Detecting Saharan Air Layer Influence on Organized Convective Systems in northern Africa

The Saharan Air Layer (SAL) is a well-mixed layer of warm, dry, and dusty air of nearly constant mixing ratio generated by the intense surface heating and strong, dry convection in the Sahara Desert. Although generated in the Sahara, this layer's influence extends throughout northern Africa and into the eastern Atlantic Ocean with both radiative and dynamical consequences on the surface energy balance and the development of organized convective systems including tropical cyclones. Given the limited coverage of radiosonde stations in and around northern Africa, we confirmed that Level 2 humidity, temperature, and aerosol optical depth data products (2003-2015) from the NASA Aqua satellite's Atmospheric Infrared Sounder (AIRS), Atmospheric Microwave Sounder Unit (AMSU), and Moderate-resolution Imaging Spectroradiometer (MODIS) are able to resolve and characterize the bulk properties of elevated dry well-mixed layers over northern Africa. AIRS and AIRS/AMSU products under detect the occurrence of well-mixed layers (0.5 average bias) between 1000 and 500 hPa as indicated by radiosondes. However, both sets of AIRS data produce a realistic number (2-4 layers) and depth (300-4500 m) of detected well-mixed layers; successful detection rates average 55% over the 12 year study period. Well-mixed layer detection rates demonstrate a strong seasonal and regional dependence with higher detection rates (up to 80%) during March, April, and May and in areas closer to the Sahara (up to 90%). The remaining layers were either below the 300 m detection threshold, missed due to instrument detection error (1°C temperature and 20% relative humidity), or averaged out within the AIRS footprint. Modern reanalysis products from data assimilation systems are also capable of well-mixed layer detection, but the AIRS and AIRS/AMSU products had greater accuracy in detecting the height and depth of individual layers in the vertical profile.

In this study, we addressed several scientific questions: 1) What is the source of dry well-mixed air? Transport from the Sahara (a true SAL), other regions such as the Mediterranean, or local boundary layer processes? 2) For those non-Saharan stations with distinct SALs in their profile, what is the potential impact on convective initiation and development? Is there a statistically identifiable impact on convective potential? Is there anything particular about the SAL versus other sources of dry air? 3) Does the sensitivity of the AIRS product (approximately 300 m) affect the assessment of the convective potential?

Since AIRS launched in 2003, there are multiple years of data allowing us to build up a robust database for statistical analysis. To address question 1, we applied NOAA's Hybrid Single Particle Lagrangian Integrated Trajectory Model (HYSPLIT) model to ERA-Interim reanalysis wind profiles to trace trajectories at 54 synoptic stations around the Sahara Desert for AIRS profiles with dry well mixed layers. The results of question 1 made it possible to categorize different sources of dry air versus season at and around the study stations. For question 2, we compared the University of Utah's TRMM Precipitation Features database and gridded rainfall products to profiles with the different sources of dry air identified in the previous analysis. In this analysis, we used both radiosonde (when available) and AIRS data and tracked dry layer type, depth, and environmental aerosol optical depth (proxy for dustiness) to evaluate how likely and how much rainfall would occur then dry layers were present.