Les ouvrages de biofiltration, systèmes végétalisés conçus pour le stockage, la filtration et éventuellement l’infiltration de l’eau, peuvent contribuer à une meilleure maîtrise du flux polluant associé au ruissellement de voirie. La capacité de ce type de système à réduire les flux de macropolluants, tels que les matières en suspension (MES) et les nutriments, via une amélioration de la qualité de l’eau combinée à une réduction de son volume, est avérée. Cependant, le comportement de micropolluants dans ces systèmes reste relativement peu documenté. Ce travail de thèse s’intéresse donc à la rétention et le devenir de micropolluants dans un système de biofiltration des eaux de voirie. Il s’articule principalement autour du suivi in situ d’un panel diversifié de micropolluants caractéristiques des eaux de voirie (éléments traces métalliques ou ETM, hydrocarbures totaux, hydrocarbures aromatiques polycycliques ou HAP, BPA, alkylphénols, phatalates) et d’indicateurs globaux dans un accotement filtrant et une noue filtrante, situés en bordure d’une route départementale. Ce suivi comporte des mesures débit métriques et de qualité d’eau en continu, un travail d’échantillonnage et d’analyse du ruissellement brut et des eaux traitées par les deux ouvrages (dix-neuf événements pluvieux étudiés au cours d’un an et demi) et un travail d’échantillonnage et d’analyse du sol. Cette démarche est couplée à une caractérisation du substrat filtrant et des matériaux de construction en laboratoire et à une approche de modélisation stochastique pour évaluer le bilan de masse annuel de polluants dans la noue filtrante. Ce suivi démontre la capacité de la biofiltration à réduire de façon significative les concentrations totales en micropolluants à l’échelle de l’événement pluvial. Le traitement est particulièrement efficace pour les contaminants associés majoritairement aux MES, tels que le Zn, le Pb et les HAP pour lesquels la réduction médiane des concentrations événementielles (EC50) est supérieure à 90% dans les deux ouvrages. Pour le Cu, le Cr, le Ni et l’OP, les EC50 sont bonnes, supérieures à 70%. L’efficacité est moindre et plus variable pour les autres micropolluants organiques. Trois événements, caractérisés par une performance dégradée vis-à-vis des MES et des polluants particulaires, ont néanmoins été observés lors de l’application du sel de déverglaçage en hiver. La nature des particules indique que ce comportement soit dû à une mauvaise filtration de particules issues de la route, probablement liée à l’abondance exceptionnelle de particules fines (<10 µm) dans les eaux de voirie pendant cette période, combiné à une fissuration du substrat filtrant, menant à des écoulements préférentiels. La rétention de la phase dissoute des micropolluants est généralement moins efficace que celle de la phase particulaire ; des concentrations élevées ont notamment été observées pour certains micropolluants dissous (BPA, alkylphénols, phtalates) en sortie de la noue filtrante pendant les premiers mois de fonctionnement. Le transport des ETM dissous semble être facilité par leur association avec le carbone organique dissous ; ils sont aussi susceptibles d’être lixiviés à partir du sol contaminé ou des particules issues de la route. La rétention des micropolluants organiques dissous est limitée essentiellement par la contamination du substrat filtrant, dont l’origine est soit antérieure à son installation dans le biofiltre (cas des HAP), soit liée à des émissions de polluants depuis les matériaux de construction (cas probable pour BPA, OP, NP, DEHP).L’établissement du bilan de masse à l’échelle annuelle montre que l’abattement du flux polluant annuel est plus faible que l’EC50, du fait d’une surverse fréquente de l’ouvrage. Il met aussi en évidence l’ampleur des émissions de certains micropolluants organiques depuis les matériaux de construction de l’ouvrage, qui dépassent largement le flux polluant intercepté au cours de la première année / Biofiltration is a technique based on vegetated systems for the temporary storage, filtration and possible infiltration of runoff water close to the source, which may be used to manage pollutant loads associated with road runoff. The ability of this type of system to reduce loads of macropollutants, such as suspended solids and nutrients, both through the improvement of water quality and the reduction of its volume, has been established previously. However, the behavior of micropollutants in these systems remains relatively uncharacterized. The objective of the present work is to evaluate the retention and fate of micropollutants in a biofiltration system treating road runoff. It is principally focused on the in situ monitoring of a wide range of micropolltuants typical of road runoff (trace metals or TMs, total petroleum hydrocarbons, polycyclic aromatic hydrocarbons or PAHs, alkylphenols, BPA, phtalates) and global water quality parameters in two biofiltration systems located beside a highway: a vegetative filter strip and a biofiltration swale. This work involves continuous flow and water quality measurements, the sampling and analysis of runoff and infiltrated water drained from both systems (nineteen rain events over a year-and-a-half period) as well as the sampling and analysis of soil. These results are combined with a laboratory characterization of the filter media and biofilter construction materials and a stochastic modeling approach used to evaluate the annual mass balance of pollutants in the biofiltration swale. Field results highlight the ability of biofiltration to significantly reduce total concentrations of micropollutants at the event scale. Treatment is particularly effective for contaminants mainly associated with suspended solids, such as Zn, Pb and PAHs, for which median concentration reductions (EC50) exceeded 90% in both systems. EC50 observed for Cu, Cr, Ni and OP are also quite good, exceeding 70%. Treatment efficiency is lower and more variable for the other organic micropollutants. Three events, characterized by a degraded performance with respect to suspended solids and particulate pollutants, were, however, observed during a winter period when deicing salt was applied to road surfaces. Particle characteristics indicate that this behavior was due to the poor filtration of road-originated particles, probably due to an exceptional abundance of fine particles (<10 µm) in the road runoff during this period, in combination with the formation of preferential flows in the filter media due to cracking. The retention of dissolved-phase micropollutants is generally less effective than that of the particulate phase; in particular, elevated dissolved concentrations of several micropollutants (BPA, alkylphenols, phthalates) were observed during the first months of operation of the biofiltration swale. The transport of dissolved TMs appears to be facilitated by their association with dissolved organic carbon. These contaminants may also be leached from contaminated soil and road sediments. The retention of dissolved organic micropollutants is essentially limited by filter media contamination, which predates installation in the biofilter for PAHs and is attributed to pollutant emissions from construction materials for BPA, OP, NP and DEHP.In order to evaluate the systems’s ability to reduce pollutant loads and characterize the fate of pollutants in the filter media, a mass balance is evaluated at the annual scale for a selection of micropollutants. This work reveals that the reduction of pollutant loads is generally lower than the EC50 observed at the event scale due to the frequent overflow of water from the device. It also shows the significance of organic micropollutant emissions from biofiltration swale construction materials, which were much greater than the pollutant loads intercepted over the first year of operation

This thesis utilizes a distributed hydrologic model to predict hydrologic effects of Low Impact Development (LID), and also analyzes runoff from small sub-areas within the watershed. City planners and developers rely on accurate hydrologic models, which enable them to design flood-proof developments and effectively mitigate flooding downstream. Common hydrologic models use a lumped approach, which averages the physical characteristics of basins for model calculations, limiting their ability to estimate runoff within the basin. In contrast, distributed hydrologic models, which divide the watershed into a grid system, can be used to predict runoff at any location within the watershed. The fully distributed hydrologic model, VfloTM, is used to model stormwater runoff in The Woodlands, TX watershed, and to demonstrate the effectiveness of the master planned community. This thesis also suggests that a calibrated VfloTM model can accurately predict stormwater runoff from small sub-areas within a watershed.

Land development transforms the natural landscape and impacts in stream ecosystems and downstream communities as it alters the natural flow regime. An increase in impervious areas results in higher volumes of storm water runoff, reduced time to peak, and more frequent flooding. Best Management Practices (BMP) and Low Impact development (LID) are a few of the set of measures which are used to mitigate the impact of urbanization. Peak flow, runoff volume are few of the conventional metrics which are used to evaluate the impact and performance of these storm water management strategies on the watershed. BMP are majorly used to control the flood runoff but results in the release of large volumes of runoff even after the flood wave passed the reach and LIDs are used to replicate the natural flow regime by controlling the runoff at the source. Therefore need to incorporate a metric which includes the timing and area being inundated needs to be considered to study the impact of these strategies on the downstream. My proposed research will focus on simulating the Low Impact Development (LID) techniques like permeable pavements and rainwater harvesting on an urbanized watershed using a curve number approach to quantify the hydrologic performance of these strategies on the watershed. LID, BMPs, and combined strategies are introduced for retrofitting existing conditions and their hydrologic performance is accessed based on the peak flow and a new metric Hydrologic Footprint Residence. A simulation optimization framework would be developed which identifies cost effective LID options that maximize the reduction of peak flow from the existing condition design storms while meeting budget restrictions. Further LID and BMP placement is included in the optimization model to study the impact of the combined scenario on the storm water management plans and their performance based on different storms and corresponding budget. Therefore a tradeoff can be illustrated between the implementation cost and the hydrological impact on the watershed based on the storm water management approach of using only LID and combination of LID and BMP corresponding to varied spectrum of design storm events.

Urbanization of watersheds leads to the degradation of watershed health, as increased areas of imperviousness produce alterations in the flow regime of receiving water bodies. While centralized infrastructure improvements, such as detention ponds, are typically implemented to manage excess runoff, a more decentralized approach that utilizes Low Impact Development (LID) design principles may better preserve the predevelopment flow regime. Peak flow is traditionally used to design both of these types of infrastructure, but this does not capture the changes in the flow regime, nor does it convey the importance of stormwater sustainability to the general public. To further the general public's understanding about stormwater sustainability, an educational tool was used to take a complicated issue and make it easier to understand by a layperson. The first purpose of this work was to explore the effectiveness of educational tools that may be developed to increase public awareness about issues of watershed sustainability and encourage adoption of sustainable stormwater controls. To increase knowledge about stormwater sustainability and encourage more sustainable practices, a new stormwater sustainability metric, the hydrologic footprint residence (HFR), was recently introduced to measure more holistically the impacts of urbanization on the downstream residence. HFR measures changes to the flow regime as the area of land inundated for one unit of time in response to one rainfall event, which is a more relatable metric than peak flow for the general public. It was the second purpose of this work is to explore the effectiveness of HFR in communicating the impacts of urbanization on watershed health, as compared to traditional stormwater metrics, such as peak flow. To test these different objectives, collaboration with the Communication and the Computer Engineering Departments at Texas A&M University was needed to create a survey, which helped evaluate the effectiveness of the educational tool in educating the general public about stormwater sustainability, and encouraging more sustainable practices. The survey was also used to evaluate and compare the use of HFR and peak flow within the quiz for communicating to the general public about stormwater sustainability. Results indicated the quiz was useful for educating the public about stormwater sustainability, encouraging more sustainable practices. In addition, results indicated the HFR was more effective than peak flow in educating the public about LIDs.

This research estimates nitrogen removal from stormwater runoff using a denitrifying bioretention system using the USEPA Storm Water Management Model Version 5 (SWMM-5). SWMM-5 has been used to help planners make better decisions since its development in 1971. A conventional bioretention system is a type of Low Impact Development (LID) technology, which designed without a media layer specifically for achieving nitrogen removal. More recently studies have showed that high TN removal efficiencies are possible when incorporating a denitrification media layer. These systems are known as denitrifying bioretention systems, or alternative bioretention systems. LID projects are currently being designed and developed in Sarasota County, Florida. These projects include a bioretention cell retrofit project on Venice East Blvd., in Venice, FL where thirteen bioretention cells will be developed. Although implementation of LID has already begun in southwest Florida, little research exists on whether these systems are effective at reducing non-point sources of nutrients. Therefore, the overall goal of this research project was to investigate the performance of a proposed bioretention system in Venice, FL to treat non-point sources of nitrogen from stormwater runoff. An alternative bioretention cell (ABC) model was designed to conceptually address water routing through a layered bioretention cell by separating the model into treatment layers- the layers where the nitrification and denitrification reactions are expected to occur within an alternative bioretention system (i.e., nitrification is assumed to occur in the sand media layer, and denitrification in the wood chip media layer). The bioretention cell configuration was based largely on the development plans provided by Sarasota County; however, the configuration incorporated the same electron donor media for denitrification that was used in a prior study (i.e., wood chips). Site-specific input parameters needed to calibrate the ABC model were obtained from laboratory analyses, the literature, and the US Geological website (websoilsurvey.nrcs.usda.gov). Using a mass balance approach, and the hydraulic residence time (HRT) values from the results of a previous study, first-order loss rate coefficients for both nitrification and denitrification (k1 and k2, respectively) were estimated. The rate coefficients were then used to develop treatment expression for nitrification and denitrification reactions. The treatment expressions were used to estimate the annual load reductions for TKN, NO3--N, and TN at the Venice East Blvd. bioretention retrofit site. Six storm events were simulated using a range of nitrogen concentrations. The simulation results showed minimal nitrification removal rates for storm events exceeding 1 inch, due to the planned bioretention system area being only 1% of the subcatchment area. A new ABC model was created (based on EPA bioretention cell sizing guidelines), to be 6% of the subcatchment area. Both systems were used to estimate TN removal efficiencies. The larger sized ABC model results showed average TKN, NO3--N and TN reductions of 84%, 96%; and 87%, respectively; these are comparable to results from similar studies. Results indicate that adequate nitrogen attenuation is achievable in the alternative bioretention system, if it is sized according to EPA sizing guidelines (5-7%).

Preliminary reconnaissance undertaken in summer 2013 identified the scale of stormwater management issues in Xiamen, having frequent storm events that overwhelm the stormwater and sewer infrastructure resulting in widespread flooding. This research explored the role that green facilities play in addressing stormwater issues through the inquiry of Low Impact Development strategies and techniques. From a long-term perspective, green infrastructure planning and implementation is inevitably linked with strong education programs, rational stormwater codes and regulations, a variety of financing and incentives, as well as an integrated and competent administration system. This research presents a design proposal for green infrastructure retrofit for a selected study block in the central area of Xiamen to help guide water sensitive urban design and development in the future. Seven recommendations based on the synthesis of the literature review, key-informant interviews, built-project studies and the retrofit design proposal are proposed. This document will be submitted to Xiamen Urban Planning & Design Institute for considerations to be integrated in city master planning policy and zoning codes and standards as needed and to inform a demonstration project to help advance long-term strategies and recommendations. / October 2015

Stormwater management is an essential aspect of urban hydrology. Urbanized areas have large amounts of impervious surface cover (ISC) and well developed sewer and drainage networks which rapidly channel water and pollutants off of streets and into local streams. This research evaluates the use of vegetated roofs as mechanisms to reduce ISC and stormwater runoff in downtown Atlanta. A 3-D model of the study site was created so that runoff rates could be measured for various rooftop scenarios under different size storm events. The results revealed a reduction in peak runoff and an increase in both the lag time and duration of response time. The results were most significant for the smallest storm event with 2/3 of the rooftops vegetated. As these experiments use a scale model for a section of downtown Atlanta, results are likely to be applicable to similar urban environments and may provide guidance for stormwater engineers.

Bioretention gardens are stormwater management practices that offer numerous water quantity and quality benefits. However, previous studies have reported inconsistent removal of nitrogen and phosphorous in these systems. The first phase of this research involved the construction and monitoring of ten vegetated, mesoscale, bioretention cells in a field setting to provide a comparison of the performance of five alternative designs intended to provide nutrient removal. Results indicated that concentrations of total nitrogen and total phosphorous may be reduced by up to 53 and 79%, respectively, in specially designed bioretention gardens. In the second phase of the research, a GIS-based site selection tool was used to identify areas suitable for bioretention implementation based on physical site requirements. Applying this tool to selected urban catchments demonstrated that bioretention gardens may be integrated into existing urban landscapes on a scale large enough to accommodate runoff and associated nutrient loads from small (<15mm) storms.

Master of Landscape Architecture / Department of Landscape Architecture/Regional and Community Planning / Richard H. Forsyth / Low Impact Development (LID) is an emerging ecological method for planning and designing stormwater management. The main goal of LID is that post development hydrology mimic predevelopment hydrology or that of the natural land pattern. The difference between LID and conventional end of pipe engineering is that LID is designed for on-site infiltration. LID was developed in Maryland and has since been adapted to other ecoregions. The introduction and background offer a review of the importance of ecological stormwater management and the basic principles and techniques of LID. Ecological stormwater management involves understanding and applying ecoregion factors such as: climate, physiography/ landform, geology, soils, and plant associations. This thesis begins with a brief overview of the ecoregional factors associated with the Topeka, Kansas (KS) area. Two case studies of LID type stormwater structures are investigated. The case studies, Jackson Street and Hillcrest, are projects of the City of Topeka Water Pollution Control Division (Topeka, KS). The primary topics investigated include: background, political approval process, condition of site before implementation, general project design, stormwater design, soil, plants, LID principles and techniques used, lessons learned, and a critique of each project. Investigation methods include: personal communication with key individuals involved, and a review of site visits, drawings, specifications, and other projects documents.

A urbanização acelerada das últimas décadas aliada à ocupação desordenada e intensa impermeabilização do solo, têm agravado as cheias nas cidades brasileiras. Na cidade de São Paulo, recentemente, os períodos chuvosos, entre os meses de outubro e março, de 2009-2010 e 2010-2011 foram especialmente relevantes, quando ocorreram precipitações intensas e de longa duração que causaram grandes inundações e alagamentos por toda a cidade. Outro agravante é a continuidade da utilização única e exclusivamente de técnicas clássicas com enfoque higienista para a solução dos problemas de drenagem urbana, como é o caso das canalizações dos córregos e rios. Diante dessa problemática surge uma nova abordagem do tema que envolve o manejo sustentável das águas pluviais por meio da utilização de um conjunto de medidas de controle da fonte, que podem ser integradas aos sistemas de drenagem existentes. Essas medidas se baseiam na retenção e infiltração das águas pluviais e na retenção de poluentes, reduzindo a probabilidade de inundações e contribuindo para a melhoria da qualidade da água dos sistemas hídricos urbanos. O objetivo deste estudo foi avaliar, por meio da modelagem matemática, o efeito que a implantação de medidas controle na fonte pode proporcionar no abatimento dos picos de cheia em áreas urbanas densamente ocupadas. Para tanto foi selecionada a bacia do córrego da Luz localizada no centro da cidade de São Paulo que apresenta altas taxas de impermeabilização do solo. As análises foram feitas com o PCSWMM, que é um sistema de suporte a decisão espacial desenvolvido pela CHI para o SWMM5 da EPA. Foram selecionadas medidas de controle na fonte que possam se adequar em locais onde a disponibilidade de espaço é pequena, como é o caso dos pavimentos permeáveis, telhados verdes e jardins de chuva. Os cenários analisados consideraram a situação atual sem implantação de medidas de controle e a implantação individual e combinada das medidas de controle na fonte. São avaliados eventos de precipitações, considerando diferentes períodos de retorno e durações de chuva e a porcentagem de área impermeável contribuinte para as medidas de controle. Uma análise geral dos resultados da modelagem permite apontar que o amortecimento dos picos de cheia foi verificado em todos os cenários, em maior ou menor intensidade, e o melhor desempenho no abatimento do pico de cheia foi verificado nos cenários que consideraram a implantação de pavimentos permeáveis. Os resultados da modelagem e os baixos custos de implantação dos pavimentos permeáveis sugerem que esta medida, dentre as três analisadas, é a que apresenta melhor custo-benefício. Contudo sabe-se que a implantação de pavimento asfáltico permeável em 100% da área disponível pode não ser viável. Logo a implantação conjunta com outras medidas de controle pode melhorar o desempenho de um sistema de controle de cheias. / The accelerated urban growth in the last decades associated with a disordered occupation and excessive impervious surfaces has worsened the floods in Brazilian cities. In the City of São Paulo, lately, the rainy season between October and March, in 2009-2010 and in 2010-2011, were especially relevant. There were intense and long-lasting rainfalls that caused flooding all over the city. Another aggravating factor is the single and exclusive use of traditional stormwater management methods for solving urban drainage problems, which is the case of stream and river canalizations. When faced with these problems, a new approach that involves the sustainable management of rainwater starts growing. These techniques include a set of low impact development practices that should be incorporated to the existing drainage systems. They are based on the detention and infiltration of stormwater, also the retention of pollutants. They also contribute to decreasing flood probability and to improving water quality of urban water systems. The goal is to evaluate, through a mathematical modeling, the low-impact development practices effect on reducing the peak runoff rates in highly urbanized areas. To this end, a subcatchment located in the center of São Paulo city, called Córrego da Luz, was chosen. This subcatchment has high rates of soil imperviousness. The analysis will be made with the PCSWMM, which is a spatial decision support system for the SWMM5 from EPA. The PCSWMM was developed by CHI. The low-impact development techniques selected for this study are permeable pavements, green roofs and rain gardens. These techniques can be implemented in places where space is no longer available. The scenarios must consider the current situation without the implementation of control measures, the individual and combined implementation of these practices. Various rainfall events were evaluated considering different durations. Also evaluated were the different percentages of impervious areas treated by low-impact development techniques. An overview of modeling results allows stating that the reduction of peak runoff was observed in all the scenarios, at greater or lower intensity. But the best performance was observed in scenarios that considered the implementation of permeable pavements. Modeling results and low-cost implementation of permeable pavements suggest that it is the most cost-effective measure, among the three analyzed. However, it is known that the implementation of permeable pavement in 100% of the available area may not be feasible. Hence, the combined implementation with other low-impact developments might improve the performance of a flood control system.