• Aguilar, E., and et al. , 2009: Changes in temperature and precipitation extremes in western Central Africa, Guinea Conakry, and Zimbabwe, 1955–2006. J. Geophys. Res.,114, D02115, doi:10.1029/2008JD011010.

  • Annamalai, H., , J. M. Slingo, , K. R. Sperber, , and K. Hodges, 1999: The mean evolution and variability of the Asian summer monsoon: Comparison of ECMWF and NCEP/NCAR reanalyses. Mon. Wea. Rev., 127, 11571186, doi:10.1175/1520-0493(1999)127<1157:TMEAVO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Burde, G. I., , A. Zangvil, , and P. J. Lamb, 1996: Estimation of the role of local evaporation in precipitation for a two-dimensional region. J. Climate, 9, 13281338, doi:10.1175/1520-0442(1996)009<1328:ETROLE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, doi:10.1002/qj.828.

    • Search Google Scholar
    • Export Citation
  • de Wasseige, C., , D. Devers, , P. de Marcken, , R. E. Atyi, , R. Nasi, , and P. Mayaux, 2009: Les forêts du Bassin du Congo—Etat des forêts 2008. Office des publications de l’Union Europenne Rep., 426 pp.

  • de Wasseige, C., and et al. , 2012: Les forêts du bassin du Congo—Etat des forêts 2010. Office des publications de l’Union Europenne Rep., 276 pp.

  • Findlater, J., 1969: A major low-level air current near the Indian Ocean during the northern summer. Quart. J. Roy. Meteor. Soc., 95, 362380, doi:10.1002/qj.49709540409.

    • Search Google Scholar
    • Export Citation
  • Fontaine, B., , and S. Janicot, 1992: Wind-field coherence and its variations over West Africa. J. Climate, 5, 512524, doi:10.1175/1520-0442(1992)005<0512:WFCAIV>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Fontaine, B., , N. Philippon, , S. Trzaska, , and P. Roucou, 2002: Spring to summer changes in the West African monsoon through the NCEP/NCAR reanalyses (1968–1998). J. Geophys. Res., 107, doi:10.1029/2001JD000834.

    • Search Google Scholar
    • Export Citation
  • Fontan, J., , A. Druilhet, , B. Benech, , and R. Lyra, 1992: The DECAFE experiments: Overview and meteorology. J. Geophys. Res., 97, 61236136, doi:10.1029/91JD02426.

    • Search Google Scholar
    • Export Citation
  • Grodsky, S. A., , J. A. Carton, , and S. Nigam, 2003: Near surface westerly wind jet in the Atlantic ITCZ. Geophys. Res. Lett., 30, 2009, doi:10.1029/2003GL017867.

    • Search Google Scholar
    • Export Citation
  • Hagos, S. M., , and C. Zhang, 2010: Diabatic heating, divergent circulation and moisture transport in the African monsoon system. Quart. J. Roy. Meteor. Soc., 136, 411–425, doi:10.1002/qj.538.

    • Search Google Scholar
    • Export Citation
  • Jackson, B., , S. E. Nicholson, , and D. Kloster, 2009: Mesoscale convective system over western equatorial Africa and their relation to large-scale circulation. Mon. Wea. Rev., 137, 12721294, doi:10.1175/2008MWR2525.1.

    • Search Google Scholar
    • Export Citation
  • James, R., , R. Washington, , and D. P. Rowell, 2013: Implications of global warming for the climate of African rainforests. Philos. Trans. Roy. Soc.,368B, 20120298, doi:10.1098/rstb.2012.0298.

  • Johnson, D. R., , R. D. Townsend, , and M. Y. Wei, 1985: The thermally coupled response of the planetary scale circulation to the global distribution of heat sources and sinks. Tellus, 37A, 106–125, doi:10.1111/j.1600-0870.1985.tb00274.x.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and et al. , 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kanamitsu, M., , W. Ebisuzaki, , J. Woolen, , S. K. Yang, , J. J. Hnilo, , M. Fiorino, , and G. L. Potter, 2002: NCEP–DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 16311643, doi:10.1175/BAMS-83-11-1631.

    • Search Google Scholar
    • Export Citation
  • Li, Z., , Y. Chao, , and J. C. McWilliams, 2006: Computation of the streamfunction and velocity potential for limited and irregular domains. Mon. Wea. Rev., 134, 33843394, doi:10.1175/MWR3249.1.

    • Search Google Scholar
    • Export Citation
  • Matsuyama, H., , T. Oki, , M. Shinoda, , and K. Masuda, 1994: The seasonal change of the water budget in the Congo River Basin. J. Meteor. Soc. Japan, 32, 281299.

    • Search Google Scholar
    • Export Citation
  • McCollum, J. R., , A. Gruber, , and M. B. Ba, 2000: Discrepancy between gauges and satellite estimates of rainfall in equatorial Africa. J. Appl. Meteor., 39, 666679, doi:10.1175/1520-0450-39.5.666.

    • Search Google Scholar
    • Export Citation
  • Molua, E. L., , and C. M. Lambi, 2007: The economic impact of climate change on agriculture in Cameroon. World Bank Policy Research Working Paper 4364, 31 pp.

  • Nicholson, S. E., 2010: A low-level jet along the Benguela coast, an integral part of the Benguela Current ecosystem. Climatic Change, 99, 613624, doi:10.1007/s10584-009-9678-z.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., , and J. P. Grist, 2003: The seasonal evolution of the atmospheric circulation over West Africa and equatorial Africa. J. Climate, 16, 10131030, doi:10.1175/1520-0442(2003)016<1013:TSEOTA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., , and A. K. Dezfuli, 2013: The relationship of rainfall variability in western equatorial Africa to the tropical oceans and atmospheric circulation. Part I: The boreal spring. J. Climate, 26, 45–65, doi:10.1175/JCLI-D-11-00653.1.

    • Search Google Scholar
    • Export Citation
  • Pokam, W. M., , L. A. T. Djiotang, , and F. K. Mkankam, 2012: Atmospheric water vapor transport and recycling in equatorial central Africa through NCEP/NCAR reanalysis data. Climate Dyn., 38, 17151729, doi:10.1007/s00382-011-1242-7.

    • Search Google Scholar
    • Export Citation
  • Pu, B., , and K. H. Cook, 2010: Dynamics of the West African westerly jet. J. Climate, 23, 62636276, doi:10.1175/2010JCLI3648.1.

  • Rienecker, M. M., and et al. , 2011: MERRA: NASA’s Modern-Era Retrospective Analysis for Research and Applications. J. Climate, 24, 3624–3648, doi:10.1175/JCLI-D-11-00015.1.

    • Search Google Scholar
    • Export Citation
  • Rozendaal, M. A., , C. B. Leovy, , and S. A. Klein, 1995: An observational study of the diurnal cycle of marine stratiform cloud. J. Climate, 8, 17951809, doi:10.1175/1520-0442(1995)008<1795:AOSODV>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sheil, D., , and D. Murdiyarso, 2009: How forests attract rain: An examination of a new hypothesis. Bioscience, 59, 341347, doi:10.1525/bio.2009.59.4.12.

    • Search Google Scholar
    • Export Citation
  • Slingo, A., , R. Brown, , and C. L. Wrench, 1982: A field-study of nocturnal stratocumulus; III. High-resolution radiative and microphysical observations. Quart. J. Roy. Meteor. Soc., 108, 145165, doi:10.1002/qj.49710845509.

    • Search Google Scholar
    • Export Citation
  • Sultan, B., , and S. Janicot, 2003: The West African monsoon dynamics. Part II: The preonset and onset of the summer monsoon. J. Climate, 16, 34073427, doi:10.1175/1520-0442(2003)016<3407:TWAMDP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C. D., , H. Nguyen, , C. Zhang, , and P. Peyrillé, 2011: Annual cycle of the West African monsoon: Regional circulations and associated water vapour transport. Quart. J. Roy. Meteor. Soc., 137, 129147, doi:10.1002/qj.728.

    • Search Google Scholar
    • Export Citation
  • Tomas, R. A., , and P. J. Webster, 1997: The role of inertial instability in determining the location and strength of near-equatorial convection. Quart. J. Roy. Meteor. Soc., 123, 14451482, doi:10.1002/qj.49712354202.

    • Search Google Scholar
    • Export Citation
  • Vondou, D. A., , A. Nzeukou, , A. Lenouo, , and F. M. Kamga, 2010: Seasonal variations in the diurnal patterns of convection in Cameroon–Nigeria and their neighboring areas. Atmos. Sci. Lett., 11, 290–330, doi:10.1002/asl.297.

    • Search Google Scholar
    • Export Citation
  • Wada, M., 1969: Concerning the mechanism of the decaying of typhoon. J. Meteor. Soc. Japan, 47, 315335.

  • Washington, R., , R. James, , H. Pearce, , W. M. Pokam, , and W. Moufouma-Okia, 2013: Congo Basin rainfall climatology: Can we believe the climate models? Philos. Trans. Roy. Soc.,368B, 20120296, doi:10.1098/rstb.2012.0296.

  • Wood, R., 2012: Stratocumulus clouds. Mon. Wea. Rev., 140, 23732423, doi:10.1175/MWR-D-11-00121.1.

  • Yanai, M., , and T. Tomita, 1998: Seasonal and interannual variability of atmospheric heat sources and moisture sinks as determined from NCEP–NCAR reanalysis. J. Climate, 11, 463482, doi:10.1175/1520-0442(1998)011<0463:SAIVOA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yanai, M., , S. Esbensen, , and J.-H. Chu, 1973: Determination of bulk properties of tropical cloud clusters from large-scale heat and moisture budgets. J. Atmos. Sci., 30, 611627, doi:10.1175/1520-0469(1973)030<0611:DOBPOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., , P. Woodworth, , and G. Gu, 2006: The seasonal cycle in the lower troposphere over West Africa from sounding observations. Quart. J. Roy. Meteor. Soc., 132, 25592582, doi:10.1256/qj.06.23.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Seasonal means of the zonal wind (m s−1) for (left)–(right) ERA-Interim, MERRA, NCEP-1, and NCEP-2 averaged between 10° and 15°E for 1989–2003.

  • View in gallery

    Seasonal means of the circulation at 925 hPa for (left)–(right) ERA-Interim, MERRA, NCEP-1, and NCEP-2 for the period 1989–2003. Contours and shading indicate the zonal wind speed. The zero contour indicates the boundaries of westerlies (positive values on solid lines).

  • View in gallery

    As in Fig. 2, but for the rotational circulation. White boxed areas indicate grid points under the topography.

  • View in gallery

    As in Fig. 3, but for the divergent circulation.

  • View in gallery

    MAM and SON of the (top) mean zonal total, (middle) divergent, and (bottom) rotational circulation in longitude–pressure section, averaged between 0° and 10°S for (a) NCEP-1, (b) NCEP-2, (c) ERA-Interim, and (d) MERRA. The vertical velocity is significantly smaller than the zonal wind; thus, for clarity in the plot the vertical velocity has been scaled up by a factor of 600. Shaded areas indicate westerly winds, greater than 1 m s−1. White boxed areas east of 10°E below 850 hPa indicates grid points under the topography. Units for westerly winds are meters per second and for vertical velocity are 101 Pa min−1.

  • View in gallery

    Diabatic heating (contours; K day−1) and the divergent zonal circulation (vectors; m s−1) in (a) NCEP-1, (b) NCEP-2, (c) ERA-Interim, and (d) MERRA in MAM and SON in longitude–pressure section, averaged between 0° and 10°S. The vertical velocity is significantly smaller than the divergent zonal wind; thus, for clarity in the plot vertical velocity has been scaled up by a factor of 600. White boxed areas east of 10°E below 850 hPa indicate grid points below the topography.

  • View in gallery

    Distributions of MAM and SON minus MAM means of the vertical integrated diabatic heating (shading; K day−1) between 700 and 925 hPa and the divergent zonal circulation (vectors; m s−1) at 925 hPa in (a) NCEP-1, (b) NCEP-2, (c) ERA-Interim, and (d) MERRA. Wind vectors below the topography are masked.

  • View in gallery

    Regions where the vertical profiles of the diabatic heating and the moisture sink are examined.

  • View in gallery

    Seasonal mean vertical profiles of the diabatic heating (K day−1) and the moisture sink (K day−1) averaged over regions in Fig. 8.

  • View in gallery

    The 1989–2003 time series of annual mean total, divergent and rotational wind at 925 hPa averaged between 5°N and 10°S and from 9° to 15°E over WEA.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 116 116 20
PDF Downloads 78 78 19

Identification of Processes Driving Low-Level Westerlies in West Equatorial Africa

View More View Less
  • 1 Department of Physics, Higher Teacher Training College, University of Yaoundé 1, and Center for International Forestry Research, Central Africa Regional Office, Yaoundé, Cameroon
  • | 2 Met Office Hadley Centre, Exeter, United Kingdom
  • | 3 Center for International Forestry Research, Central Africa Regional Office, Yaoundé, Cameroon
  • | 4 University of Mountain, Bangangté, Cameroon
© Get Permissions
Full access

Abstract

This paper investigates and characterizes the control mechanisms of the low-level circulation over west equatorial Africa (WEA) using four reanalysis datasets. Emphasis is placed on the contribution of the divergent and rotational circulation to the total flow. Additional focus is made on analyzing the zonal wind component, in order to gain insight into the processes that control the variability of the low-level westerlies (LLW) in the region. The results suggest that the control mechanisms differ north and south of 6°N. In the north, the LLW are primarily a rotational flow forming part of the cyclonic circulation driven primarily by the heat low of the West African monsoon system. This northern branch of the LLW is well developed from June to August and disappears in December–February. South of 6°N, the seasonal variability of the LLW is controlled by the heating contrast between cooling associated with subsidence over the ocean and heating over land regions largely south of the equator, where ascent prevails. The heating contrasts lead to a Walker-type circulation with development of LLW as its lower branch. Thus, evidence is presented that the LLW are driven by differential heating. This contrasts with the traditional conceptual view that the Saint Helena high is the primary driver of low-level circulation off the Atlantic Ocean to WEA. Forest cover in WEA may modulate the latent heating that helps to drive the differential heating and maintain the LLW, and this interaction should be the focus of further study.

Denotes Open Access content.

Publisher’s Note: This article was revised on 7 August 2017 to correct the name of the first author, whose given name and surname was inadvertently reversed when originally published.

Corresponding author address: Wilfried M. Pokam, Department of Physics, Higher Teacher Training College, University of Yaoundé 1, P.O. Box 47, Yaoundé, Cameroon. E-mail: wpokam@yahoo.fr

Abstract

This paper investigates and characterizes the control mechanisms of the low-level circulation over west equatorial Africa (WEA) using four reanalysis datasets. Emphasis is placed on the contribution of the divergent and rotational circulation to the total flow. Additional focus is made on analyzing the zonal wind component, in order to gain insight into the processes that control the variability of the low-level westerlies (LLW) in the region. The results suggest that the control mechanisms differ north and south of 6°N. In the north, the LLW are primarily a rotational flow forming part of the cyclonic circulation driven primarily by the heat low of the West African monsoon system. This northern branch of the LLW is well developed from June to August and disappears in December–February. South of 6°N, the seasonal variability of the LLW is controlled by the heating contrast between cooling associated with subsidence over the ocean and heating over land regions largely south of the equator, where ascent prevails. The heating contrasts lead to a Walker-type circulation with development of LLW as its lower branch. Thus, evidence is presented that the LLW are driven by differential heating. This contrasts with the traditional conceptual view that the Saint Helena high is the primary driver of low-level circulation off the Atlantic Ocean to WEA. Forest cover in WEA may modulate the latent heating that helps to drive the differential heating and maintain the LLW, and this interaction should be the focus of further study.

Denotes Open Access content.

Publisher’s Note: This article was revised on 7 August 2017 to correct the name of the first author, whose given name and surname was inadvertently reversed when originally published.

Corresponding author address: Wilfried M. Pokam, Department of Physics, Higher Teacher Training College, University of Yaoundé 1, P.O. Box 47, Yaoundé, Cameroon. E-mail: wpokam@yahoo.fr

1. Introduction

The economy of west equatorial Africa (WEA) (10°S–10°N, 9°–30°E) is dominated by natural resources and agriculture and is highly dependent on climate. Over WEA, the climate is strongly influenced by changes in low-level moisture advection throughout the year (McCollum et al. 2000; Matsuyama et al. 1994). This low-level moisture advection is dominated by moist air from the Atlantic Ocean. The strongest moisture inflow is registered during the second and main rainy season from September to November, when the highest amount of rainfall is recorded (Pokam et al. 2012). Pokam et al. (2012) found that at interannual time scales, for both year-to-year comparisons and wet minus dry composites, low-level moisture flux from the Atlantic Ocean controls the moisture content of the entire atmospheric column. Changes in associated low-level westerlies control the interannual variability of rainfall in the coastal region (Nicholson and Dezfuli 2013). Because of the important role of the water cycle in climate variability and change (Burde et al. 1996) and the heavy dependence of the economy and livelihood of the region on water cycle (Molua and Lambi 2007), it is important to explore the low-level circulation driving this moisture advection from the Atlantic Ocean to WEA. This is essential for advancing the physical understanding and modeling of climate in the region. It is also important for exploring the reasons behind the disagreement between climate model responses to expected future climate change (James et al. 2013), as Washington et al. (2013) have shown that moisture flux is a useful quantity to understand model rainfall biases over WEA.

The atmospheric circulation over WEA has been described in several previous studies. Broadly, the atmospheric circulation is dominated by a large seasonal shift in the position of the intertropical convergence zone (ITCZ), which is determined by the northeast and southwest trade winds, and the monsoon circulation from the Atlantic (Fontan et al. 1992). The regional upper-level dynamics are influenced by the high pressure cells over the Sahara and the south of Africa that drive high-altitude easterlies (Fontan et al. 1992). Nicholson and Grist (2003) found that at midlevel (around 600–700 hPa), the annual cycle of easterlies are dominated by the north component of the African easterly jet (AEJ-N) and the south component of the African easterly jet (AEJ-S).

In the lower troposphere, moist air from the Atlantic Ocean (Fontan et al. 1992), known as low-level equatorial westerlies (Nicholson and Grist 2003), are associated with the southeasterly trades on the northeastern flank of the Saint Helena (South Atlantic) high. Because of Coriolis forces, the southeasterlies become westerlies when crossing the equator. The low-level westerlies (LLW) are defined throughout the year and are well developed from July to September (Nicholson and Grist 2003). Using soundings, the atmospheric circulation was described over the west coast of WEA, including one site in the Northern Hemisphere (Douala: 4.3°N, 9.42°E) (Fontaine and Janicot 1992), one near the equator (Libreville: 0.23°N, 9.27°E), and one in the Southern Hemisphere (Luanda: 8.48°N, 13.14°E) (Zhang et al. 2006). It appears that the annual cycle of LLW varies from the southern to the Northern Hemisphere. At the northern (southern) site, LLW are well developed from July to September (October–February) and are the strongest and the deepest in August (December–January). The upper limit of westerlies migrates from around 950 hPa in July to 700 hPa in January in the southern site (Zhang et al. 2006). At the northern site, this limit moves upward from 925 hPa in April to 750 hPa in August (Fontaine and Janicot 1992). In the vicinity of the equator, LLW are defined from August to January. The rest of the year, westerlies disappear and easterlies dominate throughout the entire atmospheric column (Zhang et al. 2006).

There are well-documented examples of regions of concentrated low-level westerly flow in West Africa (Grodsky et al. 2003; Pu and Cook 2010). However, there are few documented examples in WEA. This study focuses on the investigation of mechanisms governing LLW over WEA. Very little is known about the processes that control the height of these LLW and the seasonal and interannual variability of their strength. The purpose of this study is to investigate the processes that control the depth, the intensity, and the seasonal and interannual variability of LLW using reanalysis data. We will use the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis (NCEP-1; Kalnay et al. 1996), the NCEP–U.S. Department of Energy (DOE) Atmospheric Model Intercomparison Project phase 2 (AMIP-II) reanalysis (NCEP-2; Kanamitsu et al. 2002), the European Centre for Medium-Range Weather Forecasts (ECMWF) Interim Re-Analysis (ERA-Interim; Dee et al. 2011), and the Modern-Era Retrospective Analysis for Research and Applications (MERRA; Rienecker et al. 2011). Over WEA, the density of the observation/meteorological station network is low and observations are sparse and sometimes unavailable (Aguilar et al. 2009). In the region, the reanalysis data rely strongly on the physical parameterizations in the global models used to create the reanalyses. Therefore, the reanalyses may differ because of different analysis systems and different model physics. Other discrepancies may arise from the difference in spatial resolution between the reanalyses or the number of observations used. Some aspects of the atmospheric dynamics may be more visible in ERA-Interim and MERRA because of their finescale resolution (see section 2a) compared to the coarser resolution of NCEP-1 and NCEP-2. Use of more than one reanalysis dataset reduces the susceptibility of results to errors in the underlying model used, and consistency between the reanalyses is an indicator (though not a guarantee) of robustness.

The objective of this study is to identify common features between the reanalyses in the representation of the mean characteristics of the low-level circulation over WEA, provide detailed information about the drivers of the low-level westerlies, and dissect the contributions of rotational and divergent flow. The related driving processes are investigated at both seasonal and interannual time scales. Focus is made on the drivers of the low-level zonal flow from the Atlantic Ocean to WEA. This study also revises the conventional view that LLW are driven solely by the anticyclone in the southeastern Atlantic Ocean. The paper is organized as follows: After presenting the data and methods used in section 2, section 3 presents the structure of the low-level circulation over WEA as seen from the four reanalyses. It shows the predominance of the divergent circulation on LLW. In the tropics, the divergent circulation is significantly sensitive to the diabatic heating field, particularly that associated with moist processes (Annamalai et al. 1999). The main analysis of the three-dimensional distribution of the diabatic heating is described in section 4 and is related to the structure of the divergent winds. Section 5 deals with the evaluation of interannual variability of the LLW and the drivers. Section 6 presents the summary and conclusions.

2. Data and methodology

a. Datasets

We used 15 yr of monthly-mean fields (January 1989–December 2003) from NCEP-1, NCEP-2, ERA-Interim, and MERRA. The period of study used for our investigation starts from January 1989 so that all the reanalyses products overlap with the ERA-Interim data.

The selected variables are the meridional, zonal, and vertical winds; air temperature; and specific humidity. In NCEP-1, these variables are on 17 pressure levels from 1000 to 10 hPa on a 2.5° × 2.5° spatial grid. The assimilation of data in the dynamical atmospheric model occurs through the use of three-dimensional variation data assimilation (3DVar). The 3DVar assimilates all observations that occur within a specific period at a single time. All variables are available every 6 h from 1948 to the present. NCEP-2 is an updated version of NCEP-1. NCEP-2 uses the same spatial and temporal resolution as NCEP-1. Many of the known errors in NCEP-1 are improved in NCEP-2, which uses an improved forecast model and data assimilation system. The NCEP-2 products are available from 1979 to the present. MERRA data span from 1979 to the present, and variables are available at 72 vertical levels. MERRA has a horizontal resolution of 0.5° × 0.66°. As with NCEP-1, a 3DVar approach is used to assimilate the observations into the dynamic atmospheric model. Much of the model output is archived at an hourly time scale. The ERA-Interim product is the most recent ECMWF reanalyses dataset and is available from 1989 onward. An updated version of the ECMWF forecast is used at 1.5° horizontal resolution. The ERA-Interim incorporates four-dimensional variational data assimilation (4DVar), which is a temporal extension of the 3DVar.

b. Wind decomposition and heating calculations

In section 3 we explore the processes that control the variability of LLW. Focus is made on the contribution of the divergent and the rotational (nondivergent) circulation to the total flow. To achieve this, the horizontal wind fields were partitioned into the rotational and divergent circulations. Through such an approach, predominance of a circulation type may be analyzed, allowing the identification of the driving processes across the domain. For example, the field properties of the divergent circulation will be investigated through the contribution of sensible heat flux, latent heating, and radiative cooling (e.g., as in Hagos and Zhang 2010).

According to the Helmholtz theorem, the horizontal velocity vector is partitioned into divergent and rotational components (Li et al. 2006), Vχ and Vψ, respectively,
e1
where
e2
where ψ is the streamfunction, χ is the velocity potential, is the horizontal gradient vector, and k is the unit vector in the vertical direction. Substituting (2) into (1) and applying the divergence yield,
e3
where D is the horizontal velocity divergence. Given the V(u, υ) field, we can derive the divergence D and solve the Poisson equation (3) to determine the velocity potential χ. We may then derive the divergent wind using (2) and the rotational wind as the difference between the total low and the zonal flow.
The diabatic heating Q1 (Yanai and Tomita 1998) is calculated as a residual of the thermodynamic equation and computed from
e4
and the moisture sink Q2 is computed from the equation of moisture continuity,
e5
where T is the air temperature (K) and
eq1
is the potential temperature; q is the mixing ratio of water vapor (kg kg−1), u and υ are the zonal and meridional wind components (m s−1), ω is the vertical pressure velocity (Pa s−1), P is pressure, P0 = 1000 hPa, R and Cp are the gas constant and specific heat at constant pressure of dry air, and L is the latent heat of condensation. Yanai and Tomita (1998) defined the term on the right-hand side of (4) and (5) as the apparent heating and apparent moisture sink, respectively. The diagnosed monthly rate of heating q1 (K day−1) due to apparent heating is then computed by
e6
and the rate of equivalent heating q2 (K day−1) due to the apparent moisture sink is computed by
e7
The apparent heat source Q1 consists of heating resulting from radiation, the release of latent heat resulting from condensation, and the vertical convergence of the vertical eddy transport of sensible heat. The term Q2 represents the apparent moisture sink resulting from the net condensation and the vertical divergence of the vertical eddy transport of moisture. The vertical profiles of Q1 and Q2 serve to show the occurrence of eddy vertical transport processes and in turn indicate the strength of the activity of cumulus convection (Yanai et al. 1973). In the case of strong vertical motion, leading to a strong contribution of eddy transport, the vertical profile of Q1 will differ from Q2. The levels of peak Q1 and Q2 are separated (Wada 1969), and the vertical eddy transports of heat and moisture are associated with cumulus convection and turbulent motion. The greater the difference between the height of these peaks, the stronger the convective activity (Wada 1969). The high convective activity over WEA (Jackson et al. 2009) may lead to significant contributions of the eddy vertical flux terms to Q1 and Q2.

3. The seasonal mean cycle

a. Total circulation

The mean zonal winds averaged along the coastal region, between 10° and 15°E, from all the reanalyses are shown in Fig. 1. As shown later, the core speed of the LLW in WEA is located over this area all year. During all seasons, the core speed of LLW is located around 925 hPa. LLW are slightly shallower during the rainy seasons March–May (MAM) and September–November (SON). The upper boundary of LLW is around 850 hPa during SON. It crosses above this level during December–February (DJF) and June–August (JJA). During MAM, the westerlies are weak with a mean core speed around 1 m s−1. The westerlies strengthen and extend northward during JJA. The mean core speeds reach 4 m s−1 north of 8°N and range from 2 to 3 m s−1 around the equator. During SON, the zonal winds weaken significantly north of 6°N. They strengthen around the equator and are more developed in the Southern Hemisphere in this period. The basic features of the LLW, as described above, can be seen in the mean seasonal zonal circulation from all the reanalyses. Some differences appear between the reanalyses. During MAM, one notable feature in the structure of the LLW is the two distinct cells north and south of 6°N in ERA-Interim and MERRA (first two columns of Fig. 1).

Fig. 1.
Fig. 1.

Seasonal means of the zonal wind (m s−1) for (left)–(right) ERA-Interim, MERRA, NCEP-1, and NCEP-2 averaged between 10° and 15°E for 1989–2003.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

The northern cell strengthens during JJA, with a northward extension. Its northern boundary moves to the north from 11°N during MAM and lies between 16° and 18°N in JJA. This northern cell weakens/disappears during SON and DJF. It is important to note that the seasonal cycle of this northern cell is similar to the climatological annual cycle of the West Africa monsoon (WAM) system (Fontaine et al. 2002; Thorncroft et al. 2011), whereas the seasonal cycle of the southern cell is distinct and thus likely driven by other mechanisms.

In contrast to the northern cell, the equatorial cell does not show latitudinal migration across the seasons. Its boundaries are located at 6°N in the north and between 16° and 18°S in the south all year. The differences in the annual cycle of westerlies north and south of 6°N suggest that the control mechanism in the north may differ from that in the south. This duel-cell feature is not captured by either NCEP-1 or NCEP-2. In NCEP-1 and NCEP-2, LLW appear as a single feature, with a northward migration from 7°N in DJF to a mean position of around 17°N during JJA (Fig. 1). The southern boundary also migrates during the year. NCEP-1 and NCEP-2 fail to depict the two cells potentially because of their coarse resolution compared to ERA-Interim and MERRA. During MAM, to the north of the Cameroon highlands, there is a local maximum of westerlies in ERA-Interim and MERRA. In NCEP-1 and NCEP-2, this local maximum is overlapped by the topography because of the coarse resolution. Despite this, the main characteristics of LLW are well represented by all the reanalyses, though the finescale of ERA-Interim and MERRA captures more dynamical details than NCEP-1 or NCEP-2.

To summarize, we suggest that the LLW consist of two cells, located north and south of 6°N, respectively. The cells differ in their annual cycles. The northern cell peaks during JJA, whereas the equatorial cell peaks in SON.

b. Rotational and divergent circulation

The difference in the phases of the two cells of LLW suggests that they may be controlled by different mechanisms, which are analyzed now.

1) North of 6°N

Figure 2 represents the mean seasonal circulation pattern at 925 hPa. The zonal wind in contours enables us to follow the seasonal variability of LLW. As shown before, the core peak in speed of LLW is around 925 hPa all year. During DJF, north of 6°N, the zonal wind is dominated by northeasterlies. During MAM there is a northward shift of westerlies, from the Gulf of Guinea to around 10°N. Westerlies are more defined over West Africa. This is well represented in all reanalyses. In the north of WEA (north of 5°N, between 9° and 30°E), westerlies appear in NCEP-1 and NCEP-2. In ERA-Interim and MERRA, the westerlies are limited east of 15°E. During JJA, westerlies strengthen over West Africa with a strong latitudinal extension and move eastward to the Ethiopian highlands (Fig. 2), reinforcing LLW in the north of WEA. This feature indicates that the northern cell of LLW in WEA is related to the development of the zonal circulation in the WAM. The dynamics associated with the WAM was fully described by Sultan and Janicot (2003). They state that, at the beginning of June, relative vorticity centers develop along the intertropical front (ITF), between 15° and 20°N, which is concomitant with the position of the Saharan heat low. By the end of June, the ITF moves northward, associated with the increase of the relative vorticity. This leads to the development of a cyclonic circulation in northern West Africa. The cyclonic circulation contributes to the strength of the zonal circulation south of 15°N. Around mid-July, the relative vorticity associated with the Saharan heat low is at its strongest and is located in the western part of the Northern Africa. During the same period, there is an advection of absolute vorticity from the Gulf of Guinea to the coastal region in West Africa where absolute vorticity is equal to zero (Tomas and Webster 1997). This leads to the development of zonal wind shear around the zero contour in absolute vorticity with strong low-level westerlies over West Africa driven by the rotational flow. This enhancement of the heat low rotational circulation coincides with a peak of the monsoon zonal winds over a large part of West Africa.

Fig. 2.
Fig. 2.

Seasonal means of the circulation at 925 hPa for (left)–(right) ERA-Interim, MERRA, NCEP-1, and NCEP-2 for the period 1989–2003. Contours and shading indicate the zonal wind speed. The zero contour indicates the boundaries of westerlies (positive values on solid lines).

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

The mean seasonal flow in Fig. 2 has been split into its rotational and divergent components using the method described in section 2, and they are shown in Figs. 3 and 4. The figure clearly shows that the strong zonal circulation with higher westerlies (Fig. 2) is part of the rotational circulation of the monsoon system (Fig. 3). The dominance of westerlies in the rotational circulation component appears during MAM (Fig. 3), in the early phase of the WAM. After the monsoon onset at end of June, the zonal circulation in the WAM system intensifies and is characterized by the northern extension of the westerlies (Fig. 3), ultimately controlled by the location of the Saharan heat low (Sultan and Janicot 2003).

Fig. 3.
Fig. 3.

As in Fig. 2, but for the rotational circulation. White boxed areas indicate grid points under the topography.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

Fig. 4.
Fig. 4.

As in Fig. 3, but for the divergent circulation.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

This monsoon enhancement is associated with the strengthening of westerlies in the northern WEA, primarily driven by the rotational circulation. The northern cell of LLW peaks during JJA (Fig. 3). Westerlies in the northern WEA weaken significantly in the retreat phase of the monsoon. Over West Africa, the westerlies are localized along the coastal region during SON (Fig. 3). During DJF, westerlies disappear north of 5°N, where the circulation becomes dominated by easterlies driven by the rotational winds. This feature is well captured by all the reanalyses. However, the rotational wind is slightly stronger in the ERA-Interim and NCEP-1.

2) South of 6°N

Figure 2 shows that over the Atlantic Ocean, the circulation is dominated by strong southeasterlies in all seasons. The southeasterlies are driven by the rotational circulation, which weakens from the ocean to the land (Fig. 3). The southeast–northwest orientation of this rotational flow suggests that it is related to the low-level Benguela jet (Nicholson 2010). The low-level jet (LLJ) is located on the northeastern flank of the South Atlantic high. The LLJ is primarily driven by large-scale geostrophic balance resulting from the surface pressure gradient over the southern Atlantic Ocean. From July to October, the increases in thermal contrast between land and ocean lead to the development of thermal winds, which superimposed upon the geostrophic flow strengthen the LLJ. The LLJ is well developed during this period (Nicholson 2010) and in turn acts to strengthen the rotational circulation (Fig. 3).

South of 6°N over WEA, the contribution of the rotational wind to the LLW is quite weak in the reanalyses. It is nonexistent in MERRA. Only NCEP-1 shows any substantial contribution, with core speed barely reaching 1 m s−1 in the vicinity of the equator. However, the zonal component of the divergent circulation consists primarily of westerlies (Fig. 4). Broadly, the divergent circulation strengthens from the southeast Atlantic to the continent. During MAM, the westerly component of the divergent winds weakens, with core speeds around 1 m s−1 located along the coast over the ocean sector. This feature strengthens during JJA and the area of maximum zonal wind extends over the continent, east of 20°E (Fig. 4). The core speed reaches 4 m s−1. In SON, the patterns are more like those in JJA than in the other seasons but with a slight decrease of the zonal component over the continent. The comparison of Fig. 2 with Figs. 3 and 4 reveals that, although the rotational flow dominates the total circulation over the ocean, LLW at 925 hPa, from the coast to the WEA, are driven by the divergent circulation in all seasons. The divergent zonal wind speed is about the same magnitude in all the reanalyses lending confidence to the result. However, the area of maximum winds is more longitudinally extended in NCEP-1 and NCEP-2 than in ERA-Interim and MERRA.

3) Vertical composition of the circulation

The vertical cross section of the zonal circulation is now investigated to identify the dominant flow (divergent or rotational) through the depth of westerlies. Figure 5 represents the longitude–height cross section of the mean seasonal zonal component of the winds.

Fig. 5.
Fig. 5.

MAM and SON of the (top) mean zonal total, (middle) divergent, and (bottom) rotational circulation in longitude–pressure section, averaged between 0° and 10°S for (a) NCEP-1, (b) NCEP-2, (c) ERA-Interim, and (d) MERRA. The vertical velocity is significantly smaller than the zonal wind; thus, for clarity in the plot the vertical velocity has been scaled up by a factor of 600. Shaded areas indicate westerly winds, greater than 1 m s−1. White boxed areas east of 10°E below 850 hPa indicates grid points under the topography. Units for westerly winds are meters per second and for vertical velocity are 101 Pa min−1.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

The total, divergent, and rotational circulations are averaged between 0° and 10°S. The profiles during MAM and SON represent how these seasons correspond to weak and deep westerlies, respectively, in the Southern Hemisphere over WEA. The differences between MAM and SON in strength, depth, and eastward excursion of the total flow clearly appears in all the reanalyses. In MAM, LLW are defined around 10°E and up to 925 hPa, with core speeds less than 1 m s−1.

During SON, LLW are well developed over 850 hPa with core speeds greater than 2 m s−1. Figure 5 clearly shows that the development of the LLW during SON is primarily due to the reinforcing of the divergent circulation. This is consistent with the predominance of the easterly flow in the rotational circulation throughout the tropospheric column. In all reanalyses, the divergent component of the zonal wind is far greater than the rotational component (Fig. 5, second row of each reanalysis) in the area of LLW. The divergent circulation drives a Walker-type circulation, characterized by the ascending air over the land and downward motion over the ocean. The related strengthening of low-level descent over the ocean and the divergence below 850 hPa leads to stronger westerlies around 925 hPa. In MAM, the Walker cell is weak. This leads to weak and shallow LLW. The divergent circulation in SON is associated with a Walker-type circulation. The related strengthening of the low-level branch of this circulation leads to strong and deep LLW (particularly visible in MERRA). All the reanalysis agree well on the control of the divergent circulation on the depth of the LLW.

To summarize, over WEA two cells of LLW are defined north and south of 6°N. The northern cell is well developed during JJA and is related to the development of the zonal circulation driven by the heat low of the WAM system. This cell weakens during the other seasons and disappears in DJF. The southern equatorial cell is driven by the divergent circulation through a Walker-type circulation with ascent (descent) over the continent (ocean). The divergent flow is well developed during JJA and is maintained in SON. The equatorial cell weakens in MAM.

4. Drivers of divergent circulation

In the previous section, we found that, over WEA south of 6°N, LLW are driven by the divergent circulation. The divergent circulation is modulated by the distribution of the diabatic heating (Johnson et al. 1985). In regions of cooling, which are an energy sink, the vertical energy flux is downward. In regions of heating, which are an energy source, the vertical energy flux is upward. Thus, in upper (lower) layers in heat source regions, the horizontal energy flux is divergent (convergent). However, in heat sink regions, the horizontal transport of energy is convergent (divergent) in the upper (lower) layers (Johnson et al. 1985). This leads to a low-level circulation from cooling to heating regions. These basic principles underline the direct link between the divergent circulation and the differential heating. Herein the structure of the diabatic heating in relation to the seasonal variation of the divergent circulation is investigated. Hereafter, the diabatic heating and the moisture sink refer to q1 and q2 as defined in section 2.

a. Diabatic heating and divergent flow

Figure 6 represents the longitude–pressure section of the diabatic heating averaged from the equator to 10°S. During MAM, the major heat source is found over the continent, between 10° and 30°E. Heating of 1–2 K day−1 occurs from 700 hPa upward. A low-level maximum in the heating appears to the east of WEA, between 30° and 40°E, with heating rates exceeding 2 K day−1. Below the midlevel heating over WEA, a weak cooling lower than 1 K day−1 is discernable. Cooling generally prevails over the ocean upward of 925 hPa. A well-defined cooling maximum with rate exceeding 1 K day−1 is located around 850 hPa. The cooling over the ocean is about the same magnitude in all the reanalyses. However, the heating over the continent is doubled in ERA-Interim and MERRA compared to NCEP-1 and NCEP-2. In SON, the features are like in MAM but much more pronounced. The contrast between large positive values of heating rate over the continent and the negative values over the ocean is reinforced (Fig. 6). The values of q1 are at least twice those of MAM, except in MERRA over the land. The strong low-level heating east of 30°E expands upward to the deep upper-level heating over WEA, with the peak within the layer of 500–600 hPa (Fig. 6). A heating maximum greater than 3 K day−1 occurs over the coast, east of 10°E, from the surface up to 850 hPa. This increases the low-level contrast of heating between the ocean and the continent. This is consistent with the reinforcement of the low-level peak of cooling over the ocean, with rates exceeding 2 K day−1 (Fig. 6).

Fig. 6.
Fig. 6.

Diabatic heating (contours; K day−1) and the divergent zonal circulation (vectors; m s−1) in (a) NCEP-1, (b) NCEP-2, (c) ERA-Interim, and (d) MERRA in MAM and SON in longitude–pressure section, averaged between 0° and 10°S. The vertical velocity is significantly smaller than the divergent zonal wind; thus, for clarity in the plot vertical velocity has been scaled up by a factor of 600. White boxed areas east of 10°E below 850 hPa indicate grid points below the topography.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

Also represented in Fig. 6 is the vertical section of the divergent circulation (as in the middle panels of Fig. 6). The vertical structure of the heating is associated with a Walker-type circulation with rising air over the warm continent and descent over the cool ocean. The stronger divergent circulation during SON is consistent with the greater heating (cooling) contrast over the continent (ocean). The maximum in heating over the coast may contribute to the local acceleration of the westerlies. Based on the previously mentioned link between horizontal divergent circulation and vertical distribution of the diabatic heating, we argue that the low-level peak of cooling over the ocean and the heating in the coastal area make key contributions to the strength of LLW. This increase in the strength of the LLW can feedback on the strength of the convection and diabatic heating over WEA.

To illustrate the relation between low-level contrast in the heating and the strength of LLW, Fig. 7 presents the patterns of seasonal mean rate of diabatic heating between 700 and 925 hPa and divergent circulation at 925 hPa. To represent the low-level cooling, the layer of 700–925 hPa is chosen as the cooling centers and peaks within this layer. One striking feature of the pattern in MAM is the location of a major source of heating over the north of Africa, spanning from the WAM region eastward to the Ethiopian highlands. Dry air and dry soil prevail over the Saharan region in this period. The heating shown in Fig. 7 reflects the predominance of sensible heat flux from the surface (Fontaine et al. 2002). Dry convection associated with the cyclonic region of the heat low occurs, with the peak of upward motion at 850 hPa (Sultan and Janicot 2003). This feature in combination with the subsidence over the ocean and the associated lower branch of the overturning (Fig. 6) leads to dominant meridional flow in the divergent circulation (Fig. 7). LLW weaken during this season (Figs. 4, 5). In SON, the heating decreases significantly over West Africa and is shifted south to the Gulf of Guinea. The major source of heating occurs south of 5°N over the continent with a maximum along the coast in WEA and over East Africa. Over the ocean, the cooling strengthens and extends northward to the Gulf of Guinea. The divergent circulation at 925 hPa is dominated by southwesterlies, which reinforces the zonal circulation with stronger acceleration over the coast along 12°E (Figs. 5, 7). Thus, the development of the LLW over WEA during SON is controlled by the heating contrast between the cooling over the ocean and the heating over the WEA, which strengthens a Walker-type circulation with ascent over the continent and descent over the ocean driven by the divergent wind. The related strengthening of low-level descent at 850 hPa and the lower branch of the overturning, seen previously in Fig. 6, leads to a stronger LLW at 925 hPa with the peak over the coastal region.

Fig. 7.
Fig. 7.

Distributions of MAM and SON minus MAM means of the vertical integrated diabatic heating (shading; K day−1) between 700 and 925 hPa and the divergent zonal circulation (vectors; m s−1) at 925 hPa in (a) NCEP-1, (b) NCEP-2, (c) ERA-Interim, and (d) MERRA. Wind vectors below the topography are masked.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

There are two dry seasons, DJF and JJA. During DJF, the vertical heating profile is the same as in SON but is slightly weaker (not shown). However, during JJA, when driest conditions prevail over WEA, the cooling over the Atlantic Ocean strengthens significantly, whereas heating over WEA weakens.

From the above analysis of Figs. 6 and 7, four key regions are identified as the main contributors to the development of the LLW. The regions consist of the cool region over the southeast Atlantic Ocean (Fig. 8: region A), the low-level heating over the coast (region B), the upper-level heating over WEA (region C), and the low-level heating over East Africa (region D). Region C covers the center of the Congo Basin, corresponding to the humid dense forest area of southeast Cameroon, southern Central African Republic, east of the Congo Republic, and a major part of Democratic Republic of Congo. Region B covers southern Cameroon to the northwest of Angola and covers Gabon, Equatorial Guinea, and the southeast of the Congo Republic. The regions act as follows: regions A and B modulate the lower branch of the overturning and the acceleration of LLW over the coast, while heating in region D causes wind convergence over region C, enhancing convective heating, which in turn controls the ascending branch of the Walker-type circulation.

Fig. 8.
Fig. 8.

Regions where the vertical profiles of the diabatic heating and the moisture sink are examined.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

b. Principal factors impacting heating

To identify the principal factors contributing to the cooling or heating over the four regions identified in Fig. 8, the vertical profiles of q1 and q2, defined in section 2, are analyzed. In Fig. 9 the profiles for each region between MAM and SON are shown. Although in Fig. 9 the mean vertical profiles of q1 and q2 from the surface to 300 hPa are shown, a specific layer is of interest for each location. In regions A, B, and D the lower-tropospheric heating or cooling appears to be more important, whereas in the region C it is in the upper troposphere. It is important to note that over region A (region B) only sea (land) grid points are taken into account during the computation of the vertical profiles.

Fig. 9.
Fig. 9.

Seasonal mean vertical profiles of the diabatic heating (K day−1) and the moisture sink (K day−1) averaged over regions in Fig. 8.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

The vertical distribution of q1 over the southeast Atlantic Ocean (region A; Fig. 9a) during MAM is characterized by cooling throughout the troposphere, except the layer below 925 hPa. The q2 profile indicates that moisture is being evaporated at the sea surface and then mixed through the boundary layer by turbulent processes. The peak of the apparent moisture source at 925 hPa indicates the peak of where the turbulent transports deposits moisture from the surface. This process contributes to the prevalence of low-level stratocumulus cloud cover (Rozendaal et al. 1995), with deep stratocumulus-topped boundary layer (Wood 2012). The turbulent moisture transport does not lead to a change of phase; therefore, q1 changes, except at the top of the boundary layer, where clouds form. However, this cloud formation by condensation would actually lead to a positive heating, so this would not explain the increased cooling at this level. Instead, the enhanced radiative cooling from the top of boundary layer clouds (Wood 2012; Slingo et al. 1982) leads to the peak of cooling around 850 hPa (Fig. 9a). The heating below 925 hPa is a result of vertical convergence of sensible heat flux from the surface. During SON, the negative values of q1 are reinforced. However, there is no significant change in the values of q2. This feature suggests that the longwave cooling at the top of the stratocumulus modulates the seasonal variability of the low-level cooling over the ocean.

Along the coast in WEA (region B), our main interest is the heat source below 850 hPa (Fig. 6). Figure 9b shows the mean vertical profiles of q1 and q2 over region B. During MAM, q1 is positive at low levels with a peak at 925 hPa; q2 values are also positive. This indicates the predominance of sensible heat flux and the release of latent heat by condensation in the lower troposphere. During SON, the decrease of q2 values below 850 hPa probably reflects high evaporative cooling from the wet soil resulting from an increase in rainfall over the coast (Jackson et al. 2009). However, q1 exhibits an increase of heating of around 2.5 K day−1 at the surface. This indicates that the increase in sensible heating is larger than the heat loss from increased evaporation.

The vertical profile inland over WEA (region C of Fig. 9c) during MAM is characterized by positive values of q1 and q2 throughout the troposphere, except below 850 hPa, where q2 is negative. Above 850 hPa, q1 and q2 values are positive, with the vertical separation of the peaks of q1 (500–600 hPa) and q2 (700 hPa) characterizing the presence of cumulus-convective vertical transport (Yanai et al. 1973). During SON, large positive values of q1 and q2 occur at upper levels, indicating the enhancement of the condensation process that leads to an increase in the release of latent heat. This is consistent with a strengthening in convective activity, which is most developed during this period (Jackson et al. 2009). All the reanalyses agree well on the predominance of the release of latent heat in the upper levels and the associated deeper moist convection in SON.

The profile in Fig. 9d is located over East Africa. During MAM, q1 is strongly positive in the lower-tropospheric layer and decreases with height, becoming negative above 700 hPa. The q2 values are also positive at low levels and become negative in the layer of 700–850 hPa. During SON, large positive values of q1 and q2 occur below 700 hPa. These profiles are indicative of lower-level sensible heat and latent heat release by condensation that strengthens in SON.

5. Interannual variability

In addition to the long-term seasonal mean circulation, the year-to-year variability of the contribution of divergent and rotational circulation to LLW is investigated. The patterns of standard deviation of the monthly-mean total, divergent, and rotational zonal circulation for DJF, MAM, JJA, and SON (not shown) show that there is weak interannual variability of the divergent circulation. The year-to-year variability of LLW appears to be driven by the interannual variability of the rotational circulation over both the northern and the southern part of WEA. This result is consistent among the reanalyses. As shown in the previous section, south of 6°N, the seasonal mean variability of the LLW is driven by the divergent circulation and the rotational wind is dominated by easterlies. Hence, the predominant impact of the rotational wind on LLW interannual variability is a dampening effect on the strength of the equatorial cell of LLW. This is well illustrated for MAM and SON in Fig. 10. For all reanalyses, although the divergent circulation dominates the strength of LLW, its variability is weak compared to the variability of rotational flow. As a consequence, the trend of LLW shows an increase (decrease) during years of weak (strong) rotational flow.

Fig. 10.
Fig. 10.

The 1989–2003 time series of annual mean total, divergent and rotational wind at 925 hPa averaged between 5°N and 10°S and from 9° to 15°E over WEA.

Citation: Journal of Climate 27, 11; 10.1175/JCLI-D-13-00490.1

During MAM and SON, the rotational wind over the south of WEA originates from East Africa (Fig. 3). This flow originates from the southwestern Indian Ocean and turns into northeasterlies south of the Ethiopian high. This is well illustrated in ERA-Interim and MERRA. This rotational flow may originate from the Mascarene high over the southwestern Indian Ocean (Findlater 1969). This suggests that there may be a teleconnection between LLW and Indian Ocean at interannual time scales.

6. Summary and conclusions

The seasonal means and the interannual variability of the low-level circulation over west equatorial Africa have been described in some detail using NCEP-1, NCEP-2, ERA-Interim, and MERRA. For the characterization of the processes that control the low-level circulation, a focus is made on the contribution of divergent and nondivergent circulation to the total flow.

The increased resolutions of ERA-Interim and MERRA provide a clear depiction of two distinct cells of LLW north and south of 6°N. The northern cell is related to the development of the zonal circulation driven by the heat low within the WAM system. It is well developed during JJA. South of 6°N, over the Atlantic Ocean, the circulation is dominated by strong southeasterly rotational wind in all seasons. It is related to the Benguela jet, which is primary driven by the surface pressure gradient associated with the South Atlantic high. From July to October, the increased thermal contrast between ocean and land superimposes thermal winds upon this system. The rotational circulation is the strongest during this period. Further analysis shows that, during the small (March–May) and the big (September–November) rainy seasons, the divergent circulation drives a Walker-type circulation, with ascent over the continent due to the diabatic heating. During the big rainy season, the zonal heating contrast between the diabatic cooling over the ocean and the diabatic heating over the continent intensifies. The maritime subsidence is reinforced, and the related strengthening of low-level descent around 850 hPa leads to deep and strong LLW that peak at 925 hPa. There is a good agreement between reanalysis on this increase of heating contrast and the related strengthening of LLW. These differences in the drivers of LLW north and south of 6°N are consistent with the idea that the meteorological processes controlling each region differ. The region north of 6°N is controlled by the same processes that influence Sahelian West Africa (Nicholson and Grist 2003).

Our results on the LLW over WEA contradict the conventional view that the low-level circulation from the Atlantic Ocean is associated exclusively with the South Atlantic high (Fontan et al. 1992; Nicholson and Grist 2003). This study has demonstrated that the low-level winds over the southeast Atlantic Ocean, driven by the high, remain southeasterly all year. The wind turns into southwesterlies near the coast, driven by the heating contrast between land and ocean. The seasonal variability of the strength and height of LLW is controlled by the seasonal evolution of this heating contrast.

Our hypothesis on the drivers of the LLW south of 6°N can be described as follows: During MAM, when the LLW are the weakest, the major source of heat source (sink) is located over the continent (ocean). Note that at this time low-level stratocumulus cloud is the prevalent cloud type over the southeast Atlantic Ocean. The cooling associated with the heat sink over the ocean peaks at 850 hPa, because of the enhanced radiative cooling at the top of boundary layer clouds. Over WEA, diabatic heating prevails, associated with the release of latent heat by condensation. Low-level heating over East Africa, driven by the release of latent heat by condensation and surface sensible heat, contributes to the atmospheric heating over WEA. This land–sea contrast in heating drives a Walker-type circulation with ascent over WEA and descent over the ocean. The strengthening of the descent at 850 hPa, as well as the consequent horizontal low-level flow below, leads to well-developed LLW at 925 hPa. During SON, the cooling over the Atlantic is reinforced. A low-level peak of heating appears over the coastal region in WEA as a consequence of the enhanced release of both latent heat by condensation and surface sensible heat. The low-level heating over East Africa also strengthens. This contributes to the strong heating throughout the depth of the troposphere in WEA. This reflects the strong influence of latent heat release by condensation within the deep moist convection. The strengthening of the heating contrast between the land and the ocean reinforces the Walker cell and in turn the LLW.

The LLW have been shown to have a large impact on moisture inflow and precipitation variability over WEA (Nicholson and Dezfuli 2013). The LLW enhance the zonal circulation and therefore promote ascent through the orographic lifting effect of the highlands, thus leading to the enhancement of rainfall in the highland region (Vondou et al. 2010). In addition to this effect, the strengthening of LLW in MAM intensifies the convergence associated with the ITCZ and promotes its northward excursion. This acts to link ascent associated with the ITCZ and the rain belt, which in turn contributes to the intensification of rainfall (Nicholson and Dezfuli 2013).

In WEA, 45.5% of the area is covered by the forest (de Wasseige et al. 2009, 2012), which maintains higher evaporation rates than other cover types, including open water (Sheil and Murdiyarso 2009). This suggests that forest cover may significantly contribute to the condensed water over region C and in turn to LLW.

It is important to note that, in MAM, the main source of heating over the continent is located over the WAM region and the Ethiopian highlands. During this season, southerlies dominate in the divergent flow over the ocean and LLW are weak. In SON, the major source of heating moves southward over WEA and East Africa. Over the ocean the zonal (meridional) component of the divergent flow strengthens (weakens). These changes to the divergent flow are associated with the strengthening of LLW. In conclusion, we suggest that there is a strong interaction between the WAM system and atmospheric condition over WEA.

At interannual time scales, the variability of the LLW over WEA is modulated by the rotational flow dominated by easterlies. South of 6°N, although the strength of the divergent wind dominates, its year-to-year variability is weak compare to that of the rotational flow. This flow originates from the Indian Ocean, suggesting a possible teleconnection between the interannual variability of LLW and the Indian Ocean.

Acknowledgments

This work is an output from the Climate Science Research Project (CSRP) a project funded by the U.K. Department of International Development (DFID) for the benefit of developing countries. R. S. Chadwick is supported by the Joint DECC/Defra Met Office Hadley Centre Climate Programme (GA01101). The authors would like to thank Samson Hagos for useful discussions. We also acknowledge the helpful comments provided by the reviewers.

REFERENCES

  • Aguilar, E., and et al. , 2009: Changes in temperature and precipitation extremes in western Central Africa, Guinea Conakry, and Zimbabwe, 1955–2006. J. Geophys. Res.,114, D02115, doi:10.1029/2008JD011010.

  • Annamalai, H., , J. M. Slingo, , K. R. Sperber, , and K. Hodges, 1999: The mean evolution and variability of the Asian summer monsoon: Comparison of ECMWF and NCEP/NCAR reanalyses. Mon. Wea. Rev., 127, 11571186, doi:10.1175/1520-0493(1999)127<1157:TMEAVO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Burde, G. I., , A. Zangvil, , and P. J. Lamb, 1996: Estimation of the role of local evaporation in precipitation for a two-dimensional region. J. Climate, 9, 13281338, doi:10.1175/1520-0442(1996)009<1328:ETROLE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Dee, D. P., and et al. , 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553597, doi:10.1002/qj.828.

    • Search Google Scholar
    • Export Citation
  • de Wasseige, C., , D. Devers, , P. de Marcken, , R. E. Atyi, , R. Nasi, , and P. Mayaux, 2009: Les forêts du Bassin du Congo—Etat des forêts 2008. Office des publications de l’Union Europenne Rep., 426 pp.

  • de Wasseige, C., and et al. , 2012: Les forêts du bassin du Congo—Etat des forêts 2010. Office des publications de l’Union Europenne Rep., 276 pp.

  • Findlater, J., 1969: A major low-level air current near the Indian Ocean during the northern summer. Quart. J. Roy. Meteor. Soc., 95, 362380, doi:10.1002/qj.49709540409.

    • Search Google Scholar
    • Export Citation
  • Fontaine, B., , and S. Janicot, 1992: Wind-field coherence and its variations over West Africa. J. Climate, 5, 512524, doi:10.1175/1520-0442(1992)005<0512:WFCAIV>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Fontaine, B., , N. Philippon, , S. Trzaska, , and P. Roucou, 2002: Spring to summer changes in the West African monsoon through the NCEP/NCAR reanalyses (1968–1998). J. Geophys. Res., 107, doi:10.1029/2001JD000834.

    • Search Google Scholar
    • Export Citation
  • Fontan, J., , A. Druilhet, , B. Benech, , and R. Lyra, 1992: The DECAFE experiments: Overview and meteorology. J. Geophys. Res., 97, 61236136, doi:10.1029/91JD02426.

    • Search Google Scholar
    • Export Citation
  • Grodsky, S. A., , J. A. Carton, , and S. Nigam, 2003: Near surface westerly wind jet in the Atlantic ITCZ. Geophys. Res. Lett., 30, 2009, doi:10.1029/2003GL017867.

    • Search Google Scholar
    • Export Citation
  • Hagos, S. M., , and C. Zhang, 2010: Diabatic heating, divergent circulation and moisture transport in the African monsoon system. Quart. J. Roy. Meteor. Soc., 136, 411–425, doi:10.1002/qj.538.

    • Search Google Scholar
    • Export Citation
  • Jackson, B., , S. E. Nicholson, , and D. Kloster, 2009: Mesoscale convective system over western equatorial Africa and their relation to large-scale circulation. Mon. Wea. Rev., 137, 12721294, doi:10.1175/2008MWR2525.1.

    • Search Google Scholar
    • Export Citation
  • James, R., , R. Washington, , and D. P. Rowell, 2013: Implications of global warming for the climate of African rainforests. Philos. Trans. Roy. Soc.,368B, 20120298, doi:10.1098/rstb.2012.0298.

  • Johnson, D. R., , R. D. Townsend, , and M. Y. Wei, 1985: The thermally coupled response of the planetary scale circulation to the global distribution of heat sources and sinks. Tellus, 37A, 106–125, doi:10.1111/j.1600-0870.1985.tb00274.x.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and et al. , 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Kanamitsu, M., , W. Ebisuzaki, , J. Woolen, , S. K. Yang, , J. J. Hnilo, , M. Fiorino, , and G. L. Potter, 2002: NCEP–DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 16311643, doi:10.1175/BAMS-83-11-1631.

    • Search Google Scholar
    • Export Citation
  • Li, Z., , Y. Chao, , and J. C. McWilliams, 2006: Computation of the streamfunction and velocity potential for limited and irregular domains. Mon. Wea. Rev., 134, 33843394, doi:10.1175/MWR3249.1.

    • Search Google Scholar
    • Export Citation
  • Matsuyama, H., , T. Oki, , M. Shinoda, , and K. Masuda, 1994: The seasonal change of the water budget in the Congo River Basin. J. Meteor. Soc. Japan, 32, 281299.

    • Search Google Scholar
    • Export Citation
  • McCollum, J. R., , A. Gruber, , and M. B. Ba, 2000: Discrepancy between gauges and satellite estimates of rainfall in equatorial Africa. J. Appl. Meteor., 39, 666679, doi:10.1175/1520-0450-39.5.666.

    • Search Google Scholar
    • Export Citation
  • Molua, E. L., , and C. M. Lambi, 2007: The economic impact of climate change on agriculture in Cameroon. World Bank Policy Research Working Paper 4364, 31 pp.

  • Nicholson, S. E., 2010: A low-level jet along the Benguela coast, an integral part of the Benguela Current ecosystem. Climatic Change, 99, 613624, doi:10.1007/s10584-009-9678-z.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., , and J. P. Grist, 2003: The seasonal evolution of the atmospheric circulation over West Africa and equatorial Africa. J. Climate, 16, 10131030, doi:10.1175/1520-0442(2003)016<1013:TSEOTA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., , and A. K. Dezfuli, 2013: The relationship of rainfall variability in western equatorial Africa to the tropical oceans and atmospheric circulation. Part I: The boreal spring. J. Climate, 26, 45–65, doi:10.1175/JCLI-D-11-00653.1.

    • Search Google Scholar
    • Export Citation
  • Pokam, W. M., , L. A. T. Djiotang, , and F. K. Mkankam, 2012: Atmospheric water vapor transport and recycling in equatorial central Africa through NCEP/NCAR reanalysis data. Climate Dyn., 38, 17151729, doi:10.1007/s00382-011-1242-7.

    • Search Google Scholar
    • Export Citation
  • Pu, B., , and K. H. Cook, 2010: Dynamics of the West African westerly jet. J. Climate, 23, 62636276, doi:10.1175/2010JCLI3648.1.

  • Rienecker, M. M., and et al. , 2011: MERRA: NASA’s Modern-Era Retrospective Analysis for Research and Applications. J. Climate, 24, 3624–3648, doi:10.1175/JCLI-D-11-00015.1.

    • Search Google Scholar
    • Export Citation
  • Rozendaal, M. A., , C. B. Leovy, , and S. A. Klein, 1995: An observational study of the diurnal cycle of marine stratiform cloud. J. Climate, 8, 17951809, doi:10.1175/1520-0442(1995)008<1795:AOSODV>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sheil, D., , and D. Murdiyarso, 2009: How forests attract rain: An examination of a new hypothesis. Bioscience, 59, 341347, doi:10.1525/bio.2009.59.4.12.

    • Search Google Scholar
    • Export Citation
  • Slingo, A., , R. Brown, , and C. L. Wrench, 1982: A field-study of nocturnal stratocumulus; III. High-resolution radiative and microphysical observations. Quart. J. Roy. Meteor. Soc., 108, 145165, doi:10.1002/qj.49710845509.

    • Search Google Scholar
    • Export Citation
  • Sultan, B., , and S. Janicot, 2003: The West African monsoon dynamics. Part II: The preonset and onset of the summer monsoon. J. Climate, 16, 34073427, doi:10.1175/1520-0442(2003)016<3407:TWAMDP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C. D., , H. Nguyen, , C. Zhang, , and P. Peyrillé, 2011: Annual cycle of the West African monsoon: Regional circulations and associated water vapour transport. Quart. J. Roy. Meteor. Soc., 137, 129147, doi:10.1002/qj.728.

    • Search Google Scholar
    • Export Citation
  • Tomas, R. A., , and P. J. Webster, 1997: The role of inertial instability in determining the location and strength of near-equatorial convection. Quart. J. Roy. Meteor. Soc., 123, 14451482, doi:10.1002/qj.49712354202.

    • Search Google Scholar
    • Export Citation
  • Vondou, D. A., , A. Nzeukou, , A. Lenouo, , and F. M. Kamga, 2010: Seasonal variations in the diurnal patterns of convection in Cameroon–Nigeria and their neighboring areas. Atmos. Sci. Lett., 11, 290–330, doi:10.1002/asl.297.

    • Search Google Scholar
    • Export Citation
  • Wada, M., 1969: Concerning the mechanism of the decaying of typhoon. J. Meteor. Soc. Japan, 47, 315335.

  • Washington, R., , R. James, , H. Pearce, , W. M. Pokam, , and W. Moufouma-Okia, 2013: Congo Basin rainfall climatology: Can we believe the climate models? Philos. Trans. Roy. Soc.,368B, 20120296, doi:10.1098/rstb.2012.0296.

  • Wood, R., 2012: Stratocumulus clouds. Mon. Wea. Rev., 140, 23732423, doi:10.1175/MWR-D-11-00121.1.

  • Yanai, M., , and T. Tomita, 1998: Seasonal and interannual variability of atmospheric heat sources and moisture sinks as determined from NCEP–NCAR reanalysis. J. Climate, 11, 463482, doi:10.1175/1520-0442(1998)011<0463:SAIVOA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yanai, M., , S. Esbensen, , and J.-H. Chu, 1973: Determination of bulk properties of tropical cloud clusters from large-scale heat and moisture budgets. J. Atmos. Sci., 30, 611627, doi:10.1175/1520-0469(1973)030<0611:DOBPOT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., , P. Woodworth, , and G. Gu, 2006: The seasonal cycle in the lower troposphere over West Africa from sounding observations. Quart. J. Roy. Meteor. Soc., 132, 25592582, doi:10.1256/qj.06.23.

    • Search Google Scholar
    • Export Citation
Save