Measurement of the runoff coefficient of extensive greenroof

Describes the procedure of experimental measurement of the runoff coefficient C, both of individual layers and the entire composition extensive green roofs. Experimental measurements make it possible to determine the reference behaviour of runoff characteristics, namely runoff coefficient C, with emphasis on the simulation of the real behaviour of extensive green roofs. The aim is an elementary description of the structural and physical behaviour of extensive green roofs. For the needs of experimental measurement, the dimensional and shape limits of test specimens are described, the conditions for conditioning of individual specimens, the boundary conditions of execution and individual steps of the experiment. Then is specified the method of evaluation and subsequent verification of measured data. The result of the experimental measurement is the amount of drained water from the tested specimens of the extensive green roof at time t, which shows a nonlinear behaviour. From the set of measured data, it is then possible to predict the behaviour of extensive green roofs in real conditions and to determine the runoff coefficient C of the tested specimens. These data represent reference values for the subsequent design of sub-elements and structures of buildings.

Evaluating different machine learning methods to simulate runoff from extensive green roofs

Green roofs are increasingly popular measures to permanently reduce or delay storm-water runoff. The main objective of the study was to examine the potential of using machine learning (ML) to simulate runoff from green roofs to estimate their hydrological performance. Four machine learning methods, artificial neural network (ANN), M5 model tree, long short-term memory (LSTM) and k nearest neighbour (kNN), were applied to simulate storm-water runoff from 16 extensive green roofs located in four Norwegian cities across different climatic zones. The potential of these ML methods for estimating green roof retention was assessed by comparing their simulations with a proven conceptual retention model. Furthermore, the transferability of ML models between the different green roofs in the study was tested to investigate the potential of using ML models as a tool for planning and design purposes. The ML models yielded low volumetric errors that were comparable with the conceptual retention models, which indicates good performance in estimating annual retention. The ML models yielded satisfactory modelling results (NSE >0.5) in most of the roofs, which indicates an ability to estimate green roof detention. The variations in ML models' performance between the cities was larger than between the different configurations, which was attributed to the different climatic characteristics between the four cities. Transferred ML models between cities with similar rainfall events characteristics (Bergen–Sandnes, Trondheim–Oslo) could yield satisfactory modelling performance (Nash–Sutcliffe efficiency NSE >0.5 and percentage bias |PBIAS| <25 %) in most cases. However, we recommend the use of the conceptual retention model over the transferred ML models, to estimate the retention of new green roofs, as it gives more accurate volume estimates. Follow-up studies are needed to explore the potential of ML models in estimating detention from higher temporal resolution datasets.

The Effect of Plants on Extensive Green Roofs in Urban Heat Island Mitigation Efforts in Humid Tropical Cities

One of the urban heat island mitigation strategies in reducing urban temperatures in tropical cities is the application of a green roof system. This study compares the reduction in temperature and heat flow rate provided by three types of plants on extensive green roofs (EGR). We demonstrated that a EGR constructed with three types of plants (ground cover, and shrubs) could result in a decrease in temperature relative to the normal roof (NR). The results showed that the base temperature of the EGR of the bush and ground cover was lower than the base temperature of the NR which was 10.2ºC on indoor air, 17.8ºC on the inside and 19.1ºC on the outside. The peak indoor temperature was over 50ºC for the NR prototype. In the model with pennisetum purpureum schamach as the EGR, the maximum temperature was 40.1ºC, while for portulaca grandiflora and tradescantia spathacea the peaks were 37.6ºC and 37.5ºC, respectively. This shows that plants with large leaf widths are able to reduce heat greater than plants with small leaf widths.

Nutrient Leaching in Extensive Green Roof Substrate Layers With Different Configurations

Due to substrate layers with different substrate configurations, extensive green roofs (EGRs) exhibit different rainfall runoff retention and pollution interception effects. In the rainfall runoff scouring process, nutrient leaching often occurs in the substrate layer, which becomes a pollution source for rainwater runoff. In this study, six EGR devices with different substrate layer configurations were fabricated. Then, the cumulative leaching quantity (CLQ) and total leaching rate (TLR) of NH4+, TN and TP in the outflow of nine different depth simulated rainfall events under local rainfall characteristics were evaluated and recorded. Furthermore, the impact of different substrate configurations on the pollution interception effects of EGRs for rainfall runoff was studied. Results show that a mixed adsorption substrate in the EGR substrate layer has a more significant rainfall runoff pollution interception capacity than a single adsorption substrate. PVL and PVGL, as EGRs with layered configuration substrate layers, exhibited good NH4+-N interception capacity. The CLQ and TLR of NH4+-N for PVL and PVGL were -114.613 mg and -63.43%, -121.364 mg and -67.16%, respectively. Further, the addition of biochar as a modifier significantly slowed down the substrate layer TP leaching effect and improved the interception effect of NH4+-N and TN. Moreover, although polyacrylamide addition in the substrate layer aggravated the nitrogen leaching phenomenon in the EGRs outflow, but the granular structure substrate layer constructed by it exhibited a significantly inhibited TP leaching effect.

Evaluating different machine learning methods to simulate runoff from extensive green roofs

Green roofs are increasingly popular measures to permanently reduce or delay stormwater runoff. Conceptual and physically-based hydrological models are powerful tools to estimate their performance. However, physically-based models are associated with a high level of complexity and computation costs while parameters of conceptual models are more difficult to obtain when measurements are not available for calibration. The main objective of the study was to examine the potential of using machine learning (ML) to simulate runoff from green roofs to estimate their hydrological performance. Four machine learning methods, Artificial Neural Network (ANN), M5 Model tree, Long Short-Term Memory (LSTM) and k-Nearest Neighbour (kNN) were applied to simulate stormwater runoff from sixteen extensive green roofs located in four Norwegian cities across different climatic zones. The potential of these ML methods for estimating green roof retention was assessed by comparing their simulations with a proven conceptual retention model. Furthermore, the transferability of ML models between the different green roofs in the study was tested to investigate the potential of using ML models as a tool for planning and design purposes. The ML models yielded low volumetric errors that were comparable with the conceptual retention models, which indicates good performance in estimating annual retention. The ML models yielded satisfactory modelling results (NSE > 0.5) in both training and validation data which indicates an ability to estimate green roof detention. The variations in ML models' performance between the cities was larger than between the different configurations, which was attributed to the different climatic characteristics between the four cities. Transferred ML models between cities with similar rainfall events characteristics (Bergen–Sandnes, Trondheim–Oslo) could yield satisfactory modelling performance (NSE > 0.5, |PBIAS|