Retrieval of SO₂ from thermal infrared satellite measurements: correction procedures for the effects of volcanic ash

Abstract

Abstract. The simultaneous presence of SO2 and ash in a volcanic plume can lead to a significant error in the SO2 column abundance retrieval when multispectral Thermal InfraRed (TIR) data are used. The ash particles within the plume with effective radii from 1 to 10 μm reduce the Top Of Atmosphere (TOA) radiance in the entire TIR spectral range, including the channels used for SO2 retrieval. The net effect is a significant SO2 overestimation. In this work the interference of ash is discussed and two correction procedures for satellite SO2 volcanic plume retrieval in the TIR spectral range are developed to achieve an higher computational speed and a better accuracy. The ash correction can be applied when the sensor spectral range includes the 7.3 and/or 8.7 μm SO2 absorption bands, and the split window bands centered around 11 and 12 μm required for ash retrieval. This allows the possibility of simultaneous estimation of both volcanic SO2 and ash in the same data set. The proposed ash correction procedures have been applied to the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Spin Enhanced Visible and Infrared Imager (SEVIRI) measurements. Data collected during the 24 November 2006 Mt. Etna eruption have been used to illustrate the technique. The SO2 and ash estimation is carried out by using a best weighted least squares fit method and the Brightness Temperature Difference (BTD) procedures, respectively. The simulated TOA radiance Look-Up Table (LUT) needed for the SO2 column abundance and the ash retrievals have been computed using the MODTRAN 4 Radiative Transfer Model. The results show the importance of the ash correction on SO2 retrievals at 8.7 μm, where the corrected SO2 column abundance values are less than 50% of the uncorrected values. The ash correction on SO2 retrieval at 7.3 μm is much less important and only significant for low SO2 column abundances. Results also show that the simplified and faster correction procedure underestimates the ash correction compared with the more time consuming but more accurate correction procedure. Such underestimation is greater for instruments having better ground pixel resolution, i.e. greater for MODIS than for SEVIRI.