Scientific Summary Chapter 2: Stratospheric Ozone and Surface Ultraviolet Radiation
Global Ozone Observations and Interpretation
As a result of the Montreal Protocol, ozone is expected to recover from the effect of ozone-depleting substances (ODSs) as their abundances decline in the coming decades. The 2006 Assessment showed that globally averaged column ozone ceased to decline around 1996, meeting the criterion for the first stage of recovery. Ozone is expected to increase as a result of continued decrease in ODSs (second stage of recovery). This chapter discusses recent observations of ozone and ultraviolet radiation in the context of their historical records. Natural variability, observational uncertainty, and stratospheric cooling necessitate a long record in order to attribute an ozone increase to decreases in ODSs. Table S2-1 summarizes ozone changes since 1980.
The primary tools used in this Assessment for prediction of ozone are chemistry-climate models (CCMs). These CCMs are designed to represent the processes determining the amount of stratospheric ozone and its response to changes in ODSs and greenhouse gases. Eighteen CCMs have been recently evaluated using a variety of process-based comparisons to measurements. The CCMs are further evaluated here by comparison of trends calculated from measurements with trends calculated from simulations designed to reproduce ozone behavior during an observing period.
Total Column Ozone
Average total ozone values in 2006-2009 have remained at the same level for the past decade, about 3.5% and 2.5% below the 1964-1980 averages respectively for 90°S-90°N and 60°S-60°N. Average total ozone from CCM simulations behaves in a manner similar to observations between 1980 and 2009. The average column ozone for 1964-1980 is chosen as a reference for observed changes for two reasons: 1) reliable ground-based observations sufficient to produce a global average are available in this period; 2) a significant trend is not discernible in the observations during this period.
Southern Hemisphere midlatitude (35°S-60°S) annual mean total column ozone amounts over the period 2006-2009 have remained at the same level as observed during 1996-2005, approximately 6% below the 1964-1980 average. Simulations by CCMs also show declines of the same magnitude between 1980 and 1996, and minimal change after 1996, thus both observations and simulations are consistent with the expectations of the impact of ODSs on southern midlatitude ozone.
Northern Hemisphere midlatitude (35°N-60°N) annual mean total column ozone amounts over the period 2006-2009 have remained at the same level as observed during 1998-2005, approximately 3.5% below the 1964-1980 average. A minimum about 5.5% below the 1964-1980 average was reached in the mid-1990s. Simulations by CCMs agree with these measurements, again showing the consistency of data with the expected impact of ODSs. The simulations also indicate that the minimum in the mid-1990s was primarily caused by the ozone response to effects of volcanic aerosols from the 1991 eruption of Mt. Pinatubo.
The latitude dependence of simulated total column ozone trends generally agrees with that derived from measurements, showing large negative trends at Southern Hemisphere mid and high latitudes and Northern Hemisphere midlatitudes for the period of ODS increase. However, in the tropics the statistically significant range of trends produced by CCMs (-1.5 to -4 Dobson units per decade (DU/decade)) does not agree with the trend obtained from measurements (+0.3 ± 1 DU/decade).
Northern Hemisphere midlatitude (35°N-60°N) ozone between 12 and 15 km decreased between 1979 and 1995, and increased between 1996 and 2009. The increase since the mid-1990s is larger than the changes expected from the decline in ODS abundances.
Northern Hemisphere midlatitude (35°N-60°N) ozone between 20 and 25 km declined during 1979-1995 and has since ceased to decline. Observed increases between 1996 and 2008 are statistically significant at some locations but not globally.
Northern Hemisphere midlatitude (35°N-60°N) ozone between 35 and 45 km measured using a broad range of ground-based and satellite instruments ceased to decline after the mid-1990s, consistent with the leveling off of ODS abundances. All data sets show a small ozone increase since that time, with varying degrees of statistical significance but this increase cannot presently be attributed to ODS decrease because of observational uncertainty, natural ozone variability, and stratospheric cooling. CCMs simulate the ozone response to changes in ODSs and increases in greenhouse gases; analysis of CCM results suggests that longer observational records are required to separate these effects from each other and from natural variability.
In the midlatitude upper stratosphere (35-45 km) of both hemispheres, the profile ozone trends derived from most CCMs from 1980 to 1996 agree well with trends deduced from measurements. The agreement in both magnitude and shape of the ozone trends provides evidence that increases in ODSs between 1980 and 1996 are primarily responsible for the observed behavior.
In the tropical lower stratosphere, all simulations show a negative ozone trend just above the tropopause, centered at about 18-19 km (70-80 hectoPascals, hPa), due to an increase in upwelling. The simulated trends in the lower tropical stratosphere are consistent with trends deduced for 1985-2005 from Stratospheric Aerosol and Gas Experiment (SAGE II) satellite data, although uncertainties in the SAGE II trends are large. The near-zero trend in tropical total ozone measurements is inconsistent with the negative trend found in the integrated SAGE I + SAGE II stratospheric profiles. The tropospheric ozone column does not increase enough to resolve this discrepancy.
Table S2-1. Summary of ozone changes estimated from observations.
Satellites, Umkehrs, FTIR
Northern midlatitudes 1980-1996
Declined by about 6%
Declined by about 9%
Declined by about 7%
Declined by about 10%
1992-1996 column and lower stratosphere data affected by Mt. Pinatubo
Northern midlatitudes 1996-2009
Increased from the minimum values by about 2% by 1998 and remained at the same level thereafter
Increased by about 6%
Increased byabout 2.5%
Increased by 1 to 2%, but uncertainties are large
Southern midlatitudes 1980-1996
Declined by 6%
Declined by about 7%
Declined by about 10%
Southern midlatitudes 1996-2009
Remained at approximately the same level
No statistically significant changes
No statistically significant changes
Increased by 1 to 3%, but uncertainties are large
Polar Ozone Observations and Interpretation
The Antarctic ozone hole continued to appear each spring from 2006 to 2009. This is expected because decreases in stratospheric chlorine and bromine have been moderate over the last few years. Analysis shows that since 1979 the abundance of total column ozone in the Antarctic ozone hole has evolved in a manner consistent with the time evolution of ODSs. Since about 1997 the ODS amounts have been nearly constant and the depth and magnitude of the ozone hole have been controlled by variations in temperature and dynamics. The October mean column ozone within the vortex has been about 40% below 1980 values for the past fifteen years.
Arctic winter and spring ozone loss has varied between 2007 and 2010, but remained in a range comparable to the values that have prevailed since the early 1990s. Chemical loss of about 80% of the losses observed in the record cold winters of 1999/2000 and 2004/2005 has occurred in recent cold winters.
Recent laboratory measurements of the chlorine monoxide dimer (ClOOCl) dissociation cross section and analyses of observations from aircraft and satellites have reaffirmed the fundamental understanding that polar springtime ozone depletion is caused primarily by the ClO + ClO catalytic ozone destruction cycle, with significant contributions from the BrO + ClO cycle.
Polar stratospheric clouds (PSCs) over Antarctica occur more frequently in early June and less frequently in September than expected based on the previous satellite PSC climatology. This result is obtained from measurements by a new class of satellite instruments that provide daily vortex-wide information concerning PSC composition and occurrence in both hemispheres. The previous satellite PSC climatology was developed from solar occultation instruments that have limited daily coverage.
Calculations constrained to match observed temperatures and halogen levels produce Antarctic ozone losses that are close to those derived from data. Without constraints, CCMs simulate many aspects of the Antarctic ozone hole, however they do not simultaneously produce the cold temperatures, isolation from middle latitudes, deep descent, and high amounts of halogens in the polar vortex. Furthermore, most CCMs underestimate the Arctic ozone loss that is derived from observations, primarily because the simulated northern winter vortices are too warm.
Ground-based measurements of solar ultraviolet (UV) radiation (wavelength 280-400 nanometers) remain limited both spatially and in duration. However, there have been advances both in reconstructing longer-term UV records from other types of ground-based measurements and in satellite UV retrievals. Where these UV data sets coincide, long-term changes agree, even though there may be differences in instantaneous, absolute levels of UV.
Ground-based UV reconstructions and satellite UV retrievals, supported in the later years by direct ground-based UV measurements, show that erythemal ("sunburning") irradiance over midlatitudes has increased since the late 1970s, in qualitative agreement with the observed decrease in column ozone. The increase in satellite-derived erythemal irradiance over midlatitudes during 1979-2008 is statistically significant, while there are no significant changes in the tropics. Satellite estimates of UV are difficult to interpret over the polar regions.
In the Antarctic, large ozone losses produce a clear increase in surface UV radiation. Ground-based measurements show that the average spring erythemal irradiance for 1990-2006 is up to 85% greater than the modeled irradiance for 1963-1980, depending on site. The Antarctic spring erythemal irradiance is approximately twice that measured in the Arctic for the same season.
Clear-sky UV observations from unpolluted sites in midlatitudes show that since the late 1990s, UV irradiance levels have been approximately constant, consistent with ozone column observations over this period.
Surface UV levels and trends have also been significantly influenced by clouds and aerosols, in addition to stratospheric ozone. Daily measurements under all atmospheric conditions at sites in Europe and Japan show that erythemal irradiance has continued to increase in recent years due to net reductions in the effects of clouds and aerosols. In contrast, in southern midlatitudes, zonal and annual average erythemal irradiance increases due to ozone decreases since 1979 have been offset by almost a half due to net increases in the effects of clouds and aerosols.