Dust emission mechanisms in the central Sahara: new insights from remote field observations

North Africa is the world's largest source of mineral aerosol (dust). The Fennec Project, an international consortium led by the University of Oxford, is the first project to systematically instrument the remote central Sahara Desert. These observations have, among others, provided new insights into the atmospheric mechanisms of dust emission. Bordj Badji Mokhtar, in south-west Algeria, is within kilometres of the centre of the global mean summer dust maximum. The site, operated by Fennec partners ONM Algerie, has been heavily instrumented since summer 2011. During the Intensive Observation Period (IOP) in June 2011, four main emission mechanisms were observed and documented: cold pool outflows, low level jets (LLJs), monsoon surges and dry convective plumes. Establishing the relative importance of dust emission mechanisms has been a long-standing research goal. A detailed partitioning exercise of dust events during the IOP shows that 45% of the dust over BBM was generated by local emission in cold pool outflows, 14% by LLJs and only 2% by dry convective plumes. 27% of the dust was advected to the site rather than locally emitted and 12% of the dust was residual or ';background' dust. The work shows the primacy of cold pool outflows for dust emission in the region and also the important contribution of dust advection. In accordance with long-held ideas, the cube of wind speed is strongly correlated with dust emission. Surprisingly however, particles in long-range advection (>500km) were found to be larger than locally emitted dust. Although a clear LLJ wind structure is evident in the mean diurnal cycle during the IOP (12m/s peak winds at 935hPa between 04-05h), LLJs are only responsible for a relatively small amount of dust emission. There is significant daily variability in LLJ strength; the strongest winds are produced by a relatively small number of events. The position and strength of the Saharan Heat Low is strongly associated with the development (or otherwise) of LLJs. However, the presence of a LLJ is not a guarantee of dust emission. Momentum calculations show that dust emission always occurs if momentum mixes down to the surface, but mix-down does not always happen - particularly if the surface temperature inversion is strong or ground heating is weak. Fennec findings are not only providing new insights into dust emission processes, they are also an excellent test-bed for models and satellite algorithms in a region where high-quality ';ground-truthing' measurements have been scarce. Conditions of (relatively) high water vapour appear to be a common cause of error. In one model, wind speeds in the core of monsoon LLJs are underestimated by 8.5m/s compared to observations.