Japan Geoscience Union Meeting 2022

Presentation information

[J] Poster

A (Atmospheric and Hydrospheric Sciences ) » A-AS Atmospheric Sciences, Meteorology & Atmospheric Environment

[A-AS11] Atmospheric Chemistry

Sun. May 29, 2022 11:00 AM - 1:00 PM Online Poster Zoom Room (8) (Ch.08)

convener:Risa Uchida(Japan Automobile Research Institute), convener:Yosuke Sakamoto(Kyoto University Graduate School of Global Environmental Studies), Yoko Iwamoto(Graduate School of Integrated Sciences for Life, Hiroshima University), convener:Shigeyuki Ishidoya(Advanced Industrial Science and Technology), Chairperson:Risa Uchida(Japan Automobile Research Institute), Yosuke Sakamoto(Kyoto University Graduate School of Global Environmental Studies), Yoko Iwamoto(Graduate School of Integrated Sciences for Life, Hiroshima University), Shigeyuki Ishidoya(Advanced Industrial Science and Technology)

11:00 AM - 1:00 PM

[AAS11-P02] Multi-linear regression analysis of recent trends of global total ozone

*Yuhei Suzuki1, Isao Murata2,1 (1.Graduate School of Science, Tohoku University, 2.Graduate School of Environmental Studies, Tohoku University)


Keywords:ozone, multiple linear regression models, trend

Ozone depletion started in the 1970s, when it was realized that the breakdown of chlorofluorocarbons (CFCs) in the stratosphere would release chlorine atoms that could catalytically destroy ozone [Chipperfield et al., 2017]. As a result of the 1987 Montreal Protocol and its amendments, the atmospheric loading of anthropogenic ozone-depleting substances is decreasing [Chipperfield et al., 2017]. Therefore, the total ozone is expected to recover. However, no statistically significant trend has been detected in the mid-latitudes and tropics because the increase in ozone is small relative to the other factors such as dynamically forced year-to-year variations [WMO, 2019].
To investigate the recent trend of ozone, we analyzed the temporal variations of total ozone around the world. In this study, we used total ozone data from the World Ozone and Ultraviolet Radiation Data Centre (WOUDC). We selected a total of 43 sites with few missing periods in various latitudes. We investigated trends for about 20 years after 2000 using the least squares method. The inter-annual variability was fitted with a straight line, and the seasonal variability was fitted considering up to one-half annual cycle. We used 95% confidence intervals to confirm statistical significance. When we investigated the 20-year time series, we found that the trends of many sites changed around 2015. Therefore, we divided the 20-year period into two periods around 2015. As a result, downward trends were seen after around 2015 in the Northern Hemisphere. Upward trends were seen after around 2017 in the tropics. The temporal variations in the tropics during 20-year period nearly matched the phase of the 11-year solar cycle. In this analysis we did not consider the 11-year solar cycle and other proxies that affect the total ozone.
Next, we used the multiple linear regression (MLR) model provided by the “Long-term Ozone Trends and their Uncertainties in the Stratosphere” (LOTUS) group to consider the known proxies. At first, total ozone time series at Tsukuba (36 °N) and Natal (6 °S) were analyzed using the standard proxies of LOTUS. The proxies considered were the El Niño–Southern Oscillation (ENSO), the 11-year solar cycle, the quasi-biennial oscillation (QBO), aerosols and the Equivalent Effective Stratospheric Chlorine (EESC). We calculated the standard partial regression coefficients for each proxy. QBO and EESC were effective at Tsukuba, and the 11-year solar cycle and QBO were effective at Natal. We used the adjusted R-square as an indicator of the goodness of fitting of the MLR model. This should be close to 1 when the fitting is good. The result shows 0.047 at Tsukuba and 0.586 at Natal. This means that the temporal variations of these sites could not be explained without considering other proxies that were not considered, especially at Tsukuba. In order to investigate the factors of the total ozone trend in more detail, we will add new proxies, such as potential vorticity (PV). Then, we will analyze other sites.