The nature of the surface controls many of the climatic features of the urban environment. There is much interest in using remotely-sensed data to interpret and analyze urban surface characteristics. Uncritical acceptance of such data can lead to erroneous conclusions regarding surface characteristics. For example, in areas of well-defined three-dimensional surface structure, such as cities, remote sensors viewing at nadir tend to oversample horizontal surfaces at the expense of vertical ones. Conversely, off-nadir sensors disproportionately view vertical surfaces. These biases induce a directional dependence on the measured surface emittance which in turn can lead to imprecise parameterizations of climatically important variables such as surface temperature (and the associated turbulent fluxes) if these effects are not taken into account.
The unique nature of the urban surface and its impact on the interpretation of remotely sensed data are discussed. A numerical model that attempts to portray the three-dimensional structure of the urban surface is presented, and its role in influencing sensor view factors and remotely sensed temperatures is demonstrated. This is achieved through the unique combination of ray-tracing software with an urban canopy-layer climate model. Output from the scheme is tested using observations obtained in a light industrial area of Vancouver, B.C. in August 1997.