第九届青年地学论坛

The spatiotemporal changes and driving factors of surface ozone in China since 2013 have been widely studied in recent years. However, due to a lack of long-term observations, reports on historical ozone concentration levels, their changes, and influencing factors are severely limited. In this study, we applied the XGBoost machine learning algorithm to correct the CMIP6-simulated surface ozone concentrations from 1950 to 2014. The long-term evolutions of ozone and meteorological effects on interannual ozone variations and trends in China are further analyzed. The results revealed that CMIP6 historical simulations have a large underestimation in ozone concentrations and their trends. The XGB-derived ozone are closer to observations, with R² value of 0.66 and 0.74 for daily and monthly retrievals, respectively. Both the concentrations and exceedances of ozone in most parts of China have shown increasing trends from 1950 to 2014. The higher ozone growth rates of XGB retrievals than those from the model indicate a regional surface ozone penalty due to the warming climate. The relatively significant increment in ozone are estimated in the Central and Western China. Seasonally, the ozone enhancement is largest in spring, indicating a shift in seasonal varation of ozone. Given the uncertainty in simulating historical ozone by climate model, we show that machine learning approaches can provide improved assessment of evolution in surface ozone, along with valuable information to guide future model development and formulate future ozone pollution prevention and control policies.

Surface ozone,Machine learning,CMIP6