A general methodology of urban cover classification for atmospheric modelling

To be able to run atmospheric models with an urban canopy parameterization, the land-use classifications must include specific urban covers, which characteristics and properties are different from those of natural covers. The Town Energy Balance (TEB) scheme has been recently implemented in the GEM and MC2 Canadian models, but the land cover database currently used for these models includes only water, ice, and various types of soils and vegetation covers. The objective of this work is to develop a general methodology that would provide urban land-use classifications in a semi-automatic way for the major Canadian cities, in order to represent the spatial distribution and the diversity of urban areas. This method is based on the joint analysis of satellite imagery and digital elevation models (DEMs), and the application of a decision tree model to identify the urban classes. A preliminary methodology has been developed for Oklahoma City and refined for Montreal.