Newer integrated land use-transport models require as a base-year input a representation of existing built form (building stock by type, age, density, and other variables). Available data on floor space inventory have been found to be both incomplete and inconsistent with employment and population data in three land use-transport modeling exercises in the United States: in Sacramento, California; in the state of Oregon; and in Oahu, Hawaii. This paper reports on three methods used to synthesize a built-form input at a detailed microlevel for these three models. Each method uses existing land cover data and reconciles them with population data, employment data, and data on the consumption of floor space by activity. All three approaches assign floor space quantities by type to a detailed microlevel (grid cell or parcel) inventory of land. The approaches differ in their algorithmic approach, motivations, and data requirements.