Modelling urban snowmelt runoff

Abstract

Few investigations have been made into modelling snowmelt in urban areas; hence, current urban snowmelt routines have adopted parameters and approaches intended for rural areas that are not appropriate in an urban environment. This paper examines problems with current urban snowmelt models and proposes a model that uses parameters developed from field studies focusing exclusively on urban snow. The Urban Snow Model (USM) uses an energy balance scheme at an hourly time step, changes in urban snow albedo, and incorporates eight different types of redistributed snow cover. USM is tested against observed flow data from a small residential community located in Calgary, Alberta. The degree-day method for snowmelt, the SWMM model, and a modified version of USM that incorporates a partial energy budget scheme relying only on net radiation, are also tested against the observed flow data. The full energy budget version of USM outperformed all other models in terms of time to peak, peak flowrate and model efficiency; however, the modified version of USM fared quite well and is recommended when a lack of data exists. The degree-day method and the SWMM models fared poorly and were unable to simulate peak flowrates in most cases. The tests also demonstrated the need to distribute snow into appropriate snow covers in order to simulate peak flowrates accurately and provide good model efficiency.