The present work reviews several aspects of flows that develop over surfaces that present an abrupt variation in surface roughness and temperature. In the first part, the work is concerned with the physical validation of a newly proposed formulation for the near wall temperature profile. The theory uses the concept of the displacement in origin, together with some asymptotic arguments, to propose a new expression for the logarithmic region of the turbulent boundary layer, which resorts to the dynamically determined scales to describe the temperature field. The new expressions are, therefore, of universal applicability, being independent of the type of rough surface - if of the types “K” or “D” - considered. The present formulation may be used as wall boundary conditions for two-equation differential models. The theoretical results are validated with experimental data obtained for flows that develop over flat terrain with sudden changes in surface roughness and in temperature conditions. Numerical predictions of the mean velocity and temperature profiles, and of the skin-friction coefficient and the Stanton number are compared with the experimental data. The computer code uses finite differences and the K-Є turbulence model. In the second part of the work, a laboratory study of a turbulent boundary layer which develops over surfaces that present an abrupt change in roughness is made. The cases of a uniformly smooth surface, of a uniformly rough surface and of surfaces that change from smooth to rough and from rough to smooth are investigated. Profiles of mean velocity, of turbulence intensity, of skin-friction coefficient, of the displacement in origin and of the thickness of the internal layer are presented. The skin-friction coefficient was calculated based on the chart method of Perry and Joubert (JFM, 17, 193-211, 1963) [1] and on a balance of the integral momentum equation. The same chart method was used for the evaluation of the displacement in origin. The thickness of the internal layer was calculated by two methods, the `knee` point method and the `merge` point method. Finally, in part three, the paper compares the present data with the lower atmosphere data of Bradley (QJRMS, 94, 361-379, 1968) [2] so that any possible analogy between the two works can be assessed.