Corradini, S, Spinetti, C, Carboni, E, Tirelli, C, Buongiorno, MF, Pugnaghi, S, Gangale, G (2008). Mt. Etna tropospheric ash retrieval and sensitivity analysis using Moderate Resolution Imaging Spectroradiometer measurements. JOURNAL OF APPLIED REMOTE SENSING, 2, 23550.

Abstract
A retrieval of tropospheric volcanic ash from Mt Etna has been carried out, using measurements from the Moderate Resolution Imaging Spectroradiometer (MODIS). The NASA-MODIS satellite instrument acquires images in the 0.4 to 14 mu m spectral range with a spatial resolution of 1 km at nadir. The eruption which occurred on 24 November 2006 is considered as a test case in this work. In order to derive the ash plume optical thickness, the particle effective radius and the total mass, the Brightness Temperature Difference procedure has been applied to MODIS channels 31 (centered at 11 mu m) and 32 (centered at 12 mu m). Channel 5 (centered at 1.24 mu m) has been used to refine the cloud discrimination, exploiting the distinct reflectivity of meteorological and volcanic clouds in the near infrared spectral range. The detection of volcanic ash pixels has been significantly improved by applying an atmospheric water vapor correction to MODIS data. This procedure doubles the number of pixels identified as containing volcanic ash compared to the original method. The retrieved mean ash optical thickness at 0.55 mu m, mean particle effective radius and the total ash mass in the plume are 0.4, 3.5 mu m and 3620 tons, respectively. A detailed sensitivity analysis has been carried out to investigate errors in the retrieval caused by the uncertainty in various parameters: surface temperature and emissivity, plume geometry (altitude and thickness), ash type and atmospheric water vapor. Results show that the largest contributions to retrieval errors are from uncertainty in surface parameters, aerosol type and atmospheric water vapor. The total tropospheric volcanic ash retrieval errors are estimated to be 30%, 30% and 40% for mean AOT, mean effective radius and total mass retrieval, respectively.

DOI:
10.1117/1.3046674

ISSN:
1931-3195