Saharan dust, lightning and tropical cyclones in the eastern tropical Atlantic during NAMMA-06



[1] During the summer of 2006, the downstream component of African Monsoon Multidisciplinary Analyses Campaign (NASA-AMMA (NAMMA)) examined African Easterly Waves (AEWs) emerging from the coast of Africa. Six of these disturbances went on to become named systems in the Tropical Atlantic. Two of the six systems (Tropical storm Debby and Hurricane Helene) developed in the extreme eastern Atlantic and were associated with dust outbreaks, elevated ice contents and frequent lightning. Here we show that in the early tropical cyclo-genesis stages of these systems there were thousands of cloud-to-ground (CG) lightning flashes as measured by a ground-lightning network. TRMM overpasses show high precipitation ice content above the freezing level and high latent heat release. Super-cooled water can be inferred in the lower parts of cloud systems in concert with observed high ice concentrations at high altitudes creating charge separation based on the large numbers of CG flashes.

1. Introduction

[2] During a typical Atlantic Hurricane Season (June– November) an average of 10–11 named tropical systems will occur and have the potential of impacting downstream communities along Central/North American coastlines and the Caribbean Islands. There has been considerable discussion on positive trends in the intensity of hurricanes in the Tropical Atlantic and if the trends can be attributed to natural or anthropogenic forcing [Emanuel, 2005; Webster et al., 2005; Pielke et al., 2005; Trenberth and Shea, 2006]. Sea Surface Temperatures (SSTs) act to modulate the intensity of tropical cyclones (TCs), other factors such as vertical wind shear can also play a role. The Saharan Air Layer (SAL) [Carlson and Prospero, 1972] is very likely to impact vertical shear, the large-scale thermodynamic environment and SSTs [Lau and Kim, 2007] during the Atlantic hurricane season. The SAL's influence on the Main Development Region (MDR) [6–18°N, 20–60°W] for TC evolution may be either negative or positive.

[3] The SAL is a dominant feature during Northern Hemisphere summer June–September (JJAS) and resides between 900−500 hPa in height and is bound by an enhanced trade wind inversion at its base. The SAL can negatively influences tropical cyclo-genesis (TC-genesis) in the Eastern Atlantic because of dry air, increased vertical shear and atmospheric stability [Dunion and Velden, 2004; Wong and Dessler, 2005; Evan et al., 2006]. Studies also suggest that the SAL can positively influence TC genesis through an increase in cyclonic vorticity and positive vorticity advection [Karyampudi and Pierce, 2002]. Moreover, the SAL can create a baroclinic zone which can enhance AEW development [Karyampudi and Carlson, 1988]. The SAL may also influence TC-genesis thorough aerosol -cloud microphysics interactions leading to invigorated convection [Koren et al., 2005; Khain et al., 2005; DeMott et al., 2003; Jenkins et al., 2008a]. Prior to 2006 there were few in situ or active measurements associated with the SAL and any influence that it might have on TC-genesis. The African Monsoon Multidisciplinary Analysis (AMMA) Campaign [Redelsperger et al., 2006] included a downstream component during August–September 2006. In particular NASA provided support for ground and aircraft measurements (NASA DC-8) to examine: (a) TC-genesis in the extreme Eastern Tropical Atlantic, (b) the SAL and its interaction with TC-genesis, (c) precipitation processes associated with emerging AEWs, (d) satellite comparisons to ground and aircraft measurements. The NASA component of AMMA was denoted at NAMMA-06.

[4] During the period of August 15th through September 30th there were approximately 11 AEWs that emerged from the West African coastline. Six out of the 11 waves, became named tropical systems (Debby, Ernesto, Florence, Gordon, Helene and Isaac). Two of these systems (Debby and Helene) formed very close to the West African coastline with the first aircraft measurements (NASA DC-8, NOAA G-IV) of TC-genesis and evolution in the extreme Eastern Atlantic. There were approximately 9 SAL dust outbreaks during the period of August 15th–September 30th based on Aerosol Optical Thickness values greater than 0.6 at Dakar and Aerosol Optical thickness and visible images from MODIS.

[5] Lightning which has been observed with TCs provides a measure of convective intensity (updraft strength) and is generally associated with mid-level super-cooled water/graupel and ice particles at high altitudes [Zipser et al., 2006; Black and Hallett, 1999]. Climatologically, lightning is much more frequent over land than ocean and the highest global lightning frequencies are found over Africa [Zipser et al., 2006]. Sealy et al. [2003] found the highest number of flashes during June–August (JJA) occurred over the Sahelian region using 3 years of TRMM data. Lightning is also responsible for the natural production of ozone (O3) via NOX production, potentially enhancing the middle and upper troposphere downstream of Africa during JJA [Jenkins and Ryu, 2004; Jenkins et al., 2008b]. The primary objective of this paper is to show observations of two developing tropical disturbances (TS-Debby, TD-8), the SAL, CG-lightning flashes and cloud microphysical variables derived from TRMM during the NAMMA-06 field campaign.

2. Data and Results

[6] During NAMMA-06, the UK ATD Lightning Detection Network was used, based on VLF Arrival Time Difference (ATD) principles., which measures cloud-to-ground (CG) lightning flashes was employed for continental and oceanic lightning measurements [Lee, 1986; Keogh et al., 2007]. In addition we employ latent heat estimates [Olsen et al., 2006] from the Tropical Rainfall Measurement Mission (TRMM) [Kummerow et al., 1998] overpasses for TS Debby and TD-8 which was investigated by the DC-8. TRMM overpasses of TS-Debby occur on August 22nd 2006 in the Tropical Eastern Atlantic, 1 day prior to DC-8 investigations. The TRMM overpass for TD-8 occurred on September 12th, 2006 during the DC-8 investigation. Fixes on the tropical depressions from the National Hurricane Center (NHC) and NCEP reanalysis are used. We also use visible images with false color for identifying deep convection associated with the tropical disturbances.

[7] Tropical Storm Debby was officially named a tropical depression (TD-4) on August 21st at 1800 UTC. However, the disturbance emerged from the coastline on August 20th between 0600 and 1200 UTC and was associated with cold cloud tops and frequent lightning (see below). TS-Debby reached its peak intensity on August 24th 2006 with 24.3 m-s−1 wind speeds and a sea level pressure of 999 hPa (National Hurricane Center). Hurricane Helene was officially named a tropical depression twelve on September 12th 1200 UTC approximately 12 hours after it emerged from West Africa. The system was very large in spatial extent, relative to TS-Debby, and reached hurricane status on September 16th at 1200 UTC. Hurricane Helene reached its peak intensity on September 18th at 0600 UTC with 56.4 m-s−1 wind speeds and a sea level pressure of 955 hPa.

[8] Prior to each of these AEWs emerging from the West African continent a large SAL outbreak occurred with considerable dust (Figures 1a1f) based on visible images and satellite MODIS Aerosol Optical Thickness (AOT) [Jenkins et al., 2008a]. The SAL outbreaks were associated with a mid-level ridge at 700 hPa followed by the AEW, similar to the conceptual model and simulations developed by Karyampudi and Carlson [1988]. MCSs in the form of squall lines were also generated and propagated ahead of the AEW.

Figure 1.

(a–c) Visible and False IR images of cloud pattern associated with TS-Debby and (d–f) TD-8. False colors represent the coldest cloud tops. Milky color represents dust.

[9] While the SAL can create a hostile environment for TC-genesis, Saharan dust also acts as a good source of Ice Nuclei (IN) [DeMott et al., 2003] and Cloud Condensation Nuclei (CCN). Koren et al. [2005] describes from satellite observations and Khain et al. [2005] show through simulations how aerosols would create smaller droplets and delay warm rain production. The lifting of the resulting smaller droplets to subfreezing temperatures would increase latent heating and drive stronger vertical motions. Moreover, the upper parts of the cloud would be glaciated with ice particles as a result of these strong vertical motions in concert with enhanced IN concentrations. This model is supported by satellite observations [Koren et al., 2005] and in-situ measurements showing large ice particles (>10,000 μm), very strong vertical velocities (26 m-s−1) and high cloud liquid water (1.8 g-m−3) at 10–11 km during the DC-8 flight through TD-8 on September 12th, 2006 [Jenkins et al., 2008a].

[10] Figures 2a2d show TRMM satellite overpasses of the tropical disturbances associated with TD-4 on August 22nd at 1034 UTC and TD-8 on September 12 at 1407 UTC. First, convective rain rates of 10–15 mm-hr−1 are shown near the center of these depressions (Figures 2a and 2b). Moreover the largest values of estimated latent heat from TRMM can be found at levels about the freezing level (based on radio soundings and aircraft dropsonde (4.5–5 km above the surface). In both cases latent heating greater than 8°C/hr can be found suggesting that a significant amount of freezing has occurred. The precipitation that is ice at similar or higher altitudes is estimated to be greater than 0.8 g-m−3 at 5–6 km. With TD-4 these values fall off to approximately 0.2 g-m−3 at 10–14 km in the center of Debby. Slightly higher values of precipitation ice are found in the middle and upper parts of TD-8. Moreover, the most potent convection in TD-8, associated with DC-8 measurements was NW of the TRMM overpass and not observed.

Figure 2.

(a, b) TRMM convective rain rates (mm-hr−1), (c, d) latent heat estimates (°C -hr−1) and (e, f) Precipitation ice (g-m−3) for TD-4 (Aug. 22, 1034 UTC) and TD-8 (Sept 12, 1407 UTC).

[11] In the Eastern Atlantic, lightning may serve as a precursor to TC genesis [Chronis et al., 2007]. During the NAMMA period, lightning was associated with all of the AEWs leaving the West Africa coast and traversing the eastern Atlantic from August 15th–September 20th of 2006. Furthermore Price et al. [2007] shows that most of the AEWs that transitioned to named TC in 2006 were proceeded by above average lightning in East Africa.

[12] In the genesis stages of TS-Debby a large number of CG flashes (3142, 6085) were observed on August 20th and August 21st prior to it being denoted a tropical depression (Figures 3a and 3b and Table 1). A decrease in CG lightning flashes is found on subsequent days (Figure 3c). While TS-Debby was forecasted to reach hurricane strength, it also encountered greater shear on August 23rd and encountered a second SAL outbreak. TS-Debby also began to enter into a region of cooler SSTs in the following days. In contrast, a increasing number of CG flashes (5454, 6711, 12417) are found with TD-8 during the first three days of the system. Many of the CG flashes are associated with the outer rain-bands near the dust boundary (Figures 1d1f and Table 1) and consistent with TRMM observations of lightning [Cecil et al., 2002]. Limited observations show that the highest CAPE (2000–3000 J-kg-1) were found here on September 12th, potentially enhancing convection. However the low-level shear was also very high which would tend to inhibit convection [Jenkins et al., 2008a]. The persistent and large number of CG lightning flashes implies that Saharan aerosols may have a significant impact on cloud-microphysics as was suggested by Jenkins et al. [2008a]. Table 1 summarizes the number of CG lightning flashes associated with the pre-genesis and genesis stages of Debby and Helene for the period of August 20th–24th and Sept 12th–15th. The largest numbers of CG flashes are found on 8–20 and 8–21 with TS Debby, while the greatest numbers of CG flashes are found on 9/14 when TS Helene was named. The lightning peaks occur during the genesis phases with both systems (prior to depression stage with Debby and Tropical storm stage with Helene). The lightning totals are far higher (at least 1 magnitude of order higher) than what has been observed with TCs in the Western Atlantic [Black and Hallett, 1999] and imply that Saharan dust aerosols are directly or indirectly influencing TC rain-band/microphysics (super-cooled water, graupel, ice) in the Eastern Tropical Atlantic.

Figure 3.

(a–c) CG lightning strike locations for August 20th-August 22nd associated with TD-4 and (d–f) for Sept 12th–14th associated with TD-8. The closed circle with X denotes the position of the disturbance based on analyses (purple X) and the NHC official position (red X).

Table 1. Total CG Lightning Flashes Associated With TS Debby and Helene
DateNumber of Observed Lightning FlashesLatitude-Longitude Range
8/2031428–16°N, 17°W–12°W
8/21 (TD-4)60856–16°N, 25°W–15°W
8/228056–17°N, 28°W–20°W
8/23 (TS-Debby)16014–20°N, 35°W–25°W
8/2423617–24°N, 44°W–33°W
9/12 (TD-8)54548–16°N, 30°W–17°W
9/1367118–18°N, 35°W–20°W
9/14 (TS-Helene)124178–20°N, 40°W–25°W
9/15124612–22°N, 40°W–30°W

3. Conclusion and Discussion

[13] In this paper we show that two tropical systems (TS-Debby and TD-8) were associated with high quantities of CG lightning flashes, with the maximum observed value greater than 6000 for Debby and 12,000 for Helene, prior to becoming named storms. Both systems were associated with SAL outbreaks, which produced significant amounts of ice in the upper part of the systems. The lower part of these systems probably have significant amounts of super-cooled water and graupel where high latent heating is noted above the freezing layers in both storms from TRMM overpasses.

[14] The results presented here also have implications for climate change and potential links to Saharan dust. First, we should expect that dust within the SAL will lead to reduced solar radiation at the surface potentially cooling SSTs and modulating TC evolution [Lau and Kim, 2007]. Second, the SAL does create stable, dry conditions (observed during NAMMA) that may inhibit TC genesis [Dunion and Velden, 2004]. Third, the SAL modulates cloud microphysics and potentially can invigorate developing rain bands associated with TC-genesis [Koren et al., 2005; Jenkins et al., 2008a]. The SAL can also dynamically modulate the wind field, increasing cyclonic vorticity of the AEW [Karyampudi and Pierce, 2002]. Fourth, the large number of lightning flashes over the eastern Atlantic and continental West Africa can have secondary climate impacts by enhance tropospheric O3, another greenhouse gas, through NOX production during the NH summer season [Jenkins and Ryu, 2004; Martin et al., 2007; Jenkins et al., 2008b].

[15] Many if not most of these processes are not currently included in global climate models (GCMs) at present because of uncertainties associated with aerosol-cloud microphysics-chemistry interactions and the various scales of motions (sub-grid scale) that must be resolved. Additional field campaigns are necessary to increase our understanding of processes associated with TC-genesis that can be eventually incorporated into climate models. This will increase our understanding of TC genesis and evolution at present and more importantly potential 21st century climate change. Finally, the large number of CG lightning flashes associated with developing TCs in the Eastern Atlantic pose a hazard to aircraft and marine interests and should be monitored using lightning networks.


[16] This work was supported through NASA grant NNX06AC78G. Aaron Pratt is supported by NOAA Educational Partnership Program (EPP) grant NA17AE1625.