Improvements of the dust prediction system in Japan Meteorological Agency

Japan Meteorological Agency (JMA) has been providing aeolian dust information around Japan based on numerical dust predictions by a dust prediction model and surface synoptic observations since 2004. In November 2014, we updated the dust prediction model to a new version. The new dust prediction model consists of an AGCM (called MRI-AGCM3) and a global aerosol transport model (called MASINGAR mk-2) which contains some improvements in aerosol processes (e.g. dust emission scheme, included aerosol species, etc.). Some components of this new dust prediction model are commonly used to an Earth-system model (called MRI-ESM1) for global climate change research. Statistical verification of the model against the synoptic surface observations around Japan shows the threat score (TS) for dust prediction is improved in the new model mainly for the first 3 days of the forecast period. Our verification results also suggest that the new model particularly reduces overestimates of the surface dust concentration around Japan by the upgrade of dust emission and deposition schemes. And it also indicates that the new model can predict the dust distribution better than the old one. As a result of this change, the accuracy of Asian dust predictions has been improved. JMA has a plan to increase the horizontal resolution of the prediction model from about 110 km to 40 km. Preliminary results of the model verification show the improvement of forecasting scores and we expect that this upgrade will lead to improve dust predictions. In addition, Meteorological Research Institute (MRI) of JMA has been now developing an aerosol data assimilation system using the aerosol optical depth (AOD) data of JMA's new geostationary satellite, Himawari-8 to improve dust predictions. This system shows a good performance for dust predictions in some cases. JMA also has a plan to use this assimilation system for improvement of Asian dust predictions in the next few years.