Impact of biogenic very short-lived bromine on the Antarctic ozone hole during the 21 st century

Active bromine released from the photochemical decomposition of biogenic very short-lived bromocarbons (VSL Br ) enhances stratospheric ozone depletion. Based on a dual set of 1960-2100 coupled chemistry-climate simulations (i.e. with and without VSL Br ), we show that the maximum Antarctic ozone hole depletion increases by up to 14% when natural VSL Br are considered, in better agreement with ozone observations. The impact of the additional 5 pptv VSL Br on Antarctic ozone is most evident in the periphery of the ozone hole, producing an expansion of the ozone hole area of ~5 15 million km 2 , which is equivalent in magnitude to the recently estimated Antarctic ozone healing due to the implementation of the Montreal Protocol. We find that the inclusion of VSL Br in CAM-Chem does not introduce a significant delay of the modelled ozone return date to 1980 October levels, but instead affect the depth and duration of the simulated ozone hole. Our analysis further shows that total bromine-catalysed ozone destruction in the lower stratosphere surpasses that of chlorine by year 2070, and indicates that natural VSL Br chemistry would dominate Antarctic ozone seasonality before the end of the 20 21 st century. This work suggests a large influence of biogenic bromine on the future Antarctic ozone layer.


Introduction
The detection of the springtime Antarctic ozone hole (Farman et al., 1985) has been one of the great geophysical discoveries of the 20 th century. The unambiguous scientific reports describing the active role of halogen atoms (i.e. chlorine and bromine), released from anthropogenic chlorofluorocarbons (CFCs) and halons, in depleting stratospheric ozone (Molina and 25 Rowland, 1974;McElroy et al., 1986;Daniel et al., 1999) led to the rapid and efficient implementation of the Montreal protocol in 1989 (Solomon, 1999). Since then, the consequent turnover on the anthropogenic emissions of long-lived chlorine (LL Cl ) and bromine (LL Br ) sources (Carpenter et al., 2014) has controlled the evolution of the strong springtime ozone depletion within the Antarctic vortex, and the first signs of recovery of the ozone hole became evident at the beginning of the 21 st century (WMO, 2014;Chipperfield et al., 2015;Solomon et al., 2016). 30 Several coordinated initiatives have been conducted by the scientific community to predict the future evolution of the stratospheric ozone layer and its impact on climate change (Eyring et al., 2007(Eyring et al., , 2010bAustin et al., 2010;WMO, 2014). The multi-model CCMVal-2 ozone assessment (Eyring et al., 2010a) determined that even when Antarctic ozone return date to 1980 values is expected to occur around years 2045−2060, the impact of halogenated ozone depleting substances (ODS, such as LL Cl and LL Br ) on stratospheric ozone photochemistry will persist until the end of 21 st century. Many studies show 5 that dynamical and chemical processes affect the size, strength and depth of the ozone hole formation (see Solomon et al., (2015) and references therein). Ongoing research within the Chemistry-Climate Model Initiative (CCMI) (Eyring et al., 2013;Hegglin et al., 2014) includes model experiments that consider, along with the dominant LL Cl and LL Br anthropogenic emissions, an additional contribution from biogenic very short-lived bromocarbons (VSL Br ). This additional input of bromine is required to reconcile current stratospheric bromine trends (Salawitch et al., 2010;WMO, 2014). 10 VSL Br are naturally released from biologically productive waters mainly within the tropical oceans (Warwick et al., 2006;Butler et al., 2007;Kerkweg et al., 2008), where strong convective uplifts efficiently entrain near surface air into the upper troposphere and lower stratosphere (Aschmann and Sinnhuber, 2013;Liang et al., 2014;Saiz-lopez and Fernandez, 2016).
The current contribution of VSL Br to total stratospheric inorganic bromine is estimated to be in the range of 3−8 pptv (Montzka et al., 2011;Carpenter et al., 2014;Navarro et al., 2015;Hossaini et al., 2016). The most accepted value for 15 stratospheric injection is VSL Br ≈ 5 pptv, which currently represents approximately 30% of the total contribution from LL Br substances arising from both anthropogenic and natural origins (~7.8 pptv Halons + ~7.2 pptv CH 3 Br ≈ 15-16 pptv LL Br ).
The additional stratospheric contribution of biogenic VSL Br improves the model/observations agreement with respect to stratospheric ozone trends between 1980 and present time (Sinnhuber et al., 2009), with strongest ozone depleting impacts during periods of high aerosol loading within mid-latitudes (Feng et al., 2007;Sinnhuber and Meul, 2015). Although we still 20 lack a scientific consensus with respect to the future evolution of VSL Br ocean source strength and stratospheric injection (Carpenter et al., 2014), it will probably increase in the future following the increase on sea surface temperature (SST) and oceanic nutrient supply, as well as due to the enhancement of the troposphere-to-stratosphere exchange (Hossaini et al., 2012;Leedham et al., 2013).
Previous chemistry-climate modelling studies considering VSL Br chemistry have mainly focused on improving the 25 model vs. observed ozone trends at mid-latitudes with respect to equivalent setups considering only the dominant anthropogenic LL Cl and LL Br sources (Feng et al., 2007;Sinnhuber et al., 2009). However, those previous studies lack an indepth timeline analysis of the VSL Br impact on the ozone hole evolution during the current century. More recently, Oman et al., (2016) determined that the addition of 5 pptv VSL Br to the stratosphere could delay the ozone return date to 1980 levels by as much as one decade. Their result is in agreement with that of Yang et al., (2014), who performed present-day timeslice 30 simulations to address the sensitivity of stratospheric ozone to a speculative doubling of VSL Br sources under different LL Cl scenarios. Even when those works addressed the important question of the return date, conclusions were obtained considering a unique simulation member for each case and an approximate approach of VSL Br ocean emissions. Here, using Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys. Published: 5 October 2016 c Author(s) 2016. CC-BY 3.0 License.
the CAM-Chem model Fernandez et al., 2014;Tilmes et al., 2015Tilmes et al., , 2016, we present a coherent ensemble of coupled (with an interactive ocean) chemistry-climate simulations from 1960 to 2100 with and without the contribution of oceanic VSL Br sources. We focus on natural VSL Br -driven changes in the chemical composition and evolution of the Antarctic ozone hole during the 21 st century, particularly on their influence on the seasonality and enlargement of the ozone hole area, ozone hole depth and return date to 1980s levels. The analysis shown here describes the 5 ozone hole progress distinguishing the monthly seasonality from the long-term evolution. Additionally, we present a timeline assessment of individual contribution of anthropogenic and natural chlorine and bromine species to Antarctic ozone loss during the 21 st century, recognizing the independent impact arising from LL Br and VSL Br sources to the overall halogencatalysed O 3 destruction.

Methods 10
The 3-D chemistry climate model CAM-Chem (Community Atmospheric Model with Chemistry, version 4.0) , included into the CESM framework (Community Earth System Model, version 1.1.1) has been used for this study.
The model setup is identical to the CCMI-REFC2 experiment described in detail by (Tilmes et al., 2016), with the exception that the current setup includes a full halogen chemistry mechanism from the earth surface to the lower stratosphere : i.e., instead of considering a constant lower boundary condition of 1.2 pptv for bromoform (CHBr 3 ) 15 and dibromomethane (CH 2 Br 2 ) or increasing CH 3 Br by 5 pptv, our model setup includes geographically-distributed and time-dependent oceanic emissions of six bromocarbons (VSL Br = CHBr 3 , CH 2 Br 2 , CH 2 BrCl, CHBrCl 2 , CHBr 2 Cl and CH 2 IBr) . At the model surface boundary, zonally averaged distributions of long-lived halocarbons (LL Cl = CH 3 Cl, CH 3 CCl 3 , CCl 4 , CFC-11, CFC-12, CFC-113, HCFC-22, CFC-114, CFC-115, HCFC-141b, HCFC-142b and LL Br = CH 3 Br, H-1301, H-1211, H-1202and H-2402, as well as surface concentrations of CO 2 , CH 4 , H 2 , N 2 O are specified 20 (Meinshausen et al., 2011). CAM-Chem was configured with a horizontal resolution of 1.9º latitude by 2.5º longitude and 26 vertical levels, from the surface up to 40 km. To have a reasonable representation of the overall stratospheric circulation, the integrated momentum that would have been deposited above the model top is specified by an upper boundary condition . The model includes heterogeneous processes for active halogen species in polar stratospheric clouds from MOZART-3 (Kinnison et al., 2007;Wegner et al., 2013). A full description of the CAM-Chem VSL configuration, 25 detailing both natural and anthropogenic sources, heterogeneous recycling reactions, dry and wet deposition, convective uplift and large-scale transport has been given elsewhere Fernandez et al., 2014). This model configuration uses a fully-coupled Earth System Model approach, i.e. the ocean and sea-ice are explicitly computed. included, in addition to the run LL sources, the background biogenic contribution from VSL Br oceanic sources (run LL+VSL ).
Differences between these two types of experiments allow quantifying the overall impact of natural VSL Br sources on stratospheric ozone. Please note that whenever we refer to "natural" contribution, we are pointing out to the contribution of biogenic VSL Br under a background stratospheric environment due to the dominant anthropogenic LL Cl and LL Br sources (i.e., the natural fraction of long-lived chlorine and bromine are minor). 5 Unless stated otherwise, all figures were generated considering the ensemble average (sim ens ) of each independent experiment (run LL and run LL+VSL ), which in turn were computed considering the mean of the 3 independent simulations (sim 004 , sim 005 and sim 006 ). For the case of vertical profiles and latitudinal variations, the zonal mean of each ensemble was computed to the monthly output before processing the data, while a Hamming filter with an 11 years window was applied to all long-term time-series to smooth the data. Most of the figures and values within the text include geographically averaged 10 quantities within the Southern Polar Cap (SP), defined as the region poleward of 63º S. For the case of the ozone hole area, we use the definition from the NASA Goddard Space Flight Center (GSFC), defined as the region with ozone columns below 220 DU located south of 40º S. Model results have been compared to the National Institute for Water and Atmospheric research -Bodeker Scientific (NIWA-BS) total column ozone database, which combines measurements from a number of different satellite-based instruments (Bodeker et al., 2005). 15

Contribution of LL Br and VSL Br to stratospheric bromine
The 1960-2100 evolution of the stratospheric chlorine and bromine loading is shown in Fig. 1. The dominant anthropogenic LL Cl and LL Br scenarios included in our REFC2 simulations (Tilmes et al., 2016) show a pronounced peak at the end of the 20 th century and beginning of 21 st century, respectively, after which both their abundances decline. In comparison, the 20 evolution of VSL Br sources remains constant in time, with a present-day fixed contribution of ~5 pptv .
Note that stratospheric LL Cl returns to its past 1980 levels before 2060, while the 1980 loading of LL Br is not recovered even by the end of the 21 st century. Even when biogenic VSL Br sources remain constant, their relative contribution to the total bromine stratospheric loading changes with time: while for year 2000 VSL Br represents ~24% of total bromine, by the end of the 21 st century it reaches 40% of stratospheric bromine. These values are likely lower limits of the percentage contribution 25 of biogenic sources to stratospheric bromine, as predicted increases on SST and oceanic nutrient supply are expected to enhance the biological activity and VSL Br production within the tropical oceans (Hossaini et al., 2012;Leedham et al., 2013).
Furthermore, the increase in SST and atmospheric temperature projected for the 21 st century, is expected to produce a strengthening of the convective transport within the tropics (Hossaini et al., 2012;Braesicke et al., 2013;Leedham et al., 2013), which could additionally enhance the stratospheric injection of VSL Br . The partitioning between carbon-bonded 30 (source gas) and inorganic (product gas) bromine levels injected to the stratosphere are of great importance as they strongly Atmos. Chem. Phys. Discuss., doi: 10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys. Published: 5 October 2016 c Author(s) 2016. CC-BY 3.0 License.
affect the ozone levels mostly in the lowermost stratosphere (Salawitch et al., 2005;Fernandez et al., 2014), which has implications at the altitudes where the strongest O 3 -mediated radiative forcing changes due to greenhouse gases are expected to occur (Bekki et al., 2013).

Impact of VSL Br on the ozone hole evolution and its return date
The 1960-2100 evolution of the total ozone column within the southern polar cap (TOZ SP , between 63ºS−90ºS) during 5 October is illustrated in Fig. 2. Biogenic VSL Br introduce a continuous reduction in TOZ SP that exceeds the model ensemble variability between run LL and run LL+VSL experiments, and improves the overall model-satellite agreement (Fig. 2a). The induced by VSL Br is more noticeable between 1990 and 2010, i.e., when the stratospheric LL Cl loading also maximizes (see Fig. 1). This result is in agreement to Sinnhuber and Meul (2015), who reported a faster initial decrease and an overall better agreement between past mid-latitude O 3 trends and a model simulation forced with the additional contribution from VSL Br sources. Much smaller impacts are modelled on the 2 nd quarter of the century when LL Cl constantly decreases and other ODS 25 (such as CH 4 and N 2 O) increase. Fig.2b indicate that the expected TOZ SP return date to October 1980 is approximately the same for both experiments: individual computations of the return date considering each of the independent ensemble members, show that the expected return date shift due to VSL Br lies within model uncertainties (Table 1) Agreement between model and observations for TOZ SP and ∆TOZ SP 1980 improves for all seasons when VSL Br are considered (Fig. 3). The maximum ozone difference between run LL and run LL+VSL is smaller than 10 and 5 DU for summer and fall, respectively, highlighting the much larger ozone depleting efficiency of the additional bromine from VSL Br sources during spring, when halogen chemistry dominates Antarctic ozone depletion. In all cases, the ozone return dates to 1980 seasonal TOZ SP columns lay within the model uncertainties, with shorter return dates observed for the summer (~2045) and fall 15 (<2040). Note also that the predicted springtime ∆TOZ SP 1980 will not return to their 1960 values by the end of the 21 st century for neither run LL nor run LL+VSL simulations ( Fig. 2b and Fig. 3). However, during fall positive ∆TOZ SP 1980 values are reached already by 2060, highlighting the different future seasonal behaviour of the Antarctic stratosphere (see Sect. 3.3).

Influence on the ozone hole area
We now turn to the effect of biogenic bromine on the Antarctic ozone hole area (OHA). Figure 4 indicates that the inclusion 20 of VSL Br produces total ozone reductions larger than 10 DU from 1970 to 2070. This enhanced depletion extends well outside the limits of the southern polar cap (63ºS) and into the mid-latitudes (see grey line on Fig. 4). Most notably, the maximum ozone depletion driven by biogenic bromine is not located at the centre of the ozone hole but on the ozone hole periphery, close to the outer limit of the polar vortex (see polar views on Fig. 4). This result has implications for assessments of geographical regions exposed to UV-B radiation: natural VSL Br leads to a total column ozone reduction between 20 and 25 40 DU over wide regions of the Southern Ocean near the bottom corner of South America and New Zealand. OHA larger than 5 Million km 2 , with a consequent enhancement of ~8 Million Tons on the OMD. Thus, the biogenic bromine-driven OHA enlargement is of equivalent magnitude, but opposite sign, to the chemical healing shrinkage estimated due to the current phase out of LL Cl and LL Br emissions imposed by the Montreal Protocol (Solomon et al., 2016).
Unlike the 1980-baseline ozone return date definition (which is normalized to a preceding but independent ozone column for each ensemble), the OHA and OMD definitions are computed relative to a fixed value of 220 DU. Consequently, the 5 run LL+VSL experiment shows larger ozone hole areas (white line on Fig. 4) and ozone mass deficits, but does not represent any significant extension on the size of the ozone hole by the time when the 1980-return date occurs. This supports the fact that the 1980-return date is controlled by the future evolution of the dominant LL Cl and LL Br sources. Note, however, that significant ozone depletion as large as −20 DU, and for latitudes as low as 60ºS, is still observed during 2060, i.e., after the standard 1980-return date has been reached. This indicates that the contribution from VSL Br has significant implications on 10 the baseline polar stratospheric ozone chemistry besides the above-mentioned impacts on ozone hole size, depth and return date.

Vertical distribution of the ozone hole depth
Timeline analysis of the mean October ozone vertical profile within the southern polar cap [O 3 (z) SP ] is presented in Fig. 6.
The additional O 3 (z) SP depletion due to VSL Br sources is maximized precisely at the same altitudes where the minimum O 3 number densities are found: during the 2000 th decade O 3 (z) SP densities at 100 hPa for run LL+VSL and run LL are, respectively, <1.5 and <2.5 × 10 12 molecule cm −3 , which represents ~40% enhancement on the local ozone loss. This is in agreement to the recent findings reporting that near-zero ozone concentrations in the deep Antarctic lower stratospheric polar vortex are 20 only simulated when the VSL bromine sources are included (Oman et al., 2016). Interestingly, greater ozone loss is found in the periphery of the polar vortex, and below 25 hPa, due to the larger ozone number densities prevailing at those locations (see zonal panel on Fig. 6c). Above 25 hPa, O 3 (z) SP is not significantly modified, with an overall VSL Br impact on ozone abundances smaller than 5%. This can be explained by the varying importance of bromine and chlorine chemistry at different altitudes (see Sect. 3.4). Further analysis of Fig. 6d reveals that differences larger than 25% at ~100 hPa are only 25 found between 1990 and 2010, confirming that the strongest impact of VSL Br sources occurs coincidentally with maximum LL Cl loadings (Fig. 1).
During the simulation period (i.e., 1960-2100), O 3 (z) SP densities below 100 hPa are at least 10% lower for run LL+VSL than for run LL . By year 2050, when the 1980 October return date is approximately expected to occur, the uppermost portion of the O 3 layer (above 50 hPa) shows strong signals of recovery and drives the overall TOZ SP return date, but the O 3 abundance below 30 50 hPa is still depleted relative to their pre-ozone hole era, mostly at high latitudes (Fig. 6d). It is only after year 2080 that the O 3 (z) SP vertical profile is consistent with the pre-ozone hole period, although O 3 densities above 6 × 10 12 molec. cm −3 are Atmos. Chem. Phys. Discuss., doi: 10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys. Published: 5 October 2016 c Author(s) 2016. CC-BY 3.0 License. still not recovered even by the end of the century (Fig. 6a,b). Between 2080 and 2100, inclusion of VSL Br still represents a 10% additional O 3 reduction at 100 hPa, which can be explained considering a shift from the predominant ozone destruction from chlorine to a bromine-driven depletion (whose efficiency is increased by the additional VSL Br ).

Seasonal evolution of stratospheric Antarctic ozone
Figures 7 show how the seasonal cycle of TOZ SP has changed during the modelled period, expanding from the typical solar-5 driven natural annual cycle prevailing in 1960 (Fig. 7a) to the strongly perturbed anthropogenic-induced cycle consistent with the formation of the Antarctic ozone hole (Fig. 7c, year 2000). TOZ SP July normalizations on Figs. 7 and 8 have been computed respect to the TOZ SP value on July of each year, so the aperture, closure and depth of the ozone hole at each time is computed relative to the total ozone column prevailing during the preceding winter. Figure 8 shows the evolution of the annual cycle of TOZ SP as a function of simulated year for run LL+VSL and run LL . During the pre-ozone hole era, the typical 10 southern hemisphere natural seasonality is observed, with maximum October ozone columns for run LL that exceeds the values from run LL+VSL by ~5 DU. Starting on the seventies, the natural seasonal cycle is disrupted and the natural springtime maximum is replaced by a deep ozone reduction due to the ozone hole formation (Fig. 7b) Fig. 8 represent the temporal location of the monthly TOZ SP July minimum for each simulation (white for run LL+VSL and black for run LL ). Starting on ~1981 the position of the TOZ SP July annual minimum shifts from April (the radiatively driven fall minimum) to October (the springtime ozone hole minimum) for run LL+VSL (~1984 for run LL ). Accordingly, the returning of the TOZ SP annual minimum from October to April is delayed by ~4 years when VSL Br are considered (from 2047 for run LL to 2054 for run LL+VSL ). Table 2 shows the independent values for each of the independent 20 ensemble members. Only if the baseline seasonal cycle is superposed below the long-term evolution of the polar stratospheric ozone layer (instead of considering the fixed normalization to October 1980), the inclusion of biogenic VSL Br introduces an extension on the ozone return date of ~(6.3 ± 12.2) years. Even though this value agrees with the estimations from Yang et al. (2014), it most probably represents a mere coincidence, as their timeslice computations only considered the changes in the maximum ozone hole depletion under different VSL Br loadings, while our analysis highlights the seasonal 25 TOZ SP changes within a fully coupled climatic-simulation. Note, however, that in agreement to Table 1, the modelled delay on the return date computed considering the changes in the ozone seasonal cycle also lies within the internal model variability.
The dotted lines on Fig. 8 indicates the location of the double local TOZ SP July maximums observed in Fig. 7b,d-e and allows determining how the timespan between the ozone hole formation and breaking for each year changes due to VSL Br 30 chemistry. Between mid-1970s and mid-1980s, the seasonal development of the ozone hole for each year rapidly expanded shifting from a starting point as early as July through a closing date during the summer (December and January). Most notably, the seasonal ozone hole extension during the 1 st half of the century is enlarged as much as 1 month (from January to February) for run LL+VSL between 2020 and 2040. This occurs because the additional source of VSL Br produces a deepest October ozone minimum on top of the annual seasonal cycle, displacing the 2 nd local maximum in between the minima to later dates (see Fig. 7D). During the 2000 th decade, the location of the 2 nd maxima, representing the closing end of the ozone hole, expands all the way to June of the following year because the ozone hole depletion during October is so large that its 5 impacts persist until the following winter is reached: the year-round depletion of TOZ SP July expands from 1990 to 2010 for run LL , persisting ~7 years longer, from 1990 to 2017 for the run LL+VSL case. It is worth noting that because the ozone hole seasonal extension is not tied to a fixed TOZ value (as for example 220 DU) the ozone hole seasonal duration can be computed all the way to year 2100, even after the 1980-October standard ozone return date has already been achieved. These results indicate that even when LL Cl and LL Br will control the return date of the deepest ozone levels to the 1980-baseline 10 value, the future evolution of VSL Br sources are of major importance to determine the future influence of halogen chemistry on the stratospheric Antarctic ozone seasonal cycle.

The role of chlorine and bromine ozone loss cycles (ClOx LL vs. BrOx LL+VSL )
Bromine chemical cycles play a well-known role in the halogen-mediated springtime ozone hole formation (McElroy et al., 1986;Lee and Jones, 2002;Salawitch et al., 2005). Here we have used the same definition of odd-oxygen depleting families 15 as in Table 5 from , with the exception of the iodine family which is not considered in this work. Figure 9 shows the temporal evolution of the percentage loss due to each cycle respect to the total odd-oxygen loss rate as well as the partitioning between the chlorine and bromine components within the halogen family. In the following, note that crossed ClOx-BrOx cycles have been included into BrOx LL+VSL losses because both simulations include identical stratospheric LL Cl loading but a ~5 pptv difference in total bromine (see Fig. 1). 20 Between approximately 1980 and 2060 the dominant ozone depleting family within the springtime Antarctic ozone hole is halogens: ClOx LL + BrOx LL+VSL surpass the otherwise dominant contribution from NOx and HOx cycles (Fig. 9A): e.g., during the year of largest ozone depletion (i.e. October 2003), halogens represent more than 90% of the total odd oxygen loss at 100 hPa, while NOx and HOx cycles contribute ~5% and less than 2%, respectively. By year 2050, when the 1980-October baseline ozone return date is expected to occur, the overall BrOx LL+VSL cycles represent ~45% of the total ozone loss 25 by halogens occurring at 100 hPa (Fig. 9B) and ~35% when integrated in the stratosphere (Fig. 9C). Even though ClOx LL losses represent as much as 80 % of the halogen-mediated ozone loss during the 2000 th decade, the additional contribution from VSL Br drives bromine chemistry (BrOx LL+VSL ) to dominate ozone loss by halogens approximately by year 2070. The contribution of BrOx LL+VSL cycles to ozone loss was higher than ClOx LL also before 1975, i.e. before and during the fast increase in anthropogenic CFCs occurred (Fig. 9B). This implies that, although anthropogenic chlorine has controlled and 30 will control the long-term evolution of springtime stratospheric ozone hole, its overall depleting potential in the lowermost stratosphere is strongly influenced by the total (natural + anthropogenic) stratospheric inorganic bromine, with a non-Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys. Published: 5 October 2016 c Author(s) 2016. CC-BY 3.0 License. negligible contribution (up to ~30%) from the biogenic VSL Br oceanic sources. Within the run LL experiment, BrOx LL cycles never surpass the contribution of ClOx LL losses, revealing the significant enhancement of inter-halogen ClOx LL -BrOx LL+VSL depletion due to the additional source of natural VSL Br .
There is a clear variation on the height at which ClOx and BrOx LL+VSL cycles produces its maximum destruction, as well as the period of time when the losses by each family dominate respect to the others. For example, pure ClOx LL cycles account 5 for more than 80% of the total halogen losses above 10 hPa during the whole 21 st century, while BrOx LL+VSL cycles maximize close to the tropopause. Figure 10 shows that during the Antarctic spring, stratospheric bromine chemistry below 50 hPa has been at least as important as chlorine before and after the ozone hole era. Thus, the future evolution of stratospheric LL Cl levels will control the ozone hole return date, but the role played by VSL Br by that time will be as large as the one arising from LL Br . This effect will be most evident within the lower stratospheric levels: bromine is globally ~60 times more 10 efficient than chlorine in depleting ozone (Daniel et al., 1999;Sinnhuber et al., 2009), but its efficacy relies mostly on the background levels of stratospheric chlorine and the prevailing temperature affecting the rate of the inter-halogen crossed reactions (Saiz-lopez and Fernandez, 2016). Additionally, the extent of ClOx LL depletion within the Antarctic vortex relies on the occurrence of heterogeneous activation of chlorine reservoir species on polar stratospheric clouds, which in turn depend on ambient temperature. Then, the efficiency of BrOx LL+VSL depleting cycles relative to chlorine is reduced in the 15 colder lower stratosphere at high latitudes during the 2000 th decade (see lower panels on Fig. 10), while the BrOx LL+VSL contribution is larger at mid latitudes and increase in importance as we move forward into the future.
The representation of the ClOx LL and BrOx LL+VSL contributions shown in Fig. 11 allows addressing two interesting features related to the seasonal and long-term evolution of lower stratospheric Antarctic ozone. For any fixed year during the ozone hole era, bromine chemistry reaches a minimum during austral spring, while it increases during the summer and fall months. 20 For example, the BrOx LL contribution to total halogen loss at 100 hPa for year 2000 is 25% during October, 65% in December and greater than 80% by March. Thus, if the Antarctic return date delay is computed considering the baseline 1980 value for the fall months, a greater impact from VSL Br is observed (see Fig. 3c). Accordingly, the evaluation of the long-term impact of ClOx LL and BrOx LL+VSL cycles on the evolution of Antarctic ozone changes abruptly if we focused on the fall months instead of considering the October mean. In the lower stratosphere, chlorine chemistry is dramatically enhanced 25 during October due to the formation of the Antarctic ozone hole, but during summer and fall ClOx LL losses decrease, representing less than 20% of the total halogen loss (March mean) during the 21 st century.

Discussion and Concluding Remarks
We have shown that biogenic VSL Br have a profound impact on the depth, size and vertical distribution of the springtime Antarctic ozone hole. The inclusion of VSL Br improves the quantitative 1980-2010 model/satellite agreement of TOZ SP , and 30 it is necessary to capture the lowest October mean ozone hole values. Our model results also show that, even when the Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys. Published: 5 October 2016 c Author(s) 2016. CC-BY 3.0 License. maximum springtime depletion is increased by the inclusion of VSL Br , the future recovery of Antarctic ozone to the prevailing levels before 1980 is primarily driven by the evolution of the dominant LL Cl and LL Br sources: i.e. VSL Br sources does not change significantly the estimated return date. This can be explained considering the larger impact of bromine chemistry during periods of high inorganic chlorine loading, as well as due to the background impact of the additional bromine on the past global stratosphere. Other chemistry climate modelling studies estimated a decade enlargement of the 5 expected return date based on a single member simulation (Oman et al., 2016), but those studies considered an approximate VSL Br approach increasing the CH 3 Br lower boundary condition by ~5 pptv, while here we performed 6 independent simulations including geographically-distributed time-dependent VSL Br oceanic sources. Note, however, that free-running ocean interactive simulations as the ones performed in this work possess a very large model internal variability, so more ensemble members might be required to better address the important issue of the return date. The TOZ SP minimum and the 10 ozone hole depth in the lower stratosphere are both increased by 14% and 40%, respectively, when VSL Br is considered. This effect is more pronounced in the periphery of the ozone hole and within the heights of smaller ozone densities. Interestingly, biogenic bromine produces an enlargement of the OHA of 5 million km 2 , equivalent to that of the recently estimated Antarctic ozone healing due to the implementation of the Montreal Protocol. This large effect of oceanic VSL Br on the OHA highlights the importance of including biogenic bromine in climate assessments of the future Antarctic ozone layer. As the 15 anthropogenic emissions of LL Cl and LL Br are projected to decrease in the future following the Montreal protocol, the natural VSL Br relative contribution will represent as much as 40% of stratospheric bromine throughout the 21 st century, or even more if the oceanic VSL Br source strength and deep convection tropical injection increase in the near future (Hossaini et al., 2012;Leedham et al., 2013). Indeed, enhanced bromine BrOx LL+VSL cycles will dominate the chemistry of the lowermost stratosphere over Antarctica before a complete recovery of the global ozone layer from LL Br and LL Cl has occurred. Hence, 20 oceanic VSL Br possess leverage to significantly influence the future evolution of the Antarctic ozone layer. Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys.   Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys.           Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys. Published: 5 October 2016 c Author(s) 2016. CC-BY 3.0 License.   Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-840, 2016 Manuscript under review for journal Atmos. Chem. Phys. Published: 5 October 2016 c Author(s) 2016. CC-BY 3.0 License.