Articles | Volume 6, issue 1
Research article
09 Jan 2013
Research article |  | 09 Jan 2013

Volcanic SO2 and SiF4 visualization using 2-D thermal emission spectroscopy – Part 2: Wind propagation and emission rates

A. Krueger, W. Stremme, R. Harig, and M. Grutter

Abstract. A technique for measuring two-dimensional (2-D) plumes of volcanic gases with thermal emission spectroscopy was described in Part 1 by Stremme et al. (2012a). In that paper the instrumental aspects as well as retrieval strategies for obtaining the slant column images of SO2 and SiF4, as well as animations of particular events observed at the Popocatépetl volcano, were presented. This work focuses on the procedures for determining the propagation speed of the gases and estimating an emission rate from the given image sequences. A 2-D column density distribution of a volcanic gas, available as time-consecutive frames, provides information of a projected wind field and the average velocity at which the volcanic plume is propagating. This information is valuable since the largest uncertainties when calculating emission rates of the gases using remote sensing techniques arise from propagation velocities which are often inadequately assumed. The presented reconstruction method solves the equation of continuity as an ill-posed problem using mainly a Tikhonov-like regularisation. It is observed from the available data sets that if the main direction of propagation is perpendicular to the line-of-sight, the algorithm works well for SO2, which has the strongest signals, and also for SiF4 in some favourable cases. Due to the similarity of the algorithm used here with the reconstruction methods used for profile retrievals based on optimal estimation theory, diagnostic tools like the averaging kernels can be calculated in an analogous manner and the information can be quantified as degrees of freedom. Thus, it is shown that the combination of wind field and column distribution of the gas plume can provide the emission rate of the volcano both during day and night.