Generation of urban runoff Seasonal and climate change perspective

Sammanfattning: Runoff generation in cold regions is characterized by snowmelt contributions to runoff during the periods of thawing and changing runoff patterns due to frozen ground. This thesis project aimed at addressing these challenges by advancing the procedures for winter urban runoff computations and the assessment of control measures during the winter/spring period, when the snowmelt and frozen soils dominantly impact runoff generation in the current and future climates. In such considerations, contributions of green/pervious areas to runoff and stormwater drainage systems were found particularly important and were addressed in one of the study components by conducting sensitivity analysis of the runoff modelling tool used, the MIKE SHE model. For this purpose, four runoff generation scenarios were defined, including the baseline reference scenario, a future climate scenario with up-scaled precipitation, and two scenarios with widely different infiltration rates. The results showed that the variations of infiltration capacity and the precipitation magnitude largely influenced runoff generation and impacted on the drainage system. Such impacts were measured by the number of flooded nodes and surcharged pipes, which greatly increased with decreasing infiltration capacity (described by Ks=1×10-10 m/s, which corresponds to the bedrock) and somewhat increased for increasing future precipitation (+20%). Projection of future climatological parameters to 2100 (i.e. temperature, precipitation and maximum hourly precipitation) were obtained for investigating seasonal changes in the town of Kalmar (southern Sweden). The results indicated that the seasonal precipitation patterns would become more similar in all the seasons, and the winter period would experience more changes in runoff generation, which would require more attention in stormwater management with respect to both snowmelt simulation and considerations of frozen grounds. To advance the understanding of urban snowmelt modelling, a literature review of selected snowmelt models was undertaken to identify which of them could be readily used, or easily modified, for improving the current snow modelling practice. For this purpose an urban snow cover classification (13 classes) was developed on the basis of the following considerations: human activities affecting snowmelt, land use, and the origin of deposited snow. Various snow covers in urban areas were then assessed and general recommendations were made for selecting the most appropriate model for specific studies, considering the study goals, constraints on the collection of field data, budget/time restriction, and the required accuracy. Urban runoff controls by green infrastructures, during the cold season, were studied for green roofs and infiltration facilities. Green roofs were found to be effective in warm weather, when they could counterbalance almost all the extra rainfall imposed by climate change, in a mixed land use catchment in Luleå, retrofitted with green roofs covering 30% of the catchment area. On the other hand, green roofs produced no benefits in the cold season with sub-zero temperatures and snow removal. Infiltration of runoff into two frozen engineered (sandy) soils, with slightly varying gradation, was studied in the laboratory for two values of the initial gravimetric water content (5 and 10%). Soil thawing process and restoration of infiltration capacity was slowed down by increasing water content and the content of fines in the soil. Thus, the soil with higher water content and finer gradation required more time for attaining full infiltration capacity after soil thawing. Practical implications of study results for bioretention facilities include the recommended use of coarser engineered soils, conservative estimation of infiltration rates, provision for bypassing of high flows, and fitting the facility with a valve-controlled under-drain facilitating bioretention drainage before the onset of freezing weather.