Low-cost monitoring systems for urban water management: Lessons from the field.

Sound urban water management relies on extensive and reliable monitoring of water infrastructure. As low-cost sensors and networks have become increasingly available for environmental monitoring, urban water researchers and practitioners must consider the benefits and disadvantages of such technologies. In this perspective paper, we highlight six technical and socio-technological considerations for low-cost monitoring technology to reach its full potential in the field of urban water management, including: technical barriers to implementation, complementarity with traditional sensing technologies, low-cost sensor reliability, added value of produced information, opportunities to democratize data collection, and economic and environmental costs of the technology. For each consideration, we present recent experiences from our own work and broader literature and identify future research needs to address current challenges. Our experience supports the strong potential of low-cost monitoring technology, in particular that it promotes extensive and innovative monitoring of urban water infrastructure. Future efforts should focus on more systematic documenting of experiences to lower barriers to designing, implementing, and testing of low-cost sensor networks, and on assessing the economic, social, and environmental costs and benefits of low-cost sensor deployments.

End-user perspective of low-cost sensors for urban stormwater monitoring: a review.

The large-scale deployment of low-cost monitoring systems has the potential to revolutionize the field of urban hydrology monitoring, bringing improved urban management, and a better living environment. Even though low-cost sensors emerged a few decades ago, versatile and cheap electronics like Arduino could give stormwater researchers a new opportunity to build their own monitoring systems to support their work. To find out sensors which are ready for low-cost stormwater monitoring systems, for the first time, we review the performance assessments of low-cost sensors for monitoring air humidity, wind speed, solar radiation, rainfall, water level, water flow, soil moisture, water pH, conductivity, turbidity, nitrogen, and phosphorus in a unified metrological framework considering numerous parameters. In general, as these low-cost sensors are not initially designed for scientific monitoring, there is extra work to make them suitable for in situ monitoring, to calibrate them, to validate their performance, and to connect them with open-source hardware for data transmission. We, therefore, call for international cooperation to develop uniform low-cost sensor production, interface, performance, calibration and system design, installation, and data validation guides which will greatly regulate and facilitate the sharing of experience and knowledge.

Practice makes the model: A critical review of stormwater green infrastructure modelling practice.

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.

Classifying pollutant flush signals in stormwater using functional data analysis on TSS MV curves.

Pollution levels in stormwater vary significantly during rain events, with pollutant flushes carrying a major fraction of an event pollutant load in a short period. Understanding these flushes is thus essential for stormwater management. However, current studies mainly focus on describing the first flush or are limited by predetermined flush categories. This study provides a new perspective on the topic by applying data-driven approaches to categorise Mass Volume (MV) curves for TSS into distinct classes of flush tailored to specific monitoring location. Functional Data Analysis (FDA) was used to investigate the dynamics of MV curves in two large data sets, consisting of 343 measured events and 915 modelled events, respectively. Potential links between classes of MV curves and combinations of rain characteristics were explored through a priori clustering. This yielded correct class assignments for 23-63% of the events using different combinations of MV curve clustering and rainfall characteristics. This suggests that while global rainfall characteristics influence flush, they are not sufficient as sole explanatory variables of different flush phenomena, and additional explanatory variables are needed to assign MV curves into classes with a predictive power that is suitable for e.g. design of stormwater control measures. Our results highlight the great potential of the FDA methodology as a new approach for classifying, describing, and understanding pollutant flush signals in stormwater.

SIPIBEL observatory: Data on usual pollutants (solids, organic matter, nutrients, ions) and micropollutants (pharmaceuticals, surfactants, metals), biological and ecotoxicity indicators in hospital and urban wastewater, in treated effluent and sludge from wastewater treatment plant, and in surface and groundwater.

The Bellecombe pilot site - SIPIBEL - was created in 2010 in order to study the characterisation, treatability and impacts of hospital effluents in an urban wastewater treatment plant. This pilot site is composed of: i) the Alpes Léman hospital (CHAL), opened in February 2012, ii) the Bellecombe wastewater treatment plant, with two separate treatment lines allowing to fully separate the hospital wastewater and the urban wastewater, and iii) the Arve River as the receiving water body and a tributary of the Rhône River and the Geneva aquifer. The database includes in total 48 439 values measured on 961 samples (raw and treated hospital and urban wastewater, activated sludge in aeration tanks, dried sludge after dewatering, river and groundwater, and a few additional campaigns in aerobic and anaerobic sewers) with 44 455 physico-chemistry values (including 15 pharmaceuticals and 14 related transformation products, biocides compounds, metals, organic micropollutants), 2 193 bioassay values (ecotoxicity), 1 679 microbiology values (including microorganisms and antibioresistance indicators) and 112 hydrobiology values.

Modelling daily and hourly loads of pharmaceuticals in urban wastewater.

Pharmaceuticals are known contaminants of the environment. Assessing and managing the risk associated to this contamination has become an important field of study in environmental sciences. Accurately sampling and measuring pharmaceuticals concentrations in wastewater or in the environment is still costly and difficult. Thus only a few studies have looked at the temporal variability of the concentrations. In parallel, models have been proposed to predict the occurrence of pharmaceuticals. They usually assume that the loads of pharmaceuticals entering a wastewater treatment plant (WWTP) are proportional to the pharmaceuticals sales. However, most of the time, the results are difficult to interpret. The main problem with those models is the lack of data at WWTP. In this context, a stochastic model predicting daily and hourly loads is proposed to better understand the processes influencing the occurrence of pharmaceuticals in wastewater. Using accurate statistical pharmaceutical sales data, the mass consumed daily in the catchment is randomly picked from empirical probability density functions. Then, patients are randomly generated until the picked total mass is reached. For each patient, posology, metabolism and excretion rates of pharmaceuticals to the sewer system are randomly generated according to relevant literature data. In particular, time-use data are used to generate the time-use of patients to pick the times when patients consume pharmaceuticals and also when they excrete them in toilets. Applied to 9 molecules in a French catchment, results show that the model accurately and reliably predicts both the daily and hourly loads of pharmaceuticals at the inlet of the WWTP.

Performance and uncertainties of TSS stormwater sampling strategies from online time series.

The event mean concentrations (EMCs) that would have been obtained by four different stormwater sampling strategies are simulated by using total suspended solids (TSS) and flowrate time series (about one minute time-step and one year of data). These EMCs are compared to the reference EMCs calculated by considering the complete time series. The sampling strategies are assessed with datasets from four catchments: (i) Berlin, Germany, combined sewer overflow (CSO); (ii) Graz, Austria, CSO; (iii) Chassieu, France, separate sewer system; and (iv) Ecully, France, CSO. A sampling strategy in which samples are collected at constant time intervals over the rainfall event and sampling volumes are pre-set as proportional to the runoff volume discharged between two consecutive sample leads to the most representative results. Recommended sampling time intervals are of 5 min for Berlin and Chassieu (resp. 100 and 185 ha area) and 10 min for Graz and Ecully (resp. 335 and 245 ha area), with relative sampling errors between 7% and 20% and uncertainties in sampling errors of about 5%. Uncertainties related to sampling volumes, TSS laboratory analyses and beginning/ending of rainstorm events are reported as the most influent sources in the uncertainties of sampling errors and EMCs.

Revisiting conceptual stormwater quality models by reconstructing virtual state variables.

This work proposes a Bayesian non-informative reconstruction of virtual state variables in the representation of stormwater total suspended solids pollutographs by the traditional wash-off models, based on 255 rainfall events measured in a 185 ha French urban catchment. Results from event-based analyses revealed the missing representation of an essential process in the traditional rating curve (RC) model (simplest wash-off model) for 56% of the rainfall events. The unsatisfactory performances of the RC model are found to be not necessarily linked to antecedent dry weather conditions, as assumed by a great number of accumulation/wash-off models. Statistical tests suggest that non-representable rainfall events by the RC model are randomly distributed in time. The proposed Bayesian reconstructions of a potential process missed by the RC model exhibit a suitable identifiability at an intra-event scale. However, these reconstructions are not interpretable from the traditional accumulation/wash-off notions, i.e. in terms of a unique state of virtual available mass over the catchment that is decreasing over time, due to their high unrepeatability regarding their shape and their low prediction capacity for other rainfall events.

Evaluation of two statistical approaches for estimating pollutant loads at adjacent combined sewer overflow structures.

Quantifying pollutant loads from combined sewer overflows (CSOs) is necessary for assessing impacts of urban drainage on receiving water bodies. Based on data obtained at three adjacent CSO structures in the Louis Fargue catchment in Bordeaux, France, this study implements multiple linear regression (MLR) and random forest regression (RFR) approaches to develop statistical models for estimating emitted loads of total suspended solids (TSS). Comparison between hierarchical clustering selection and random selection of CSO events for model calibration is included in model development. The results indicate that selection of the model's explanatory variables depends on both the type of approach and the CSO structure. By using the cluster technique to select representative events for model calibration, model predictability is generally improved. For the available dataset, MLR may have advantages over RFR in terms of verification performance and lower range of error due to splitting events for calibration and verification. But RFR model uncertainty bands are considerably narrower than the MLR ones.

Assessment of sampling strategies for estimation of site mean concentrations of stormwater pollutants.

The estimation of stormwater pollutant concentrations is a primary requirement of integrated urban water management. In order to determine effective sampling strategies for estimating pollutant concentrations, data from extensive field measurements at seven different catchments was used. At all sites, 1-min resolution continuous flow measurements, as well as flow-weighted samples, were taken and analysed for total suspend solids (TSS), total nitrogen (TN) and Escherichia coli (E. coli). For each of these parameters, the data was used to calculate the Event Mean Concentrations (EMCs) for each event. The measured Site Mean Concentrations (SMCs) were taken as the volume-weighted average of these EMCs for each parameter, at each site. 17 different sampling strategies, including random and fixed strategies were tested to estimate SMCs, which were compared with the measured SMCs. The ratios of estimated/measured SMCs were further analysed to determine the most effective sampling strategies. Results indicate that the random sampling strategies were the most promising method in reproducing SMCs for TSS and TN, while some fixed sampling strategies were better for estimating the SMC of E. coli. The differences in taking one, two or three random samples were small (up to 20% for TSS, and 10% for TN and E. coli), indicating that there is little benefit in investing in collection of more than one sample per event if attempting to estimate the SMC through monitoring of multiple events. It was estimated that an average of 27 events across the studied catchments are needed for characterising SMCs of TSS with a 90% confidence interval (CI) width of 1.0, followed by E.coli (average 12 events) and TN (average 11 events). The coefficient of variation of pollutant concentrations was linearly and significantly correlated to the 90% confidence interval ratio of the estimated/measured SMCs (R2 = 0.49; P < 0.01) as well as the number of events required to achieve certain accuracy, and hence could be a promising surrogate for determining the sampling frequency needed to accurately estimate SMCs of pollutants.

The SIPIBEL project: treatment of hospital and urban wastewater in a conventional urban wastewater treatment plant.

Hospital wastewater (HWW) receives increasing attention because of its specific composition and higher concentrations of some micropollutants. Better knowledge of HWW is needed in order to improve management strategies and to ensure the preservation of wastewater treatment efficiency and freshwater ecosystems. This context pushed forward the development of a pilot study site named Site Pilote de Bellecombe (SIPIBEL), which collects and treats HWW separately from urban wastewater, applying the same conventional treatment process. This particular configuration offers the opportunity for various scientific investigations. It enables to compare hospital and urban wastewater, the efficiency of the two parallel treatment lines, and the composition of the resulting hospital and urban treated effluents, as well as the evaluation of their effects on the environment. The study site takes into account environmental, economic, and social issues and promotes scientific and technical multidisciplinary actions. ᅟ.

Calibration of UV/Vis spectrophotometers: A review and comparison of different methods to estimate TSS and total and dissolved COD concentrations in sewers, WWTPs and rivers.

UV/Vis spectrophotometers have been used for one decade to monitor water quality in various locations: sewers, rivers, wastewater treatment plants (WWTPs), tap water networks, etc. Resulting equivalent concentrations of interest can be estimated by three ways: i) by manufacturer global calibration; ii) by local calibration based on the provided global calibration and grab sampling; iii) by advanced calibration looking for relations between UV/Vis spectra and corresponding concentrations from grab sampling. However, no study has compared the applied methods so far. This collaborative work presents a comparison between five different methods. A Linear Regression (LR), Support Vector Machine (SVM), EVOlutionary algorithm method (EVO) and Partial Least Squares (PLS) have been applied on various data sets (sewers, rivers, WWTPs under dry, wet and all weather conditions) and for three water quality parameters: TSS, COD total and dissolved. Two criteria (r(2) and Root Mean Square Error RMSE) have been calculated - on calibration and verification data subsets - to evaluate accuracy and robustness of the applied methods. Values of criteria have then been statistically analysed for all and separated data sets. Non-consistent outcomes come through this study. According to the Kruskal-Wallis test and RMSEs, PLS and SVM seem to be the best methods. According to uncertainties in laboratory analysis and ranking of methods, LR and EVO appear more robust and sustainable for concentration estimations. Conclusions are mostly independent of water matrices, weather conditions or concentrations investigated.

Influence of sampling intake position on suspended solid measurements in sewers: two probability/time-series-based approaches.

Total suspended solid (TSS) measurements in urban drainage systems are required for several reasons. Aiming to assess uncertainties in the mean TSS concentration due to the influence of sampling intake vertical position and vertical concentration gradients in a sewer pipe, two methods are proposed: a simplified method based on a theoretical vertical concentration profile (SM) and a time series grouping method (TSM). SM is based on flow rate and water depth time series. TSM requires additional TSS time series as input data. All time series are from the Chassieu urban catchment in Lyon, France (time series from 2007 with 2-min time step, 89 rainfall events). The probability of measuring a TSS value lower than the mean TSS along the vertical cross section (TSS underestimation) is about 0.88 with SM and about 0.64 with TSM. TSM shows more realistic TSS underestimation values (about 39 %) than SM (about 269 %). Interquartile ranges (IQR) over the probability values indicate that SM is more uncertain (IQR = 0.08) than TSM (IQR = 0.02). Differences between the two methods are mainly due to simplifications in SM (absence of TSS measurements). SM assumes a significant asymmetry of the TSS concentration profile along the vertical axis in the cross section. This is compatible with the distribution of TSS measurements found in the TSM approach. The methods provide insights towards an indicator of the measurement performance and representativeness for a TSS sampling protocol.

Using a Bayesian approach to improve and calibrate a dynamic model of polycyclic aromatic hydrocarbons degradation in an industrial contaminated soil.

A novel kinetics model that describes the dynamics of polycyclic aromatic hydrocarbons (PAHs) in contaminated soils is presented. The model includes two typical biodegradation pathways: the co-metabolic pathway using pseudo first order kinetics and the specific biodegradation pathway modeled using Monod kinetics. The sorption of PAHs to the solid soil occurs through bi-phasic fist order kinetics, and two types of non-extractible bounded residues are considered: the biogenic and the physically sequestrated into soil matrix. The PAH model was developed in Matlab, parameterized and tested successfully on batch experimental data using a Bayesian approach (DREAM). Preliminary results led to significant model simplifications. They also highlighted that the specific biodegradation pathway was the most efficient at explaining experimental data, as would be expected for an old industrial contaminated soil. Global analysis of sensitivity showed that the amount of PAHs ultimately degraded was mostly governed by physicochemical interactions rather than by biological activity.

Influence of local calibration on the quality of online wet weather discharge monitoring: feedback from five international case studies.

This paper reports about experiences gathered from five online monitoring campaigns in the sewer systems of Berlin (Germany), Graz (Austria), Lyon (France) and Bogota (Colombia) using ultraviolet-visible (UV-VIS) spectrometers and turbidimeters. Online probes are useful for the measurement of highly dynamic processes, e.g. combined sewer overflows (CSO), storm events, and river impacts. The influence of local calibration on the quality of online chemical oxygen demand (COD) measurements of wet weather discharges has been assessed. Results underline the need to establish local calibration functions for both UV-VIS spectrometers and turbidimeters. It is suggested that practitioners calibrate locally their probes using at least 15-20 samples. However, these samples should be collected over several events and cover most of the natural variability of the measured concentration. For this reason, the use of automatic peristaltic samplers in parallel to online monitoring is recommended with short representative sampling campaigns during wet weather discharges. Using reliable calibration functions, COD loads of CSO and storm events can be estimated with a relative uncertainty of approximately 20%. If no local calibration is established, concentrations and loads are estimated with a high error rate, questioning the reliability and meaning of the online measurement. Similar results have been obtained for total suspended solids measurements.

Simplified hydraulic model of French vertical-flow constructed wetlands.

Designing vertical-flow constructed wetlands (VFCWs) to treat both rain events and dry weather flow is a complex task due to the stochastic nature of rain events. Dynamic models can help to improve design, but they usually prove difficult to handle for designers. This study focuses on the development of a simplified hydraulic model of French VFCWs using an empirical infiltration coefficient--infiltration capacity parameter (ICP). The model was fitted using 60-second-step data collected on two experimental French VFCW systems and compared with Hydrus 1D software. The model revealed a season-by-season evolution of the ICP that could be explained by the mechanical role of reeds. This simplified model makes it possible to define time-course shifts in ponding time and outlet flows. As ponding time hinders oxygen renewal, thus impacting nitrification and organic matter degradation, ponding time limits can be used to fix a reliable design when treating both dry and rain events.

Computational fluid dynamics modelling of flow and particulate contaminants sedimentation in an urban stormwater detention and settling basin.

Sedimentation is a common but complex phenomenon in the urban drainage system. The settling mechanisms involved in detention basins are still not well understood. The lack of knowledge on sediment transport and settling processes in actual detention basins is still an obstacle to the optimization of the design and the management of the stormwater detention basins. In order to well understand the sedimentation processes, in this paper, a new boundary condition as an attempt to represent the sedimentation processes based on particle tracking approach is presented. The proposed boundary condition is based on the assumption that the flow turbulent kinetic energy near the bottom plays an important role on the sedimentation processes. The simulated results show that the proposed boundary condition appears as a potential capability to identify the preferential sediment zones and to predict the trapping efficiency of the basin during storm events.

Input variable selection and calibration data selection for storm water quality regression models.

Storm water quality models are useful tools in storm water management. Interest has been growing in analyzing existing data for developing models for urban storm water quality evaluations. It is important to select appropriate model inputs when many candidate explanatory variables are available. Model calibration and verification are essential steps in any storm water quality modeling. This study investigates input variable selection and calibration data selection in storm water quality regression models. The two selection problems are mutually interacted. A procedure is developed in order to fulfil the two selection tasks in order. The procedure firstly selects model input variables using a cross validation method. An appropriate number of variables are identified as model inputs to ensure that a model is neither overfitted nor underfitted. Based on the model input selection results, calibration data selection is studied. Uncertainty of model performances due to calibration data selection is investigated with a random selection method. An approach using the cluster method is applied in order to enhance model calibration practice based on the principle of selecting representative data for calibration. The comparison between results from the cluster selection method and random selection shows that the former can significantly improve performances of calibrated models. It is found that the information content in calibration data is important in addition to the size of calibration data.

Accuracy of different sensors for the estimation of pollutant concentrations (total suspended solids, total and dissolved chemical oxygen demand) in wastewater and stormwater.

Many field investigations have used continuous sensors (turbidimeters and/or ultraviolet (UV)-visible spectrophotometers) to estimate with a short time step pollutant concentrations in sewer systems. Few, if any, publications compare the performance of various sensors for the same set of samples. Different surrogate sensors (turbidity sensors, UV-visible spectrophotometer, pH meter, conductivity meter and microwave sensor) were tested to link concentrations of total suspended solids (TSS), total and dissolved chemical oxygen demand (COD), and sensors' outputs. In the combined sewer at the inlet of a wastewater treatment plant, 94 samples were collected during dry weather, 44 samples were collected during wet weather, and 165 samples were collected under both dry and wet weather conditions. From these samples, triplicate standard laboratory analyses were performed and corresponding sensors outputs were recorded. Two outlier detection methods were developed, based, respectively, on the Mahalanobis and Euclidean distances. Several hundred regression models were tested, and the best ones (according to the root mean square error criterion) are presented in order of decreasing performance. No sensor appears as the best one for all three investigated pollutants.