Inaug.-diss.--Rostock. / "Literatur-verzeichnis": p. [56]-57.

In boreal summer, satellite measurements show that the central Sahara is the dustiest region of the planet. However, ground-based observations of the central Sahara have been limited to its outer edges, leaving a void in observations approximately 1 million km² in area. The Fennec Project has been the first campaign to instrument this remote but climatologically important region. This thesis uses these new observations to detect and explain the atmospheric mechanisms that make the central Sahara the summer global dust maximum. Four atmospheric mechanisms are found to cause dust storms in the central Sahara in June 2011 and June 2012. These are cold pool outflows, low-level jets (LLJs), monsoon surges and dry convective plumes. Dust may be emitted locally by these phenomena, or be advected, principally by cold pools. In both field seasons, dust emission by cold pool outflows is the most important dust mechanism, causing roughly half of the total dust loadings at the Fennec supersite of Bordj-Badji Mokhtar (BBM), the closest station to the dust maximum. The second most important mechanism is dust advection by cold pools (roughly 30% dust at BBM), followed by dust emission by monsoon surges, LLJs and finally dry convective plumes (only 2% dust at BBM). Although June 2012 was significantly more dusty than June 2011, the relative importance of the different atmospheric dust mechanisms at BBM did not change. At the automatic weather stations (AWSs) across the remote desert, cold pools and LLJs are by far the most frequently detected atmospheric dust mechanisms. LLJs are particularly common in the Atlantic Inflow in western Mauritania and in the north-easterly Harmattan in western Algeria. Cold pools are much more frequent at BBM, the station under the greatest moist monsoon influence, than at the AWSs to the north. Detection of advected dust is a particular difficulty without dedicated dust-detection instrumentation or human observers (e.g. at the AWSs). Detection of dust emission mechanisms can be very successful with only routine ground observations and satellite measurements, but quantifying the associated dust burden without dedicated dust instruments is problematic. The choice of instrumentation for dust measurement is crucial. Because cold pool outflows - the most important dust mechanism - frequently occur at night or under cloud, sun photometers miss about half of cold pool dust. Lidars have the advantage of providing height resolved dust profiles, but they suffer from attenuation in thick dust. The nephelometer proved to be the most reliable dust instrument. Although LLJs occurred on 21/28 mornings at BBM in June 2011, only five of these jets led to dust emission. Calculations of momentum exchanges through the atmospheric column show that momentum mix-down from the jet core is the cause of dust emission on these occasions, but that the LLJ has to be particularly strong (≥ 16 m s^-1) to result in dust emission at the surface. Met Office Africa-LAM underestimates monsoon LLJ wind profiles and ERA-Interim reanalysis underestimates both monsoon and Harmattan LLJ wind profiles. At the surface, the Met Office Africa-LAM and GLOBAL models significantly underpredict the frequency of observed wind speeds &gt6 m s^-1. This will cause them to significantly underestimate dust emission, as emission is a threshold process proportional to the cube of wind speed. A particularly interesting implication of the research presented here is that the central Sahara is likely much more dusty than previously thought. This is because almost all of the techniques currently used to study dust in the region are systematically biased to result in underestimates of dust burden. Cold pools are the most important dust mechanism but, since they rarely occur during the daytime or in cloud-free conditions they are often missed by sun photometers. Many will be missed by satellites that cannot retrieve below cloud and satellites that pass over the Sahara in daylight hours (e.g. the A-train). A commonly-used satellite dust detection algorithm often misses dust under moist (i.e. cold pool) conditions. Cold pools cannot be simulated by models without explicit convection, which requires very high spatial resolution. Finally, the numerical models assessed here significantly underpredict the frequency of wind speeds over the dust emission threshold. The Sahara is probably much dustier than current estimates suggest.