A proper description of the computational domain is needed for a Computational Fluid Dynamic (CFD) that models Atmospheric Transport and Dispersion (ATD) in urban settings. Urban setting descriptions are not always provided in a suitable format for fast processing for a CFD model. In certain occasions, the only data available does not even provide the minimum requirements to be used as input for CFD modeling. Those requirements are related to the need of tessellating the computational domain where the Navier-Stokes equations are discretized. Some of the requirements are well defined surfaces to describe building geometry, non-manifold surfaces, and overall, assure that the surfaces conform a watertight geometry. The digital description of urban settings is usually in ShapeFile format, and the terrain of such urban area is in Digital Elevation Model (DEM). The combination of these two data sets into one as input for a CFD model is not straight forward. Therefore, there has been attempts in the past to simplified these modeling constrains for using CFD modeling. This paper will compare four of the most common modeling approaches in the case of the Oklahoma City data set. The modeling approaches compared are: watertight, immersed, embedded, and porosity approach. The watertight approach needs a seamless surface of complex building structures jointly with the terrain. The geometrical description of both terrain and buildings together is challenging and existing approaches are laborious, time-consuming and error-prone. We used an efficient and robust new procedure using computational geometry techniques to derive triangulated building surfaces from 2D polygon data with height attribute. We also used a novice method to merge the resultant building surfaces with the triangulated terrain surface to produce a seamless surface for the entire study area. The immersed and embedded approaches require only a simple description of the buildings as boxes that do not need to be intersected. The main computational domain is a big box and the buildings will define inside and outside points to the domain where the Navier-Stokes equations are calculated. Finally, in the porosity approach we applied a porosity coefficient to describe the buildings. We applied all four approaches to the Oklahoma City experiment and revealed the main differences and the gains and losses for each of them.