Turbulence is an important phenomenon affecting safety when operating aircraft. It is difficult to recognize and avoid clear air turbulence (CAT), which occurs in clear air, because the weather radar equipped on the airplane cannot detect CAT. Therefore, in order to avoid CAT, accumulation of information such as observations and reports from aircrafts in flight and accurate forecasting technology are important. In order to investigate when and where turbulence frequently occur over Japan, the pilot reports (PIREPs) for 5 years from 2012 to 2016 were analyzed statistically. In addition to statistical analysis, which indexes can express the occurrence potential of CAT was examined by calculating diagnostic indexes using numerical model.

In this research, we focus on the moderate or greater (MOG) turbulence because PIREPs include enough MOG reports and MOG turbulence is more remarkable. MOG turbulence occurred frequently in winter and spring, especially around Okinawa in winter and spring and around Niigata in summer especially in the upper level above 20 000 ft. The seasonal variation of the horizontaldistribution was obvious in the latitude direction than in the longitude direction. MOG turbulence in the summer season in the high latitude region and the winter in the low latitude region accounted for a large proportion of the MOG turbulence in the year. Since it corresponds to the seasonal variation of the jet stream over Japan, it is supposed that much of the MOG turbulence occurs around the jet stream. On the other hand, in the lower level, the number of MOG turbulence is concentrated around the domestic major airports, and seasonal variation in the latitude direction is not seen, so it is supposed that much MOG turbulence occurred around the time of takeoff and landing.

The CAT cases were picked up by checking the clouds by infrared images of weather satellites among severe turbulence above 20,000 ft. The number of CAT cases was largest in November of all months and in spring of all seasons. The seasonal variation of CAT distribution in the latitude direction were also seen. We examined the distance between the jet stream and CAT occurrence points, and it was found that 76.7% of all CAT occur around the jet stream.

In order to investigate which indexes can express the occurrence potential of CAT, 9 indicators were calculated for each CAT case from 2014 to 2016 by weather forecast experiment with CReSS (Cloud Resolving Storm Simulator). As a result of calculation of each indicator, it was found that frontogenesis function and TI 1 (Turbulence index 1), which express wind shear and front around the jet stream, are available to express the occurrence potential of CAT. It correspond with that much of CAT occurred around jet stream. On the other hand, the Brant-Vaisala frequency and the vertical gradient of potential temperature which represent the stability of the atmosphere cannot express the occurrence potential of CAT. The possible reasons are that the vertical resolution was low to calculatethese indexes and that it was not frequently to occur the CAT due to KH instability or mountain waves.