2008 Volume 86 Issue 6 Pages 951-967
This paper studies an evaluation of climatological reproducibility and one of future climate projections of extreme precipitation in the Baiu season (from mid-May to July) around Japan using data of numerical experiments. The models are a non-hydrostatic cloud-system-resolving atmospheric model with a horizontal grid size of 5 km (5km-NHM) utilized as a regional climate model (RCM), and an atmospheric general circulation model (AGCM) with a horizontal resolution of about 20 km (20km-AGCM) which provided the boundary condition of 5km-NHM. Ten-year runs were performed by the time-slice method. The mean precipitation amount of the three heaviest precipitation events that ranked in order of precipitation amount in every year (PTop3) at each grid point was used as an index to examine quantitatively extreme precipitation within a specific term with various precipitation accumulation periods (APs).
The 5km-NHM experiment reproduced PTop3 much better than that of 20km-AGCM, especially for shorter APs. In the projection of PTop3 in the future climate with 5km-NHM, ratio in two climate states of PTop3 for longer APs and that of precipitation amount show the roughly same pattern. The increase areas of those were localized around Kyushu, Japan. In particular, the PTop3 for longer APs was projected to increase 30.70% in the mountainous areas along the Pacific coast of western Japan. This increase was due to a cyclonic circulation of an incremental low-pressure generated on the western side of Kyushu. On the other hand, large increases of PTop3 for shorter APs spread widely with a scattered pattern to the area where precipitation decreases. This characteristic was associated with the intensification of precipitation due to global warming.
