Bromo activity over the last decade: consistent passive degassing and source magma evolution

Bromo is among the very active volcanoes in Indonesia and is known for its recurrent and long-lasting eruptive manifestations. Past volcanic gas studies have revealed Bromo as one of the principal sources of volcanic degassing in Indonesia. This high degassing from Bromo volcano is further characterized in this work, based on more than 10 years of intermittent ground-based gas measurements, combined with daily SO2 mass, captured by the OMI sensor. Over the past decade, Bromo has released 0.7 Tg of SO2 into the atmosphere, representing 3% of the volcanic degassing budget of Indonesia and 0.3% of the global volcanic SO2 emission budget outside eruptive periods. Results also reveal that 18.8 Tg of H2O, 2.0 Tg of CO2, 0.1 Tg of H2S, and 0.005 Tg of H2 were released from the Bromo volcano in one decade. About 81% of these gases are released passively between eruptive events. The chemistry of the eruptive products, sampled between 2001 and 2019, indicate that Bromo volcanic activity is sustained by a basaltic-andesite to basalt trachy-andesite magma source with a transition from medium-K to high-K composition. Such an evolution associated to a C-rich gas likely resulted from a low partial melting and sediment contribution to the genesis of the source magma. New magma injections into the reservoir and fractional crystallization have further amplified the changes of magma composition. Finally, we speculate that the shallow reservoir replenishment, in response to the continuous strong degassing is the driving mechanism behind the Bromo frequent eruptive events.


Introduction
Bromo is one of the most active volcanoes in Indonesia, located on the eastern part of Java Island ( Fig. 1) (Coord. 7.942°S/112.953°E). Over the last two centuries, more than 50 eruptions have been recorded on this volcano (GVP 2013), which broadly correspond to one eruption every three years. It is still unclearly when the volcanic activity commenced on Bromo, but according to 14 C datings, it has been active since the seventeenth century (Zaenudin 1990;Mulyadi 1992) and it is the most recent cone, formed after the Tengger Caldera formation, more than 45,000 years ago (van Gervan and Pichler 1995). Bromo is currently the only active cone in Tengger caldera with an open-vent activity that channels a continuous degassing into the atmosphere. It constitutes, with 4 other cones, namely Watangan, Widodaren, Sagarawedi, and Batok (Fig. 1), a volcanic complex in the caldera whose tephra deposits have filled the caldera floor and formed the so-called "sand sea caldera" (Abidin et al. 2004). Few available gas studies have evidenced Bromo as a strong source of volcanic degassing with a daily SO 2 emission rate varying between 14 and 166 t d −1 (Andres and Kasgnoc 1998;Bani et al. 2013;Aiuppa et al. 2015). The gas composition with its magmatic signature indicates a water-rich magma, typical of arc volcanoes and a CO 2 /S T ratio of 3.2 (Aiuppa et al. 2015), comparable to the mean value of 4.3 for the Indonesian magmatic gas (Hilton et al. 2002), and suggests a contribution of carbon-rich sediments to the magma genesis .
This work aims to characterize the evolution of volcanic degassing from Bromo and to determine how it relates to the subsurface magmatic processes based on the chemistry of the erupted products.

Methods
The first ground-based SO 2 flux measurements on Bromo were carried out in March 1995 using the correlation spectrometer (COSPEC) (see McGonigle and Oppenheimer 2003 for technical details). The instrument was positioned 2 km northeast of the cone, on the rim of the sand sea caldera, and performed a scan across the rising plume. Over the last 15 years, Bromo's SO 2 emission rate has regularly been measured, using a flyspec (e.g., Horton et al., 2006) since 2005, a DOAS (differential optical absorption spectroscopy; Galle et al. 2003) since 2013 (Bani et al. 2013), and a UV-Camera (Mori and Burton 2006) since 2014 (Aiuppa et al. 2015). These different approaches were carried out on a static mode at 2 km northeast of the active crater, at a similar position of COSPEC measurements in 1995 (Fig. 1). Wind speed was obtained either from the handheld anemometer for the early measurements or from cameras in the recent recordings . The most recent SO 2 flux measurements were carried out in 2018 and 2019 using a scanning DOAS, on a fix-position at 2 km northeast of the Bromo cone. The spectrometer used was an Ocean Optic USB2000 + with a spectral range of 280-440 nm and 0.5 nm FWHM (Bani et al. 2013. The scanning step angle was 1°. The SO 2 column amounts were retrieved using DOAS standard analysis procedures (Platt and Stutz 2008). The reference spectra included in the nonlinear fit were obtained by convolving high-resolution SO 2 (Bogumil et al. 2003) and O 3 (Voigt et al. 2001) cross-sections with the instrument line shape. A Fraunhofer reference spectrum and a ring spectrum, calculated in DOASIS (Krauss 2005), were also included in the fit. The total column amount is then multiplied with the mean plume speed (estimated at 1 m/s using a thermal camera) to obtain the SO 2 flux. To assess the fluctuation of the SO 2 emission from Bromo over a longer time series (Fig. 2), the SO 2 mass captured by the ozone monitoring instrument (OMI), available online (https:// so2. gsfc. nasa. gov/ index. html) were compiled and analyzed following the approach in Bani et al (2016) (see Additional file 1: 1).
Bromo gas composition was first obtained in 2014 (Aiuppa et al. 2015) using the multicomponent gas analyzer system (Multi-GAS) (Aiuppa et al. 2005;Shinohara 2005). The same instrument was again deployed in 2018 (this work) and positioned downwind side of the Bromo crater rim. The unit simultaneously measured the concentrations of H 2 O, CO 2 , SO 2 , H 2 S, and H 2 at 0.5 Hz. Water (H 2 O) and CO 2 were detected by nondispersive infrared spectroscopy (LI-COR LI-840A; 0-60,000 ppm range). While SO 2 , H 2 S, and H 2 were measured using electrochemical sensors, respective models 3ST/F, EZ3H, and EZT3HYT Easy Cal from City Technology with the measurement range of 0-200 ppm. The data were processed and analyzed for the molar ratios between different volatile species using the Ratiocalc software (Tamburello 2015).
To gain further insights into the magmatic source behind the Bromo activity, ash and scoria deposits from 2000-2001, 2010-2011, 2015-2016, and 2019 eruptions were sampled then analyzed for major elements in the laboratory using X-ray fluorescence (XRF). Results were then compared to available data in the literature.

Persistent strong passive degassing
Ground-based SO 2 flux measurement results on Bromo since the first recording in 1995 are summarized in Additional file 1: Table S2, highlighting a persistent passive degassing into the atmosphere. The first mean value of 14 t day −1 obtained from COSPEC recordings is one to two orders of magnitude lower than the recent estimates but comparable to the incomplete DOAS traverse results from Bani et al. (2013). Using this unique SO 2 flux value available in the mid-1990s, Andres and Kasgnoc (1998) have classified Bromo (Tengger Caldera) in the 46th position out of 49 main volcanic gas contributors into the atmosphere. If we assume that such a relatively small gas emission can be representative of Bromo activity in the 1990s, then the corresponding magma source must be less-rich in gas content. However, if the COSPEC and DOAS traverse results are excluded, then the daily emission rate outside the eruptive periods from Bromo fluctuates between 82 and 280 t with a mean value of 181 ± 60 t day −1 . This degassing value is consistent over the last 15 years of ground-based measurements (Additional file 1: Table S2, Fig. 3), which thus confirms Bromo as one of the strong and persistent degassing sources in Indonesia. When integrating our result into the recent global volcanic SO 2 emission budget from Carn et al. (2017), Bromo would rank 68th out of the 91 listed volcanoes, comparable to the degassing of Stromboli volcano (181 t day −1 ). Assuming that our result is representative, then the mean annual SO 2 contribution into the atmosphere from Bromo is ~ 0.07 Tg, representing 2.3% of the annual contribution from Indonesian volcanoes ) and 0.3% of the annual global volcanic SO 2 emission budget (Carn et al. 2017).

Eruptive gas contribution
In the recent inventory of volcanic gas emissions, based on OMI satellite data, Carn et al. (2017) rank the combined SO 2 emission budget from Bromo and Semeru volcanoes on the 20th position out of 91 identified volcanic degassing sources worldwide, with a daily mean value of 775 t, much higher than the mean passive degassing of 181 t obtained in this work. It is also higher than the ground-based SO 2 fluxes obtained during the winding down phases of the 2015-2016 eruption (Additional file 1: Table S2, Fig. 3) that fluctuated between 130 and 440 t day −1 . When comparing the SO 2 mass obtained from OMI and the ground-based DOAS measurements (Fig. 3) outside the eruptive periods, the OMI values are systematically lower than the DOAS results. Moreover, Smekens et al. (2015) estimated a daily SO 2 release from Semeru to be 21 − 71 t using a UV-Camera. If the ground-based estimates of Bromo were combined with Semeru results, assuming that these figures are representative, then the total contribution outside the eruptive period from these two volcanoes would be about 200 − 250 t day −1 , with Bromo being the main degassing source, representing > 70% of the total SO 2 output. One alternative explanation of this difference would be an overestimation of SO 2 mass discharges from the OMI data, due to the eruptive discharges that can last from a few to more than ten months at Bromo.
Between 2009 and 2019, Bromo experienced three distinct eruptive events: in 2010 − 2011 (8 months), in 2015 − 2016 (11 months), and 2019 (3 months) with the corresponding SO 2 mass discharges of 68 kt, 53 kt, and 2 kt, respectively (Fig. 3). Thus a total of 123 kt of SO 2 mass was released from Bromo during the eruptive events over the last decade, a figure that corresponds to 19% of the total SO 2 contribution into the atmosphere. In comparison, more than 532 kt of SO 2 were released through passive degassing over the same period, representing 81% of the total SO 2 emission budget. The main SO 2 contribution to the atmosphere from Bromo is thus through passive degassing, as already seen on other volcanoes (e.g., Andres and Kasgnoc 1998;Bani et al. 2012).

Gas composition and emission budget
Multi-GAS measurement results obtained in October 2018 are comparable to Aiuppa et al. (2015) findings (Table 1; Fig. 4) with H 2 S/SO 2 , CO 2 /SO 2 , H 2 /SO 2 , and H 2 O/SO 2 ratios, obtained from the best-fit regression lines of 0.2 − 0.5, 3.6 − 5.0, 0.1 − 0.4, and 61 − 142 over 3 days of field measurements. Note that on the 3rd day  and nickel-nickel oxide buffers (NNO) (oxygen fugacity of 10 -16 bars). The mean CO 2 /S T ratio is 3.0 ± 1.6, S T being the combined H 2 S and SO 2 molar values. This value is above the mean CO 2 /S T ratio (~ 2) for persistently degassing at open-vent arc volcanoes (Shinohara 2013), but is well within the range of volcanic gas CO 2 /S T ratios of 3 − 6 of Java volcanoes (Aiuppa et al. 2015). This C-rich gas composition implies some substantial carbon contribution from either the slab or the crustal-derived fluids  into the source mechanism, however further work is required to better constrain this C-enrichment. By converting the measured volatile ratios into molar percentages (in the assumption that no other major volcanogenic gas is present in addition to those determined), it is found that Bromo releases a water-rich gas with 94.4% H 2 O molar proportion, typical for arc volcanic gases (Fischer 2008). The inferred concentrations for the other gases are 4.0%, 0.9%, 0.4%, and 0.2% for CO 2 , SO 2 , H 2 S, and H 2 , respectively (Table 1). When combining the gas ratios and the mean SO 2 flux of 181 ± 60 t day −1 , the daily passive emission for each gas component would be 5141 ± 658 t, 532 ± 70 t , 37 ± 12 t, and 1.0 ± 0.2 t for H 2 O, CO 2 , H 2 S, and H 2 , respectively. If these results are extrapolated over the last 10 years, then the total magmatic gas contribution into the atmosphere from Bromo over the last decade would be 18.

Degassing source
The scoria and ash samples collected during four past eruptive events (2000-2001, 2010-2011, 2015-2016, and 2019 eruptions) were analyzed for their major elements. Results (Additional file 1: Table S3) indicate a basaltic-andesite to basalt trachy-andesite melt source beneath the Bromo volcano, with a transition from a medium-K to a high-K composition (Fig. 5) (Fig. 5), and they emphasize a tholeiitic affinity while the material from the other eruptions plot along the transition line between the tholeiitic and calc-alkaline series (Fig. 5). Such an evolution of rock composition was reported in the past works (Whitford et al. 1979;Van Gerven and Pichler 1995;Santoso et al. 2017) and the medium-K rocks were attributed to pre-caldera activities while in contrast, the high-K affinity samples were considered to be induced by the younger and post-caldera magma. Van Gerven and Picher (1995), who also observed the co-occurrence of these two magma affinities on Bromo, indicate that the calc-alkaline rocks would not be expected as a function of the depth of the Waditi-Benioff zone.
Over the last 20 years, the SiO 2 obtained from the eruptive products is virtually unchanged, between 53 and 57 wt%. Such character is generally encountered on volcanoes with a tholeiitic affinity (e.g., Rogers 2015; Chin et al. 2018). When referring to MgO as a fractional crystallization index, knowing that its content in the melt decreases with the cooling and crystal removals from the gradually solidifying melt (e.g., Rogers 2015), it appears that the 2000-2001 eruption was likely sustained by an evolved magmatic source, given its low MgO content (1.4-2.0 wt%) and a relatively high P 2 O 5 (0.12-0.49 wt%), SiO 2 (53.8-57.21 wt%), and CaO (8.5-9.9 wt%). In an open-vent system, such a scenario is symptomatic of a magma that has released much of its gas. The low SO 2 flux (6-23 t/day) recorded in 1995 (Additional file 1: Table S2) was likely sustained by this same degassed melt source. In contrast, the 2010-2011 eruption displays a higher MgO content (1.8-3.2 wt%) but relatively low CaO (6.2-7.2 wt%), and SiO 2 (53.2-56.6 wt%) contents indicating a less evolved magmatic source. Such a change in magma composition suggests a possible new magma supply into the reservoir. The wide ranges of CaO, Al 2 O 3 , TiO 2 , SiO 2 , Na 2 O, P 2 O 5 , and even K 2 O, indicate that this new magma supply was injected into a relatively evolved melt source (e.g., Cassidy et al. 2018). The high amount of SO 2 discharge (68 kt) (Fig. 3) (Irvine and Baragar 1971) highlights the transition between the tholeiitic and calc-alkaline sources. Data from Van Gervan and Pitchler (1995) and Carn and Pyle (2001) are plotted as references. Note that there are no data from the 2004 eruption compared to 2001 is another evidence of a new magma injection but with the assumption that the magma genesis involves a low degree of partial melting (Schiano et al. 1998), knowing that K can be easily mobilized by incoming fluid. In such a scenario, a variable degree of partial melting may be responsible for the transition from medium-to high-k magma source (Whitford et al. 1979) on Bromo. The fluid responsible for the partial melting is likely induced by the dehydration of the subduction slab with a contribution of subducted sediments that are possibly enriched in carbon, given the relatively high CO 2 /S T ratio. A sediment contribution to the magma genesis is described by Gertisser and Keller (2003) to be responsible for the compositional contrast between medium-K and high-k magma on Merapi volcano. The sediment contribution and the variability of partial meting may contribute to the changes of magma composition on Bromo volcano as well, but further isotopic analyses would be needed to corroborate such interpretation.
The SO 2 flux obtained between 2005 and 2009 (Additional file 1: Table S2) fluctuates between 112 and 281 t d −1 , which is an order of magnitude higher than the 1995 values. Furthermore, according to the history of Bromo eruptive activities (GVP 2013), there was a strong eruptive event in June 2004 that built a large and dense column of ash up to 3000 m above the crater GVP 2004). Similarly to 2010, such an intense manifestation requires a new magma supply into the reservoir. Unfortunately the tephra products from this event are not available for this study. However given the low SO 2 flux recorded before the eruption (14 t day −1 in 1995), the high intensity of the 2004 eruption followed by a persistent high SO 2 flux, and also the evolved and degassed source observed during 2000-2001 eruption, it is likely that Bromo volcano has also witnessed a new magma injection in 2004. Both these two new magma supplies (in 2004 and 2010-2011) are consequently responsible for the persistent strong degassing that we observed over the last decade. The compositions of the ejected products from the 2015-2016 and 2019 eruptions fall in the range of 2010-2011 products suggesting a unique magmatic source, in agreement with the unchanged gas composition obtained in 2014 (Aiuppa et al. 2015) and 2018 (Table 1). The relative increase in concentrations of oxides, between the 2015-2016 and the 2019 eruptions, highlighted by K 2 O (from 2.6-3.2 to 3.0-3.5 wt%), P 2 O 5 (from 0.31-0.35 to 0.37-0.47 wt%), SiO 2 (from 55.3-55.9 to 54.6-56.9 wt%) or TiO 2 (1.01-1.09 to 1.09-1.13 wt%) indicate a relative fractional crystallization. The decrease of eruptive SO 2 mass discharge, from 53 kt during the 2015-2016 eruption, to 2 kt during the 2019 event, also reflect the evolution within the same melt source (Additional file 1: Table S2, Fig. 3).
Considering the hyperactivity of this volcano (50 eruptions in 2 centuries), frequent pressurization of the shallow reservoir is required. Although it is obvious that a new magma injection plays a major role in Bromo hyperactivity, it is unlikely that such magma supply constitutes the unique triggering factor of all the eruptions. We speculate that the progressive replenishment of a shallow reservoir, in response to the persistently high degassing following the model of Girona et al. (2015) would contribute to the hyperactivity of Bromo while maintaining the continuous strong degassing. In such a scenario the pressure in the shallow reservoir decreases because of degassing. But subsequently, it induces a pressure difference with a deeper reservoir. Thus the depressurization is simultaneously compensated by the pressurization induced by the replenishment process of the yield-stress magma. In such a case, the dike between the two reservoirs is regularly cleared, allowing the deep and shallow reservoirs to be connected leading to a regular pressure buildup in the shallow reservoir and subsequent eruptive events.

Conclusion
Ground-based gas measurements combined with daily satellite recording of SO 2 mass above Bromo Volcano indicate that over the last decade about 18.8 Tg of H 2 O, 2.0 Tg of CO 2 , 0.7 Tg of SO 2 , 0.1 Tg of H 2 S, and 0.005 Tg of H 2 were released into the atmosphere. Bromo is thus one of the notable magmatic volatile contributors to the atmosphere, representing 2.3% and 3% of the annual SO 2 and CO 2 emission budgets from Indonesian volcanoes and about 0.3% and 0.4% of the annual global volcanic SO 2 and CO 2 degassing budget outside eruptive period.
Despite the recurrent and long eruptive manifestations, the main volatile contribution into the atmosphere from Bromo is through passive degassing, representing about 81% of the total gas released. The magmatic source that feeds Bromo activity is of basaltic-andesite to basalt trachy-andesite composition with a transition from medium-K to high-K over the last two decades due to the low degree of partial melting and the C-rich sediment contributions to the magma genesis. The moderate fractional crystallization processes and the new magma injections also contributed to the evolution of source magma. Finally, the hyperactivity of the Bromo volcano and its persistent strong degassing is likely sustained by the new magma injections and a progressive replenishment of the shallow reservoir in response to depressurization and pressure buildup initiated by the continuous