Rapid chemical evolution of tropospheric volcanic emissions from Redoubt Volcano, Alaska, based on observations of ozone and halogen-containing gases

Abstract

We report results from an observational and modeling study of reactive chemistry in the tropospheric plume emitted by Redoubt Volcano, Alaska. Our measurements include the first observations of Br and I degassing from an Alaskan volcano, the first study of O₃ evolution in a volcanic plume, as well as the first detection of BrO in the plume of a passively degassing Alaskan volcano. This study also represents the first detailed spatially-resolved comparison of measured and modeled O₃ depletion in a volcanic plume. The composition of the plume was measured on June 20, 2010 using base-treated filter packs (for F, Cl, Br, I, and S) at the crater rim and by an instrumented fixed-wing aircraft on June 21 and August 19, 2010. The aircraft was used to track the chemical evolution of the plume up to ~ 30 km downwind (2 h plume travel time) from the volcano and was equipped to make in situ observations of O₃, water vapor, CO₂, SO₂, and H₂S during both flights plus remote spectroscopic observations of SO₂ and BrO on the August 19th flight. The airborne data from June 21 reveal rapid chemical O₃ destruction in the plume as well as the strong influence chemical heterogeneity in background air had on plume composition. Spectroscopic retrievals from airborne traverses made under the plume on August 19 show that BrO was present ~ 6 km downwind (20 min plume travel time) and in situ measurements revealed several ppbv of O₃ loss near the center of the plume at a similar location downwind. Simulations with the PlumeChem model reproduce the timing and magnitude of the observed O₃ deficits and suggest that autocatalytic release of reactive bromine and in-plume formation of BrO were primarily responsible for the observed O₃ destruction in the plume. The measurements are therefore in general agreement with recent model studies of reactive halogen formation in volcanic plumes, but also show that field studies must pay close attention to variations in the composition of ambient air entrained into volcanic plumes in order to unambiguously attribute observed O₃ anomalies to specific chemical or dynamic processes. Our results suggest that volcanic eruptions in Alaska are sources of reactive halogen species to the subarctic troposphere.

Introduction

Understanding the chemistry of tropospheric volcanic plumes is important for evaluating eruption hazards and to assess the atmospheric and environmental impacts of volcanic emissions. Recent studies have shown that tropospheric volcanic plumes are chemically active and that relatively non-reactive hydrogen halide gases (e.g. hydrogen bromide, HBr) emitted from volcanoes can be rapidly converted into reactive halogen species (e.g. bromine monoxide, BrO) by in-plume heterogeneous chemical processes (Bobrowski et al., 2003, Gerlach, 2004, Oppenheimer et al., 2006, Bobrowski and Platt, 2007, Bobrowski et al., 2007, Bani et al., 2009, Kern et al., 2009). The formation of BrO in volcanic plumes is notable because it is linked to families of chemical reactions known to profoundly impact tropospheric chemistry (von Glasow et al., 2009, von Glasow, 2010). Reactive halogen species (X, XO, X₂, XY, OXO, HOX, XONO₂, XNO₂, where X, Y = Cl, Br, or I), especially those that contain bromine, can rapidly destroy ozone (O₃), alter common atmospheric oxidation pathways, and increase the deposition of toxic metals like mercury to the surface (e.g. Simpson et al., 2007, von Glasow, 2010).

The main chemical reaction sequence thought to be responsible for activation of Br and formation of BrO in volcanic plumes is conceptually similar to the autocatalytic “bromine explosion” mechanism that leads to polar tropospheric O₃ depletion events (see Roberts et al., 2009 and von Glasow et al., 2009 for more thorough descriptions of reactive halogen chemistry in volcanic plumes and Simpson et al., 2007 for a review of polar halogen chemistry), as summarized below:

$${\text{HOBr}}_{\left(\text{gas}\right)}\to {\text{HOBr}}_{\left(\text{aq}\right)}$$

$${\text{HBr}}_{\left(\text{gas}\right)}\to {\text{Br}}^{-}{}_{\left(\text{aq}\right)}+{\text{H}}^{+}{}_{\left(\text{aq}\right)}$$

$${\text{HOBr}}_{\left(\text{aq}\right)}+{\text{Br}}^{-}{}_{\left(\text{aq}\right)}+{\text{H}}^{+}{}_{\left(\text{aq}\right)}\to {\text{Br}}_{2\left(\text{gas}\right)}+{\text{H}}_2{\text{O}}_{\left(\text{aq}\right)}$$

$${\text{Br}}_2+\text{h}\nu \to 2\text{Br}$$

$$2\left(\text{Br}+{\text{O}}_3\to \text{BrO}+{\text{O}}_2\right)$$

$$\text{BrO}+\text{BrO}\to 2\text{Br}+{\text{O}}_2$$

$$\text{BrO}+\text{BrO}\to {\text{Br}}_2+{\text{O}}_2$$

$$\text{BrO}+{\text{HO}}_2\to \text{HOBr}+{\text{O}}_2$$

According to this mechanism, gaseous HOBr and HBr are taken up onto acidic particles (pH < 6; R1 and R2) and an acid-catalyzed aqueous-phase reaction results in Br₂ that is released to the gas phase (R3). Rapid photolysis of Br₂ (R4) and subsequent reaction with O₃ yields BrO (R5). Further O₃ destruction can be sustained by the BrO self-reaction (R6a, R6b), and reaction of BrO with HO₂ (R7) allows the cycle to repeat. Each autocatalytic cycle doubles the BrO concentration which can lead to rapid, non-linear increases of BrO in volcanic plumes and significant O₃ destruction.

The idea that in-plume BrO formation and O₃ destruction could be important aspects of volcanic plume chemistry was first proposed by Bobrowski et al. (2003), evaluated from a high-temperature thermodynamic perspective by Gerlach (2004), and assessed from a kinetic standpoint by Oppenheimer et al. (2006). The presence of BrO has been detected by ultraviolet differential optical absorption spectroscopy (UV-DOAS) in the plumes of at least 10 volcanoes worldwide (see Boichu et al. (2011) and references therein for data compiled from 9 volcanoes; more recently Heue et al. (2011) reported detection of BrO in the plume from Eyjafjallajökull volcano). In addition, Theys et al. (2009) reported satellite identification of BrO in the upper tropospheric/lower stratospheric eruptive plume from the 2009 eruption of Kasatochi volcano, Alaska. Numerical models have also predicted extensive O₃ destruction in volcanic plumes due mainly to in-plume BrO formation (Bobrowski et al., 2007, Roberts et al., 2009, von Glasow, 2010).

Despite these recent advances, observations of O₃ in volcanic plumes have been relatively sparse and empirical correlations of BrO formation and O₃ destruction have been challenging to obtain. A few older studies each reported O₃ depletion in volcanic plumes (Stith et al., 1978, Fruchter et al., 1980, Hobbs et al., 1982) but did not offer much detail or discussion to accompany their observations, which limited their utility. Several recent studies have reported observations of low O₃ in tropospheric volcanic plumes (Oppenheimer et al., 2010, Vance et al., 2010, Schumann et al., 2011) and have speculated that their observations were a consequence of in-plume bromine chemistry, although supporting measurements of BrO or other reactive halogens were not included. Boichu et al. (2011) presented a study where BrO was identified in the young Mt. Erebus plume (aged 3–7 min) in spectra collected on Dec. 3, 2005, and based on this finding and model results, showed that bromine chemistry could have been responsible for the low in-plume O₃ Oppenheimer et al. (2010) observed 26–39 km downwind (4–6 h plume travel time) during their research flight a week later, on Dec. 10, 2005. However, in another study, Heue et al. (2011) identified BrO in a 1–2 day old plume from Eyjafjallajökull but found no measureable O₃ depletion. Further complexity in establishing the links between O₃ and bromine chemistry in volcanic plumes was discussed by Carn et al. (2011), who reported O₃ deficits of 20–30% in volcanic plumes from Ecuador and Colombia aged from ~ 22–48 h and ~ 2 h, respectively, but concluded that insufficient knowledge of ambient O₃ during the plumes' evolution prevented them from unambiguously attributing the observed anomalies to reactive halogen chemistry.

To further examine the fast halogen-mediated chemical reactions occurring in volcanic plumes, we conducted an airborne and ground-based investigation of the tropospheric plume of Redoubt Volcano, Alaska. The data reported here relate to rapid chemical processes occurring in the plume after volcanic volatiles were released to the atmosphere and aged up to ~ 2 h during the daytime. We report new observations of halogen-containing (F, Cl, Br, I) gases collected using filter packs from the rim of the volcano on June 20, 2010 that constrain the composition of the young plume (< 5 min old) as well as detailed airborne observations from flights on June 21st and August 19th, 2010 where the chemical evolution of the plume was tracked with an instrumented fixed-wing aircraft. We describe a method for mitigating SO₂ interference in ultraviolet (UV) absorption-based O₃ measurements and report in situ observations of O₃, water vapor, CO₂, SO₂, and H₂S from both flights plus airborne remote UV spectroscopic observations of SO₂ and BrO from the August 19th flight.

Using airborne measurements we first characterize the composition of ambient air entrained into the plume and estimate the degree to which entrainment of air from chemically distinct ambient air layers influences the plume's O₃ composition. Next, we quantify the extent and duration of chemical O₃ loss in the plume. Finally we compare our measurements with results from a state-of-the-art chemical model of volcanic plume chemistry (PlumeChem, Roberts et al., 2009) tuned to the environmental conditions at Redoubt on June 21 and initialized using the filter pack data from June 20. Our results highlight important aspects of in-plume chemical and dynamic processes and provide insights that will be useful for better understanding O₃ anomalies and reactive halogen chemistry in volcanic plumes.