An extensive set of airborne and satellite observations of volcanic ash from the Eyjafjallajökull eruption have been analysed for a cloud-free case study over the southern North Sea on 17 May 2010. The Facility for Airborne Atmospheric Measurements (FAAM) BAe-146 aircraft made both in situ and remote measurements of the ash layers. Satellite retrievals of aerosol loadings and effective radii from observed Infrared Atmospheric Sounding Interferometer (IASI) spectra have recently been developed and are applied to two overpasses available on the 17 May 2010. The ‘radiative closure' study shows that the particle size distributions and optical properties derived from the in-situ aircraft observations are entirely consistent with the airborne solar and terrestrial radiation measurements and IASI satellite retrievals. The study also shows that for this case, refractive index and particle shape assumptions originally developed for mineral dust can be applied successfully in the derivation of ash optical properties and mass concentrations.

Particle size distributions derived from the FAAM aircraft wing-mounted optical particle counters (OPC) showed an ash mode between 0.6 – 35µm with a peak at ~4µm. Aerosol optical depths derived from the OPCs (ranging from 0.2-0.5) were consistent with an on-board nephelometer measuring aerosol scattering coefficients at three visible wavelengths as well as retrievals from a nadir-viewing lidar operating at 355 nm. Ash mass concentrations estimated from the OPC data peaked at around 500 µg/m3, though higher values of around 800 µg/m3 were inferred from the lidar as it surveyed the ash layers from above. The OPC and lidar data provided consistent column loadings of 0.1 – 0.75 g/m2 but with a high degree of spatial variability. IASI retrievals based on the measured OPC size distribution and mineral dust refractive index show similar column loadings on average but also a few outlying upper estimates reaching 1.7 g/m2 that may reflect differences in the timing and spatial sampling of the observations. The Airborne Research Interferometer Evaluation System (ARIES) showed decreases in measured infrared upwelling radiances, equivalent to several degrees in brightness temperature, over the ash layers. Broadband Eppley radiometers (BBR) showed an increase of around 60 Wm-2 in upwelling short-wave irradiance over the same ash layer, equivalent to an estimated radiative forcing efficiency of -130 Wm-2/τ. The spectrally resolved Short-wave Hemispherical Integrating Measurement System (SHIMS) showed this additional reflection to be well spread across the UV to near-IR (0.3 µm and 1.7 µm) suggesting little wavelength dependency to AOD across that wavelength interval. The observed solar irradiances and terrestrial radiances were well reproduced by radiative transfer models incorporating the optical properties derived from the OPC data and the observed AOD from the lidar data.

The sensitivity of modelling and retrievals to particle shape and refractive index assumptions was also tested. The default study was to assume irregular shapes for the coarse mode and the mineral dust refractive index in analysing the particle size distribution. Assuming the same refractive index but spheres rather than irregulars resulted in the size distribution peak increasing from 3.6 µm to 4.0 µm along with a slight broadening of the distribution. This yielded a 30% increase in ash mass derived from the OPCs while radiative retrievals were found to be fairly insensitive to the assumed shape across both the short-wave and the infrared.

Various refractive indices have been tested in the IASI retrieval, including that of desert dust used in the OPC analysis and those of other volcanic rocks. The desert dust refractive index was found to yield the best fit, giving confidence in its validity as an appropriate assumption in this case study.

The Deutsches Zentrum für Luft- und Raumfahrt (DLR) Falcon aircraft made measurements of the same ash cloud 1-3 hours later than the FAAM BAe-146. As a further test of the radiative closure, the size distribution reported by the DLR Falcon was applied to the simulations. The DLR Falcon coarse mode had a peak at 10 µm and could be represented by a log-normal with standard deviation 4.0. This was significantly larger and broader than the FAAM OPC size distribution. Even if the same assumptions were adopted in analysing the FAAM OPC data, the two could not be reconciled, pointing to spatial and/or temporal differences in the ash cloud and varying instrument sensitivities. Applying the DLR size distribution to simulations of the radiation measurements resulted in significantly larger residuals in all cases, across both the short-wave and infrared.