Kelly, P.J., Werner, C., Evans, W., Ingebritsen, S. and Tucker, D., 2012, Observations of Gas Emissions from Cascade Range Volcanoes (USA) using a Portable Real-Time Sensor Package and Evacuated Flasks, 2012 AGU Fall Meeting, abstract 1497591

Observations of Gas Emissions from Cascade Range Volcanoes (USA) using a Portable Real-Time Sensor Package and Evacuated Flasks, 2012 AGU Fall Meeting, abstract 1497591

Degassing from most Cascade Range Volcanoes, USA, is characterized by low-temperature hydrothermal emissions. It is important to monitor these emissions as part of a comprehensive monitoring strategy yet access is often difficult and most features are sampled by the USGS only once per year at best. In an effort to increase the sampling frequency of major gas species and in preparation for building permanent, autonomous units, we built a portable sensor package capable of measuring H2O, CO2, SO2, and H2S in volcanic gas plumes. Here we compare results from the portable sensor package with gas analyses from direct samples obtained using a titanium tube and evacuated glass flasks collected at the same time. The sensor package is housed in a small, rugged case, weighs 5 kg, and includes sensors for measuring H2O (0-16 parts per thousand), CO2 (0-5000 ppmv), SO2 (0-100 ppm), and H2S (0-20 ppm) gases. Additional temperature and pressure sensors, a micro air pump, datalogger, and an internal battery are also incorporated. H2O and CO2 are measured using an infrared spectrometer (Licor 840) and sulfur-containing gases are measured using electrochemical sensors equipped with filters to mitigate cross-sensitivities. Data are collected at a 1 Hz sampling rate and can be recorded and displayed in real-time using a netbook computer or can be saved to the onboard datalogger. The data display includes timeseries of H2O, CO2, SO2, and H2S mixing ratios, the four-component bulk composition of the plume, and automated calculation of gas ratios commonly used in volcanic gas monitoring, such as H2O/CO2, CO2/SO2, and CO2/H2S . In the Cascade Range, the sensor package has been tested at Mt. Baker, Mt. St. Helens, Mt. Hood, and in Lassen Volcanic National Park. In each case, the instrument was placed 5 to 30 meters from the fumarole or fumarole field and emissions were sampled for 5 to 30 minutes. No SO2 was detected at any location. At Mt. Hood the sensor package yielded average CO2/H2S ratios from 10 to 16 in fumarole plumes versus flask CO2/H2S ratios (n = 2) of 13 and 16 on 9 July 2011, and on 28 July 2012 the sensor package yielded an average CO2/H2S ratio of 12 versus flask ratios (n = 2) of 13 (both sets of flask samples obtained in the Crater Rock area). At Mt. Baker, the sensor package yielded average CO2/H2S ratios from 19 to 22 whereas flask ratios (n = 3) were higher, from 25 to 32 (both fumarole-plume and flask samples obtained in the Sherman Crater area) on 22 July 2011. The mismatch falls slightly outside expected analytical uncertainty for the sensor package (about 20% relative for CO2/H2S ratios). However, flask samples collected in Sherman Crater in 2006 and 2007 (n = 5) yielded CO2/H2S ratios from 18 to 29, which nearly spans the range of observations in 2011. Therefore, one explanation for the small mismatch between the results of the sensor package and direct samples is that the sensor package measures bulk plume compositions that may integrate emissions from several chemically distinct fumaroles and the direct samples better represent the composition of discrete vents. Overall, the sensor package and evacuated flask data show good agreement and demonstrate that the real-time technique is a viable means for monitoring major volcanic gas species.