Long-term ozone trends at Uccle

We will shortly describe the temporal variation of the time series of ozone measurements that are available at Uccle.

Total ozone

The evolution of the total ozone concentrations measured at Uccle is inherently linked with the time evolution of the abundances of ozone depleting species like CFCs and halons containing bromine in the stratosphere. In the first years of the observations, started in 1971, there were almost no man-made ozone depleting substances (ODS) in the stratosphere. However, from about 1980 onwards the concentration of ODS gradually increased, to peak in the second half of the 1990s (in particular in 1997). Thanks to the industrial phase-out and ban of the CFCs and halons containing bromine (the Montreal Protocol, signed in 1987), their stratospheric concentrations are decreasing slowly since 1997. Therefore, we will use the data before 1980 as the reference period for pre-anthropogenic influence. To see whether there is a difference in the ozone changes, trends with respect to this reference period are calculated, and for two periods (before and after 1997) in the next Fig. 1.

Fig. 1: Evolution of the ozone column at Uccle as observed with Dobson 40 (1972-1989), Brewer 16 (1990-2001), and Brewer 178 (2002-present). The times of major volcanic erutions affecting the ozone layer are indicated on the time axis.

The trends are -0.23+/-0.05 % per year and +0.13+/-0.02 % per year for the periods 1980-1997 and 1997-2023, respectively. Although this can be interpreted as a sign of the recovery of the ozone layer, it should however be noted that the last period is too short for final conclusions about the recovery of the ozone layer. Moreover, the large year to year variability during the last decade is striking, which is attributed to the dynamical behaviour of the stratosphere itself. 

To calculate the total ozone trends as a function of season (see Fig. 2), the total ozone column time series has been split up into monthly mean ozone values (in DU). Then a least square linear regression is applied to these time series. During the entire time range, there is only a significant negative total ozone column trend present in the data during June, and a significant positive one in November, as shown in Fig. 2 below.

Fig. 2: Monthly total ozone trends at Uccle in percent per year derived from the ozone observations between 1971 and 2023. The vertical bars represent the 95 % confidence level of the calculated trends.

Top

Vertical ozone distribution

Information on the vertical distribution of ozone is obtained from ozone soundings. The vertical profile time series has been homogenized and is now of the highest possible quality for further research. The more than 50 years of Uccle ozone sounding data is visualized in Fig. 3 below. The vertical coordinate is the altitude relative to the tropopause (tropopause location = 0), which marks the border between the troposphere and the stratosphere. So positive values in the altitude level are in the stratosphere, negative values in the troposphere. As ozone is formed by different processes in the troposphere and the stratosphere, this distinction allows for a better study of the ozone concentrations at those different atmospheric layers. 

With an average tropopause location around 11 km at Uccle, it should be obvious from this Fig. 3 that the maximum ozone concentrations (blue-violet colours) occur between altitudes of 20 to 25 km.

Fig. 3: The time series of the vertical ozone profiles (ozone partial pressures in mPa) measured at Uccle by ozonesondes. The altitude level is expressed in km relative to the tropopause height.

To have a better idea of the time variability of the atmospheric ozone concentrations, we calculated the mean ozone profile for the different subsequent decades from the dataset of ozone profiles shown here above. These different mean ozone profiles are presented by different colors in Fig. 4 here below, with the altitude relative to the tropopause height again as vertical coordinate.

From this graph, it can be noted that the first decade of observations (1969-1978, in black) had the lowest mean tropospheric ozone amounts (negative values of the vertical coordinate), while the highest stratospheric ozone amounts were measured during this decade. The minimum mean stratospheric ozone concentrations were measured in Uccle during the 1989-1998 decade (green colored graph), a decade also characterized with a high amount of (harmful) tropospheric ozone concentrations. During this decade, the amounts of chlorine and bromine substances (like chlorofluorocarbons or CFCs and halons) in the stratosphere reached peak  values (with its maximum in 1997), and these substances deplete very efficiently the ozone amount in the stratosphere.

Since the decreasing concentrations of ozone depleting substances in the stratosphere thanks to the Montreal Protocol emission regulations, the ozone layer is recovering (see blue, grey, and orange curves, corresponding to the periods 1999-2008, 2009-2018, and 2019-2023 respectively), but the tropospheric ozone concentrations also remain high. We can also clearly see that the stratospheric ozone amounts measured at the beginning of our time series are still not reached during the most recent years. Full ozone layer recovery is predicted to take place around 2050, and also depends on climate change, which interacts with stratospheric ozone amounts.

Fig. 4: Mean vertical ozone profiles calculated for the subsequent decades (+ the remaining years) in which ozone sounding data are available at Uccle.

We now calculate the relative trends in the ozone concentrations by a least square linear regression on the ozone partial pressures at the different altitude levels. For the entire 1969-2023 time period, presented in black in Fig. 5 below, the ozone concentrations have been decreasing throughout the stratosphere between -1 to -2 %/decade, while increasing in the troposphere at a rate close to +2%/decade. However, it should be noted that these trends in the stratosphere are the result of a strong decreasing ozone amounts in the period 1969-1996 (around -5%/decade, in red in Fig. 5), followed by increasing ozone amounts  (around +2%/decade, in green) around the ozone layer maximum in the period 1997-2023. The choice of splitting the time period in the year 1997 stems from the peak concentrations of ozone depleting substances in the stratosphere (the same trend turning point is chosen in the total ozone trends, see Fig. 1), but also coincides with the change of ozonesonde sensor type at Uccle.

Fig. 5: Vertical ozone concentration trends in % per decade, for different time periods. The error bars denote the 2σ standard errors of the linear regression slope determination and can be considered as a rough estimate of the trend uncertainty.

Finally, the linear trends are obtained on the monthly mean ozone partial pressures, and we show these trends for the different months separately in Fig. 6 below, for our entire time period of more than 50 years of measurements. The most significant decrease in the ozone concentrations around the ozone maximum (between 10 to 15 km in altitudes relative to the tropopause) took place in spring and early summer (March to June, dark red colored), when the concentrations at the ozone maximum are high (highest from February to April, see here). At altitudes above the ozone maximum, the ozone amounts also decreased significantly in November-December.

Another interesting finding from Fig. 6 is that the less significant increases in tropospheric ozone concentrations (light blue colors) arise during the summer months, when tropospheric ozone production is highest due to photochemical reactions involving sunlight and ozone precursor gases from anthropogenic fossil fuel and biofuel combustion, crop burning, chemical solvents, etc.

Fig. 6: Season-height cross-section of ozone trends in percent per year at Uccle for the period January 1969 to December 2023. Areas where the trend is statistically significant at the 95 % level are colored darker (red for negative and blue for positive trends). The altitude level is in units (km) relative to the tropopause height.

Top

Cookies saved