Transatlantic Dust Transport and Deposition: An Evaluation of the GEOS-5 Model with Satellite and Aircraft Observations

Massive dust emitted from deserts and semi-arid regions in North Africa can transport long distances across the tropical Atlantic Ocean, reaching the Americas. Dust deposition along the transit adds essential nutrients to marine and terrestrial ecosystems, which could increase the productivity of the ecosystems and CO₂ uptake, modulate biogeochemical cycle, and influence climate. Assessment of the dust-biogeochemistry–climate interactions has been in part hindered by the paucity of dust deposition measurements, particularly in open oceans, and large uncertainties associated with representing dust processes in models. The objectives of our study are to quantify the dust transport and deposition and to assess and improve model representations of dust processes through exploring the use of a suite of satellite remote sensing measurements and aircraft observations. Based on the 8-year (2007-2014) record of aerosol observations from CALIOP/CALIPSO, we estimate that the yearly dust deposition is 102, 20, and 28 Tg into the tropical Atlantic Ocean, Caribbean Basin, and Amazon Basin, respectively. The CALIOP observations also yields a dust residence time of about 10 days in the trans-Atlantic region. The GEOS-5 model overestimates the dust deposition and gives a factor of 2 shorter residence time in the atmosphere. We further evaluate the GEOS-5 model representations of size-dependent dust processes through comparing the model results against Fennec aircraft observations and satellite observations of dust optical depth (DOD) in the mid-visible (VIS) and thermal infrared (TIR). The GEOS-5 model considering dust particles of 0.2 – 20 μm in diameter only accounts for about 50% of observed dust mass over the Sahara. The model also substantially overestimates the mass fraction for small particles (< 6 μm) but underestimates the fraction for large particles (> 6 μm). The DOD gradient along the trans-Atlantic transit reflects the removal efficiency of dust during the transport. The larger the DOD gradient, the faster the dust removal is. We found that the VIS DOD gradient as revealed by MODIS, MISR, and CALIOP is significantly smaller than the TIR DOD gradient by IASI. This satellite-observed VIS and TIR difference appears to be consistent with the understanding of preferential removal of large particles from the atmosphere, because IASI is mainly sensitive to large particles. However, the GEOS-5 model shows negligible difference in the DOD gradient between VIS and TIR. This satellite-model discrepancy may suggest model deficiencies in representing the size dependence of dust removal processes.