When cold air outbreaks occur from the Eurasian Continent, snow clouds develop over the Sea of Japan. The snow clouds are frequently intensified along the coastal region. But the understanding of its real situations and mechanisms is not enough. In particular, the stationary snow clouds along the coastal region have not been investigated at all. An observation of snow clouds was performed in Hokuriku, Japan in winter of 2000 to 2001 using a X-band Doppler radar of Nagoya University. The radar range was overlapped by another C-band dual-polarization radar. From 15 to 16 January 2001, two adjacent snowbands stayed along the coast of Hokuriku for 20 hours. The purpose of this study is to clarify the structure and maintenance process of the stationary snowbands along the coast.

In the middle of January, the cold air outbreak occurred over the Sea of Japan. On 15 January, the cold air less than -33 C was present at 500 hPa over Hokuriku. But the north-westerly monsoon wind in the lower level was weaker than in typical cold air outbreaks. A mixing layer with a thickness of 4 km was formed. In the mixing layer, two snowbands were formed along the coast; Snowband 1 and Snowband 2. Snowband 1 was located over the sea and had strong radar reflectivity. Snowband 2 was formed over the land and had relatively weak reflectivity. The convective cells in both snowbands showed successive developments. The development of a cell in Snowband 2 followed the decay of a cell in Snowband 1. The horizontal wind derived from VAD method showed that the south-easterly wind developed gradually and had a thickness of 400 m. The north-westerly monsoon wind and the south-easterly land breeze made strong convergence which resulted in strong updraft. A strong divergence zone was present over the strong convergence zone in Snowband 1 and made the weak convergence zone in the middle layer in Snowband 2. The weak convergence contributed to the formation of weak updraft. The Z and ZDR distributions were investigated in 7 boxes which were located along the trajectories of cellular echoes in both snowbands. The ZDR of Snowband 1 concentrated negative values when the reflectivity became strong. This indicates the developments of graupels. On the other hand, below Snowband 2, all pictures of snow particles taken at the ground showed aggregates. The values of about 0 dB in ZDR and its variance larger than Snowband 1 are consistent with the result.

The strong convergence between the monsoon wind and the land breeze maintained Snowband 1. The formation of snow particles in Snowband 1 was mainly by the riming process. The strong divergence in Snowband 1 made the weak convergence in Snowband 2. Therefore the weak updraft was produced in Snowband 2. In the weak updraft, the aggregates were formed by the aggregation and deposition processes.