Saharan dust and biomass burning aerosols during ex-hurricane Ophelia: validation of the new UK lidar and sun-photometer network

Abstract

On 15-16 October 2017, ex-hurricane Ophelia passed to the West of the British Isles, bringing dust from the Sahara and smoke from Portuguese forest fires that was observable to the naked eye and reported in the national press. We report here detailed observations of this event using the UK operational lidar and sunphotometer network, established for the early detection of aviation hazards. We also use ECMWF ERA5 wind field data, and MODIS imagery to examine the aerosol transport. The observations, taken continuously over a period of 30 hours, show a complex picture, dominated by several aerosol layers at different times, and clearly correlated with the passage of different air-masses associated with the intense cyclonic system. A similar evolution was observed at several sites, with a time delay between them explained by their different location with respect to the storm. The event commenced with a shallow dust layer at 1-2 km in altitude, and culminated in a deep and complex structure that lasted 12 hours at each site, correlated with the storm’s warm sector. For most of the time, the aerosol detected was dominated by mineral dust mixtures, as highlighted by depolarisation measurements, but an intense biomass burning aerosol (BBA) layer was observed towards the end of the event, lasting around 3 hours at each site. The aerosol optical depth at 355nm (AOD355) during the whole event ranged from 0.2 to 2.9, with the larger AOD correlated to the intense BBA layer. Such a large AOD is unprecedented in the United Kingdom according to AERONET records for the last 20 years. The Raman lidars permitted the measurement of the aerosol extinction coefficient at 355 nm, the particle depolarisation ratio (PLDR) and the lidar ratio (LR), and made possible the separation of the dust (depolarising) aerosol from other aerosol types. A specific extinction has also been computed to provide an estimate of the atmospheric concentration of both aerosol types separately, which peaked at 420±200 µgm−3 for the dust and 558±232 µgm−3 for the smoke. Back-trajectories computed using the Numerical Atmospheric dispersion Modelling Environment (NAME) were used to identify the sources and strengthen the conclusions drawn from the observations. The UK network represents a significant expansion of the observing capability in Northern Europe, with instruments evenly distributed across Great Britain, from Camborne in Cornwall to Lerwick in the Shetland islands, and this study represents the first attempt to demonstrate its capability and validate the methods in use. Its ultimate purpose will be the detection and quantification of volcanic plumes, but the present study clearly demonstrates the advanced capabilities of the network.