Robust effects of springtime Arctic ozone depletion on surface climate

Massive spring ozone loss due to anthropogenic emissions of ozone depleting substances is not limited to the austral hemisphere, but can also occur in the Arctic. Previous studies have suggested a link between springtime Arctic ozone depletion and Northern Hemispheric surface climate, which might add surface predictability. However, so far it has not been possible to isolate the role of stratospheric ozone from dynamical downward impacts. For the ﬁrst time, we quantify the impact of springtime Arctic ozone depletion on surface climate using observations and targeted chemistry-climate model experiments to isolate the effects of ozone feedbacks. We ﬁnd that springtime stratospheric ozone depletion is followed by surface anomalies in precipitation and temperature resembling a positive Arctic Oscillation. Most notably, we show that these anomalies, affecting large portions of the Northern Hemisphere, cannot be explained by dynamical variability alone, but are to a signiﬁcant degree driven by stratospheric ozone. The surface signal is linked to reduced shortwave absorption by stratospheric ozone, forcing persistent negative temperature anomalies in the lower stratosphere and a delayed breakup of the polar vortex - analogous to ozone-surface coupling in the Southern Hemisphere.These results suggest that Arctic stratospheric ozone actively forces springtime Northern Hemispheric surface climate and thus provides a source of predictability on seasonal scales. 11

positive phase of the Arctic Oscillation (AO) 13, 14 , similar to observations in the SH. Some analyses argue that this ozone-surface 23 climate connection could be useful for statistical predictions of NH seasonal climate 15,16 . 24 However, it remains difficult to disentangle the potential downward influence of ozone extremes from extreme dynamical 25 events in the stratosphere, for which a surface impact is well established [17][18][19] . Surface patterns coincident with ozone depletion 26 might be caused entirely by dynamical variability in the lower stratosphere, with ozone simply acting as a passive tracer of 27 such dynamical variability 20, 21 . Conversely, some studies based on models and observations conclude that ozone extremes 28 actively influence surface climate 13, 14 . These conflicting results arise from both a lack of simulations which explicitly isolate 29 the ozone feedbacks, and the specific analysis methods used in past studies. Until now, there is neither robust evidence for 30 a causal link between springtime stratospheric ozone and NH surface climate, nor have the feedback processes driven by 31 ozone been quantitatively assessed. Moreover, past studies focused on monthly averaged climate variables and chose fixed 32 reference months (March/April) to define springtime ozone extremes, ignoring inter-annual variations in the timing of those 33 events. In addition, they assessed the ozone-surface climate link by contrasting years with high and low springtime Arctic 34 ozone concentrations 13-15 , which does not allow them to isolate the surface signal forced by ozone depletion due to the mutual 35 dependence of ozone and stratospheric dynamics. 36 Here, we shed new light on the surface impacts of Arctic ozone depletion by (1) isolating ozone feedbacks from dynamical 37 contributions in model simulations and by (2) improving the detection of ozone depletion events and their surface signature 38 through consideration of their relative timing. A better understanding of the ozone contribution is expected to improve surface 39 forecasts 15,16,22 , which further motivates investigation of the effects of Arctic ozone depletion on surface climate. 40 Ozone-surface climate connection in observations 41 We start by revisiting springtime ozone depletion and associated surface patterns from 1980 to 2020 in the MERRA2 reanalysis 42 dataset 23 . We define springtime ozone minima based on partial stratospheric ozone column (30-70 hPa) over the polar cap 43 (60-90°N). For each year, the minima in daily ozone values within March and April are identified and ranked, and the day 44 exhibiting the lowest ozone value is termed as "central ozone minimum date". In the following, we use the terms "ozone 45 minima" and "ozone depletion" interchangeably (see supplementary material and Fig. S9). For further analysis, the 10 years -46 25% of the 41 year-long period -with the lowest springtime ozone values are considered. This detection method allows for a 47 better alignment of stratospheric ozone depletion and associated surface effects compared to previous studies 13, 14 . 48 In the 30 days following the central ozone minimum date we predominantly find negative sea level pressure (SLP) anomalies 49 in the polar region and positive SLP anomalies in midlatitudes, especially over northwestern Europe (Fig. 1 a-c). This pattern 50 resembles the positive phase of the AO, consistent with previous results in the context of ozone extremes 13 . The positive AO is 51 associated with regional temperature anomalies; we find warming over Siberia and large parts of Eurasia (up to 2 K) as well as 52 over western Europe (up to 1 K), which is adjacent to cooling over southeastern Europe. As expected for a positive AO, we find 53 that Arctic ozone depletion is followed by reduced precipitation over large parts of Europe and central Asia, and increased 54 precipitation over the Arctic.

55
Even though we find a strong connection between stratospheric ozone depletion and a positive AO at the surface, the spread 56 in the mean AO index averaged over the month after the central date between individual events is large, and for 2 out of the 57 10 events the AO is negative (Fig. 2). This spread is similar to the variable tropospheric response after Sudden Stratospheric 58 Warmings (SSWs 24 ) 25-27 . Yet, we see a clear shift towards a predominantly positive phase of the AO in the aftermath of 59 stratospheric ozone depletion. Our results not only confirm the robust statistical connection between springtime stratospheric 60 ozone values and surface climate reported previously, but the new detection method used here reveals an even larger surface 61 signal following stratospheric ozone depletion than reported by past studies 13, 14 (see Fig. S8).

62
Isolating the influence of ozone on surface climate 63 In order to establish the causality of the ozone-surface coupling, we perform targeted model experiments designed to isolate 64 ozone feedbacks, meaning the ozone impact on stratospheric dynamics and surface climate. We use two chemistry-climate-65 models, WACCM4 28 and SOCOL-MPIOM 29 . These models have different dynamical cores and chemistry modules (see 66 methods). With both models, we perform two simulations: one with fully interactive ozone chemistry (INT_O3) and one with 67 prescribed climatological ozone (CLIM_O3). Both integrations employ present-day boundary conditions. Unlike other studies 68 using a similar set-up 30-32 , experiments with prescribed ozone climatologies (CLIM_O3) still employ the chemistry scheme. 69 However, in these experiments, the three-dimensional ozone field is radiatively inactive. Thus, ozone purely acts as a passive 70 tracer in CLIM_O3 and ozone feedbacks on the atmospheric circulation are disabled. A similar set-up has already been used in 71 the context of SH ozone depletion 33 . Here, this approach allows us to apply the same definition for detecting the 25% most  In runs with specified ozone chemistry (CLIM_O3) in WACCM, springtime ozone minima are followed by significant 76 negative SLP anomalies over the polar cap, positive temperature anomalies over large parts of Eurasia and increased precipitation 77 close to the pole ( Fig. 1 g-i). Since ozone anomalies do not exert any radiative-dynamical feedback in the CLIM_O3 setting, 78 these springtime surface anomalies are solely due to dynamical variability and are linked to an exceptionally strong polar vortex 79 and a cold stratosphere. Surface anomalies in CLIM_O3 following Arctic ozone depletion are thus comparable to observations 80 in their sign and pattern for large parts of the NH, but are substantially weaker (cp. to Fig. 1 a-c). This suggests that surface 81 patterns found in the observations cannot be explained by dynamical variability alone.

82
Runs with interactive ozone (INT_O3), which additionally include ozone feedbacks, show significantly enhanced surface 83 anomalies compared to CLIM_O3 in the 30 days after springtime Arctic ozone depletion. More specifically, negative SLP 84 anomalies over the pole are up to 4 hPa larger ( Fig. 1 d), temperature anomalies in Eurasia are enhanced by more than 1 K 85 ( Fig. 1 e) and precipitation anomalies over the Arctic are increased (Fig. 1 f)  is reproduced by both model experiments (Fig. 3 a, c). In the lower stratosphere, a stronger polar vortex is associated with 105 reduced transport of ozone-rich air into the polar regions 32 , which contributes to the reduced Arctic ozone abundance. This 106 dynamical contribution can also be seen in the gradual evolution of the ozone minima starting in early winter, when chemical 107 ozone depletion is still negligible (see Fig. S7). to experiments with specified ozone (see Table S1). Arctic ozone anomalies therefore actively extend stratospheric winter

125
The ozone feedback mechanism presented here is the first description of the downward impact of springtime ozone depletion 126 in the NH in a mechanistic way and is consistent with our understanding of the dynamical impacts of the ozone hole in the SH:

127
In the Antarctic, stratospheric cooling caused by ozone depletion in spring is usually followed by a strengthening of the polar 128 vortex, which in turn facilitates planetary wave propagation, strengthens the BDC and results in a dynamical heating over the 129 pole ( 38, 39 and references therein). Even though ozone depletion is much less frequent in the Arctic than in the Antarctic, we 130 conclude that the large contribution of ozone depletion to springtime surface climate as well as the mechanism by which ozone 131 affects the stratospheric circulation is analogous on both hemispheres. The results presented here contribute to this discussion in two ways: First, they shed new light on the nature of the 146 ozone-surface climate connection in the NH -a relationship previously discussed controversially. Our modelling experiments 147 show in a robust manner that springtime NH surface patterns in the aftermath of strong stratospheric ozone depletion cannot be 148 explained by dynamical variability alone. Rather, ozone feedbacks represent an important contribution to the surface response. 149 We therefore conclude that interactive ozone chemistry is essential for weather and climate models to realistically reproduce 150 Northern Hemispheric spring conditions. A second contribution is that these novel findings create new incentives to explore the 151 value of stratospheric ozone for subseasonal to seasonal prediction, and they should serve as a motivation to explore ways to 152 include a more realistic representation of stratospheric ozone in forecast models and to further investigate the prediction skill 153 arising from stratospheric ozone depletion in current and future climate for both hemispheres. Despite the projected Arctic 154 ozone recovery, large dynamical variability will ensue in the future, leading to large episodic springtime depletion 50 . Hence, 155 Arctic ozone will continue playing an important role in future climate variability.

278
Extending to the lower thermosphere (5.1 · 10 −6 hPa) in altitude with 66 vertical levels and a well resolved stratosphere 28 , 279 WACCM has been documented to capture stratospheric trends and variability reasonably well and has been used in many 280 recent studies analysing interannual stratospheric variability (e.g. 9, 31, 32 53 ). WACCM has a horizontal resolution of 1.9°281 latitude and 2.5°longitude 28 , while the ocean has a nominal latitude-longitude resolution of 1°. Being coupled to an interactive 282 chemistry scheme 54 , WACCM calculates ozone concentrations over a set of chemical equations between a total of 59 species 283 28 and therefore actively simulates feedbacks between ozone and dynamics. In addition, WACCM can be run in a "specified  To assess the impact of ozone feedbacks, we contrast runs with fully interactive and specified ozone chemistry in both models.
where χ O 3 (p) denotes the ozone mixing ratio at pressure level p, ∆p describes the distance to the next pressure level in Pa,  standard deviations from the mean of the randomly created distribution, which is equivalent to a significance at the 95.4% level.

349
The procedure for a 2-sample bootstrapping significance test is conducted accordingly: To compare two composites, random 350 composites are created as described above in both datasets and the difference of both random composites was calculated.

351
Repeating this procedure 500 times, we create a distribution of 500 random samples. The difference between both samples is 352 considered significant if it differs more than 2 standard deviations from the mean value of the random distribution.  Calculation of dynamical heating rate 362 We calculate the vertical (w * ) and meridional (v * ) component of the residual circulation according to the transformed Eulerian whereT is the zonal mean temperature and S the atmospheric stability parameter 61 which can be calculated according to  2005,2002,2000,1997,1996,1995,1993,1990.