Since the emergence of the Antarctic ‘ozone hole’ in the 1980s¹ and elucidation of the chemical mechanisms^{2, 3, 4, 5} and meteorological conditions⁶ involved in its formation, the likelihood of extreme ozone depletion over the Arctic has been debated. Similar processes are at work in the polar lower stratosphere in both hemispheres, but differences in the evolution of the winter polar vortex and associated polar temperatures have in the past led to vastly disparate degrees of springtime ozone destruction in the Arctic and Antarctic. We show that chemical ozone loss in spring 2011 far exceeded any previously observed over the Arctic. For the first time, sufficient loss occurred to reasonably be described as an Arctic ozone hole.

In the winter polar lower stratosphere, low temperatures induce condensation of water vapour and nitric acid (HNO₃) into polar stratospheric clouds (PSCs). PSCs and other cold aerosols provide surfaces for heterogeneous conversion of chlorine from longer-lived reservoir species, such as chlorine nitrate (ClONO₂) and hydrogen chloride (HCl), into reactive (ozone-destroying) forms, with chlorine monoxide (ClO) predominant in daylight^{5, 7}.

In the Antarctic, enhanced ClO is usually present for 4–5 months (through to the end of September)^{8, 9, 10, 11}, leading to destruction of most of the ozone in the polar vortex between ~14 and 20 km altitude⁷. Although ClO enhancement comparable to that in the Antarctic occurs at some times and altitudes in most Arctic winters⁹, it rarely persists for more than 2–3 months, even in the coldest years¹⁰. Thus chemical ozone loss in the Arctic has until now been limited, with largest previous losses observed in 2005, 2000 and 1996^{7, 12, 13, 14}.

The 2010–11 Arctic winter–spring was characterized by an anomalously strong stratospheric polar vortex and an atypically long continuously cold period. In February–March 2011, the barrier to transport at the Arctic vortex edge was the strongest in either hemisphere in the last ~30 years (Fig. 1a, Supplementary Discussion).

The persistence of a strong, cold vortex from December through to the end of March was unprecedented. In the previous years with most ozone loss, temperatures (T) rose above the threshold associated with chlorine activation (T_act, near 196 K, roughly the threshold for the potential existence of PSCs) by early March (Fig. 1b, Supplementary Figs 1, 2). Only in 2011 and 1997 have Arctic temperatures below T_act persisted through to the end of March, sporadically approaching a vortex volume fraction similar in size to that in some Antarctic winters (Fig. 1b). In 1996–97, however, the cold volume remained very limited until mid-January and was smaller than that in 2011 at most times during late January through to the end of March (Fig. 1b, Supplementary Figs 1, 2).

Daily minimum temperatures in the 2010–11 Arctic winter were not unusually low, but the persistently cold region was remarkably deep (Supplementary Figs 1, 2). Temperatures were below T_act for more than 100 days over an altitude range of ~15–23 km, compared to a similarly prolonged cold period over only ~20–23 km altitude in 1997; below ~19 km altitude, T < T_act continued for ~30 days longer in 2011 than in 1997 (Supplementary Fig. 1b). In 2005, the previous year with largest Arctic ozone loss⁷, T < T_act occurred for more than 100 days over ~17–23 km altitude, but all before early March.

The winter mean volume of air in which PSCs may form (that is, with T < T_act), V_psc, is closely correlated with the potential for ozone loss^{7, 15, 16, 17}. In 2011, V_psc (as a fraction of the vortex volume) was the largest on record (Fig. 1c). Both large V_psc and cold lingering well into spring are important in producing severe chemical loss^{7, 15, 16}, and 2010–11 was the only Arctic winter during which both conditions have been met. Much lower fractional V_psc in 1997 than in 1996, 2000, 2005 or 2011 (Fig. 1c) is consistent with less ozone loss that year^{16, 17}. Factors playing secondary parts in governing interannual variability in ozone destruction, including vortex strength, structure and position relative to the cold region, also favour large loss in 2011 (Supplementary Figs 2, 3, Supplementary Discussion). However, despite the fraction of the vortex with T < T_act and mid-March temperatures sporadically approaching those seen in the Antarctic (Fig. 1b, Supplementary Fig. 1a), even in 2011 temperatures were much higher, and the cold regions much smaller, than those in most Antarctic winters.

Satellite trace-gas and PSC measurements highlight the stark contrast between polar processing in 2010–11 and that in typical Arctic winters, and the parallels with Antarctic conditions (Figs 2, 3). In 2011, PSCs or aerosols were abundant until mid-March (Fig. 3a; consistent with a deep region with T < T_act, Fig. 3b), much later than usual in the Arctic^{18, 19, 20}, with vortex-average amounts at some altitudes similar to those in the Antarctic and dramatically larger than the near-zero values at that time in most Arctic winters. Furthermore, PSCs in 2011 spanned an altitude range comparable to that in the Antarctic, an uncommon occurrence in the Arctic^{18, 19, 20}. Particles in long-lasting PSCs can grow large enough to sediment, resulting in denitrification, permanent removal of HNO₃ from the stratosphere^{7, 12}. By late March 2011 no PSCs remained (Fig. 3a), yet HNO₃ mixing ratios were much lower than observed in any previous Arctic winter (Fig. 2a). The continuing depression in HNO₃ after PSCs had evaporated indicates denitrification. Albeit less severe than in typical Antarctic winters (Fig. 2b, c, 3c), the extent and degree of denitrification in 2011 were unmatched in the Arctic, approaching the range of Antarctic conditions for the first time.

Decreasing HCl and increasing ClO signify chlorine activation (Fig. 2d–i). Some ClO enhancement has occurred in all recent Arctic winters, but has never been as prolonged and extensive as that in 2011. In late February, high ClO pervaded the sunlit portion of the vortex. The 2011 values vastly exceed the range previously observed in the Arctic from late February through to the end of March. They also briefly lie outside the Antarctic seasonal envelope, primarily because the higher solar zenith angles of the Antarctic measurements used here lead to ~30% lower ClO under fully activated conditions. In late February, HCl values (unaffected by solar zenith angle issues) fall along the lower boundary of the Antarctic envelope, confirming the picture seen in ClO. The vertical extent of chlorine activation was also comparable to that in the Antarctic (Fig. 3d, e).

In previous cold Arctic winters, chlorine was deactivated (converted from ozone-destroying forms into less reactive reservoir species) by mid-March¹¹; even in 1997, ClO started to decline by late February (Fig. 2g). In 2011, by contrast, ClO began decreasing rapidly only about a week earlier than is typical in the Antarctic. ClO data in late February 1997 indicate that not only were maximum values lower than those in early March 2011, but also the vertical range of enhancement was shallower, with weaker activation at low altitudes than in 2011 (Fig. 3e), consistent with the higher altitudes and decreasing extent (Figs 1b, 3b, Supplementary Fig. 2) of T < T_act.

When chlorine is deactivated, whether it is converted first into HCl or ClONO₂ depends sensitively upon HNO₃ and ozone abundances. In the Arctic, chlorine is normally deactivated through initial reformation of ClONO₂. In the severely denitrified and ozone-depleted Antarctic vortex, production of ClONO₂ is suppressed and that of HCl highly favoured^{11, 12, 21}. In March 2011, the recovery of HCl followed a much more Antarctic-like pathway than has been observed in any other Arctic winter.

The largest Arctic chemical ozone loss was previously observed in 2005, followed closely by 2000 and 1996^{7, 12, 13, 14}. Although low temperatures persisted until the end of March 1997, the ozone loss in that year was far less. No previous year rivals 2011, when the evolution of Arctic ozone more closely followed that typical of the Antarctic (Fig. 2j). Ozone profiles in late March 2011 resemble typical Antarctic late-winter profiles much more strongly than they do the average Arctic one (Fig. 3f). Because mixing in April 2011 (for example, lamination events larger than that shown in Fig. 3f) entrained ozone-rich air into the vortex, the slight decrease in vortex-averaged ozone at a potential temperature of 485 K from 26 March to 20 April (from ~1.8 to ~1.6 p.p.m.v., Fig. 2j) indicates continuing chemical loss during this interval.

Chemical loss is difficult to quantify in the Arctic, where transport from above replenishes ozone in the lower stratospheric vortex, obscuring the signature of chlorine-catalysed destruction^{12, 22, 23}. The evolution of the long-lived trace gas nitrous oxide (N₂O) reflects steady downward transport throughout the 2010–11 winter–spring, indicating that subsidence partially masked chemical loss. Horizontal transport can also confound the signature of chemical loss, bringing air into the vortex that has either higher²⁴ or lower¹⁴ concentrations of ozone, depending on the altitude and latitude from which it originates.

Representative results from two types of chemical loss calculations^{24, 25, 26, 27, 28} based on balloon-borne and satellite observations are shown in Fig. 4. The differences (up to ~0.4 p.p.m.v. at the end of March 2011) in estimates derived from the various methods and data sets imply some uncertainty in the chemical loss determination. Year-to-year differences in the amount of ozone loss are very similar when obtained from any method/data set combination, however, indicating a high degree of precision in the relative amount of calculated loss between different years. Chemical destruction was severe between ~16 and 22 km altitude, with the largest loss exceeding 2.5 p.p.m.v. by 26 March 2011 (Fig. 4a). By 31 March 2011, chemical loss was nearly double that in 2005 from ~18 to above 22 km, and similar to that in 2005 at lower altitudes (Fig. 4b, c). From ~18 to 20 km, more than 80% of the ozone present in January had been chemically destroyed by late March. Chemical removal in 1996 and 2000 started at a rate similar to that in 2011 (Fig. 4c), but ceased by late March; maximum losses in 2000 approached those in 2011, but extended over a much smaller vertical range (Fig. 4b). Loss in 1996, 2000 and 2005 considerably exceeded that in 1997, with greater destruction at lower altitudes in those years contributing more to total column loss^{7, 12, 13}. Chemical loss in 2011 was two to three times larger than that in 1997, and about twice that in 1996 and 2005 above ~16 km; from ~15 to 23 km it was comparable to that in the Antarctic ozone hole in 1985²⁹. Single ozone-sonde station measurements in early April 2011 suggest continuing ozone loss (Fig. 4c).

Although the meteorology during March–April was similar in 1997 and 2011, ozone loss was much more pronounced in 2011. Photochemical box model simulations (Supplementary Fig. 4, Supplementary Discussion) elucidate how early winter conditions set the stage for record springtime ozone destruction in 2011. Chlorine activation brought on by enduring cold from December through to the end of February led to ~0.7–0.8 p.p.m.v. lower ozone at the beginning of March 2011 (Figs 2j, 4c). The early onset of continuous cold also facilitated formation of PSC particles large enough to sediment, resulting in ~4 p.p.b.v. less HNO₃ by March in 2011 than in 1997 (Fig. 2a). The degree of denitrification has a profound impact on the severity of springtime Arctic ozone loss³⁰. By delaying chlorine deactivation, lower HNO₃ by 1 March was responsible for ~0.6 p.p.m.v. more ozone loss after that date in 2011 than in 1997 (Supplementary Fig. 4, Supplementary Discussion). The effects of denitrification and early-winter loss together account for the disparity in ozone depletion in these two winters (~1.5 p.p.m.v. more loss at 460 K in 2011 than in 1997, Fig. 4c, Supplementary Fig. 4). Loss as severe as that in 2011 thus requires T < T_act, with consequent chlorine activation and ozone destruction, early in winter (as in 1996, 2000 and 2005, but not in 1997), a cold period and region before March sufficient to allow widespread denitrification, and the persistence of a cold polar vortex into April (as in 1997, but not in 1996, 2000 or 2005).

Total column ozone is a predominant factor determining exposure of Earth’s surface to ultraviolet radiation^{7, 12}. In the context of previous Arctic winters, 2011 was truly remarkable: the fraction of the Arctic vortex in March with total ozone less than 275 Dobson units (DU) is typically near zero, but reached nearly 45% in 2011 (Fig. 5a). Because of the dynamically-driven correlation between total ozone and lower-stratospheric temperature^{23, 31, 32, 33, 34} (Supplementary Discussion), the abiding cold in 1997 and 2011 would have led to lower March total ozone than in other Arctic winters even without chemical loss; dynamical conditions in March–April 1997 particularly favoured low total ozone³³ (Supplementary Discussion). In March 2011, however, the area of low total ozone covered more than twice as much of the vortex as in 1997, and the daily vortex ‘ozone deficit’ (Supplementary Fig. 5a) was 30–50 DU larger, consistent with the greater chemical loss (Fig. 4). Maximum 2011 vortex fractions of low ozone approached those in early Antarctic ozone holes (Fig. 5a). The close correspondence between the vortex and both low total ozone and the large Arctic total ozone deficit (Fig. 5b, d) implies that low total ozone in March 2011 resulted primarily from chemical loss^{31, 32} (Supplementary Discussion). The ozone deficit in the Antarctic (Fig. 5e) shows a maximum over 0–90° W, and a minimum over 90–200° E, reflecting a vortex position in 2010 different to that in the reference state (which is less robust than that for the Arctic). Differences in morphology deep in the vortex are, however, minimal. The 2011 Arctic ozone deficit was at least comparable to that in the 2010 Antarctic vortex core at an equivalent time.

In the absence of chemical ozone loss, downward transport during winter results in a springtime maximum in total ozone; because this transport is stronger in the Arctic, background ozone levels there are ~100 DU higher than those in the Antarctic^{7, 23}. Therefore Arctic spring total ozone could, even after chemical destruction comparable to that in an Antarctic ozone hole (commonly defined by values less than 220 DU; refs 7, 12), exhibit only a weak maximum in total ozone rather than a well-defined minimum. Examination of the long-term ozone-sonde record in the Arctic shows that abundances near 250 DU or less are well below typical autumn values, thus appearing as a ‘hole’ in total ozone. Dynamical processes can result in transient regions of very low total ozone (Supplementary Discussion, Supplementary Figs 5, 6) and/or local minima in lower-stratospheric ozone profiles (for example, via ozone-poor extra-vortex air transported into the polar vortex^{14, 24}). For an interhemispheric comparison of chemical loss, it is thus important to verify that observed Arctic ozone decreases were primarily related to chemical, rather than dynamical, processes.

Figure 4 shows that the precipitous decline in Arctic ozone in February–March 2011 resulted from chemical loss of similar magnitude to that in the Antarctic in the mid-1980s. Observed ozone between ~15 and 20 km altitude decreased to values matching the minima in early Antarctic ozone holes and those reached at the corresponding time in some recent Antarctic winters (Figs 2j–l; 3f). In late March–early April, most ozone-sonde profiles in the vortex had mixing ratios less than 1 p.p.m.v., with values ~0.7 p.p.m.v. over an approximately 2-km altitude region, and some dipping to 0.5 p.p.m.v. (Supplementary Fig. 7). Minimum total ozone in spring 2011 was continuously below 250 DU for ~27 days (Supplementary Fig. 5b), with a maximal area below that level of ~2 × 10⁶ km² (roughly five times the area of Germany or California). Values dropped to ~220–230 DU for about a week in late March 2011.

In these respects, chemical ozone destruction in the 2011 Arctic polar vortex attained, for the first time, a level clearly identifiable as an Arctic ozone hole. On the other hand, although the magnitude of chemical depletion was comparable to that in the Antarctic, total ozone values remained higher and, because the areal extent of the Arctic vortex was much smaller (~60% the size of a typical Antarctic vortex), the low-ozone region was more confined.

The Arctic winter stratosphere exhibits striking interannual variability. The past decade has included the four most dynamically active (hence among the warmest) Arctic winters in the past 32 years (ref. 35) and now the two coldest winters with largest ozone loss^{7, 12, 13, 14}, extending the previously noted trend of the coldest winters becoming colder^{13, 16}. Had implementation of the Montreal Protocol not curbed the increase in stratospheric halogen loading, formation of an Arctic ozone hole would have already become common even in moderately cold winters³⁶. Even with the lower anthropogenic halogen levels actually reached, the potential for Antarctic-like ozone loss in the Arctic in the event of a persistently cold winter–spring such as that in 2010–11 has been recognized for decades^{5, 22}. Despite temperatures that were generally far higher than those in Antarctic winter, Arctic chemical ozone destruction in 2011 rivalled that in some Antarctic ozone holes. The development of an Arctic ozone hole under conditions only slightly more extreme than those in some previous Arctic winters raises the possibility of yet more severe depletion as lower-stratospheric temperatures decrease. More acute Arctic ozone destruction could exacerbate biological risks from increased ultraviolet radiation exposure, especially if the vortex shifted over densely populated mid-latitudes, as it did in April 2011.

Our present understanding of what drives variability in the Arctic winter stratosphere is incomplete. Stratospheric temperatures and vortex evolution depend on the atmosphere’s radiative properties and propagation of wave activity^{37, 38}, which are being modified by increasing greenhouse gas concentrations. Day-to-day tropospheric disturbances can lead to stratospheric warming or cooling, depending on their geographical location and the stratospheric vortex structure, which controls their upward propagation^{39, 40}. Current climate models do not fully capture either the observed short-timescale patterns of Arctic variability or the full extent of the observed longer-term cooling trend in cold stratospheric winters; nor do they agree on future circulation changes that affect trends in transport^{41, 42}. Our ability to predict when conditions similar to, or more extreme than, those in 2011 may be realized is thus very limited. Improving our predictive capabilities for Arctic ozone loss, especially while anthropogenic halogen levels remain high, is one of the greatest challenges in polar ozone research. Comprehensive stratospheric data sets, such as those used here, are critical to meeting that challenge.

Data sets

Modern Era Retrospective-analysis for Research and Applications (MERRA)⁴³ fields, from the Goddard Earth Observing System Version 5.2.0 (GEOS-5) data assimilation system, are used for the temperature and vortex analysis. The Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) on the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) satellite⁴⁴ provides PSC/aerosol information. CALIOP measurements began in April 2006. Trace gas profile measurements are from the Microwave Limb Sounder (MLS)⁴⁵ on NASA’s Aura satellite, and the predecessor MLS instrument²⁶ on the Upper Atmosphere Research Satellite (UARS). Total column ozone data are from the Dutch-Finnish Ozone Monitoring Instrument (OMI)⁴⁶ on board Aura. The historical total ozone record comprises data from Nimbus-7 and Earth Probe Total Ozone Mapping Spectrometer (TOMS)⁵⁰. Aura MLS and OMI measurements are available from August 2004 through to the present. UARS MLS measurements were obtained from September 1992 through to early 2000, with increasingly sparse sampling in the later years²⁶. TOMS data are available beginning in 1979, but no TOMS instrument was taking measurements during the 1995–96 Arctic winter.

Measurements from the Match network of balloon-borne ozone soundings (ozone sondes)⁴⁷ are used in some of the chemical ozone loss estimates.

Temperature and vortex analysis

Potential vorticity⁴⁹ (PV) is used to define the vortex, with a contour of ‘scaled’ PV of 1.4 × 10^–4 s⁻¹ (in vorticity units) demarking the vortex edge^{51, 52}. Vortex strength is diagnosed as the maximum daily gradient in PV as a function of equivalent latitude (the latitude that would enclose the same area between it and the pole as a given PV contour)^{51, 52, 53}. Scaled PV multiplied by 10⁴ is used in the calculation, resulting in units for its gradient of 10⁻⁴ (s degrees equivalent latitude)⁻¹.

The temperature threshold for chlorine activation, T_act, is estimated using the formula for nitric acid trihydrate formation⁵⁴, which depends on pressure, HNO₃ and H₂O. Climatological HNO₃ and H₂O profiles are used, derived from UARS data. The area with T < T_act is calculated on seven isentropic surfaces in the lower stratosphere: 390, 410, 430, 460, 490, 520 and 550 K; T_act on these levels is 197.5, 197.2, 196.8, 196.5, 195.9, 195.3 and 194.5 K, respectively. To get the volume with T < T_act from 380 through 565 K, the areas at each of the seven levels are multiplied by the estimated altitude associated with that layer and summed. The altitude range associated with each layer is obtained from a standard potential temperature profile as a function of altitude derived from high latitude temperature soundings taken during the 1988–89 through to 2001–02 winters (the same profile was used for V_psc calculations in refs 13, 16 and 48). These thicknesses are 1.29088, 1.19995, 1.36770, 1.46281, 1.30554, 1.18199 and 1.07382 km for the seven levels listed above. Vortex volume is calculated from vortex area in the same manner. Winter mean V_psc is calculated over 16 December through to 15 April. Previous studies have shown that V_psc scaled by the vortex area is a good proxy for chlorine activation and ozone loss potential¹⁷. Additional temperature and vortex diagnostics are described in Supplementary Information.

Polar stratospheric cloud and aerosol information

Particulate backscatter averaged over the polar vortex derived from CALIOP data is used to provide PSC/aerosol information. Total attenuated backscatter at 532 nm, b(z), is one of the basic CALIOP Level 1B data products. b(z) is the sum of the particulate backscatter (due to liquid aerosol and PSCs), b_p(z), and molecular backscatter, b_m(z). b_m(z) is calculated using GEOS-5 molecular density profiles (included in the CALIOP Level 1B data files) and a theoretical value for the molecular scattering cross-section⁵⁵. Profiles of b_p(z) are then produced by subtracting b_m(z) from b(z). Vortex-averaged profiles of b_p(z) are produced by averaging all CALIOP b_p(z) profiles located inside the vortex edge (defined using information available in GEOS-5 Derived Meteorological Product (DMP) files for the nearly-coincident Aura MLS data⁵²) over the selected time interval.

MLS trace gas profile measurements and analysis

Trace gas profile measurements of HNO₃, HCl, ClO, ozone and N₂O (a long-lived tracer used to assess descent) are from Aura MLS⁴⁵ version 3 retrievals; data quality screening is as recommended in the MLS data quality document⁵⁶. MLS data are retrieved on pressure surfaces; potential temperature as a function of pressure from MLS DMPs⁵² calculated from GEOS-5 analyses is used to interpolate to isentropic surfaces. Vortex averages of MLS data are calculated using the 1.4 × 10⁻⁴ s⁻¹ scaled PV contour to define the vortex edge, using PV values from the MLS DMPs⁵². Active chlorine is in the form of ClO mainly during the daytime, and thus measured ClO amounts vary with the solar zenith angle (SZA) at which the measurements are taken. Only daytime ClO measurements are used here. Northern high latitudes are sampled near midday local time, southern high latitudes are sampled in late afternoon, thus the SZA of Aura MLS Antarctic measurements is ~7° higher on average than that in the Arctic. Reactive chlorine partitioning shifts away from ClO at higher SZAs^{7, 12}, leading to ~30% lower ClO measured by Aura MLS in the Antarctic than in the Arctic under fully activated conditions. MLS measurements are unavailable from 27 March through to 20 April 2011 because of an instrument anomaly. Upper Atmosphere Research Satellite (UARS) MLS measurements, used for analysis of 1995–96 and 1996–97, are sparse because of the UARS yaw cycle and other measurement gaps²⁶. The time of day of UARS measurements varied through the yaw cycle, in the middle of which no daytime ClO measurements were obtained¹⁰; thus ClO values shown in 1995–96 and 1996–97 near those dates (including the mid-February 1996 measurements shown in Fig. 2g) are not representative of the degree of chlorine activation.

Chemical loss calculations

Chemical ozone loss is quantified by two methods, both widely used for such calculations^{7, 12, 24, 25, 26, 27, 28, 47, 48}. In the ‘passive subtraction’ method^{25, 26, 27}, a transport model is used to calculate the evolution of ozone in the absence of chemical changes (‘passive’ ozone). The difference between passive ozone and observed ozone provides an estimate of chemical loss.

Here, passive ozone is obtained in three different ways. First, MLS observations of N₂O, a long-lived species unaffected by chemical processes, are used to calculate vertical motion, and that estimate of descent is then used to calculate how initial MLS ozone profiles would have evolved in the absence of chemical loss¹⁴. Second, a ‘reverse trajectory’ transport model^{25, 26} is used to transport an initial state based on MLS-observed ozone with no chemistry. Finally, the ATLAS (Alfred Wegener Institute Lagrangian Chemistry/Transport System) chemistry and transport model is run in passive mode²⁸, initialized with MLS ozone.

Vortex ozone is also examined in relation to the surfaces on which it is subsiding^{12, 14, 28, 48}. The descent rates used here are obtained by averaging radiative heating/cooling rates from the radiation calculation used in the ATLAS model over the polar vortex⁴⁸. These rates are then used to examine vortex-averaged MLS and ozone-sonde data on surfaces of ‘spring equivalent potential temperature’⁴⁸, defined as the potential temperature at which air originating at a given level arrived at the end of March. Since the air descended on these surfaces, ozone would have been constant on each such surface in the absence of chemical loss.

The ozone-sonde data used here are all from electrochemical concentration cell (ECC) sondes, made by different manufacturers. Ozone-sonde data quality was assessed in an intercomparison experiment⁵⁷ and is discussed in ref. 47. For chemical loss calculations using ozone-sonde data, the profiles are first examined using a procedure for detecting lamination in the profiles; such lamination (an example is shown in Fig. 3f) is associated with mixing in of extra-vortex air, which may obscure the signature of chemical loss. Profiles that have been significantly altered by mixing processes, as indicated by lamination, are excluded from the vortex averages used in the chemical loss calculations. 2010–11 Arctic ozone-sonde data are provided as Supplementary Information.

Results from the ATLAS model passive subtraction calculations, and from the calculations on spring equivalent potential temperature surfaces using the Match network ozone-sonde data, are shown in Fig. 4; all panels show vortex averages. These results have been compared with the results from the other methods described above. While absolute ozone values obtained from different methods/data sets vary significantly (up to ~0.4 p.p.m.v. at the end of March 2011), the year-to-year variations in chemical loss calculated using all three methods agree closely, indicating a high degree of precision in the relative amount of calculated loss between different years.

The Alfred Wegener Institute chemical box model, also used as the chemical module in ATLAS, simulates 175 reactions between 48 chemical species in the stratosphere^{27, 58} This model was used to perform conceptual runs (Supplementary Fig. 4), started on 1 March with identical initial mixing ratios of all species except HNO₃ and O₃. For these two species values corresponding to 1997 (3 p.p.m.v. O₃, 10 p.p.b.v. HNO₃) and 2011 (2.2 p.p.m.v. O₃, 6 p.p.b.v. HNO₃) (compare Figs 2a and 4c) were combined to yield four sets of initial conditions. Initial ClO_x was 2 p.p.b.v., corresponding to the vortex-averaged ClO_x derived by ATLAS from MLS ClO measurements on 1 March 2011. An air parcel at 70° N, 460 K potential temperature, with a temperature of 193 K throughout March, was used. Heterogeneous reactions took place on liquid aerosols, rather than solid (nitric acid trihydrate, NAT) PSCs, since the widespread existence of the latter is inconsistent with MLS observations of gas-phase HNO₃ values (Fig. 2a) larger than those the microphysical module predicts if NAT is present. A sensitivity run showed that sporadically occurring solid PSCs did not change the results significantly.

Column ozone and ozone deficit calculation

OMI total ozone data were processed with version 8.5 of the TOMS algorithm and have been extensively validated⁵⁹. TOMS data were processed with version 8 of the algorithm. The OMI and TOMS total ozone data used in this study were averaged on a fixed global 1° × 1° latitude × longitude grid. Averages were computed by area-weighting observations based on the overlap of their instantaneous field-of-view with each grid cell. Only data that satisfy quality criteria based on measurement path length and algorithm diagnostic criteria were included in the averaged samples. Individual total ozone retrievals included in the samples are expected to have a root-mean-squared error of 1–2%.

Total ozone ‘deficit’ is calculated as the difference between daily values and a reference that is minimally affected by chemical ozone loss. The reference for the Arctic is the daily mean over all Arctic winters from 1978–79 through to 2009–10, from OMI starting in 2004–05 and from TOMS for earlier years⁵⁰. The Antarctic reference state is the daily mean of TOMS measurements for 1979 through to 1981. Because the Antarctic reference state is based on only three years’ data for each day, variations in vortex position are not effectively averaged out; this reference is thus less robust than that for the Arctic, so patterns in daily maps may partially reflect differences in vortex position between the reference and the focus day.

1. Farman, J. C., Gardiner, B. G. & Shanklin, J. D. Large losses of total ozone in Antarctica reveal seasonal ClO_x/NO_x interaction. Nature 315, 207–210 (1985)
   • ChemPort
   • ISI
   • Article
2. Solomon, S., Garcia, R. R., Rowland, F. S. & Wuebbles, D. J. On the depletion of Antarctic ozone. Nature 321, 755–758 (1986)
   • ChemPort
   • ISI
   • Article
3. Molina, L. T. & Molina, M. J. Production of Cl₂O₂ from the self-reaction of the ClO radical. J. Phys. Chem. 91, 433–436 (1987)
   • ChemPort
   • ISI
   • Article
4. Anderson, J. G., Brune, W. H. & Proffitt, M. H. Ozone destruction by chlorine radicals within the Antarctic vortex: the spatial and temporal evolution of ClO-O₃ anticorrelation based on in situ ER-2 data. J. Geophys. Res. 94, 11465–11479 (1989)
   • ChemPort
   • Article
5. Solomon, S. Stratospheric ozone depletion: a review of concepts and history. Rev. Geophys. 37, 275–316 (1999)
   • ChemPort
   • ISI
   • Article
6. Schoeberl, M. R. & Hartmann, D. L. The dynamics of the stratospheric polar vortex and its relation to springtime ozone depletions. Science 251, 46–52 (1991)
   • ChemPort
   • ISI
   • PubMed
   • Article
7. World Meteorological Organization. Scientific Assessment of Ozone Depletion: 2010 (Report 52, Global Ozone Research and Monitoring Project, 2011)
8. Solomon, P. M. et al. High concentrations of chlorine monoxide at low altitudes in the Antarctic spring stratosphere: secular variation. Nature 328, 411–413 (1987)
   • ChemPort
   • ISI
   • Article
9. Waters, J. W. et al. Stratospheric ClO and ozone from the Microwave Limb Sounder on the Upper Atmosphere Research Satellite. Nature 362, 597–602 (1993)
   • ChemPort
   • ISI
   • Article
10. Santee, M. L., Manney, G. L., Waters, J. W. & Livesey, N. J. Variations and climatology of ClO in the polar lower stratosphere from UARS Microwave Limb Sounder measurements. J. Geophys. Res.. 108, 4454, http://dx.doi.org/10.1029/2002JD003335 (2003)
    • ChemPort
    • Article
11. Santee, M. L. et al. A study of stratospheric chlorine partitioning based on new satellite measurements and modeling. J. Geophys. Res.. 113, D12307, http://dx.doi.org/10.1029/2007JD009057 (2008)
    • ChemPort
    • Article
12. World Meteorological Organization. Scientific Assessment of Ozone Depletion: 2006 (Report 50, Global Ozone Research and Monitoring Project, 2007)
13. Rex, M. et al. Arctic winter 2005: implications for stratospheric ozone loss and climate change. Geophys. Res. Lett.. 33, L23808, http://dx.doi.org/10.1029/2006GL026731 (2006)
    • ChemPort
    • Article
14. Manney, G. L. et al. EOS MLS observations of ozone loss in the 2004–2005 Arctic winter. Geophys. Res. Lett.. 33, L04802, http://dx.doi.org/10.1029/2005GL024494 (2006)
    • ChemPort
    • Article
15. Harris, N. R. P., Lehmann, R., Rex, M. & von der Gathen, P. A closer look at Arctic ozone loss and polar stratospheric clouds. Atmos. Chem. Phys. 10, 8499–8510 (2010)
    • ChemPort
    • ISI
    • Article
16. Rex, M. et al. Arctic ozone loss and climate change. Geophys. Res. Lett.. 31, L04116, http://dx.doi.org/10.1029/2003GL018844 (2004)
    • ChemPort
    • Article
17. Tilmes, S., Müller, R., Engel, A., Rex, M. & Russell, J. M., III Chemical ozone loss in the Arctic and Antarctic stratosphere between 1992 and 2005. Geophys. Res. Lett.. 33, L20812, http://dx.doi.org/10.1029/2006GL026925 (2006)
    • ChemPort
    • Article
18. Poole, L. R. & Pitts, M. C. Polar stratospheric cloud climatology based on Stratospheric Aerosol Measurement II observations from 1978 to 1989. J. Geophys. Res. 99, 13083–13089 (1994)
    • Article
19. Fromm, M. D. et al. An analysis of Polar Ozone and Aerosol Measurement (POAM) II Arctic stratospheric cloud observations, 1993–1996. J. Geophys. Res. 104, 24341–24357 (1999)
    • ChemPort
    • Article
20. Pitts, M. C., Poole, L. R. & Thomason, L. W. CALIPSO polar stratospheric cloud observations: second-generation detection algorithm and composition discrimination. Atmos. Chem. Phys. 9, 7577–7589 (2009)
    • ChemPort
    • ISI
    • Article
21. Douglass, A. R. et al. Interhemispheric differences in springtime production of HCl and ClONO₂ in the polar vortices. J. Geophys. Res. 100, 13967–13978 (1995)
    • Article
22. Manney, G. L. et al. Chemical depletion of ozone in the Arctic lower stratosphere during winter 1992–93. Nature 370, 429–434 (1994)
    • ChemPort
    • ISI
    • Article
23. Tegtmeier, S., Rex, M., Wohltmann, I. & Krüger, K. Relative importance of dynamical and chemical contributions to Arctic wintertime ozone. Geophys. Res. Lett. 35 L17801 http://dx.doi.org/10.1029/2008GL034250 (2008)
    • ChemPort
    • Article
24. Rex, M. et al. In situ measurements of stratospheric ozone depletion rates in the Arctic winter of 1991/1992: a Lagrangian approach. J. Geophys. Res. 103, 5843–5853 (1998)
    • ChemPort
    • Article
25. Manney, G. L. et al. Lagrangian transport calculations using UARS data. Part II: ozone. J. Atmos. Sci. 52, 3069–3081 (1995)
    • ISI
    • Article
26. Manney, G. L. et al. Variability of ozone loss during Arctic winter (1991–2000) estimated from UARS Microwave Limb Sounder measurements. J. Geophys. Res.. 108, 4149, http://dx.doi.org/10.1029/2002JD002634 (2003)
    • ChemPort
    • Article
27. Wohltmann, I., Lehmann, R. & Rex, M. The Lagrangian chemistry and transport model ATLAS: simulation and validation of stratospheric chemistry and ozone loss in the winter 1999/2000. Geosci. Model Dev. 3, 585–601 (2010)
    • ISI
    • Article
28. von der Gathen, P. et al. Observational evidence for chemical ozone depletion over the Arctic in winter 1991–92. Nature 375, 131–134 (1995)
    • ChemPort
    • Article
29. Gernandt, H. The vertical ozone distribution above the GDR-research base, Antarctica in 1985. Geophys. Res. Lett. 14, 84–86 (1987)
    • ChemPort
    • ISI
    • Article
30. Rex, M. et al. Prolonged stratospheric ozone loss in the 1995–96 Arctic winter. Nature 389, 835–838 (1997)
    • ChemPort
    • ISI
    • Article
31. Manney, G. L., Froidevaux, L., Santee, M. L., Zurek, R. W. & Waters, J. W. MLS observations of Arctic ozone loss in 1996–97. Geophys. Res. Lett. 24, 2697–2700 (1997)
    • ChemPort
    • ISI
    • Article
32. Manney, G. L., Santee, M. L., Froidevaux, L., Waters, J. W. & Zurek, R. W. Polar vortex conditions during the 1995–96 Arctic winter: meteorology and MLS ozone. Geophys. Res. Lett. 23, 3203–3206 (1996)
    • ChemPort
    • ISI
    • Article
33. Petzoldt, K. The role of dynamics in total ozone deviations from their long-term mean over the Northern Hemisphere. Ann. Geophys. 17, 231–241 (1999)
    • ChemPort
    • Article
34. Hood, L. L., Soukharev, B. E., Fromm, M. & McCormack, J. P. Origin of extreme ozone minima at middle to high northern latitudes. J. Geophys. Res. 106, 20925–20940 (2001)
    • ChemPort
    • Article
35. Manney, G. L. et al. Aura Microwave Limb Sounder observations of dynamics and transport during the record-breaking 2009 Arctic stratospheric major warming. Geophys. Res. Lett.. 36, L12815, http://dx.doi.org/10.1029/2009GL038586 (2009)
    • ChemPort
    • Article
36. Newman, P. A. et al. What would have happened to the ozone layer if chlorofluorocarbons (CFCs) had not been regulated? Atmos. Chem. Phys. 9, 2113–2128 (2009)
    • ChemPort
    • ISI
    • Article
37. Newman, P. A., Nash, E. R. & Rosenfield, J. E. What controls the temperatures of the Arctic stratosphere during the spring? J. Geophys. Res. 106, 19999–20010 (2001)
    • ISI
    • Article
38. Polvani, L. M. & Saravanan, R. The three-dimensional structure of breaking Rossby waves in the polar wintertime stratosphere. J. Atmos. Sci. 57, 3663–3685 (2000)
    • ISI
    • Article
39. Orsolini, Y. J., Karpechko, A. Y. & Nikulin, G. Variability of the Northern Hemisphere polar stratospheric cloud potential: the role of North Pacific disturbances. Q. J. R. Meteorol. Soc. 135, 1020–1029 (2009)
    • ISI
    • Article
40. Woollings, T., Charlton-Perez, A., Ineson, S., Marshall, G. & Masato, G. Associations between stratospheric variability and tropospheric blocking. J. Geophys. Res.. 115, D06108, http://dx.doi.org/10.1029/2009JD012742 (2010)
    • Article
41. Butchart, N. et al. Multimodel climate and variability of the stratosphere. J. Geophys. Res.. 116, D05102, http://dx.doi.org/10.1029/2010JD014995 (2011)
    • Article
42. Charlton-Perez, A. et al. The potential to narrow uncertainty in projections of stratospheric ozone over the 21st century. Atmos. Chem. Phys. 10, 9473–9486 (2010)
    • ChemPort
    • ISI
    • Article
43. Reinecker, M. M. et al. MERRA — NASA’s modern-era retrospective analysis for research and applications. J. Clim. 24 3624–3648 http://dx.doi.org/10.1175/JCLI-D-11-00015.1 (2011)
    • Article
44. Hunt, W. H. et al. CALIPSO lidar description and performance assessment. J. Atmos. Ocean. Technol. 26, 1214–1228 (2009)
    • ISI
    • Article
45. Waters, J. W. et al. The Earth Observing System Microwave Limb Sounder (EOS MLS) on the Aura satellite. IEEE Trans. Geosci. Rem. Sens. 44, 1075–1092 (2006)
    • Article
46. Levelt, P. F. et al. The Ozone Monitoring Instrument. IEEE Trans. Geosci. Rem. Sens. 44, 1093–1101 (2006)
    • Article
47. Rex, M. et al. Chemical ozone loss in the Arctic winter 1994/95 as determined by the Match technique. J. Atmos. Chem. 32, 35–59 (1999)
    • ChemPort
    • ISI
    • Article
48. Rex, M. et al. Chemical depletion of Arctic ozone in winter 1999/2000. J. Geophys. Res.. 107, 8276, http://dx.doi.org/10.1029/2001JD000533 (2002)
    • ChemPort
    • Article
49. Hoskins, B. J., McIntyre, M. E. & Robertson, A. W. On the use and significance of isentropic potential-vorticity maps. Q. J. R. Meteorol. Soc. 111, 877–946 (1985)
    • ISI
    • Article
50. McPeters, R. D. et al. Earth Probe Total Ozone Mapping Spectrometer (TOMS) Data Products User’s Guide (NASA Technical Publication 1998-206895, 1998)
51. Manney, G. L., Zurek, R. W., Gelman, M. E., Miller, A. J. & Nagatani, R. The anomalous Arctic lower stratospheric polar vortex of 1992–1993. Geophys. Res. Lett. 21, 2405–2408 (1994)
    • ISI
    • Article
52. Manney, G. L. et al. Solar occultation satellite data and derived meteorological products: Sampling issues and comparisons with Aura MLS. J. Geophys. Res.. 112, D24S50, http://dx.doi.org/10.1029/2007JD008709 (2007)
    • ChemPort
    • Article
53. Butchart, N. & Remsberg, E. E. The area of the stratospheric polar vortex as a diagnostic for tracer transport on an isentropic surface. J. Atmos. Sci. 43, 1319–1339 (1986)
    • ISI
    • Article
54. Hanson, D. & Mauersberger, K. Laboratory studies of the nitric acid trihydrate: implications for the south polar stratosphere. Geophys. Res. Lett. 15, 855–858 (1988)
    • ChemPort
    • ISI
    • Article
55. Hostetler, C. A. et al. CALIOP algorithm theoretical basis document. Calibration and Level 1 data products (Technical Report, NASA Langley Research Center, 2006); available at http://www-calipso.larc.nasa.gov/resources/pdfs/PC-SCI-201v1.0.pdf.
56. Livesey, N. J. et al. Version 3.3 Level 2 data quality and description document. (Technical Report JPL D-33509, Jet Propulsion Laboratory, 2010); available at http://mls.jpl.nasa.gov/data/v3-3_data_quality_document.pdf.
57. Smit, H. G. et al. Assessment of the performance of ECC-ozonesondes under quasi-flight conditions in the environmental simulation chamber: insights from the Jülich Ozone Sonde Intercomparison Experiment (JOSIE). J. Geophys. Res.. 112, D19306, http://dx.doi.org/10.1029/2006JD007308 (2007)
    • ChemPort
    • Article
58. Krämer, M. et al. Intercomparison of stratospheric chemistry models under polar vortex conditions. J. Atmos. Chem. 45, 51–77 (2003)
    • ISI
    • Article
59. McPeters, R. et al. Validation of the Aura Ozone Monitoring Instrument total column ozone product. J. Geophys. Res.. 113, D15S14, http://dx.doi.org/10.1029/2007JD008802 (2008)
    • ChemPort
    • Article

Download references

We thank the MLS (especially A. Lambert, D. Miller, W. Read, M. Schwartz, P. Stek, J. Waters), OMI (especially P. K. Bhartia, G. Jaross, G. Labow), CALIPSO and Match science teams, as well as A. Douglass, J. Joiner and the Aura project, for their support. We also thank W. Daffer and R. Fuller for programming assistance at JPL; the many observers whose work went into obtaining the ozone-sonde measurements; the ozone scientists who participated in the discussion of the 2011 Arctic ozone loss and appropriate definition of an Arctic ozone hole (including, but not limited to, N. Harris, G. Bodeker, G. Braathen, M. Kurylo, R. Salawitch); and especially P. Newman and K. Minschwaner for discussions and comments. Meteorological analyses were provided by NASA’s Global Modeling and Assimilation Office (GMAO) and by the European Centre for Medium-Range Weather Forecasts. We thank S. Pawson of GMAO for advice on usage of the MERRA reanalysis. Ozone-sonde measurements at Alert, Eureka, Resolute Bay, Churchill and Goose Bay were funded by Environment Canada. Additional ozone sondes were flown at Eureka as part of the Canadian Arctic Atmospheric Chemistry Experiment (ACE) Validation Campaign and were funded by the Canadian Space Agency. Academy of Finland provided partial funding for performing and processing ozone-sonde measurements in Jokioinen and Sodankylä. Ozone soundings and work at AWI were partially funded by the EC DG Research through the RECONCILE project. Work at the Jet Propulsion Laboratory, California Institute of Technology, and at Science Systems and Applications Inc., was done under contract with NASA.

CALIOP data are publicly available at http://eosweb.larc.nasa.gov/PRODOCS/calipso/table_calipso.html, MLS data at http://disc.sci.gsfc.nasa.gov/Aura/data-holdings/MLS, OMI data at http://disc.sci.gsfc.nasa.gov/Aura/data-holdings/OMI/omto3_v003.shtml, and GEOS-5 MERRA analyses through http://disc.sci.gsfc.nasa.gov/mdisc/data-holdings/merra/. The balloon-borne Antarctic ozone-sonde data recorded in 1985 and the following years are publicly available at http://dx.doi.org/10.1594/PANGAEA.547983.

Subscribe to comments