A stratospheric prognostic ozone for seamless Earth system models: performance, impacts and future

. We have implemented a new stratospheric ozone model in the European Centre for Medium-Range Weather Forecasts (ECMWF) system and tested its performance for different timescales to assess the impact of stratospheric ozone on meteorological ﬁelds. We have used the new ozone model to provide prognostic ozone in medium-range and long-range (seasonal) experiments, showing the feasibility of this ozone scheme for a seam-less numerical weather prediction (NWP) modelling approach. We ﬁnd that the stratospheric ozone distribution provided by the new scheme in ECMWF forecast experiments is in very good agreement with observations, even for unusual meteorological conditions such as Arctic stratospheric sudden warmings (SSWs) and Antarctic polar vortex events like the vortex split of year 2002. To assess the impact it has on meteorological variables, we have performed experiments in which the prognostic ozone is interactive with radiation. The new scheme provides a realistic ozone ﬁeld able to improve the description of the stratosphere in the ECMWF system, as we ﬁnd clear reductions of biases in the stratospheric forecast temperature. The seasonality of the Southern Hemisphere polar vortex is also signiﬁcantly improved when using the new ozone model. In medium-range simulations we also ﬁnd improvements in high-latitude tropospheric winds during the SSW event considered in this study. In long-range simulations, the use of the new ozone model leads to an increase in the correlation of the winter North Atlantic Oscillation (NAO) index with respect to ERA-Interim and an increase in the signal-to-noise ratio over the North Atlantic sector. In our study we show that by improving the description of the stratospheric ozone in the ECMWF system, the stratosphere–troposphere coupling improves. This highlights the potential beneﬁts of this new ozone model to exploit stratospheric sources of predictability and improve weather predictions over Europe on a range of timescales.


Introduction
The new emerging generation of seamless Earth System Models (ESMs) needs to be developed in ways that allow accurate performance in timescales from weather to climate, including seasonal and subseasonal scales. This requires slow evolving pro-20 cesses that influence the troposphere, like stratospheric processes, to be realistically included. The links between stratospheric ozone, polar vortex dynamics, and extreme winter weather over Europe are being increasingly recognised (e.g. Kolstad et al., 2010;Waugh et al., 2017;Kretschmer et al., 2018). Lack of detail in the description of the stratosphere is also linked to an unrealistic representation of stratosphere-troposphere coupling, which makes most existing models unable to exploit all potential sources of predictability deriving from the stratosphere (e.g. Scaife et al., 2016). 25 The stratospheric ozone layer accounts for 90 % of the total existing atmospheric ozone and plays fundamental roles for atmospheric processes and for life on Earth. Ozone in this atmospheric region provides a vital shield against harmful ultraviolet (UV) radiation, preventing the most energetic UV-C and UV-B wavelengths (wavelength bands below 300 nm) from reaching the Earth's surface. UV radiation is absorbed by ozone in the stratosphere via very exothermic reactions, therefore ozone is the main player in shaping the vertical profile of temperature in the stratosphere and has a fundamental role in the interactions 30 between radiation and dynamics in this region and the exchange of air masses with the troposphere. Unlike ozone in the troposphere, where its influence on physics and dynamics is dwarfed by the influence of other meteorological phenomena, a realistic distribution of ozone in the stratosphere is essential to correctly model the dynamical behaviour in this region.
Interannual dynamical variability of the polar vortex, in both hemispheres, causes large differences in the amounts of ozone depletion from year to year. In the Northern Hemisphere (NH), the occurrence of sudden stratospheric warming (SSW) events, 35 with temperatures in the polar stratosphere experiencing very rapid increases, lead to significantly less Arctic ozone loss than during cold Arctic years without SSW disturbances (e.g. Monge-Sanz et al., 2011;Solomon et al., 2014;Strahan et al., 2016).
Over Antarctica, the formation of the ozone hole every year is caused by the presence of ozone depleting substances (ODS) in the atmosphere, but its extent and duration depends on the particular dynamics of the polar vortex each year. The amount of ozone depletion then feeds back to temperature and winds through radiative interactions. Thus, correctly simulating the amount 40 of polar stratospheric ozone depletion during late winter/spring and allowing it to interact with radiation has also the potential to improve the way models reproduce stratosphere-troposphere coupling.
Stratospheric ozone research has been very active during the past 30 years (WMO, 2019); however, modelling tools required to answer remaining and emerging questions at different timescales regarding links between ozone, climate change and weather extremes, are not yet available. New seamless Earth System Models (ESMs) that integrate climate and weather elements of the 45 Earth System are starting to be developed and will provide valuable tools to address such questions. How to most efficiently incorporate appropriate descriptions of stratospheric processes, including stratospheric ozone, is still an open question. Such descriptions will need to exhibit the right compromise between realism and computational cost to be able to adequately perform at all timescales.
The Antarctic ozone hole was first discovered in the mid 1980s, and understanding processes regulating the amount and 50 distribution of ozone in the stratosphere became a high scientific priority for societal needs. For this reason, and to monitor the effectiveness of the Montreal Protocol, a set of complex atmospheric models was specifically developed to address stratospheric ozone related questions: chemistry-transport models (CTMs) and chemistry-climate models (CCMs) became the best modelling tools to understand links between chemical and dynamical factors governing the formation, distribution and destruction of stratospheric ozone. Nowadays, these modelling tools include very detailed atmospheric chemistry processes, based on the 55 most up-to-date scientific knowledge, and can provide very accurate simulations of stratospheric ozone (e.g. Eyring et al., 2007;Morgenstern et al., 2017).
Nevertheless, such full-chemistry level of detail is not affordable for high resolution multiple ensemble weather forecasting simulations, because of computational costs. Alternative stratospheric descriptions that are both realistic and affordable for all time scales are key needs for emerging seamless systems. In this work we assess the feasibility and performance of a linear 60 model for stratospheric ozone that can be implemented in any global circulation model (GCM) within an ESM at very low computational cost, yet providing quality comparable to the ozone field from world leading full-chemistry models.
The first linear ozone model was formulated by Cariolle and Déqué (CD) (Cariolle and Déqué, 1986) when the heterogeneous chemistry of the ozone hole was still unknown, and therefore the scheme parameterised only the effects of ozone 65 gas-phase chemistry, ignoring the heterogenous chemistry processes reponsible for polar ozone loss. Susbsequent versions of the CD model kept the initial approach but included an additional term to take into account the polar destruction of ozone at low temperatures (Cariolle and Teyssèdre, 2007). This is the ozone model currently used by the Integrated Forecast System (IFS) of the European Centre for Medium-Range Weather Forecasts (ECMWF). However, this CD approach has significant limitations in the way it represents heterogeneous ozone loss. The BMS model, as the CD one, is a linear representation of stratospheric ozone sources and sinks as a function of ozone concentrations and temperature. But the BMS model, unlike the CD one or any other previous linear ozone model, consistently includes both gas-phase and heterogenous chemistry for stratospheric ozone. Monge-Sanz et al. (2011) tested this new ozone model within the SLIMCAT 3D chemistry transport model (CTM) used to obtain the linear scheme, showing the superiority 75 of the BMS scheme over the CD scheme in a multiannual run covering the period 1991-2002. In their study they showed the capacity of the new ozone scheme to provide a stratospheric ozone field of comparable quality to the ozone field from the SLIMCAT full-chemistry model. More details on the differences between the BMS and the CD ozone models are given below in Sect. 2.1, and a full discussion comparing both formulations can be found in Monge-Sanz et al. (2011).
Besides the above mentioned decadal simulations within the SLIMCAT CTM, the BMS scheme has already been adopted 80 by global models for different applications, from numerical weather forecasting to tropospheric air-quality (Jeong et al., 2016;Badia et al., 2017), because of the more realistic simulation of stratospheric ozone it provides, compared to other available options like observation-based monthly climatologies or the CD scheme, for models that cannot afford stratospheric fullchemistry modules.
For the present study we have implemented the new BMS ozone in the ECMWF general circulation model (GCM) and 85 compared the performance of the new BMS and the default CD ozone model schemes in terms of the ozone distributions they provide. Then we have evaluated the way stratospheric ozone impacts meteorological fields at different timescales. This is the first time that the performance of an ozone model has been assessed for different timescales in a GCM, with the goal of evaluating its feasibility for seamless Earth systems simulations.
The structure of this article is as follows: Section 2 describes the model configuration used and the set of experiments 90 designed for this study, as well as giving an overview of the observations datasets used for validation of our model results.
Section 3 shows the ozone distribution results obtained for different experiments, regions and case studies. Then the impacts on meteorological fields are discussed in Sect. 4. The summary of results and conclusions for this study is found in Sect. 5, which also provides a discussion of future work and recommendations.

New ozone model in the IFS
We have implemented the stratospheric ozone model described by Monge-Sanz et al. (2011), the BMS ozone scheme, within the ECMWF Integrated Forecast System (IFS). The scheme represents the effects of stratospheric ozone sources and sinks following a linear approach, so that a CTM or GCM model can simulate time evolution of ozone by including an advected tracer whose concentration f evolves in time according to the following equation: where the coefficients c i (i = 0, 1, 2, 3) are tendencies derived from the full-chemistry CTM runs, and the termsf ,T ,c O3 are climatological reference values (in this case obtained from the full-chemistry output fields) for the ozone concentration f , temperature T and partial column of ozone above the considered location c O3 . The coefficients c i and the climatological terms where the tendency coefficients c i (i = 0, 1, 2, 3) include only gas-phase chemistry effects, making it necessary to add the 5 th term and the corresponding coefficient c 4 to account for ozone destruction related to heterogenous chemistry processes.

115
Cl EQ is the equivalent chlorine content of the stratosphere, and varies from year to year. This fifth term is only active when temperature falls below 195 K in daytime for latitudes polewards of 45 • , and the coefficient c 4 is calculated with different localised temperature term like the one in the default ECMWF scheme.
The treatment of heterogeneous chemistry is one fundamental difference in the new (BMS) scheme, which includes stratospheric ozone chemistry processes, both gas-phase and heterogeneous, in a consistent embedded way for all locations using only the first four linear coefficients c i (i = 0, 1, 2, 3). Unlike previous stratospheric ozone linear schemes, the BMS scheme is the first one to include heterogeneous chemistry effects and gas-phase chemistry in a consistent, implicit way in all terms of allows for more realistic interactions between parameterised ozone, radiation and temperature, and therefore better response and feedbacks to meteorological conditions, than previous ozone parameterisation approaches.

ECMWF model experiments
We have used the Integrated Forecast System (IFS) of ECMWF to run forecast experiments at medium-range (10 days) and vertical levels, up to 0.01 hPa. Ozone concentrations are initialised from operational analyses at the start of each experiment (00UTC on day 1) and then, unless otherwise stated, ozone is left to evolve freely along the duration of the experiment using either the CD or the BMS scheme.
In the default IFS configuration the radiation scheme does not employ the prognostic ozone, instead it uses an ozone climatology in the form of zonal-mean monthly-mean ozone values. In the IFS version used in this study, the ozone climatology is 150 derived from the MACC reanalysis (Inness et al., 2013), a dataset that covers the period 2003-2011 at T255 horizontal resolution on 60 vertical levels. However, for some of our experiments the standard IFS model has been adapted so that the prognostic ozone provides the input to the radiation scheme (Table 1), thus allowing for feedbacks between the ozone scheme and model dynamics.

Datasets for validation
In   Interannual variability is more realistic with the new BMS scheme compared to the reanalysis, while the default scheme does not show enough variability, especially for the fraction area covered by the ozone hole, which manifests that the description of heterogenous chemistry processes in the default scheme is not realistic enough to deal with meteorological interannual 230 variability in the Antarctic region. Since CAMSiRA uses the CD scheme in the stratosphere, Fig. 3 also suggests that with the new BMS scheme, assimilation increments in the reanalysis would be reduced for the ozone field.
The formation of the hole commences earlier when using the new ozone scheme, this is at least partly due to the fact that the new ozone scheme is able to capture the ozone loss that in late winter starts to occur at the edge of the polar vortex, while the heterogeneous treatment in the scheme currently used by ECMWF cannot reproduce this process. The ozone hole closure It is also worth noting that these ozone hole diagnostics are related to the total ozone column (TOC), and that realistic TOC values do not necessarily imply that the depletion in the model is occurring at the right altitudes. For similar TOC values over the Antarctic, we have shown that the new BMS scheme provides a much more realistic ozone vertical profile than the default ozone scheme in the ECMWF model (Sect. 3.1), also in agreement with previous studies using the BMS scheme Jeong et al., 2016;Badia et al., 2017).
Regarding the duration and extent of the ozone hole, several studies have shown how the Antarctic stratospheric ozone hole feeds back to dynamics and radiation causing changes in tropospheric winds and climate (e.g. Kang et al., 2011;Polvani et al., 2011;Orr et al., 2012;Haase et al., 2020). It is therefore important that future Earth System Models use a stratospheric ozone description that realistically captures the ozone hole intensity and evolution, to correctly simulate tropospheric trends.

Midlatitudes
One of the longest ozone records in Europe corresponds to the Alpine station of Hohenpeissenberg (Fig. 4).

295
SSW events like this one in winter 2015/2016 are good case studies to assess the new stratospheric ozone scheme. Important feedbacks between ozone and dynamics take place in this type of events: The occurrence of a SSW reduces the amount of ozone loss over the Arctic, therefore more ozone is available to absorb UV radiation, which contributes to a further increase of temperature in the region. The fact that past versions of the operational ECMWF forecast model reproduced SSW events overall weaker (colder) than observed (e.g. Diamantakis, 2014) is a further indication that a forecast model cannot properly account 300 for these feedbacks when using an ozone climatology in the radiation scheme. A realistic scheme for prognostic stratospheric ozone is therefore able to contribute to a better reproduction of SSW events and their feedbacks within the model.

Impact on meteorological fields
The previous sections have shown the improved stratospheric ozone distribution and variability obtained when using the new BMS ozone model in the ECMWF system. From a weather and climate modelling perspective, we are interested in how the 305 new representation of stratospheric ozone affects meteorological fields. To evaluate this, the prognostic ozone scheme has been made interactive with the radiation scheme in the ECMWF model. The corresponding impact on meteorological variables has been compared with results from the default ECMWF ozone configuration in which the radiation scheme uses an ozone climatology.

Impacts on medium-range forecasts 310
The mean error in the temperature field is shown in Fig. 7 for two forecast model experiments in which the prognostic ozone has been made interactive with the radiation scheme. Both experiments are 10-day forecast covering the period August 2012-July 2013. The only difference between the configuration of the two experiments is in the prognostic ozone model: one of them uses the CD ozone, the other one the new BMS ozone. In the stratosphere (above 100 hPa), the new ozone clearly reduces the model temperature bias by up to 1 K. Smaller improvements can also be seen for lower levels in the extratropics (Fig. 7).

315
The tropical region at 100 hPa is the only region in which the use of the new ozone scheme increases the temperature mean error. This is most probably attributable to the negative bias in tropical ozone exhibited by the BMS scheme. Although the negative ozone bias is found above 100 hPa, it means that more UV can reach lower altitudes (less ozone above), therefore less ozone below (more dissociation by UV); and with less ozone there is also a decrease in local T, leading to the larger negative T bias over the tropics at 100 hPa shown in Fig. 7. For tropospheric levels there is an overall improvement in the T mean error, 320 although not everywhere for all leadtimes, but results for the troposphere are not statistically significant.
To display the statistical significance of these results, the differences in normalised error change in the temperature field, and corresponding significance 95% bars, are shown in Fig. 8. The improvement in the model error in the stratosphere above 100 hPa is clearly evident and statistically significant with the new ozone. For altitudes below 100 hPa, it shows small increases in model error with the new scheme but these are not statistically significant, except in the tropical UTLS. Interpreting results in 325 the tropical UTLS is not straightforward due to the many interplaying factors in this region, but the negative concentrations bias exhibited at higher altitudes over the tropics is most probably playing a role. In summary, the new scheme, therefore, improves the temperature behaviour in the stratosphere without degrading the temperature field in the troposphere; the only atmospheric region where there is some degradation in the T field is the tropical LS and this is at least partly due to a known bias in the new ozone scheme.

330
A similar comparison can be performed between the forecast experiment using the prognostic BMS ozone interactive with radiation and the default ECMWF operational configuration in which the radiation scheme uses an ozone climatology. Fig. 9 shows the mean error in temperature for these two experiments and Fig. 10 shows the corresponding differences in normalised error change with the 95% significance bars. These results show a similar pattern to those from the comparison of the two prognostic schemes, i.e. the BMS scheme provides an improved temperature field in the stratosphere compared to the default 335 climatology, with the exception of the tropical region at 50 hPa and the NH extratropics at the same 50 hPa level. The new ozone also provides mean error improvements for the troposphere outside the tropics although results are only marginally statistically significant or not significant. Note also that differences in the troposphere are one order of magnitude smaller than in the stratosphere. In the tropical troposphere a small degradation can be found, which unlike in the comparison of both schemes now becomes statistically significant. This is related to the fact that this type of ozone linear models is not designed 340 for tropospheric use, and in the troposphere the use of a realistic climatology could be considered a good alternative for NWP purposes.
The rest of this paper focuses on the evaluation of meteorological impacts for experiments using the new BMS prognostic ozone compared to the control experiments using the default climatology configuration. It is worth noting that the ECMWF 345 operational model has gone through version updates after our study, and the ozone climatology has been updated following Hogan et al. (2017). Although beyond the scope of this paper, comparison of biases for the currently operational default configuration should be a matter of future investigation.

Impacts during SSW events
The winter stratospheric variability in the NH high latitudes is dominated by the occurrence of sudden stratospheric warmings  Figure 11 shows that, with the new prognostic ozone scheme, temperature at 5 hPa becomes warmer over the Eastern Arctic region (up to 20 K warmer) compared to the default model configuration (Fig. 11a,b,d), bringing it closer to the operational 360 analysis (Fig. 11c). We have compared t+240 h in the forecast experiments, from the 28 th of January forecast, against the corresponding operational analysis for the 7 th February, to allow for the maximum ozone response. The improvement seen at 5 hPa is also seen at lower altitude levels (figure not shown).
For these two experiments we have also examined the impact on the wind field. Figure 12 shows the change in RMS error for the meridional wind velocity for different lead times, averaged over the experiment duration. The improvement in wind 365 errors when BMS ozone is used consistently increases with lead time, and is transferred from the stratosphere down to the troposphere for high latitudes from day 6. By day 10 in the forecast the error reduction in the wind field is statistically significant in the troposphere. During SSW events, when both temperature and ozone distributions change rapidly in the stratosphere, a climatology-based ozone field cannot pass information to the radiation code that resembles the actual atmospheric situation, therefore the model misses the potential source of tropospheric predictability that comes from the downward propagation of 370 the stratospheric signal.

Long-range impacts
We have performed two seasonal experiments with start dates in May and November, covering the period 1981-2010, using an horizontal resolution of T255, 91 vertical levels and three members. The control experiment uses the MACC ozone cli-matology inside the radiation code, while the BMS experiment uses the new prognostic ozone interactive with the radiation code. Temperature differences with respect to ERA-Interim for these two experiments are shown in Figure 13, averaged over DJF (upper panels) and MAM (lower panels). There is a clear improvement around 50 hPa when using the new ozone scheme for both seasons and all latitudes, especially over the SH mid and high latitudes. In these SH regions the BMS prognostic ozone reduces differences by up to 4.0 K. Also for levels above 20 hPa differences with respect to ERA-Interim are reduced, especially in the summer hemisphere, by more than 1.0 K during DJF and MAM. These results have shown that a prognostic 380 ozone field contributes to more realistic temperatures in the SH stratosphere than a climatology, also for the seasons following the Antarctic ozone hole months. Figure 14 shows the zonal averaged differences in zonal wind between the two experiments for the SON season. The new ozone experiment shows stronger zonal winds over the Antarctic vortex edge latitudes between 20-400 hPa, which is physically linked to the lower concentrations of ozone simulated by the new scheme over this region compared to the default climatology.

385
This strengthening of winds is in overall agreement with the findings in Son et al. (2008). Seasonal experiments performed under the SPECS 1 EU project also showed improvements in the equatorial winds and the QBO signal when the new BMS prognostic ozone was made interactive with radiation (Knight et al., 2016).

North Atlantic Sector
Moving to the North Atlantic Sector winter, we find that the new ozone scheme has a positive impact on the signal to noise 390 ratio; accordingly, using ERA-Interim as reference, the correlation value for the winter North Atlantic Oscillation (NAO) index is almost doubled compared to the default configuration (Fig.15), increasing from 0.25 in the control experiment (default ozone climatology) to 0.44 in the experiment with the new BMS prognostic ozone. These experiments, where the only difference is the stratospheric ozone representation, allow us to attribute this increase in NAO model performance to stratospheric sources.
A more realistic stratospheric ozone distribution improves the ozone concentration gradients between the Pole and the equator, 395 which modifies the latitudinal heating gradient in the LS region compared to the default model configuration. This affects LS wave breaking and winds and has also an effect on the altitude distribution of the tropopause. The combination of these effects impacts the tropospheric pressure gradient between low and high latitudes and therefore the NAO signal.
These results show that a more realistic stratospheric ozone field contributes to a more realistic stratosphere-troposphere coupling in the model. The links between the NAO and winter time weather over Europe are well established (e.g. Cattiaux et al., ration (ozone climatology in the radiation scheme), the second SEAS5 experiment uses the same configuration except that ozone is replaced by the BMS prognostic ozone model interactive with radiation. The interannual variability of the SH polar vortex is shown in Fig. 16 for these seasonal experiments initialised on the 1 st of August over the period 1993-2015 (2002 has been excluded from the analysis shown in this figure). From the top panel in Fig. 16 it can be seen that the seasonality of 410 the stratosphere in this region is not realistic with the default SEAS5 compared to ERA-Interim; the vortex shift-down occurs too early compared to the reanalysis. When the new BMS prognostic ozone is used the timing of the SH polar vortex is in much better agreement with ERA-Interim, and the interannual variability increases. Byrne and Shepherd (2018) in their study using ERA-Interim reanalysis, found that Antarctic ozone depletion has caused a seasonal delay in the breakdown of the SH polar vortex along the period 1980-present, and they also pointed out that feedbacks between ozone and dynamics may be

Discussion and Conclusions
In this section we summarise the main findings of our study, and discuss further work plans and recommendations deriving 435 from our investigations.

Summary
We have implemented the stratospheric ozone model by Monge-Sanz et al. (2011) (also known as BMS model) in the ECMWF system, compared its performance to that of the default ozone used by ECMWF, and assessed its impacts on meterological fields at medium-range and seasonal time scales. The BMS scheme is the first stratospheric ozone linear model that consistently 440 accounts for heterogeneous chemistry (e.g. ozone destruction due to polar stratospheric clouds), instead of using a separate adhoc term, providing a more realistic link with temperature and radiation. The new approach is in better agreement with the current scientific knowledge of chemical and physical processes that affect stratospheric ozone (WMO, 2019) than approaches adopted by previous linear ozone models (McLinden et al., 2000;McCormack et al., 2006;Cariolle and Teyssèdre, 2007).
The present study is, to the best of our knowledge, the first time that the impacts of a stratospheric ozone model are assessed 445 at different NWP timescales to evaluate its performance towards its implementation in a seamless model system. We have shown that the new scheme provides significantly better ozone distribution and variability in the ECMWF model than the currently default ozone configuration, showing particularly good agreement with observations over the high southern latitudes and the ozone hole season, even during the unusual atmospheric conditions of the 2002 Antarctic vortex split.
When used interactively with radiation in the ECMWF model, the BMS ozone scheme reduces stratospheric temperature 450 biases both in medium-range and seasonal time scales, compared to the default ECMWF model configuration in which the radiation scheme uses an ozone climatology, and improves temperature and wind fields during Arctic SSWs. We have also shown that the BMS ozone improves the NAO signal in seasonal model runs, therefore contributing to a more realistic stratospheretroposphere coupling in the model. The interannual variability and seasonality of the SH polar vortex is also improved when using the BMS ozone model in runs performed with the ECMWF seasonal system (SEAS5), compared to the default SEAS5 455 configuration which uses an ozone climatology. All this demonstrates that the BMS scheme is realistically linked to temperature and dynamics and therefore well prepared to adapt and feedback to rapid changes in meteorology, the same adaptability cannot be achieved with an ozone climatology in the radiation scheme.
Our results also provide evidence for the need of a realistic prognostic stratospheric ozone field in ESMs for these models to perform more accurately at different time scales. The realistic stratospheric concentration values obtained with the BMS 460 scheme, its high adaptability to usual and unusual meteorological conditions at different time scales, together with its low computational cost, make the BMS scheme an excellent option to model stratospheric ozone within ESMs. The BMS scheme is able to model stratospheric ozone with a degree of complexity that provides realistic stratospheric ozone distributions, of comparable quality to the ozone field from world-leading full-chemistry models, while keeping low computational costs suitable for resolution and production times required for weather forecasting (both at medium-range and seasonal time scales).

465
Future work should also investigate the performance of the BMS scheme within other global models and exploit its benefits for different ESMs applications beyond medium-range NWP and seasonal prediction, into climate time scales.
Next, we briefly discuss ongoing further developments of the BMS model version we have used here, and how they are expected to improve the results obtained with the current version. We also provide a summary of benefits that the implementation of a realistic prognostic stratospheric ozone brings to seamless Earth System Models, from medium-range NWP forecasts to seasonal prediction and Reanalysis production.

Ongoing and Future work
To

Tropospheric ozone treatment
The ozone model discussed in this study, as other ozone linear schemes like the default one used by the ECMWF system, are 490 designed for the stratosphere; their use is not recommended for the tropospheric region, where ozone is affected by highly nonlinear processes involving pollutants and ozone precursors. A realistic representation of tropospheric ozone is the full-chemistry approach, but this is still unviable for operational high resolution models due to high computational costs. Alternatives for representing tropospheric ozone include the use of an up-to-date climatology based on observations or reanalysis that would be merged to the prognostic ozone in the stratosphere.

Benefits for seamless Earth System Models
A realistic stratosphere is increasingly recognised as one of the keys to develop seamless Earth System Models, due to the role it plays for tropospheric processes at different time scales, from weather to climate. The ozone model in our study is a valuable contribution to achieve seamless use of emerging ESMs, as it offers similar accuracy to a full-chemistry model for stratospheric ozone, allowing for ozone-climate feedbacks that level with those provided by current chemistry-climate models (CCMs) with interactive stratospheric ozone, while keeping the computational cost affordable for weather forecast applications and resolutions.
Additionally, by implementing a realistic scheme for prognostic ozone, and ideally also for other radiative active gases in the stratosphere (see e.g. Monge-Sanz et al., 2013), the system will be better prepared for the eventual operational use of interactive full-chemistry, as several feedbacks within the model will already have been investigated with the ozone model we propose, 505 allowing for compensation errors to be identified and possibly eliminated.

Benefits for long Reanalyses
Long reanalyses have become an essential part of weather and climate scientific research and applications. Recent major international projects, like the SPARC 2 Reanalysis Intercomparison Project (SRIP), part of the World Climate Research Program (WCRP) core activities, have identified areas that will need more attention in the production of future reanalyses in order to 510 represent a more realistic stratosphere (Fujiwara et al., 2017). The representation of stratospheric ozone in the atmospheric models used to produce the reanalyses is one of these aspects (Monge-Sanz et al., under review, 2020). The adaptability the new BMS stratospheric ozone scheme shows to very different meteorological situations makes it an excellent candidate for use in oncoming Earth System Reanalysis, where climate forcings and their feedbacks will need to be accurately accounted for.
Author contributions. BMMS designed the study, implemented the new ozone scheme in the ECMWF system, performed experiments, 515 results validation and interpretation, wrote the paper and coordinated co-authors' contributions. AB, NB, MD, JF, LM, IP and AW contributed to data analysis and results interpretation. LM and IP also contributed to performance of seasonal experiments. LJ designed the visualization tool used for validation against ozone profile observations. LJG, RJH, NW, MPC and TGS provided insightful feedback and contributed to review draft versions of the manuscript.
Competing interests. The authors declare that they have no conflict of interest.
Acknowledgements. This study was partially funded by the MACC-II and SPECS FP7 EU projects. BMMS and LJG also acknowledge funding from the UK Natural Environment Research Council (NERC) through the ACSIS project (North Atlantic Climate System Integrated Study) led by the National Centre for Atmospheric Science (NCAS). The first author is very grateful to Adrian Simmons, Agathe Untch and Jean-Jacques Morcrette for many helpful discussions and their valuable initial guidance with the ECMWF system; special thanks also to Franco Molteni for useful discussions during the preparation of this manuscript. We also thank Paul Burton, Gabor Radnoti and the ECMWF

525
User Support Team for their help with the IFS technical environment.  Table 2. Seasonal experiments using the new ozone scheme ('BMS'), or the default ozone configuration in which the radiation sees an ozone climatology ('CLIM').
For each experiment information on model version, resolution, period covered, as well as the number of ensemble members is also included.