Can Climate Models Reproduce the Decadal Change of Dust Aerosol in East Asia?

Dust aerosol plays an important role in the Earth System. As a natural aerosol, dust aerosol is often calculated interactively in global climate models and temporal variations of dust emission in the past century are far less constrained compared to those of anthropogenic aerosol emissions. Here we evaluate dust emission in East Asia simulated by 15 climate models participating in the Coupled Model Intercomparison Project Phase 5. The results show that none of the models can reproduce the observed decline of dust event frequency during 1961-2005 over East Asia. The models tend to simulate either much less decline or even increase of dust emission. The discrepancy is mainly ascribed to weaker or opposite trends of surface wind speeds and precipitation in the models. These results cast a doubt on the interpretation of long-term variations of dust-affected fields in climate models and highlight the need for further improvements of the models. Plain Language Summary Atmospheric aerosol is one of key factors that influence climate change. Aerosols include anthropogenic aerosols (such as sulfate aerosol and black carbon from fossil fuel burning) and natural aerosols (such as dust that is emitted by strong winds over bare soil). Climate models are important tools we use to predict the climate in the future, and the reliability of climate models lies in their ability to reproduce the change of climate system in the past. Here we examine the ability of climate models in reproducing the long-term change of dust storm frequency in East Asia. First, historical records of dust events show the dust activities decreased greatly during 1961 to 2005 over East Asia. Compared to this observation, current climate models are unable to reproduce the large decline of dust activities during 1961 to 2005. The reason is that climate models cannot capture the decrease of surface wind speed and increase of precipitation. These results imply that climate models may not represent well the change of meteorological elements (e.g., temperature and clouds) that will be affected by dust. Our results highlight urgent need to improve the performance of climate models in simulating long-term dust change.