Water retention performance of newly constructed green roofs in cold climates

Infiltration rates in urban areas are low due to a high proportion of impervious structures. Impermeability results in increased rates of urban runoff, which often leads to degradation in receiving waters. Stormwater retention in urban areas can be increased, for instance, with green, vegetated roofs. While impervious, normal roofs produce the runoff immediately, studies with green roofs have shown that they cause delays in peak runoff and reduce the runoff rate and volume by water retention and attenuation. The water retaining capacity of green roofs vary due to local weather conditions and roof characteristics. Several studies and experiments considering stormwater management and other ecosystem services that green roofs provide have been performed worldwide, mainly in temperate regions, while more studies are needed in cold climates especially to quantify the performance of green roofs in winter. The purpose of this thesis was to evaluate the effects of various types of precipitation events on runoff rates, timing and water retention in locally built new green roofs from late summer to early winter in southern Finland. Data on precipitation and green roof runoff as well as soil temperature and humidity were monitored automatically. Green roofs included a) precultivated readymade vegetation mats and b) built on site novel substrate mixture with plug plants and seedlings. My principal aim was to study the capability of green roofs in retaining and attenuating water in relation to the intensity and duration of precipitation, the length of the antecedent dry weather period as well as to temperature and moisture of the green roof substrate. I hypothesised that green roofs retain low intensity events better than high intensity events and more effectively in dry and warm than in wet and cold weather, being negligible at temperatures below 0 °C. Furthermore, I studied whether retention capacity can be improved by the amendment of biochar. Finally, readymade green roofs with dense vegetation was hypothesised to have better retention capacity than the newly created roofs with very sparse vegetation but only in summer due to evaporation. Based on cumulative runoff, green roofs retained 52 % of rainfall, which is close to the retention capability found in previous studies. Retention was generally higher at warm temperatures and for biochar-amended roofs, in agreement with my hypothesis. Against expectations, roofs with readymade vegetation mats had lower retention than those built on site. In summer and autumn, before freezing temperatures occurred, results were generally according to the hypotheses and previous research: retention rates decreased as rain depth or rain peak intensity increased. When the amount of rain preceding the measurement event was low, and the substrate moisture content was low, retention was better. During wintertime, results were contrary to my hypotheses: Total retention rates increased with the amount of rain and rain intensity, or when substrate moisture content increased. However, a long antecedent dry weather period resulted in better retention, especially during winter. Mean peak flow attenuation for rain events in this study was 64 % and results are in accordance with my hypothesis and previous research. Furthermore, delay times from when the rain event started to when runoff started and from rain peak to runoff peak were detected in this study, mean values of more than 1 h being comparable to what has been reported in the literature. Results from my thesis can be used to improve hydrological models for local stormwater management purposes. Furthermore, results can be compared with those of other ecological stormwater treatment methods. Possible future research topics include the functioning of green roofs during different seasons and especially during freezing and melting periods with assumedly complex hydrological interactions