Observations of an 11 September Sahelian Squall Line and Saharan Air Layer Outbreak during NAMMA-06

1 Program in Atmospheric Sciences, Howard University, Washington, DC 20059, USA 2 NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA 3 Atmospheric Science Group, Department of Geosciences, Texas Tech University, Lubbuck, TX 79409, USA 4 Earth Resources Technology, Inc., Laurel, MD 20707, USA 5 Laboratory of Atmospheric Physics—Simeon Fongong, Cheikh Anta Diop University, Dakar, Senegal 6 Rosenstiel School of Marine and Atmospheric Science, Cooperative Institute for Marine and Atmospheric Science Studies, University of Miami, Miami, FL 33149, USA 7 Environmental Protection Agency, Region 4, Atlanta, GA 30303, USA 8 Department of Mechanical Engineering, Howard University, Washington, DC 20059, USA 9 Department of Environmental Science, Policy, and Management, University of California Berkeley, Berkeley, CA 94720, USA 10Department of Environmental Sciences, University of Virginia, Charlottesville, VA 22903, USA


Introduction
Westerly propagating squall lines (SLs) are frequent phenomena in the West African Sahel during the Northern Hemisphere summer [1,2].This SL formed and propagated ahead of the trough region of an African Easterly Wave (AEWs) [3,4].The development of SLs involves a combination of atmospheric conditions including conditional instability, low-level convergence, and vertical wind shear between relatively cool, moist, monsoon lower tropospheric air and warmer, drier Harmattan winds driven by the 650 hPa AEJ [3,5].
The structure of SLs in West Africa was observed and analyzed during the Global Atmospheric Research Program's (GARP's) Atlantic Tropical Experiment (GATE) and Convection Profonde Tropicale (COPT 81) field campaigns [3,[5][6][7][8].Squall lines typically have a convective leading edge that produces a short period of intense rainfall.Observations show that intense precipitation extended 5 to 6 km above the surface (see reflectivity values of ≥50 dBZ in [7], Figure 8, [8], Figure 7).Observations and modeling studies suggest that SLs in the Sahel were embedded in the troughs of AEWs [3,9,10].The axis of the AEW trough often contains two distinct mesoscale features: a low-level Saharan heat low to the north of the AEJ and a circulation at 650 to 700 hPa around 10 • N [11][12][13][14].The trough of the AEW highlighted as "Wave 7" of the 2006 NASA-African Monsoon Multidisciplinary Analyses (NAMMA-06) [15] took on these coupled dynamical features.The NAMMA-06 campaign extended from 15 August through 30 September 2006.The NAMMA-06 Campaign's goals were to find differences between AEWs that produce hurricanes and those that do not, the specific role of the Saharan Air Layer (SAL) has on AEWs and tropical cyclones, and to examine the role of cloud microphysics on precipitation amounts and contribution.The NAMMA-06 study accomplished the goals by observing seven AEWs and SAL outbreaks [15].The SL in this study was one of two that developed with an AEW propagating westward out of Mali during the 9-11 September 2006 period and was a precursor to Hurricane Helene.Moreover, NAMMA-06 included an investigation of five SAL outbreaks and associated cloud microphysical properties [16].

International Journal of Geophysics
The primary goals of this paper are to use data from the NAMMA-06 period to examine observations of an 11 September SL, SAL outbreak, and its interaction with the SL.Data locations and sources are discussed in Section 2 and the synoptic analysis and observations in Section 3. Observations of the 11 September SAL outbreak are presented in Section 4. A brief synthesis of the SL and SAL interaction is given in Section 5. Section 6 gives a summary of the major results and ideas for future work.

Data Locations and Sources
The locations of sites within Senegal where meteorological data were collected near the SL and SAL outbreak are presented in Figure 1     and six hours, respectively [17].The data and locations include flux tower measurements in Kawsara (14.7 • N, 17.1 • W) (Figure 1), surface observations at the Cap Skirring Airport (GOGS) (12.4 • N, 16.7 • W) (Figure 1), rawindsonde data from Kawsara, Leopold Sedar Senghor International Airport in Dakar (GOOY) (14.7 • N, 17.5 • W), and Tambacounda Airport (GOTT) (13.8 • N, 13.7 • W) (Figure 1).The National Centers for Environmental Prediction/Climate Prediction Center (NCEP/CPC) 4 km globally merged Infrared cloud top brightness temperature dataset [18] are used to evaluate the location and intensity of the SL convection.The 0000 UTC analysis on 11 September 2006 depicts several large-scale dynamical features over West Africa.The Tropical Easterly Jet (TEJ) can be seen at 200 hPa extending from central Africa westward along the Gulf of Guinea coast, and then northwestward just offshore of Guinea-Bissau and Senegal (Figure 3(a)).This jet helped support upper-level divergence, which aided in repeated squall line development with this AEW.

Squall Line Environmental Characteristics
A closed circulation exists within the AEW trough axis at 650 hPa, with the core of the AEJ just to its north (Figure 3(b)).This closed circulation is also noted at 850 hPa (Figure 3(c)) and 925 hPa (Figure 3(d)).A sharp gradient in specific humidity can be seen over Senegal at 850 hPa, separating the dry SAL regime from the much moister (specific humidity > 0.012 kg kg −1 ) regime near the vicinity of the incipient squall line.A Saharan heat low is also present at 925 hPa (Figure 3(d)) over southern Mauritania; the flow around the heat low advects dust aerosols south and east toward central and eastern Senegal.This general circulation pattern is characteristic of some AEWs [11][12][13][14].The same general circulation features were observed for this AEW with the Global Data Assimilation System analyses [15].At 0600 UTC (Figures 4(a)-4(c)), the low-level vortex associated with the AEW at 850 hPa and 925 hPa is located over southern Guinea (12 • W, 10 • N).These features, along with the collocated 650 hPa AEW vortex and 925 hPa heat low over southern Mauritania, are propagating westward.The drier SAL regime has pushed further to the south (Figure 5(b)), while dust continues to be advected into the storm environment by the 925 hPa circulation.At 1200 UTC (Figure 5(a)-5(d)), the vortex at 850 hPa has moved further west with wind speeds increasing as suggested by the streamlines.The Saharan heat low (Figure 5(d)) moves southwestward to the northern coast of Senegal and the compact streamlines at 650 hPa suggest that the circulation is also strengthening.This circulation is centered 75-100 km east of the coast of Guinea and continues moving westward (Figure 5(b)).The position of these coupled features shows that the northern end of the AEW trough axis is moving faster than its southern end.Moreover, the 650 hPa circulation is located northwest of the 850 hPa circulation, indicating that the AEW circulation is tilted from southeast to northwest with height (Figure 5(c)).

Satellite and Surface
Observations.The merged IR satellite observations (Figure 6) show the development and evolution of the convection associated with the AEW.This particular AEW had a history of producing numerous squall lines; one such squall line can be seen moving offshore at 0000 UTC 11 September [15].A second SL (the purpose of this study) is initiated in southeastern Senegal at 0200 UTC (Figure 6(b)).As this SL developed IR, its cloud-top temperatures cool to 200-220 K south of the 14 • N latitude line (Figures 6(c)-6(f)); there is also an expansion of the cloud canopy associated with this system.The cooling cloud top temperatures are indicative of the SL intensifying; the cloud tops are coolest at 0600 UTC (Figure 6(d)).This system moved offshore between 0900-1000 UTC.
Pregust front/SL conditions in Kawsara and Cap Skirring (Figure 7(a)) show a steady west (270 • ) to west-northwest (300 • ) wind at both locations throughout the morning, providing moist low-level onshore flow into the region.The wind speeds range from 3.0 to 8.0 m s −1 .Meanwhile, the SL continued to propagate westward into a thermodynamic regime conducive for the maintenance of strong convection.Convective Available Potential Energy (CAPE) value of 3237 J kg −1 at 0000 UTC is observed at Dakar and 4443 J kg −1 at 0549 UTC at Kawsara (Table 1).At both locations, the CAPE exceeds one of the highest COPT 81 values of 2810 J kg −1 [9].
There was a marked temperature drop from 28.0 • C at 0900 to 22.0 • C at 1100 UTC at Cap Skirring (Figure 7(a))    7(b)).Between the hours of 0900 and 1100 UTC, the pressure rises from 1011 to 1014 hPa in Cap Skirring after the SL passage and from 1002 to 1006 hPa in Kawsara after the gust front passed.These observations are typical of squall lines and suggest a mesohigh followed the SL and gust front [7,8].In both locations, the pressure dropped quickly as the SL approaches and then rose afterward due to the post-SL mesohigh.Moreover, at both Kawsara and Cap Skirring, the wind shifts to an easterly direction and the speed decreases to 1.0-4.0m s −1 after SL passage.The CAPPI images in the top panels of Figures 9(a)-9(c) depict the westward progression of the SL off the coast of Senegal.The radar scans focus on the northern portion of the SL in western Senegal.These observations coupled with the surface data in Figure 7 and the surveillance scans in Figure 8 provide additional evidence that a dry gust front passes through Kawsara during the late morning of 11 September 2006.

Radar and Lightning
The dashed lines on the CAPPI images are the horizontal locations of RHI cross-section plots in the bottom panels of Figures 8(a up to nearly 15 km AGL, which is consistent with higher cloud top heights at 0900 UTC (see Figure 6).However, while the 48 dBZ contour extends horizontally for 10 km, it only reached an altitude of 3 km AGL.This horizontally extensive but shallow convection is consistent with the general lack of convective intensity seen in this SL at 0900 UTC.At this time, there is no trailing stratiform rain region associated with this SL.Surveillance scans to the south (see Figure 8) further reveal this underdeveloped stratiform region.As the squall line transitions from a continental to maritime regime, the convection associated with it became shallower and less intense (Figure 9(b)).The reduced amount of CAPE offshore relative to land most likely contributed to the decrease in convective intensity; dropsondes released offshore on 12 September indicated maximum CAPE values of approximately 2500 J kg −1 (not shown), considerably less than the values observed over land the previous day.A study investigating an mesoscale convective system moving over West Africa on 31 August 2006 also noted less favorable thermodynamic and dynamic conditions offshore that contributed to convective weakening [19].Although there were a few small cells with 54 dBZ contours at 4 km AGL observed in the horizontal cross section, the vertical cross section taken through the leading edge shows the 42 dBZ contour reaching only slightly above 4 km AGL.A secondary reflectivity maximum at approximately 47 to 50 km distance along the dashed line cross section is also observed.This suggests that decaying cells from the convective leading edge move rearward once they have reached a mature stage [20].However, at 1000 UTC (Figure 9), the stratiform rain region is even less cohesive with the leading edge to the west than it was at 0900 UTC.This lack of cohesion of the stratiform region is uncharacteristic of SLs observed in GATE and COPT 81 [3,[5][6][7][8].Figure 9(c) shows that the northern end of the SL is approximately 50 km southwest of the NPOL radar location at 1100 UTC.
Figure 10 depicts the UK-ATD lightning data [21,22] between 0400-1200 UTC on 11 September 2006.Lightning flashes show that the SL extended southward from approximately 14.5 • N to 12.5 • N from 0400 to 0600 UTC as shown by purple crosses.Lightning associated with the SL moved westward from 0600 to 0800 UTC as shown by blue circles.Between 0800-1000 UTC, the lightning flashes associated with the SL as shown by red squares approached the West African coastline.The lightning flashes associated with the SL moved westward into the Atlantic between 1000-1200 UTC shown by the yellow triangles.The region of lightning flashes to the northeast of the SL persisted throughout this time period.

Observations of Saharan Air Layer Outbreak
A significant SAL outbreak, one of five observed between 15 August and 15 September 2006 [23], preceded the SL on 11 September 2006.SAL outbreaks are defined as the onset of a well-mixed, warm, dry, and dusty air mass 2.0 to 6.0 km deep that originates from Saharan Africa and propagates westward toward the Eastern Atlantic Ocean [24].This air mass (bounded by temperature inversions at its base and top) International Journal of Geophysics is usually 5 to 6 • C warmer than the more humid and shallow (e.g., 1.2 to 1.8 km) marine boundary layer air below [25][26][27].Typically the AEJ forms around 650 hPa on the southern edge of the SAL primarily due to the temperature contrast between the warm SAL to the north and the relatively cooler marine air to its south [24,28,29].
The ECMWF Reanalyses of 850 hPa specific humidity on 11 September (see Figures 3(c Rawindsonde data also show the presence of the SAL in the vicinity of the SL on 11 September 2006.At 0000 UTC, the SAL (though weak and shallow) is present at Tambacounda, Senegal (Figure 11(a)) and (more prominently) in Dakar (Figure 11(b)) and at 0549 UTC in Kawsara (Figure 11(c)).The depth of the SAL is denoted by the vertical extent of the SAL between its lower and upper temperature inversions.At Dakar (Figure 11(b)), the SAL extends from 935 hPa to 580 hPa, while at Kawsara (Figure 9(c)) it starts at 900 hPa and is capped at 600 hPa.At both locations (Figures 11(b) and 11(c)), there is a robust dewpoint depression on Two-day back-trajectories from the HYSPLIT Model at Dakar, Kawsara, and Tambacounda (Figure 14) illustrate the origins of the air mass at various altitudes over western Senegal on 11 September.The air mass at 3.0 km in altitude in Tambacounda at 1200 UTC 10 September originated out of the Sahara; by contrast, the origin of the air mass at 0.5 km and 1.5 km originates over the Eastern Atlantic basin and northern Senegal, respectively (Figure 14(a)).Dakar and Kawsara exhibit airflow from the Sahara at both 1.5 km and 3.0 km at 0000 UTC 11 September (Figure 14).This deeper airflow out of the Sahara explains the greater vertical extent of the SAL and stronger temperature inversion at its base over Dakar and Kawsara as compared to Tambacounda.These conditions also contributed to enhanced atmospheric stability, which lead to weakening of the northern portion of the squall line.

Discussion
The SL was generated by an AEW in southeastern Senegal at 0000 UTC on 11 September 2006 (Figure 6).A SAL outbreak was building into the western Senegal and the adjacent Atlantic Ocean waters.The SL approached the coast and was several hundred kilometers ahead of the AEW vortex by 0900 UTC.The southern two-thirds of the SL maintained its strength.This is noted by its NPOL reflectivity and the robust lightning flashes observed by WWLLN south of 14 • N. The northern third weakened in the presence of abundant dust noted by the lack of lightning north of 14 • N (Figure 10), the Kawsara and Dakar radiosondes (Figures 11(a) and 11(b), 12(a) and 12(b)), increasingly high AOT at Mbour (Figure 13), no recorded rainfall by the AMMA22 and AMMA23 rain gauges (not shown), and precipitationfree boundary in Kawsara during the 0900 to 1200 UTC hour.Moreover, HYSPLIT modeling (Figure 14) provides evidence that particles in coastal Senegal at 0800 to 1000 UTC originated from the Sahara implying that Saharan dust was present in this region.
The reflectivity values in Figure 8 indicate that the southern part of the SL weakened as it propagated off the coast, particularly by 1100 UTC; however, the reflectivity is the only observation that reveals the southern portion of the SL weakens.This is similar to the transition the 31 August SL in DeLonge et al. [19].No solid conclusion can be made on the SL and AEW transition to Hurricane Helene.There are not aircraft observations or TRMM overpasses over this SL on 11 September to make the connection.More aircraft-and surface-based measurements are needed to tie this AEW and SL to other coastally transitioning storms.The next set of aircraft observations are not until 12 September 2006.At this time, the AEW develops into Hurricane Helene by then (Jenkins et al. [23] and Zawislak and Zipser [15]).

Conclusion
The NAMMA-06 campaign provided measurements for coastal Senegal.These observations provided insight into the dynamics and thermodynamics of the SL and SAL.The ECMWF Reanalyses show that the AEW and SL were in a favorable environment for continued strengthening and maturing.These observations included ground-based measurements such as flux tower measurements, radiosonde launches, and AERONET.There was remote sensing data from the NPOL radar, TRMM MERG, and MODIS satellite products.These observations revealed that the westerly propagating SL and the SAL in the lower troposphere interacted.The interaction was evident through the weakening of the northern portion of the SL.The HYSPLIT back-trajectory model gives indication of impending dust.Even though CAPE was high and an SL-related boundary passed through Kawsara, there was not precipitation on its northern end.
There are avenues for additional work on this paper.Additional plausible ideas for future research include aircraft observations that might identify the role SAL aerosols have  on clouds, precipitation, and developing tropical systems.Despite there being observations and arguments to support the negative interaction of the two (e.g., Wong and Dessler [30]), there is evidence that dust and aerosol entrainment into convective systems acts as cloud condensation nuclei subsequently forming precipitation and banding.Wong and Dessler [31] concluded that Saharan dust increased the number of cumulus cloud particles in the North Atlantic

Figure 1 :Figure 2 :
Figure 1: Map of cities and villages in Senegal where meteorological data relevant to this study was obtained.
. The European Center for Medium Range Weather Forecasting (ECMWF) 40-year (ERA-40) Reanalysis data are used to examine the synoptic-scale and mesoscale conditions in the vicinity of the SL.The ERA-40 Reanalysis covers the years of 1957-2010 and has a spatial and temporal resolution of approximately

Figure 7 :
Figure 7: (a) Double y-axis plots containing an 11 September 2006 time series of wind speed in m s −1 and wind direction in degrees from north for Kawsara and Cap Skirring, Senegal.(b) Double y-axis plots containing an 11 September 2006 time series of pressure in hPa denoted by the solid line and 1.5 m temperature in • C denoted by the dashed line.
), 4(b), and 5(c)) indicate a tongue of low specific humidity (values less than 0.009 kg kg −1 ) extending south into coastal and northern Senegal.The low moisture measurements are indicative of drier air and dust intruding from the Saharan Desert into northern Senegal.The moisture steadily increases to the south and east over southeast Senegal, Mali, and Guinea-Bissau.

Figure 9 :
Figure 9: (a) A horizontal constant altitude plan position indicator (CAPPI) image at 4.0 km altitude and radar height indicator RHI plots of the NPOL radar reflectivity (dBZ) along the leading convective line on 11 September 2006 at 0900 UTC.The negative sign in the CAPPI images indicates distance in the westward direction.The dashed line in the CAPPI image represents the location of the cross section.In the RHI images, ordinate axis is height z in km and the azimuthal is the distance in km along the cross-section.(b) A horizontal constant altitude plan position indicator (CAPPI) image at 4.0 km altitude and radar height indicator RHI plots of the NPOL radar reflectivity (dBZ) along the leading convective line on 11 September 2006 at 1000 UTC.The negative sign in the CAPPI images indicates distance in the westward direction.The dashed line in the CAPPI image represents the location of the cross section.In the RHI images, ordinate axis is height z in km and the azimuthal is the distance in km along the cross-section.(c) A horizontal constant altitude plan position indicator (CAPPI) image at 4.0 km altitude and radar height indicator RHI plots of the NPOL radar reflectivity (dBZ) along the leading convective line on 11 September 2006 at 1100 UTC.The negative sign in the CAPPI images indicates distance in the westward direction.The dashed line in the CAPPI image represents the location of the cross section.In the RHI images, ordinate axis is height z in km and the azimuthal is the distance in km along the cross-section.

Figure 10 :Figure 11 :Figure 12 :
Figure 10: An image of cloud-to-ground lightning flashes on 11 September 2006 between 0400 and 1200 UTC.Purple crosses are for CG flashes from 0400 to 0600 UTC; blue circles are for flashes from 0600 to 0800 UTC; red squares are for flashes from 0800 to 1000 UTC; yellow triangles are for flashes from 1000 to 1200 UTC.

Figure 13 :
Figure 13: (a) A time series plot of 440 Angstrom AOT for 8 September 2006 at Mbour, Senegal.(b) A time series plot of 440 Angstrom AOT for 9 September 2006 at Mbour, Senegal.(c) A time series plot of 440 Angstrom AOT for 10 September 2006 at Mbour, Senegal.(d) A time series plot of 440 Angstrom AOT for 11 September 2006 at Mbour, Senegal.

Table 1 :
Depiction of the CAPE and CIN for Dakar, Kawsara, and Tambacounda, Senegal, prior to the SL propagating into the vicinity of these cities.