Variability of the hydrologic performance of green infrastructures due to climatic regimes across Sweden

In recent years, increasing scientific evidence has demonstrated the potential of urban green infrastructures (GI) for coping with increased stormwater runoff caused by urbanization, by limiting both runoff flows and volumes (Pennino et al., 2016). However, their hydrologic performance is influenced by climatic conditions (Johannessen et al., 2018), and a better understanding of variations in performance between different geographic situations with differing climatic regimes is needed to ensure that design criteria are adapted to local conditions. A recent cluster analysis of rainfall data across Sweden has identified four distinct regions with respect to short-duration extreme events (Olsson et al., 2017) demonstrating that rainfall is not homogenous throughout the country. However, it is not yet clear how this variability, along with another climatic variability, affects the hydrologic performance of GI. The current guidelines for the design and dimensioning of stormwater control measures (Larm & Blecken, 2019) use unique design criteria for the entire country. The objective of this work is to evaluate variability in the expected hydrological performance of GI facilities under different Swedish climatic regimes in order to consider whether the design and dimensioning recommendations should vary between regions. Using a hydrologic model with long-term rainfall and climate data for eleven urban areas in Sweden, hydrologic performance, both in terms of volume reduction and peak-flow attenuation, was evaluated for three types of GI: (1) biofiltration cells, (2) green roof and (3) detention ponds. Climate data were obtained from the SHMI-national meteorological network covering 23 years(1997–2019) at a 15-min time resolution for each of the urban areas, which were selected to include at least two areas within each of the four extreme-rainfall regions identified by Olsson et al. (2019). Modeled facility configurations were based on the existing recommendations (Larm & Blecken,2019). Estimation of daily reference Penman-Monteith evapotranspiration values was used as input for the hydrological model. In addition, uncertainties in hydrologic outputs of the model due to parameter estimation were quantified using a stochastic approach. Preliminary results show variations in annual stormwater volume reduction caused by differences in rainfall volumes, dry durations between events, and air temperatures, which all influence antecedent soil moisture, leading to differences in available retention capacity at the beginning of events. All locations showed considerable stormwater runoff volume reductions using a biofilter cell, ranging from 85% to 92%, with the highest values in Stockholm and Eskilstuna. The highest volume reductions were found for locations with the warmest conditions, which also had the highest average ET. This may be explained by the fact that in GI, the soil’s retention capacity is renewed as stored water is lost to evapotranspiration between events. A higher variability is observed for peak flow reduction. This indicates that while it may be reasonable to have national recommendations for cases where volume reduction is the only objective when peak-flow reduction is of importance, it may be important to account for regional variability. Future work will consider a greater variety of facilities and account for uncertainties in order to identify whether differences between regions are significant.