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.
Green infrastructures (GIs) have in recent decades emerged as sustainable technologies for urban stormwater management, and numerous studies have been conducted to develop and improve hydrological models for GIs. This review aims to assess current practice in GI hydrological modelling, encompassing the selection of model structure, equations, model parametrization and testing, uncertainty analysis, sensitivity analysis, the selection of objective functions for model calibration, and the interpretation of modelling results. During a quantitative and qualitative analysis, based on a paper analysis methodology applied across a sample of 270 published studies, we found that the authors of GI modelling studies generally fail to justify their modelling choices and their alignments between modelling objectives and methods. Some practices, such as uncertainty analysis, were also found to be limited, despite their necessity being widely acknowledged by the scientific community and their application in other fields. In order to improve current GI modelling practice, the authors suggest the following: i) a framework, called STAMP, designed to promote the standardisation of the documentation of GI modelling studies, and ii) improvements in modelling tools for facilitating good practices, iii) the sharing of data for better model testing, iv) the evaluation of the suitability of hydrological equations for GI application, v) the publication of clear statements regarding model limitations and negative results.
8 June 2022 in Hydrology and Earth System Sciences
A strategy to evaluate the suitability of different multiplicative random cascades to produce rainfall time series, taking into account climate change, inputs for green infrastructures models. The multiplicative random cascades reproduce a (multi)fractal distribution of precipitation through an iterative and multiplicative random process. In the current study, the initial model, a flexible cascade that deviates from multifractal scale invariance, was improved with (i) a temperature dependency and (ii) an additional function to reproduce the temporal structure of rainfall. The structure of the models with depth and temperature dependency was found to be applicable in eight locations studied across Norway and France. The resulting time series from both reference period and projection based on RCP 8.5 were applied to two green roofs with different properties. The different models led to a slight change in the performance of green roofs, but this was not significant compared to the range of outcomes due to ensemble uncertainty in climate modelling and the stochastic uncertainty due to the nature of the process. The hydrological dampening effect of the green infrastructure was found to decrease in most of the Norwegian cities due to an increase in precipitation, especially Bergen (Norway), while slightly increasing in Marseille (France) due to decrease in rainfall event frequency.
28 September 2023 in Journal of Environmental Management
Historically, green infrastructure for stormwater management has been event-based designed. This study aims to realign the green infrastructure design strategies with principles for robust decision making, through the example of green roofs design with the variational method and exemplified using the Norwegian context of the 3-step approach (3SA) for stormwater management. The 3SA consists of planning solutions to handle day-to-day rain at site scale through infiltration (step 1) and detention (step 2), and extreme events with safe floodways (step 3). An innovative framework based on downscaling of rainfall timeseries is suggested as follows: (i) long duration continuous simulation for retention variation and day-to-day discharge, corresponding to step 1 in the 3SA; (ii) intensive sampling of local extreme events to estimate reliability and robustness of solutions, corresponding to steps 2 and 3 in the 3SA. Comparing the traditional variational method to Highly-Informed-Design-Evaluation-Strategy (HIDES), it was found that the variational method possibly leads to incorrect decisions while the suggested novel approach was found to give more informed and reliable results by suggesting a design based on both operating mode and failure mode. It allows to embed solutions within the urban water system by facilitating the link between the steps of the 3SA. Such a framework was found to be data-wise applicable in the Norwegian context.
In this study, we show that pollutants of emerging concern are, by nature, prone to the emergence of epistemic uncertainty. We also show that the current uncertainty quantification methods used for pollutant modelling rely almost exclusively on parameter uncertainty, which is not adequate to tackle epistemic uncertainty affecting the model structure. We, therefore, suggest a paradigm shift in the current pollutant modelling approaches by adding a term explicitly accounting for epistemic uncertainties. In a proof-of-concept, we use this approach to investigate the impact of epistemic uncertainty in the fluctuation of pollutants during wet-weather discharge (input information) on the distribution of mass of pollutants (output distributions). We found that the range of variability negatively impacts the tail of output distributions. The fluctuation time, associated with high covariance between discharge and concentration, is a major driver for the output distributions. Adapting to different levels of epistemic uncertainty, our approach helps to identify critical unknown information in the fluctuation of pollutant concentration. Such information can be used in a risk management context and to design smart monitoring campaigns.