SUMMARY: The overall structure of changes in total column ozone levels over a 50-year sample period from 1966 to 2015 and across a range of latitudes from -90 to +71 shows that the data from Antarctica prior to 1995 represent a peculiar outlier condition specific to that time and place and not a representation of long term trends in global mean ozone concentration. The finding is inconsistent with the Rowland-Molina theory of chemical ozone depletion and with the use of the periodic „ozone hole“ condition at the South Pole as supporting evidence for this theory first proposed in Farman etal 1985.
KEY EVENTS IN THE GENESIS OF THE MONTREAL PROTOCOL AND THE RISE OF THE UN AS GLOBAL ENVIRONMENTAL PROTECTION AGENCY
In 1971, environmentalist James Lovelock studied the unrestricted release of halogenated hydrocarbons (HHC) into the atmosphere from their use as aerosol dispensers, fumigants, pesticides, and refrigerants and found HHC in the air in the middle of the Atlantic Ocean. He was concerned that these chemicals were man-made and they did not otherwise occur in nature and that they were chemically inert and that therefore their atmospheric release could cause irreversible accumulation. In a landmark 1973 paper by he presented the discovery that air samples above the Atlantic ocean far from human habitation contained measurable quantities of HHC. It established for the first time that environmental issues could be framed on a global scale and it served as the First of three Key Events that eventually led to the Montreal Protocol and its worldwide ban on the production, sale, and atmospheric release of HHC. However, since HHCs were non-toxic and, as of 1973, environmental science knew of no harmful effects of HHC, the their accumulation in the atmosphere remained an academic curiosity.
This situation changed in the following year with the publication of a paper by Mario Molina and Frank Rowland in which is contained the foundational theory of ozone depletion and the rationale for the Montreal Protocol’s plan to save the ozone layer.
The Rowland-Molina theory of ozone depletion (RMTOD), the extreme volatility and chemical inertness of the HHCs ensure that there
is no natural sink for these chemicals in the troposphere and that therefore once emitted they may remain in the atmosphere for 40 to 150 years and be transported by diffusion and atmospheric motion to the stratospheric ozone layer where they are subjected to solar radiation at frequencies that will
cause them to dissociate into chlorine atoms and free radicals. Chlorine atoms can then act as a catalytic agent of ozone destruction in a chemical reaction cycle described in the paper and reproduced in Figure 1 below taken from Rowland and Molina, 1974.
Ozone depletion poses a danger to lie on earth because the ozone layer protects life on the surface of the earth from the harmful effects of UVB radiation. The Rowland-Molina paper, is the second key event
that led to the Montreal Protocol. It established that the atmospheric accumulation of HHC is not harmless and provided a theoretical framework that links HHC to ozone depletion that exposes life on the surface of the earth to the harmful impacts of UVB radiation. The Rowland-Molina paper is the Second Key Event that led to the Montreal Protocol. It established that the atmospheric accumulation of HHC is not harmless and provided a theoretical framework that links HHC to harmful ozone depletion.
The third key event in the genesis of the Montreal Protocol was the paper by Farman, Gardiner, and Shanklin that is taken as empirical evidence for the kind of ozone depletion described by the RMTOD (Farman, 1985). The essential finding of the Farman paper is contained in the top frame of the paper’s Figure 1 which is reproduced here as Figure 2. Ignoring the very light lines in the top frame of Figure 2, we see two dark curves one darker than the other. The darker curve contains average daily values of total column ozone in Dobson units for the 5-year test period 1980-1984. The lighter curve shows daily averages for the 16-year reference period 1957-1973. The conclusions the authors draw from the graph are that (1) atmospheric ozone levels are lower in the test period than in the reference period and (2) that the difference is more dramatic in the two spring months of October and November than it is in the summer and fall. The difference and the seasonality of the difference between the two curves are
interpreted by the authors in terms of the ozone depletion chemistry and their kinetics described by Molina and Rowland (Molina, 1974). The Farman paper was thus hailed as empirical evidence of RMTOD and the science of ozone depletion due to the atmospheric release of HHC appeared to be well
established by these three key papers. First, atmospheric release of HHC caused them to accumulate in the atmosphere on a planetary scale because they are insoluble and chemically inert (Lovelock). Second, their long life and volatility ensure that they will end up in the stratosphere where HHC will be dissociated by radiation to release chlorine atoms which will act as catalytic agents of ozone depletion (Molina-Rowland). And third, the Farman etal 1985 paper provides the empirical evidence and validates the depletion of ozone and therefore the RMTOD. The Montreal Protocol was put in place on this basis. LINK TO RELATED POST ON FARMAN ETAL 1985: https://tambonthongchai.com/2019/03/12/ozone1966-2015/
DATA AND METHODS: Total column ozone (TCO) measurements made with Dobson spectrophotometers at ground stations are used in this study. Twelve stations are selected to represent a large range of latitudes. The selected stations, each identified with a three-character code, are listed below. The locations of these stations are described and identified with global coordinates. Ozone data are prom these stations are provided online by the NOAA /ESRL and by the British Antarctic Survey. Most stations provide daily mean values of total column ozone in Dobson units. The time span of the data ranges from 1957 to 2015. The first year of data available varies from station to station in the range of 1957 to 1987, and the last month from August 2013 to December 2015. Some months and some years in the span of measurements do not contain data for many of the stations. The core study period is somewhat arbitrarily defined as consisting of ten Lustra (5-year periods) from 1966 to 2015. The Farman etal 1985 paper provides a precedence for the use of changes in 5-year means in the evaluation of long term trends. The period definitions are not precise for the first and last Lustra. The first Lustrum is longer than five years for some stations and shorter than five years for others. The last Lustrum is imprecise because of the variability in the last month of data availability.
The calendar month sequence is arranged from September to August in the tables and charts presented to maintain seasonal integrity. The seasons are roughly defined as follows: September-November (northern autumn and southern spring), December-February (northern winter and southern summer), March-May (northern spring and southern autumn), and June-August (northern summer and southern winter). Daily and intraday ozone data are averaged into monthly means for each period. These monthly means are then used to study trends across the ten Lustra for each calendar month and also to examine the average seasonal cycle for each Lustrum. Trends in mean monthly ozone and seasonal cycles are compared to examine the differences among latitudes. These patterns are then used to compare and evaluate the chemical and transport theories for changes in atmospheric ozone. The chemical explanation of these changes rests on the destruction of ozone by chlorine atoms derived from HHC (Molina, 1974) while the transport theory describes them in terms of the Brewer-Dobson circulation (BDC) and polar vortices that transport ozone from the tropics where they are formed to the greater latitudes where they are more stable.
TABLE 1: LIST OF STATIONS
DATA ANALYSIS STATION BY STATION: AMS SOUTH POLE ANTARCTICA
The data for AMS are summarized above. The first panel shows the number of observations reported in the dataset for each month of each Lustrum. We see in this panel that data are sparse for the months of September and March. The second panel contains the average value of total column ozone in Dobson Units (DU) for each month of each Lustrum. The columns in this panel represent long term trends for each month across the ten Lustra and the rows represent the average seasonal cycle in each Lustrum across the twelve calendar months. These trends and seasonal cycles are depicted graphically below.
Visually, Figure 3 indicates that the most extreme gradients in long term trends and the most extreme differences in ozone levels among months are seen in the southern spring months of September, October, and November. In the month of October ozone levels declined steeply losing more than 127from Lustrum#2 (266 DU) to Lustrum#6. A similar long term decline is seen for November where the decline persists from Lustrum#1 (344 DU) to Lustrum#9 (173 DU). In both of these months total column ozone at AMS fell to levels well below the arbitrary threshold of 200 DU where
atmospheric ozone concentration is described as an „ozone hole„. September data, though patchy, appear to mirror the October decline. The December decline is not as steep as those in October and November. In addition, the the data show large month to month differences in ozone levels among the
spring months exceeding 50 DU. For the rest of the year, ozone levels are
generally flat at about 280 DU throughout the study period with only small differences among the Lustra and among the months. A gradual decline in ozone levels from 280 DU to 240 DU is evident in the in the era prior to 1996 from Lustrum#1 to Lustrum#6. The ozone level appears to be stable in the post 1996 era at above 230 DU. The ozone hole appears to be a seasonal phenomenon peculiar to the spring month of October.
The average seasonal cycle for each Lustrum in Figure 4 shows ozone levels from September to August. In general, the ozone level tends to fall to its lowest level of well below 200 DU in October and then to rise sharply during November and December to above 300 DU with a gradual decline thereafter to 230 DU in late winter (August) before sinking back into ozone hole conditions in spring (October). The range of the seasonal cycle is about 120 DU. The seasonal cycle differs greatly among the Lustra prior to 1991
(left panel of Figure 4) but these differences appear to have narrowed since then (right panel of Figure 4.
HLB HALLEY BAY, ANTARCTICA.
The data for HLB are summarized in Table 4 above. The first panel shows that data are sparse for the months of May, June, and July. The seasonal cycle for each Lustrum and the long term trend for each month across the Lustra derived from Table 4 are depicted graphically in Figures 5&6. In Figure 5 the spring months (September, October, November) show a significant decline in ozone levels from Lustrum#1 to Lustra #6 and #7 with large differences in ozone concentration among the months. Mean October ozone levels fell almost 170 DU from Lustrum#1 (300 DU) to Lustrum#6 (132
DU). In the same period November ozone levels fell 150 DU. Similar magnitudes of decline are seen in the spring months of September (140 DU) and December (80 DU) and in the winter month of August (90 DU).
These data are historically important because the decline in October and November ozone levels by 80 DU or more from Lustrum#2 to Lustrum#4 reported by Farman et al (Farman, 1985) first alerted the world to what was thought to be catastrophic anthropogenic ozone depletion and served to validate the chemical theory of ozone depletion (RMTOD) attributed to HHC emissions (Molina, 1974). These data are therefore the proximate cause that triggered the Montreal Protocol of 1987 and its worldwide ban on HHC. The banned chemicals are described in the Protocol as ozone depleting substances.
In addition, the left panel of Figure 5 shows large differences in ozone concentration among the months that vary in the long term across the Lustra. The range of monthly values doubles from 75 DU in Lustrum#1 to 150 DU in Lustrum#6 before shrinking back to 90 DU in Lustrum#10. However, the great differences among months seen in the spring are mostly absent in the summer months of December, January, and February where we see only a modest decline in ozone levels with differences among the months and the rate of decline eroding with time across the Lustra. The data also show that ozone levels appear to have stabilized since Lustrum#6. Figure 6 shows that, as in AMS, the September to August seasonal cycle shows a steep rise during the southern spring months of October and November with a gradual decline during summer, autumn, and winter. Also in common with AMS are that large differences in ozone concentration among the Lustra are seen only in the seasonal cycles prior to 1995. These differences are greatly reduced in the period since 1995. Taken together, the data do not indicate that the sharp decline reported by Farman for the period Lustrum#2 to Lustrum#4 can be generalized as a phenomenon across the sample period.
LDR: Lauder, New Zealand
Total column ozone data from LDR are available for a relatively short period. Data for the first four Lustra are not available and the fifth and tenth Lustra are abbreviated. Unlike the Antarctica data, the graphical display of the LDR data in Figure 7 does not show a trend in ozone concentration for any calendar month. Also the lowest levels of ozone at LDR are generally higher than those at the Antarctica stations by about 100 DU. Yet another distinction from Antarctica is that the seasonal cycles for all Lustra in the dataset appear to converge into a single coherent pattern shown in Figure 8. In an 80 DU seasonal cycle, the ozone level is highest in the southern spring month of October at about 350 DU falling to a low of 270 in March. This pattern is the exact reverse of the seasonal cycle in Antarctica.
PTH: Perth Australia
Total column ozone data for PTH are displayed in Table 6 and in Figures 9&10. No long term trend in the ozone levels is apparent for any calendar month. A 50 DU seasonal cycle shows a high of 320 DU in the southern spring falling to a low of 270 DU in summer – in sync with LDR but shallower.
SMO: American Samoa
The data for SMO show a steady ozone level of approximately 250 DU with a standard deviation of 6 DU for all months of all Lustra. There is no evidence of trends. The seasonal cycle is almost flat.
MLO: Mauna Loa Hawaii
MLO ozone data show no trends across the ten Lustra. A very shallow 40 DU seasonal cycle fluctuates from a low of 240 DU in January to a high of 280 DU in May for all Lustra. The seasonal cycle is not in sync with those observed in the southern hemisphere.
WAI Wallops Island
Total column ozone data at WAI contain no apparent long term trend. They show a seasonal cycle with an amplitude of 70 DU running from a low of 280 DU in October to a high of 350 in April.
BDR: Boulder Colorado
Total column ozone data from the BDR station show no long term trends. An 80DU seasonal cycle rises from 270DU in October to 350DU in April similar to the seasonal cycle at WAI.
CAR: Caribou Maine
The data show a modest decline in ozone levels at CAR in the month of December at a rate of about 5DU per Lustrum on average. A 100DU seasonal cycle runs from a low of around 300DU in October to a high of 400DU in February. The seasonal cycle is fairly uniform across the ten Lustra.
BIS: Bismark, South Dakota
There are no sustained trends in the ozone data for BIS although a modest decline in the range of 2 to 4DU per Lustrum on average is seen in the months of March and April. A 100DU seasonal cycle runs from a low of 280DU in October to a high of 380DU in March. The seasonal cycle is fairly uniform for the ten Lustra and in sync with the seasonal cycle observed in BIS and in Antarctica.
FBK: Fairbanks, Alaska
Total column ozone data from FBK summarized in Table 13 show missing data for Lustrum#3 and for the northern fall and winter months of November, December, and January. The graphical depiction of the long term trends for each month (Figure 23) and for the average seasonal cycle for each Lustrum (Figure 24) contain gaps corresponding to the missing data. Still it is possible to discern in these graphs the absence of patterns or trends in ozone levels. In fact we find comparatively high ozone levels in the range of 300DU to 400DU corresponding to a 100DU seasonal cycle that goes from a low of less than 300DU in September to a high above 400DU in March. The seasonal cycle is in sync with those observed at CAR, BIS, and in Antarctica.
BRW: Barrow Alaska
Ozone levels are generally high compared with those in the lower latitudes and in the southern hemisphere. A 100DU seasonal cycle is evident in Figure 26 with the ozone level rising from a low of 300DU in September and October to a high of over 400DU in March and April. The seasonal cycle is similar to the ones observed in Antarctica and in the higher northern latitudes. No long term trends are evident.
SUMMARY AND COMPARISONS
The annual cycle in total column ozone at different latitudes
Figure 27 shows that the range of observed ozone levels is a strong function of latitude. It reaches a minimum of about 20DU in the tropics and increases asymmetrically toward the two poles. The hemispheric asymmetry has two dimensions. The northward increase in range is gradual and the southward increase in range is steep. Also, the northward increase in range is achieved mostly with rising maximum values while southward increase in range is achieved mostly with falling minimum values. The midpoint between the HIGH and LOW values is symmetrical within ±45 from the equator but diverges sharply beyond 45 with the northern leg continuing to rise while the southern leg changes to a steep decline as seen in Figure 28. Hemispheric asymmetry in atmospheric circulation patterns is well known (Butchart, 2014) (Smith, 2014) and the corresponding asymmetry in ozone levels is also recognized (Crook, 2008) (Tegtmeier, 2008) (Pan, 1997). These asymmetries are also evident when comparing seasonal cycles among the ground stations (Figure 29). The observed asymmetries are attributed to differences in land-water patterns in the two hemispheres with specific reference to the existence of a large ice covered land mass in the South Pole (Oppenheimer, 1998) (Kang, 2010) (Turner, 2009). The climactic uniqueness of Antarctica is widely recognized (Munshi, Mass Loss in the Greenland and Antarctica Ice Sheets, 2015) (NASA, 2016) (NASA, 2015).
The left panel of Figure 30 represents the southern hemisphere from AMS (-90o) to SMO (-14o). The right panel represents the northern hemisphere from MLO (+19.5o) to BRW (+71o). The x-axis in each panel indicates the calendar months of the year from September = 1 to August = 12. The ordinate measures the average rate of change in total column ozone for each calendar month among adjacent Lustra for all Lustra estimated using OLS regression of mean total column ozone against Lustrum number for each month. For example, in the left panel we see that in the month of September
(x=1) ozone levels at HLB (shown in red) fell at an average rate of 15DU per Lustrum for the entire study period; and in the right panel we see that in the month of July (x=11) ozone levels at FBK (shown in orange) rose at an average rate of more than 2DU per Lustrum over the entire study period. The full study period is 50 years divided into 10 Lustra but it is abbreviated for some stations according to data availability.
The concern about ozone depletion is derived from the finding by Farman et al in 1985 that ozone levels at HLB fell more than 100DU from the average value for October in 1957-1973 to the average value for October in 1980-1984. In comparison, changes of ±5DU from Lustrum to Lustrum seem inconsequential. In that light, and somewhat arbitrarily if we describe ±5DU per Lustrum as insignificant and perhaps representative of random natural variability, what we see in Figure 30 is that, except for the two Antarctica stations (AMS and HLB), no average change in monthly mean ozone from Lustrum to Lustrum falls outside this range.
It is therefore not likely that the HLB data reported by Farman et al can be generalized globally. We conclude from this analysis that the Farman etal study, the only empirical evidence thought to validate the Rowland Molina theory of ozone depletion, is flawed and therefore does not serve as evidence of anthropogenic ozone depletion. And yet, Farman etal 1985 served and still serves to this day as the sole empirical support for the ozone crisis that created the role for the UN in global environmentalism.