• Abdou, K., , D. J. Parker, , B. Brooks, , N. Kalthoff, , and T. Lebel, 2010: The diurnal cycle of lower boundary-layer wind in the West African monsoon. Quart. J. Roy. Meteor. Soc., 136 (Suppl.), 6676.

    • Search Google Scholar
    • Export Citation
  • Bain, C. L., , D. J. Parker, , C. M. Taylor, , L. Kergoat, , and F. Guichard, 2010: Observations of the nocturnal boundary layer associated with the West African monsoon. Mon. Wea. Rev., 138, 31423156.

    • Search Google Scholar
    • Export Citation
  • Blackadar, A. K., 1957: Boundary layer wind maxima and their significance for the growth of nocturnal inversions. Bull. Amer. Meteor. Soc., 38, 283290.

    • Search Google Scholar
    • Export Citation
  • Bock, O., and Coauthors, 2008: West African Monsoon observed with ground-based GPS receivers during African Monsoon Multidisciplinary Analysis (AMMA). J. Geophys. Res., 113, D21105, doi:10.1029/2008JD010327.

    • Search Google Scholar
    • Export Citation
  • Crewell, S., , and U. Löhnert, 2007: Accuracy of boundary layer temperature profiles retrieved with multifrequency multiangle microwave radiometry. IEEE Trans. Geosci. Remote Sens., 45, 21952201.

    • Search Google Scholar
    • Export Citation
  • Douglas, M., , R. Beida, , and A. Dominguez, 2010: Developing high spatial resolution daytime cloud climatologies for Africa. Preprints, 29th Conf. on Hurricanes and Tropical Meteorology, Tucson, AZ, Amer. Meteor. Soc., P2.23. [Available online at http://ams.confex.com/ams/pdfpapers/168263.pdf.]

  • Farquharson, J. S., 1939: The diurnal variation of wind over tropical Africa. Quart. J. Roy. Meteor. Soc., 65, 165184.

  • Hamilton, R. A., , and J. W. Archbold, 1945: Meteorology of Nigeria and adjacent territory. Quart. J. Roy. Meteor. Soc., 71, 231264.

  • Jacoby-Koaly, S., , B. Campistron, , S. Bernard, , B. Bénech, , F. Ardhuin-Girard, , J. Dessens, , E. Dupont, , and B. Carissimo, 2002: Turbulent dissipation rate in the boundary layer via UHF wind profiler Doppler spectral width measurements. Bound.-Layer Meteor., 103, 361389.

    • Search Google Scholar
    • Export Citation
  • Knippertz, P., , A. H. Fink, , R. Schuster, , J. Trentmann, , J. M. Schrage, , and C. Yorke, 2011: Ultra-low clouds over the southern West African monsoon region. Geophys. Res. Lett., 38, L21808, doi:10.1029/2011GL049278.

    • Search Google Scholar
    • Export Citation
  • Lensky, I. M., , and D. Rosenfeld, 2008: Clouds-Aerosols-Precipitation Satellite Analysis Tool (CAPSAT). Atmos. Chem. Phys., 8, 67396753, doi:10.5194/acp-8-6739-2008.

    • Search Google Scholar
    • Export Citation
  • Lothon, M., , F. Saïd, , F. Lohou, , and B. Campistron, 2008: Observation of the diurnal cycle in the low troposphere of West Africa. Mon. Wea. Rev., 136, 34773500.

    • Search Google Scholar
    • Export Citation
  • National Weather Service, 2010: Rawinsonde observations. National Weather Service Manual 10-1401, 208 pp. [Available online at http://www.weather.gov/directives/sym/pd01014001curr.pdf.]

  • Parker, D. J., and Coauthors, 2005: The diurnal cycle of the West African monsoon circulation. Quart. J. Roy. Meteor. Soc., 131, 28392860.

    • Search Google Scholar
    • Export Citation
  • Parker, D. J., and Coauthors, 2008: The AMMA radiosonde program and its implications for the future of atmospheric monitoring over Africa. Bull. Amer. Meteor. Soc., 89, 10151027.

    • Search Google Scholar
    • Export Citation
  • Pospichal, B., , and S. Crewell, 2007: Boundary layer observations in West Africa using a novel microwave radiometer. Meteor. Z., 16, 513523.

    • Search Google Scholar
    • Export Citation
  • Redelsperger, J.-L., , C. D. Thorncroft, , A. Diedhiou, , T. Lebel, , D. J. Parker, , and J. Polcher, 2006: African Monsoon Multidisciplinary Analysis: An international research project and field campaign. Bull. Amer. Meteor. Soc., 87, 17391746.

    • Search Google Scholar
    • Export Citation
  • Schrage, J. M., , S. Augustyn, , and A. H. Fink, 2007: Nocturnal stratiform cloudiness during the West African monsoon. Meteor. Atmos. Phys., 95, 7386.

    • Search Google Scholar
    • Export Citation
  • Stephens, G., and Coauthors, 2002: The CloudSat mission and the A-train. Bull. Amer. Meteor. Soc., 83, 17711790.

  • WMO, 1995: Alphanumeric codes. Manual on codes, Vol. I.1, Part A, WMO Publ. 306.

  • View in gallery

    “Climatology of stratus and stratocumulus” at 0600 UTC June–October 2006. (a) Frequency with which stratiform clouds were observed. (b) Average coverage of stratiform clouds when such clouds are present. The size of the marker suggests how often the station reports were available. The star indicates the location of Nangatchori, “L” denotes Lagos (Nigeria), “C” denotes Cotonou (Benin), “A” denotes Ada (Ghana), “D” denotes Daloa, and “G” denotes Gagnoa (all Ivory Coast).

  • View in gallery

    Seasonal cycle of stratus cloud average in 2006 for (a) the entire region, (b) stations between 6.5° and 9°N, and (c) stations between 9° and 11°N. Markers indicate the percentage of stations that reported stratus, and bars reflect the average coverage of stratus (octas) at the stations that reported stratus. The geographic region and the stations used are shown in Fig. 1.

  • View in gallery

    (a) MSG false color image at 0130 UTC 20 Aug 2006 produced from a combination of three IR channels: blue, channel 9 (10.8 μm); red, channel 10 (12 μm) minus channel 9; and green, channel 9 minus channel 4 (3.9 μm). (b) Cross section along the white line in (a) of the perpendicular backscatter coefficient at 532 nm from the CALIOP lidar on board the CALIPSO satellite. The units are 1 (steradian × km)−1. (c) MSG visible (0.8 μm) from 0700 UTC 20 Aug 2006.

  • View in gallery

    (a) MSG visible (0.8 μm) satellite image for 0700 UTC 4 Aug 2006, showing widespread nocturnal stratiform clouds across the Soudanian and Guinea Coast climate zones of tropical West Africa. (b) Visible satellite image for 0700 UTC 4 Jul 2006, showing clear conditions across large parts of Ghana, Togo, Benin, and Nigeria.

  • View in gallery

    Mean difference in cloudiness in percent between Aqua and Terra for the period July–August 2005–10. Aqua (Terra) overpasses are at 1030 (1330) LT and red (blue) colors indicate decreasing (increasing) cloudiness. The contours indicate altitudes of 200 m (black), 400 m (purple), 600 m (green), 800 m (orange), and 1000 m (red).

  • View in gallery

    Backscatter coefficient at 905 nm in 1 (steradian × km × 10 000)−1 (i.e., the fraction of power that is scattered back to the receiver per unit length and per unit solid angle and multiplied by a scaling constant) from the ceilometer at Nangatchori. (a) A cloudy night (19–20 Aug 2006) and (b) a clear night (3–4 Jul 2006). Both nights are included in Tables 1 and 2.

  • View in gallery

    Time–height cross sections of various meteorological parameters for the period between 1600 UTC 8 Aug 2006 and 0700 UTC 9 Aug 2006. (a) The backscatter coefficient from the ceilometer observations at Nangatchori in units of 1 (steradian × km × 10 000)−1. (b) The wind speed in m s−1 as determined by the profiler at Nangatchori (shading) and the associated wind vectors; the wind barbs denote surface winds from the Nangatchori weather station. (c) The static stability in K km−1 based on temperature retrievals from the HATPRO profiler at Nangatchori. The bold closed line indicates an isothermal layer. (d) The gradient Richardson number (Ri, no units) calculated from the winds in (b) and temperature gradients in (c). In (d), the light gray shaded area marks the transition at a value of 0.25 < Ri < 1 between stable (higher Ri values) and turbulent flow (lower Ri values).

  • View in gallery

    As in Fig. 7, but for the period from 1600 UTC 24 Aug 2006 to 0700 UTC 25 Aug 2006.

  • View in gallery

    As in Fig. 7, but for the period from 1600 UTC 21 Sep 2006 to 0700 UTC 22 Sep 2006.

  • View in gallery

    As in Fig. 7, but as composites over the following nights: 8–9, 18–19, 19–20, 22–23 Aug; 21–22 Sep; and 5–6 and 6–7 Oct (i.e., the cloudy nights for which we have wind profiler data). Data have been composited such that hour 0 corresponds to the time of onset of low clouds at Nangatchori. (c) The vertical temperature profiles were only taken before sunrise. Therefore the x axis is truncated at 4 h after low cloud onset. (d) The contour colors merely provide visual contrast, and the dashed contour corresponds to a value of 0.25.

  • View in gallery

    Time series of mean wind speed for the 200–400 m AGL layer, as retrieved by the wind profiler at Nangatchori on nights when low cloud decks formed. Here the onset of low clouds (as determined by the ceilometer at Nangatchori) is defined as hour 0.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 117 117 9
PDF Downloads 53 53 7

Nocturnal Continental Low-Level Stratus over Tropical West Africa: Observations and Possible Mechanisms Controlling Its Onset

View More View Less
  • 1 Department of Atmospheric Sciences, Creighton University, Omaha, Nebraska
  • | 2 Institute of Geophysics and Meteorology, University of Cologne, Cologne, Germany
© Get Permissions
Full access

Abstract

Some spatiotemporal characteristics and possible mechanisms controlling the onset of the widespread, low-level nocturnal stratiform clouds that formed during May–October 2006 over southern tropical West Africa are investigated using cloudiness observations from surface weather stations, data from various satellite platforms, and surface-based remote sensing profiles at Nangatchori in central Benin. It is found that the continental stratus is lower than the maritime stratus over the Gulf of Guinea and persists well into the noon hours. For the study period, a clear seasonal cycle was documented, as well as a dependence on latitude with the cloudiest zone north of the coastal zone and south of approximately 9°N. It is also shown that nonprecipitating clear and cloudy nights observed at Nangatchori in central Benin often reflect clearer and cloudier than normal conditions over a wide region of southern West Africa. At Nangatchori, on average the stratus developed at 0236 UTC (about local time) with an extremely low cloud base at 172 m (above ground level) when averaged over all cloudy nights. About 2–3 h before cloudiness onset, a distinct nighttime low-level jet formed that promoted static destabilization and a low Richardson number flow underneath it. The ensuing vertical upward mixing of moisture that accumulated under the near-surface inversion after sunset caused the cloud formation. It is argued that a strong shear underneath the nighttime low-level jet is the major process for cloud formation, but the low-level static stability and the time scale of the shear-driven mixing are other potential factors.

Corresponding author address: Dr. Jon M. Schrage, Department of Atmospheric Sciences, Creighton University, 2500 California Plaza, Omaha, NE 68178. E-mail: jon@creighton.edu

Abstract

Some spatiotemporal characteristics and possible mechanisms controlling the onset of the widespread, low-level nocturnal stratiform clouds that formed during May–October 2006 over southern tropical West Africa are investigated using cloudiness observations from surface weather stations, data from various satellite platforms, and surface-based remote sensing profiles at Nangatchori in central Benin. It is found that the continental stratus is lower than the maritime stratus over the Gulf of Guinea and persists well into the noon hours. For the study period, a clear seasonal cycle was documented, as well as a dependence on latitude with the cloudiest zone north of the coastal zone and south of approximately 9°N. It is also shown that nonprecipitating clear and cloudy nights observed at Nangatchori in central Benin often reflect clearer and cloudier than normal conditions over a wide region of southern West Africa. At Nangatchori, on average the stratus developed at 0236 UTC (about local time) with an extremely low cloud base at 172 m (above ground level) when averaged over all cloudy nights. About 2–3 h before cloudiness onset, a distinct nighttime low-level jet formed that promoted static destabilization and a low Richardson number flow underneath it. The ensuing vertical upward mixing of moisture that accumulated under the near-surface inversion after sunset caused the cloud formation. It is argued that a strong shear underneath the nighttime low-level jet is the major process for cloud formation, but the low-level static stability and the time scale of the shear-driven mixing are other potential factors.

Corresponding author address: Dr. Jon M. Schrage, Department of Atmospheric Sciences, Creighton University, 2500 California Plaza, Omaha, NE 68178. E-mail: jon@creighton.edu

1. Introduction

At night during the West African summer monsoon season between May and October, the Soudanian (9°–12°N) and Guinea Coast (south of 9°N) climate zones of West Africa are often blanketed in a shallow layer of clouds that are typically described in synoptic observations as either stratus or stratocumulus. These shallow cloud layers are extremely difficult to detect in an analysis of infrared satellite pictures because of their relatively warm temperatures (Knippertz et al. 2011). While these stratiform cloud decks seem to be quite common, they are by no means ubiquitous. Rather, sometimes it is the case that large regions of the Soudano–Guinean zone exhibit clear conditions at low levels at night during the rainy season. Schrage et al. (2007) used a combination of in situ, satellite, and European Centre for Medium-Range Weather Forecasts (ECMWF) operational analysis data from the summer monsoon season in West Africa in 2002 to compare the structure of the atmosphere on nights in which this stratiform cloud layer formed at a relatively early hour (viz., prior to 0000 UTC) with that observed on nights in which the same region remained relatively cloud free. That study focused on the radiosonde station at Parakou, Benin (9.35°N, 2.62°E), as this was the only operational upper-air station in the Soudanian climate zone at that time.

Schrage et al. (2007) demonstrated that the cloudy nights in Parakou were characterized by strong moisture convergence at 925 hPa in the ECMWF operational analyses. This moisture convergence was driven by enhanced convergence in the wind field, primarily caused by friction near the surface. Schrage et al. (2007) proposed that cloudy nights were maintained through a positive feedback in which the clouds maintain the low static stability through their radiative impact, coupling the surface to the 925-hPa level. For the clear nights, the authors proposed that radiative cooling of the surface shortly after sunset decouples the near-surface layer from the ground through the formation of a shallow stable layer. Monsoon southwesterlies at 925 hPa in these cases experience less friction, yielding much less moisture convergence and a positive feedback that maintained the clear conditions.

The present work suggests that the development of a nocturnal low-level jet (NLLJ) across this region may play a key role in the development of these cloud systems. The diurnal cycle of the flow over tropical Africa has been explored by a number of studies, most notably Farquharson (1939), Hamilton and Archbold (1945), and Parker et al. (2005). These studies and others are in agreement that there is a nocturnal peak in the low-level flow a few hundred meters above the surface. These winds appear to be broadly consistent with the structure and evolution of a NLLJ first suggested by Blackadar (1957), although the low latitude calls into question the assumption that the winds can achieve supergeostrophic velocities, even in the reduced-friction regime observed at night.

Recent studies on the nocturnal boundary layer and the NLLJ in West Africa used data gathered during the African Monsoon Multidisciplinary Analyses (AMMA) 2006 special observing periods (SOPs) or the 2005–07 extended observing period (cf. Redelsperger et al. 2006). Several investigations (e.g., Lothon et al. 2008; Abdou et al. 2010; Bain et al. 2010) focused on the conditions in the Sahel zone (e.g., north of 12°N), which is located several hundred kilometers north of the typical extent of the nocturnal stratiform cloud decks. Bain et al. (2010) used tethered balloon soundings up to a height of 200 m at the Sahelian site of Gourma (15.2°N, 1.3°W) to show the expected evolution of a nighttime boundary layer, with a near-surface inversion decoupling the overlying atmosphere from the ground, a NLLJ developing, and mechanically driven turbulence reducing the static stability before sunrise.

Lothon et al. (2008) compared the evolution and characteristics of the NLLJs at Niamey (13.48°N, 2.17°E, Niger, Sahel zone) and at the site of the present study, Nangatchori (9.70°N, 1.68°E, Benin, Soudanian zone). At Niamey, the NLLJ peaked at 350 m AGL with wind speeds around 15 m s−1, set in at 2000 UTC, and peaked at about 0600–0700 UTC. In April 2006, when the transition from the dry to the humid season (i.e., persistently high surface dewpoint temperatures) at Nangatchori was observed, the NLLJ formed more rapidly after sunset, was less thick (1000 m instead of 1500 m), and stopped earlier in the course of the day than in Niamey. Lothon et al. (2008) linked larger turbulent kinetic energy (TKE) values and negative skewness of surface vertical velocity at nighttime to the downward transport of momentum from the NLLJ. Using these indirect measures of a NLLJ existence, Lothon et al. (2008) concluded that the NLLJ is a regular feature of all seasons at Nangatchori. Its occurrence and strength peaked in the premonsoon season (March–May), and was less during both the dry (November–February) and wet seasons (June–September). Some very low TKE values in some nights during the wet season led them to speculate that no NLLJ was established during those nights.

While these AMMA studies helped to shed light on the evolution of the nighttime boundary layer and the NLLJ, the relation to the extensive overnight stratus cloud cover, which is often observed between about May and October over the densely vegetated southern zone and that has received very little attention in the literature to date, has not been explored. Thus, the first major goal of the present study is to provide new observational evidence of the nocturnal low stratiform cloud deck using classic cloud observations from weather stations in combination with recently available data from satellite platforms, and data from the AMMA 2006 summer monsoon SOPs. The second major aim of the present study is to explore a possible connection between the NLLJ, its impact on stability and mixing, and the development of low stratiform cloud decks at Nangatchori using data from various in situ instruments deployed during the AMMA 2006 SOPs. Section 2 describes the datasets used in this present study. The methodology for selecting and compositing data in this study is explained in section 3. A number of specific cases of nights without clouds or with clouds that formed at different times during the night are presented in section 4, whereas the various cases are composited with respect to the time of onset of stratocumulus clouds in section 5. Results are summarized and discussed in section 6.

2. Data

a. HATPRO microwave profiler and lidar ceilometer at Nangatchori

Estimates of the static stability of various layers of the atmosphere were calculated from temperature and humidity retrievals made by a Humidity and Temperature Profiler (HATPRO) microwave profiler operating at Nangatchori in 2006. The physical configuration and geographic setting of this profiler has been described in detail by Pospichal and Crewell (2007). The instrument observes the temperature and humidity structure of the atmosphere every 15 min. For the present study, calculations were based only on the temperature and humidity estimates on levels at or below 1000 m above ground level (AGL), and the vertical resolution of the data is 50 m below 250 m, 75 m below 700 m, and 100 m until an altitude of 1 km AGL. Crewell and Löhnert (2007) have shown that the temperature estimates obtained from such a profiler are accurate to within 1 K in the lowest 1.5 km of the atmosphere and that these data are particularly useful for assessing the height and strength of low-level temperature inversions. A collocated lidar ceilometer operating at a wavelength of 905 nm was used to detect cloud-base heights. Pospichal and Crewell (2007) provide a description of the configuration of this instrument, which had a temporal resolution of 15 s and a vertical resolution of 30 m.

b. Wind profiler at Nangatchori

As part of the AMMA field campaign, a UHF wind profiler was installed at Nangatchori, Benin. The use and interpretation of data from such a wind profiler has been summarized by Jacoby-Koaly et al. (2002). Wind profiles were derived from the low mode of the instrument approximately 14 times per hour on average at a gate spacing of 75 m (i.e., available vertical resolution) starting at 74 m AGL. This instrument was put into operation in April of 2006, but most of the data gathered prior to 7 August 2006 are missing as a result of a variety of technical problems on site. As discussed in Jacoby-Koaly et al. (2002), the lowest levels of wind profiler data are affected by ground clutter contamination. As a consequence, the two lowest levels of the UHF low mode data, 74 and 149 m AGL, have not been used. The lowermost level is 224 m AGL and the near-surface vertical wind shear has been estimated between the surface wind at the Nangatchori ground station and the UHF wind at this level.

c. Radiosondes at Parakou

Wind profiles at a vertical resolution of 5 m were taken from radiosondes launched at Parakou, about 100 km southeast of Nangatchori, to carry out consistency checks between the UHF and radiosonde wind profiles. During different phases of the AMMA field campaign, nominally two, four, or eight radiosondes were launched daily. Between 8 June, the restart of radiosounding at Parakou during AMMA SOP1, and 30 September 2006, the end of the AMMA SOPs, the launching frequency was 4 times day−1 at 0000, 0600, 1200, and 1800 UTC, with additional soundings at 0300, 0900, 1500, and 2100 UTC during the two intensive observing periods: 20–30 June and 1–15 August 2006 (cf. Parker et al. 2008). In October 2006, only twice-daily soundings at 0000 and 1200 UTC were performed. As is standard practice (e.g., National Weather Service 2010), the balloons are actually launched 60–90 min prior to the official synoptic time (e.g., the balloon of the 0000 UTC sounding is launched between 2230 and 2300 UTC); since the balloon crosses the levels of interest within the first few minutes of the flight, the radiosonde wind profile nominally associated with conditions at 0000 UTC is assigned to 2300 UTC instead. Similarly, other radiosonde launches were also associated with an offset of one hour.

d. Synoptic reports

Most synoptic weather stations across West Africa are manned and operational 24 h day−1. The trained observers report eye observations of cloudiness on an hourly basis according to World Meteorological Organization (WMO) regulations. These cloudiness reports are transmitted into the Global Telecommunication Network (GTS) of WMO on a 3- or 6-hourly basis. Cloud information (i.e., cloud types, cloud cover in octas, and height of cloud base AGL) is coded in parts I and III of the WMO FM12 SYNOP code (WMO 1995, see also http://www.wmo.int/pages/prog/www/WMOCodes.html). Whereas cloud cover of low-, or if not present, midlevel clouds and types of low-, midlevel, and high-level clouds are provided in the eighth group of part I in the SYNOP code, more details on cloudiness is given in the eighth group of part III. In this part, multiple occurrences of the eighth group are possible if various cloud types are present. Thus, for example, it is possible to distinguish between cumulus and stratocumulus cloud characteristics that have a common code (i.e., CL= 8) in the first part of the SYNOP. The cloud information in the third part of the SYNOP was used in the present study and only eighth groups with CL= 6 (stratus, St) and CL= 7 (stratocumulus, Sc) coding were considered to describe stratiform cloudiness at 0600 UTC.

Various archives of SYNOP data were utilized. The primary source was the AMMA Database (http://amma-international.org/database), which contains decoded SYNOP reports that were available in real time on the GTS. Unfortunately, the synoptic reports are not always available on the GTS and the extent of the data gaps vary from country to country. In the present study, data gaps could be filled in completely for Benin using the raw SYNOP code available at the “Direction Nationale de la Météorologie” for six Beninese stations. Missing cloud information for 22 SYNOP stations in Ghana was furnished by the Ghana Meteorological Service (GMET) in decoded form, but only from part I of the SYNOP. As a consequence, the frequency and cloud cover of stratiform clouds at Ghanaian stations is somewhat underestimated as reports of cumulus and stratocumulus (CL = 8) in part I were not counted as stratiform cloudiness. Only CL= 5 (Sc) and CL= 6 (St nebulosus or fractus) were considered. Reports of CL= 4 (Sc cumulonimbogenitus) were absent and CL= 7 (Stratus fractus and cumulus fractus) were very infrequent in the GMET data.

e. Satellite imagery

To visualize the low stratiform clouds over the West African continent, Meteosat Second Generation (MSG) images that are produced from a combination of three Spinning Enhanced Visible and Infrared Imager (SEVIRI) infrared channels were used: blue, channel 9 (10.8 μm); red, channel 10 (12 μm) minus channel 9; and green, channel 9 minus channel 4 (3.9 μm). The raw data were extracted from the AMMA database and the resolution at nadir is 3 km. The “Night Microphysical” scheme described in Lensky and Rosenfeld (2008) and on the EUMETSAT web page (http://oiswww.eumetsat.int/~idds/html/doc/fog_interpretation.pdf) uses the same channel combination. However, the parameter ranges in the red–greenblue (RGB) composite were modified slightly to meet tropical conditions—red: −4 to +2 K, green: 0 to +3 K, and blue: 253 to 313 K. Values lower (higher) than the ranges were set to zero (saturation). The adaption was tested against cloud observations from surface synoptic stations. Low-level clouds or fog appear in greenish colors in the RGB composites. MSG visible (0.8 μm) satellite images at an approximate resolution of 5-km resolution were used to detect stratus clouds after sunrise.

The perpendicular backscatter coefficient at 532 nm from the Cloud Aerosol Lidar with Orthogonal Polarization (CALIOP) lidar on board the Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) satellite has been used to illustrate examples of stratiform clouds. CALIPSO is part of the A-train satellite constellation (Stephens et al. 2002) and passes over the region between approximately 0115 and 0130 LT. The CALIOP lidar is a nonscanning instrument with a very narrow footprint of 100 m.

3. Observational evidence of nocturnal continental stratus

Maps of frequency of stratocumulus occurrence and cover for June–October 2006 were compiled from the synoptic reports for 0600 UTC [this is local standard time (LST) in Ivory Coast, Ghana, Burkina Faso, and Togo, and LST + 1 h in Benin and Nigeria]. Figure 1a shows how frequently stratiform cloudiness is reported and Fig. 1b exhibits the mean stratiform cloud cover in octas over times when stratus was reported. Cotonou, a coastal station in Benin, is an example where all 0600 UTC reports had stratus reports (Fig. 1a), but the mean coverage of three octas indicates that the coverage is often not particularly extensive (Fig. 1b). A high frequency of occurrence of stratiform clouds paired with relatively low coverage is typical for other coastal stations. Lagos and Ada observe stratiform clouds 81%–94% of the time, but with four octas mean coverage between May and October 2006. Frequent (81%–100%) and spatially extensive (5.5–7.5 octas) stratus decks are observed inland of the coastal strip. For example, Dalao and Gagnoa in the west-central Ivory Coast have more than 7.5 octas of stratus (Fig. 1b) averaged over 94% (Fig. 1a) of the available observations. The frequency and extent of cloudiness decrease for stations north of about 9°N. Thus, the SYNOP reports indicate extensive morning stratus, especially inland of the coastal strip, but also that stratiform cloud base is generally low, mostly between 200 and 400 AGL (not shown).

Fig. 1.
Fig. 1.

“Climatology of stratus and stratocumulus” at 0600 UTC June–October 2006. (a) Frequency with which stratiform clouds were observed. (b) Average coverage of stratiform clouds when such clouds are present. The size of the marker suggests how often the station reports were available. The star indicates the location of Nangatchori, “L” denotes Lagos (Nigeria), “C” denotes Cotonou (Benin), “A” denotes Ada (Ghana), “D” denotes Daloa, and “G” denotes Gagnoa (all Ivory Coast).

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

The 2006 seasonal cycle of low-level stratiform clouds in the synoptic observations at 0600 UTC is depicted in Fig. 2. Using all the stations in the region, Fig. 2a shows that both the number of stations reporting such clouds and the fraction of the sky covered by such clouds reached a peak in approximately mid-August. Focusing on just the stations between 6.5° and 9°N (i.e., the cloudiest portion of the total domain), Fig. 2b shows that it was often the case that 100% of the stations reported such low clouds at 0600 UTC during a long period of the rainy season. Even the drier portion of the total region (represented by the average of the stations between 9° and 11°N in Fig. 2c) had several nights in which such clouds were very widespread.

Fig. 2.
Fig. 2.

Seasonal cycle of stratus cloud average in 2006 for (a) the entire region, (b) stations between 6.5° and 9°N, and (c) stations between 9° and 11°N. Markers indicate the percentage of stations that reported stratus, and bars reflect the average coverage of stratus (octas) at the stations that reported stratus. The geographic region and the stations used are shown in Fig. 1.

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

As will be shown below, the low stratiform cloud deck is shallow and its infrared brightness temperature in the atmospheric window channel (10–12 μm) is barely distinguishable from cloud-free pixels. In contrast to the marine stratus over the Gulf of Guinea, the continental stratus is therefore neither evident in the International Satellite Cloud Climatology Project (ISCCP) D1 daily “number of cloudy pixels” below 800 hPa at 0600 UTC nor in ISCCP D2 monthly mean “low cloud amount” at 0600 UTC (Knippertz et al. 2011). An example of the MSG RGB channel composite that highlights low level nocturnal cloud/fog features is given in Fig. 3a for 0130 UTC 20 August 2006. The extensive stratus north of the Guinea Coast, especially north of the Gulf of Benin, is clearly visible by the greenish areas. There are also large patches of stratus over the waters of the Gulf of Guinea. However, the presence of mid, upper-level, or deep convective clouds obscure the detection of stratus by this method and unbiased stratus climatologies cannot be inferred. Some more insight into the characteristics of the continental stratus can be gained by using the active lidar on board the CALIPSO satellite. The track of the CALIPSO lidar is indicated by the black line in Fig. 3a and the corresponding cross section of the perpendicular 532-nm backscatter coefficient is shown in Fig. 3b. The continental stratus between 6° and 8°N is clearly visible below an altitude of 1 km. The marine stratus over the Gulf of Guinea is higher with tops around 2 km. The maritime stratus can be seen in the CloudSat 94-GHz cloud radar, whereas the continental stratus is not seen in the CloudSat radar because of the contamination of the lowest kilometer with the strong backscatter signal from the surface (cf. Knippertz et al. 2011). Moreover, the presence of mid- and upper-level clouds can again obscure the continental stratus detection by CALIOP. Thus, climatologies of the spatiotemporal distribution of continental stratus over West Africa at night can presently not be inferred from satellite instruments. The stratus deck, seen in the RGB composite north of the Gulf of Benin at 0130 UTC (Fig. 3a), is still discernible in the 0700 UTC visible satellite image (Fig. 3c).

Fig. 3.
Fig. 3.

(a) MSG false color image at 0130 UTC 20 Aug 2006 produced from a combination of three IR channels: blue, channel 9 (10.8 μm); red, channel 10 (12 μm) minus channel 9; and green, channel 9 minus channel 4 (3.9 μm). (b) Cross section along the white line in (a) of the perpendicular backscatter coefficient at 532 nm from the CALIOP lidar on board the CALIPSO satellite. The units are 1 (steradian × km)−1. (c) MSG visible (0.8 μm) from 0700 UTC 20 Aug 2006.

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

While extensive stratiform cloud decks form frequently at night during the height of the monsoon season, these cloud structures do not necessarily form on all nights (cf. Schrage et al. 2007). Figure 4a shows the extensive stratus in the Guineo–Sahelian zone in the visible satellite image from 0700 UTC 4 August 2006. At 0130 UTC, the detection of the low stratus in the RGB composite was largely obscured by the presence of midlevel cloud decks (not shown). Thus, parts of the cloudiness in the southern part of West Africa seen in Fig. 4a may also be altostratus or altocumulus. On the contrary, Fig. 4b shows a much lower amount of stratiform clouds in the Soudanian and Guinea Coast climate zones of West Africa at 0700 UTC 4 July 2006. Extensive cloud-free areas exist across much of Ghana, Togo, Benin, and Nigeria. The early morning visible satellite image shown in Fig. 3c is an intermediate example with extensive stratus coverage (e.g., centered over Benin), but also large cloud-free areas over Nigeria and parts of Ghana.

Fig. 4.
Fig. 4.

(a) MSG visible (0.8 μm) satellite image for 0700 UTC 4 Aug 2006, showing widespread nocturnal stratiform clouds across the Soudanian and Guinea Coast climate zones of tropical West Africa. (b) Visible satellite image for 0700 UTC 4 Jul 2006, showing clear conditions across large parts of Ghana, Togo, Benin, and Nigeria.

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

The hour when the stratus disappears in the course of the day is significant for the radiation budget of the region. Figure 5 shows the difference in cloudiness between Aqua and Terra for overpasses at about 1030 and 1330 LT and averaged over July–August 2005–10. The method to infer the cloudiness percentage from the Moderate Resolution Imaging Spectroradiometer (MODIS) visible imagery is described in Douglas et al. (2010). Here, blue (red) colors indicate regions in which cloud cover increases (decreases) over this 3-h period. Inland from the coastal regions with increasing cloudiness at the sea-breeze front (blue colors), extensive areas in central Ivory Coast, Benin, and western Nigeria show a tendency to clear off in the 3 h between 1030 and 1330 LT (red colors). The altitude contour lines in Fig. 5 indicate that these regions are mostly low-lying (i.e., 200–400 m) flatlands. Absolute mean changes in cloudiness are on the order of 5%–15%. This might appear small. However, the time of the day when the stratiform cloud deck breaks off vary over time, the stratus is not always present in the morning hours at 1030 LT, and days are included that see an increase in cloudiness by 1330 LT by, for example, the arrival of a convective cloud cluster. All these cases entered the map in Fig. 5.

Fig. 5.
Fig. 5.

Mean difference in cloudiness in percent between Aqua and Terra for the period July–August 2005–10. Aqua (Terra) overpasses are at 1030 (1330) LT and red (blue) colors indicate decreasing (increasing) cloudiness. The contours indicate altitudes of 200 m (black), 400 m (purple), 600 m (green), 800 m (orange), and 1000 m (red).

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

The ground-based backscattering lidar at Nangatchori gives some more insight into how the stratus dissipates. Figure 6a shows that the stratus was already present an hour before midnight (2300 UTC) on the night 19–20 August 2006, broke up around 0200 UTC (0300 LST), reformed and thickened later in the night. The stratus seemed to rest on the surface as fog and in fact, at an altitude of about 431 m, Nangatchori experiences very low cloud (<200 m) decks, and sometimes mist or fog. The cloud base starts to lift 2 h after sunrise, but overcast conditions prevail until 1000 UTC (1100 LST) when the stratus breaks up and cumuliform clouds start forming. The clearing off of the stratus can occur earlier than 1100 LST, but also examples with a later disappearance of the stratus are in the sample. The night of 3–4 July 2006 is an example of a night when the nocturnal stratus did not form (Fig. 6b). It is the example shown in the visible satellite image in Fig. 4b. The following two sections are concerned with what controls the onset of nocturnal stratus at Nangatchori.

Fig. 6.
Fig. 6.

Backscatter coefficient at 905 nm in 1 (steradian × km × 10 000)−1 (i.e., the fraction of power that is scattered back to the receiver per unit length and per unit solid angle and multiplied by a scaling constant) from the ceilometer at Nangatchori. (a) A cloudy night (19–20 Aug 2006) and (b) a clear night (3–4 Jul 2006). Both nights are included in Tables 1 and 2.

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

4. Cases

Using the lidar ceilometer data at Nangatchori, nights between May and October 2006 during which a stratiform cloud deck formed and persisted into the morning hours were separated from nights during which no near-surface clouds were detected. To minimize the impact of processes related to moist convection (e.g., convective outflow boundaries), infrared satellite imagery was visually inspected for the presence of convective clusters at night in the vicinity of Nangatchori; such nights and nights with intermittent low cloudiness were excluded from further analysis. The summary statistics for 21 cloudy and 16 clear nights are given in Tables 1 and 2, respectively. For the cloudy nights (Table 1), the onset time and average base of low clouds are given. The mean onset of a persistent stratiform cloud deck at Nangatchori was at 0236 UTC at a height of 172 m AGL. It turns out that most of the cloudy nights at Nangatchori also show extensive cloudiness over the Guineo–Soudanian zone, as noted in the remaining three columns of Table 1. Here it is seen that on nights with stratus indicated by the ceilometer at Nangatchori, 70% of all stations in the study region (cf. Fig. 1) also reported low stratus clouds at 0600 UTC. The 3.6 octas average sky coverage of stratus at stations reporting stratus demonstrates that the low clouds are not only widely observed, but also cover on average half of the sky over stations in the whole study region (Table 1). For the cloudiest region just to the south of the ceilometer (i.e., 6.5°– 9°N, cf. Fig. 1b), the corresponding values are 90% and 5.5 octas.

Table 1.

Summary statistics on time of stratiform cloud formation and the cloud base for selected nights in which nocturnal stratocumulus clouds were detected by the ceilometer at Nangatchori. Note that local sunrise occurs between 0530 and 0545 UTC. For three geographical regions, the percentage of stations reporting stratus clouds at 0600 UTC is given, as well as the average sky coverage of stratus at locations reporting stratus. (Ceilometer information at Nangatchori on the night of 6–7 August is unavailable. The low clouds were already present at 0000 UTC 7 Aug 2006 when the ceilometer began archiving data again after an 18-h outage.)

Table 1.
Table 2.

Summary statistics on coverage of stratus clouds for selected nights in which nocturnal stratocumulus clouds were not detected by the ceilometer at Nangatchori. For three geographical regions, the percentage of stations reporting stratus clouds at 0600 UTC is given, as well as the average sky coverage of stratus at locations reporting stratus. For the averages, an italic (bold) font indicates that the difference between the value for the cloudy (cf. Table 1, bottom row) and clear nights exceeds the 95% (99%) confidence interval for the appropriate t test.

Table 2.

As can be expected from the seasonal cycle of cloudiness (Fig. 2), widespread clear conditions are more often observed in May and October (e.g., 24–25 May and 14–15 October in Table 2). Some clear nights at Nangatchori, however, show extensive stratus in the Guineo–Sahelian zone (e.g., 14–15 August and 24–25 August in Table 2). Despite this case-to-case variability, however, the differences between the averages for both percentage of stations reporting stratus at 0600 UTC and the average coverage of stratus at stations reporting stratus are statistically significant when comparing the clear nights to the cloudy nights in Tables 1 and 2. For example, only 59% of all stations in the entire study region report stratiform clouds and the average sky coverage is reduced to 2.85 octas when averaged over all 0600 UTC observations of the 16 clear nights at Nangatchori (Table 2, last row). In summary, the 0600 UTC synoptic cloud observations suggest that cloudy (clear) nights at Nangatchori are often, but not always, indicative of more (less) extensive low-level cloudiness over wider areas of southern West Africa. In sections 4ac, the near-surface stability and wind conditions at Nangatchori for three selected nights are examined.

a. A night with early stratiform cloud formation: 8–9 August 2006

Figure 7 depicts the meteorological events of an example of a night in which stratocumulus clouds formed about 3 h earlier than the mean onset time of 0236 UTC. While the ceilometer detected some scattered clouds below 1000 m in the first half of the night, a thick layer of low clouds was observed starting shortly before 0000 UTC, and these clouds persisted for the remainder of the night (Fig. 7a). This night was generally windier than was the case on the other two nights presented below, and the wind profiler at Nangatchori identified the formation of a nocturnal low level jet that peaked with a magnitude of more than 8 m s−1 at about 400 m AGL at around 2300 UTC 8 August (Fig. 7b). The available radiosonde observations from Parakou (not shown) provide supporting documentation that these were the strongest low-level winds of the night and that they represented an enhancement of the flow in the monsoon layer. The low-level stability that developed immediately after sunset and in the early night was quite weak compared to the other two examples discussed below. Figure 7c displays the vertical temperature gradient based on retrievals from the HATPRO profiler at Nangatchori and it shows that the near-surface temperature gradient −∂T/∂z remained positive (i.e., never reached an isothermal layering throughout the night), except for a short period and small layer shortly after midnight. The vertical temperature gradient evolution in Fig. 7c also indicates the destabilization below the NLLJ close to the surface after midnight due to mechanically induced mixing.

Fig. 7.
Fig. 7.

Time–height cross sections of various meteorological parameters for the period between 1600 UTC 8 Aug 2006 and 0700 UTC 9 Aug 2006. (a) The backscatter coefficient from the ceilometer observations at Nangatchori in units of 1 (steradian × km × 10 000)−1. (b) The wind speed in m s−1 as determined by the profiler at Nangatchori (shading) and the associated wind vectors; the wind barbs denote surface winds from the Nangatchori weather station. (c) The static stability in K km−1 based on temperature retrievals from the HATPRO profiler at Nangatchori. The bold closed line indicates an isothermal layer. (d) The gradient Richardson number (Ri, no units) calculated from the winds in (b) and temperature gradients in (c). In (d), the light gray shaded area marks the transition at a value of 0.25 < Ri < 1 between stable (higher Ri values) and turbulent flow (lower Ri values).

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

Throughout most of the night (including the periods both before and after the onset of low clouds), the Richardson number was below 0.25 for the layers below 200 m AGL (Fig. 7d). Throughout this study, discussions of the Richardson number refer to the so-called gradient Richardson number:
e1
computed from HATPRO temperature and humidity, and UHF wind profiler data. To obtain the virtual equivalent potential temperature in Eq. (1), the pressure–height profile based on a standard atmosphere was assumed. Due to the fact that UHF winds below about 200 m AGL were rejected (cf. section 2b), changes in the Richardson number below this level are due to changes in the HATPRO temperature and humidity profiles, with the vertical shear held constant and calculated from the UHF wind at 224 m and the wind speed from the surface station. Assuming such a linear wind profile (in contrast to a log wind profile) potentially introduces some error in the estimates of the Richardson number at a given height in this near-surface layer, but the values calculated should be representative of the overall stability of the lowest 224 m at these times.

The fact that the lowest layers of the atmosphere did not get particularly stable on this night (cf. Fig. 7c) probably both reflects the fact that turbulent mixing continued throughout the night and provides an explanation for the low values of the Richardson number, especially at 0600 UTC (Fig. 7d). Overall, this night strongly resembles the patterns discussed by Schrage et al. (2007), which is not surprising as that study focused on the nights in which the stratocumulus clouds were already noted in the 0000 UTC synoptic reports in the Soudanian zone.

b. A clear night: 24–25 August 2006

Although it is true that the ceilometer at Nangatchori did detect the passage of some scattered clouds below 1500 m AGL on the night of 24–25 August 2006 (see Fig. 8a), these clouds were both too isolated and too high to be examples of the kinds of nocturnal stratocumulus decks that are the subject of the present work. In the minutes before sunrise (at approximately 0545 UTC), the ceilometer did register some lidar returns from some very low hydrometeors, but it is not clear whether these are stratocumulus clouds, simply fog, or mist. Unfortunately, the synoptic reports from nearby stations at this hour are missing, but it seems reasonable to state that long-lasting stratocumulus decks did not form that evening over Nangatchori.

Fig. 8.
Fig. 8.

As in Fig. 7, but for the period from 1600 UTC 24 Aug 2006 to 0700 UTC 25 Aug 2006.

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

The UHF wind profiler data show that the wind speeds remained quite weak throughout the night of 24–25 August (Fig. 8b). The observation is corroborated by the radiosonde wind observations from Parakou (not shown). The absence of a distinct NLLJ combined with the moderately strong stability of the near-surface layer (Fig. 8c) limited the circulation to laminar flow on this night, with the Richardson number remaining above the critical range of 1.0 around and after midnight (Fig. 8d).

c. A night with very late stratus formation: 21–22 September 2006

Very low stratocumulus clouds were observed at Nangatchori starting around 0445 UTC on the morning of 22 September 2006 (Fig. 9a). Prior to that time, the ceilometer shows that the sky had been mostly clear below 2 km. The onset of a nocturnal low level jet around 0000 UTC led to a peak wind of more than 9 m s−1 around 0315 UTC at about 225 m AGL at Nangatchori (Fig. 9b). The only available radiosonde wind profile at Parakou (viz., 2300 UTC) is in agreement that the jet had not yet become established at that early hour (not shown). From sunset until approximately 0000 UTC, the near-surface layer became increasingly stable, with lapse rates of less than −12 K km−1 detected for several hours (Fig. 9c). As soon as the NLLJ set in after midnight (Fig. 9b), the low-level temperature inversion became weaker, suggesting that the mechanical mixing caused by the jet was contributing to the destabilization of the near-surface layer. The decreased static stability and increased vertical wind shear during the second part of the night caused a substantial decrease in the Richardson number to values below the critical value of 0.25, suggesting the onset of turbulent mixing in the near-surface layer.

Fig. 9.
Fig. 9.

As in Fig. 7, but for the period from 1600 UTC 21 Sep 2006 to 0700 UTC 22 Sep 2006.

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

The example of the night 21–22 September exemplifies a time lag between the onset of the NLLJ just after midnight and the onset of the stratus some 4–5 h later. A specific feature of the present case is the low-level dryness in the afternoon hours of 21 September 2006 that persisted into the night (not shown). It was associated with low-level winds coming from northerly directions that only turned to monsoonal southwesterlies after midnight (Fig. 9b). It should be stressed that no indication was found that the northerly wind components were due to an outflow boundary from a mesoscale convective system. The dryness and clear skies might have caused the strong cooling and temperature inversion in the early night hours (Fig. 9c). The period between 0000 and 0400 UTC is then characterized by a growth of the inversion layer, which itself became less stable at the same time (Fig. 9c). The present case suggests that the strong NLLJ that only developed late in the night can overcome strong surface inversions to cause mixing and stratiform cloud formation. However, the lag between NLLJ and stratus formation might grow.

5. Composite evolution of the onset of nocturnal stratocumulus

Figure 10 depicts the composite evolution of profiles of wind speed at Nangatchori on seven nights for which UHF wind data were fully available. Here hour 0 was defined as the time of the onset of low clouds (as given in Table 1), and therefore the plot depicts the average wind speeds from 6 h before the onset of clouds to 4 h after. The contouring and shading schemes are the same as in Figs. 79, except for the Richardson number where contours instead of vertical profiles are shown in Fig. 10d. Figure 10a is a composite ceilometer plot for the seven nights that were employed in this calculation. The scattered clouds shown before time 0 are not meaningful; they are clouds with very high backscatter values in the ceilometer dataset, and even when averaged with the values of approximately zero backscatter in the other nights, the effects of these clouds are still seen. Figure 10a clearly confirms that the base of the low-level cloud deck at Nangatchori is often situated very close to the surface. The station of Parakou, only 100 km southeast of Nangatchori, is located on a hill at a comparable altitude of 391 m and reports low-level cloud heights of 200–400 m. It is not clear what causes the extremely low cloud base in the Nangatchori area.

Fig. 10.
Fig. 10.

As in Fig. 7, but as composites over the following nights: 8–9, 18–19, 19–20, 22–23 Aug; 21–22 Sep; and 5–6 and 6–7 Oct (i.e., the cloudy nights for which we have wind profiler data). Data have been composited such that hour 0 corresponds to the time of onset of low clouds at Nangatchori. (c) The vertical temperature profiles were only taken before sunrise. Therefore the x axis is truncated at 4 h after low cloud onset. (d) The contour colors merely provide visual contrast, and the dashed contour corresponds to a value of 0.25.

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

In Fig. 10b, wind speeds at Nangatchori peak between 200 and 300 m AGL about 30 min before the onset of low clouds. This wind maximum is part of a longer period of enhanced flow that began more than 2 h earlier. These results are supported by Fig. 11, which depicts time series of the mean wind speed over the 200–400-m layer with respect to the onset of low clouds, as well as some suggestion of the case-to-case variability of these data. Despite the considerable variation in the times of onset of the low clouds within this sample of nights, the results show a robust tendency for the winds to be increasing in the hours prior to the onset of cloudiness, peaking shortly before the low clouds are observed. While the onset of the NLLJ cannot be objectively defined, the composites in Figs. 10 and 11 suggest some 2–3-h time lag between the onset of strong winds in the boundary layer and the cloudiness onset.

Fig. 11.
Fig. 11.

Time series of mean wind speed for the 200–400 m AGL layer, as retrieved by the wind profiler at Nangatchori on nights when low cloud decks formed. Here the onset of low clouds (as determined by the ceilometer at Nangatchori) is defined as hour 0.

Citation: Monthly Weather Review 140, 6; 10.1175/MWR-D-11-00172.1

In Fig. 10c, the greatest static stability is detected approximately 4 h prior to the onset of stratocumulus. As winds in the nocturnal low-level jet increase over the course of the night, the composite stability slowly decreases (i.e., the temperature lapse rate increases) due to low-level mixing driven by the wind shear. In fact, by about 2 h after the onset of cloudiness, the near-surface lapse rate is nearly equal to the moist adiabatic lapse rate for this range of pressures and temperatures, suggesting that shallow moist convection is adjusting the temperature structure of the lowest few tens of meters of the atmosphere.

Figure 10d depicts the evolution of the composite Richardson number on these seven cloudy nights. Richardson numbers are very high for the first few hours of the night, but these values fall as the lower atmosphere becomes progressively less stable in the hours before low cloud formation, coinciding with an increase in the NLLJ. Near the surface, Richardson numbers decrease rapidly to below the critical threshold of 1.0 shortly before the onset of clouds, and the values decrease to less than 0.25 in the ensuing hours.

6. Summary and discussion

The present study had two major goals—an examination of some of the local factors that may control the time of onset of nocturnal stratiform clouds during nonprecipitating nights in tropical West Africa, and a documentation of some of the spatiotemporal characteristics of these clouds. With respect to the first goal, a period of destabilization was found to occur in the near-surface layer over Nangatchori (central Benin) just prior to the formation of these cloud decks. This phenomenon seems to happen on most nights in which a nocturnal low-level jet forms, producing mechanical turbulence and vertical mixing that tends to reduce the static stability of the layer between the jet and the surface. Taking place in the context of a moisture profile that in which humidity decreases with respect to height (cf. Schrage et al. 2007, their Fig. 3), the turbulent eddies transports moisture upward and contributes to the cloud formation. The near-surface moisture will likely at least in part stem from evapotranspiration from the lush vegetation that accumulates in the near-surface layer when the daytime mixing is shut off after sunset and the surface inversion forms. This NLLJ-driven destabilization naturally competes with the tendency of the near-surface layer to become increasingly stable over the course of the night by purely radiative mechanisms, suggesting that the timing and magnitude of the jet, the radiative processes, and the ensuing time scale of shear-driven mixing and destabilization are somewhat critical for the formation of nocturnal stratocumulus clouds. After all, a survey of the ceilometer data revealed that at least 26% of the nights at Nagatchori remained clear during the May–October 2006 period.

The present study suggests that the NLLJ is the major mechanism explaining stratus formation by destabilizing the near-surface layer and causing vertical mixing. This leads to the question of which mechanisms control the timing and strength of the NLLJ. Schrage et al. (2007) note an enhanced low-level meridional pressure gradient for their composite of 11 cloudy nights at Parakou during the summer monsoon 2002. These nights were also associated with enhanced frictional convergence and decreased stability in the surface layer. This may suggest that larger pressure gradient may cause a stronger NLLJ, which in turn enhances friction-related moisture convergence in the surface monsoon flow over a wider region in southern West Africa. The NLLJ-related explanation of stratus formation presented in this study does not answer the questions why (i) the NLLJ is observed on so many nights that remain clear and (ii) why the NLLJ did not form on some nonprecipitating cloudy nights. Differences in near-surface humidity seemed not be a decisive factor for cloud formation at Nangatchori. The environmental relative humidity sensor mounted at the HATPRO profiler showed that the lowest layer was saturated, even on the clear nights (not shown). However, vertical moisture profiles from HATPRO are not very accurate and moisture profiles obtained from MODEM sondes at Parakou are known to have a wet bias, especially at night (Bock et al. 2008). Thus, this study leaves the following question unanswered: what role the daytime drying of the mixing layer and existing dryer layers above the NLLJ level may have on stratus formation.

In terms of the second goal, assessment of the spatial extent of these cloud decks was performed based on eye observations from manned synoptic stations. For the study period, a clear seasonal cycle was documented, as well as a dependence on latitude. Figure 2 demonstrated that on many nights these stratiform clouds were observed at most or all of the stations in the domain, particularly for the portion of the domain with the wetter climate. The stratus cloud cover often persisted well into daylight hours, clearing off by the formation of cumuliform clouds before and around noon. For a sample of 21 cloudy and 16 clear nights at Nangatchori, cloud observation from synoptic stations show that these nights showed similar cloudiness anomalies in extensive regions of the Guineo–Soudanian zone of West Africa in the majority of the cases, indicating a nonlocal phenomenon (see Tables 1 and 2).

Given the areal extensiveness of the stratiform cloudiness, the presence or nonpresence of such low clouds will impact the moisture budget and the radiation balance, particularly after sunrise due to the high albedo of the stratus as compared to the low surface albedo of the green vegetated surface. Knippertz et al. (2011) suggest that the lack of low-level clouds over southern West Africa in almost all of their investigated 19 global climate models explains the overestimation of solar irradiance by these models when compared to the few available surface measurements in southern West Africa. Moreover, the mechanically driven low-level turbulence at night and in the early morning hours mixes moisture from the vegetation upward into the NLLJ level. It is speculated here that any misrepresentation of the degree of decoupling of the NLLJ from the surface in climate models may be a source of error in the simulated moisture budget of the West African monsoon.

Acknowledgments

This research of the first author was made possible through the financial support of the Office of the Dean of the Graduate School at Creighton University. The research of the second author was partly supported under the IMPETUS Project (BMBF Grant 01LW06001A, North Rhine-Westphalia Grant 313-21200200). We acknowledge the help of Robert Schuster in plotting Figs. 3a,b. Michael Douglas and Rahama Beida from NOAA/NSSL kindly provided Fig. 5. The microwave HATPRO profiler, the ceilometer, the UHF profiler, and surface meteorological data at Nangatchori have been obtained from the AMMA data base (http://database.amma-international.org/). We acknowledge the respective principal investigators, Bernhard Pospichal and Susanne Crewell for the HATPRO and ceilometer, Bernard Campistron for the UHF profiler, and Dominique Serça for the surface meteorological data from the Nangatchori flux station. We thank Bernhard Pospichal for helpful discussions on the use and interpretation of the ceilometer and HATPRO measurements. We are grateful to Charles Yorke from the Ghana Meteorological Agency and Francis Didé from the Beninese Weather Service who provided missing synoptic data for their respective countries. We thank the two anonymous reviewers whose comments helped to greatly improve the manuscript.

REFERENCES

  • Abdou, K., , D. J. Parker, , B. Brooks, , N. Kalthoff, , and T. Lebel, 2010: The diurnal cycle of lower boundary-layer wind in the West African monsoon. Quart. J. Roy. Meteor. Soc., 136 (Suppl.), 6676.

    • Search Google Scholar
    • Export Citation
  • Bain, C. L., , D. J. Parker, , C. M. Taylor, , L. Kergoat, , and F. Guichard, 2010: Observations of the nocturnal boundary layer associated with the West African monsoon. Mon. Wea. Rev., 138, 31423156.

    • Search Google Scholar
    • Export Citation
  • Blackadar, A. K., 1957: Boundary layer wind maxima and their significance for the growth of nocturnal inversions. Bull. Amer. Meteor. Soc., 38, 283290.

    • Search Google Scholar
    • Export Citation
  • Bock, O., and Coauthors, 2008: West African Monsoon observed with ground-based GPS receivers during African Monsoon Multidisciplinary Analysis (AMMA). J. Geophys. Res., 113, D21105, doi:10.1029/2008JD010327.

    • Search Google Scholar
    • Export Citation
  • Crewell, S., , and U. Löhnert, 2007: Accuracy of boundary layer temperature profiles retrieved with multifrequency multiangle microwave radiometry. IEEE Trans. Geosci. Remote Sens., 45, 21952201.

    • Search Google Scholar
    • Export Citation
  • Douglas, M., , R. Beida, , and A. Dominguez, 2010: Developing high spatial resolution daytime cloud climatologies for Africa. Preprints, 29th Conf. on Hurricanes and Tropical Meteorology, Tucson, AZ, Amer. Meteor. Soc., P2.23. [Available online at http://ams.confex.com/ams/pdfpapers/168263.pdf.]

  • Farquharson, J. S., 1939: The diurnal variation of wind over tropical Africa. Quart. J. Roy. Meteor. Soc., 65, 165184.

  • Hamilton, R. A., , and J. W. Archbold, 1945: Meteorology of Nigeria and adjacent territory. Quart. J. Roy. Meteor. Soc., 71, 231264.

  • Jacoby-Koaly, S., , B. Campistron, , S. Bernard, , B. Bénech, , F. Ardhuin-Girard, , J. Dessens, , E. Dupont, , and B. Carissimo, 2002: Turbulent dissipation rate in the boundary layer via UHF wind profiler Doppler spectral width measurements. Bound.-Layer Meteor., 103, 361389.

    • Search Google Scholar
    • Export Citation
  • Knippertz, P., , A. H. Fink, , R. Schuster, , J. Trentmann, , J. M. Schrage, , and C. Yorke, 2011: Ultra-low clouds over the southern West African monsoon region. Geophys. Res. Lett., 38, L21808, doi:10.1029/2011GL049278.

    • Search Google Scholar
    • Export Citation
  • Lensky, I. M., , and D. Rosenfeld, 2008: Clouds-Aerosols-Precipitation Satellite Analysis Tool (CAPSAT). Atmos. Chem. Phys., 8, 67396753, doi:10.5194/acp-8-6739-2008.

    • Search Google Scholar
    • Export Citation
  • Lothon, M., , F. Saïd, , F. Lohou, , and B. Campistron, 2008: Observation of the diurnal cycle in the low troposphere of West Africa. Mon. Wea. Rev., 136, 34773500.

    • Search Google Scholar
    • Export Citation
  • National Weather Service, 2010: Rawinsonde observations. National Weather Service Manual 10-1401, 208 pp. [Available online at http://www.weather.gov/directives/sym/pd01014001curr.pdf.]

  • Parker, D. J., and Coauthors, 2005: The diurnal cycle of the West African monsoon circulation. Quart. J. Roy. Meteor. Soc., 131, 28392860.

    • Search Google Scholar
    • Export Citation
  • Parker, D. J., and Coauthors, 2008: The AMMA radiosonde program and its implications for the future of atmospheric monitoring over Africa. Bull. Amer. Meteor. Soc., 89, 10151027.

    • Search Google Scholar
    • Export Citation
  • Pospichal, B., , and S. Crewell, 2007: Boundary layer observations in West Africa using a novel microwave radiometer. Meteor. Z., 16, 513523.

    • Search Google Scholar
    • Export Citation
  • Redelsperger, J.-L., , C. D. Thorncroft, , A. Diedhiou, , T. Lebel, , D. J. Parker, , and J. Polcher, 2006: African Monsoon Multidisciplinary Analysis: An international research project and field campaign. Bull. Amer. Meteor. Soc., 87, 17391746.

    • Search Google Scholar
    • Export Citation
  • Schrage, J. M., , S. Augustyn, , and A. H. Fink, 2007: Nocturnal stratiform cloudiness during the West African monsoon. Meteor. Atmos. Phys., 95, 7386.

    • Search Google Scholar
    • Export Citation
  • Stephens, G., and Coauthors, 2002: The CloudSat mission and the A-train. Bull. Amer. Meteor. Soc., 83, 17711790.

  • WMO, 1995: Alphanumeric codes. Manual on codes, Vol. I.1, Part A, WMO Publ. 306.

Save