Title: Dynamics of Outgassing and Plume Transport Revealed by Proximal Unmanned Aerial System (UAS) Measurements at Volcán Villarrica, Chile
Abstract: Geochemistry, Geophysics, GeosystemsVolume 20, Issue 2 p. 730-750 Research ArticleFree Access Dynamics of Outgassing and Plume Transport Revealed by Proximal Unmanned Aerial System (UAS) Measurements at Volcán Villarrica, Chile Emma J. Liu, Corresponding Author Emma J. Liu [email protected] orcid.org/0000-0003-1749-9285 Department of Earth Sciences, University of Cambridge, Cambridge, UK Correspondence to: E. J. Liu, [email protected] for more papers by this authorKieran Wood, Kieran Wood orcid.org/0000-0002-5804-7704 Department of Aerospace Engineering, University of Bristol, Bristol, UKSearch for more papers by this authorEmily Mason, Emily Mason orcid.org/0000-0002-7050-6475 Department of Earth Sciences, University of Cambridge, Cambridge, UKSearch for more papers by this authorMarie Edmonds, Marie Edmonds orcid.org/0000-0003-1243-137X Department of Earth Sciences, University of Cambridge, Cambridge, UKSearch for more papers by this authorAlessandro Aiuppa, Alessandro Aiuppa orcid.org/0000-0002-0254-6539 Dipartimento DiSTeM, Università di Palermo, Palermo, Italy Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, ItalySearch for more papers by this authorGaetano Giudice, Gaetano Giudice orcid.org/0000-0002-9410-4139 Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, ItalySearch for more papers by this authorMarcello Bitetto, Marcello Bitetto orcid.org/0000-0003-0460-9772 Dipartimento DiSTeM, Università di Palermo, Palermo, ItalySearch for more papers by this authorVincenzo Francofonte, Vincenzo Francofonte Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, ItalySearch for more papers by this authorSteve Burrow, Steve Burrow Department of Aerospace Engineering, University of Bristol, Bristol, UKSearch for more papers by this authorThomas Richardson, Thomas Richardson orcid.org/0000-0001-7767-452X Department of Aerospace Engineering, University of Bristol, Bristol, UKSearch for more papers by this authorMatthew Watson, Matthew Watson School of Earth Sciences, University of Bristol, Wills Memorial Building, UKSearch for more papers by this authorTom D. Pering, Tom D. Pering orcid.org/0000-0001-6028-308X Department of Geography, University of Sheffield, Winter Street, UKSearch for more papers by this authorThomas C. Wilkes, Thomas C. Wilkes orcid.org/0000-0002-3448-6067 Department of Geography, University of Sheffield, Winter Street, UKSearch for more papers by this authorAndrew J. S. McGonigle, Andrew J. S. McGonigle orcid.org/0000-0002-0234-9981 Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, Italy Department of Geography, University of Sheffield, Winter Street, UK School of Geosciences, University of Sydney, Sydney, AustraliaSearch for more papers by this authorGabriela Velasquez, Gabriela Velasquez Observatorio Volcanológico de los Andes del Sur (OVDAS), Red Nacional de Vigilancia Volcánica (RNVV), Servicio Nacional de Geología y Minería, Temuco, ChileSearch for more papers by this authorCarlos Melgarejo, Carlos Melgarejo Observatorio Volcanológico de los Andes del Sur (OVDAS), Red Nacional de Vigilancia Volcánica (RNVV), Servicio Nacional de Geología y Minería, Temuco, ChileSearch for more papers by this authorClaudia Bucarey, Claudia Bucarey Observatorio Volcanológico de los Andes del Sur (OVDAS), Red Nacional de Vigilancia Volcánica (RNVV), Servicio Nacional de Geología y Minería, Temuco, ChileSearch for more papers by this author Emma J. Liu, Corresponding Author Emma J. Liu [email protected] orcid.org/0000-0003-1749-9285 Department of Earth Sciences, University of Cambridge, Cambridge, UK Correspondence to: E. J. Liu, [email protected] for more papers by this authorKieran Wood, Kieran Wood orcid.org/0000-0002-5804-7704 Department of Aerospace Engineering, University of Bristol, Bristol, UKSearch for more papers by this authorEmily Mason, Emily Mason orcid.org/0000-0002-7050-6475 Department of Earth Sciences, University of Cambridge, Cambridge, UKSearch for more papers by this authorMarie Edmonds, Marie Edmonds orcid.org/0000-0003-1243-137X Department of Earth Sciences, University of Cambridge, Cambridge, UKSearch for more papers by this authorAlessandro Aiuppa, Alessandro Aiuppa orcid.org/0000-0002-0254-6539 Dipartimento DiSTeM, Università di Palermo, Palermo, Italy Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, ItalySearch for more papers by this authorGaetano Giudice, Gaetano Giudice orcid.org/0000-0002-9410-4139 Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, ItalySearch for more papers by this authorMarcello Bitetto, Marcello Bitetto orcid.org/0000-0003-0460-9772 Dipartimento DiSTeM, Università di Palermo, Palermo, ItalySearch for more papers by this authorVincenzo Francofonte, Vincenzo Francofonte Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, ItalySearch for more papers by this authorSteve Burrow, Steve Burrow Department of Aerospace Engineering, University of Bristol, Bristol, UKSearch for more papers by this authorThomas Richardson, Thomas Richardson orcid.org/0000-0001-7767-452X Department of Aerospace Engineering, University of Bristol, Bristol, UKSearch for more papers by this authorMatthew Watson, Matthew Watson School of Earth Sciences, University of Bristol, Wills Memorial Building, UKSearch for more papers by this authorTom D. Pering, Tom D. Pering orcid.org/0000-0001-6028-308X Department of Geography, University of Sheffield, Winter Street, UKSearch for more papers by this authorThomas C. Wilkes, Thomas C. Wilkes orcid.org/0000-0002-3448-6067 Department of Geography, University of Sheffield, Winter Street, UKSearch for more papers by this authorAndrew J. S. McGonigle, Andrew J. S. McGonigle orcid.org/0000-0002-0234-9981 Istituto Nazionale di Geofisica e Vulcanologia, Sezione di Palermo, Palermo, Italy Department of Geography, University of Sheffield, Winter Street, UK School of Geosciences, University of Sydney, Sydney, AustraliaSearch for more papers by this authorGabriela Velasquez, Gabriela Velasquez Observatorio Volcanológico de los Andes del Sur (OVDAS), Red Nacional de Vigilancia Volcánica (RNVV), Servicio Nacional de Geología y Minería, Temuco, ChileSearch for more papers by this authorCarlos Melgarejo, Carlos Melgarejo Observatorio Volcanológico de los Andes del Sur (OVDAS), Red Nacional de Vigilancia Volcánica (RNVV), Servicio Nacional de Geología y Minería, Temuco, ChileSearch for more papers by this authorClaudia Bucarey, Claudia Bucarey Observatorio Volcanológico de los Andes del Sur (OVDAS), Red Nacional de Vigilancia Volcánica (RNVV), Servicio Nacional de Geología y Minería, Temuco, ChileSearch for more papers by this author First published: 28 December 2018 https://doi.org/10.1029/2018GC007692Citations: 30AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract Volcanic gas emissions are intimately linked to the dynamics of magma ascent and outgassing and, on geological time scales, constitute an important source of volatiles to the Earth's atmosphere. Measurements of gas composition and flux are therefore critical to both volcano monitoring and to determining the contribution of volcanoes to global geochemical cycles. However, significant gaps remain in our global inventories of volcanic emissions, (particularly for CO2, which requires proximal sampling of a concentrated plume) for those volcanoes where the near-vent region is hazardous or inaccessible. Unmanned Aerial Systems (UAS) provide a robust and effective solution to proximal sampling of dense volcanic plumes in extreme volcanic environments. Here we present gas compositional data acquired using a gas sensor payload aboard a UAS flown at Volcán Villarrica, Chile. We compare UAS-derived gas time series to simultaneous crater rim multi-GAS data and UV camera imagery to investigate early plume evolution. SO2 concentrations measured in the young proximal plume exhibit periodic variations that are well correlated with the concentrations of other species. By combining molar gas ratios (CO2/SO2 = 1.48–1.68, H2O/SO2 = 67–75, and H2O/CO2 = 45–51) with the SO2 flux (142 ± 17 t/day) from UV camera images, we derive CO2 and H2O fluxes of ~150 t/day and ~2,850 t/day, respectively. We observe good agreement between time-averaged molar gas ratios obtained from simultaneous UAS- and ground-based multi-GAS acquisitions. However, the UAS measurements made in the young, less diluted plume reveal additional short-term periodic structure that reflects active degassing through discrete, audible gas exhalations. Key Points We present high-resolution gas compositional data acquired using a multirotor Unmanned Aerial System (UAS) at Volcán Villarrica, Chile We derive SO2, CO2, and H2O fluxes of ~162, ~150, and ~2850 t/day, respectively UAS gas measurements made in the young, undiluted plume reveal short-term periodic active degassing associated with audible gas exhalations 1 Introduction The gases released by volcanoes at the Earth's surface are a window into the magmatic systems beneath. Abrupt changes in gas composition have been shown to occur immediately prior to large “paroxysmal” eruptions at several arc volcanoes worldwide, with gas ratios such as CO2/SO2 identified in hindsight as timely forecasts of shifts in eruptive behavior (Aiuppa et al., 2007; Aiuppa, Bertagnini, et al., 2010; Aiuppa, Burton, et al., 2010; Aiuppa, Fischer, et al., 2017; de Moor et al., 2016; Shinohara, 2005; Shinohara et al., 2008; Werner et al., 2008). Cyclical variations in gas flux, particularly when cross-correlated with other monitoring parameters such as seismicity or ground deformation, provide critical insight into the mechanisms governing the time scales of recurrent eruptive activity (Flower & Carn, 2015; Ilanko et al., 2015; Nicholson et al., 2013; Odbert et al., 2014; Tamburello et al., 2013; Werner et al., 2008) and the data required to test hypotheses generated by numerical conduit models (e.g., Costa et al., 2007). On geological time scales, quantification of degassing budgets on a regional- or global-scale offers important constraints on volatile cycling through the Earth system (Aiuppa, Bitetto, et al., 2017; Mason et al., 2017). However, there remain significant gaps in our global inventories of volcanic gas emissions, particularly for CO2 and other species which usually require proximal sampling of a concentrated plume, for those volcanoes where the near-vent region is hazardous or inaccessible. Measurements of volcanic gas emissions can be either direct or remote. Direct measurements requiring placement of an instrument close to the vent source by volcanologists are hazardous to acquire and the instruments may often be destroyed during eruptions (e.g., Aiuppa, Fischer, et al., 2017). However, the high gas concentrations and limited atmospheric entrainment in young, proximal plumes yield high quality data that record faithfully primary degassing processes at a temporal resolution approaching that of geophysical data (≤1 Hz). Remote measurements, such as can be acquired using imaging or spectral techniques (e.g., UV/IR-cameras and COSPEC/scanning-differential optical absorption spectroscopy (DOAS)/FlySpec/mobile-DOAS, respectively), are typically performed several kilometers from the volcanic gas source and therefore pose little to no direct risk to volcanologists or equipment (e.g., Bluth et al., 2007; Edmonds et al., 2003; Galle et al., 2003; Holland et al., 2011; Horton et al., 2006; McGonigle et al., 2002, 2003; Moffat & Millan, 1971; Mori & Burton, 2006; Oppenheimer et al., 1998; Platt et al., 2018; Tamburello et al., 2012, 2013; Weibring et al., 1998; Wilkes et al., 2016). Note that, although significant progress has been made in the use of Light Detection And Ranging (LIDAR) for remote sensing of CO2 (Santoro et al., 2017), SO2 is currently the only gas species able to be measured routinely by remote methods due to its strong absorption at UV wavelengths and the negligible concentration in ambient air. Globally, volcanic SO2 emissions are routinely monitored from space using satellite-based instrumentation (such as the Ozone Monitoring Instrument), or by ground-based networks at specific volcanoes (for example, the NOVAC network; Galle et al., 2010). In contrast, data for other gas species such as H2O and CO2 are lacking for many volcanic systems where proximal measurements are challenging, particularly in a monitoring capacity. Unmanned Aerial Vehicles (UAVs) are now bridging the gap between direct and remote measurements by enabling proximal sampling from a safe and accessible distance. Recent developments in drone technology (in terms of both physical capability and user-accessibility) have been matched by a drive toward increasingly lightweight and compact sensor payloads, such that the resulting Unmanned Aerial Systems (UAS) are rapidly becoming “go-to” solutions for a wide range of volcanological applications. UAS are driving the greatest advances in those fields requiring either proximal measurements in extreme environments or large areal coverage, including lava flow mapping, constructing topographic models, and eruptive volume estimations (Darmawan et al., 2018; Favalli et al., 2018; Moussallam et al., 2016; Müller et al., 2017; Nakano et al., 2014; Turner et al., 2017), post-eruption visual observation (Koyama et al., 2013), thermal imaging (Di Stefano et al., 2018), aeromagnetic surveys (Hashimoto et al., 2014; Kaneko et al., 2011), DOAS traverses for SO2 flux determination, and volcanic gas measurements and sampling (Diaz et al., 2015; Di Stefano et al., 2018; McGonigle et al., 2008; Mori et al., 2016; Pieri et al., 2013; Rüdiger et al., 2018; Shinohara, 2013; Stix et al., 2018). Plume compositions measured by UAS 1–2 km from the vent following the 2014 phreatic eruption of Mt. Ontake demonstrated a dominantly hydrothermal degassing source (low SO2/H2S molar ratios, combined with SO2 concentrations <1 ppm) and were critical to safe monitoring of the post-eruptive state of the volcano (Mori et al., 2016). At Turrialba volcano, Costa Rica, SO2 concentrations measured by UAS in the dilute distal plume (up to 3 km from the vent, measuring 0.3–20 ppm SO2) were used to derive estimates of SO2 emission rates using an inverse Bayesian modeling approach incorporating meteorological wind fields (Xi et al., 2016). A recent comparison of SO2 fluxes obtained by traditional ground-based DOAS traverses to those from drone-mounted DOAS (DROAS) demonstrated the utility of UAS for this application, with the additional ability to constrain wind speeds at plume altitude based on passive drift speed of the UAS (Stix et al., 2018). Furthermore, comparison of CO2/SO2 molar ratios obtained by ground-based multi-GAS and simultaneously by a UAS-based gas sensor unit at Masaya, Nicaragua (Rüdiger et al., 2018; Stix et al., 2018), showed that downwind measurements using UAS in dilute plumes yield robust correlations between gas species that are comparable to those derived from proximal ground-based measurements. However, molar ratios in distal plumes (<5 ppm maximum SO2) can be characterized by larger standard deviations than proximal crater rim measurements (<38 ppm SO2; Rüdiger et al., 2018), highlighting that, although feasible, challenges still remain in acquiring high quality data from dilute plumes. This is particularly true for gas species such as CO2, which have much higher background concentrations in ambient air (e.g., Aiuppa et al., 2009). Recently, Rüdiger et al. (2018) demonstrate that UAS now enable the measurement of reactive halogen species in previously inaccessible downwind plume regions and present proof-of-concept data that suggest an interesting potential relationship between Brx/SO2 and CO2/SO2 at Stromboli, Italy. In situ measurements of volcanic gas composition are particularly suited to a UAS-based approach as sampling young, high-temperature plumes minimizes the effects of atmospheric interactions that dilute and chemically modify the gas composition, and thus reduce the fidelity of the signal. Here we investigate short-timescale spatial and temporal variability in CO2, H2O, SO2, H2S, and HCl emissions within a very young plume at Volcán Villarrica, Chile. Villarrica is historically the most active volcano in Chile, and is one of only six volcanoes worldwide to host an open lava lake at its summit. It is a persistent emission source, and exhibited a notable compositional excursion in gas ratios prior to the last major eruption in March 2015 (Aiuppa, Fischer, et al., 2017). The accessibility to the open summit vent, together with the permanent multi-GAS instrument located on the crater rim for data comparison, makes Villarrica an ideal natural laboratory for UAS field testing. Specifically, we present UAS-measured gas molar ratios in combination with SO2 flux estimates derived from remote UV camera measurements, to quantify periodicities in gas composition and flux (on time scales of tens of seconds to minutes). We describe the development and deployment of two instrumented UAS: a customized Vulcan Black Widow multirotor UAV carrying a live-telemetered multi-GAS payload, designed specifically to withstand challenging environmental conditions during static hovers in the concentrated plume, and a modular sensor attachment (Airgraph Aeris commercial prototype) to a DJI Phantom 3 Pro quadcopter offering agile mobility in the plume for large-scale SO2 mapping. The results presented in this study highlight previously unrecognized periodic degassing behavior at Volcán Villarrica and demonstrate further the validity of UAS for high-resolution studies and routine monitoring of volcanic emissions, more generally. Together with a suggested pre-flight checklist for UAV safe operation published as supporting information, this work contributes a solid foundation on which to further advance the use of UAS for volcanic gas sensing, with future research focused on achieving long-range missions (>3 km) and coordinated multi-UAS experiments at strongly degassing but inaccessible volcanoes. 2 Geological Setting Volcán Villarrica (2,847 m AMSL) is a partially glaciated stratovolcano within the Southern Volcanic Zone (SVZ) of the Andes of Chile (Figure 1). The SVZ is a relatively carbon-poor volcanic arc segment compared to other arcs globally (Aiuppa, Bitetto, et al., 2017; Shinohara & Witter, 2005), with limited involvement of subducting slab-derived fluids in magma genesis (Jacques et al., 2013; Wehrmann et al., 2014). Villarrica's volcanic edifice hosts a persistently degassing open lava lake at its summit, which is periodically perturbed by Strombolian explosions and transient lava fountaining (Calder et al., 2004; Palma et al., 2008). Erupted magma compositions range from basaltic to basaltic andesite (50–57 wt% SiO2; Hickey-Vargas et al., 2004; Moreno et al., 1994; Witter et al., 2004), and include several mafic ignimbrites emplaced during Plinian eruptions during the Holocene (Costantini et al., 2011; Parejas et al., 2010). Historical eruptions, documented since 1558, have been predominantly characterized by mild/moderate explosive fountaining with occasional lava effusion. The surface of the lava lake fluctuates from <50 to >200 m below the crater rim on monthly time scales, and indicates the top of the magma column that resides in the main conduit (Calder et al., 2004; Moussallam et al., 2016; Palma et al., 2008; Richardson et al., 2014). Variations in lava lake level are broadly correlated with both seismicity and degassing flux: periods when the lava level is high in the conduit are typically associated with more vigorous bubble bursting activity, higher SO2 fluxes, and elevated Real-time Seismic-Amplitude Measurement (RSAM) seismic amplitude (Palma et al., 2008). Figure 1Open in figure viewerPowerPoint Location and volcanological setting; (a) Volcán Villarrica is the easternmost in a chain of three volcanoes (Lanin, Quetrapillan) striking obliquely to the main N-S axis of volcanism; (b, c) The magma level was extremely low in the conduit during the measurement campaign, with the lake surface only visible as several pixels in aerial imagery; (d) Unmanned Aerial Systems (UAS) were launched from a sheltered plateau on the northern rim of the crater, with the semipermanent multi-GAS station visible on the eastern rim; (e) location map of the region, showing the position of UV camera. The green shaded region delimits the extent of the national park. Inset: Aerial map of the summit region shown in (d). The summit crater is ~200 m in diameter; (e) Two instrumented multirotor vehicles were used in this campaign, the Vulcan octocopter with multi-GAS (left) and DJI Phantom 3 Pro with Aeris gas sensor (right); (f) Vulcan UAS in flight on 20 March 2018. UAV = Unmanned Aerial Vehicle. Villarrica is a persistent emission source within the SVZ and has maintained an SO2 flux on the order of a few hundred tons per day (t/day) during several campaign measurements since 2005 (Mather et al., 2004; Moussallam et al., 2016; Sawyer et al., 2011; Shinohara & Witter, 2005), and by a permanent DOAS network since 2010 maintained by the Observatorio Volcanológico de los Andes del Sur (OVDAS). From January 2010 through May 2012, the lava lake surface was continuously visible and SO2 emission rates averaged 926 t/day (OVDAS, personal communication, April 1, 2018). Following subsidence of the lava lake out of view in June 2012, the emission rate reduced to an average of 386 t/day until February 2015 (OVDAS, personal communication). Prior to the paroxysmal eruption of 3 March 2015, volcanic infrasound signals recorded a rapid increase in the level of the magma free surface from a stable level at >120 m below the crater rim to <70 m in the days prior to the eruption (Johnson et al., 2018). Following the eruption, the lake level remained high, maintaining an average SO2 flux of 714 t/day until December 2017 when the magma column again withdrew out of sight (OVDAS, personal communication). A permanent, fully autonomous multi-GAS station installed on the eastern side of the summit crater detected a significant compositional excursion in molar gas ratios toward more CO2-rich compositions immediately prior to the last major eruption in March 2015 (Aiuppa, Fischer, et al., 2017). Although the instrument was destroyed during the intense lava fountaining activity of 3 March, a replacement station was reinstalled in late November 2017, and acquires data during four 30-min sampling windows each day. Previous campaign measurements using ground-based multi-GAS instruments at Villarrica (Aiuppa, Fischer, et al., 2017; Moussallam et al., 2016; Shinohara & Witter, 2005) have not identified any clear periodic gas compositional variations associated with observed lava lake dynamics, but it remains unresolved whether this reflects the true degassing signature or is the result of homogenization and dilution during plume transport to the crater rim. 3 Methods 3.1 UAS Design Volcanic gas sensing requires a vehicle that is robust and resilient in the field, resists acid gas corrosion, and has sufficient propulsion and mobility to remain stable in strong winds. In response to these criteria, together with the need for modular attachment of sensor packages, we developed a bespoke UAS based on a multirotor platform. The multirotor configuration was selected over a fixed-wing design for this field campaign due to the need to maintain a static hover at a single position in the plume and to approach the vent in close proximity. The vehicle was an octocopter in the X8 configuration based on a Vulcan “Black Widow” frame with hub-to-hub diameter of 120 cm (Vulcan UAV, United Kingdom). Lift was provided by eight 16-inch (407 mm) propellers with a hub-to-hub dimension of 140 cm. The mass of the frame, instrument payload, and batteries were 6.2, 0.8, and 3.5 kg respectively, resulting in a combined take-off weight of 10.5 kg. Maximum take-off weight is 16 kg. The vehicle used six 4,250 mAh capacity batteries, each with a nominal voltage of 24 V (6S). When testing at an altitude of 250 m AMSL prior to summit ascent, the vehicle comfortably achieved a 13-min flight duration with large capacity margins remaining in the battery. Given the performance degradation expected with the increased altitude (2,847 m) and wind speed at the summit, flight durations were conservatively limited to 13 min or until the battery voltage dropped below 22 V, whichever occurred sooner. The avionics comprised several commercially available products, selected for their reliability and long-range capabilities. The flight computer is a Pixhawk 2.1 auto-pilot (Hex Technology, Hong Kong) with associated GPS module. All flight critical electronics were housed in sealed enclosures to reduce exposure to acidic volcanic gases and prevent corrosion. For the flight computer, a small hole was required in the case to allow for pressure equalization for the barometric altimeter, however the hole was sufficiently small to prevent significant airflow. Three separate radio-frequency links were used to communicate with the vehicle. The primary pilot control link used a transceiver set (Dragon Link, United States), operating on the 433 MHz frequency. An on-board video link used a transmitter and receiver set (ImmersionRC) on the 2.4 GHz frequency. The live video stream was made available to both the pilot and ground station via a first-person-view headset and handheld screen, thus allowing the vehicle to be visually positioned in the dense plume. The third link provided a stream of live flight data using RFD868+ radio modems (RFDesign, Australia) operating on the 868 MHz frequency. Transmitted live data included information on the vehicle status, such as battery voltage and altitude, and real-time gas concentrations from the on-board multi-GAS sensor. The transmission of live sensor data to the ground station was achieved by transcoding the digital multi-GAS serial messages into the commonly used MAVLink protocol using a Teensy 3.6 microcontroller. 3.2 Multi-GAS Concentrations of CO2, SO2, and H2S (along with pressure, temperature, and relative humidity) were measured at a 1 Hz sampling rate within the volcanic plume using a miniaturized multicomponent gas analyzer (multi-GAS; see Table S1 for detailed specifications of all components; Aiuppa et al., 2007, 2009; Shinohara, 2005), customized to be flown on a multirotor UAS (section 3.1). The CO2 spectrometer unit (non-dispersive infrared; NDIR) was wrapped in brass foil to shield the sensor board from radio-frequency interference from the UAS transmission system. H2O concentrations were calculated from records of temperature and relative humidity measured on-board the UAS, using a time-average pressure of 724 ± 0.5 mbar. Air was sampled through a 1 μm particle filter exposed to ambient air, at pump rate of 1.0 L/min. The multi-GAS was calibrated with standard references gases at INGV Palermo 2 weeks prior to the field campaign, and again 2 weeks after. No significant sensor drift requiring data correction was identified. All sensor data were logged on-board, and also telemetered directly to the ground station where it could be visualized in real-time. Full specifications of the permanent ground-based multi-GAS station are given in Aiuppa, Bitetto, et al. (2017). Multi-GAS concentration time series were post-processed using Ratiocalc software (Tamburello, 2015). CO2 concentrations were corrected internally for temperature (±0.2% full span per degree Celsius) and pressure (±0.15% per hPa). No pressure correction was applied to SO2 or H2S time series; however, applying the manufacturer compensation of 0.01% (SO2) and 0.008% (H2S) signal per mbar to a subset of the data shows a +2.9% increase in SO2 concentrations (Figure S1; Kelly, 2017). This pressure effect translates to a maximum underestimation of 3.5 ppm at 120 ppm SO2, and <1 ppm underestimation at <35 ppm SO2. Importantly, however, barometric pressure varied by <2 mbar over the duration of the flight, so the temporal properties of the time series cannot be attributed to pressure fluctuations. Volcanogenic CO2 was resolved from atmospheric background by subtracting the CO2 concentration in ambient air (450 ± 5 ppmv; measured outside the plume where SO2 = 0) from the raw CO2 time series. No baseline drift correction was required for any gas species. H2S concentrations were corrected for 13% cross-sensitivity to SO2 (H2Scorrected; Tamburello, 2015), where the magnitude of the cross-sensitivity was determined from laboratory tests using standard reference gases. Molar ratios (CO2/SO2, H2O/CO2, and CO2/H2O) were derived from gas-gas scatterplots by calculating the gradient of the best-fit linear regression line through the data. Datapoints where SO2 is present at <5 ppmv were excluded from the regression due to the greater error associated with very dilute plumes (e.g., Aiuppa et al., 2009); and >120 ppm due to the breakdown of the calibration curve above this concentration (the specific SO2 sensor model used here is expected to exhibit a linear response in the 0–100 ppm ± 20% range). Uncertainties in derived molar gas ratios are ≥6.4% at >10 ppm SO2 level and 12.5% at <10