The Urban Heat Island (higher temperatures in the urban areas respect to the surrounding rural areas) is a phenomenon that has been studied since more than a century. Many modelling studies can be found in the literature, mainly from the last 3-4 decades, that focus on this phenomenon in a variety of cities around the world. However, it is becoming more and more evident that cities do not behave uniformly, and that a large microclimate variability can be found within the same city. This is the result of the interaction between microscale features, determined by the urban morphology and human activity of each specific neighborhood, and the mesoscale features that are generated at the scale of the whole city and the surrounding rural areas, that influence the development of the Planetary Boundary Layer above the city itself. The next challenge of mesoscale modelling in urban areas is to be able to proper capture this intra-urban variability. To do this, urban parametrizations embedded in mesoscale models must be able to distinguish between different urban morphologies. This is an essential capability in order to use mesoscale models to inform urban planning, e. g. to choose the best urban structure in the context of a changing climate. In this seminar I will present the work that has been done in CIEMAT (Spain) to improve this capability. This is based on the extensive use of CFD microscale models over different types of obstacle configurations (representing different urban morphologies) to provide the needed parameters (essentially drag coefficients and turbulent length scales) for the urban canopy parametrization. The achievements reached so far, and the challenges linked to this approach will be highlighted. An example of the impact that the new parameters, derived with this technique, can have on urban mesoscale modeling will be presented for the city of New York.