Urban Snowmelt Processes: Modelling and Observation

Sammanfattning: In northern Scandinavia and other cold regions, urban drainage systems are often unable to cope with the high volumes of melt water which can be sustained for several weeks. High waste water treatment plant inflows, combined system overflows and low water quality are a few problems associated with snowmelt. However, reports detailing urban snowmelt processes are lacking in the hydrological literature. This thesis reports research into the snow hydrology of the City of Luleå, Sweden. There are four objectives: to investigate (a) the relationship between snow and landuse; (b) the effects of urbanisation on snowmelt processes; (c) the effect of melt water on urban hydrology; and (d) the feasibility of modelling urban snowmelt processes. These are met in a series of five appended papers. Papers I and II investigate snowmelt sensitivity to the urban environment. Paper III concerns seasonal patterns of the town water balance. Paper IV presents a snow survey that reaffirms the assumptions about snow cover and characteristics made in Papers II and III. Paper V is a literature review that questions the validity of urban runoff routing models on the basis of findings from the earlier papers. Urbanisation is associated with extreme heterogeneity and diminished surface permeability. The urban hydrograph shows high peak flows and rapid responses after short, low intensity events. Snowmelt processes are the same in both rural and urban areas, but climate and snowpack properties differ greatly. Urban snow, for instance, is subject to ploughing. Snowmelt induced runoff generation is further complicated by a more pronounced thermal component of the water balance compared to rain events. Snowpack energy fluxes are greatly influenced by the urban environment. Melt is earlier and more intense due to enhanced radiation to the south of buildings in comparison to snowpacks in open ground or to the shaded north. Urban snow has low albedo and high density than that of rural snow leading to speeded melt. The role of surface type and snow cover characteristics upon storm- and waste water inflow is demonstrated. Predictions of waste water inflow are improved by spatially weighting snow types according to a priori assumptions of snow properties and coverage. For instance, the proportion of undisturbed snow coverage decreases as landuse intensity increases. Snow piles in suburbs tend to be on permeable ground, but they are more likely to be on sealed surfaces in the central city. There are seasonal differences in runoff volumes and flow pathways to the waste water treatment plant and receiving waters. Increased volumes of runoff, reduced concentration times and long event duration lead to high waste water loads at the treatment plant. The surface water component of sewage originates from direct flow into pipe inlets and infiltration into sewer pipes. Current methods of estimating urban snowmelt, such as the degree-day method, and runoff are found to be problematic. However, a lack of data for developing, testing or operating more sophisticated models restricts urban engineers.