Adaptation of a leaf wetness model to estimate dewfall amount on a roof surface

Most leaf wetness and dew studies are rural, however, there is growing scientific interest in urban dew. This study investigates the feasibility of adapting an existing leaf wetness (dew) model for urban use. The approach predicts dewfall amount for (a) a maple leaf, and (b) an asphalt-shingle roof, using a computed surface energy balance (which includes a roof subsurface heat flux), latent heat fluxes, and an inferred surface water balance. Simulations are run and verified to a first-order using data collected in Vancouver, BC, Canada, in 1996. Too few data are available to adequately test the ‘leaf’ model. The skill of the ‘roof’ model to predict dewfall amount is confirmed by several statistical indices, e.g. a root mean square error of 0.04 mm per night, and Willmott's index of agreement of 0.87. Skill is seen also in the timing of dewfall onset and maxima. It seems feasible to estimate dewfall on an artificial surface, such as a roof, using a ‘leaf wetness’ approach, with suitable modifications. Approaches could readily be applied to other surfaces, and are a first step towards a comprehensive urban dew model.