The major objective of this study is to re-evaluate the ocean–land transport of moisture for rainfall in West Africa using 1979–2008 NCEP–NCAR reanalysis data. The vertically integrated atmospheric water vapor flux for the surface–850 hPa is calculated to account for total low-level moisture flux contribution to rainfall over West Africa. Analysis of mean monthly total vapor fluxes shows a progressive penetration of the flux into West Africa from the south and west. During spring (April–June), the northward flux forms a “moisture river” transporting moisture current into the Gulf of Guinea coast. In the peak monsoon season (July–September), the southerly transport weakens, but westerly transport is enhanced and extends to 20°N owing to the strengthening West African jet off the west coast. Mean seasonal values of total water vapor flux components across boundaries indicate that the zonal component is the largest contributor to mean moisture transport into the Sahel, while the meridional transport contributes the most over the Guinea coast. For the wet years of the Sahel rainy season (July–September), active anomalies are displaced farther north compared to the long-term average. This includes the latitude of the intertropical front (ITF), the extent of moisture flux, and the zone of strong moisture flux convergence, with an enhanced westerly flow. For the dry Sahel years, the opposite patterns are observed. Statistically significant positive correlations between the zonal moisture fluxes and Sudan–Sahel rainfall totals are most pronounced when the zonal fluxes lead by 1–4 pentads. However, although weak, they still are statistically significant at lags 3 and 4 for meridional moisture fluxes.
The West African monsoon (WAM) is an integral component of Earth’s climate system, involving complex interactions between the atmosphere, the hydrosphere, and the biosphere and over many time scales during the boreal summer (e.g., Redelsperger et al. 2002; Nicholson and Grist 2003; Redelsperger et al. 2006). It encompasses developing countries that are particularly vulnerable to climate variability and climate change. The variability of water availability is one of the most limiting parameters for sustaining life, agriculture, and economic development in the sub-Saharan West African countries. Therefore, the role played by the atmospheric moisture transfer and its phase transitions through evaporation, latent heat release, and associated energy transports and exchanges are of central importance for the WAM dynamics and variability. This is because hydrological processes play an important role in determining the scales of the major circulation patterns (e.g., Webster 1994) and also because the natural variability of weather and climate at both regional and global scales is regulated by the water cycle (e.g., Eltahir and Bras 1996). Accordingly, research on water vapor flux and its convergence has both scientific and societal value.
The WAM rainfall results from the moisture fluxes originating from many sources during the summer season. Over 80% of the annual precipitation falls during June–September when the intertropical front (ITF) extends farther north, but the total precipitation has large year-to-year variations (Nicholson et al. 2000; Le Barbé et al. 2002). The WAM region frequently suffers from droughts, which cause water shortages and disrupt the agriculture sector; this is the only sector that provides both food and income for the majority of rural households. The impact of these droughts and the controversy concerning their causes has prompted climatologists to offer a variety of hypotheses—including changes in the ITF latitude position, tropical Atlantic sea surface temperature anomalies, the El Niño–Southern Oscillation (ENSO) phenomenon, and regional-scale atmospheric features—to account for the observed variations in rainfall. An overview is provided by Nicholson (2013).
Given the proximity of the extent of monsoon layer, the ITF latitude, and the rain belt over West Africa (WA), the concept of an interaction between the three has an intuitive appeal since the northern boundary of the moisture flux position determines the northernmost edge of the habitable area. However, a surprisingly small number of previous studies addressed the vertically integrated moisture flux (VIMF) over WA as a rainfall predictor. The few studies (reviewed below) related to this subject have mainly focused on extreme hydrologic events over the Sahelian zone. In most studies, the validity of the results was questioned, because of the limited use of the atmospheric station data. This limitation involved not only insufficient spatial sampling but also insufficient resolution in time.
Several long-term goals justify an analysis of the atmospheric moisture flux in the WAM region. First, such an analysis contributes to the understanding of the processes responsible for the variability of rainfall and therefore the occurrence of multiyear drought periods. However, the causes remain elusive. Furthermore, such an analysis can contribute to an explanation of the changes from wet to drought conditions that began in the late 1960s. In this study, the VIMFs are derived using the NCEP–NCAR reanalysis data. From the viewpoint of moisture transport, this is a valuable resource to study interannual variability in the hydrologic cycle, with the possibility of using several decades of data.
Attention is given to the phases of WAM life cycle, from buildup and onset to maturity and withdrawal. The goal is to understand better the relation between large-scale atmospheric circulation, moisture flux and associated convergence, and precipitation events.
Few studies have investigated the sources of moisture transport over WA using both observations and modeling. For example, Kidson (1977) analyzed upper-air data over the region and suggested that the widespread reduction in rainfall during the drought years of 1972/73 in the Sahel is associated with the disappearance of the 850-hPa trough near 8°N and a weakening of the easterly jet above it. This is an indication that the westerly flow in the surface–850-hPa layer is an important moisture source for the Sahel.
Lamb (1983) investigated the interannual and intraseasonal variability of the monsoon layer thickness and moisture content over WA. That study found that sub-Saharan drought did not appear in conjunction with unusually dry southerly surface air from the tropical Atlantic. Lamb (1983) also found that during the extremely dry year, the northward moisture flux across the Gulf of Guinea was shallow compared with the much deeper monsoon layer, during the less severe drought years. Cadet and Nnoli (1987) used one summer (1979) of the European Centre for Medium-Range Weather Forecasts (ECMWF) data to study the water vapor transport over Africa. They analyzed biweekly fields of water vapor flux between the surface and 850 hPa and showed that the Gulf of Guinea and central Africa supply most of the moisture for rainfall over WA. Furthermore, the authors showed a northward penetration of southerly monsoon flow up to 20°N and the presence of a large belt of westerlies around 10°N.
Druyan and Koster (1989) investigated the sources of water vapor fluxes into WA using the Goddard Institute for Space Studies (GISS) climate model. They found that the tropical North Atlantic Ocean contributes the most to rainfall over western Sahel, whereas the Gulf of Guinea and the South Atlantic Ocean contribute the most over the central Sahel. Local evaporation is the second largest contributor to rainfall for both regions.
Long et al. (2000) examine the large-scale forcing mechanisms in relation to initiation and maintenance of the Sahelian long-term drought. They analyzed rainfall, moisture flux, and vertical motion data and concluded that changes in the general circulation were important in initiating drought in the Sahel, although other mechanisms may be responsible for its persistence.
The advent of assimilated data products in the mid-1990s, such as the NCEP–NCAR reanalysis data, has provided opportunities for atmospheric research including the water vapor transport and associated moisture flux convergence studies. Fontaine et al. (2003) studied the atmospheric water cycle and associated moisture fluxes in the WAM region using the NCEP–NCAR reanalysis data. The study identifies the Mediterranean Sea, central Africa, and the Gulf of Guinea as sources of low-level moisture for the WAM region. The authors also noted the importance of the meridional and zonal moisture fluxes for interannual rainfall variability, with the zonal moisture flux being most important for the Sahelian region. Couvreux et al. (2010) used observations and the ECMWF dataset to underscore the variability of the northward excursions of the moisture flux at a 3–5-day time scale. Similarly, Grams et al. (2010) studied the sea breeze–like inflow from eastern Atlantic via Mauritania into the southwestern part of the Saharan heat low (SHL). They found that the so-called Atlantic inflow regulates the penetration of moist air into WA. Based on a combination of data that provides the best estimate of the terms of the moisture budget equation, Meynadier et al. (2010a) presented a hybrid dataset for the atmospheric water budget in WA and explored its variability during 2002–07. The analyses show that WA is a moisture source rather than a sink during the summer season. Bock et al. (2011) suggested that the representation of moist convection in numerical models is a major problem simulating the WAM. Agustí-Panareda et al. (2010a) assessed the water cycle as represented in the ECMWF African Monsoon Multidisciplinary Analyses (AMMA) reanalysis data with and without radiosonde humidity bias correction in comparison to the Meynadier et al. (2010a) hybrid dataset and highlight the presence of large biases in the forecast model associated with the parameterization of physical processes.
The annual cycle of water vapor transport over WA associated with the regional circulation was investigated by Thorncroft et al. (2011). Their study found that the monsoon southerly flow is established between the Atlantic cold tongue and the SHL. The study also found that the cold tongue regulates the timing and intensity of the coastal rainfall in the spring, while the SHL brings moisture into the continent. Mera et al. (2014) studied the variability and transport of surface moisture in WA using NCEP global model analyses, satellite, and stations observations during the spring. Their surface air parcel trajectory analysis, both horizontally and vertically, revealed that the northeastern Atlantic, the Gulf of Guinea, the South Atlantic, North Africa, and the Mediterranean Sea are moisture sources for WA during the February–June period. Mera et al. (2014) also found the levels of the atmosphere from which the air originated as the seasons progressed. Using the ECMWF global dataset at 600 hPa, Spinks and Lin (2015) investigated the variability of the Saharan and Arabian highs, the African easterly jet (AEJ) and associated local wind maxima (LWM) over West and East Africa, and easterly waves during the August 1979–2010 period. They suggested that, while baroclinicity explains the formation of the LWM, the maximum geopotential gradient located to the south of the Saharan high center explains the increased easterly wind flow that results from the LWM being centered on average at 16.2°N, 12.3°W.
3. Data and methodology
The daily and monthly rainfall records covering 1979–2008 over the WAM region are used for 180 stations (Fig. 1) in this study. The station rainfall totals were provided by the national weather services of countries across the region. The data were quality controlled, with values beyond physically reasonable limits being excluded. The majority of the station records were incomplete, with considerable variation in the length of continuous observations and the presence of gaps in the records. Because no attempt was made in this study to estimate missing values, the 180 stations selected were those with the most complete and reliable records. In addition, a regional-scale time series of average normalized April–October (1941–2012) rainfall departure (referred to as the Lamb index; e.g., Boyd et al. 2013) is also used. The Lamb index results from an average normalized departure for 20 stations in the West African Sudan–Sahelian zone (11°–18°N, 17.5°W–8.5°E).
The Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP; Xie and Arkin 1997) and the Global Precipitation Climatology Project (GPCP; Xie et al. 2003) satellite datasets also were used. Both the CMAP and GPCP datasets blend rain gauge observations and satellite estimations. They are gridded in 2.5° latitude × 2.5° longitude boxes and are available from 1979 to the present. Over land areas, the CMAP and the GPCP datasets were shown to have similar major precipitation patterns (e.g., Xie and Arkin 1997; Yin, et al. 2004).
The NCEP–NCAR reanalysis data (Kalnay et al. 1996) provide global atmospheric information from 1948 to the present. It is important to note that satellite soundings were not available before 1979; for this reason, this study will focus on the period from 1979 to 2008, when the data are most reliable. These data are divided into three categories. The most reliable variables are type A variables and are strongly influenced by the observations. These include zonal and meridional wind components. Type B variables, however, are influenced by both the observations and the model. Among this category are the low-level winds and the specific humidity. Type C variables (e.g., precipitation) are completely determined by the model. This study uses the zonal (u) and the meridional (υ) wind components, the specific humidity (q), and the surface air pressure (ps) to calculate moisture fluxes. However, for data-sparse regions like WA, the NCEP–NCAR reanalysis reliability is questionable, and many studies have found substantial errors in the WAM region in global weather and climate models (Agustí-Panareda et al. 2010a; Xue et al. 2010; Meynadier et al. 2010b; Bock et al. 2011). Over Africa, more emphasis frequently is placed on model forecasts because of the limitations impose by the lack of observations, especially radiosonde observations. Moreover, biases in both the observations and the models are known to introduce spurious variability and trends into the NCEP–NCAR reanalysis (e.g., Bock and Nuret 2009; Agustí-Panareda et al. 2010b). Consequently, results from such analyses should be considered with due caution.
b. Vertically integrated moisture flux
The total horizontal mean flux components (zonal and meridional) of water vapor are calculated by vertical integration, using the following equations:
where is the zonal component and is the meridional component of the vertically integrated total mean vapor flux. The variable g is the gravitational acceleration. The value of g is 9.81 m s−2, so the units for and are kg s−1 m−1. No attempt was made to analyze the contribution of the eddy flux and mean flux terms separately. The column pressure extends from the surface of the earth ps to a pressure surface pu, above which the water vapor content is negligible. As moisture flux in the WAM region is concentrated in the lower altitudes (e.g., Lamb 1983; Cadet and Nnoli 1987; Bielli and Roca 2010; Pu and Cook 2012), the low-level moisture flux of the surface–850-hPa layer will be considered.
a. Seasonal cycle of rainfall
The origins of the seasonal cycle in WA are linked to the annual cycle of solar declination. Peyrillé et al. (2007) and Thorncroft et al. (2011) have described the seasonal circulation that characterizes the WAM system. It is a pronounced seasonal wind shift produced by the thermal contrast between the hot Sahara and the Atlantic Ocean and the associated southwesterly monsoon flow. This occurs during the boreal spring (April–June), when the temperatures are high over the Sahel and where the highest surface temperatures coincides with the lowest surface pressure, forming the SHL. A cold tongue develops over the equatorial Atlantic. The low pressure system in the SHL creates a steady wind that blows toward the land carrying the near-surface moist air from the Atlantic Ocean, which acts to shift the ITF farther north and initiates the onset of the WAM.
Spatial distributions of monthly mean rainfall interpolated from rain gauges and associated ITF latitude positions from the NCEP–NCAR reanalysis are shown in Fig. 2. The ITF is represented by the 15°C isodrosotherm (Lélé and Lamb 2010), which is located around 10°N during the spring season and gradually moves northward. It reaches its northernmost latitude of approximately 20°N in August over the Sahelian region and then abruptly retreats southward thereafter. Concomitant with the ITF, abundant rainfall first appears over the Gulf of Guinea coast in March and peaks in June. As the season progresses, precipitation extends northward over the Sudan zone (~8°–11°N), but the most significant rainfall during April–June still occurs in the Gulf of Guinea (Figs. 2b,d). In July, the rainfall band suddenly moves northward over the Sahel (Fig. 2e), where the most intense rainfall is observed in August, characterizing the peak of the WAM (Fig. 2f). At that time, precipitation decreases over the Guinea coast; this is referred to as the “little dry season.” Precipitation progressively retreats southward in September–October (Figs. 2g,h), marking the end of the monsoon season, the beginning of a long dry period over the Sahel, and the start of a second rainfall season over the Gulf of Guinea. This results in a bimodal rainfall distribution over the Guinean coast with the first peak occurring in June and the second peak in late October. The difference in the Sahelian and the Guinean coast rainfall regimes indicates that there may be two different physical processes modulating them (e.g., Gu and Adler 2004).
b. Characteristic features of moisture flux field over West Africa
The large-scale characteristics of the WAM low-level water vapor flux field during 1979–2008 are illustrated in Figs. 3 and 4. The contours in these figures represent the moisture flux vector magnitudes, and the shading represents the absolute values of moisture flux magnitudes equal to or greater than 100 kg m−1 s−1, with a 40 kg m−1 s−1 interval. The dashed line indicates the latitude of the ITF.
1) Spring season (April–June)
During March–April (Figs. 3a,b), the ITF is centered near the Gulf of Guinea coast, between the southern coast of WA and south of 10°N. On average it reaches 10°N in April in the western portion of the region but is slightly farther south in the eastern portion. Even in these premonsoonal months, some cross-equatorial southwesterly moisture flow is observed, much of it redirected onto the Gulf of Guinea coast under the influence of the quasi-permanent anticyclonic circulation centered south of 10°S. Although the monthly mean vectors are relatively small off the Gulf of Guinea coast, there is a substantial mean transport of approximately 120–160 kg m−1 s−1 onto the Gulf of Guinea coast. This enhanced moisture flux is associated with the northward shift of the ITF from its winter position near the Guinea coast to around 10°N in April. It is also associated with the local sea-breeze/land-breeze regime and the onset of the first rainy season, which occurs over the Guinean coast during the spring season (e.g., Sultan and Janicot 2003).
Another notable feature in March and April is the prominent interhemispheric, northeasterly moisture transport off the west coast of Africa between 20° and 30°N. Although this zone of moisture transport has a weaker flow pattern at this time, it merges with the cross-equatorial southwesterly flow south of the ITF in the region centered at 0°–5°N. This interhemispheric moisture transport increases considerably in April (Fig. 3b) and feeds the increased zonal moisture flux over the Gulf of Guinea region. This northeasterly moisture transport is in agreement with Mera et al. (2014), who found that during April moisture sources over WA are scattered from 650 hPa at 30°N to lower than 850 hPa south of the equator. Meanwhile, neither the cross-equatorial southwesterly monsoon flow nor the northwesterly monsoonal flow affects the Sahelian region of WA (i.e., north of 10°N). Instead, northeasterly trade winds emanating from the Libyan anticyclone dominate this period.
The salient features in May and June are the progression and strengthening of the northward advancing southwest inflow of moisture and the sudden northward latitude shift of the ITF from its position around 10°N in April to 15°N in June (Figs. 3c,d), resulting in a strong moisture buildup over the Guinea coast and Sudan zones in the spring season. During these months, the latitude belt from 15°S to the equator in the Atlantic Ocean is the primary source of water vapor for the region. The mean southwest transport from these latitudes reaches a maximum around 12°N. It has a well-demarcated region of strong flow, referred to in this study as the “moisture river.” This strong flow originates from the equatorial South Atlantic, moves northeastward across the equator, passes over the Gulf of Guinea region, and then supplies moisture to the central Sudan region. The enhanced southwesterly transport is associated with both a strengthening and poleward shift of the St. Helena high pressure system and the rainfall peak observed during the spring season in June over the Gulf of Guinea region. The moisture river inflow continues to strengthen and advances northward, leading to a moistening of the Sudan–Sahel zone, where a maximum moisture buildup is reached and the transport exceeds 200 kg m−1 s−1 in June. This strengthening of the southerly flux south of the equator during spring, as well as its associated moisture buildup, has previously been documented (e.g., Omotosho 1990; Omotosho et al. 2000; Thorncroft et al. 2011; Mera et al. 2014) and is suggested as being linked to the development of a cold tongue over the tropical Atlantic Ocean (e.g., Thorncroft et al. 2011), with a source being lower than 850 hPa and south of the equator (e.g., Mera et al. 2014).
2) Summer season (July–September)
The most active phase of the WAM in the Sahel occurs during the summer, by which time the seasonal cycle of precipitation north of 8°N latitude consists of a single rainy season maximum, which could be brief as three months, from July to September. Other important features occur during the summer, which is the season when the atmosphere–ocean–land surface interactions are strong, and when the Sahelian precipitations are abundant. Figure 4 displays the climatology of moisture flow fields during the summer season months. As the monsoon extends into these boreal summer months, the moisture river weakens and becomes less extensive. However, the boundary between the northerly and the southwesterly flow just south of 15°N in June shifts northward during the early summer in the Sahelian zone. This accounts for the seasonal change in sign of rainfall across the Gulf of Guinea and Sahel, with rainfall (no rainfall) over the Gulf of Guinea (Sahel) region during the spring and little or no rainfall in the so-called little dry season during the summer. The strong inflow through the Guinean coast, previously observed in the spring season, now exhibits two low-level westerly currents entering the continent. The first is a southwesterly flow across the Guinean coast, which is known as the WAM flow. Almost all of its northward flow into the Sahel originates from the Gulf region and is associated with the Sahelian rainy season. For instance, Lamb (1983) found that the highly deficient 1972 sub-Saharan rainy season is associated with a shallow southwesterly flow that contains less than half the water vapor in the entire tropospheric column. This implies that the Gulf of Guinea region is an important source of moisture for the Sahel. Hence, changes in the sea surface temperature (SST) in the region will impact Sahelian rainfall variability, as pointed out in some earlier studies (e.g., Lamb 1978a,b; Folland et al. 1986; Ward 1998; Giannini et al. 2003). The northward extent of the southwesterly moisture flux increases as the season progresses and reaches its northernmost latitude of around 20°N in August, corresponding to the peak of the Sahelian rainy season. As indicated by Thorncroft et al. (2011), the northward extension of the southwesterly moisture flux in July–September (JAS) is determined by the location of the SHL.
The central–eastern portions of the Atlantic Ocean also act as a distribution zone for moisture entering from the Atlantic. This second westerly moisture flow is established near 8°N latitude and has a major inflow directed from the eastern Atlantic onto the west coast of the Sahelian zone. This is essentially the flow pattern found in other studies that analyzed the zonal component of moisture transport over WA and recognized it as a fundamental feature of the rainfall cycle over the Sahelian region (e.g., Kidson 1977; Cadet and Nnoli 1987; Druyan and Koster 1989; Cook 1999; Grams et al. 2010; Thorncroft et al. 2011; Pu and Cook 2012). For example, Druyan and Koster (1989) use the GISS climate model to compare the water vapor transport into WA and found that the westerly water vapor from the North Atlantic is the largest moisture current for the western portion of the Sahel, while that from the Gulf of Guinea contributes to the moistening of the central Sahel. Later, Pu and Cook (2012) associated this westerly moisture current to the West African westerly jet (WAWJ). They suggested that Sahel drought years are characterized by negative anomalies of moisture transport from the eastern Atlantic due to a weak WAWJ, while during more abundant rainy years the jet becomes stronger and associated with an enhanced eastward transport of moisture flux.
The peaks of the westerly and southwesterly fluxes coincide with the peak of Sahelian precipitation maximum, in August, with flux magnitudes over 160 kg m−1 s−1 (Fig. 4b). However, the highest rainfall totals are reported over the Guinea highlands and Cameroon mountains instead, where the mean monthly rainfall in August exceeds 12 mm day−1 (Fig. 2f). These regional-scale precipitation differences may be due in part to both the zonal and meridional terrain distortions of the westerlies. The mountain ranges force moist air to ascend, causing the air to cool and the excess moisture to condense, producing heavy precipitation. Thereafter, the southward moisture retreat across the Sahel occurs rapidly between September and October. This withdrawal proceeds most rapidly in the western region of Africa. Lélé and Lamb (2010) obtained a similar withdrawal pattern in their ITF variability study. By October, the monsoon southwesterly and the westerly flows have retreated southward from the Sahelian zone. The flow pattern for October resembles the premonsoon month of March, also characterized by the equatorward retreat of the ITF and reestablishment of the prevailing near-surface stream of dry and hot northeasterly winds, which is part of the continental trade wind system.
c. Annual cycle of moisture flux convergence
The annual cycle for the total moisture flux convergence, averaged between 10°W and 10°E and over the years 1979–2008, is shown in Fig. 5 as a Hovmöller time–latitude diagram. Also shown in the diagram is the latitudinal position of the ITF. Over the entire WAM domain, the total field presents all essential features of the climatological rainfall distribution (Fig. 2). Notably, a large area of moisture flux convergence is observed over land between 10°W and 10°E, whereas a moisture flux divergence area is present to the south, over the Atlantic Ocean. The northward advance of flux convergence across the Gulf of Guinea coast begins in March, by which time the convergence zone has migrated north of 8°N from its winter position just north of the equator. Along the Gulf of Guinea there is some evidence of a belt of increased convergence in the spring season, similar to the belt of higher rainfall seen in the annual cycle. During the summer season, the belt gradually advances over the Sahel. This advance, which reaches its poleward location south of 20°N in early August, is enhanced as the ITF proceeds rapidly northward during July, when both the SHL and the westerly flow across the west coast of Africa intensify. By this time, in the Gulf of Guinea the very small moisture flux convergence is consistent with the observed very low or zero accumulated rainfall (i.e., the little dry season). Water vapor convergence generally prevails north of the equator, with the center of the maximum flux convergence located south of the ITF latitude. This is consistent with the observation that appreciable rainfall over WA generally occurs more than 400 km south of the ITF (Hastenrath 1991, p. 169; Lélé and Lamb 2010).
Area averaged, mean daily rate of total moisture flux convergence for the Gulf of Guinea is about 8 mm day−1 during April–June and about 4–8 mm day−1 in the Sahel during July–August. These figures are quantitatively in good agreement with the observed rate of precipitation shown in Fig. 2 and therefore indicate that the transport of moisture flow from both the east and South Atlantic is the primary source for rainfall over WA. Local evaporation resulting from the interaction between local groundwater and climate constitutes the second largest contributor to precipitation in WA, as suggested in other studies (e.g., Druyan and Koster 1989; Gong and Eltahir 1996). The strongest moisture buildup, however, occurs during the April–June spring season rather than during the peak summer monsoon season in July–September, suggesting that the moistening of the atmospheric column during the WAM season occurs during the premonsoon season. Beginning late August and September, the convergence belt gradually weakens and abruptly retreats southward. A much weaker moisture flux convergence belt is observed over the Guinea coast. This secondary zone of convergence of moisture prevails through the autumn season, consistent with the start of the second rainfall season in the Guinea coast.
However, during the Sahelian phase of the WAM (late July–early September) the vertically integrated moisture flux has a west-southwest–east-northeast orientation (Fig. 4b) and likely is responsible for the dual-maximum moisture flux convergence depicted in Fig. 5 during the same period. This set of dual maxima, centered south of 20°N, near 7°N, and separated by a weak minimum, results from the wide range of longitudinal averaging in which peak values near 10°W are at lower latitudes than peak values near 10°E. The belt of convergence over land is accompanied by a large area of moisture flux divergence over the ocean. This area of flux divergence comprises two maxima, each one located below the maxima of flux convergence over land. The two maxima are separated by a divergence minimum. This juxtaposition reveals that moisture transport over WA is regulated by an enhanced evaporation over the tropical Atlantic Ocean. Others studies (e.g., Lamb 1983; Cadet and Nnoli 1987; Fontaine et al. 2003; Thorncroft et al. 2011) also found that that the Atlantic Ocean is the main source of water vapor to the WAM and that the transport is essentially poleward.
d. Moisture transport and associated rainfall pattern
To evaluate the temporal and spatial evolution of the circulation and precipitation associated with moisture fluxes, two composites were formed based on the meridional flux (). The first composite consists of southerly flow at latitude 5°N, between 15°W and 15°E, which occurred during April–June, and exceeds one standard deviation greater than the long-term mean meridional flux. The second is similar to the first but consists of fluxes at latitude 10°N in July–September. Both composites are characterized by strong flows into the Gulf of Guinea and the Sahel, respectively. A Student’s t test is carried out on the composites at each grid point to assess whether the anomaly fields are statistically different from zero. However, the flux vector fields are plotted if any of the components are significant at the 95% confidence level. The spatiotemporal northward migration of the moisture flux is displayed in Fig. 6, where the vector amplitude is proportional to the vector length. The shading indicates the large-scale rainfall pattern during April–June and July–September. In general, the seasonal cycle depicted in Fig. 6 is consistent with the characteristics of the seasonal cycle of the measured precipitation. During spring, the moisture flux is largest in the Guinea coast, consistent with the observed large-scale region of enhanced precipitation (Fig. 6a) over the area. At the same time, precipitation is zero over the Sahelian area. The strong southerly flow into the Sahel is observed in the summer season, by which time the largest fluxes exhibits double inflow currents. One current is the monsoon flow from the Gulf of Guinea and the other is the westerly from the west coast, as previously indicated. They originate from the south and eastern Atlantic Ocean and penetrate through the location of the rainband. This northward advance of the southwesterly monsoon fluxes was accompanied by a northward shift of the rain producing system from the Guinea coast to the Sahel. The two rainfall maxima are observed during the summer season. One is over the west coast of Africa and another is over the eastern portion of WAM domain. These rainfall maxima coincide with the immediate regions where the two strongest currents enter into the Sahel. Figure 6 indicates that the largest southwesterly moisture transport occurs in the spring, as previously mentioned, rather than in summer. Above 15°N along the west coast, a southward transport of moisture flux off the coast of Africa is observed: some of which is redirected into the western Sahel under the influence of the low-level cyclonic circulation centered inland at about 18°N, 15°W. North of 20°N, the low-level moisture flux into the Sahel is rather weak and consists of a southward transport. This moisture current was investigated by Fontaine et al. (2003), who suggested that, in addition to the Gulf of Guinea, the low-level moisture into the Sahel originates from the Mediterranean Sea.
e. Case studies of wet and dry years
To examine the relationship between moisture transport and convection in individual extreme years, the extreme 1984 drought year (DRY) and the 1988 wet year (WET; which is the first wet year after the 1984 drought) were chosen. The outgoing longwave radiation (OLR) is used as a proxy for tropical convection. OLR also is controlled by the presence of water vapor in the atmosphere and cloud cover, and reveals some information on the temperature, humidity, and cloudiness. Figure 7 shows the seasonal average (July–September) of moisture flux (vectors) and OLR (shading) for 1984 DRY (Fig. 7a) and 1988 WET (Fig. 7b). Note that the 1984 Sudan–Sahel drought was associated with a classic tropical Atlantic dipole SST anomaly pattern [negative (positive) departures north (south) of ~10°N; Lamb and Peppler 1992] and a dipole West African rainfall anomaly pattern in which the Gulf of Guinea rainfall was above average (Ward 1998; Nicholson and Grist 2001; Nicholson and Webster 2007; Nicholson 2008). The tropical Atlantic dipole is characterized by opposing SST anomalies between the tropical North and South Atlantic Oceans. In the Sahel region, the 1984 moisture flux shows a pronounced southward extent, especially over its western and central regions. The southward displacement of the ITF also was pronounced compared with its long-term average latitude position. Over the west coast, the cyclonic cell climatologically centered at about 18°N, 15°W (Fig. 7b), shifted southward, suppressing the strong westerly flux current into the Sahelian region initially found during the summer season. In contrast, strong fluxes dominate the Guinea coast. Consequently, areas of low OLR, synonymous with enhanced convection are seen in the Guinea coast rather than the usual summer location in the Sahel. This produced the widespread and severe drought of 1984 across the Sahel and produced large-scale water and food deficits, famine, and human fatalities. Thus, the southward moisture flux limit, the southward ITF displacement, and the associated Sahel rainfall suppression were more pronounced when the tropical Atlantic dipole SST anomaly pattern was observed.
During the 1988 wet season, the moisture flux transport extends farther north because of the strengthening of the westerly and southwesterly inflow across the west coast and the gulf regions. The ITF migrated northward and reached its northernmost latitude position around 20°N. The cyclonic cell along the western coast deepened and reached a northernmost location at about 22°N, 12.5°W. Hence, the WAWJ was enhanced and brought increased moisture into the west coast. The subsequent zone of moisture flux convergence lay across the Sahel region (Fig. 7b). Accordingly, a large area convective activity, as well as associated abundant monsoon rainfall, was observed over the Sahel, whereas relatively dry conditions persist in the Guinea coast. During the dry 1984 rainy season, there were less than average (with a more southward extent) low-level southwesterly and westerly moisture flows as they crossed the West African coast. The strong southwesterly flow in the Guinea coast during the 1984 season can be traced a few hundred kilometers upstream off the coast. Their southward extent seemingly resulted from the warm SST anomaly identified in several other studies (e.g., Lamb and Peppler 1992; Ward 1998; Nicholson and Grist 2001; Nicholson and Webster 2007; Nicholson 2008). However, during the 1988 wet year, the northward extent of the fluxes resulted from the cold SST anomaly in the South Atlantic Ocean.
f. Ocean-to-land moisture transport variations during dry and wet years
The breakdown of the ocean–land moisture transport into its zonal and meridional components was evaluated for four very WET years (1988, 1994, 1998, and 1999) and four very DRY years (1982, 1983, 1984, and 1987) across WAM boundaries. The selection of these years is based exclusively on the Lamb index, where they appear as pronounced wet and dry years. While the meridional flux is evaluated along southern (5° and 10°N) and northern (20°N) boundaries, the zonal flux is evaluated along the western (15°W) and eastern (15°E) boundaries. These boundaries were chosen according to WA climatic airflow patterns, and they account for most of the total inflow from the moisture sources. Figure 8 displays the seasonal average inflow to WA across the lateral boundaries for the long-term mean and for the WET and DRY composites. Each bar indicates the magnitude and direction of the water vapor flux at region boundaries, taken at 2.5° latitude–longitude intervals.
Figure 8 reveals that different mechanisms are responsible for the summer flux transport into the Sahelian zone during the WET and DRY composites. During the Sahelian July–September WET regime, strong westerly flow, extending from 7.5° to 15°N, dominates water vapor transport over WA (Fig. 7a). The zonal outflows are a magnitude smaller than the zonal inflows. However, there is still a large outward zonal flow along 15°E compared to the long term, indicating that the maximum low-level moisture transport is from the west (Cadet and Nnoli 1987; Thorncroft et al. 2011; Pu and Cook 2012). The southerly fluxes sustained from the Guinea coast to the Sahelian region with a slightly decreasing intensity to the north. Nevertheless, it is noted that the meridional fluxes across the Sahelian boundary are weaker over its western portion and larger over its central region, between approximately 2.5°W and 7.5°E. This shows that during WET years a large amount of moisture, but less than the zonal moisture fluxes, comes from the Gulf and is transported northward to the Sahel. Overall, net positive moisture flux characterizes the Sahel during WET regimes, owing to the fact that the inflows exceed the outflows across much of the grid points along 15°W, 15°E, 10°N, and 20°N.
During DRY regimes, both the zonal and meridional fluxes are less than the long-term average. The weakening of the zonal flows into the Sahelian zone is accompanied by large zonal outflows (Fig. 8a: red bars at 15°W), which are locally greater than the climatological outflow. A similar situation is observed for the meridional fluxes, which is an indication of a net negative moisture transport flux in the region during DRY regimes. The differences in moisture transport crossing the lateral boundaries reflect the differences in the transport patterns described for contrasting individual years. The mechanism for the alternation between WET and DRY regimes is described by Hastenrath and Polzin (2014). As they indicated, the strong easterly flow observed during a WET regime is favored by warmer (colder) SST located to the Atlantic northwest (southeast) of the band of the warmest waters compared to the long-term mean, lower pressure to the north of the long-term mean low pressure trough, and enhanced (reduced) southerly (northerly) wind. Therefore, DRY years are characterized by a weakening of the westerlies off the west coast of WA, consistent with previous studies (e.g., Cadet and Nnoli 1987; Thorncroft et al. 2011; Pu and Cook 2012; Grams et al. 2010; Meynadier et al. 2010a). It also is consistent with both a weakening of the southerly winds off the coast of Africa and with the southward shift of the ITF latitude (Lélé and Lamb 2010) favored by warming in the tropical South Atlantic rather than the North Atlantic, consistent with earlier analysis (e.g., Lamb 1978a).
g. Correlation between moisture transport components and Sahelian precipitation
To investigate the strength of the relationship between moisture transport components and Sahelian rainfall, a moisture transport index is defined as an area average of the zonal and the meridional moisture transport along the south and west lateral boundaries. These boundaries were characterized by low-level wind fluctuations tied to changing in large-scale circulation. Figure 9 illustrates the lagged correlations between the Sahelian precipitation on pentad (5 day) scale and components of moisture flux. Only correlation coefficients exceeding the 95% confidence level are displayed. Figures 9a and 9b show the correlations between the GPCP and the zonal and meridional components, respectively. Figures 9c,d are similar to Figs. 9a,b, except that the correlation is with the CMAP data. In general, significant positive correlations extend across the Sahel, but they are considerably lower at zero lead time. However, they continue to be larger as the lead time increases. For instance, when the zonal moisture transport leads by one to four pentads, significant positive correlations extend across the Sahel region, suggesting that the zonal component of moisture transport exerts a significant control on rainfall over the Sahel. Significant correlation coefficients also are depicted between the GPCP precipitation and the meridional component. However, these correlation coefficients appear to be significant only from lags 3 and 4, and they are not as large as the correlation coefficients between the zonal moisture transport and its rainfall counterpart. These two correlation patterns between components of moisture transport and rainfall over the Sahelian zone suggests that, more than the southwesterly moisture transport from the Gulf of Guinea, the zonal transport of moisture from the eastern Atlantic is the major contributor to the July–September Sahelian rainfall variability; this result is in agreement with previous analysis.
A similar correlation analysis is performed using the components of moisture transport time series at the western and southern boundaries of Sahelian zone and CMAP data (Figs. 9c,d). Similar correlation patterns as those when the GPCP precipitation was used emerge, indicating that the relationship between moisture transport anomalies and Sahelian rainfall is robust and consistent across the data used. Although the simultaneous correlation between the zonal flux component and precipitation (lag 0) is positive, it is weak and localized compared with the other lag correlations. This suggests that the precipitation over the Sahelian region lag the moisture inflow into the region, in agreement to the fact that the moistening of the atmospheric column over WA occurs during the premonsoon season and during the spring season. The extent and magnitude of the correlation between Sahelian precipitation and the zonal component of moisture transport, for other lags, supports the previous claim that the eastern Atlantic Ocean is a major source of moisture for the Sahelian rainfall regime. This source of moisture for the Sahelian region is also noted by Cadet and Nnoli (1987) and more recently by Long et al. (2000), Thorncroft et al. (2011), Grams et al. (2010), and Pu and Cook (2012), who found that the easterly flow dominates water vapor transport patterns over sub-Saharan Africa.
5. Summary and conclusions
The water cycle is an integral part of the global energy cycle and plays a fundamental role in determining the large-scale circulation and precipitation patterns. Over WA, it has been long recognized that the southwesterly monsoon flux in the Gulf of Guinea is a key source of water vapor for Sudan–Sahel summer precipitation (e.g., Lamb 1978a; Cadet and Nnoli 1987; Gong and Eltahir 1996). The Mediterranean basin, local evaporation, and evaporation in areas east of the WAM domain have also been identified as sources of water vapor (e.g., Gong and Eltahir 1996; Fontaine et al. 2003; Nieto et al. 2006). However, the variations of these fluxes, along with alternations between WET and DRY regimes, are poorly understood, despite evidence from previous studies showing that the ITCZ, with its north–south progression, is responsible for much of the rainfall variability in the region (e.g., Long et al. 2000).
In this study, the origins and mechanisms of the surface–850-hPa transport of moisture and moisture convergence into WA during the summer monsoon season and their impact on rainfall variability are examined using NCEP–NCAR reanalysis data. The interannual variability of the moisture fluxes entering the WA was examined, and a comparison of composite of dry years (DRY) in Sudan–Sahel was made with patterns for their counterpart wet composite (WET). The primary objective was to improve our understanding of the underlying atmospheric moisture variability on which WAM rainfall depends.
The seasonal evolution of the moisture fluxes, their convergence, the ITF displacement, and associated rainfall is strongly affected by the intensity and latitude position of the equatorial Atlantic cold tongue, the Atlantic semipermanent high pressure systems, the WAWJ, and the SHL. The transport of moisture occurs mainly in phases and starts from the premonsoon season in the South Atlantic Ocean, continuing to the fully developed monsoon phase inland. The premonsoon onset in March–April is characterized by cross-equatorial southwesterly moisture flow over the Gulf of Guinea. This relatively weak southwesterly flow transports a substantial amount of moisture into the Guinea coast, and is associated with the northward shift of the ITF from its winter position near the Guinea coast to around 10°N in April. Two main flux fields observed entering WA characterized the major features during the spring season. One is the intense southwesterly monsoon current associated with the extremity of the subtropical high pressure belt of the Southern Hemisphere, generally between 20°S and the equator and between 30° and 10°W, and the sudden northward latitude shift of the ITF. The second major inflow area is the northeastward flow associated with the southwestern extremity of the Atlantic subtropical high pressure belt. These two moisture currents appear to be responsible for a strong moisture buildup along the Guinea coast and the associated April–June rainy season. As the monsoon season further develops during the summer, the evolution of the low-level monthly fields of mean fluxes of water vapor show a progressive penetration of moisture over the Sahelian zone. A gradual northward shifting of the location of the ITF is noted, whereas the northward moisture current is weakened. However, there is an increase of zonal moisture transported by the trade winds north of the equator, off the west coast of WA. During the peak of the monsoon season (August), the southwesterly transport from the Gulf of Guinea further weakens, while the transport by the trade winds over the west coast intensify and feed the Sahelian region with substantial water vapor. As recently indicated by Pu and Cook (2012), this eastward transport of moisture is attributed to the development of the WAWJ.
Two latitudinal bands of moisture convergence were found in this study. One is a large convergence zone extending from the Guinea coast to the Sahel. It has its maximum during the spring season, which suggests that the moistening of the atmospheric column in WA mainly occurs during the spring rather than during the peak summer season. It is also clear that convergence of water vapor generally prevails north of the equator, and the center of the maximum flux convergence is located south of the ITF latitude. This is consistent with observations that show substantial rainfall over WA generally occurs more than 400 km south of the ITF (Hastenrath 1991; Lélé and Lamb 2010). The maximum moisture convergence during March–May period is about 8 mm day−1 over the Gulf of Guinea, whereas it is 0–2 mm day−1 over the Sahel. However, during the full development of the monsoon, moisture convergence increases to more than 8 mm day−1 over the Sahelian zone. This is consistent with a dipole rainfall pattern observed in WA.
On an interannual time scale, observations reveal a significant interannual variability in moisture transport. Drought years are generally associated with large decreases in moisture content in the low-level due to the weakening of the southwesterly flow, a much farther southward extent of moisture across WA, and a southward ITF position compared to its long-term average latitude. This is in agreement with Lamb (1978a), who also found that during a DRY regime the tropical Atlantic near-equatorial trough, the kinematic axis, and the zone of maximum SST are displaced 200–300 km south of their long-term average locations. Moreover, during a WET regime, the transport of moisture flux extends farther north, because of the strengthening of the eastward and northward inflows, and the ITF reaches its northernmost latitude position around 20°N. The breakdown of the ocean–land moisture transport into its zonal and meridional components suggests that WET monsoon regimes over WA are characterized by enhanced eastward zonal flux across the west boundary of WA and a decreased northward transport across the southern boundary. Also, the outflow along the eastern and northern limits of the Sahel during WET years is typically of smaller magnitude than their inflow counterparts, regardless of the season.
The correlations between components of moisture transport and rainfall confirm the strong association between the zonal moisture transport and Sahelian precipitation. While significant positive correlation coefficients exist between the zonal flux component and rainfall starting from zero lag, they appear only at lag 3 between the meridional component and Sahelian rainfall. The statistically significant positive correlations (beyond 5% level) between the zonal moisture flux and Sudan–Sahel rainfall is most pronounced when the zonal flux lead by 1–4 pentads (r from +0.44 to +0.50). Although weak, they still are statistically significant at the 95% confidence level and at lags 3–4 when the meridional moisture flux is employed (r = +0.3). Comparison of June–September simultaneous correlations (lag 0) with other lag relationships reveals that the precipitation occurrences over the Sahel lag the moisture inflow into the region, which is in agreement with the fact that most moistening of the atmospheric column over WA occurs during the spring season.
The main question addressed here is the origin and variability of the water vapor fluxes which are the source of precipitation over the Sudan–Sahel zone of WA. The mechanisms regulating the alternation between WET and DRY years found in this study mirrors the general circulation patterns for contrasting years found in earlier studies (Lamb 1978a,b, 1983; Nicholson 1985; Hastenrath 1990; Hastenrath and Polzin 2011; Lélé and Lamb 2010; Hastenrath and Polzin 2014). Generally, a WET year is favored by warm (cold) SST in the Atlantic northwest (southeast) of the band of warmest water compared with the long-term mean, lower pressure to the north, enhanced (reduced) southerly (northerly) wind, a northward shift of the ITF latitude, enhanced easterly wind. For DRY years, in contrast, intense atmosphere–ocean anomalies are confined to the Gulf of Guinea coast.
This paper is dedicated to the memory of Professor Peter J. Lamb, under whose scientific guidance and teaching the first author earned his graduate degrees. Dr. Lamb’s belief in African institutions is leading the way in transforming African climate research development and commitment to enhancing the frontiers of African meteorology. The authors also thank Drs. Zewdu T. Segele and Michael Douglas for valuable suggestions and editorial comments. Finally, the authors express their gratitude to the anonymous reviewers whose comments and suggestions considerably improved the manuscript. The research was supported by the Cooperative Institute for Mesoscale Meteorological Studies, University of Oklahoma.
This article is included in the In Honor of Peter J. Lamb special collection.