Portal de Periódicos da CAPES

Comment on “Atmospheric controls on the annual cycle of North African dust” by S. Engelstaedter and R. Washington

2008; American Geophysical Union; Volume: 113; Issue: D23 Linguagem: Inglês

10.1029/2008jd009930

2156-2202

Earle Williams,

Atmospheric Ozone and Climate

[1] The recent review of natural dust production in North Africa by Engelstaedter and Washington [2007] (hereinafter referred to as EW) gives no attention to a prevalent meteorological phenomenon south of the Intertropical Discontinuity (ITD) in the Sahel called 'haboob'. These vigorous dust-producers are essentially dry microbursts, and are the products of moist convection over sparsely vegetated terrain. The inattention to this phenomenon can be attributed to shortcomings of climatological TOMS satellite observations, blocked by upper tropospheric cirrus cloud and strongly attenuated in the lower boundary layer. [2] Engelstaedter and Washington [2007] analyze the annual cycle of dust production in North Africa using one principle remote sensing instrument: Total Ozone Mapping Spectrometer (TOMS). Studies of a similar nature appeared earlier in the work by Engelstaedter et al. [2006]. The thesis of this comment is that TOMS is limited in detecting another potent source of dust production in North Africa during a similar phase of the annual cycle: haboobs [Sutton, 1925, 1931]. These gust front outflows, ubiquitous in the Sahel, are caused by moist convection in the northern fringe of the Intertropical Convergence Zone (ITCZ), which is also producing upper level cirrus. The large cirrus canopy serves to block TOMS's column-integrated "view" of the lower tropospheric dust repository, since both cirrus and dust serve as scatterers in the ultraviolet. An additional limitation of TOMS pertains to the detection of low-altitude (200–300 m) sources [Mahowald and Dufresne, 2004]. Recent observations with the Massachusetts Institute of Technology radar during African Monsoon Multidisciplinary Analysis (AMMA) [Williams et al., 2006] show that virtually every cold outflow (microburst) from moist convection in Niamey, Niger (a representative location for the Sahel), produces a deep dust cloud (2–5 km) that can propagate tens to hundreds of kilometers from the parent storm. EW restricts their scope of investigation to dry convection as a dust raiser in Africa and gives no mention to haboobs in the Sahel. This restricted view has evidently led the authors to certain assumptions and interpretations whose validity invites debate and discussion. Examples of such assertions and interpretations are given below, in each case followed by an alternative interpretation based on the reality of potent Sahelian dust sources, as well as on recent observations in the Saharan Air Layer (SAL) on the opposite side of the ITD. [3] EW (paragraph 5) state "most meteorological stations are located in the Sahel… and not near the major dust sources in the central Sahara." [4] Recent studies by Williams et al. [2006, 2008] have shown that major gust front–producing dust storms are located in the Sahel, where the more numerous meteorological stations are present to document them. These events are associated with moist convection predominantly on the northern fringe of the ITCZ and in a phase of the annual cycle for North Africa consistent with what is shown by EW. The mass loading of dust in these events measured at the surface generally exceeds like measurements in dry dust storms in the Sahara [D'Almeida, 1986]. These Sahelian gust fronts satisfy EW's requirement of "small-scale, high-wind" events needed for dust raising. Indeed, the local observations in Niamey, Niger, show a strongly nonlinear response of lofted dust to local wind speed. Comparisons of dry deposition in western Niger [Drees et al., 1993] by dust events during the monsoon season and during the wintertime Harmattan show a predominance in the monsoon. [5] Comparisons with MeteoSat imagery during AMMA have shown gust fronts propagating northward into the Sahara Desert more than 1000 km from their convective origins [Williams et al., 2006]. [6] EW (paragraphs 13 and 14) state "from January onward, the ITCZ starts moving north, bringing the monsoon rain to the Sahel and thereby scavenging the atmosphere of aerosols…. This north-south gradient in dust occurrence might be explained by mesoscale convective systems to the south of the AEW which are associated with precipitation and the reduction in atmospheric dust." [7] Scavenging of dust by rain is EW's explanation for the paucity of TOMS observations of dust in the Sahel. Such an explanation requires that the dust produced by gust fronts propagating tens and hundreds of kilometers from the parent storms [Sutton, 1925, 1931; Williams et al., 2006] be reingested by the parent storm and subsequently returned to Earth in precipitation. This seems unlikely for multiple reasons. The air that descends from the cloud to feed the gust front outflow and mix with the dust probably originates in a region of low theta-e aloft and is, therefore, not thermodynamically predisposed to reingestion by the parent storm. Furthermore, the precipitation efficiency of these strongly continental Sahelian storms is probably quite low because boundary layer aerosol is so abundant, and so cloud droplet sizes will remain small and serve to suppress coalescence [Williams et al., 2002]. Consequently, even that portion of the dust raised by the gust fronts and reingested by the convection may not be returned to Earth in rain. [8] EW (paragraph 13) state "in the Sahel… emissions are highest in winter when the Sahel is driest…" [9] True, the TOMS maps of aerosol index (AI) in EW's Figure 2 show the largest values in the winter months of December, January, and February, but these are also the months when the Sahel is located in the subsidence of the local Hadley circulation, and so upper tropospheric cirrus cloud is generally absent (O. Torres, personal communication, 2007). In such conditions, TOMS is most likely to have a clear view to the lower troposphere, and so this wintertime dust is readily detected. In the wintertime, one must be concerned with aerosol from biomass burning in addition to mineral dust. [10] The limitation of TOMS in detecting dust beneath tropospheric cirrus cloud provides a viable alternative explanation for the paucity of dust observations in the Sahel and the pronounced latitudinal gradient in the TOMS AI in EW's Figure 2 in the summer months when Sahelian cirrus production is most prevalent. The paucity of dust observations in the Sahelian region in summer TOMS observations is a systematic result and is not unique to this study by EW [see, e.g., Prospero et al., 2002]. [11] EW (2007, paragraph 15) state "given the assumption that erodibility factors that can limit dust emissions (such as the availability of deflatable sediments or vegetations cover) do not change significantly in these regions over the course of the year." [12] In the Sahel, there is abundant evidence that the wetting of the soil by rainfall suppresses the lofting of dust [Williams et al., 2006] as the wet season progresses. For this reason, the production of dust in the Sahel probably maximizes in the early wet season (June and July), consistent with early studies in Khartoum [Sutton, 1925, 1931]. The annual phase is well matched to results shown by EW and also with the annual variation of dust arriving in Barbados from Africa [Prospero and Lamb, 2003]. This role of rainfall in "keeping the dust down" is also a viable explanation for the inverse correlation between Sahel rainfall and dust arriving in Barbados on the interannual time scale documented by Prospero and Lamb [2003]. [13] EW (paragraph 16) state "convergence with southerly winds from the equator occurs in the ITCZ… the convergence belt moves north until August when it reaches its most northerly position between 16–22°N." [14] The sharply defined line of surface convergence discussed above and illustrated in EW's Figure 6 is not the ITCZ but rather the ITD (or sometimes referred to as the Intertropical Front (ITF)). This key meteorological feature, which plays a critical role in dust raising and transport in recent studies by Flamant et al. [2007] and Bou Karam et al. [2008], is not specifically addressed by EW. The ITCZ generally lies some 500 km south of the ITD [Hamilton and Archibald, 1945; Lélé and Lamb, 2007] and represents the core region for upper tropospheric cirrus production. The ITD separates the moist monsoon layer to the south (and its haboob-producing storms) from the dry Harmattan boundary layer to the north. Recent observations by Flamant et al. [2007] have shown abundant dust north of the ITD and progressing southward and upward over the monsoon layer in the northeasterly Harmattan flow. This mechanism is consistent with the traditional picture of the SAL [Carlson and Prospero, 1972] but departs from EW's picture of dry vertical convection alone. [15] EW (paragraph 19) state "dry deep convection associated with surface convergence plays an important role in dust generation in the West Africa source regions." [16] Niamey, Niger, in the Sahel lies at a similar latitude (∼13°N) to Barbados in the Caribbean. There is little question that the surface material in Niamey available for deflation is abundant in April (the hottest, driest month there) because prodigious deep dust clouds are produced 1–2 months later when isolated moist convection is available to raise it. Furthermore, cirrus cloud is largely absent at this latitude in April. The ITD crosses the latitude of Niamey in April, heading north. Why then is there no substantial TOMS signal at the latitude of Niamey in April, and a strong burst of dust arriving in Barbados in the same month, instead of two months later? [17] EW (paragraph 19) state "where rainfall is sufficient to support vegetation, the lack of emissions during the southward advance of the convergence zone can be explained by a reduction in available sediments as a result of increased vegetation cover. Reasons for the lack of symmetry in the hyperarid Sahara are less clear…" [18] Figures 4 and 8 from EW show the seasonal march of TOMS dust observations and the climatological rainfall. Climatology on the position of the ITD (at the longitude of Niamey, Niger) [see Lélé and Lamb, 2007] shows that the ITD progresses in latitude approximately in step with the 250 mm a−1 rainfall contour depicted by EW's Figure 4. [19] EW's interpretation of Figures 4 and 8, showing a diminishment of TOMS-observed dust in the latitudinal rainfall gradient, is again based on scavenging of dust by rain. An alternative interpretation, discussed earlier, is based on obscuration of the dust by cirrus cloud, which is also increasingly prevalent in the rainfall region (where moist convection is available to produce the high-altitude cirrus). The latter interpretation also allows for the two aspects not satisfactorily explained by EW's interpretation, namely, (1) the predominance of TOMS-observed dust earlier than the most northward ITD advances and (2) [EW, 2007, paragraph 19] "peak dust emissions are restricted only to the crossing of the convergence belt on its northward advance and not when the convergence zone retreats south 2–3 months later." [20] This alternative interpretation relies on the presence of strong dust emissions beneath the condensate mask, south of the latitudinal maxima documented in the TOMS observations. In addition, this interpretation relies on rainfall-mediated surface conditions (both surface moisture and vegetation) that mitigate against deflation. The rainfall put down during the northward advance of the ITCZ can suppress dust production during its southward retreat, even in the presence of abundant haboobs. This interpretation cannot be valid, however, over the "hyperarid Sahara" because rainfall is absent there, as EW agree. While other satellite methods have been developed to distinguish cirrus from dust over West Africa [Chaboureau et al., 2007], even these methods are not likely to work when cirrus cloud lies directly over dust cloud, as is the case during the wet season in the Sahel. [21] In summary, the interpretations from EW and in this comment are quite distinct. In one case Sahelian dust is eliminated by precipitation in the ITCZ, and in the other case the ITCZ cloudiness merely masks the dust from satellite observation. We are in accord with EW (2007, paragraph 1) that the "understanding of West African sources is limited because of the remoteness of sources and lack of surface observations." Sahelian haboobs, however, have been recognized in surface observations for many decades [Sutton, 1925, 1931; Tetzlaff and Peters, 1986; Pye, 1987]. Why then did they not figure into this important review on sources of North African dust over the annual cycle? Additional light will be shed on this issue through the use of new satellite assets Cloudsat and Calipso which, unlike TOMS, are height-resolving observations. [22] Discussions on this topic with C. Flamant, D. Bou Karam, J.-P. Chaboureau, I. Lélé, L. Machado, N. Mahowald, N. Renno, B. Russell, C. Thorncroft, S. Tanelli, P. Lamb, J. Prospero, and O. Torres are much appreciated. Radar field studies in Niger for AMMA have been supported by NASA Hydrology (J. Entin) and by funds from RIPIECSA (A. Diedhiou).