorsbeing favorable to convection, the convective inhibitioncould not last a long time, which could contribute tothe abrupt summer monsoon onset.

Once the summer monsoon onset occurs, the atmo-spheric system shows large modifications. We previ-ously saw that t0 corresponds to the release of the inertialinstability, and to a huge increase of the monsoon layer

and of the precipitable water height as the ITF goes onrapidly to the north and the ITCZ ‘‘jumps’’ to 108N.Figure 8c shows the state of the atmospheric system att0 1 20 when the ITCZ has arrived at 108N. We clearlysee that the whole system has moved northward, theheat low as well as the ITF being located at 208N, theITCZ at 108N, the AEJ 58 northward, and the monsoonzonal winds having increased. The intensity of the heatlow associated transverse circulation is weaker than att0, as well as the AEJ wind speed. Another striking pointis the large development of the northern Hadley-type


FIG. 9. Composite time sequence of the 10–60-day filtered NCEP–NCAR 925-hPa wind field averaged over the period 1968–90 by usingas the reference date the ITCZ shift date for each year. The compositeunfiltered rainfall field is superimposed (colored; mm day21). Thescale of the vectors is indicated below (m s21).

cell since its northern subsiding branch is not more per-turbed by the heat low circulation because of its morenorthern latitudinal location, enabling its extension inthe whole troposphere depth and leading it to mergewith the northern subsiding branch of the heat low. Thisdisorganizes the northern part of the heat low transversecirculation, by the reversal of the meridional winds fromsoutherly to northerly in the whole troposphere and es-pecially in the midlevels. This development of the north-ern Hadley-type cell can be explained by the intensi-fication of convection in the ITCZ and its northern pro-gression, as well as we observe an increase of the TEJspeed by more than 2 m s21. So once the summer mon-soon is settled, the associated atmospheric circulationseems to be more controlled by the deep convection andthe Hadley-type meridional circulation than before.However the heat low appears to still be able to mod-ulate the atmospheric circulation in the low and mid-levels. It may have some impact on the moist air ad-vection from the southwest but also on intrusion of dryair coming from high levels of the northern midlatitudes,modulating convection and rainfall at intraseasonalscales (Sultan et al. 2003).

b. A zonal view

We said that we can consider at the first order theWest African monsoon as a zonally symmetric atmo-spheric system. It enables us to highlight the heat lowdynamics and suggests that it could have a significantimpact on the summer monsoon onset specific to WestAfrica. On the other hand we have learned a few thingsabout the heat low dynamics itself, and, in particular,we cannot explain why it enhances at t0. Moreover, itseems that the onset is also characterized by an en-hancement of the zonal wind component of the monsoonwhile the meridional component decreases from t0 southof 108N (Fig. 7e). So we further investigate the mech-anisms of the monsoon onset by looking at it along azonal axis.

In Fig. 4d the composite rainfall time series at 108Ncan be considered as being constituted by a smoothseasonal cycle and superimposed short-term fluctuationsclearly evident between t0 2 20 and t0 1 20. To betteranalyze these short-term fluctuations, which belong tothe intraseasonal timescale, and to separate them for asmoother seasonal signal, we have investigated the com-posite fields of the atmospheric signals filtered between10 and 60 days around t0. Figure 9 shows the timesequence of the mean composite filtered 925-hPa windfield from t0 2 12 to t0 1 21, and the superimposedcomposite unfiltered rainfall field. The following com-ments take into account the superimposition of thesefiltered wind fields on the mean wind field. At t0 2 9,two weak circulation centers are evident along 208N inthe filtered wind field, an anticyclonic one centered at108W and a cyclonic one at 12.58E. These centers inducenortherly wind anomalies over West Africa between

108W and 108E while rainfall has begun to decreasesince t0 2 10 (Fig. 4d). These two centers develop andmove westward in the following days. At t0, the anti-cyclonic center has moved over the northern tropicalAtlantic basin, and the cyclonic center is now locatedat 08 and 17.58N. It is associated to the greatest dryconvection in the heat low located at this latitude (Fig.8) and the weakest rainfall values between 108W and108E (Fig. 4d), whereas this center now induces south-westerly wind anomalies over a part of West Africa. Att0 1 3, this cyclonic circulation extends greatly west-ward as well as the area of southwesterly wind anom-alies. However, it is the time where rainfall over thecentral part of West Africa is the weakest, except westof 58W and east of 7.58E where rainfall begins to en-hance, the first sign of the northern extension of theITCZ. This is consistent with our interpretation of Fig.8 suggesting that the heat low dynamics can simulta-neously increase the convective inhibition by dry sub-sident air intrusion in the midlevels and increase theconvective potential below by sustaining more humidair advection from the southwest. At t0 1 6, the ITCZbegins its northward move by its northern boundary(isoline 4 mm day21) while the cyclonic center, nowlocated at about 108W, induces southwesterly windanomalies over the large part of West Africa, moisturebeing imported not only from the Guinea gulf but alsofrom the northern tropical Atlantic. At t0 1 9, the heartof the ITCZ has clearly moved northward but has notreached 108N yet. The moisture advection enhancementis now mostly controlled through westerly winds whilethe cyclonic center has moved to the north, being located


FIG. 10. Composite time–lon diagram of the 10–60-day filtered NCEP–NCAR 925-hPa relativevorticity (1026 s21) at 17.58N, averaged over the period 1968–90 by using the ITCZ shift dateas the reference date for each year. Values are presented from t0 (the shift date) minus 20 daysto t0 1 20 days. (b) Same as (a) but for the 850-hPa vertical velocity at 17.58N (unit in 100 Pas21). Negative values represent upward velocity. (c) Same as (b) but for 400-hPa vertical velocityat 108N.

along the western coast of West Africa between 208 and22.58N. At t0 1 12 and t0 1 15, the westerly windadvection is maintained by a rather weak cyclonic cir-culation centered along 108N while the northern cy-clonic center is now visible over the ocean. The ITCZget its final location at 108N around t0 1 21 with theappearance of rainfall minima along the Guinean coast.So the examination of the time series of the filtered windfield patterns enables to suggest that the summer mon-soon onset is not only due to a meridional dynamicsassociated to the heat low but that it is also associatedwith a westward-propagating signal at an intraseasonaltimescale, which induces the enhancement of the heatlow rotational and transverse circulation.

Figure 10 helps to quantify these observations bydepicting composite longitude–time diagrams of the 10–60-day filtered 925-hPa relative vorticity at 17.58N (Fig.10a), 850-hPa vertical velocity at 17.58N (Fig. 10b), and400-hPa vertical velocity at 108N (Fig. 10c), computedbetween t0 2 20 and t0 1 20. The diagrams show be-tween t0 2 10 and t0 1 10 the westward propagationof positive values of relative vorticity at 17.58N, at anapproximate mean phase speed of 48 longitude per day,concomitant with the enhancement of dry convection inthe heat low at the same latitude and with a decreaseof the 400-hPa upward velocity at 108N. These signalsare maintained mainly during the phase of decreasingthen increasing rainfall at 108N characterizing the spe-cific summer monsoon onset over West Africa.

Figure 11a shows the composite unfiltered NCEP–NCAR 925-hPa wind field at t0 2 15. Colored areasdepict relative vorticity values and the solid line is thezero isoline of the zonal wind component depicting the

ITF over the Sahel. The different positive maxima ofvorticity along the ITF localize the semipermanent cy-clonic vorticity centers. Four maxima are identified overland at 308, 158E, 08, and 158W. Fifteen days after, att0 (Fig. 11b), the ITF has moved a bit northward. Figure11d shows the composite wind field of the differencebetween t0 and t0 2 15. Colored areas depict OLR dif-ferences and isolines relative vorticity differences. Thenorthward ITF location at t0 is depicted in Fig. 11d bypositive relative vorticity anomalies between 158 and258N and southwesterly wind anomalies on the southernside. However, the positive anomalies of vorticity aregreater east of 08. Associated with this vorticity field,negative OLR anomalies, meaning greater convection,are located mostly east of 58E. They are located along12.58N, on the northern boundary of the ITCZ still lo-cated at 58N. Small negative OLR anomalies also occurover the ocean north of the ITCZ. Positive OLR anom-alies along 2.58N sign the northward progression of theITCZ southern boundary. Between approximately 108Wand 58E, only positive OLR anomalies are observable,signaling the convection decrease at these longitudesjust before the monsoon onset (Fig. 6).

Figure 11e shows the difference fields between t0 210 and t0 1 10. The OLR pattern is similar to the pre-vious one but amplified. However, the relative vorticityfield is different. The greatest anomalies are now locatedon the western part of northern Africa, and especiallywest of 58E. This leads to developed southwesterlyanomaly winds over West Africa and over the northerntropical Atlantic. At the same time, we can observe anamplification of an anticyclonic circulation along 308Nnorth of the Atlas Mountains (see Fig. 12), and of as-


FIG. 11. (a) Composite unfiltered NCEP–NCAR 925-hPa wind field at t0 2 15 averaged over the period 1968–90 by using as the referencedate the ITCZ shift date for each year. Colored areas depict relative vorticity values (10 26 s21). The solid line is the zero isoline of the zonalwind component. (b) Same as (a) but for t0. (c) Same as (a) but for t0 1 20. (d) Composite unfiltered NCEP–NCAR 925-hPa wind field[vector scale (m s21) is displayed below] for the difference between t0 and t0 2 15, averaged over the period 1968–90 by using the ITCZshift date as the reference date for each year. Shaded areas depict OLR differences (W m22). Isolines represent relative vorticity differences(1026 s21). (e) Same as (d) but for the difference between t0 1 10 and t0 2 10. (f ) Same as (d) but for the difference between t0 1 20 andt0.

sociated northeasterly winds. Figure 11c shows the windfield pattern at t0 1 20, and Fig. 11f the difference fieldsbetween t0 1 20 and t0. The positive vorticity anomaliesare the greatest west of 58E (Fig. 11f ) and signal themost northward movement and amplification of the ITFat these longitudes. This induces southwesterly anomalywinds over the northern tropical Atlantic and over thelarge part of West Africa north of 108N. Negative vor-ticity differences are still present north of the AtlasMountains. The negative OLR differences are now cov-ering all of West Africa signaling, in particular, theabrupt displacement of the ITCZ from 58 to 108N onthe longitudes 108W–58E. So, between t0 2 15 and t0

1 20 the cyclonic activity of the ITF amplifies, firsteast of 58E over land, and then west of 58E over landand over the ocean, concomitant with the abrupt shiftof the ITCZ around the Greenwich meridian area. Theheat low circulation over land may also have an impacton the atmospheric circulation over the northern tropicalAtlantic and on convection in the oceanic ITCZ.

5. A possible mechanism for the monsoon onset

In the previous section we have suggested that theabrupt shift of the ITCZ at the time of the summermonsoon onset could be due to an interaction between


FIG. 12. Orography of northern Africa (m; areas higher than 400 m are shaded by step of400 m).

the deep convection and the meridional transverse cir-culation associated with the heat low. When the heatlow significantly increases in an area located west of58E and south of the Atlas Mountains, this enhancementis accompanied by a similar increase of the anticyclonicvorticity north of the Atlas Mountains. We now discussthe hypothesis that the orography in North Africa couldbe responsible for this acceleration of the seasonal cycleof the monsoon over West Africa.

This hypothesis has been deduced from a work ofSemazzi and Sun (1997). Although most of West Africasouth of the Sahara is generally free of steep terrain,nearby mountains ranges, such as the Atlas Mountains,the Ahaggar Plateau, and the Tibesti highs, could exertsignificant control of the atmospheric circulation, andthen of the convection in the ITCZ. Figure 12 showsthat this orography is oriented along a northwest–south-east axis, so being nearer to the ITF trough in the eastthan in the west. In a previous work, Viltard et al. (1990)showed that the ITF over West Africa during northernsummer is characterized by three semipermanenttroughs (see our Fig. 11c for instance) and that someof them could be induced by orography. Semazzi andSun (1997) examined the impact of orography on theWest African summer climate in the context of differentsea surface temperature anomalies through GCM sim-ulations with and without orography. They found thatthese orographic effects are composed of (i) a quasi-stationary orographic ridge-trough dipole generated bythe passage of the low-level prevailing easterlies overthe Atlas–Ahaggar Mountain complex over northern Af-rica, and (ii) an elongated zonal windward orographicridge generated as the cross-equatorial summer mon-soons from the South Atlantic Ocean basin ascends over

the elevated landmass of West Africa. This results inwetter conditions over the Sahel and drier conditionsfarther south along the Guinea coast region, whichwould have been the case in the absence of orography,leading to a permanent orographic-induced rainfall di-pole pattern over West Africa that we can associate toa more northward location of the ITCZ.

We can think about applying this hypothesis in thecontext of the seasonal cycle of the ITCZ. In the ex-periment of Semazzi and Sun, the orography-inducedforcing leads to a northward location of the ITCZ duringthe summer due to the monsoon circulation enhance-ment because of the leeward effect. In our investigation,the onset is also characterized by a northward shift ofthe ITCZ in combination with the enhancement of theheat low circulation. This is the main reason why Se-mazzi and Sun’s results and ours can be associated. Ifat a certain time during the northward course of theITCZ, the interaction between orography and the low-level atmospheric circulation amplifies, the orographic-induced more northward location of the ITCZ couldexplain its abrupt shift concomitant with the onset ofthe summer monsoon. Figure 13 is a reproduction ofSemazzi and Sun’s Figs. 7 and 8, that is the differencebetween a control run with orography and a control runwithout orography of the July–September geopotentialheight and wind field patterns at 850 hPa. We see adistinct windward high pressure and leeward low pres-sure distribution across the Atlas–Ahaggar Mountainrange. The low-level cyclonic circulation associatedwith this trough is the largest over the western part ofAfrica but extends far eastward over the land and con-tributes to enhance westerly advection of moisture fromthe northern tropical Atlantic to the Sahel. We can ob-


FIG. 13. (a) Difference between a control run with orography anda control run without orography of the 850-hPa geopotential height(contour interval 2 m) and (b) 850-hPa streamline wind patterns inthe experiment of Semazzi and Sun (1997).

FIG. 14. (a) Composite difference between t0 1 12 and t0 2 5 ofthe 925-hPa geopotential height (mgp) and streamline wind patternsaveraged over the period 1968–90 by using the ITCZ shift date asthe reference date for each year. The location of the three centersused to compute time series are displayed. The dashed line representsthe crest line of the orography. (b) Composite time series of the 925-hPa geopotential height difference between values at 308N, 7.58E(point 2) and 22.58N, 108W (point 1; thick curve; mgp), of the 925-hPa geopotential height difference between values at 308N, 7.58E(point 2) and 17.58N, 158E (point 3; thin curve; mgp), of the averaged108W–108E 600-hPa vertical velocity at 27.58N (dashed curve; unitin 100 Pa s21). Values are presented from t0 2 90 days to t0 1 140days, and filtered to remove variability lower than 10 days. The twoscales in the left ordinate correspond, respectively, from right to leftto the thick and thin curves; the scale of the right ordinate correspondsto the dashed curve. (c) Composite time–latitude diagrams of theNCEP–NCAR daily 600-hPa geopotential height (mgp), averagedover 108W–108E, filtered to remove variability lower than 10 days,and averaged over the period 1968–90 by using the ITCZ shift dateas the reference date for each year. Values are presented from t0 minus90 days to t0 1 140 days.

serve also that these patterns are oriented as the orog-raphy on a northwest–southeast axis and that they ex-tend farther to the west over the northern tropical At-lantic, inducing a rainfall dipole over this part of theocean similar to the one over West Africa (Fig. 9a ofSemazzi and Sun). So the orography of northern Africacan have also a significant impact of the oceanic ITCZ.

Figure 14a shows the composite difference betweent0 1 12 and t0 2 5 of the 925-hPa geopotential heightand streamline wind patterns. It shows a very high sim-ilarity with the results of Semazzi and Sun on the twosides of the Atlas–Ahaggar Mountain axis (similar fieldshave been obtained at 850 hPa, the level chosen bySemazzi and Sun; not shown). We can observe the pos-itive geopotential height anomalies centered at 7.58E onthe windward side of the mountains (point labeled 2)and the two negative anomalies centered at 158E and108W on the leeward side (labeled 3 and 1, respectively).These geopotential anomalies are associated respective-ly with anticyclonic circulation on the north leading toenhanced northerly winds from the Mediterranean Sea,and with cyclonic circulation on the south inducing en-hanced monsoon winds over land and also over the

northern tropical Atlantic basin east of 308W (see Fig.11f). The orography-induced low-level circulation overWest Africa could then favor the enhancement of theheat low both east and west of the Greenwich meridianand play a role in the ITCZ abrupt shift, as well as inthe northward progression of the oceanic ITCZ.


Figure 14b shows the composite time series from t0

2 90 to t0 1 140 of the 925-hPa geopotential heightdifference between values at 308N, 7.58E (point labeled2 in Fig. 14a) and 22.58N, 108W (point 1; thick curve),of the 925-hPa geopotential height difference betweenvalues at 308N, 7.58E (point 2) and 17.58N, 158E (point3; thin curve), and of the averaged 108W–108E 600-hPavertical velocity at 27.58N (dashed curve). The thickand the thin curves again highlight the abrupt changeof the atmospheric circulation with their steep slopeoccurring from t0 2 10 to t0 1 20, following a ratherstable state before t0 2 10. The time series of the geo-potential heights associated to the anticyclonic centerand to the cyclonic centers separately show an abruptbreak around t0 2 10 (not shown). The dashed curvein Fig. 14b shows the downward vertical velocity at27.58N, that is the location of both the northern sub-siding branch of the heat low and the northern Hadley-type cell (Fig. 8). The subsidence in this area begins toincrease around t0 2 35, then more rapidly from t0 220 to t0 1 40. This enhancement is due first to thenorthern extension of the subsiding part of the transversecell associated to the heat low and to its intensificationaround t0, second to the large development of the north-ern Hadley-type cell after t0, in particular of its northernsubsiding part. So the windward subsidence and highgeopotentials begin to increase first, increasing then theassociated northeasterly winds, leading finally to theleeward cyclonic circulation enhancement.

Figure 14c shows the latitude–time cross section ofthe 600-hPa geopotential heights averaged over 108W–108E. Figure 7f has highlighted the meridional abruptshift of the zonal winds at 600 hPa between 208 and308N signing a similar shift of the axis of the highgeopotentials at this pressure level. We see in Fig. 14cthat it is concomitant with a strong increase of the geo-potential heights to the highest values reached duringthe summer. This anticyclonic center is located after themonsoon onset above and northward of the North Africahighlands, and it is associated with the enhancement ofthe subsiding motions of the northern Hadley-type cir-culation. This signal of the monsoon onset can be putin parallel with the Asian monsoon onset where a similarnorthward abrupt shift of the subtropical westerly jetand the associated subtropical high pressures have beenidentified in the high troposphere above the Tibet high-lands (Hahn and Manabe 1975).

Figure 14a shows a background signal specific to theWest African summer monsoon onset over the period1968–90. We have investigated the interannual vari-ability associated with this pattern by performing anautomatic classification (Janicot 1992) of these geo-potential fields (not shown). We obtained four classeswith similar spatial patterns, retaining the three centers(labeled 1, 2, and 3 in Fig. 14a) at about the samelocations, with similar geopotential height gradientsigns, but with different height amplitudes associatedwith these centers. However we could not define any

significant characterization of the ITCZ shifts related tothese different classes (the four averaged rainfall fieldsdepict a similar abrupt shift), nor any correlation withthe summer rainfall amount over the Sahel. This sug-gests 1) the independency between the mechanisms as-sociated to the interannual rainfall variability of the WestAfrican monsoon and the mechanisms induced in themonsoon onset, and 2) the hypothesis that the orogra-phy-induced mechanisms shown by Semazzi and Sunon the mean strength of the West African monsoon canwork in the monsoon onset of individual years.

We suggest that the northern progression of the heatlow and its transverse circulation leads to the enhance-ment of the anticyclonic circulation on the windwardside of the Atlas–Ahaggar Mountain crest, inducing en-hanced northeasterly winds that interact with the orog-raphy and lead to the increase of the leeward trough andassociated westerly advection of moisture inland, con-tributing to the enhancement of the heat low intensityat t0. Then the ITCZ moves to the north and intensifiesafter breaking the convective inhibition induced by theheat low. This leads to the enhancement of the northernHadley-type cell and to the development of its northernsubsiding branch, which also contributes to increase thehigh geopotentials north of the mountains and then theorography-induced lee-trough southward. The abruptshift of the ITCZ could then be explained by a positivefeedback between the atmospheric circulation of theheat low, the convection in the ITCZ, and the Hadley-type circulation, through the control by the orography.The apparent westward signal seen in the filtered windfields could be due to the northwest–southeast orien-tation of the orography axis, which induces the inter-action with the heat low first in the east where the moun-tains are nearer to the ITF than in the west, and secondin the west. The fact that we observe an abrupt shift ofthe ITCZ only in the western part of West Africa mightresult from the enhancement of moisture advection thatcomes from the west and has a stronger impact west ofthe Greenwich meridian.

6. Synthesis and conclusions

As the seasonal cycle advances in the first half of theyear, a land–sea temperature gradient grows betweenWest Africa and the equatorial and southern Atlanticbasin, and the resulting low-level meridional pressuregradient moves northward. Over West Africa, the highnorthward temperature gradient and the high southwardmoisture gradient between the Sahara and the Guineancoast induce a meridional lag between the surface drystatic energy maximum (located more northward) andthe moist static energy (located more southward). Thisshows that the deep convection in the ITCZ controllingthe two Hadley-type cells is located to the south of thedry convection of the heat low, limited in the low andmidlevels of the troposphere, and capped by the sub-siding branch of the northern Hadley-type cell. The


ground location of the heat low is also characterized bya positive relative vorticity maximum and by the con-fluence of the moist southwesterly winds of the monsoonand the dry northeasterly Harmattan, that is the ITF. Atthe preonset of the summer monsoon, that is at ap-proximately t0 2 40 days (where t0 means the date ofthe summer monsoon onset), the ITF is located at 158Nand the ITCZ at 58N. Between t0 2 40 and t0 2 10,the heat low dynamics dominates north of the ITCZ anda part of its northern subsiding area is located aboveand north of the Atlas–Ahaggar crest. During that time,isolated convective systems begins to occur south of theITF, contributing to the northern progression of theITCZ northern boundary, and may be to a slight increaseof the northern Hadley-type cell dynamics and its sub-siding motions, reinforcing the subsidence over the At-las–Ahaggar orography. These two factors contribute toenhance the high geopotentials north of these mountainsand the associated northeasterly winds.

Between t0 2 10 and t0, the interaction between orog-raphy and northeasterly winds on the windward side ofthese mountains enhances and lead to the developmentof a leeward trough that reinforces the heat low dynam-ics. This has opposite consequences. On one hand, themidlevel subsidence of the southern branch of the heatlow transverse circulation increases and may be accom-panied by intrusion of dry air from the north. This canincrease convective inhibition in the ITCZ and explainthe observed temporary convection and rainfall de-crease. On the other hand, cyclonic vorticity increasesat low levels contributing to a greater inland moistureadvection by stronger southwesterly winds and a deepermonsoon layer. In the midlevels, the anticyclonic cir-culation increases leading to the enhancement of theAEJ. Then the vertical wind shear between the low andmidlevels becomes higher. Deeper monsoon and highervertical wind shear enhance the local potential insta-bility. The observed weaker convection also leads tohigher downward solar radiation received at the surface(seen in the NCEP reanalysis; not shown), a favorablecondition for potential instability increase. Finally asthe heat low transverse circulation also maintains theinstability of the AEJ by imposing a reversal of themeridional gradient of potential vorticity (Thorncroftand Blackburn 1997), the enhanced heat low dynamicsincreases the instability character of the AEJ (negativemeridional gradient of potential vorticity is enhanced at108N between t0 2 10 and t0; not shown), which mayfavor easterly wave development and then convection.

From t0 to t0 1 20, the accumulated potential insta-bility breaks the convective inhibition, the inertial in-stability of the monsoon circulation is released and theassociated regional scale circulation increases, leadingto the abrupt shift of the ITCZ. Then the ITCZ movesto the north, where thermodynamical conditions are fa-vorable, up to 108N. The northern Hadley-type circu-lation drastically enhances, which leads to the devel-opment of subsidence in the whole troposphere between

258 and 308N. Then geopotential highs increases northand above the Atlas–Ahaggar Mountains, which en-hances the orography-induced windward high and lee-ward trough again, maintaining an active convectiveITCZ during the summer monsoon. The abrupt shift ofthe ITCZ could then be explained by a positive feedbackbetween the atmospheric circulation of the heat low, theconvection in the ITCZ, and the northern Hadley-typecirculation, through an interaction with the orography.

The northwest–southeast orientation of the Atlas–Ahaggar crest can induce the interaction with the heatlow first in the east where the mountains are nearer tothe ITF than in the west, and second in the west. Thefact that we observe an abrupt shift of the ITCZ onlyin the western part of West Africa may result from theenhancement of moisture advection that comes from thewest and has a stronger impact west of the Greenwichmeridian.

Another consequence of the orography-induced in-teraction with the atmospheric circulation is that theinduced leeward through, increasing the cyclonic vor-ticity in the heat low, may stimulate moisture conver-gence in the oceanic ITCZ near the western coast ofWest Africa. However, a similar ITCZ abrupt shift isnot observed over this part of the northern tropical At-lantic.

Finally, once the monsoon onset occurs, the devel-oped northern Hadley-type circulation can favor intru-sion of dry air from the north and the upper levels intothe ITCZ. These intrusions can be also controlled bythe northerly winds of the heat low transverse circula-tion. This question has been addressed in Sultan et al.(2003) through the study of the intraseasonal timescalemodulation of convection observed during the West Af-rican summer monsoon. This mechanism should notwork efficiently before the summer monsoon onset sinceat this time the northern Hadley-type cell is not devel-oped enough to the north to allow northerly advectionsfar from the ITCZ.

The hypothesis of the abrupt shift of the ITCZ hasbeen deduced from the observations. It must be con-firmed through other ways of investigation. In particularatmospheric GCM experiments similar to that of Se-mazzi and Sun, but in the context of the seasonal cycle,would be a good way to test this hypothesis. However,these types of models have to first accurately reproducethe dynamics of the West African monsoon describedhere to be validated (Vernekar and Ji 1999). On the otherhand, the hypothesis presented here is not necessarilythe only explanation for the ITCZ abrupt shift. For in-stance, interactions between land surface processes anddeep convection, especially through the impact of soilmoisture and of the sensible heat input in the boundarylayer, have been previously involved to explain north-ward-propagating rain bands in the Indian monsoon(Webster 1983; Gueremy 1990; Ferranti et al. 1999).Similar mechanisms could be involved in the ITCZ lat-itudinal shift over West Africa. Xie and Saiki (1999)


have proposed another mechanism to explain delayedabrupt onsets observed in the Asian summer monsoonsby the concept of ‘‘geostrophic monsoon’’: in a rotatingatmosphere, the land–sea temperature difference doesnot necessarily lead to a direct overturning cell, butinstead can be balanced by adjusting its vertical shearso as to be in a thermal wind balance with the meridionaltemperature gradient; so the onset is delayed and canbe initiated by the explosive growth of a traveling dis-turbance. Plumb and Hou (1992) also showed that sucha thermal wind adjustment is possible for an off-equa-torial heating, and that an intensive meridional circu-lation occurs when the subtropical heating rate exceedsa certain threshold, another source for an abrupt changeof the monsoon circulation and its onset. These typesof mechanisms must be investigated further throughocean–atmosphere coupled GCMs since the monsoononset could have a significant impact on the oceanicITCZ near the western coast of West Africa.

In this work, by using gridded rainfall and NCEP–NCAR reanalysis data during the period 1968–90 (OLRwas used since 1979), we have described the arrival ofthe summer monsoon over West Africa and character-ized two steps: its preonset and its onset. We have shownthat these steps are not correlated in time nor correlatedwith the summer rainfall amount over the Sahel. Theresults on the definition of these steps have been ob-tained through the computation of composite meansbased on the preonset or the onset date for each year.This method assumes that the rate of development be-fore and after each of these steps is similar each year.We have characterized the preonset stage of the summermonsoon by the arrival of the ITF at 158N around mid-May while the ITCZ is still centered at 58N. The onsethas been pointed out by an abrupt latitudinal shift ofthe ITCZ in late June from 58 to 108N. This compositeapproach enables to highlight the atmospheric processesassociated with these steps of the West African monsoonseasonal cycle. It could be thought that the compositeprocedure introduces artificial signals since, for themonsoon onset for instance, as we select t0 so as todetect a shift, it is logical to find such a shift. HoweverLe Barbe et al (2002) found a similar meridional shiftof the ITCZ by computing a classic average of the rain-fall seasonal cycle. So our composite approach, not onlydoes not create the ITCZ shift, but enables us to focusmore precisely on the associated dynamical mecha-nisms.

This is not only an interesting meteorological andclimatological issue, it is also a major point for agri-culturists. A better knowledge of the mechanisms of thearrival of the summer monsoon should help providingefficient forecast of the beginning of the growing season.This point has not been considered here since we haveused a smoothing procedure to remove high-frequencytransient signals from the composite rainfall and dy-namical patterns associated to the summer monsoon on-set. Forecasting this onset in an operational context need

to take these transients into account, in addition to thecomposite scenario proposed here.

Acknowledgments. We are thankful to the NOAA–CIRES Climate Diagnostics Center (Boulder, CO) forproviding the NCEP–NCAR reanalysis dataset and theinterpolated OLR dataset from their Web site (http://www.cdc.noaa.gov/). We also thank the two reviewersfor their help improving this paper. This research wasin part supported by the EC Environment and ClimateResearch Programme (Contract: EVK2-CT-1999-00022).

REFERENCES

Ati, O. F., C. J. Stigter, and E. O. Oladipo, 2002: A comparison ofmethods to determine the onset of the growing season in northernNigeria. Int. J. Climatol., 22, 731–742.

Diedhiou, A., S. Janicot, A. Viltard, P. de Felice, and H. Laurent,1999: Easterly wave regimes and associated convection overWest Africa and the tropical Atlantic: Results from NCEP/NCARand ECMWF reanalyses. Climate Dyn., 15, 795–822.

Duvel, J. P., 1989: Convection over tropical Africa and the AtlanticOcean during northern summer. Part I: Interannual and diurnalvariations. Mon. Wea. Rev., 117, 2782–2799.

Ferranti, L., J. M. Slingo, T. N. Palmer, and B. J. Hoskins, 1999: Theeffect of land-surface feedbacks on the monsoon circulation.Quart. J. Roy. Meteor. Soc., 125, 1527–1550.

Gibson, J. K., P. Kalberg, S. Uppala, A. Hernandez, A. Nomura, andE. Serrano, 1997: ERA description. ECMWF Re-Analysis Proj-ect, Rep. Series 1, ECMWF, Reading, United Kingdom, 72 pp.

Gruber, A., and A. F. Krueger, 1984: The status of the NOAA outgoinglongwave radiation data set. Bull. Amer. Meteor. Soc., 65, 958–962.

Gueremy, J. F., 1990: Heat and moisture fluxes on the time scale of20 to 60 days over the Indian monsoon area. Meteor. Atmos.Phys., 44, 219–250.

Hahn, D. G., and S. Manabe, 1975: The role of mountains in theSouth Asian monsoon circulation. J. Atmos. Sci., 32, 1515–1541.

Hastenrath, S., 1995: Climate Dynamics of the Tropics. Kluwer, 488pp.

Janicot, S., 1992: Spatiotemporal variability of West African rainfall.Part I: Regionalizations and typings. J. Climate, 5, 489–497.

——, S. Trzaska, and I. Poccard, 2001: Summer Sahel-ENSO tele-connection and decadal time scale SST variations. Climate Dyn.,18, 303–320.

Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Re-analysis Project. Bull. Amer. Meteor. Soc., 77, 437–471.

Kanamitsu, M., W. Ebisuzaki, J. Woollen, S.-K. Yarg, J. J. Hnilo, M.Fiorino, and G. L. Potter, 2002: NCEP-DOE AMIP-II reanalysis(R-2). Bull. Amer. Meteor. Soc., 83, 1631–1643.

Lamb, P. J., 1983: West African water vapor variations between recentcontrasting Subsaharan rainy seasons. Tellus, 35A, 198–212.

Le Barbe, L., T. Lebel, and D. Tapsoba, 2002: Rainfall variability inWest Africa during the years 1950–90. J. Climate, 15, 187–202.

Liebmann, B., and C. A. Smith, 1996: Description of a complete(interpolated) outgoing longwave radiation dataset. Bull. Amer.Meteor. Soc., 77, 1275–1277.

Plumb, R. A., and A. Y. Hou, 1992: The response of a zonally sym-metric atmosphere to subtropical thermal forcing: Threshold be-havior. J. Atmos. Sci., 49, 1790–1799.

Poccard, I., S. Janicot, and P. Camberlin, 2000: Comparison of rainfallstructures between NCEP/NCAR reanalysis and observed dataover tropical Africa. Climate Dyn., 16, 897–915.

Richman, M. B., 1986: Rotation on principal components. J. Cli-matol., 6, 293–335.


Semazzi, F. H. M., and L. Sun, 1997: The role of orography in de-termining the Sahelian climate. Int. J. Climatol., 17, 581–596.

Sow, C. S., 1997: Diurnal rainfall variations in Senegal. Secheresse,8, 157–162.

Sultan, B., 2002: Study of the monsoon onset in West Africa and ofthe convection intraseasonal convection. Application to the cropyield sensitivity (in French). Ph.D. thesis, 283 pp.

——, and S. Janicot, 2000: Abrupt shift of the ITCZ over West Africaand intra-seasonal variability. Geophys. Res. Lett., 27, 3353–3356.

——, ——, and A. Diedhiou, 2003: West African monsoon dynamics.Part I: Documentation of intraseasonal variability. J. Climate,16, 3389–3406.

Thorncroft, C. D., and M. Blackburn, 1997: On the maintenance ofthe African Easterly Jet. Quart. J. Roy. Meteor. Soc., 123, 763–786.

Tomas, R. A., and P. J. Webster, 1997: The role of inertial instabilityin determining the location and strength of the near-equatorialconvection. Quart. J. Roy. Meteor. Soc., 123, 1445–1482.

Vernekar, A. D., and Y., Ji, 1999: Simulation of the onset and intra-seasonal variability of two contrasting summer monsoons. J.Climate, 12, 1707–1725.

Viltard, A., H. Laurent, and P. de Felice, 1990: Lower-troposphericcyclonic vorticities in northern Africa in summer. Mon. Wea.Rev., 118, 818–823.

Webster, P. J., 1983: Mechanisms of monsoon low-frequency vari-ability: Surface hydrological effects. J. Atmos. Sci., 40, 2110–2124.

Xie, S. P., and N. Saiki, 1999: Abrupt onset and slow seasonal evo-lution of summer monsoon idealized GCM simulation. J. Meteor.Soc. Japan, 77, 949–968.

The West African Monsoon Dynamics. Part II: The ‘‘Preonset ... · monsoon, deﬁning the...

Documents

Transcript of The West African Monsoon Dynamics. Part II: The ‘‘Preonset ... · monsoon, deﬁning the...