A comparison of precipitation intensity from radar with model results around coastal topography during cold-air outbreak periods

The effect of topography is relatively large on shallow convection. Snow clouds, forming and developing over the central part of the Japan Sea in winter, usually have a cloud top height of 4000 m to 6000 m. The cloud top height is restricted by a stable (often inverse) layer at the top of the convective mixed layer. Such stratification results in the strong topographic effect to the convective cloud systems, which influences the snowfall distribution on the land and the possibility of snow-related disasters. The convective mixed layer behaves as a shallow-water layer when it passes over a low mountain barrier of a height of 600 m, and supercooled droplets generated by the orographic lifting can enhance snowfall in the lower convective mixed layer. Partial blocking and hydraulic jump may also occur when the convective mixed layer is shallower and/or the mountain range is higher.

Heavy snowfall lasted accompanying outbreak of the strong winter monsoon from December 2005 to February 2006. Many convective cloud systems brought continuous strong snowfall, resulting in the record snow depth of more than 3 m at inland cities. The monsoon outbreak was strong especially in December 2005, when many convective cloud systems traveled into the land. They gave us an opportunity to analyze the basic processes of coastal and orographic effect on the convective cloud systems. In this paper, the precipitation variation of the convective cloud systems within onshore monsoon wind is examined in relation to the coastal and orographic effect, using radar observation and numerical simulation data.

Radar observations were carried out at the Snow and Ice Research Center (SIRC, at 20 km distance from coastline), National Research Institute for Earth Science and Disaster Prevention (NIED), almost continuously from 10 to 29 December by 12-elevation volume scans in 3- to 4-minute intervals. The observed data were projected on a Cartesian coordinate with horizontal/vertical resolution of 1 km / 500 m. Equivalent reflectivity factor (Ze) and radial velocity (Vr) data are used for the analysis. The precipitation variation was investigated using snow water equivalent (SWE) calculated assuming an experimental Ze-SWE relation.

Numerical simulations of snowfall in December 2005 were carried out at the Meteorological Research Institute (MRI) of the Japan Meteorological Agency (JMA) by the way of double-nesting of JMANHM. JMANHM is a nonhydrostatic mesoscale model for research and operational use, developed at the Numerical Prediction Division (NPD) of JMA and MRI. The simulations started at every 6 hours. Outer model was executed using regional objective analysis data (RANAL) of JMA as initial and boundary conditions. The 3-hour forecast and the succeeding forecast data of the outer model was used as initial and boundary conditions of the inner model. The resolutions of the RANAL, outer and inner models are 10 km, 5 km, and 1 km, respectively. An explicit 2-moment bulk-type cloud microphysics scheme was used in both outer and inner models. The Kain-Fritsch convective parameterization scheme was additionally applied only to the outer model. The integration time of the inner model was 9 hours and the data output interval was 1 hour in the forecast time. Four-hour to 9-hour forecast data were used for the analysis. The observed and simulated precipitation variation showed qualitatively similar characteristics. It can be said that the simulation reproduced the feature of the precipitating cloud system successfully.

Variations of the snowfall intensity were examined on the convective systems with respect to the distance from the coastline. The analysis was performed for seventeen cases with the northwesterly or north-northwesterly prevailing winds; defined by periods with the similar feature of the precipitation distribution. The cases were classified into the 3 types of convective cloud systems, that is, longitudinal lines, transversal lines, and meso-beta scale vortices. The precipitation variation showed different characteristics among the three types. Precipitation of the longitudinal lines was moderately enhanced on the land. In cases of the transversal lines, the precipitation enhancement was significant. The difference between longitudinal and transversal lines is outstanding, in spite of their similarity in the linear shape. These features could be attributed to the structure of the convective systems and the effect of the coastal topography. The precipitation variation of the meso-beta scale vortices was characterized by both of the low-level SWE maximum and echo-top maximum in the coastal area. The numerical simulation confirmed that the variation in precipitation intensity around coastal topography depends on the type of the convective cloud systems. The detailed structures of the convective cloud systems, especially the difference among the three types, are examined in our future issues.