Record High March 2024 Arctic Total Column Ozone

Abstract

Observations of March 2024 Arctic (63°N–90°N) total column ozone set a record high of 477 Dobson Units (DU) against the 1979–2023 satellite era time series. It was about 60 DU higher than average and 6 DU higher than the previous March 1979 471 DU record. Daily Arctic ozone was above average for every day in March 2024, and set record highs from 11–26 March 2024. Microwave Limb Sounder data show this record ozone anomaly was concentrated in the lower stratosphere (10–30 km). These record values developed over the 2023–2024 winter and can be associated with vertically propagating planetary-scale wave events that caused significant stratospheric warmings. These wave events forced poleward and downward ozone advection into the lower stratosphere, leading to record column ozone levels. The above average levels persisted through August 2024 and across the northern hemisphere.

Key Points

• Arctic total column ozone in March 2024 set a record high for the 1979-present period

• Polar lower stratosphere temperatures also set a record high in March 2024 in the MERRA-2 reanalysis data

• A record amount of Rossby waves propagating upward from the troposphere caused the record total ozone and lower stratospheric temperature

Plain Language Summary

Man-made chlorofluorocarbons (CFCs) depleted the Earth ozone layer. The 1987 Montreal Protocol curbed CFC growth, but because CFCs have multi-decadal lifetimes, Arctic ozone is not expected to recover back to 1980 levels until ∼2045. Current high CFC levels combined with persistent and cold polar vortices led to severe Arctic ozone spring depletion in 1997, 2011, and 2020. Contrary to expectations, March 2024 Arctic ozone showed a record high level, dramatically contrasting against the severe depletion events. Meteorological and ozone profile information show that the exceptional 2024 ozone was mainly found in the lowermost Arctic stratosphere, in association with record high lowermost stratospheric temperatures. The ozone levels incrementally increased during the 2023–2024 winter because of large-scale weather systems that propagated from the troposphere into the stratosphere. Collectively, these weather systems also were at a record level, moving higher ozone concentrations from the mid-latitudes and upper stratospheric into the Arctic region. This record high ozone would likely have not occurred if CFC levels had not begun slowly declining in response to the Montreal Protocol. Given the absence of high Arctic ozone since the 1970s, the March 2024 record high should be considered a positive harbinger of the future Arctic ozone layer.

1 Introduction

Evidence for Arctic ozone depletion driven by man-made ozone depleting substances (ODSs) is overwhelming (WMO, 2023). Early studies showed large declines of Arctic (63°N–90°N) total column ozone in March satellite averages during the 1980–1990 period (e.g., Gleason et al., 1993). Airborne campaigns and satellite observations revealed that these Arctic total ozone declines resulted from high levels of stratospheric chlorine and bromine that directly depleted lower stratospheric ozone. In response to this ozone depletion, the 1987 Montreal Protocol (MP) controlled ODS consumption and production. Stratospheric levels of inorganic chlorine and bromine from these ODSs stopped increasing in the early 2000s, and are now on a slow decline (Laube & Tegtmeier, 2023).

In spite of the slow ODS decline, Arctic stratospheric chlorine and bromine levels have remained relatively high because ODSs have multi-decadal lifetimes. These high ODS levels have enabled some extreme Arctic ozone depletion events (e.g., 1997, 2011, and 2020) during persistently stable winter polar vortex conditions, resulting in cold polar lower stratospheric temperatures that are conducive to high levels of reactive chlorine to cause ozone depletion (WMO, 2023). In particular, the 2020 March Arctic total ozone column was exceptionally low because of the cold and persistent vortex, surpassing the previous record-breaking spring 2011 depletion event (WMO, 2023). The 1980–2010 Arctic ozone downward trend and the large depletion events of 1997, 2011, and 2020 were generally consistent with ODS growth into the mid-1990s and the subsequent slow decline of ODSs since the early-2000s.

Coupled chemistry and climate models (CCMs) project a steady increase of Arctic column ozone because of both the declining levels of ODSs and increasing levels of greenhouse gases (GHGs). WMO (2023) estimated that Arctic column ozone would return to 1980 levels in ∼2045 (much earlier than the 2066 recovery of the Antarctic ozone hole). It was also noted that dynamical changes in these models were associated with increasing GHGs, and this leads to an earlier Arctic recovery to 1980 values (e.g., Dhomse et al., 2018; Morgenstern et al., 2018). However, because of large year-to-year variability during the northern hemisphere (NH) winter, an identification of increasing Arctic ozone is not apparent. On the contrary, mid-latitude ozone trend studies have suggested ozone continues to decline as a result of dynamically forced transport changes (Ball et al., 2019; Bognar et al., 2022).

In this paper, we document the extreme high levels of ozone in the 2024 Arctic spring stratosphere. Research emphasis on the very large Arctic depletions of 1997, 2011 and 2020 was warranted because they resulted in increased surface UV, causing harmful human and biological effects. High Arctic ozone levels have not received similar attention because these highs decrease UV. Section 2 details the data used herein: total column ozone from a long series of NASA/NOAA backscatter ultraviolet satellite instruments, Microwave Limb Sounder (MLS) data from the NASA Aura satellite, and MERRA-2 meteorological data from the NASA Goddard Global Modeling and Assimilation Office. Section 3 shows ozone and meteorological analysis results. These results show the multi-decadal evolution of Arctic ozone and the vertical structure of the 2024 Arctic ozone distribution. Section 4 provides information on factors influencing polar ozone and wave forcing, while Section 5 summarizes this work and discusses the implications.

2 Data

Total column ozone: Total ozone data are from the daily global gridded measurements obtained from a succession of UV-backscatter mapper satellite instruments: (1979–1983) Nimbus-7 Total Ozone Mapping Spectrometer (TOMS) version 8 (TOMS Science Team, 1993); (1993–1994) Meteor-3 TOMS version 8 (TOMS Science Team, 1996); (1996–2004) Earth Probe TOMS version 8C (TOMS Science Team, 2006); (2004–2016) Aura Ozone Monitoring Instrument (OMI) enhanced version 8 (Bhartia, 2012); and (2016–2024) the Ozone Mapping and Profiler Suite (OMPS) Nadir Mapper version 2.1 (Jaross, 2017).

An area-weighted average was taken over the polar cap (63°N to 90°N), with missing data (e.g., from areas in the polar night) filled in using GEOS FP assimilation (2013–2015 and 2018–2024), MERRA reanalysis (1979–2013), or MERRA-2 reanalysis (2016–2017), interpolated in space and time to missing grid points at local noon. Comparisons of total columns from OMPS Nadir Mapper and Aura OMI with a network of ground-based Brewer and Dobson instruments have shown that zonal average biases are within 0.5% (McPeters et al., 2008; McPeters et al., 2019). Aura OMI total column has also shown an upward trend of ∼+2 DU/per decade that increased over time particularly after 2016 (Frith et al., 2023). In this study we limit Aura OMI to the 2004–2016 period. Hereafter, we refer to this TOMS/OMI/OMPS data simply as mapper polar ozone.

Merged Ozone Data (MOD): The Solar Backscatter Ultraviolet (SBUV) MOD version 8.7 data is comprised of a series of backscatter UV nadir profiler instruments that are cross calibrated on radiance level and processed with the same retrieval algorithm, yielding very stable column ozone decadal records (Labow et al., 2013; Frith et al., 2014, 2020). These nadir profilers do not have scanning capability, so March scans only extend to ∼75°N–80°N. Despite sampling differences (polar cap data poleward of 65°N are not filled to the pole), the MOD variability compares very well to the TOMS/OMI/OMPS mapper data, with an RMS difference of 4.6 DU between the data sets.

Microwave Limb Sounder (MLS): MLS on board the NASA Aura satellite provides measurements of vertical ozone distribution with a vertical resolution of about 2.5–3.5 km in the stratosphere (Livesey et al., 2022; Schwartz et al., 2021). Herein we use the MLS version 5 data, while filtering retrievals according to the MLS Science Team recommendations (Livesey et al., 2022). We convert mixing ratio profiles into ozone partial columns to a regular geopotential height grid (DU/km) using MERRA-2 temperature and pressure profiles. We calculate stratospheric ozone columns by integrating ozone profiles between 316 hPa and 0.1 hPa. The MLS August 2004 to November 2023 values are used to calculate a climatology and statistics. Anomalous MLS ozone retrievals occurred from February 27 to March 6, which we exclude from our analysis. We backfilled by using MERRA-2 vertical ozone distribution with a simple regression model for stratospheric columns.

Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2): MERRA-2 is an ongoing long-term reanalysis produced by NASA Goddard's Global Modeling and Assimilation Office (GMAO). This is generated (starting in 1980) using a consistent data assimilation system (Gelaro et al., 2017). The polar analysis was improved over the original MERRA by using a cubed-sphere model grid, eliminating the polar singularity. Comparisons of zonal and annual mean temperature and winds with another reanalysis (Bosilovich et al., 2015, their Figures 3.2, 3.5, 3.8) shows biases less than 0.4 m/s, 0.3 K, and 0.1 m/s (zonal wind, temperature, and meridional wind respectively) in the polar lower stratosphere. Herein we use MERRA-2, 3-hourly output interpolated to constant pressure levels (GMAO, 2015).

3 Results

March 2024 Arctic mapper polar ozone exceeded the highest value observed since 1979. The TOMS/OMI/OMPS time series of polar cap mapper total ozone values was averaged for March over the polar cap (63°N–90°N) from 1979 through 2024. Figure 1a (black dots and line) shows this time series, with the horizontal dashed gray line showing the 2024 maximum of 477 Dobson Units (DU). This 2024 March value was 60 DU (14.5%) higher than the 1979–2023 average of 416 DU and was 6 DU higher than the previous 1979 record of 471 DU. March polar cap averages from the Nimbus-4 satellite backscatter ultraviolet instrument (BUV, 1970–1976, not shown herein) also showed large ozone values greater than 450 DU, but limited spatial coverage and lack of solid calibration precludes comparisons. Nevertheless, this March 2024 total ozone was also higher than the March 1970-1976 polar cap BUV values.

Figure 1

March (a) Total column ozone (polar cap averaged 63°N–90°N) from (TOMS/OMI/OMPS, black line, and points), Aura MLS satellite instrument partial columns integrated between 316 hPa and 0.1 hPa (red points, averaged 65°N–82°N), and SBUV MOD data (blue points, averaged 65°N–80°N). An additional 25 DU has been added to MLS to account for tropospheric ozone. (b) Temperature ($$\overline{T}$$, polar cap 60°N–90°N, 150 hPa). (c) Zonal mean wind ($$\overline{u}$$, 45°N–75°N, 70 hPa) - note this zonal wind y-axis is inverted. (d) Nov. 15 to March 15 (entire winter) averaged eddy heat flux ($$\overline{v{\rhook}T{\rhook}}$$, 45°N–75°N, 100 hPa) for each year. The horizontal dashed lines highlight the 2024 values.

MLS vertically integrated ozone (316–0.1 hPa) also showed a 2024 record for its 2005–2024 time series (Figure 1a, red points). MLS data are averaged over a slightly different spatial domain (65°N–82°N) than the mapper ozone, and the MLS data have not been filled. Nevertheless, the 2005–2024 MLS averages are correlated (r = 0.996) with the polar mapper ozone. There is an RMS difference of 6 DU, with the MLS stratospheric column biased 25 DU lower than the total column from the polar mappers (likely due to tropospheric ozone not included in the MLS averages). To represent the March value uncertainty for the 1979–2024 Arctic ozone columns we use a standard deviation of 6 DU, which captures both the SBUV MOD and MLS variability. The 6 DU difference between the 2024 average and the previous 1979 record suggests that there is a 76% chance that 2024 is higher than 1979. It is likely that March 2024 is the new record for Arctic ozone.

The March 2024 MOD data also show a record high of 480 DU. We compared the SBUV MOD record (Figure 1a, blue points) with the mapper Arctic ozone (black points and line). The MOD data is highly correlated with TOMS/OMI/OMPS data (r = 0.992), with an 8.1 DU bias, and an RMS difference of 4.6 DU. The 2024 March MOD average exceeds the previous 1979 high by 8.1 DU. Given the long-term accuracy of the MOD data, it is highly likely that 2024 was a record high for the 1979-2024 March period.

In addition to the March averaged record, daily values of Arctic averaged ozone had record levels over most of the month. Figure 2a shows daily column ozone values for December 2023 to March 2024 (blue line), the 1979–2023 averaged daily mean (black line), while shaded values show the single-day minimums and maximums in light gray, the 10% and 90% percentiles in a somewhat darker gray, and the 30% and 70% percentiles in dark gray. MLS daily partial column averages (offset upward by 25 DU) are in red. These MLS column estimates are very consistent with the total ozone data. During March 2024, 17 of 31 days set total ozone Arctic records, beginning on March 9, and ending on March 26. Arctic ozone on March 20 was 499 DU, exceeding the previous March 1980 record by 30 DU. There was a steep Arctic ozone increase between March 3 and March 10 (see magenta vertical line on 3 March). Other steep increases of Arctic ozone occurred around December 21 and 30, January 13, and February 13 (denoted by the magenta vertical lines in Figure 2a).

Figure 2

Daily December through March (a) Total column ozone (TOMS/OMI/OMPS, polar cap averaged 63°N–90°N). The blue line shows 2023–2024 values, while the red line shows the vertically integrated Aura MLS satellite partial column (averaged 65°N–82°N, with an additional 25 DU added to account for tropospheric ozone). The gray bar to the right shows the range of values from the daily 1979–2023 values (light gray shading shows maximum and minimum values). The March blue values above the light gray shading reveal record high Arctic ozone. See Figure 1a for the 2024 March average. (b) Vertical profiles from MLS (units are DU km⁻¹) with the 2004–2023 climatology subtracted. (c) Zonal mean temperature ($$\overline{T}$$, 60°N–90°N, 150 hPa). See Figure 1b for the 2024 March average. (d) Zonal mean zonal wind ($$\overline{u}$$, 45°N–75°N, 10 hPa). Note this zonal wind y-axis is inverted. (e) Eddy heat flux ($$\overline{v{\rhook}T{\rhook}}$$, 45°N–75°N, 100 hPa). See Figure 1d for the 15 Nov. to 15 March (entire winter) average. The vertical magenta lines are set onto the peaks of the eddy heat flux events on 12/21/2023, 12/30/2023, 1/13/2024, 2/13/2024, and 3/03/2024, while the background shaded pink highlights the duration of these eddy events ($$\overline{v{\rhook}T{\rhook}} > 20\,K\,m{s}^{-1}$$). The eddy heat flux values have been temporally smoothed with a Gaussian low-pass filter (½ amplitude = 7.7 days).

The anomalous March 2024 excess ozone was mainly found in the mid-to-lower stratosphere. Figure 2b shows the 2024 MLS vertical profile deviations from the daily ozone climatology derived from MLS data averaged over August 2004 to November 2023 (contour increments are 1 DU km⁻¹). Excess ozone during March 2024 is mainly found between 200 and 30 hPa, with a peak value of 6 DU km⁻¹ in mid-March. On isentropic surfaces, this excess ozone is found between the 380 K isentropic level and the tropopause (lowermost stratosphere). Because the lifetime of stratospheric ozone is quite long in the lowermost stratosphere (small photochemical production), the source of this excess ozone is primarily caused by ozone transport (not shown herein). Further, the long lifetime of ozone has resulted in persistent high ozone values into the 2024 summer.

In addition to the extreme Arctic March ozone, Arctic temperatures in the lowermost stratosphere were also at record high values in the MERRA-2 reanalyses. Daily Arctic averaged (60°N–90°N) temperatures are shown in Figure 2c for the 150 hPa level. Again, most of these 2024 average MERRA-2 temperatures (blue line) are at record values (23 of 31 days beginning on March 7 and persisting to the end of the month). The monthly mean Arctic averaged temperatures for the 150 hPa level (Figure 1b, 224.3 K horizontal dotted line) is 6.4 K above the 1979–2023 average, and 2.3 K above the previous record in 2009. Daily temperatures (Figure 2c) in the lowermost stratosphere also underwent sharp increases, coherent with the ozone increases in Figure 2a.

The March mid-stratosphere polar night jet (polar vortex) was very weak, but not a record low. Figure 1c shows the 70 hPa (lower stratosphere) zonal mean zonal wind averaged over the 45°N–75°N (inverted y-axis to show the strong correlations with Figure 1b polar cap temperature and Figure 1a total ozone). Previous years (e.g., 1984 and 1999) also had weak polar night jets, and these are also associated with warmer Arctic lower stratospheric temperatures. Warm March periods (lower-to-mid stratosphere) are related to weak polar night jets (mid-stratosphere), theoretically consistent via the thermal wind equation ($$\boldsymbol{f}\frac{\partial \overline{\boldsymbol{u}}}{\partial \boldsymbol{z}}=-\frac{\boldsymbol{R}}{\boldsymbol{H}}$$ $$\frac{\partial \overline{\boldsymbol{T}}}{\partial \boldsymbol{y}}$$). Herein, all mathematical terms shown follow norms shown in Andrews et al. (1987), that is, overbars are zonal means, f is the Coriolis parameter, u is the zonal mean wind, etc. The weak March polar vortices in these years are driven by stratospheric sudden warmings. The 70 hPa zonal winds are highly correlated with the 150 hPa Arctic temperatures (r = −0.94, while the 10 hPa winds and 150 hPa T correlation is −0.63). The 2024 daily winds (Figure 2d) also show sharp wind speed declines (again, note inversion of the y-axis) in association with sudden warmings of the Arctic stratosphere (Figure 2c) and the increases of polar total ozone (Figures 2a and 2b).

The 2024 eddy heat flux ($$\overline{\boldsymbol{v}{\rhook}\boldsymbol{T}{\rhook}}$$) also set a record. Figure 1d shows the 45 year time series of average (November 15 to March 15) eddy heat flux (100 hPa, 45°N–75°N, 1979 not available). $$\overline{\boldsymbol{v}{\rhook}\boldsymbol{T}{\rhook}}$$ is proportional to the Rossby wave vertical group velocity times the wave activity ($${\boldsymbol{C}}_{\boldsymbol{z}}^{\boldsymbol{g}}\overline{{\boldsymbol{A}}^{\mathbf{2}}}$$), providing a quantitative estimate of the wave energy propagating into the stratosphere (Andrews et al., 1987). The vertical propagation of these waves into the stratosphere deposits easterly momentum in the stratosphere, decelerating the polar night jet and forcing a poleward and downward transport circulation response. The waves' transport effects forces advection of ozone from the mid-latitudes into the polar region. This 100 hPa $$\overline{\boldsymbol{v}{\rhook}\boldsymbol{T}{\rhook}}$$ is well correlated with March polar ozone (r = 0.77, after removing the longer-term trend on ozone forced by ODSs). The record high of the 2023–2024 winter eddy heat flux was the primary factor in driving the record March 2024 polar cap ozone.

Vertically propagating eddy wave events increased polar ozone levels, warmed the lower stratosphere, and decelerated the polar vortex on short timescales during the 2024 winter season. Figure 2e shows the daily 100 hPa $$\overline{\boldsymbol{v}{\rhook}\boldsymbol{T}{\rhook}}$$ values (smoothed with a low-pass 7.7 day half-amplitude Gaussian filter that mainly removes day-to-day variability). During the 2023–2024 winter, $$\overline{\boldsymbol{v}{\rhook}\boldsymbol{T}{\rhook}}$$ was above average, except for a short mid-to-late January period. Prominent eddy events were centered on 12 December 2023, 30 December 2023, 13 January 2024, 13 February 2024, and 3 March 2024 (marked by vertical magenta lines on all Figure 2 panels). The peaks of these $$\overline{\boldsymbol{v}{\rhook}\boldsymbol{T}{\rhook}}$$ events are associated with increased column ozone (Figure 2a), build-up of ozone in the lower stratosphere (Figure 2b), polar cap warming (Figure 2c), and deceleration of the stratospheric polar night jet stream (Figure 2d).

4 Analysis: Factors Influencing Polar Ozone and Wave Forcing

Ozone advection: The vertical integral of ozone transport (the sum of mean advection and mixing) set a record high for influx of ozone into the polar cap in 2024. The column ozone in the polar region is impacted by the vertical integrals of a (a) residual circulation mean ozone advection term ($${\left[{{\overline{\boldsymbol{v}}}^{\ast }\boldsymbol{\rho }}_{\mathbf{0}}\,\boldsymbol{cos}\,\boldsymbol{\phi }\overline{\boldsymbol{\chi }}\right]}^{{\boldsymbol{\phi }}_{\mathbf{0}}}$$) and (b) an ozone mixing term ($${\left[\overline{\boldsymbol{v}{\rhook}\boldsymbol{\chi }{\rhook}}+\overline{\boldsymbol{\chi }}{\boldsymbol{\rho }}_{\mathbf{0}}^{-\mathbf{1}}{\left(\frac{{\boldsymbol{\rho }}_{\mathbf{0}}\overline{{\boldsymbol{v}}^{\prime }{\boldsymbol{T}}^{\prime }}}{\boldsymbol{S}}\right)}_{\boldsymbol{z}}\right]}^{{\boldsymbol{\phi }}_{\mathbf{0}}}$$) (Andrews et al., 1987). The mean residual circulation advection ($${\overline{\boldsymbol{v}}}^{\ast }$$ term) is weighted by ozone density ($${\boldsymbol{\rho }}_{\mathbf{0}}\overline{\boldsymbol{\chi }}$$) and this product peaks at ∼30 hPa. The vertical integral of this $${{\boldsymbol{\rho }}_{\mathbf{0}}\overline{\boldsymbol{\chi }}\overline{\boldsymbol{v}}}^{\ast }$$ term set a record high in 2024 averaged over the 15 November 2023–15 March 2024 period (calculated from MERRA-2 reanalysis meteorology and ozone). The ozone mixing term (b) is almost always negative. It decreases polar ozone by mixing low mid-latitude ozone into the polar region. In 2024, this mixing term was near zero - almost a record high value for this typically negative term. As expected, ozone transport is highly correlated with the eddy heat flux (r = 0.64). This correlation is evidence of how vertically propagating Rossby wave energy controls March polar ozone levels, leading to the 2024 March record.

Ozone production and loss: During mid-winter, the lack of sunlight precludes photochemical ozone production in the stratosphere. Ozone loss rates are complicated by the large ozone depletion events that have occurred in the Arctic when polar stratospheric temperatures have been cold (1997, 2011, and 2020). However, in 2024, temperatures were generally too warm to form polar stratospheric clouds in the late-winter, and ozone depletion was very weak.

Eddy heat flux forcing: The eddy heat flux (100 hPa, 45°N–75°N, 11/15–3/15 average) is directly correlated with March polar ozone (r = 0.69). Because of changing ODS levels, we can include effective equivalent stratospheric chlorine (EESC) to account for changing polar ozone depletion over the 1980 to 2024 period (Newman et al., 2004). The two-parameter fit of March ozone to EESC (4 year mean age), and the eddy heat flux increases the correlation with total column ozone ($${\Omega }={\boldsymbol{a}}_{\mathbf{0}}\boldsymbol{E}\boldsymbol{E}\boldsymbol{S}\boldsymbol{C}+{\boldsymbol{a}}_{\mathbf{1}}\overline{\boldsymbol{v}{\rhook}\boldsymbol{T}{\rhook}})$$ to 0.82. In this equation, a₀ = −46 DU ppb⁻¹, and a₁ = 14 DU K⁻¹ m⁻¹ s.

External factors: Because the Quasi-biennial oscillation (QBO) modulates mid-latitude total ozone via its secondary circulation, the QBO can influence polar ozone in winter. During the 2023–2024 winter the easterly QBO phase increased mid-latitude ozone values, possibly contributing to an advective increase of polar ozone. The QBO can also shift the stratospheric zero mean zonal wind line (critical line), modulating planetary scale wave activity (Holton & Tan, 1980). Despite these known QBO factors, the inclusion of a QBO term in the multiple parameter regression of March Arctic column ozone only improved the explained variance by a few per cent. Inclusion of the El Nino/Southern Oscillation (ENSO) with the QBO using the multi-variate ENSO Index Version 2 values for December-January also yielded a minor improvement (correlation increases slightly to 0.83). ENSO and the QBO did not provide significant improvements to the fitting of March polar ozone, particularly for March 2024.

Eurasia snow extent: Large October snow extent over Eurasia has been associated with strong winter eddy heat flux events (Cohen et al., 2007). As noted earlier, strong wave events lead to increased advection of ozone into the polar region. Wang et al. (2024) connect increased March ozone levels with these large Eurasian snow events. We use the MERRA-2 analysis snow coverage parameter averaged in the Siberia region to test this snow-to-Arctic-ozone relationship. The MERRA-2 snow average correlates with the Wang et al. (2024) snow estimate parameter over the 2000–2024 period. However, the 2023 October MERRA-2 snow coverage estimate was not an exceptional value. While Eurasian snow coverage may play a role in March polar ozone levels, the 2023 snow coverage did not seem to have a large impact on March 2024 ozone.

5 Conclusions

The March 2024 Arctic column ozone set a record high of 477 DU for the modern satellite era (1979-present). These high levels developed over the course of the 2023–2024 winter period, culminating in extremely high March levels with multiple single-day records. The March 2024 Arctic column ozone was 60 DU (14.5%) higher than the 1979–2023 average of 416 DU. This record high emerges from the increased ozone concentrations in the lower stratosphere (30–300 hPa). The record ozone results from a series of large-scale wave events that propagated upward from the troposphere into stratosphere. These wave events decelerated the polar night jet - weakening the polar vortex. They also warmed the Arctic stratosphere and induced an increased poleward circulation. This poleward circulation increased ozone advection into the Arctic region, increasing the Arctic column. While these wave events were not particularly extreme, they individually (stepwise) increased Arctic ozone over the course of the 2023–2024 winter, resulting in the record high March ozone levels.

The eddy heat flux and ODS levels only explain about 67% (r = 0.82) of the total column ozone year-to-year variance, as derived from the two-parameter (EESC and $$\overline{v{\rhook}T{\rhook}}$$) fit of the March Arctic column ozone. The extremely high averaged eddy heat flux in 2024 led to an extremely high level of Arctic total ozone and lower-stratospheric Arctic temperature. Based upon the ODS decline in the two-parameter fit and strong 2024 eddy heat flux, we estimate that the average 2024 Arctic column ozone has increased by 24 DU since the late-1990s. The strong 2024 eddy forcing increased ozone by 48 DU, and an additional 11 DU increase is unexplained by our fit.

There remains considerable year-to-year variability that is not explained by dynamical forcings. Better correlations between the eddy heat flux and total ozone could be obtained herein by optimizing averaging intervals in latitude, pressure, and time. In addition, levels of halogens are different for air mixed into the polar region from mid-latitudes versus halogen levels observed in a relatively undisturbed polar vortex with air descended from the upper stratosphere. Further, ozone depletion rates are different for years with very cold temperatures having extensive polar stratospheric cloud coverage that activate reactive chlorine.

CCMs project that Arctic ozone will recover by 2045 (WMO, 2023). Based upon these 3-D CCMs, the March Arctic ozone multi-model average has increased 10–30 DU between 2000 and 2025 (Figure 3 of Dhomse et al., 2018). Of this average 10–30 DU, ∼20 DU comes from the ODS decline, and <10 DU is from the GHG increases, consistent with our two-parameter fit above. The CCM simulations shown in Dhomse et al. (2018) also exhibit considerable interannual variability of ∼20 DU (one sigma). The observed 2024 March ozone was on the upper-edge of these CCM simulations year-to-year variations. These model mean increases from ODSs and GHGs are smaller than the dynamically induced 48 DU increase that we calculate above. Nevertheless, it is highly likely that the decreased ODSs and increased GHGs aided the March 2024 ozone record level.

The record high March Arctic ozone has been felt across the NH during the 2024 summer, since spring ozone anomalies persist through summer (Fioletov & Shepherd, 2003). The NH mid-latitude summer ozone is connected to polar anomalies (Hadjinicolaou et al., 1997; Millard et al., 2003). Daily Arctic record high values (not shown herein) have been observed from April through August in the Arctic. Ozone in the NH mid-latitudes (30°N–60°N) has also been extremely high as a result of the exceptional transport during the 2023–2024 winter. The May 2024 averaged Arctic ozone is also a record high of 417 DU over the 1979–2024 period. Low Arctic ozone anomalies have been connected to high UV in the NH mid-latitudes (Karpechko et al., 2013) and the Arctic (Bernhard et al., 2020). Following P. Newman and McKenzie (2011) we use an average total column ozone of 417 DU and a May average of 388 DU at 70°N to calculate a local-noon clear-sky UV index (May 15) that is 8% below the 1979–2023 average. A similar calculation for May in the NH mid-latitudes yields the UV index to be 5% below the 1979–2023 average. The record 2024 Arctic ozone results in UV levels across the NH that are substantially below average.

Acknowledgments

This work was supported by NASA funding from the Aura satellite project, the Upper Atmosphere Research Program, the Atmospheric Composition Modeling and Analysis Program, the Modeling and Analysis Prediction Program, and the Long-term Measurement of Ozone Program.

Open Research

Data Availability Statement

The total column ozone observations are from a satellite instruments combination (see data section). The ozone column data are archived and freely available with registration at the NASA Goddard Earth Sciences Data and Information Services Center (GES DISC). More specifically: TOMS Nimbus-7 Level 3 (TOMS Science Team, 1993); TOMS Meteor-3 Ozone Level 3 (TOMS Science Team, 1996); TOMS Earth Probe, Ozone Level 3 (TOMS Science Team, 2006); Aura OMI TOMS-like total column ozone Level 3 (Bhartia, 2012); and the NASA/NOAA Suomi OMPS total ozone Level 3 (Jaross, 2017). Aura MLS version 5 (Schwartz et al., 2021) ozone data and the NASA Goddard Global Modeling and Assimilation Office (GMAO) MERRA-2 analysis fields (GMAO, 2015) used to quantify the meteorology in this study are also publicly available from the GES DISC. The SBUV Merged Ozone Data set (Frith et al., 2014, 2020) is publicly accessible at https://acd-ext.gsfc.nasa.gov/Data_services/merged/index.html.