• Bister, M., , and K. A. Emanuel, 1997: The genesis of Hurricane Guillermo: TEXMEX analysis and a modeling study. Mon. Wea. Rev., 125 , 26622682.

    • Search Google Scholar
    • Export Citation
  • Bosart, L. F., , and J. A. Bartlo, 1991: Tropical storm formation in a baroclinic environment. Mon. Wea. Rev., 119 , 19792013.

  • Bracken, W. E., , and L. F. Bosart, 2000: The role of synoptic-scale flow during tropical cyclogenesis over the North Atlantic Ocean. Mon. Wea. Rev., 128 , 353376.

    • Search Google Scholar
    • Export Citation
  • Briegel, L. M., , and W. M. Frank, 1997: Large-scale influences on tropical cyclogenesis in the western North Pacific. Mon. Wea. Rev., 125 , 13971413.

    • Search Google Scholar
    • Export Citation
  • Burpee, R. W., 1972: The origin and structure of easterly waves in the lower troposphere of North Africa. J. Atmos. Sci., 29 , 7790.

  • Carlson, T. N., , and J. M. Prospero, 1972: The large-scale movement of Saharan air outbreaks over the northern equatorial Atlantic. J. Appl. Meteor., 11 , 283297.

    • Search Google Scholar
    • Export Citation
  • Carlson, T. N., , and S. G. Benjamin, 1980: Radiative heating rates of Saharan dust. J. Atmos. Sci., 37 , 193213.

  • Chen, F., , and J. Dudhia, 2001: Coupling an advanced land surface–hydrology model with the Penn State–NCAR MM5 modeling system. Part I: Model implementation and sensitivity. Mon. Wea. Rev., 129 , 569585.

    • Search Google Scholar
    • Export Citation
  • Chen, S-H., , and W-Y. Sun, 2002: A one-dimensional time dependent cloud model. J. Meteor. Soc. Japan, 80 , 99118.

  • Colón, J. A., , and W. R. Nightingale, 1963: Development of tropical cyclones in relation to circulation patterns at the 200-millibar level. Mon. Wea. Rev., 91 , 329336.

    • Search Google Scholar
    • Export Citation
  • Cook, K. H., 1999: Generation of the African easterly jet and its role in determining West African precipitation. J. Climate, 12 , 11651184.

    • Search Google Scholar
    • Export Citation
  • DeMaria, M., , J. A. Knaff, , and B. H. Connell, 2001: A tropical cyclone genesis parameter for the tropical Atlantic. Wea. Forecasting, 16 , 219233.

    • Search Google Scholar
    • Export Citation
  • Dunion, J. P., , and C. S. Velden, 2004: The impact of the Saharan air layer on Atlantic tropical cyclone activity. Bull. Amer. Meteor. Soc., 85 , 353365.

    • Search Google Scholar
    • Export Citation
  • Dunkerton, T. J., , M. T. Montgomery, , and Z. Wang, 2008: Tropical cyclogenesis in a tropical wave critical layer: Easterly waves. Atmos. Chem. Phys. Discuss., 8 , 1114911292.

    • Search Google Scholar
    • Export Citation
  • Emanuel, K. A., 1989: The finite amplitude nature of tropical cyclogenesis. J. Atmos. Phys., 46 , 34313456.

  • Fett, R. W., 1966: Upper-level structure of the formative tropical cyclone. Mon. Wea. Rev., 94 , 918.

  • Frank, W. M., 1988: Tropical cyclone formation. A Global View of Tropical Cyclones, R. L. Elsberry, Ed., Office of Naval Research, 53–90.

    • Search Google Scholar
    • Export Citation
  • Frank, W. M., , and P. E. Roundy, 2006: The role of tropical waves in tropical cyclogenesis. Mon. Wea. Rev., 134 , 23972417.

  • Fujita, T., , K. Watanabe, , and T. Izawa, 1969: Formation and structure of equatorial anticyclones caused by large-scale cross-equatorial flow determined by ATS-1 photographs. University of Chicago Research Rep. SMRP-78, 37 pp.

    • Search Google Scholar
    • Export Citation
  • Gallina, G. M., , and C. S. Velden, 2002: Environmental vertical wind shear and tropical cyclone intensity change utilizing enhanced satellite derived wind information. Preprints, 25th Conf. on Hurricanes and Tropical Meteorology, San Diego, CA, Amer. Meteor. Soc., 172–173.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., 1968: Global view of the origin of tropical disturbances and storms. Mon. Wea. Rev., 96 , 669700.

  • Gray, W. M., 1988: Environmental influences on tropical cyclones. Aust. Meteor. Mag., 36 , 127139.

  • Gray, W. M., 1998: The formation of tropical cyclones. Meteor. Atmos. Phys., 67 , 3769.

  • 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
  • Holland, G. J., 1995: Scale interaction in the western Pacific monsoon. Meteor. Atmos. Phys., 56 , 5279.

  • Jenkins, G. S., , A. S. Pratt, , and A. Heymsfield, 2008: Possible linkages between Saharan dust and tropical cyclone rain band invigoration in the eastern Atlantic during NAMMA-06. Geophys. Res. Lett., 35 , LO8815. doi:10.1029/2008GL034072.

    • Search Google Scholar
    • Export Citation
  • Kain, J. S., , and J. M. Fritsch, 1990: A one-dimensional entraining/detraining plume model and its application in convective parameterization. J. Atmos. Sci., 47 , 27842802.

    • Search Google Scholar
    • Export Citation
  • Kain, J. S., , and J. M. Fritsch, 1993: Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 165–170.

    • Search Google Scholar
    • Export Citation
  • Lee, C-S., , R. Edson, , and W. M. Gray, 1989: Some large-scale characteristics associated with tropical cyclone development in the north Indian Ocean during FGGE. Mon. Wea. Rev., 117 , 407426.

    • Search Google Scholar
    • Export Citation
  • Miyakoda, K., , J. Sheldon, , and J. Sirutis, 1982: Four-dimensional analysis experiment during the gate period. Part II. J. Atmos. Sci., 39 , 486506.

    • Search Google Scholar
    • Export Citation
  • Molinari, J., , S. Skubis, , and D. Vollaro, 1995: External influences on hurricane intensity. Part III: Potential vorticity structure. J. Atmos. Sci., 52 , 35933606.

    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., , M. E. Nicholls, , T. A. Cram, , and A. B. Saunders, 2006: A vertical hot tower route to tropical cyclogenesis. J. Atmos. Sci., 63 , 355386.

    • Search Google Scholar
    • Export Citation
  • Palmen, E., 1948: On the formation and structure of tropical cyclones. Geophysics, 3 , 2638.

  • Powell, M. D., 1990: Boundary layer structure and dynamics in outer hurricane rainbands. Part II: Downdraft modification and mixed-layer recovery. Mon. Wea. Rev., 118 , 918938.

    • Search Google Scholar
    • Export Citation
  • Pytharoulis, I., , and C. Thorncroft, 1999: The low-level structure of African easterly waves in 1995. Mon. Wea. Rev., 127 , 22662280.

  • Ramage, C. S., 1959: Hurricane development. J. Meteor., 16 , 227237.

  • Reasor, P. D., , M. T. Montgomery, , and L. F. Bosart, 2005: Mesoscale observations of the genesis of Hurricane Dolly (1996). J. Atmos. Sci., 62 , 31513171.

    • Search Google Scholar
    • Export Citation
  • Reed, R. J., , D. C. Norquist, , and E. E. Recker, 1977: The structure and properties of African disturbances as observed during phase III of GATE. Mon. Wea. Rev., 105 , 317333.

    • Search Google Scholar
    • Export Citation
  • Reed, R. J., , A. Hollingsworth, , W. A. Heckley, , and F. Delsol, 1988a: An evaluation of the ECMWF operational system in analyzing and forecasting easterly wave disturbances over Africa and the tropical Atlantic. Mon. Wea. Rev., 116 , 824865.

    • Search Google Scholar
    • Export Citation
  • Reed, R. J., , E. Klinker, , and A. Hollingsworth, 1988b: The structure and characteristics of African easterly wave disturbances as determined from the ECMWF operational analysis/forecast system. Meteor. Atmos. Phys., 38 , 2233.

    • Search Google Scholar
    • Export Citation
  • Simpson, J., , E. Ritchie, , G. J. Holland, , J. Halverson, , and S. Stewart, 1997: Mesoscale interactions in tropical cyclone genesis. Mon. Wea. Rev., 125 , 26432661.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., , J. B. Klemp, , J. Dudhia, , D. O. Gill, , D. M. Barker, , W. Wang, , and J. G. Powers, 2005: A description of the advanced research WRF version 2. NCAR Tech. Note NCAR/TN-408+STR, 88 pp.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C., , and K. Hodges, 2001: African easterly wave variability and its relationship to Atlantic tropical cyclone activity. J. Climate, 14 , 11661179.

    • Search Google Scholar
    • Export Citation
  • Wu, L., 2007: Impact of Saharan air layer on hurricane peak intensity. Geophys. Res. Lett., 34 , L09802. doi:10.1029/2007GL029564.

  • Zehr, R. M., 1991: Observational investigation of tropical cyclogenesis in the western North Pacific. Preprints, 19th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., 235–240.

    • Search Google Scholar
    • Export Citation
  • Zehr, R. M., 1992: Tropical cyclogenesis in the western North Pacific. NOAA Tech. Rep. NESDIS 61, 181 pp.

  • Zipser, E. J., and Coauthors, 2009: The Saharan air layer and the fate of African easterly waves: The NAMMA field program. Bull. Amer. Meteor. Soc., 90 , 11371156.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Model domain utilized for this study. Shading over land denotes topography in meters as given by the shading bar. Contours over the ocean denote the NASA TMI SSTs (°C) from 18 Aug 2006. SST values greater than 26°C are shaded in light gray. The box denotes the longitudinal constraints selected for formulation of temporal composites.

  • View in gallery

    Debby wave vortex mean sea level pressure (SLP) every 6 h between 1800 UTC 20 Aug and 0000 UTC 24 Aug 2006. Solid bold line denotes WRF model simulation analyzed further in this paper; the long-dashed (short-dashed) bold line is the ECMWF operational reanalysis (NHC observed intensity). The thin solid lines denote simulation results of other physical parameterization combinations defined in the text.

  • View in gallery

    Pre-Ernesto wave vortex (a) mean sea level pressure (hPa), (b) 850-hPa height (m), and (c) 700-hPa height (m) every 6 h between 1800 UTC 18 Aug and 0000 UTC 20 Aug 2006. Solid line denotes WRF model simulation; the dashed line is the ECMWF operational reanalysis.

  • View in gallery

    Pre-Ernesto wave vortex center positions at sea level, 850 hPa, and 700 hPa every 6 h between 1800 UTC 18 Aug and 0000 UTC 20 Aug 2006 from the (a) ECMWF operational reanalysis and (b) model simulation.

  • View in gallery

    Debby wave vortex (a) 850- and (b) 700-hPa heights (m) every 6 h between 1800 UTC 20 Aug and 0000 UTC 24 Aug 2006. Solid line denotes WRF model simulation; the long-dashed line is the ECMWF operational reanalysis.

  • View in gallery

    As in Fig. 4 but for Tropical Storm Debby between 1800 UTC 20 Aug and 0000 UTC 23 Aug 2006.

  • View in gallery

    Averaged precipitation rates (mm day−1), 700-hPa winds (m s−1), and geopotential heights (m) for the pre-Ernesto wave from (a) WRF simulation and (b) ECMWF operational reanalysis and TRMM 3B43 rainfall rates averaged from 1200 UTC 18 Aug to 0000 UTC 20 Aug 2006. (c),(d) As in (a),(b) but for the Debby wave averaged from 1200 UTC 20 Aug to 1800 UTC 22 Aug 2006. Shading denotes rainfall rates according to the grayscale bar.

  • View in gallery

    The 600-hPa zonal wind component (m s−1) at Praia, Cape Verde (14.9°N, 23.5°W), from the pre-Ernesto wave simulation (solid line), the Debby wave simulation (long-dashed line), and the 700-hPa zonal wind component from the ECMWF operational reanalysis (short-dashed lines). Crosshair marks are observed 600-hPa winds from NASA AMMA radiosonde measurements.

  • View in gallery

    WRF pre-Ernesto wave 200–850-hPa vertical wind shear (m s−1) for (a) 1200 UTC 18 Aug and (b) 1200 UTC 19 Aug 2006. Additionally, 600–925-hPa vertical wind shear for (c) 1200 UTC 18 Aug and (d) 1200 UTC 19 Aug 2006. Values greater than 15 m s−1 are shaded. Bolded mark denotes the location of the 700-hPa vortex. Vectors are in m s−1.

  • View in gallery

    WRF Debby wave 200–850-hPa vertical wind shear (m s−1) for (a) 1200 UTC 20 Aug and (b) 1200 UTC 21 Aug 2006. Additionally, 600–925-hPa vertical wind shear for (c) 1200 UTC 20 Aug and (d) 1200 UTC 21 Aug 2006. Values greater than 15 m s−1 are shaded; bolded mark denotes the location of the 700-hPa vortex; vectors in m s−1.

  • View in gallery

    Vertical cross section at 20°W of the zonal wind (shaded; m s−1) and the meridional and vertical circulation (vectors and streamlines) for (a) pre-Ernesto and (b) Debby wave WRF simulation composites.

  • View in gallery

    WRF pre-Ernesto wave simulation composite geopotential heights (m) and winds (m s−1) at (a) 850 hPa, (b) 700, (c) 500, and (d) 200 hPa.

  • View in gallery

    WRF Debby wave simulation composite geopotential heights (m) and winds (m s−1) at (a) 850, (b) 700 hPa, (c) 500 hPa, and (d) 200 hPa.

  • View in gallery

    WRF 925-hPa streamlines for the pre-Ernesto wave simulation at 0300 UTC 18 Aug 2006.

  • View in gallery

    As in Fig. 14 but for the Debby wave simulation at (a) 0000 and (b) 1600 UTC 20 Aug 2006.

  • View in gallery

    Vertical cross section of relative vorticity (×10−5 s−1) at 20°W for (a) pre-Ernesto and (b) Debby wave WRF simulation composites and (c) the Debby minus pre-Ernesto difference. Shading denotes negative values.

  • View in gallery

    Cross section of relative humidity (%) at 20°W for the (a) Ernesto and (b) Debby wave WRF composites and (c) the Debby minus Ernesto RH difference. Shading in (c) denotes negative values.

  • View in gallery

    Area-averaged (10°–15°N, 15°–25°W) Debby wave minus pre-Ernesto wave WRF composite difference of MSE (bold solid line), sensible energy (bold dashed line), latent energy (dotted line), and geopotential energy (long/short-dashed line) from the WRF model. Units are all 104 J kg−1.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 40 40 3
PDF Downloads 21 21 2

Tropical Storm Development from African Easterly Waves in the Eastern Atlantic: A Comparison of Two Successive Waves Using a Regional Model as Part of NASA AMMA 2006

View More View Less
  • 1 Institute for Geophysics, The University of Texas at Austin, Austin, Texas
  • | 2 Jackson School of Geosciences, The University of Texas at Austin, Austin, Texas
© Get Permissions
Full access

Abstract

Two successive African easterly waves (AEWs) from August 2006 are analyzed utilizing observational data, the European Centre for Medium-Range Weather Forecasts reanalysis, and output from the National Center for Atmospheric Research–National Oceanic and Atmospheric Administration Weather Research and Forecasting model (WRF) to understand why the first wave does not develop over the eastern Atlantic while the second wave does. The first AEW eventually forms Hurricane Ernesto over the Caribbean Sea, but genesis does not occur over the eastern Atlantic. The second wave, although weaker than the first over land, leaves the West African coast and quickly intensifies into Tropical Storm Debby west of the Cape Verde islands. This study shows that the environmental conditions associated with the first AEW’s passage inhibited development. These conditions include strong low- and midtropospheric vertical wind shear owing to a stronger than normal African easterly jet, lower than normal relative humidity, and increased atmospheric stability. All of these are characteristics of an intensification of the Saharan air layer (SAL), or SAL outbreak, over the eastern Atlantic. The environmental conditions were more favorable for genesis 2½ days later when the second wave left the African coast. Additionally, a strong low-level southwesterly surge develops over the eastern North Atlantic in the wake of the passage of the first wave. This westerly surge is associated with an enhancement of the low-level westerly flow, low-level cyclonic vorticity, large-scale low-level wind convergence, and vertical motion conducive for development over the region. While the initial westerly surge is likely associated with the passage of the first wave, over time (i.e., by 1600 UTC 20 August 2006) the development of the second wave becomes influential in maintaining the low-level westerly surge. Although SAL outbreaks are also associated with the addition of dust, the different cyclogenesis histories of the two systems are simulated without including dust in the regional model.

Corresponding author address: Edward K. Vizy, Institute for Geophysics, The University of Texas at Austin, J. J. Pickle Research Campus, Bldg. 196, 10100 Burnet Road (R2200), Austin, TX 78758-4445. Email: ned@ig.utexas.edu

This article included in the TCSP NAMMA special collection.

Abstract

Two successive African easterly waves (AEWs) from August 2006 are analyzed utilizing observational data, the European Centre for Medium-Range Weather Forecasts reanalysis, and output from the National Center for Atmospheric Research–National Oceanic and Atmospheric Administration Weather Research and Forecasting model (WRF) to understand why the first wave does not develop over the eastern Atlantic while the second wave does. The first AEW eventually forms Hurricane Ernesto over the Caribbean Sea, but genesis does not occur over the eastern Atlantic. The second wave, although weaker than the first over land, leaves the West African coast and quickly intensifies into Tropical Storm Debby west of the Cape Verde islands. This study shows that the environmental conditions associated with the first AEW’s passage inhibited development. These conditions include strong low- and midtropospheric vertical wind shear owing to a stronger than normal African easterly jet, lower than normal relative humidity, and increased atmospheric stability. All of these are characteristics of an intensification of the Saharan air layer (SAL), or SAL outbreak, over the eastern Atlantic. The environmental conditions were more favorable for genesis 2½ days later when the second wave left the African coast. Additionally, a strong low-level southwesterly surge develops over the eastern North Atlantic in the wake of the passage of the first wave. This westerly surge is associated with an enhancement of the low-level westerly flow, low-level cyclonic vorticity, large-scale low-level wind convergence, and vertical motion conducive for development over the region. While the initial westerly surge is likely associated with the passage of the first wave, over time (i.e., by 1600 UTC 20 August 2006) the development of the second wave becomes influential in maintaining the low-level westerly surge. Although SAL outbreaks are also associated with the addition of dust, the different cyclogenesis histories of the two systems are simulated without including dust in the regional model.

Corresponding author address: Edward K. Vizy, Institute for Geophysics, The University of Texas at Austin, J. J. Pickle Research Campus, Bldg. 196, 10100 Burnet Road (R2200), Austin, TX 78758-4445. Email: ned@ig.utexas.edu

This article included in the TCSP NAMMA special collection.

1. Introduction

During the period from 18 to 23 August 2006, two African easterly waves (AEWs) moved westward off the West African coast over the eastern Atlantic. The first wave, which originated over central Africa approximately 4 days earlier, left the coast on 18 August 2006 as a relatively strong wave but did not develop into Hurricane Ernesto until reaching the windward islands of the Caribbean on 24 August 2006. The second wave, which originated over West Africa as a weak wave, left the West African coast 2½ days later on 20 August 2006 following a strong convective flare up near the coastline (Zipser et al. 2009) and quickly developed into Tropical Storm Debby over the Cape Verde region. The purpose of this paper is to explain why one AEW developed over the eastern Atlantic, whereas the other did not, by understanding the differences in the synoptic-scale environmental conditions associated with the two waves. To do so, we utilize high-resolution regional modeling results from the National Center for Atmospheric Research–National Oceanic and Atmospheric Administration (NCAR–NOAA) Advanced Research Weather Research and Forecasting model (ARW-WRF) and National Aeronautics and Space Administration observations taken in support of the African Monsoon Multidisciplinary Analyses (AMMA) field campaign.

Background on the conditions necessary for tropical cyclogenesis is reviewed in section 2. Section 3 describes the regional model and the modeling approach. In section 4, the regional model results are validated with the European Center for Medium-Range Weather Forecasts (ECMWF) reanalysis, NASA Tropical Rainfall Measuring Mission (TRMM) rainfall estimates, and NASA AMMA dropsonde and radiosonde data; the results are presented in section 5. Conclusions are in section 6.

2. Background

Previous research on tropical cyclone development over the past 50 years has identified some necessary environmental conditions for tropical cyclogenesis. One condition relates to the underlying ocean temperature. Studies (e.g., Palmen 1948; Gray 1968, 1988) suggest that average SSTs in the ocean mixed layer of at least 26°C are required for development of deep convection associated with tropical cyclones to occur.

Another condition needed for tropical development is a synoptic low-level cyclonic vorticity maximum stronger than the background planetary vorticity (e.g., Gray 1968, 1988). Off the West African coast such a cyclonic relative vorticity maximum can be provided by AEWs (e.g., Reed et al. 1977) or in some instances by squall lines propagating westward via the formation of mesoscale convective vortices (e.g., Bister and Emanuel 1997; Simpson et al. 1997; Montgomery et al. 2006; Dunkerton et al. 2008). There also needs to be sufficiently large background planetary vorticity (i.e., greater than 1.3 × 10−5 s−1—the value of the Coriolis parameter 5° off the equator) for tropical cyclogenesis.

Tropical cyclone development is favored by weak vertical wind shear since strong vertical shear is found to be associated with the low-level circulation/vortex centers lagging behind the mid- and upper-tropospheric vortex centers (Dunion and Velden 2004). This is especially relevant over the eastern North Atlantic where tropical cyclones often interact with the warm Saharan air layer (SAL) to the north. Vertical shear is often measured over the 200–850-hPa layer; although there is no definitive threshold value defining “weak,” it is generally thought that tropical development is unlikely when the shear is greater than 15 m s−1 (e.g., Gray 1968; DeMaria et al. 2001). Other studies (e.g., Zehr 1991, 1992; Gallina and Velden 2002) suggest lower threshold values ranging between 7.5 and 15 m s−1.

Over the eastern North Atlantic off the West African coast, strong lower-to-middle tropospheric vertical shear exists in association with the midlevel African easterly jet (AEJ). The AEJ develops over western Africa as the result of strong meridional soil moisture gradients that lead to strong positive meridional temperature gradients at the surface and in the lower troposphere. The positive meridional temperature gradient at the surface induces easterly flow over the surface monsoon westerlies, including at the level of the AEJ (Cook 1999).

The AEJ can extend westward over the adjacent eastern North Atlantic and intensify, especially during a SAL outbreak. The SAL is a layer of warm, dry Saharan air and dust that advances westward from West Africa over the eastern North Atlantic Ocean, generally between 500 and 925 hPa and from 500 to 850 hPa in the central and western North Atlantic (Carlson and Prospero 1972). Over the ocean, the SAL overlies the cooler, moist marine surface air, resulting in the formation of a strong temperature inversion at its base (Carlson and Benjamin 1980; Dunion and Velden 2004). Warm temperatures above the inversion base are maintained by the absorption of solar radiation by the suspended dust (Carlson and Benjamin 1980) as well as subsidence. By this mechanism, these intensifications of the SAL, or SAL outbreaks, can be maintained across the North Atlantic, reaching as far as the western Caribbean Sea, approximately 7000 km from the African coast (Dunion and Velden 2004).

Some studies (e.g., Ramage 1959; Colón and Nightingale 1963; Fett 1966; Fujita et al. 1969; Gray 1988, 1998; Lee et al. 1989; Bosart and Bartlo 1991; Zehr 1991; Molinari et al. 1995) suggest the need for an external influence acting upon the developing disturbance to promote genesis. Frank (1988) and Bracken and Bosart (2000) proposed that tropical cyclone development consists of two phases, genesis and intensification, each of which is dominated by different physical processes. In the genesis phase the mesoscale vortex forms rapidly within a region of loosely organized tropical convection, and external forcing may be required to produce deep convection (Briegel and Frank 1997). In the intensification phase the developing vortex reaches a threshold intensity and becomes capable of intensifying into a tropical cyclone due to its own interactions with the ocean without further external forcing. Although external forcing (e.g., SAL outbreaks) may still influence the intensification process (Wu 2007), the vortex is capable of self-intensification (Briegel and Frank 1997).

Another factor needed for tropical cyclogenesis is high values of relative humidity (RH) in the lower to middle troposphere to inhibit the negative effects of evaporatively driven downdrafts (Bister and Emanuel 1997). Evaporatively driven downdrafts are weakened by the entrainment of high RH air, allowing for the development of deep convection (Bister and Emanuel 1997). If low RH environmental air is entrained instead, such as in the case of a SAL outbreak (e.g., Dunion and Velden 2004), convection can be suppressed through the enhancement of evaporatively driven downdrafts (Emanuel 1989; Powell 1990; Dunion and Velden 2004).

3. Model description and experimental design

The model used to conduct the simulations for this study is the NCAR–NOAA WRF model (Skamarock et al. 2005). It is a fully compressible, nonhydrostatic model whose governing equations are cast in flux form and solved using a time-split integration scheme that accounts for both low and high frequency modes. The vertical coordinate system is terrain following. Here, we use 31 vertical levels, with the top of the atmosphere set at 50 hPa, and a 30-km grid spacing with a model time step of 1 min. The domain is large (Fig. 1) to encompass much of northern and tropical Africa and the Atlantic Ocean.

The model is run in synoptic mode with initial and lateral boundary conditions for temperature, horizontal wind, geopotential height, relative humidity, land surface temperature, and soil moisture taken from the 6-hourly 1.125° ECMWF operational reanalysis. SSTs are prescribed and updated every 6 h as derived from the NASA TRMM Microwave Imager (TMI) 0.25°-resolution 3-day running mean SSTs product. The 3-day running mean product was chosen to minimize missing SSTs since the TMI has a fairly narrow swath over the tropics. For any given day, each daily 3-day running mean SST field is assumed to be the 1200 UTC field, and 6-hourly values are then linearly interpolated between the successive 1200 UTC values. Missing SSTs within the domain are interpolated from the nearest SST points where data is available. Figure 1 shows the SSTs averaged between 0300 UTC 18 August and 1800 UTC 19 August 2006, with values greater than 26°C shaded.

Two WRF simulations are discussed in detail. The Ernesto wave simulation is initialized on 0000 UTC 18 August and runs until 0000 UTC 24 August 2006, and the Debby wave simulation is initialized on 0000 UTC 20 August and runs out through 0000 UTC 25 August 2006. Our initial intent was to have one simulation encompass the passage of both easterly waves through the Cape Verde region; however, in a simulation that began on 18 August 2006 and ran through 25 August 2006, the second AEW (e.g., the wave associated with Debby) was not accurately simulated. A second wave is simulated, but closer examination showed that it moved off the West African coast approximately 18 h after Debby was observed to cross the West African coastline; consequently, the simulated wave did not intensify into a tropical storm. This result is not too surprising owing to the large domain (Fig. 1). Away from the lateral edges, synoptic variations are generated within the model domain that are essentially independent of the initial conditions after a few days. By running a second simulation initialized on 0000 UTC 20 August 2006 for the Debby wave, we are able to circumvent this issue and accurately simulate the development of Debby over the eastern North Atlantic.

Extensive testing and validation were used to select the physical parameterizations that provide an optimal simulation for the region. In this process, a series of model simulations were run for each case study with various parameterization combinations. Physical parameterizations tested include the new Kain–Fritsch (KF) (Kain and Fritsch 1990, 1993) and Betts–Miller–Janjic (BMJ) cumulus schemes; the Purdue–Lin (LIN) (Chen and Sun 2002), WRF single-moment 3-class (WSM-3), and WRF single moment 5-class (WSM-5) microphysics schemes; and the NCAR Community Atmosphere Model (CAM) and Rapid Radiative Transfer model (RRTM) radiation schemes. The simulated position and magnitude of the vortex center at the surface, 850 hPa, and 700 hPa is then compared with the ECMWF operational reanalysis and—in the case for the second wave—National Hurricane Center (NHC) observations to evaluate the simulation performance. Although all six of the simulations were reasonably able to produce the storm track, Fig. 2 shows that only one parameterization combination realistically captured the intensification for the second wave, Tropical Storm Debby, so those parameters were selected for use here. Parameterizations chosen include the Purdue–Lin microphysics scheme, new Kain–Fritsch cumulus convection scheme; Monin–Obukhov surface layer scheme; unified National Centers for Environmental Prediction (NCEP), Oregon State, Air Force, Hydrologic Research Laboratory of the National Weather Service (NOAH) land surface model (Chen and Dudhia 2001); the NCAR CAM radiation parameterization; and the Yonsei University planetary boundary layer scheme. Note that there is currently no parameterization that accounts for the influence of the variability of aerosol forcing in the model.

4. Validation of the WRF model

Surface, 850-hPa, and 700-hPa vortex center intensity and track information from the simulated and ECMWF operational reanalysis pre-Ernesto AEW are shown in Figs. 3 and 4, respectively. Overall, WRF simulates the intensity of the vortex center at these three levels reasonably well compared to the ECMWF reanalysis, with differences of ±3 hPa at the surface to −5 to +20 m at 850 and 700 hPa. Although the 850- and 700-hPa vortex center heights are generally higher in magnitude than in the lower-resolution reanalysis, WRF does capture the appropriate direction of the height change between each 6-h interval.

Note that both the reanalysis and WRF simulation reveal an equatorward tilt with height of the vortex center (Fig. 4), which is particularly pronounced from 1800 UTC 18 August to 0600 UTC 19 August 2006. After this time, the 850- and 700-hPa vortex centers begin to align vertically. Other studies (e.g., Pytharoulis and Thorncroft 1999; Thorncroft and Hodges 2001) have observed a vertical tilt of 10°–15° latitude of the vortex column for AEWs over continental Africa and their subsequent alignment over the eastern North Atlantic Ocean. The alignment of the vortex centers occurs sooner in the WRF model, because the 700-hPa vortex starts to propagate northwestward about 12 h earlier than in the reanalysis, whereas the surface vortex center propagates equatorward faster between 1200 and 1800 UTC 19 August 2006.

Analysis of the vortex intensity and track information for the Debby wave (Figs. 2, 5, 6) indicates that, when compared to the ECMWF reanalysis, WRF reasonably simulates the intensity of the vortex centers through 1200 UTC 22 August. The intensity differences between the simulation and reanalysis increase after 1200 UTC 22 August 2006, with surface, 850-hPa, and 700-hPa differences on 0000 UTC 23 August 2006 of −7 hPa, −55 m, and −50 m, respectively. NHC observational data (Fig. 2) reveals that WRF simulates the pressure drop associated with the development of Debby more realistically than the ECMWF reanalysis. While this discrepancy is not too surprising, as it may be an artifact of the coarseness of the resolution of the ECMWF reanalysis (1.125° resolution) owing to its global coverage, this difference emphasizes the value in using WRF as opposed to the ECMWF reanalysis to conduct this study.

Upon leaving the African coast, the WRF simulated vortex centers for the Debby wave (Fig. 6) quickly align along 13°N. The storm system at all levels moves approximately 30%–40% slower than the Ernesto wave. The vortex centers also align quickly in the ECMWF operational reanalysis, in the same general location and at approximately the same time as in the simulations, but the simulated Debby vortex propagates approximately 10%–20% slower to the northwest than the reanalyzed vortex.

Simulated storm track and intensity errors are calculated by comparing the Debby simulation with the NHC observed track data between 1800 UTC 21 August (f42) and 0000 UTC 24 August 2006 (f96) and are listed in Table 1. Track errors between 1800 UTC 21 August (f42) and 0600 UTC 22 August 2006 (f54) range between 96 and 163 km, which is less than the NHC storm-specific track data errors for f36 (135 km) and about the same as the errors for f48 (161 km). Track errors increase after 0600 UTC 22 August 2006, with magnitudes approximately 80 km greater than the NHC track error of 206 km for the f72 forecast. Similarly, simulation intensity errors for Tropical Storm Debby are generally lower than the NHC storm-specific intensity errors, which are 3.9, 5.4, 9.2, and 15.4 m s−1 for f36, f48, f72, and f96, respectively. The NHC report indicates that Debby achieves tropical storm strength around 0000 UTC 23 August 2006. The simulated surface winds reach tropical storm strength by 0100 UTC 23 August 2006 in the WRF simulation.

Figure 7a displays the simulated WRF rainfall rate and 700-hPa circulation for the Ernesto wave averaged between 1200 UTC 18 August and 0000 UTC 20 August 2006, whereas Fig. 5b shows the NASA TRMM estimated rainfall and 700-hPa circulation from the ECMWF operational reanalysis over the same averaging period. The WRF model simulates a rainfall band just south of 10°N between 15° and 35°W, which is approximately 2°–3° latitude farther north than observed in TRMM. East of 15°W the simulated rainfall band is much stronger and is located just south of 10°N as opposed to north of 10°N in TRMM. Generally, WRF has a wet bias with the intensity of the rainfall maxima being approximately twice as large as TRMM.

The reanalyzed and simulated 700-hPa flows (Figs. 7a,b) are very similar, with strong 700 hPa easterly flow from the West African coast to approximately 30°W and a broad region of diffluence between 25° and 35°W. The vortex associated with the Ernesto wave is at approximately 8°N, 23°W in both the ECMWF reanalysis and the simulation.

Rainfall and 700-hPa circulation fields for the Debby simulation are shown in Fig. 7c averaged from 1200 UTC 20 August to 1800 UTC 22 August 2006. A band of strong convection is centered just south of 9°N, with a well-defined cyclonic center at 11°N, 22°W. Easterly flow north (south) of this rain belt is strong (weak).

Figure 7d displays the corresponding NASA TRMM 3B43 estimated rainfall and 700-hPa circulation from the ECMWF reanalysis over the same time-averaging period. Similar to the simulation (Fig. 7c), a region of organized strong convection is located over the Cape Verde region centered at 9°N, 23°W. Over and to the west of this vortex, rainfall rates are lower and less concentrated than those simulated by WRF. Again, WRF has a wet bias in the regions where the strongest convection is located.

To evaluate WRF’s ability to capture the atmospheric moisture field, simulation results are compared to NASA AMMA radiosonde measurements from Kawsara, Senegal (14.7°N, 17.1°W), and Praia, Cape Verde Islands (14.9°N, 23.5°W), and dropsonde measurements from aircraft missions. A comparison between the WRF simulated and observed precipitable water (PW) values is given in Table 2. Overall, total water vapor amounts in WRF are approximately 10% lower than the observed values. The differences for individual sites range from near zero up to 43.22%. Note that the dropsonde measurements have different launch altitudes, so because of this the values shown in Table 2 may underestimate precipitable water to some degree. However, the WRF precipitable water amounts are still lower than the observations. The ECMWF operational reanalysis has a similar, but more pronounced, dry bias (not shown). Although WRF is able to simulate the basic profile of temperature, dewpoint temperature, and RH as in the sounding profile, it has difficulty capturing small-scale structures (e.g., moist/dry intrusions), and this failure accounts for most of the precipitable water differences. It is unclear whether increasing the WRF’s vertical resolution and/or the vertical resolution of the initialization data could improve this difference.

Another important feature of the large-scale flow over the eastern North Atlantic is the midtropospheric AEJ. Figure 8 shows a comparison of the simulated and observed 600-hPa zonal wind and the 700-hPa zonal wind from the ECMWF reanalysis at Praia, Cape Verde Islands. Note a limitation of the ECMWF operational reanalysis is that 600 hPa is not a standard output level, but this is closest to the level where the AEJ is usually the strongest (e.g., Burpee 1972; Cook 1999). Overall, both WRF simulations capture the fluctuations in the zonal flow compared to the observations but have difficulties simulating the magnitude of the AEJ during periods when the easterly flow is particularly strong. During these strong easterly flow periods, the simulated jet is 3–12 m s−1 weaker than in the observations. At other times, there is good agreement (i.e., less than 2 m s−1 difference) between simulated and observed easterly flow. The ECMWF reanalysis has similar issues, including a particularly large difference (i.e., 8–12 m s−1 weaker than the observed zonal flow) on 23 August 2006. Much of this discrepancy may be associated with the difficulties of the reanalysis in capturing the development/intensification of Tropical Storm Debby (see Figs. 2, 5).

Results from Figs. 2 –8 suggest that the WRF simulations are capturing the AEW movement, large-scale circulation, and convection for the two wave cases. Furthermore, WRF does reproduce the observed nondevelopment of the Ernesto wave and development of the Debby wave in the eastern North Atlantic. In the next section, we will use these two WRF simulations to examine why the first AEW does not develop whereas the second AEW does.

5. Results

To characterize environmental conditions and AEW structures, a temporal compositing technique is used in addition with synoptic analysis. For each wave, the location of the 700-hPa vortex center is used to determine when the vortex passes the 15° and 25°W meridians (i.e., into and out of the boxed area in Fig. 1), which we call the Cape Verde region. Atmospheric fields (i.e., geopotential height, temperature, winds, vertical velocity, relative humidity, precipitation) are averaged over this period to establish the environmental conditions in the Cape Verde region. The 700-hPa vortex is selected as a base since vortices at this level are often associated with AEWs (e.g., Miyakoda et al. 1982; Reed et al. 1988a,b; Pytharoulis and Thorncroft 1999). By employing this averaging technique, we can accurately compare the environmental conditions between the two cases, eliminating other shorter temporal factors such as diurnal variations.

Table 3 lists the estimated movement and propagation rates for each simulated case. Note that both 700-hPa vortex centers move to the northwest, but the pre-Ernesto AEW moves approximately 70% faster than the second wave. This is an approximately 4.5 m s−1 increase in the easterly wind speed between the pre-Ernesto and Debby waves, which based on the results shown in Fig. 8 is reasonable and realistic. The estimated difference in the propagation rates between the two waves is small (<0.70 m s−1), consistent with satellite imagery estimates.

Differences in SSTs over the eastern North Atlantic are not responsible for the differences in development for the two successive AEWs. Figure 1 shows the composite SSTs for the Ernesto AEW simulation, with shaded values over the ocean denoting SST values greater than 26°C. Within the Cape Verde region, the satellite-derived SSTs meet this threshold for tropical cyclogenesis. The SSTs for the Debby AEW composite (not shown) are similar to those in Fig. 1, including the region where SSTs are greater than 26°C.

Three primary factors are identified that explain why the wave associated with Ernesto does not form into a tropical storm in the Cape Verde region while the wave associated with Debby undergoes tropical cyclogenesis: the environmental wind shear, influence of low-level westerly wind surges, and environmental moisture/stability.

a. Wind shear

The wave associated with Debby has a much weaker vertical tilt because the vertical wind shear over the eastern Atlantic is weaker than that associated with the pre-Ernesto AEW. The south/southwest vertical tilt below 700 hPa of the wave associated with Debby is weaker by approximately a factor of 5 in both the WRF model output and the ECMWF reanalysis.

Figures 9a,b show the WRF pre-Ernesto wave 200–850-hPa vertical wind shear for 1200 UTC 18 August and 1200 UTC 19 August 2006, respectively. The first time corresponds to when the pre-Ernesto wave is simulated to leave West Africa; the second time is selected to be one day later. Overall, the 200–850-hPa vertical wind shear is less than 15 m s−1 over the eastern North Atlantic on 1200 UTC 18 August 2006. In the vicinity of the 700-hPa vortex center, denoted by the black dot, the shear is less than 10 m s−1. One day later (Fig. 9b), the 200–850-hPa vertical wind shear increases to 15–25 m s−1 over the eastern Atlantic between 5° and 10°N, whereas the 700-hPa vortex is located in this stronger vertical shear region.

The above measure of vertical shear, between 200 and 850 hPa, does not isolate the influence of the AEJ. Instead, the 600–925-hPa vertical wind shear needs to be examined. On 1200 UTC 18 August 2006 (Fig. 9c), an area of strong (i.e., >15 m s−1) mid-to-low-level vertical shear extends from Africa westward to 27°W. The 700-hPa vortex is southeast of this strong vertical shear in a region where the shear is between 12 and 15 m s−1. One day later (Fig. 9d), the strong 600–925-hPa vertical wind shear expands to the south and west, while the 700-hPa vortex center has moved into this region of relatively strong vertical shear.

Figure 10 shows the vertical shear analysis for the WRF Debby wave simulation. On 1200 UTC 20 August 2006 (Fig. 10a), the simulated 700-hPa wave associated with Debby is located along 15°W, but with no identifiable center. The 200–850-hPa vertical wind shear is greater than 15 m s−1 between the equator and 10°N from 15°W to 25°W, associated with strong low-level southwesterly and upper-level northeasterly flow. North of 10°N the vertical shear is less than 10 m s−1. This pattern of strong (weak) vertical shear south (north) of 10°N persists over the eastern Atlantic one day later (Fig. 10b) with some intensification of the strong vertical shear between 15° and 25°W. Unlike the pre-Ernesto wave case, the 700-hPa vortex center is positioned in the relatively weaker shear area. However, it is within couple degrees of latitude of an area of strong vertical shear. Such close proximity to a region of strong vertical shear for a developing system is not uncommon (Frank and Roundy 2006).

A strong core of 600–925-hPa vertical wind shear is also simulated near 15°N when the Debby wave is at 15°W (Fig. 10c), but this core is not as latitudinally wide as in Fig. 9c. Additionally, the vertical shear in the vicinity of the wave (i.e., near 9°N, 15°W) is approximately 3–5 m s−1 weaker compared to the pre-Ernesto wave conditions. One day later, the strong 600–925-hPa vertical wind shear region has passed to the east, and vertical shear weakens to under 12 m s−1.

Note that, in the case of Debby, there is also a contribution to the vertical shear from the development of the vortex center. In this case study, the largest vertical shear values are south and west of the vortex center. Interestingly, the storm track of Debby is to the northwest (Fig. 6b), which is away from these areas of stronger vortex-induced vertical wind shear. This could suggest that vortex-associated vertical wind shear may be influencing the storm track and the development of Debby to some degree (recall that Debby never achieves hurricane strength).

For comparison, satellite-estimated vertical wind shear from the University of Wisconsin—Madison Cooperative Institute for Meteorological Satellite Studies (CIMSS; not shown) reveals that the vertical shear environment for Debby (the pre-Ernesto AEW) was generally less than 11 m s−1 (greater than 15 m s−1). Our simulation results are in line with the CIMSS satellite vertical wind shear estimates.

The vertical shear from the ECMWF reanalysis (not shown) resembles the pattern shown in Figs. 9 and 10 but with reduced magnitudes. The low-level to midlevel vertical shear in the Ernesto wave is not nearly as large as in the WRF simulation, with the core of strongest vertical shear (i.e., >15 m s−1) confined to a much smaller region between 12° and 16°N, 15° and 25°W. Vertical shear estimates from NASA AMMA dropsondes agree more closely with the WRF results for the pre-Ernesto wave. This difference may be at least partially associated with the stronger development of the vortex in the case of the wave associated with Debby.

To better understand the vertical shear patterns, a cross section of the zonal wind and meridional/vertical circulation along 20°W for each composite is shown in Fig. 11. For the pre-Ernesto wave composite (Fig. 11a), key features include a westerly maximum [i.e., the low-level westerly jet (e.g., Reed et al. 1977; Grodsky et al. 2003)] of 6 m s−1 centered at 925 hPa and 9°N, an easterly maximum (i.e., the AEJ) of 18 m s−1 centered at 600 hPa and 14°N, and an easterly maximum [i.e., the tropical easterly jet (TEJ)] of 9 m s−1 centered at 300 hPa and 4°N. The AEJ is stronger and broader than average with wind speeds greater than 10 m s−1 extending from 10° to 20°N. This intensification of the AEJ is a common characteristic of a SAL outbreak. Rising vertical motion associated with the marine ITCZ is confined south of the AEJ between 6° and 10°N, whereas over the AEJ there is subsidence between 600 and 200 hPa. The marine ITCZ tilts equatorward with height above 700 hPa, associated with the stronger midlevel AEJ. The strong shear between 5° and 10°N over the Atlantic in Fig. 9b is associated with the difference between the upper tropospheric easterly and low-level westerly flow, whereas the strong mid-to-low-level vertical shear in Fig. 9d is primarily associated with the strength of the AEJ at 600 hPa.

Figure 11b shows a similar cross section at 20°W for the WRF Debby wave composite. Compared with the Ernesto wave composite, the low-level westerly jet is approximately 33% stronger (the TEJ is approximately 23% stronger) and accounts for the strong 200–850-hPa vertical shear between 5° and 10°N south of the 700-hPa vortex center (Fig. 10b). The stronger westerly jet is associated with Debby’s well-established low-level circulation over the eastern Atlantic. The AEJ for the Debby composite is approximately 22% weaker than the Ernesto composite AEJ, with its core located north of 15°N. The low-level to midlevel vertical shear (Fig. 10d) over the eastern North Atlantic remains relatively strong (i.e., >10 m s−1) despite the weakening of the AEJ, because the low-level westerly jet near the surface intensifies in the Debby AEW simulation. Additionally, the marine ITCZ tilts in the opposite direction than in the Ernesto AEW simulation.

The influence of vertical shear is apparent in the vortex structure at various levels. Figure 12 shows geopotential heights and winds at various levels for the Ernesto AEW composite and indicates that the pre-Ernesto vortex center slopes southwestward with decreasing pressure between the surface and the level of the AEJ. At 850 hPa (Fig. 12a), a well-defined vortex center is positioned just off the southern Senegal coast with cyclonic flow associated with strong northerly flow just west of the vortex center. To the south of this vortex there is westerly flow. At 700 hPa (Fig. 12b), the vortex center is positioned farther to the south and west of the vortex at 850 hPa. Farther west, a ridge along 38°W is associated with diffluence between 25° and 40°W.

Above the level of the AEJ, the Ernesto wave vortex center is not discernable in the circulation field. The 500-hPa flow (Fig. 12c) is easterly, becoming diffluent by 25°W. West of 25°W, there is a broad jet entrance region along 5°–10°N that favors upper-level divergence. A strong meridional height gradient is located over the Cape Verde region, associated with a strengthening of the AEJ on 19–20 August 2006 (see Fig. 8) and a SAL outbreak. At 200 hPa (Fig. 12d) flow is predominantly easterly, associated with an anticyclone over the eastern North Atlantic. Closer inspection of the wind vectors in the area bounded by 9°–15°N, 15°–25°W reveals a region of upper-tropospheric wind convergence associated with subsidence above the strong AEJ in the Ernesto composite (Fig. 11a). South of 9°N there is weak wind divergence coinciding with the upward vertical motions associated with the ITCZ (Fig. 11a).

Figure 13 displays the geopotential heights and winds at the same pressure levels for the Debby wave composite. In this case, the vortex remains vertically aligned, extending into the upper troposphere above the AEJ level.

Figure 13a shows the 850-hPa circulation composite. Compared to the pre-Ernesto composite, the vortex center is centered at 12°N, 23°W with approximately the same minimum height value. Flow is northerly west of the vortex center, while a strong westerly/southwesterly flow is simulated south.

At 700 hPa (Fig. 13b) the vortex center is aligned vertically with the 850-hPa vortex. The meridional height gradient and easterly flow associated with the AEJ is positioned north of this center. The jet core is shifted northward to 15°N and has a maximum magnitude of 14 m s−1, approximately 4 m s−1 weaker than the AEJ core for the Ernesto wave.

Unlike the Ernesto wave composite, a vortex center is clearly identifiable at 500 hPa (Fig. 13c), as Debby is a tropical depression for over half of the compositing period. The flow is predominantly easterly except around the trough axis where there is a northerly (southerly) component to the flow west (east) of the trough. Recall from Fig. 8 that the observations indicate that the AEJ intensified over Praia, Cape Verde, on 22 August 2006. The strengthening of the midtropospheric easterly flow over Praia in this case is associated with the tropical development of Debby (e.g., Figs. 13b,c), and not a SAL outbreak, which was the case for the pre-Ernesto wave. Analysis of the ECMWF operational reanalysis and radiosonde data from Praia, Cape Verde, and Dakar, Senegal, confirms that the midlevel moisture was larger (i.e., RH greater than 70%) and that there was not a robust jet maximum leaving West Africa, at least of the size and magnitude associated with the passage of the first wave two days earlier.

At 200 hPa (Fig. 13d) the flow is still anticyclonic north of 13°N; however, the anticyclone over the eastern North Atlantic has weakened. Although difficult to interpret from Fig. 13d, wind divergence is stronger over the Cape Verde region between 8° and 12°N and reflects a deeper, better organized storm system with stronger convection compared to the Ernesto AEW composite. To quantify the upper-level divergence between the two composites, the area-averaged divergence from 8° to 12°N, 15° to 25°W is calculated for each case. For the WRF Ernesto AEW composite (Fig. 12d), the divergence is 0.75 × 10−5 s−1; for the Debby wave composite it is 2.51 × 10−5 s−1, which is over three times stronger.

b. Low-level westerly wind surges

Figures 11 –13 suggest that the development of the Debby AEW over the Cape Verde region was associated with a low-level westerly jet located just south of the 850-hPa vortex center. This surge of the low-level westerly flow prior to tropical development has been noted as being an influential mechanism in providing the external forcing (e.g., low-level cyclonic vorticity and large-scale vertical ascent) necessary for genesis (e.g., Gray 1988, 1998; Lee et al. 1989; Zehr 1992; Briegel and Frank 1997).

The 925-hPa streamlines for the pre-Ernesto AEW simulation at 0300 UTC 18 August 2006 (Fig. 14) indicates that there is no low-level westerly wind surge occurring to the west of the pre-Ernesto wave between 5° and 12°N. At 925 hPa, the vortex associated with Ernesto is positioned at 17°N, 14°W with cyclonic flow around its center. Farther west an anticyclone is centered at 8°N, 32°W, and the low-level flow between these two features is northerly over the Cape Verde region.

The 925-hPa streamlines for the Debby AEW simulation at 0000 and 1600 UTC 20 August 2006 are displayed in Figs. 15a,b. The 925-hPa vortex associated with Debby begins to develop between 0000 and 1600 UTC, forming a defined 925-hPa vortex center at 15°N, 15°W by 1600 UTC. The vortex associated with the pre-Ernesto AEW is located downstream, approximately 2000 km to the west of the Debby vortex. At 0000 UTC (Fig. 15a), low-level flow is southerly to the west of the pre-Ernesto vortex between 20° and 30°W. Between 8° and 12°N near 20°W, some of this southerly flow diverted eastward toward West Africa, forming a low-level westerly jet. At this time, it appears that the formation of the westerly flow is primarily associated with the prior passage of the Ernesto disturbance, with southerly flow converging in the vicinity of the monsoon trough ahead of the Debby wave, initially enhancing low-level convergence early on 20 August 2006 within the monsoon trough region over the eastern Atlantic. The low-level flow is deflected eastward due to Coriolis constraints, resulting in the strong westerly low-level jet on the southern flank of the monsoon trough near 10°N.

The westerly low-level surge is further enhanced as the developing second wave begins to strengthen around 1600 UTC (Fig. 15b). Convergence continues to intensify off the coast of Guinea in association with the developing of the second wave.

Although the discussion in the previous subsection highlighted the importance of the vertical shear, horizontal shear is also relevant, as the location and magnitudes of the low-level westerly jet and the midtropospheric AEJ can influence the generation of cyclonic relative vorticity. Recall from Fig. 11a that in the pre-Ernesto wave environment the AEJ was stronger and extended farther equatorward in association with a SAL outbreak, whereas the low-level westerly jet was weaker, resulting in a weaker vertical tilt of the zonal isotachs between 10° and 15°N and 900 and 700 hPa. The “flattening” of the zonal isotachs is associated with weaker cyclonic relative vorticity due to a weaker meridional gradient of the zonal flow (i.e., ∂u/∂y is weaker) in the lower troposphere underneath the AEJ. The center of the vortex column is at 15°N at the surface and tilts equatorward along the 0 m s−1 zonal wind contour to 700 hPa, where it is centered at 8°N.

Figures 16a,b show the vertical cross section of the relative vorticity field along 20°W for the Ernesto and Debby wave simulation composites, and Fig. 15c shows their difference. Between 12° and 16°N, positive relative vorticity values in the pre-Ernesto environment are located near the surface, whereas between 800 and 300 hPa magnitudes are generally between ±1 × 10−5 s−1. Stronger positive relative vorticity is located near the surface around 17°N and between 700 and 400 hPa near 9°N. The former is associated with strong easterly flow near the surface at 19°N; the latter is associated with the equatorward flank of the AEJ.

Lower-tropospheric cyclonic relative vorticity between 12° and 16°N is stronger in the Debby environment (Fig. 16b) and extends higher into the atmosphere. Cyclonic relative vorticity between 12° and 16°N increases by at least 33% near the surface up to 900 hPa, while positive values greater than 1 × 10−5 s−1 extend up from 900 to 600 hPa, an increase of 2 × 10−5 to 3 × 10−5 s−1 compared to the pre-Ernesto environment. The 600–300-hPa cyclonic relative vorticity maximum associated with the southern flank of the AEJ is 33% weaker than in the pre-Ernesto wave conditions (Fig. 16a). It is also located approximately 2° latitude farther north (i.e., at 10.5°N and 400 hPa for Debby compared to 8.5°N and 400 hPa for the Ernesto wave), consistent with the weakening and retreat of the AEJ (Fig. 11).

Thus, in the Debby environment, the weaker, northward-shifted AEJ and the stronger, deeper low-level westerly jet are associated with an increase in the meridional gradient of the low-level zonal flow and low-level cyclonic relative vorticity between 12° and 16°N. The vertical slope of the zonal isotachs between 10° and 15°N from the surface to 700 hPa is greater than in the pre-Ernesto case (Fig. 11) and is associated with the decrease in the equatorward tilt of the vortex column for Debby as the vortex column for Debby is more vertically aligned.

This examination of the low-level flow indicates that the passage of the AEW that later developed into Ernesto was one factor that preconditioned the lower troposphere, making the environment favorable for the development of the Debby wave by enhancing low-level southwesterly flow, convergence, horizontal low-level zonal shear, and cyclonic relative vorticity ahead of Debby. The increased low-level southwesterly/westerly flow generated to the east of the first wave’s vortex near 10°N is associated with an increase in the low-level horizontal wind shear gradients over the eastern North Atlantic, particularly the meridional gradient of the zonal wind (i.e., ∂u/∂y). Between 10° and 15°N, low-level cyclonic relative vorticity increases by 15%–30% just prior to Debby’s development, because the stronger low-level westerly flow aids in “spinning up” the lower troposphere. Although most of the strengthening of the westerly flow, low-level convergence, and cyclonic relative vorticity after 1200 UTC 20 August 2006 can be directly associated with the development and intensification of the second wave, the occurrence of these conditions prior to 1200 UTC suggests that the passage of the first wave is important initially. These results are consistent with those of Lee et al. (1989) and Briegel and Frank (1997), who note the importance of wind surges prior to tropical cyclogenesis over the northern Indian and western North Pacific Oceans, respectively.

c. Environmental moisture/stability

Another factor that made the environment favorable for the development of Debby was the availability of moisture and its effects on the vertical stability. Figures 17a,b show the vertical profile of relative humidity along 20°W for the Ernesto and Debby composites, whereas Fig. 17c shows their difference. The location of 20°W is selected because it is the midpoint between the bounding longitudes used to formulate the temporal composite (Fig. 1 and Table 3). Between 10° and 20°N, relative humidity values are larger for the Debby composite by 20% below 750 hPa and 40% above 500 hPa. Between the equator and 10°N the pattern reverses, with smaller relative humidity values in the Debby composite. Inspection of vertical cross sections at other longitudes between 17° and 25°W (not shown) indicates results similar to Fig. 17. The ECMWF climatological August position of the maximum relative humidity at this longitude is 10°N, which indicates that, in the Ernesto case, the relative humidity maximum was positioned anomalously south by approximately 5° latitude. This relative humidity anomaly is a characteristic of a SAL outbreak in which dry air is observed off the West African coast in association with stronger midtropospheric easterly flow (Carlson and Prospero 1972; Wu 2007). The precipitation field (Fig. 7) also reflects this southward displacement.

The implications of the moisture distribution for the vertical stability of the environment are evident in its effects on the moist static energy (MSE) field. MSE is defined as the sum of the sensible, latent, and geopotential energy according to
i1520-0469-66-11-3313-e1
where cp is the specific heat of air at constant pressure, T is air temperature, L the latent heat of vaporization of water, q is specific humidity, g the acceleration due to gravity, and z is height. MSE increasing with decreasing pressure denotes a stable atmosphere, so increases in low-level MSE destabilize the vertical column, while increases in the upper-level MSE stabilize the vertical column.

Figure 18 displays differences (Debby wave minus Ernesto wave) in the MSE profile for the composites averaged over the eastern North Atlantic development region for Debby: namely, 10°–15°N, 15°–25°W. Levels where the slope of the MSE difference are negative (i.e., the 850–750- and 500–200-hPa layers) indicate locations in the vertical profile where the environment is more unstable in the Debby wave simulation composite. The increased instability in the Debby environment between 850 and 750 hPa corresponds to the general location of where the SAL is typically located. This layer is also where large differences in the zonal wind (Fig. 10) and cyclonic relative vorticity (Fig. 16) between the pre-Ernesto and Debby wave environments are noted. From the surface to 850 hPa and from 750 to 500 hPa, the slope of the MSE difference is positive, indicating locations in the vertical profile where the environment is more stable in the Debby wave simulation composite. These changes in the MSE profile are related to fluctuations of the moisture profile (i.e., Lq) rather than changes in the thermal characteristics (i.e., cpT), as the shape of the moisture profile difference closely resembles the MSE difference, only with larger magnitudes to offset the negative contribution of cpT difference.

Another way to evaluate the differences in the stability of the large-scale environment between the two cases is to examine the differences in the convective available potential energy (CAPE). CAPE is a measure of the amount of energy a parcel of air would have if lifted between the parcel’s level of free convection to its level of neutral buoyancy. Large values of CAPE (i.e., greater than 2000 J kg−1) imply greater atmospheric instability and, therefore, the likelihood for strong convection.

Area-averaged (10°–15°N, 15°–25°W) daily CAPE is calculated when the waves are approximately located over the Cape Verde region. For the Ernesto (Debby) wave simulation, this is the value from 18 (21) August 2006. Daily values are used to eliminate diurnal variations for the comparison, whereas area-averaging is utilized to present the results on the spatial scales representative of the environmental scale. Results show that the pre-Ernesto CAPE of 1682 J kg−1 is approximately 602 J kg−1 lower than Debby’s CAPE value of 2283 J kg−1 and reflect a more stable environment with reduced convective activity over this region (Fig. 6a).

The combination of lower CAPE, relative humidity, and latent energy values over the Cape Verde region in the WRF model indicate that the environment associated with the passage of the Ernesto AEW is a more stable, drier environment that is less conducive to the development of convection north of 10°N. These characteristics, along with the strong AEJ and increased low-level to midlevel vertical shear, are associated with an intensification of the SAL that coincided with the passage of the Ernesto wave over the Cape Verde region.

6. Conclusions

The passage of two successive AEWs over the eastern North Atlantic between 18 and 23 August 2006 is investigated using the NCAR–NOAA ARW-WRF model to understand why the first AEW does not undergo cyclogenesis over the Cape Verde region, while the subsequent wave develops into tropical storm Debby in the same region. Although the origin location and each wave’s structural characteristics may differ, especially over land, both of these systems move over the eastern Atlantic as AEWs.

Simulation results are first compared with the ECMWF operational reanalysis, NASA AMMA radiosonde and dropsonde measurements, and NHC observations to evaluate the WRF model’s performance. Results from this comparison reveal that WRF better reproduces the observed deepening of the vortex associated with Debby compared to the ECMWF operational reanalysis. The simulations also capture the observed variability of the AEJ, though magnitudes of the strength of the jet during periods when it is most intense are underestimated by 3–10 m s−1. WRF also appears to have a dry bias because precipitable water values are generally 10%–40% lower than radiosonde and dropsonde observations. The WRF model can realistically distinguish the observed nondevelopment of the Ernesto AEW in contrast with the Debby wave, which develops into a tropical storm over the eastern North Atlantic.

Results from this study suggest that the environmental conditions associated with the first AEW’s passage were unfavorable for development. The passage of the first wave over the Cape Verde region is associated with strong low to-middle tropospheric vertical shear owing to a stronger than normal AEJ, lower than normal relative humidity values throughout the troposphere between 10° and 20°N, and an overall more stable atmosphere between 15° and 25°W. All of these environmental conditions are characteristics generally associated with an intensification of the SAL, or a SAL outbreak, which coincided with the passage of the first AEW from the West African coast to the central North Atlantic. Jenkins et al. (2008) also characterize this as an outbreak period.

Strong low-level to midlevel vertical shear suppresses the vertical development of the vortex center above the AEJ level (e.g., 600 hPa). Below 600 hPa, a southwest tilt of the vortex center with height is simulated. In the upper troposphere, convergence and resultant subsidence above the strong AEJ core caps any vertical development above the AEJ level.

The SAL outbreak is also associated with the transport of dry, low relative humidity air over the eastern Atlantic from northern Africa during the passage of the Ernesto AEW. This intrusion of dry air is associated with reduced CAPE values and enhanced subsidence (Fig. 11a) over the Cape Verde region between 10° and 18°N, implying greater atmospheric stability and the suppression of convection poleward of 10°N. Convection that does form during this wave’s passage occurs equatorward of 10°N (Figs. 7a,b), in a region with higher relative humidity values, weaker low-level to midlevel vertical wind shear, and rising vertical motions. This constitutes an equatorward shift of the eastern Atlantic marine ITCZ in association with the SAL outbreak.

The environmental conditions associated with the passage of the second wave disturbance, Debby, were conducive for tropical development. By the time this wave left the West African coast about 1200 UTC 20 August 2006, the SAL outbreak and AEJ had weakened over the eastern North Atlantic. The 600–925-hPa vertical wind shear decreased over the Cape Verde region by 40% compared to the first wave, with shear values less than 15 m s−1 near the vortex center. The 200–850-hPa vertical wind shear values near the vortex center were less than 8 m s−1, falling well within the threshold range necessary for tropical development identified by other studies (e.g., Gray 1998; DeMaria et al. 2001). The vortex center quickly becomes vertically aligned as it propagates into the eastern Atlantic (Figs. 5, 13), and there is strong upward vertical motion throughout the troposphere and strong upper-tropospheric wind divergence between 10° and 15°N. These characteristics are consistent with the development of deep, organized ascent in the atmosphere (Frank and Roundy 2006).

The atmospheric moisture content over the eastern North Atlantic was higher and the environment less stable for the passage of Debby, particularly between 850 and 750 hPa and above 500-hPa layers. Relative humidity values between 10° and 18°N rebound after the SAL outbreak, increasing by 20% (40%) in the lower (upper) troposphere. The added moisture coincides with an approximate 35% increase in CAPE values, reflecting a destabilization of the environment over the eastern North Atlantic off the West African coast.

Finally, the passage of the first wave is one factor that initially helps precondition the environment to be more favorable for tropical development for Debby. By the time the Debby wave disturbance leaves the West African coast, the first AEW was approximately 2000 km to the west (Fig. 15). Low-level flow from the equator to 10°N in the wake of this first wave disturbance is southerly/southwesterly between 15° and 35°W, resulting in the development of a strong southwesterly wind surge south of Debby’s vortex center that intensifies the low-level westerly jet along 9°N prior to genesis (Fig. 11). The initial surge is likely associated with the passage of the first wave; however, by 1200 UTC 20 August 2006, the development/strengthening of the second wave vortex becomes important in the maintenance and continued intensification of the low-level westerly jet. Associated with the stronger, deeper low-level westerly flow, zonal shear ∂u/∂y and, hence, low-level cyclonic relative vorticity increases between 12° and 16°N, resulting in an environment more favorable for development. Briegel and Frank (1997) and Gray (1998) suggest that low-level wind surges southwest of the genesis region can force the large-scale low-level convergence and strong vertical motions necessary for tropical cyclogenesis. Here, we find that the wind surge amounts to a nonlinear interaction between these two consecutive wave disturbances. Furthermore, Briegel and Frank note that these lower-tropospheric surges are often related to the presence of a prior circulation approximately 2000 km to the west of the genesis location, which is the approximate distance between these two disturbances in this study.

Note that the development characteristics of the two disturbances were captured realistically in WRF without accounting for the presumably high dust content of the SAL. This suggests either that other mechanisms are more important in the AEW development process besides aerosol forcing or that other errors, such as the model’s dry bias, may be contributing to produce the realistic response, but for the wrong reasons. Further examination of this issue utilizing a regional model that includes the effects of aerosols is needed.

Finally, the focus here has been on evaluating the synoptic-scale environmental conditions associated with the passage of the two waves. The influence of smaller, mesoscale processes that may also influence cyclogensis (e.g., Holland 1995; Simpson et al. 1997; Reasor et al. 2005) are not explored since they are not resolved sufficiently at 30-km resolution. The next step is to examine the development of the wave associated with Debby at spatial scales suited to resolve these small-scale processes to understand their role in the development of Debby and how they interact with the synoptic environmental conditions documented in this work.

Acknowledgments

This research was supported by NASA NAMMA06 Grant NNX07AI92G. The rainfall data used in this study were acquired as part of the Tropical Rainfall Measuring Mission (TRMM). TRMM is an international project jointly sponsored by the Japan National Space Development Agency (NASDA) and the U.S. NASA Office of Earth Sciences. Acknowledgement is made to the National Center for Atmospheric Research, which is sponsored by the National Science Foundation, for the computing time used in this research.

REFERENCES

  • Bister, M., , and K. A. Emanuel, 1997: The genesis of Hurricane Guillermo: TEXMEX analysis and a modeling study. Mon. Wea. Rev., 125 , 26622682.

    • Search Google Scholar
    • Export Citation
  • Bosart, L. F., , and J. A. Bartlo, 1991: Tropical storm formation in a baroclinic environment. Mon. Wea. Rev., 119 , 19792013.

  • Bracken, W. E., , and L. F. Bosart, 2000: The role of synoptic-scale flow during tropical cyclogenesis over the North Atlantic Ocean. Mon. Wea. Rev., 128 , 353376.

    • Search Google Scholar
    • Export Citation
  • Briegel, L. M., , and W. M. Frank, 1997: Large-scale influences on tropical cyclogenesis in the western North Pacific. Mon. Wea. Rev., 125 , 13971413.

    • Search Google Scholar
    • Export Citation
  • Burpee, R. W., 1972: The origin and structure of easterly waves in the lower troposphere of North Africa. J. Atmos. Sci., 29 , 7790.

  • Carlson, T. N., , and J. M. Prospero, 1972: The large-scale movement of Saharan air outbreaks over the northern equatorial Atlantic. J. Appl. Meteor., 11 , 283297.

    • Search Google Scholar
    • Export Citation
  • Carlson, T. N., , and S. G. Benjamin, 1980: Radiative heating rates of Saharan dust. J. Atmos. Sci., 37 , 193213.

  • Chen, F., , and J. Dudhia, 2001: Coupling an advanced land surface–hydrology model with the Penn State–NCAR MM5 modeling system. Part I: Model implementation and sensitivity. Mon. Wea. Rev., 129 , 569585.

    • Search Google Scholar
    • Export Citation
  • Chen, S-H., , and W-Y. Sun, 2002: A one-dimensional time dependent cloud model. J. Meteor. Soc. Japan, 80 , 99118.

  • Colón, J. A., , and W. R. Nightingale, 1963: Development of tropical cyclones in relation to circulation patterns at the 200-millibar level. Mon. Wea. Rev., 91 , 329336.

    • Search Google Scholar
    • Export Citation
  • Cook, K. H., 1999: Generation of the African easterly jet and its role in determining West African precipitation. J. Climate, 12 , 11651184.

    • Search Google Scholar
    • Export Citation
  • DeMaria, M., , J. A. Knaff, , and B. H. Connell, 2001: A tropical cyclone genesis parameter for the tropical Atlantic. Wea. Forecasting, 16 , 219233.

    • Search Google Scholar
    • Export Citation
  • Dunion, J. P., , and C. S. Velden, 2004: The impact of the Saharan air layer on Atlantic tropical cyclone activity. Bull. Amer. Meteor. Soc., 85 , 353365.

    • Search Google Scholar
    • Export Citation
  • Dunkerton, T. J., , M. T. Montgomery, , and Z. Wang, 2008: Tropical cyclogenesis in a tropical wave critical layer: Easterly waves. Atmos. Chem. Phys. Discuss., 8 , 1114911292.

    • Search Google Scholar
    • Export Citation
  • Emanuel, K. A., 1989: The finite amplitude nature of tropical cyclogenesis. J. Atmos. Phys., 46 , 34313456.

  • Fett, R. W., 1966: Upper-level structure of the formative tropical cyclone. Mon. Wea. Rev., 94 , 918.

  • Frank, W. M., 1988: Tropical cyclone formation. A Global View of Tropical Cyclones, R. L. Elsberry, Ed., Office of Naval Research, 53–90.

    • Search Google Scholar
    • Export Citation
  • Frank, W. M., , and P. E. Roundy, 2006: The role of tropical waves in tropical cyclogenesis. Mon. Wea. Rev., 134 , 23972417.

  • Fujita, T., , K. Watanabe, , and T. Izawa, 1969: Formation and structure of equatorial anticyclones caused by large-scale cross-equatorial flow determined by ATS-1 photographs. University of Chicago Research Rep. SMRP-78, 37 pp.

    • Search Google Scholar
    • Export Citation
  • Gallina, G. M., , and C. S. Velden, 2002: Environmental vertical wind shear and tropical cyclone intensity change utilizing enhanced satellite derived wind information. Preprints, 25th Conf. on Hurricanes and Tropical Meteorology, San Diego, CA, Amer. Meteor. Soc., 172–173.

    • Search Google Scholar
    • Export Citation
  • Gray, W. M., 1968: Global view of the origin of tropical disturbances and storms. Mon. Wea. Rev., 96 , 669700.

  • Gray, W. M., 1988: Environmental influences on tropical cyclones. Aust. Meteor. Mag., 36 , 127139.

  • Gray, W. M., 1998: The formation of tropical cyclones. Meteor. Atmos. Phys., 67 , 3769.

  • 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
  • Holland, G. J., 1995: Scale interaction in the western Pacific monsoon. Meteor. Atmos. Phys., 56 , 5279.

  • Jenkins, G. S., , A. S. Pratt, , and A. Heymsfield, 2008: Possible linkages between Saharan dust and tropical cyclone rain band invigoration in the eastern Atlantic during NAMMA-06. Geophys. Res. Lett., 35 , LO8815. doi:10.1029/2008GL034072.

    • Search Google Scholar
    • Export Citation
  • Kain, J. S., , and J. M. Fritsch, 1990: A one-dimensional entraining/detraining plume model and its application in convective parameterization. J. Atmos. Sci., 47 , 27842802.

    • Search Google Scholar
    • Export Citation
  • Kain, J. S., , and J. M. Fritsch, 1993: Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 46, Amer. Meteor. Soc., 165–170.

    • Search Google Scholar
    • Export Citation
  • Lee, C-S., , R. Edson, , and W. M. Gray, 1989: Some large-scale characteristics associated with tropical cyclone development in the north Indian Ocean during FGGE. Mon. Wea. Rev., 117 , 407426.

    • Search Google Scholar
    • Export Citation
  • Miyakoda, K., , J. Sheldon, , and J. Sirutis, 1982: Four-dimensional analysis experiment during the gate period. Part II. J. Atmos. Sci., 39 , 486506.

    • Search Google Scholar
    • Export Citation
  • Molinari, J., , S. Skubis, , and D. Vollaro, 1995: External influences on hurricane intensity. Part III: Potential vorticity structure. J. Atmos. Sci., 52 , 35933606.

    • Search Google Scholar
    • Export Citation
  • Montgomery, M. T., , M. E. Nicholls, , T. A. Cram, , and A. B. Saunders, 2006: A vertical hot tower route to tropical cyclogenesis. J. Atmos. Sci., 63 , 355386.

    • Search Google Scholar
    • Export Citation
  • Palmen, E., 1948: On the formation and structure of tropical cyclones. Geophysics, 3 , 2638.

  • Powell, M. D., 1990: Boundary layer structure and dynamics in outer hurricane rainbands. Part II: Downdraft modification and mixed-layer recovery. Mon. Wea. Rev., 118 , 918938.

    • Search Google Scholar
    • Export Citation
  • Pytharoulis, I., , and C. Thorncroft, 1999: The low-level structure of African easterly waves in 1995. Mon. Wea. Rev., 127 , 22662280.

  • Ramage, C. S., 1959: Hurricane development. J. Meteor., 16 , 227237.

  • Reasor, P. D., , M. T. Montgomery, , and L. F. Bosart, 2005: Mesoscale observations of the genesis of Hurricane Dolly (1996). J. Atmos. Sci., 62 , 31513171.

    • Search Google Scholar
    • Export Citation
  • Reed, R. J., , D. C. Norquist, , and E. E. Recker, 1977: The structure and properties of African disturbances as observed during phase III of GATE. Mon. Wea. Rev., 105 , 317333.

    • Search Google Scholar
    • Export Citation
  • Reed, R. J., , A. Hollingsworth, , W. A. Heckley, , and F. Delsol, 1988a: An evaluation of the ECMWF operational system in analyzing and forecasting easterly wave disturbances over Africa and the tropical Atlantic. Mon. Wea. Rev., 116 , 824865.

    • Search Google Scholar
    • Export Citation
  • Reed, R. J., , E. Klinker, , and A. Hollingsworth, 1988b: The structure and characteristics of African easterly wave disturbances as determined from the ECMWF operational analysis/forecast system. Meteor. Atmos. Phys., 38 , 2233.

    • Search Google Scholar
    • Export Citation
  • Simpson, J., , E. Ritchie, , G. J. Holland, , J. Halverson, , and S. Stewart, 1997: Mesoscale interactions in tropical cyclone genesis. Mon. Wea. Rev., 125 , 26432661.

    • Search Google Scholar
    • Export Citation
  • Skamarock, W. C., , J. B. Klemp, , J. Dudhia, , D. O. Gill, , D. M. Barker, , W. Wang, , and J. G. Powers, 2005: A description of the advanced research WRF version 2. NCAR Tech. Note NCAR/TN-408+STR, 88 pp.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C., , and K. Hodges, 2001: African easterly wave variability and its relationship to Atlantic tropical cyclone activity. J. Climate, 14 , 11661179.

    • Search Google Scholar
    • Export Citation
  • Wu, L., 2007: Impact of Saharan air layer on hurricane peak intensity. Geophys. Res. Lett., 34 , L09802. doi:10.1029/2007GL029564.

  • Zehr, R. M., 1991: Observational investigation of tropical cyclogenesis in the western North Pacific. Preprints, 19th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., 235–240.

    • Search Google Scholar
    • Export Citation
  • Zehr, R. M., 1992: Tropical cyclogenesis in the western North Pacific. NOAA Tech. Rep. NESDIS 61, 181 pp.

  • Zipser, E. J., and Coauthors, 2009: The Saharan air layer and the fate of African easterly waves: The NAMMA field program. Bull. Amer. Meteor. Soc., 90 , 11371156.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Model domain utilized for this study. Shading over land denotes topography in meters as given by the shading bar. Contours over the ocean denote the NASA TMI SSTs (°C) from 18 Aug 2006. SST values greater than 26°C are shaded in light gray. The box denotes the longitudinal constraints selected for formulation of temporal composites.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 2.
Fig. 2.

Debby wave vortex mean sea level pressure (SLP) every 6 h between 1800 UTC 20 Aug and 0000 UTC 24 Aug 2006. Solid bold line denotes WRF model simulation analyzed further in this paper; the long-dashed (short-dashed) bold line is the ECMWF operational reanalysis (NHC observed intensity). The thin solid lines denote simulation results of other physical parameterization combinations defined in the text.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 3.
Fig. 3.

Pre-Ernesto wave vortex (a) mean sea level pressure (hPa), (b) 850-hPa height (m), and (c) 700-hPa height (m) every 6 h between 1800 UTC 18 Aug and 0000 UTC 20 Aug 2006. Solid line denotes WRF model simulation; the dashed line is the ECMWF operational reanalysis.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 4.
Fig. 4.

Pre-Ernesto wave vortex center positions at sea level, 850 hPa, and 700 hPa every 6 h between 1800 UTC 18 Aug and 0000 UTC 20 Aug 2006 from the (a) ECMWF operational reanalysis and (b) model simulation.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 5.
Fig. 5.

Debby wave vortex (a) 850- and (b) 700-hPa heights (m) every 6 h between 1800 UTC 20 Aug and 0000 UTC 24 Aug 2006. Solid line denotes WRF model simulation; the long-dashed line is the ECMWF operational reanalysis.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 6.
Fig. 6.

As in Fig. 4 but for Tropical Storm Debby between 1800 UTC 20 Aug and 0000 UTC 23 Aug 2006.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 7.
Fig. 7.

Averaged precipitation rates (mm day−1), 700-hPa winds (m s−1), and geopotential heights (m) for the pre-Ernesto wave from (a) WRF simulation and (b) ECMWF operational reanalysis and TRMM 3B43 rainfall rates averaged from 1200 UTC 18 Aug to 0000 UTC 20 Aug 2006. (c),(d) As in (a),(b) but for the Debby wave averaged from 1200 UTC 20 Aug to 1800 UTC 22 Aug 2006. Shading denotes rainfall rates according to the grayscale bar.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 8.
Fig. 8.

The 600-hPa zonal wind component (m s−1) at Praia, Cape Verde (14.9°N, 23.5°W), from the pre-Ernesto wave simulation (solid line), the Debby wave simulation (long-dashed line), and the 700-hPa zonal wind component from the ECMWF operational reanalysis (short-dashed lines). Crosshair marks are observed 600-hPa winds from NASA AMMA radiosonde measurements.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 9.
Fig. 9.

WRF pre-Ernesto wave 200–850-hPa vertical wind shear (m s−1) for (a) 1200 UTC 18 Aug and (b) 1200 UTC 19 Aug 2006. Additionally, 600–925-hPa vertical wind shear for (c) 1200 UTC 18 Aug and (d) 1200 UTC 19 Aug 2006. Values greater than 15 m s−1 are shaded. Bolded mark denotes the location of the 700-hPa vortex. Vectors are in m s−1.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 10.
Fig. 10.

WRF Debby wave 200–850-hPa vertical wind shear (m s−1) for (a) 1200 UTC 20 Aug and (b) 1200 UTC 21 Aug 2006. Additionally, 600–925-hPa vertical wind shear for (c) 1200 UTC 20 Aug and (d) 1200 UTC 21 Aug 2006. Values greater than 15 m s−1 are shaded; bolded mark denotes the location of the 700-hPa vortex; vectors in m s−1.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 11.
Fig. 11.

Vertical cross section at 20°W of the zonal wind (shaded; m s−1) and the meridional and vertical circulation (vectors and streamlines) for (a) pre-Ernesto and (b) Debby wave WRF simulation composites.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 12.
Fig. 12.

WRF pre-Ernesto wave simulation composite geopotential heights (m) and winds (m s−1) at (a) 850 hPa, (b) 700, (c) 500, and (d) 200 hPa.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 13.
Fig. 13.

WRF Debby wave simulation composite geopotential heights (m) and winds (m s−1) at (a) 850, (b) 700 hPa, (c) 500 hPa, and (d) 200 hPa.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 14.
Fig. 14.

WRF 925-hPa streamlines for the pre-Ernesto wave simulation at 0300 UTC 18 Aug 2006.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 15.
Fig. 15.

As in Fig. 14 but for the Debby wave simulation at (a) 0000 and (b) 1600 UTC 20 Aug 2006.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 16.
Fig. 16.

Vertical cross section of relative vorticity (×10−5 s−1) at 20°W for (a) pre-Ernesto and (b) Debby wave WRF simulation composites and (c) the Debby minus pre-Ernesto difference. Shading denotes negative values.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 17.
Fig. 17.

Cross section of relative humidity (%) at 20°W for the (a) Ernesto and (b) Debby wave WRF composites and (c) the Debby minus Ernesto RH difference. Shading in (c) denotes negative values.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Fig. 18.
Fig. 18.

Area-averaged (10°–15°N, 15°–25°W) Debby wave minus pre-Ernesto wave WRF composite difference of MSE (bold solid line), sensible energy (bold dashed line), latent energy (dotted line), and geopotential energy (long/short-dashed line) from the WRF model. Units are all 104 J kg−1.

Citation: Journal of the Atmospheric Sciences 66, 11; 10.1175/2009JAS3064.1

Table 1.

Track and wind speed intensity error for the Tropical Storm Debby wave in August.

Table 1.
Table 2.

Precipitable water (PW) comparison for instrument type: radiosonde (R) and Dropsonde (D).

Table 2.
Table 3.

Estimated 700-hPa movement rates for two simulated waves.

Table 3.
Save