The urban environment is very heterogeneous in terms of e.g. building layout, building material and geometry. The appropriate description and initialization of all these urban parameters affects the surface energy balance considerably and pays back on the near surface prognostic variables like temperatures, humidity and wind. Even though increasing computing power enables us to perform high-resolution simulations of urban areas incorporating detailed information of urban building properties and layout, a simplification of the description of the urban environment is desirable in order to keep the necessary input information at a minimum. The impact of urban input parameters and some simplifications in the description of the physical processes on the surface energy balance and the resulting near surface temperatures is investigated using the Single Column Reading Urban Model SCRUM with a four facet approach and a simpler 2 facet approach. This methodology now also forms the basis of the urban parameterisation being implemented into the Met Office Unified Model