EGU General Assembly 2020
© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Trends in ambient ozone concentrations at twelve sites of the Czech Republic over the past three decades

Iva Hunova1, Marek Brabec2,3, and Marek Malý2,3
Iva Hunova et al.
  • 1Czech Hydrometeorological Institute, Ambient Air Quality Department, Prague, Czechia (
  • 2National Institute of Public Health, 100 42, Prague, Czech Republic
  • 3Institute of Computer Science, Academy of Sciences of the Czech Republic, 182 07, Prague, Czech Republic

Ambient ozone (O3) remains a serious air pollution problem (O3) of Northern Hemisphere, and still represents a considerable threat both for human health and ecosystems. In Europe, the critical levels of O3 are permanently exceeded over vast areas (EEA, 2019). In the Czech Republic (CR), monitoring of O3 has been operated since 1993, currently at 50 sites, including both rural and urban stations covering the country (CHMU, 2019). O3 exposures in the CR are relatively high (Hůnová, Schreiberová, 2012; Hůnová et al., 2016), and may result in negative endpoints, both regarding human health (Hůnová et al. 2013) and vegetation (Hůnová et al., 2011). O3 is highly meteorology dependent and shows considerable year-to-year variations (Hůnová et al., 2019 a, b). Two to three-decade time series allows for a sound trend analysis, hence O3 concentrations for trends at Czech long-term monitoring sites were already analysed using Mann-Kendall non-parametric test (Hůnová, Bäumelt, 2018).

This time, however, our approach for time analysis was different. We applied a generalized additive model, GAM (Wood, 2017; Hastie & Tibshirani, 1990) framework as a flexible, semiparametric regression approach to address nonlinear trend shapes in a formalized and unified way. In particular, we employed penalized spline approach with cross-validated penalty coefficient estimation.  We have examined daily mean O3 concentrations measured at twelve Czech sites representing different environments, geographical areas, and altitudes across the country; four urban, for rural and four mountain sites. We used long-term data series from the time period of 1994–2018.

Our results show inconsistent behaviour of sites before 1998 when the strict emission limits were introduced with an immediate consequence of substantial decrease in O3 precursor emissions. The highest concentrations and the most dynamic O3 decrease in this time period was recorded at the Praha 4-Libus urban background site, the lowest concentrations and the steepest increase in O3 were recorded at the Rudolice mountain site in the former Black Triangle Area. Two local maxima – around 2003 for some sites and 2006 for other sites – and a local minimum around 2013 are indicated. Steady increase in O3 concentrations for all sites is evident after 2014 up to now, most likely due to recent five hot and dry summer seasons. Seasonal O3 course averaged for the entire measuring period is similar for all sites, with clear maximum in May-June. The highest O3 in summer and lowest in winter were observed at the Usti nad Labem-Kockov site, relatively most flat curve, with the least differences between summer and winter was recorded at the Churanov site, in the Sumava Mts. More interesting is to compare the seasonal O3 curves for individual years.

In contrast with Mann-Kendall test standardly used for this kind of analysis, the GAM approach offers a detailed view on both time trend and seasonality curve and facilitates the analysis and interpretation of the results.



How to cite: Hunova, I., Brabec, M., and Malý, M.: Trends in ambient ozone concentrations at twelve sites of the Czech Republic over the past three decades, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-6973,, 2020

This abstract will not be presented.