Because of different radiative properties of the urban and rural surfaces, surface temperatures over urban areas tend to be higher than the ambient air temperatures. This is particularly true for large metropolitan areas. Thus, recent rapid urbanization over the Washington-Baltimore metropolitan region has motivated us to conduct this project evaluating the effects of changing vegetation and residential coverage on the surface climatology of the region. Some preliminary results on the “urban heat island” effects have obtained.

The urban heat island effects are studied using the next generation numerical weather prediction model, i.e., the Weather Research and Forecast (WRF) model coupled with a single-layer urban canopy model. The coupled urban canopy model used in this simulation is developed by NCAR's scientists, based on a single-layer urban canopy parameterization of Kusaka et al. (2001),wherein the effect of roofs, walls, roads and even the anthropogenic activity are incorporated. The four-nested domains are used with the horizontal resolutions of 13.5, 4.5, 1.5 and 0.5 km, respectively. The model integration is performed from 12 UTC July 7, 2007 to 12 UTC July 10, 2007, during which period ozone concentration was peaked and daytime surface temperature was abnormally high. The NCEP 1°×1°Final Analysis (FNL) data are used to generate the initial and lateral boundary conditions. In the urban canopy model, the urban land-use map over the Washington-Baltimore area is categorized to the following 3 groups: 1) low density residential; 2) high density residential; and 3) commercial/industry/transportation.

A comparison of the modeled skin temperature to NASA's MODIS satellite data shows that the model captures very well the hot skin temperature at major cities and towns as well as major roads on the early afternoons. The hottest skin temperature is as high as 44°C in the central portion of Washington and Baltimore. The skin temperature distribution resembles well that of the land use. The skin temperature difference between urban and rural areas is as high as 12°C. Some differences between the observations and model data could be attributed to the lack of land-use information after 2001. Results indicate that many surface meteorological variables could be reasonably predicted in the future when high-resolution land-use data becomes available.