Kai et al. (1997) collected lidar and radiosonde measurements to examine the cause for formation of cumulus clouds (so called "Kampachi Cloud") frequently observed over the Metropolitan Loop Road 8 in Tokyo. >From the analysis of the surface wind and temperature measurements, they found that Kampachi Cloud was located over the surface wind convergence zone, which also coincided with the boundary of the heat island of Tokyo.
In this study, we used a three-dimensional mesoscale model HOTMAC (Yamada & Bunker, 1989)to simulate Kampachi Cloud to understand the cause of formation and investigate the influence of urbanization on the cloud. We employed nested grids: the outer grid domain was 240 km x 260 km x 7800 m (vertical) with horizontal grid spacing of 4 km and the inner grid domain was 20 km x 40 km x 7800 m (vertical) with horizontal grid spacing of 1 km.
Simulation started at 0000 lt (local time), August 8, 1994 (Julian day 220). Initially, wind speed of 5 m/s and south-southwesterly wind direction (210 degrees) were assumed. Sea breezes began to develop approximately at 1000 lt from both Tokyo Bay and Sagami Bay. Sea breezes from two bays converged and produced cloud street. The convergence zone approximately coincided with the location of the Loop Road 8. The simulated clouds began to appear approximately at 1000 lt, developed well in the afternoon, and dissipated by 2000 lt, which approximately agree with the observed sequence.
Sensitivity studies were conducted. First, the inner grid was removed to examine the effect of horizontal grid spacing on the cloud formation. Initial values and boundary conditions remained unchanged from the control run. Sea breezes from Tokyo Bay and Sagami Bay were underestimated considerably due to a coarse horizontal grid spacing (4 km). Consequently, no cloud street formed over the Loop Road 8. Second, surface boundary conditions were modified taking the urbanization effect into consideration. Distributions of anthropogenic heat fluxes, soil moisture, roughness height, and albedo were obtained from the land use distributions. The control run assumed the distributions were uniform. Other conditions remained the same as in the control run. The cloud street formed, but the amount of clouds reduced considerably compared with the control run. The main reason for the cloud reduction was drier surface due to urbanization.
In summary, numerical simulations appear to support the speculation by Kai et al. that the main reason for the cause of Kampachi Cloud is the convergence of sea breezes developed from Tokyo Bay and Sagami Bay. Numerical results also suggest that the urbanization may have decreased the cloud amount due to drier surface.