Reactive bromine plays a crucial role in the Arctic troposphere (Barrie et al., 1988). In polar spring, the release of large amounts of reactive bromine species (Br and BrO) during bromine explosion events (BEEs) (Wennberg, 1999; Abbatt et al., 2012; Simpson et al., 2007a) can trigger ozone depletion events (ODEs), in which surface ozone volume mixing ratios (VMRs) fall below 10 ppbv and sometimes approach near-zero values (Oltmans and Komhyr, 1986; Bottenheim et al., 1986; Halfacre et al., 2014). Reactive bromine also influences key photochemical species in the boundary layer, including OH/HO2, NO/NO2, and volatile organic compounds (Barrie and Platt, 1997; von Glasow and Crutzen, 2007; Simpson et al., 2015). Furthermore, enhanced reactive bromine promotes the oxidation of gaseous elemental mercury (GEM) and increases the deposition of particulate mercury, potentially impacting Arctic wildlife and human health through bioaccumulation in the food chain (Schroeder et al., 1998; Steffen et al., 2008).
In ozone depletion chemistry, the three main reaction cycles involving reactive bromine – self-reaction (BrO + BrO), cross-reaction (BrO + IO/ClO), and the catalytic cycle (BrO + HO2) – do not modify the total abundance of the Brx family (Simpson et al., 2007a). BEEs are widely believed to occur through heterogeneous reactions on low-pH saline surfaces containing sufficient bromide (Br−), which release reactive bromine into the gas phase via an autocatalytic cycle (Fan and Jacob, 1992; Wennberg, 1999; Abbatt et al., 2012):
In the multiphase Reaction (R1), HOBr reacts with Br−, releasing a bromine molecule. Photolysis of Br2 (Reaction R2) produces two bromine atoms, leading to a rapid buildup of reactive bromine (Lehrer et al., 2004). In addition to this pathway, other bromine release mechanisms have been reported. For example, at night, the reaction between O3 and bromide ions on saline surfaces can generate gaseous bromine molecules (Oum et al., 1998). Laboratory studies have shown that bromide ions, H3O+, and O2 can produce bromine radicals through heterogeneous reactions under nighttime conditions (Cao et al., 2024). Furthermore, the release of reactive bromine could also be influenced by sea salt Cl− (Vogt et al., 1996) and OH radicals (Halfacre et al., 2019).
Potential sources of reactive bromine include: open-ocean sea spray (Sander et al., 2003); snow over first-year sea ice (FYI) (Jones et al., 2006; Simpson et al., 2007b) and over multi-year sea ice (MYI) (Peterson et al., 2019; Huang et al., 2020); frost flowers (Kaleschke et al., 2004; Nghiem et al., 2012); sea salt aerosol (SSA) from polynyas or open leads (Kirpes et al., 2019; Criscitiello et al., 2021); tundra snowpack photochemistry (Pratt et al., 2013); blowing-snow-sourced SSA (Yang et al., 2008, 2010; Jones et al., 2009; Choi et al., 2018; Huang et al., 2020); and stratosphere-to-troposphere transport of BrO (Salawitch et al., 2010). Field studies in Alaska and laboratory experiments have shown that acidic saline snowpacks can release reactive bromine when exposed to sunlight and ozone (Wren et al., 2013; Pratt et al., 2013). Reactive bromine fluxes from snowpack in Alaska, ranging from 7 × 107 to 1.2 × 109 molec. cm−2 s−1, based on direct measurements of BrCl and Br2, were reported. These are in line with those needed in models to reproduce BEEs and ODEs (Custard et al., 2017). In contrast, Yang et al. (2024) derived a much lower average snowpack release flux of 1 × 107 molec. cm−2 s−1 or below from field measurements in Eureka, Canada (86.4° W, 80.1° N), suggesting that coastal snowpack is a weak reactive bromine source. Moreover, based on a mass balance approach, Yang et al. (2024) estimated that the lifetime of atmospheric reactive bromine as a family is 17-42 d, which is longer than the 4-10 d reported in previous studies (von Glasow et al., 2004; Yang et al., 2005). Fresh frost flowers, although highly alkaline and saline, likely make only minor or local contributions due to their limited spatial extent on open leads (Obbard et al., 2009; Lieb-Lappen and Obbard, 2015). Stratospheric influence on tropospheric BrO columns also appears limited (Theys et al., 2011).
Among these potential bromine sources, FYI and MYI differ in their physical and chemical properties, which may influence the efficiency of bromine activation. MYI is generally thicker and less porous than FYI, and these physical characteristics limit brine connectivity within the ice and its upward transport, resulting in distinct roles in Arctic bromine chemistry (Haas et al., 2006, 2010). After undergoing multiple summer melt-refreeze cycles, MYI is desalinated through gravity drainage and flushing, whereas FYI typically retains higher salinity (Krnavek et al., 2012). Although brine can migrate upward through the ice-snow interface and supply salts to overlying snow, this process is highly sensitive to snow depth, with observations suggesting an effective upper limit of approximately 17 cm (Domine et al., 2004). MYI regions generally accumulate deeper snowpacks, limiting the upward transport of salts from the ice and resulting in lower snow salinity compared to FYI regions (Webster et al., 2014; Blanchard-Wrigglesworth et al., 2015). Under such low-salinity conditions, atmospheric deposition and recycling may become increasingly important in controlling bromide availability in the snow (Krnavek et al., 2012; Nandan et al., 2017). Surface snow over MYI regions typically exhibits low salinity, with a typical median value around 0.01 practical salinity unit (psu), whereas snow over FYI regions is generally more saline, with median values ranging from 0.1 to 0.7 psu (Krnavek et al., 2012). Analysis of bromine enrichment factors indicates that bromide depletion is more frequently observed in snow over MYI compared to snow over FYI, suggesting that, in addition to FYI regions, snow over MYI regions may actively participate in Arctic boundary layer bromine chemistry (Peterson et al., 2019). This explains the enhanced tropospheric BrO observed over MYI regions (Peterson et al., 2016; Burd et al., 2017). However, it is commonly assumed that snow over FYI regions represents the primary source of reactive bromine, while the contribution from snow over MYI regions is relatively limited. Most current chemistry models do not account for the effects of MYI.
Investigating reactive bromine release mechanisms in the polar troposphere requires robust observational evidence, which has been partially provided by satellite measurements offering key insights into BrO distribution. Wagner and Platt (1998) and Richter et al. (1998) reported elevated BrO in the polar spring from GOME measurements, with BrO-rich air masses generally located near sea ice regions. Jacobi et al. (2006) reported enhanced tropospheric BrO detected by the SCIAMACHY satellite during an ODE near the Arctic sea ice edge. Jones et al. (2009) observed elevated BrO with SCIAMACHY at Halley Station, Antarctica, and found that both stable boundary layers under low wind speeds and turbulent boundary layers under high wind speeds could lead to ODEs. Subsequent studies using GOME-2 satellite observations linked Arctic BrO enhancement to high-wind cyclonic systems, blowing-snow-generated SSA, and FYI contact (Begoin et al., 2010; Theys et al., 2011; Bougoudis et al., 2020). Seo et al. (2019) detected BrO enhancements with TROPOMI in diverse regions, including Arctic sea ice, volcanic areas, and salt lakes. However, limited spatial resolution and vertical sensitivity hinder a full understanding of reactive bromine activation mechanisms (Sihler et al., 2012).
Ground-based Differential Optical Absorption Spectroscopy (DOAS) observations provide minute-level temporal resolution and good vertical coverage, offering significant advantages for studying boundary layer BEEs (Hausmann and Platt, 1994; Hönninger et al., 2004; Frieß et al., 2004, 2011; Liao et al., 2011; Stutz et al., 2011; Peterson et al., 2015; Zhao et al., 2016; Simpson et al., 2017). Enhanced BrO has been detected during polar spring using long-path DOAS and MAX-DOAS instruments (Hausmann and Platt, 1994; Hönninger et al., 2004). MAX-DOAS observations also facilitate investigations of BrO vertical distribution within the boundary layer (Peterson et al., 2015; Simpson et al., 2017; Frieß et al., 2023; Brockway et al., 2024). Additionally, BrO has been measured using airborne DOAS instruments and chemical ionization mass spectrometers, although such measurements are generally limited to specific campaigns (Liao et al., 2011, 2012; General et al., 2014; Peterson et al., 2017). In Ny-Ålesund, several research groups have conducted BrO observations over the years. Tuckermann et al. (1997) detected BrO VMR exceeding 30 pptv using a DOAS instrument, which was associated with severe ODEs. Langendörfer et al. (1999) reported that filterable bromine concentrations measured by ion chromatography were consistent with BrO variations observed by DOAS. Luo et al. (2018) reported that enhanced BrO detected by MAX-DOAS in April 2015 was linked to sea ice cover in the Kings Bay region. Chen et al. (2022) documented a BEE event in March 2017, observed by MAX-DOAS, and attributed it primarily to SSA generated by blowing snow over sea ice under high wind conditions. Zilker et al. (2023) analyzed surface ozone and satellite-retrieved tropospheric BrO column data during spring seasons from 2010 to 2021, showing that enhanced BrO frequently occurred near Svalbard during ODEs, with the strongest enhancements typically observed in March.
Based on these observations, various reactive bromine emission mechanisms have been incorporated into chemical models. For example, Toyota et al. (2011, 2014) implemented a chemical scheme representing reactive bromine release from snowpack on sea ice, successfully reproducing enhanced BrO observed by the GOME satellite during Arctic spring. The same mechanism has also been applied to simulate BrO in both the Arctic and Antarctic (Falk and Sinnhuber, 2018). Yang et al. (2008, 2010) developed an alternative scheme in which reactive bromine is released from SSA generated by blowing snow, rather than directly from snowpack emissions. This mechanism is supported by recent field observations confirming SSA production from blowing snow (Yang et al., 2019; Frey et al., 2020; Gong et al., 2023). Models using this scheme have successfully reproduced BEEs and ODEs in both hemispheres. For instance, Choi et al. (2018) reported that the spatial distribution of model-simulated SSA from blowing snow closely matched tropospheric BrO columns observed by the OMI satellite during Arctic spring. Marelle et al. (2021) modelled Arctic ODEs in spring 2012 by applying both the snowpack and blowing-snow SSA schemes, and their results indicated that each contributed to ODE occurrence. Although these models can reproduce BEEs and ODEs, challenges remain due to low spatial resolution, limited ability to simulate boundary layer dynamics, and a lack of field data to constrain key parameters such as snow salinity and snow age, making quantitative comparison with in situ observations and mechanism validation difficult.
In this study, we retrieved 0-4 km BrO partial columns from observations made using a MAX-DOAS deployed at the Yellow River Station in Ny-Ålesund during 2017-2023. These data were combined with overpassing GOME-2B satellite measurements, p-TOMCAT model simulations, near-surface ozone observations, local meteorological records, and backward trajectories to study the possible sources and causes of variability in BrO levels in Ny-Ålesund. We further quantified the relative contributions of open ocean and sea ice to reactive bromine. Section 2 describes the methods and data used in this analysis. The results and discussions are presented in Sect. 3, and the conclusions are summarized in Sect. 4.
