Analysis of urban databases with respect to mesoscale modeling requirements

A number of different parameterization schemes are used in mesoscale models to approximate the effects of the urban canopy on meteorological flow field development. At a minimum, urban landuse information is needed to help prescribe roughness and surface energy balance parameters. More complex urban canopy parameterizations require morphological information cross-correlated with landuse. In this presentation, we look at two primary issues: 1) what is the utility/accuracy of readily-available USGS landuse data and 2) how are urban landuse categories related to particular building morphological characteristics. The USGS landuse data is free, available online, and covers the entire U.S. at 200 m resolution, hence it is a valuable resource to mesoscale modelers. However, the data was derived from analysis of satellite images dating from 1973 and only 7 urban landuse categories were characterized, with several being ambiguous (e.g., "other urban or built-up land"). In the first part of this presentation, we compare the USGS land use data to newer, more detailed landuse datasets collected in the Los Angeles and Phoenix metropolitan areas. We show differences in the areal extent of urban coverage due to growth of the cities over the 25 year period since the USGS data was obtained. We also ascertain the fractional make-up of the 7 USGS landuse categories in terms of the corresponding 110 Los Angeles and 18 Phoenix urban landuse categories in order to better characterize the somewhat ambiguous USGS urban landuse types. In the second part of the presentation, we will show how the urban landuse categories correlate to building morphology for the Los Angeles area. Building morphological characteristics include average height, plan area density, and building area density vs. height, for example. The characteristics of other cities will be investigated if 3-d building datasets can be obtained.