Mesoscale modeling of urban-rural complex with the inclusion of land surface processes

We have been conducting a research that seeks for practical and cost-effective countermeasures to mitigate adverse effects of the urban climate. Measures under consideration include cooling pavements, reduction of anthropogenic heat, roof vegetation, increasing water surface and others.

In this study, a meso-scale model was developed to apply for the urban heat island simulation based on the PSU/MM5 that is a popular model being used in the United States for both research and operational forecasting. The major changes made to MM5 are the parameterization of surface heterogeneity to deal with complex urban land-use, inclusion of a land surface model and the anthropogenic heat release with spatial and temporal variation into the surface layer. The land surface model predicts water contents within surface soil layers and is capable of representing surface moisture availability depending on the weather conditions and soil types. In addition, AVHRR composite data were used to derive better estimates of surface parameters, i.e. albedo and fraction of vegetation. Both parameters have significant effects on the surface energy balance.

The model was first applied to the Sendai City, Japan and its suburbs to approve the capability of the model in simulating the urban heat island. The Sendai City is one of the major coastal cities with a population of more than one million. Simulations were conducted for actual events in the summer, 1995 under clear sky conditions. Two domains are used for the simulations. The resolution of the mother domain and the nested domain are 6km and 2km, respectively. The outputs from the model such as near-surface air temperature, wind velocity, wind direction as well as their vertical profiles were compared with observed variables obtained at 6 automated weather stations. Assuming the middle initial moisture content, the simulated air temperature agreed well with the observations. At some points, however, the model overestimated the minimum air temperatures in the early morning, especially in the suburban areas. This can be attributed to the modeling of radiation cooling.

Using the model verified in this study, the effects of the soil moisture content on the near-surface air temperature, humidity and mesoscale interaction between the city and suburban areas were investigated to find their sensitivity to the land surface processes. The middle and high initial moisture conditions produced similar air temperatures over 3 days, while the lower soil moisture increases the air temperature in the city center by 3 degrees Celsius during the daytime due to the mesoscale heat and vapor transport. This implies that more realistic simulations of the urban heat island can be achieved by considering land surface processes in the suburban areas as well as the inclusion of the urban canopy model. In this paper, the simulation results obtained by the use of satellite derived data will be also compared with those given using the conventional method in which the surface parameters are estimated from the land-use types in each computation grid.