Using volcanic gases to understand open-vent volcanoes

Abstract

Open-vent volcanoes include the most active on Earth and are characterized by their persistent outgassing and the occurrence of moderate explosive activity between major eruptions. Forecasting eruptive activity at open-vent volcanoes is particularly challenging owing to the small changes in geophysical parameters which typically proceed eruptions. Changes in the emission rate and composition of plume gases however, presents a critical opportunity to detect changes before the arrival of magma at the surface. Subsequent identification of periodic trends implies that periodicities are an underlying component of volcanic plume degassing which could provide fundamental insights into the processes controlling volcanic degassing, and in turn, aid our ability to forecast eruptions. One of the only ways to obtain a continuous record of the plume gas is by ground-based remote sensing using UV-spectrometers where the favoured approach for quantifying emissions of SO2 is by Differential Optical Absorption Spectroscopy, a technique based on the Beer-Lambert-Bouguer law which relates the attenuation of light to properties of the material through which it has travelled. The challenge for volcanologists arises due to the absorption signatures of several trace gases in the spectra, which are present in various unknown quantities, many of which absorb light at shared wavelengths. To isolate the absorption due to SO2 from that of other trace gases, as well as atmospheric and instrumentals effects, a complex fitting routine is required where expertise is not only essential for the selection of suitable fit parameters but also for visual inspection of the fit results. Here, I present a novel approach which instead exploits the well-defined spatial frequencies in UV spectra. Although the wavelength ranges of trace gas absorption cross sections overlap, the occurrence of their narrow-band absorption features is not uniform and when viewed in the wavenumber domain their absorption signatures are distinct. I present a simple linear model to quantify SO2 absorption by detection of its wavelet transform and demonstrate its ability to return expected slant column densities of SO2 which are comparable to those achieved using a traditional DOAS algorithm. I further capitalise on the spatial frequencies contained in the recorded UV spectra to track volcanic SO2 without the need for external reference spectra. By correlating spectra in the frequency domain, I provide a simple means to identify spectra which contain the absorption signature of the volcanic plume. Use of wavelet coherence provides an alternative means to verify fit results and to establish a suitable wavelength range for spectral analysis. I test the possibility to analyse spectra for trace gases which are typically present at much lower concentrations, and explore the use of the linear model to quantify bromine monoxide (BrO) and chlorine dioxide (OClO), which offer the potential to detect magma emplaced at shallow depth. Finally, I test the existence of periodicities in the SO2 emission rate from a range of volcanoes with vastly different characteristics. I use time series data acquired by the Network of Volcanic and Atmospheric Change (NOVAC) and the Lomb-Scargle periodogram, to identify significant periodicities in the SO2 emission rates of 17 of the 28 volcanoes analysed. However, I find that most of these periodicities are also present in the plume speeds used to determine SO2 emission rates and are related to intra- and inter-seasonality in global trade winds and not volcanic in origin.

Kindly be aware of the following typographical error. The designation 'Network of Volcanic and Atmospheric Change (NOVAC)' is intended to be expressed as 'Network for Observation of Volcanic and Atmospheric Change (NOVAC).'