Assessing AOD simulations from 16 CMIP6 models based on satellite-derived AOD from 2010 to 2019 over China

Abstract:

In recent years, China has implemented a set of effective pollution control policies to tackle the severe air pollution problems so that the obvious decline of AOD has been captured by MODIS-Terra and in-situ observation. However, it is still unknown how well CMIP6 models simulate AOD and its decline trend in China. Here we comprehensively assess the performance of CMIP6 models in simulating climatology and the trend of annual mean AOD over eastern and central China (ECC) from 2010 to 2019. The climatology of AOD is the largest over Sichuan Basin and North China Plain, which can be reasonably reproduced by CMIP6 multi-model ensemble (MME), outperforming individual models over Sichuan Basin (SCB) with larger correlation coefficient (CC) and more equivalent standardized deviation (SD) to the observation. MODIS exhibits two maximum centers at 23°N in spring and 40°N in summer, but CMIP6 MME shows the maximum center at 34°N in spring. A large uncertainty exits among CMIP6 models in simulated AOD trend and MME improves the trend simulation. UKESM1-0-LL, GFDL-CM4, HadGEM3-GC31-LL, and ACCESS-CM2.0 perform well over the ECC with the high CC of 0.7. Even though some models show similar verdicts with the satellite-derived results, seasonality of the AOD trend cannot be simulated by the majority of models over these regions.