Remotely sensed temperatures provide a means to assess spatial variability of heat fluxes over urban areas. However, modeling of surface sensible heat fluxes are restricted because application of surface micrometeorological theory requires that a thermal roughness length or equivalently the parameter kB^-1 be known a priori. Little observational work has been completed over urban surfaces that can allow estimation of kB^-1 beyond that of the theoretical calculations for bluff-rough elements by Brutsaert (1982) and initial observational results for a simple, mostly non-vegetated surface by Voogt and Grimmond (2000). Here we extend our initial results to a wider range of urban surfaces that incorporate both bluff-rough and permeable-rough urban surfaces, taking into particular account the anisothermal nature of the urban surface. We also use our datasets to examine alternatives to the specification of an urban thermal roughness length including the estimation of the aerodynamic temperature and application of dual source models.