Geophysical Research Letters

Severe 2011 ozone depletion assessed with 11 years of ozone, NO2, and OClO measurements at 80°N

Authors


Abstract

[1] Unusually cold conditions in Arctic winter 2010/11 led to large stratospheric ozone loss. We investigate this with UV-visible measurements made at Eureka, Canada (80.05°N, 86.42°W) from 1999–2011. For 8–22 March 2011, OClO was enhanced, indicating chlorine activation above Eureka. Ozone columns were lower than in any other year in the record, reaching minima of 237 DU and 247 DU in two datasets. The average NO2 column inside the vortex, measured at visible and UV wavelengths, was 46 ± 30% and 45 ± 27% lower in 2011 than the average NO2column from previous years. Ozone column loss was estimated from two ozone datasets, using a modeled passive ozone tracer. For 12–20 March 2011, the average ozone loss was 27% and 29% (99 DU and 108 DU). The largest percent ozone loss in the 11-year record of 47% (250 DU and 251 DU) was observed on 5 April 2011.

1. Introduction

[2] In spring 2011, chemical ozone loss in the Arctic was comparable to that observed over Antarctica for the first time on record [Manney et al., 2011]. This resulted from an unusually prolonged period with a strong, cold polar vortex. Due to these persistent low temperatures, polar stratospheric clouds (PSCs) were observed until mid-March and activated chlorine was observed until late March. This resulted in a record ozone loss [Balis et al., 2011; Manney et al., 2011].

[3] The polar vortex was above the Polar Environment Atmospheric Research Laboratory (PEARL), located at Eureka, Canada (80.05°N, 86.42°W) for a large part of spring 2011. A suite of instruments, operated by the Canadian Network for the Detection of Atmospheric Change (CANDAC), take continuous measurements at PEARL. We present results from four differential optical absorption spectroscopy (DOAS) instruments. DOAS instruments can measure under both clear and cloudy conditions and therefore present a more continuous timeseries of ozone and NO2 than solar tracking Fourier Transform Infrared (FTIR) spectrometers. Furthermore, DOAS instruments can also measure OClO, which is a good qualitative indicator of chlorine activation [Sessler et al., 1995]. However, quantification of ClO from OClO measurements is difficult due to uncertainties in model predictions, particularly under strong chlorine activation [e.g., Oetjen et al., 2011]. Ozone, NO2, and OClO measurements can be combined with stratospheric parameters in order to identify ozone depletion, chlorine activation, and denitrification within the polar vortex [e.g., Tornkvist et al., 2002; Tétard et al., 2009].

2. Datasets

[4] Measurements included in this study were taken by four ground-based DOAS instruments: the PEARL and University of Toronto ground-based spectrometers (GBSs) [Fraser et al., 2009] and two System D'Analyse par Observations Zenithales (SAOZ) instruments [Pommereau and Goutail, 1988]. These instruments are part of the Network for the Detection of Atmospheric Composition Change (NDACC) and their Eureka ozone and NO2 datasets are described in detail by Adams et al. [2012]. Measurements from the two GBS (two SAOZ) instruments are nearly identical and therefore were combined to create a single GBS (SAOZ) dataset.

[5] The GBSs are UV-visible Triax-180 triple-grating spectrometers, built by Instruments S.A. / Jobin Yvon Horiba, with cooled charge-coupled device detectors and a 2° field-of-view. The resolution varies from 0.2–2.5 nm and the wavelength range varies from 320–600 nm depending on the selected grating and target wavelength. GBS ozone and NO2columns were retrieved above Eureka in the spring for 1999–2011, except for 2001 and 2002. The SAOZ instruments are grating spectrometers, which measure in the 270–620 nm range with a 1.0-nm resolution and a 10° field-of-view and record spectra on uncooled 1024-pixel linear diode array detectors. SAOZ instruments took spring-time measurements at Eureka for 2005–2011.

[6] The DOAS (GBS and SAOZ) ozone measurements were analyzed in the 450–550 nm range, using the NDACC guidelines [Hendrick et al., 2011]. NO2partial columns were retrieved for the GBS instruments in two different wavelength regions: 425–450 nm (GBS-vis) and 350–380 nm (GBS-UV), depending on the selected measurement grating. The GBS NO2 partial columns were calculated for 17 km to the top of the atmosphere for the validation of satellite partial column measurements, using AMFs that were set to zero below 17 km [Adams et al., 2012]. SAOZ NO2 total columns were retrieved in the 410–510 nm range. The SAOZ NO2 total columns are qualitatively consistent with the GBS partial columns, but are not presented here because they cover a different altitude range. The SAOZ and GBS ozone and NO2columns have been shown to agree well with other ground-based and satellite measurements [Fraser et al., 2008, 2009; Adams et al., 2012]. OClO differential slant column densities (DSCDs) at solar zenith angle 90° were also retrieved from spring 2007, 2008, and 2011 GBS spectra in the 350–380 nm range. The OClO retrievals are described in the auxiliary material.

[7] Derived meteorological products [Manney et al., 2007] were calculated along the lines-of-sight of the DOAS instruments [Adams et al., 2012] for 1999–2003 using the Met Office analysis and 2004–2011 using the GEOS-5.1.0/GEOS-5.2.0 analysis. Stratospheric temperatures and scaled potential vorticity (sPV) were interpolated to the 490-K potential temperature level (∼ ozone concentration maximum, ∼19 km) and are referred to here as T490K and sPV490K. The inner and outer vortex edges are identified by sPV490K values of 1.6 × 10−4 s−1 and 1.2 × 10−4 s−1, respectively [Manney et al., 2007].

3. Timeseries of Ozone, NO2, and OClO

[8] The 1999–2011 timeseries of ozone, NO2, OClO, T490K, and sPV490K are shown in Figure 1. In 2000, 2005, 2007 and 2011, low ozone columns were measured above Eureka when the polar vortex was overhead. These years are shown in color, while the other measurement years are shown in grey. Low ozone coincides with low NO2, low T490K, and time periods when the instruments are sampling inside the polar vortex.

Figure 1.

Timeseries of measurements and dynamical parameters along the DOAS line-of-sight for 1999–2011 versus day of year. Year 2000 is shown in orange, 2005 in cyan, 2007 in blue, 2011 in magenta, and all other years are shown in grey. (a) Ozone total columns measured by the GBS (closed diamonds) and SAOZ (open squares). (b) NO2partial columns (17 km to top of atmosphere) measured by GBS-vis (closed diamonds) and GBS-UV (open circles). (c) OClO DSCDs measured by the GBS. (d) T490K and (e) sPV490K.

[9] For 23 February to 21 March 2011 (days 54–80), the DOAS instruments sample lower stratospheric air inside the polar vortex. OClO DSCDs of 0.8–2.0 × 1014 mol/cm2 are within the range of previous elevated OClO measurements [e.g., Tornkvist et al., 2002], suggesting chlorine activation. All elevated OClO DSCDs, from 8–22 March 2011 (days 67–81) and 2–5 March 2007 (days 61–64), are measured inside the polar vortex when the high-latitude minimum temperature (calculated byManney et al. [2011], not shown here) is below the threshold for PSC formation (TNAT). High OClO measurements do not always correspond with local T490K < TNAT (Figure 1d), because the time-scale for vortex mixing (∼5–7 days) is smaller than the time-scale for chlorine deactivation (∼weeks). During the period of elevated OClO in 2011, ozone, NO2, and T490Kall reach minima in the 11-year record, with ozone values of 247 DU (237 DU) measured by the GBS (SAOZ) on 18 March (day 77).

[10] After 22 March 2011 (day 81), the instruments primarily sample the lower stratosphere outside the polar vortex. During this period, ozone and NO2increase to levels that are normal in the context of the 11-year data record. On 5 April (day 87) and 28 March (day 95), ozone and NO2 columns and T490K decrease sharply, as the instruments sample air masses inside the vortex. After 5 April (day 95), T490K and NO2increase to maxima in the 11-year dataset. This increase is the subject of a companion study.

4. Dynamical and Chemical Contributions to Low Ozone

[11] As is evident in the DOAS timeseries (Figure 1), 2011 is extremely different from previous years. Ozone, NO2, and OClO measurements taken inside the polar vortex (sPV490K > 1.6 × 10−4 s−1) for days 55–80 (24 February to 19/20 March) were selected to investigate this further. The time-period was limited in order to reduce the impact of seasonal variation on the results. NO2 measurements were scaled to local solar noon using a photochemical model [McLinden et al., 2000] initialized with temperature and ozone from Eureka ozonesonde profiles from the nearest available date.

[12] Figure 2 shows histograms of ozone, NO2, and OClO for 1999–2010 (grey) and 2011 (transparent with thick black outline). In 2011, the mean vortex ozone column measured by the GBS (SAOZ) is 28 ± 13% (32 ± 14%) lower than the mean column from other years, where the error denotes the 1σstatistical uncertainty. Similarly, GBS-vis (GBS-UV) NO2 is 46 ± 30% (45 ± 27%) lower and GBS OClO is three times higher in 2011 than in previous years.

Figure 2.

Histograms of (a) GBS ozone, (b) SAOZ ozone, (c) GBS-vis NO2, (d) GBS-UV NO2, and (e) GBS OClO. Measurements were taken inside the vortex for days 55–80 (24 February to 19/20 March), with 1999–2010 in gray and 2011 transparent with thick black lines. N ± M denotes the average (N) and 1σ standard deviation (M) in the measurements.

[13] The unusual 2011 ozone and NO2columns are a result of both chemistry and transport, which contribute approximately equally to year-to-year total ozone column variability in the Arctic [Tegtmeier et al., 2008]. Figure 3 shows the correlation between vortex ozone/NO2 measurements and the local T490K. Correlation between ozone and local lower stratospheric temperature has been observed in previous studies and points to replenishing of ozone through vertical descent, horizontal mixing across the vortex edge, and adiabatic compression of the column, which all increase with higher stratospheric temperatures (e.g., supplementary material of Manney et al. [2011, and references therein]). In the present study, the strongest correlation between ozone and T490K was calculated when data were excluded from years with few vortex measurements above Eureka (grey) and 12–20 March 2011 (days 71–79, red). The outliers for years with few vortex measurements may result from errors in matching T490K and sPV490K to measurements both spatially and temporally when the vortex edge is near Eureka. For 12–20 March 2011, the ozone columns remain low, despite the rise in the local T490K. This deviation from the correlation between T490K and ozone suggests chemical depletion (supplementary material of Manney et al. [2011]). NO2 is also correlated with local lower stratospheric temperature, as has been observed in previous studies [e.g., Pommereau and Goutail, 1988; Dirksen et al., 2011]. The correlation for NO2 is weaker than for ozone, likely due to the seasonal increase in NO2as it is released from night-time reservoirs.

Figure 3.

Correlation between T490Kand (a) GBS ozone, (b) SAOZ ozone, (c) GBS-vis NO2, and (d) GBS-UV NO2. Measurements were taken inside the vortex for days 55–80 (24 February to 19/20 March). Data are shown for 2000, 2005, 2007, and 24 February to 11 March 2011 (blue); 12–20 March 2011 (red); and other years (grey). R and Rall are correlation coefficients for data indicated by blue only and for all data in the figure, respectively.

[14] Investigation of complementary datasets provides further evidence of chemical ozone depletion and denitrification above Eureka in 2011. For 9–18 March (days 68–77), low HNO3 and ClONO2 columns over Eureka were measured by the CANDAC Bruker FTIR [Lindenmaier et al., 2012]. This suggests that the extremely low NO2 columns measured during the same period are not caused by conversion to HNO3 or ClONO2. During this period of low ClONO2, HNO3, and NO2, OClO DSCDs are elevated, reinforcing that chlorine remains activated. Furthermore, PSCs were measured above Eureka with the CANDAC Rayleigh-Mie-Raman Lidar between 8–18 March (days 67–77) [Lindenmaier et al., 2012]. These measurements agree with photochemical model runs in supplementary material of Manney et al. [2011], which indicate that prolonged denitrification by sedimentation of PSCs delayed chlorine deactivation, leading to the record ozone loss.

[15] In order to isolate chemical ozone depletion from dynamical features, the passive subtraction method [e.g., Manney et al., 1995; World Meteorological Organization, 2003; Feng et al., 2007] was employed using SLIMCAT [Chipperfield, 2006], a three-dimensional off-line chemical transport model. These ozone loss estimates are described in detail in the auxiliary material. The average ozone loss for 12–20 March 2011 was 27% (29%) or 99 DU (108 DU), as estimated from GBS (SAOZ) data. The maximum percent ozone loss in the 11-year data record was calculated from GBS (SAOZ) data on 5 April 2011 at 47% (47%) or 250 DU (251 DU). A similar maximum ozone loss of 266 DU was observed byLindenmaier et al. [2012] on 5 April 2011 above Eureka.

5. Conclusion

[16] Unprecedentedly low ozone and NO2 columns were measured in 2011 and correspond to elevated OClO, suggesting chlorine activation and ozone depletion. Vortex ozone and NO2 total columns from 1999–2011 are correlated with the lower stratospheric temperature above Eureka, indicating that transport also contributes to the low ozone and NO2 measurements. Using the SLIMCAT passive tracer model, a maximum percent ozone loss of 47% was observed on 5 April 2011.

Acknowledgments

[17] The 2006–2011 GBS measurements were made at PEARL by CANDAC. CANDAC is supported by the Atlantic Innovation Fund/Nova Scotia Research Innovation Trust, Canada Foundation for Innovation, Canadian Foundation for Climate and Atmospheric Sciences (CFCAS), Canadian Space Agency (CSA), Environment Canada (EC), Government of Canada International Polar Year funding, Natural Sciences and Engineering Research Council (NSERC), Northern Scientific Training Program (NSTP), Ontario Innovation Trust, Polar Continental Shelf Program, and Ontario Research Fund. Ozonesonde measurements were made by EC. The spring 2004–2011 GBS, SAOZ, and ozonesonde measurements were also supported by the Canadian Arctic ACE Validation Campaigns, which were funded by CSA, NSERC, NSTP, EC, and the Centre for Global Change Science. The spring 1999–2000 GBS measurements were supported by NSERC and the University of Toronto and the 2001–2003 GBS measurements were supported by CFCAS and NSTP. SAOZ participation in the campaigns was supported by the Centre National D'Études Spatiales. The authors wish to thank PEARL site manager Pierre F. Fogal, the CANDAC operators, and the staff at EC's Eureka weather station for their contributions to data acquisition, and logistical and on-site support. Work carried out at the Jet Propulsion Laboratory, California Institute of Technology was done under contract with the National Aeronautics and Space Administration. The QDOAS data analysis software and ozone/NO2air-mass factors were provided by IASB-BIRA.

[18] The Editor thanks Hideaki Nakajima and an anonymous reviewer for their assistance in evaluating this paper.