Success of Montreal Protocol Demonstrated by Comparing High-Quality UV Measurements with "World Avoided" Calculations from Two Chemistry-Climate Models

Abstract

The Montreal Protocol on Substances that Deplete the Ozone Layer has been hailed as the most successful environmental treaty ever ( https://www.unenvironment.org/news-and-stories/story/montreal-protocol-triumph-treaty ). Yet, although our main concern about ozone depletion is the subsequent increase in harmful solar UV radiation at the Earth's surface, no studies to date have demonstrated its effectiveness in that regard. Here we use long-term UV Index (UVI) data derived from high-quality UV spectroradiometer measurements to demonstrate its success in curbing increases in UV radiation. Without this landmark agreement, UVI values would have increased at mid-latitude locations by approximately 20% between the early 1990s and today and would approximately quadruple at mid-latitudes by 2100. In contrast, an analysis of UVI data from multiple clean-air sites shows that maximum daily UVI values have remained essentially constant over the last ~20 years in all seasons, and may even have decreased slightly in the southern hemisphere, especially in Antarctica, where effects of ozone depletion were larger. Reconstructions of the UVI from total ozone data show evidence of increasing UVI levels in the 1980s, but unfortunately, there are no high-quality UV measurements available prior to the early 1990s to confirm these increases with direct observations.

Figure 1

Comparison between projected ozone values in the GEOS-CCM (blue lines) and NIWA UKCA (red lines) World Avoided models for high (a,d), mid (b,e), and low (c,f) latitudes. Thicker lines are 5-year running means. Green lines indicate the smoothed difference between the two models (i.e., NIWA-UKCA minus GEOS-CCM). Plots for latitude 15° are representative of all equatorial latitudes, while those for 75° are representative for higher polar latitudes.

Figure 2

UVI calculated from World Avoided and World Expected simulations for the period 1960 to 2100 for high (a,d), mid (b,e), and low (c,f) latitudes. The shaded area shows the seasonal range of predictions calculated from the NIWA-UKCA World Avoided simulation. The upper envelope of the shaded range refers to peak values predicted for the summer and the lower envelope to minimum values predicted for winter. The red and blue lines indicate the 5-year running means for the NIWA-UKCA and GEOS-CCM World Avoided model simulations, respectively. The green line is the 5-year running mean of the NIWA-UKCA World Expected simulation, which assumes full compliance with the Montreal Protocol.

Figure 3

Total ozone column (a) and daily maximum UVI within 1 hour of local noon (b) at Lauder. Panels (c–f) show UVI changes for winter (c), spring (d), summer (e) and autumn (f) determined from measurements of the NDACC instruments (blue), calculated from total ozone column (black), and projected by the two World Avoided (red and magenta) and World Expected (green) CCM model runs. UVI changes were normalized as described in the text and the “UVI ratios” shown in Panels (c–f) are ratios relative to these normalizations. Note that for the season that spans two years (summer at this southern hemisphere site), the year label refers to the year at the start of the season. For example, the summer of Dec 2017 to Feb 2018 is plotted at 2017.

Figure 4

Comparison of normalized UVI ratios for summer months at all sites between measurements, calculations for clear skies based on the ozone assimilation, and clear-sky models for the World Avoided and World Expected scenarios. Note that the vertical axis scale varies between panels.

Figure 5

Comparison of UVI in spring at polar sites between measurements, calculations for clear skies based on the ozone assimilation, and clear-sky models for the World Avoided and World Expected scenarios.

Figure 6

Calculated decadal trends in measured UVI since 1996 (or from the data start year, if later than 1996) as a function of site latitude, compared with those calculated for clear skies from observed ozone, and as calculated by the two World Avoided model runs and the World Expected run for each season (a–d). Sites where the time series spans 20 years or more are denoted by bold text and solid symbols. The number of years of data included in the trend analysis at each site is indicated beside the site name. If data from some seasons are missing, this number can be less than the total number of years. Error bars shown are 2-σ uncertainties of the regression model.