Recently, there have been some efforts in combining the CFD and atmospheric models capabilities to address effects on air flows from a building to terrain scales. This is what is required to simulate air flows over the urban areas in complex terrain and/or coastal areas. This paper discusses how an atmospheric model HOTMAC was improved to simulate air flows around buildings under the influence of mesoscale wind variations.

We simulated diurnal variations of air flows around a cluster of buildings, which were bound by the ocean and hills. Large cities are often located in a coastal area or near complex terrain. Prediction of transport and diffusion of air pollutants and toxic materials is a considerable interest to the safety of the people living in urban areas.

The governing equations for mean wind, temperature, mixing ratio of water vapor, and turbulence are similar to those used by Yamada and Bunker (1988). Turbulence equations were based on the Level 2.5 Mellor-Yamada second-moment turbulence-closure model (1974, 1982). Five primitive equations were solved for ensemble averaged variables: three wind components, potential temperature, and mixing ratio of water vapor. In addition, two primitive equations were solved for turbulence: one for turbulence kinetic energy and the other for a turbulence length scale (Yamada, 1983).

Pressure variations are caused by the changes in wind speeds, and the resulted pressure gradients subsequently affect wind distributions. We adopted the HSMAC (Highly Simplified Marker and Cell) method (Hirt and Cox, 1972) for pressure computation because the method is simple yet efficient. The method is equivalent to solving a Poisson equation, which is commonly used in non-hydrostatic atmospheric models.

Two inner domains were nested in a large domain. The first domain was 6560 m x 8960 m with a horizontal grid spacing of 160 m. The second domain was 1280 m x 1440 m with a horizontal grid spacing of 40 m and the third domain was 360 m x 400 m with a horizontal grid spacing of 10 m.

Domain 1 includes topographic features such as the ocean, coastal area, plains, and hills. Domain 2 is a transition area between Domain 1 and Domain 3. Buildings were located in Domain 3.

There were significant interactions between air flows generated by topographic variations and a cluster of buildings. The winds were blocked and the sea breeze fronts were retarded by buildings. Winds were calm in the area surrounded by buildings. Winds diverged in the upstream side and converged in the downstream side of the building cluster. Wind speeds and wind directions around buildings changed as the winds in the outer domains encountered diurnal variations.

We also found that the surface temperatures in the urban area were several degrees higher than those in the surrounding non-building areas. The warming was caused mainly by the reduction of wind speeds by building blocking. When the roofs of the buildings were heated by the sun, the surface temperatures increased additional one or two degrees. In summary, we were successful in simulating qualitatively both mechanical (blocking) and thermal (heating) effects of buildings in a coastal area where sea- and land-breeze circulations were significant.