Votre recherche

AbstractThe frequency and severity of floods has increased in different regions of the world due to climate change. Although the impact of floods on human health has been extensively studied, the increase in the segments of the population that are likely to be impacted by floods in the future makes it necessary to examine how adaptation measures impact the mental health of individuals affected by these natural disasters. The goal of this scoping review is to document the existing studies on flood adaptation measures and their impact on the mental health of affected populations, in order to identify the best preventive strategies as well as limitations that deserve further exploration. This study employed the methodology of the PRISMA-ScR extension for scoping reviews to systematically search the databases Medline and Web of Science to identify studies that examined the impact of adaptation measures on the mental health of flood victims. The database queries resulted in a total of 857 records from both databases. Following two rounds of screening, 9 studies were included for full-text analysis. Most of the analyzed studies sought to identify the factors that drive resilience in flood victims, particularly in the context of social capital (6 studies), whereas the remaining studies analyzed the impact of external interventions on the mental health of flood victims, either from preventive or post-disaster measures (3 studies). There is a very limited number of studies that analyze the impact of adaptation measures on the mental health of populations and individuals affected by floods, which complicates the generalizability of their findings. There is a need for public health policies and guidelines for the development of flood adaptation measures that adequately consider a social component that can be used to support the mental health of flood victims.

Traditional stormwater control measures are designed to handle system loadings induced by fixed-size storm events. However, climate change is predicted to alter the frequency and intensity of flooding events, stimulating the need to explore another more adaptive flooding solution like real-time control (RTC). This study assesses the performance of RTC to mitigate impacts of climate change on urban flooding resilience. A simulated, yet realistic, urban drainage system in Salt Lake City, Utah, USA, shows that RTC improves the flooding resilience by up to 17% under climatic rainfall changes. Compared with green stormwater infrastructure (GSI), RTC exhibits a lower resistibility, lower flooding failure level, and higher recovery rate in system performance curves. Results articulate that keeping RTC's performance consistent under ‘back-to-back’ storms requires a tradeoff between upstream dynamical operation and downstream flooding functionality loss. This research suggests that RTC provides a path towards smart and resilient stormwater management strategy.

Les inondations de 2017 et 2019 au Québec ont affecté respectivement 293 et 240 municipalités. Ces inondations ont généré une cascade d’évènements stressants (stresseurs primaires et secondaires) qui ont eu des effets sur la santé mentale de la population et retardé le processus de rétablissement des individus. Cette période de rétablissement peut s’échelonner sur plusieurs mois voire plusieurs années. Cette étude s’inscrit dans la spécificité de la recherche mixte mise de l’avant à travers trois stratégies de recherche, réalisées de façon séquentielle : 1) sondage populationnelle réalisé auprès de 680 personnes, 2) analyse de documents produits par les organisations participant au processus de rétablissement social des sinistrés, ou sur des analyses externes portant sur ces interventions de rétablissement et 3) entrevues semi-dirigées auprès de 15 propriétaires occupants ayant complété une demande d’indemnisation à la suite des inondations de 2019 et auprès de 11 professionnels et gestionnaires participant au processus de rétablissement social. Les entrevues semi-dirigées et les questionnaires complétés par les personnes sinistrées lors des inondations de 2019 démontrent que les principales sources de stress ayant des impacts sur la santé et le bien-être des répondants sont : 1) l’absence d’avertissement et la vitesse de la montée des eaux; 2) l’obligation de se relocaliser et la peur d’être victime de pillage; 3) le manque de solidarité et d’empathie de la part de certains employés du MSP; 4) la gestion des conflits familiaux; 5) la gestion de problèmes de santé nouveaux ou préexistants; 6) la complexité des demandes d’indemnisation; 7) la lourdeur et les délais des travaux de nettoyage ou de restauration; 8) les indemnités inférieures aux coûts engendrés par l’inondation; 9) les pertes matérielles subies, particulièrement ceux d’une valeur de plus de 50 000 $; et 10) la diminution anticipée de la valeur de sa résidence. À cela s’ajoute l’insatisfaction à l’égard du programme d’indemnisation du gouvernement du Québec (PGIAF) qui fait plus que doubler la prévalence des symptômes de stress post-traumatique. Les inondations entraînent également une perte de satisfaction ou de bien-être statistiquement significative. La valeur monétaire de cette perte de jouissance peut être exprimée en équivalent salaires. En moyenne, cette diminution du bien-être équivaut à une baisse de salaire de 60 000$ pour les individus ayant vécu une première inondation et à 100 000$ pour les individus ayant vécu de multiples inondations. Ces résultats suggèrent que les coûts indirects et intangibles représentent une part importante des dommages découlant des inondations. Ce projet de recherche vise également à analyser l’application du PGIAF et son influence sur les stresseurs vécus par les sinistrés dans le contexte de la pandémie de COVID-19. La principale recommandation de cette étude repose sur une analyse de documents, un sondage populationnel et des entrevues semi-dirigées. Ainsi, s’attaquer à la réduction de principaux stresseurs nécessite 1) d’améliorer la gouvernance du risque d’inondation, 2) d’intensifier la communication et le support aux sinistrés, et 3) de revoir les mécanismes d’indemnisation existants.

Abstract Measuring freshwater submerged aquatic vegetation (SAV) biomass at large spatial scales is challenging, and no single technique can cost effectively accomplish this while maintaining accuracy. We propose to combine and intercalibrate accurate quadrat‐scuba diver technique, fast rake sampling, and large‐scale echosounding. We found that the overall relationship between quadrat and rake biomass is moderately strong (pseudo R 2  = 0.61) and varies with substrate type and SAV growth form. Rake biomass was also successfully estimated from biovolume (pseudo R 2  = 0.57), a biomass proxy derived from echosounding. In addition, the relationship was affected, in decreasing relevance, by SAV growth form, flow velocity, acoustic data quality, depth, and wind conditions. Sequential application of calibrations yielded predictions in agreement with quadrat observations, but echosounding predictions underestimated biomass in shallow areas (< 1 m) while outperforming point estimation in deep areas (> 3 m). Whole‐system quadrat‐equivalent biomass from echosounding differed by a factor of two from point survey estimates, suggesting echosounding is more accurate at larger scales owing to the increased sample size and better representation of spatial heterogeneity. To decide when an individual or a combination of techniques is profitable, we developed a step‐by‐step guideline. Given the risks of quadrat‐scuba diver technique, we recommend developing a one‐time quadrat–rake calibration, followed by the use of rake and echosounding when sampling at larger spatial and temporal scales. In this case, rake sampling becomes a valid ground truthing method for echosounding, also providing valuable species information and estimates in shallow waters where echosounding is inappropriate.

Climate change and more frequent severe storms have caused persistent flooding, storm surges, and erosion in the northeastern coastal region of the United States. These weather-related disasters have continued to generate negative environmental consequences across many communities. This study examined how coastal residents’ exposure to flood risk information and information seeking behavior were related to their threat appraisal, threat-coping efficacy, and participation in community action in the context of building social resilience. A random sample of residents of a coastal community in the Northeastern United States was selected to participate in an online survey (N = 302). Key study results suggested that while offline news exposure was weakly related to flood vulnerability perception, online news exposure and mobile app use were both weakly associated with flood-risk information seeking. As flood vulnerability perception was strongly connected to flood severity perception but weakly linked to lower self-efficacy beliefs, flood severity perception was weakly and moderately associated with response-efficacy beliefs and information seeking, respectively. Furthermore, self-efficacy beliefs, response efficacy beliefs, and flood-risk information seeking were each a weak or moderate predictor of collective efficacy beliefs. Lastly, flood risk information-seeking was a strong predictor and collective efficacy beliefs were a weak predictor of community action for flood-risk management. This study tested a conceptual model that integrated the constructs from risk communication, information seeking, and protection motivation theory. Based on the modeling results reflecting a set of first-time findings, theoretical and practical implications are discussed.

Abstract. The amount and phase of cold season precipitation accumulating in the upper Saint John River basin are critical factors in determining spring runoff, ice-jams, and flooding in downstream communities. To study the impact of winter and spring storms on the snowpack in the upper Saint John River (SJR) basin, the Saint John River Experiment on Cold Season Storms (SAJESS) utilized meteorological instrumentation, upper air soundings, human observations, and hydrometeor macrophotography during winter/spring 2020–21. Here, we provide an overview of the SAJESS study area, field campaign, and existing data networks surrounding the upper SJR basin. Initially, meteorological instrumentation was co-located with an Environment and Climate Change Canada station near Edmundston, New Brunswick, in early December 2020. This was followed by an intensive observation period that involved manual observations, upper-air soundings, a multi-angle snowflake camera, macrophotography of solid hydrometeors, and advanced automated instrumentation throughout March and April 2021. The resulting datasets include optical disdrometer size and velocity distributions of hydrometeors, micro rain radar output, near-surface meteorological observations, and wind speed, temperature, pressure and precipitation amounts from a K63 Hotplate precipitation gauge, the first one operating in Canada. These data are publicly available from the Federated Research Data Repository at https://doi.org/10.20383/103.0591 (Thompson et al., 2022). We also include a synopsis of the data management plan and data processing, and a brief assessment of the rewards and challenges of utilizing community volunteers for hydro-meteorological citizen science.

Airborne LiDAR scanning is a promising approach to providing high-resolution products that are appropriate for different applications, such as flood management. However, the vertical accuracy of airborne LiDAR point clouds is not constant and varies in space. Having a better knowledge of their accuracy will assist decision makers in more accurately estimating the damage caused by flood. Data producers often report the total estimation of errors by means of comparison with a ground truth. However, the reliability of such an approach depends on various factors including the sample size, accessibility to ground truth, distribution, and a large enough diversity of ground truth, which comes at a cost and is somewhat unfeasible in the larger scale. Therefore, the main objective of this article is to propose a method that could provide a local estimation of error without any third-party datasets. In this regard, we take advantage of geostatistical ordinary kriging as an alternative accuracy estimator. The challenge of considering constant variation across the space leads us to propose a non-stationary ordinary kriging model that results in the local estimation of elevation accuracy. The proposed method is compared with global ordinary kriging and a ground truth, and the results indicate that our method provides more reliable error values. These errors are lower in urban and semi-urban areas, especially in farmland and residential areas, but larger in forests, due to the lower density of points and the larger terrain variations.

Hydrological time series often present nonstationarities such as trends, shifts, or oscillations due to anthropogenic effects and hydroclimatological variations, including global climate change. For water managers, it is crucial to recognize and define the nonstationarities in hydrological records. The nonstationarities must be appropriately modeled and stochastically simulated according to the characteristics of observed records to evaluate the adequacy of flood risk mitigation measures and future water resources management strategies. Therefore, in the current study, three approaches were suggested to address stochastically nonstationary behaviors, especially in the long-term variability of hydrological variables: as an overall trend, shifting mean, or as a long-term oscillation. To represent these options for hydrological variables, the autoregressive model with an overall trend, shifting mean level (SML), and empirical mode decomposition with nonstationary oscillation resampling (EMD-NSOR) were employed in the hydrological series of the net basin supply in the Lake Champlain-River Richelieu basin, where the International Joint Committee recently managed and significant flood damage from long consistent high flows occurred. The detailed results indicate that the EMD-NSOR model can be an appropriate option by reproducing long-term dependence statistics and generating manageable scenarios, while the SML model does not properly reproduce the observed long-term dependence, that are critical to simulate sustainable flood events. The trend model produces too many risks for floods in the future but no risk for droughts. The overall results conclude that the nonstationarities in hydrological series should be carefully handled in stochastic simulation models to appropriately manage future water-related risks.

Research in hydrological sciences is constantly evolving to provide adequate answers to address various water-related issues. Methodological approaches inspired by mathematical and physical sciences have shaped hydrological sciences from its inceptions to the present day. Nowadays, as a better understanding of the social consequences of extreme meteorological events and of the population’s ability to adapt to these becomes increasingly necessary, hydrological sciences have begun to integrate knowledge from social sciences. Such knowledge allows for the study of complex social-ecological realities surrounding hydrological phenomena, such as citizens’ perception of water resources, as well as individual and collective behaviors related to water management. Using a mixed methods approach to combine quantitative and qualitative approaches has thus become necessary to understand the complexity of hydrological phenomena and propose adequate solutions for their management. In this paper, we detail how mixed methods can be used to research flood hydrology and low-flow conditions, as well as in the management of these hydrological extremes, through the analysis of case studies. We frame our analysis within the three paradigms (positivism, post-positivism, and constructivism) and four research designs (triangulation, complementary, explanatory, and exploratory) that guide research in hydrology. We show that mixed methods can notably contribute to the densification of data on extreme flood events to help reduce forecasting uncertainties, to the production of knowledge on low-flow hydrological states that are insufficiently documented, and to improving participatory decision making in water management and in handling extreme hydrological events.

Earthquakes pose potentially substantial risks to residents in the Western Quebec seismic zone of eastern Canada, where Ottawa and Montreal are located. In eastern Canada, the majority of houses are not constructed to modern seismic standards and most homeowners do not purchase earthquake insurance for their homes. If a devastating earthquake strikes, homeowners would be left unprotected financially. To quantify financial risks to homeowners in the Western Quebec seismic zone, regional earthquake catastrophe models are developed by incorporating up-to-date public information on hazard, exposure and vulnerability. The developed catastrophe models can quantify the expected and upper-tail financial seismic risks by considering a comprehensive list of possible seismic events as well as critical earthquake scenarios based on the latest geological data in the region. The results indicate that regional seismic losses could reach several tens of billions of dollars if a moderate-to-large earthquake occurs near urban centres in the region, such as Montreal and Ottawa. The regional seismic loss estimates produced in this study are useful for informing earthquake risk management strategies, including earthquake insurance and disaster relief policies.

Natural calamities like floods and droughts pose a significant threat to humanity, impacting millions of people each year and incurring substantial economic losses to society. In response to this challenge, this thesis focuses on developing advanced machine learning techniques to improve water height prediction accuracy that can aid municipalities in effective flood mitigation. The primary objective of this study is to evaluate an innovative architecture that leverages Long Short Term Networks - neural networks to predict water height accurately in three different environmental scenarios, i.e., frazil, droughts and floods due to snow spring melt. A distinguishing feature of our approach is the incorporation of meteorological forecast as an input parameter into the prediction model. By modeling the intricate relationships between water level data, historical meteorological data and meteorological forecasts, we seek to evaluate the impact of meteorological forecasts and if any inaccuracies could impact water-level prediction. We compare the outcomes obtained by incorporating next-hour, next-day and next-week meteorological data into our novel LSTM model. Our results indicate a comprehensive comparison of the usage of various parameters as input and our findings suggest that accurate weather forecasts are crucial in achieving reliable water height predictions. Additionally, this study focuses on the utilization of IoT sensor data in combination with ML models to enhance the effectiveness of flood prediction and management. We present an online machine learning approach that performs online training of the model using real-time data from IoT sensors. The integration of live sensor data provides a dynamic and adaptive system that demonstrates superior predictive capabilities compared to traditional static models. By adopting these advanced techniques, we can mitigate the adverse impacts of natural catastrophes and work towards building more resilient and disaster-resistant communities.

L’estimation du débit en rivières est un paramètre clé pour la gestion des ressources hydriques, la prévention des risques liés aux inondations et la planification des équipements hydroélectriques. Lorsque le débit d’eau est très élevé lors d'évènements extrêmes, les méthodes de jaugeage traditionnelles ne peuvent pas être utilisées. De plus, les stations du réseau hydrométrique sont généralement éparses et leur répartition spatiale n’est pas optimale. Par conséquent, de nombreuses sections de rivières ne peuvent être suivies par des mesures et observations du débit. Pour ces raisons, pendant la dernière décennie, les capteurs satellitaires ont été considérés comme une source d’observation complémentaire aux observations traditionnelles du niveau d’eau et du débit en rivières. L’utilisation d’une telle approche a fourni un moyen de maintenir et d’étendre le réseau d'observation hydrométrique. L’approche avec télédétection permet d’estimer le débit à partir des courbes de tarage qui met en relation le débit instantané (Q) et la géométrie d’une section transversale du chenal (la largeur ou la profondeur effective de la surface d’eau). En revanche, cette méthode est associée à des limitations, notamment, sa dépendance aux courbes de tarage. En effet, en raison de leurs natures empiriques, les courbes de tarage sont limitées à des sections spécifiques et ne peuvent être appliquées dans d’autres rivières. Récemment, des techniques d’apprentissage profond ont été appliquées avec succès dans de nombreux domaines, y compris en hydrologie. Dans le présent travail, l’approche d’apprentissage profond a été choisie, en particulier les réseaux de neurones convolutifs (CNN), pour estimer le débit en rivière. L’objectif principal de ce travail est de développer une approche d’estimation du débit en rivières à partir de l’imagerie RADARSAT 1&amp;2 à l’aide de l’apprentissage profond. La zone d’étude se trouve dans l’ecozone du bouclier boréal à l’Est du Canada. Au total, 39 sites hydrographiques ont fait l’objet de cette étude. Dans le présent travail, une nouvelle architecture de CNN a été a été proposée, elle s'adapte aux données utilisées et permet d’estimer le débit en rivière instantané. Ce modèle donne un résultat du coefficient de détermination (R²) et de Nash-Sutcliffe égale à 0.91, le résultat d’erreur quadratique moyenne égale à 33 m³ /s. Cela démontre que le modèle CNN donne une solution appropriée aux problèmes d’estimation du débit avec des capteurs satellites sans intervention humaine. <br /><br />Estimating river flow is a key parameter for effective water resources management, flood risk prevention and hydroelectric facilities planning. In cases of very high flow of water or extreme events, traditional gauging methods cannot be reliable. In addition, hydrometric network stations are often sparse and their spatial distribution is not optimal. Therefore, many river sections cannot be monitored using traditional flow measurements and observations. For these reasons, satellite sensors are considered as a complementary observation source to traditional water level and flow observations in the last decades. The use of this kind of approach has provided a way to maintain and expand the hydrometric observation network. Remote sensing data can be used to estimate flow from rating curves that relate the instantaneous flow (Q) to the geometry of a channel cross-section (the effective width or depth of the water surface). On the other hand, remote sensing is also associated with limitations, notably its dependence on the rating curves. Indeed, due to their empirical nature, rating curves are limited to specific sections and cannot be applied in other rivers. Recently, deep learning techniques have been successfully applied in many fields, including hydrology. In the present work, the deep learning approach has been chosen, in particular convolutional neural networks (CNN), to estimate river flow. The main objective of this work is to develop an approach to estimate river flow from RADARSAT 1&amp;2 imagery using deep learning. In this study, 39 hydrographic sites of the Boreal Shield ecozone in Eastern Canada were considered. A new CNN architecture was developed to provide a straightforward estimation of the instantaneous river flow rate. The achieved results demonstrated a coefficient of determination (R²) and Nash-Sutcliffe values of 0.91, and a root mean square error of 33m³ /s. This indicates the effectiveness of CNN in automatic flow estimation with satellite sensors.

Streamflow forecasting is important for managing water resources in sectors like agriculture, hydropower, drought management, and urban flood prevention planning. Our study examines short and long lead-times to create a framework for streamflow forecasting that can benefit water resource management and related sectors. To improve streamflow forecasts for up to ten days of lead-time, the study first focuses on improving initial conditions using an ensemble Kalman filter as a data assimilation method. The goal is to regulate the hyperparameters of the ensemble Kalman filter for each season to produce more accurate forecasts. A sensitivity analysis is conducted to identify the best hyperparameter sets for each season, including uncertainty in temperature, precipitation, observed streamflow, and the water content of three state variables - vadose zone, saturated zone, and snowpack - from the CEQUEAU model. Results indicate that improving initial conditions with the ensemble Kalman filter produces more skillful forecasts until a 6-day leadtime. Temperature uncertainty is particularly sensitive and varies across seasons. The vadose zone state variable was identified as the most important and sensitive state variable, and updating all state variables systematically may not be necessary for improving forecast skill. Recent machine learning advances are improving short-term streamflow forecasting. One such method is the Long Short-Term Memory (LSTM) model. In general, neural networks learn from regression as relationships exist between input-output. However, LSTM models have a feature named ‘forget gate’, which enables them to learn the relationship between inputs (e.g., temperature and precipitation) and output (streamflow), and also to capture temporal dependencies in the data. The study aimed to compare the performance of the Long ShortTerm Memory (LSTM) model with data assimilation-based and process-based hydrological models in short-term streamflow forecasting. All three models were tested using the same ensemble weather forecasts. The LSTM model demonstrated good performance in forecasting streamflow, with a Kling-Gupta efficiency (KGE) greater than 0.88 for 9 lead-times. The LSTM model did not incorporate data assimilation, but it benefited from observed streamflow until the last day before the forecast. This is because the LSTM model learned and incorporated knowledge from the previous days while issuing forecasts, similar to how data assimilation updates initial conditions. The study results also showed that the LSTM model had better performance up to day 6 of lead-time compared to the data assimilation-based models. However, training the LSTM model separately for each lead-time is a time-consuming process and is a disadvantage compared to the data assimilation-based methods. Nonetheless, the study demonstrated the potential of machine learning techniques in improving streamflow forecasting. The forecasting of streamflow for long lead-times such as a month usually involves the use of historical meteorological data to create probable future scenarios, as meteorological forecasts become unreliable beyond this lead-time. In this study, we proposed a novel method for streamflow forecasting based on ensemble streamflow forecasting (ESP) filtering, using a Genetic Algorithm (GA) to filter forecast scenarios. This method quantifies the potential of historical data for each basin. This potential could be utilized to enhance the accuracy of streamflow forecasts. We sorted the selected and unselected scenarios to find out the common features between them, but the results did not help distinguish between the two groups. Nonetheless, the GA method can be used as a benchmark for future studies to improve longterm streamflow forecasting. This method can also be used to compare different forecast methods based on the potential shown by the GA method for a specific size of ESP members. For instance, if a method uses large-scale climate signals to filter ESP members, the forecast skill result could be compared with the potential of historical data for that particular size of ESP members.

Les instances responsables d’assurer la gestion des risques d’inondations reliés aux embâcles de glace sont toujours à la recherche d’outils visant à prévenir les risques et réduire les conséquences sur les populations et les infrastructures. Grâce à la modélisation de certains paramètres jouant un rôle dans la formation des embâcles de glace comme la prédisposition géomorphologique et les conditions hydrométéorologiques, la prévention de ceux-ci s’est grandement améliorée. Les modèles axés sur la force de mobilisation de la rivière et sa capacité à contraindre l’écoulement gagnerait en pertinence s’ils pouvaient inclure la résistance du couvert de glace. Les outils de télédétection sont une manière efficace de connaitre l’état du couvert de glace tant sur l’ensemble de la rivière qu'à différents endroits ciblés. Ceux-ci peuvent générer différents produits cartographiques utiles avant, pendant et après les événements. La présente thèse vise à intégrer le suivi du couvert de glace dans les méthodes de prévention des embâcles à l’aide d’outils provenant de la télédétection. Pour ce faire, quatre sous-objectifs ont été accomplis 1) créer une approche de suivi du couvert de glace à grande échelle en exploitant les données de télédétection optique, radar et acquises par drone, 2) développer une méthode de cartographie automatique du type de glace par estimation d’ensemble à partir d’imagerie radar, 3) concevoir un modèle de détection automatique des lieux à risque de débâcle en utilisant les connaissances de personnes expertes pour interpréter les cartes du type de glace et 4) intégrer les outils développés aux autres modèles conçus dans le cadre de DAVE (Dispositif d’alertes et de vigilance aux embâcles). Les contributions originales découlant de cette thèse touchent plusieurs aspects du suivi du couvert de glace. Elles incluent la démonstration de la pertinence des indicateurs de suivi de la glace par la télédétection, la conceptualisation d’une méthode de segmentation de la rivière en secteurs de production, transport et accumulation de la glace, l’élaboration d’un modèle de cartographie du type de glace par estimation d’ensemble plus performant et polyvalent que les classificateurs originaux, la construction d’une base de données de dégradation du couvert de glace à partir des connaissances de personnes expertes en cartographie de la glace et un modèle de classification de la dégradation du couvert de glace. Cette thèse se conclut par l’intégration conceptuelle à l’aide d’une analyse multicritères hiérarchique des différents outils développés au sein de DAVE. <br /><br /> The authorities responsible for ice jam flood risk management are always looking for tools to prevent harm and reduce consequences on populations and infrastructure. Ice jam prevention has been greatly improved by modelling certain parameters—such as geomorphic predispositions and hydrometeorological conditions—that are central to ice jam formation. These models focusing on the strength of the river current and its ability to constrain the flow would gain in relevance if they could include ice cover strength. Remote sensing tools are an effective way of knowing the state of the ice cover over the whole river and at different target locations. These can be used to generate different map products to include in monitoring before, during and after hazard events. This dissertation therefore aims to integrate ice cover monitoring into ice jam prevention methods using remote sensing tools. To do so, three main sub-objectives were accomplished: 1) to create a large-scale ice cover monitoring approach using remote sensing data such as optical, radar and drone images, 2) to develop a method for automatic mapping of ice type by ensemble estimation from radar imagery, 3) to design an automatic detection model for breakup risk locations using ice type maps and expert judgment and 4) to integrate the tools developed with other models created within the Dispositif d’alertes et de vigilance des embâcles de glace (DAVE). The original contribution from this work covers multiple aspects of ice cover monitoring. This dissertation demonstrates the relevance of direct indicators of ice monitoring by remote sensing, conceptualizes a method for river segmentation into areas of ice production, transport and accumulation, develops an ensemble-based estimation ice type mapping model more efficient and versatile than the original algorithms, constructs an ice cover degradation database derived from the knowledge of ice mapping experts, and proposes a classification model of ice cover degradation. This dissertation concludes with the conceptual integration by analytic hierarchy process of the different tools developed within DAVE.

Aujourd'hui, la cartographie des réseaux hydrographiques est un sujet important pour la gestion et l'aménagement de l'espace forestier, la prévention contre les risques d'inondation, etc. Les données sources pour cartographier les cours d'eau sont des nuages de points obtenus par des lidars aéroportés. Cependant, les méthodes d'extraction des réseaux usuelles nécessitent des opérations de découpage, de rééchantillonnage et d'assemblage des résultats pour produire un réseau complet, altérant la qualité des résultats et limitant l'automatisation des traitements. Afin de limiter ces opérations, une nouvelle approche d'extraction est considérée. Cette approche propose de construire un réseau de crêtes et de talwegs à partir des points lidar, puis transforme ce réseau en réseau hydrographique. Notre recherche consiste à concevoir une méthode d'extraction robuste du réseau adaptée aux données massives. Ainsi, nous proposons d'abord une approche de calcul du réseau adaptée aux surfaces triangulées garantissant la cohérence topologique du réseau. Nous proposons ensuite une architecture s'appuyant sur des conteneurs pour paralléliser les calculs et ainsi traiter des données massives.

Au Québec, les conditions printanières extraordinaires de 2017 et 2019 ont incité le gouvernement provincial à commander une mise à jour des cartes des zones inondables. La plupart des cartes existantes ne reflètent pas adéquatement l’aménagement actuel du territoire, ni l’aléa associé. Généralement, pour la cartographie, les modèles hydrodynamiques tel que HEC-RAS sont utilisés, mais ces outils nécessitent une expertise significative, des données hydrométriques et des relevés bathymétriques à haute résolution. Étant donnée la nécessité de mettre à jour ces cartes tout en réduisant les coûts financiers associés, des méthodes conceptuelles simplifiées ont été développées. Ces approches, y compris l’approche géomatique HAND (Height above the nearest drainage), qui reposent uniquement sur un modèle numérique d’élévation (MNE), sont de plus en plus utilisées. HAND permet de calculer la hauteur d’eau nécessaire pour inonder chaque pixel du MNE selon la différence entre son élévation et celle du pixel du cours d’eau dans lequel il se déverse. Les informations sur la géométrie hydraulique dérivées par HAND ainsi que l’application de l’équation de Manning permettent la construction d’une courbe de tarage synthétique (CTS) pour chaque tronçon de rivière homogène. Dans la littérature, cette méthode a été appliquée pour établir une cartographie de la zone inondable de première instance de grands fleuves aux États-Unis avec un taux de correspondance de 90% par rapport à l’utilisation de HEC-RAS. Elle n’a toutefois pas été appliquée sur de petits bassins versants, car ceux-ci engendrent des défis méthodologiques substantiels. Ce projet s’attaque à ces défis sur deux bassins versants Québécois, ceux des rivières à la Raquette et Delisle. Les conditions frontières des modèles sont dérivées d’un traitement statistique empirique des séries de débits simulés avec le modèle hydrologique HYDROTEL. Étant donnée l’absence de stations météorologiques sur le territoire à l’étude, des chroniques du système Canadien d’Analyse de la précipitation (CaPA) ont été utilisées pour cette modélisation hydrologique. Les résultats de ce projet pointent vers des performances satisfaisantes de l’approche géomatique HAND-CTS en comparaison avec le modèle hydrodynamique HEC-RAS (1D/2D et 2D au complet), avec des taux de correspondance entre les étendues des inondations supérieurs à 60 % pour les bassins versants de Delisle et à la Raquette. Les comparaisons étaient effectuées sur une gamme de débit allant d’un débit de période de retour de 2 ans jusqu’à un débit de plus de 350 ans. On notera que l’application sur la rivière à la Raquette a été développée dans les règles de l’art, incluant un processus de calage développé dans le cadre d’un projet de maitrise en sciences de l’eau connexe à ce mémoire, relativement à la longueur du tronçon, le calage vertical de la CTS en considérant la hauteur d’eau présente dans le cours d’eau lors du relevé LiDAR et sa précision verticale. Les résultats ont montré que le coefficient de précision globale le plus bas était de 98 % pour un débit de 350 ans, avec une précision de plus que 99 % pour les autres périodes de retour, ce qui représente une très bonne performance du modèle. Et par ailleurs, le coefficient de Kappa conditionnel humide variait entre 58 % et 28 %. Alors, que pour la rivière Delisle, l’application se veut naïve, c’est-à-dire sans calage préalable de la méthode HANDCTS. La précision globale a varié entre 83 % et 96 %, ce qui est considéré comme "très approprié" et une variation du coefficient Kappa conditionnel humide de 35,2 à 64,3 %. Alors que pour une différence d’élévations d'eau entre les élévations de référence et simulées, la performance était quantifiée par un RMSE qui variait pour les périodes de retour de 100 ans et de 350 ans respectivement de 4,5 m et de 7,1 m. Enfin, la distribution spatiale des différences d’élévations montre une distribution gaussienne avec une moyenne qui est à peu près égale à 0 où la plupart des erreurs se situent entre -0,34 m et 1,1 m La cartographie des zones inondables dérivée de HAND-CTS présente encore certains défis associés notamment à la présence d’infrastructures urbaines complexes (ex. : ponceaux, ponts et seuils) dont l’influence hydraulique n’est pas considérée. Dans le contexte où l’ensemble du Québec (529 000 km²) dispose d’une couverture LiDAR, les résultats de ce mémoire permettront de mieux comprendre les sources d’incertitude associées à la méthode HAND-CTS tout en démontrant son potentiel pour les bassins versants dépourvus de données bathymétriques et hydrométéorologiques. <br /><br />The 2017 and 2019 extraordinary spring conditions prompted the Quebec government to update flood risk maps, as most of them do not adequately reflect current land use and associated hazard. Generally, hydrodynamic models such as HEC-RAS are used for flood mapping, but they require significant expertise, hydrometric data, and high-resolution bathymetric surveys. Given the need to update these maps while reducing the associated financial costs, simplified conceptual methods have been developed over the last decade. These methods are increasingly used, including HAND (height above the nearest drainage), which relies on a Digital Elevation Model (DEM) to delineate the inundation area given the water height in a river segment. Furthermore, the river geometry derived from HAND data and the application of Manning’s equation allow for the construction of a synthetic rating curve (SRC) for each homogeneous river segment. In the scientific literature, this framework has been applied to produce first-instance floodplain mapping of large rivers. For example, in the Continental United States 90% match rates were achieved when compared to the use of HEC-RAS. However, this framework has not been validated for small watersheds, as substantial methodological challenges are anticipated. This project addresses these underlying challenges in two Quebec watersheds, the à la Raquette and Delisle watersheds. The boundary conditions of the HECRAS models were derived from an empirical statistical treatment of flow time series simulated by HYDROTEL, a hydrological model, using Canadian Precipitation Analysis Product (CaPA) time series. The results of this project point towards satisfactory performances, with match rates greater than 60 % for both watersheds. It should be noted that the application on the Delisle River is naive, that is without prior calibration of the HAND-SRC method. The overall accuracy ranged from 83.4 % to 96.2 % while the water surface elevation difference was quantified by an RMSE that was for the 100-year and 350-year return periods of 4.5 m and 7.1 m respectively and where most errors are between -0.34 m and 1.1 m representing a very good model comparing to similar studies. For à la Raquette, the application showed an overall accuracy coefficient of 98 % for a 350-year flow, with an accuracy of over 99 % for other return periods. The mapping of flood risk areas using HAND-SRC still faces certain challenges, notably the presence of complex urban infrastructures (e.g., culverts, bridges, and weirs) whose hydraulic influences are not considered by this geomatic approach. Given that most of Quebec (529,000 km²) topography has been digitized using LiDAR data, the results conveyed in this MSc thesis will allow for a better understanding of the sources of uncertainty associated with the application of the HAND-SRC method while demonstrating its potential for watersheds lacking hydrometeorological and high-resolution bathymetric data.

This thesis examines the main socio-environmental relationships related to recurrent floodings in the Sainte-Anne River watershed by mobilizing the history of land use, local knowledge and risk management policies. From a political ecology perspective, these relationships are part of both social representations of nature, power dynamics associated with the appropriation of ressources and multiple temporalities. By also mobilizing the theoreticals frameworks of the anthropology of disasters and amphibian anthropology, this study allow to retrace step by step what “flood zone” is as a socially constructed space by the authorities, and go back to the source of the advent of “flood” as a catastrophic event, where rising waters have been part of the characteristics of these territories since the beginning of the sedentarization of its inhabitants. Based on 93 semi-directed interviews (76 residents and 17 institutions stakeholders) realized between February to October 2019, non-participant observation and documentary research in municipal and regional archives, this ethnography of the Sainte-Anne River watershed allows a unique incursion with Quebec riverside residents who live with recurrent rising waters. Through four case studies (Saint-Raymond, Saint-Alban, Saint-Casimir and Sainte-Anne-de-la-Pérade), historical contexts of occupation of the territory were documented and significant events were described by focusing on local residents adaptation strategies and anticipated management by institutional stakeholders. The result is a portrait, in a temporal perspective, of the relationship of cohabitation between residents and the river and its overflows. This cohabitation is characterized by tensions and paradoxes associated with different social representations of water and temporality that coexist within the actors, as well as changes in power relations towards the environment. Cette thèse examine les principaux rapports socio-environnementaux liés aux inondations récurrentes dans le bassin versant de la rivière Sainte-Anne en mobilisant l’histoire de l’occupation du territoire, les savoirs locaux et les politiques de gestion des risques. Dans une perspective d’écologie politique, ces rapports s’inscrivent à la fois dans les représentations sociales de la nature, les dynamiques de pouvoir associées à l’appropriation des ressources et des temporalités multiples. En puisant également dans les cadres théoriques de l’anthropologie des catastrophes et de l’anthropologie amphibienne, cette étude permet notamment de retracer pas à pas ce qu’est la « zone inondable » en tant qu’espace construit socialement par les autorités, et de remonter à la source de l’avènement de « l’inondation » comme étant un événement catastrophique, alors que la montée des eaux fait partie des caractéristiques de ces territoires depuis le début de la sédentarisation des habitants. Basée sur 93 entrevues semi-dirigées (76 riverains et 17 acteurs institutionnels) menées de février à octobre 2019, de l’observation non participante et une recherche documentaire dans les archives municipales et régionales, cette ethnographie du bassin versant de la rivière Sainte-Anne permet une incursion unique auprès de citoyens québécois qui vivent avec la montée récurrente des eaux. Pour quatre municipalités riveraines (Saint-Raymond, Saint-Alban, Saint-Casimir et Sainte-Anne-de-la-Pérade), les contextes historiques de l’occupation du territoire ont été documentés et les événements significatifs ont été décrits en focalisant sur les stratégies d’adaptation des résidents et la gestion menée par des acteurs institutionnels. En résulte un portrait, dans une perspective temporelle, de la relation de cohabitation entre les riverains et la rivière et ses débordements. Cette cohabitation est caractérisée par des tensions et des paradoxes associés aux différentes représentations sociales de l’eau et de la temporalité qui coexistent au sein des acteurs, ainsi qu’aux changements dans les rapports de pouvoir envers l’environnement.

Abstract Worldwide, there has been an increase in the presence of potentially toxic cyanobacterial blooms in drinking water sources and within drinking water treatment plants (DWTPs). The objective of this study is to validate the use of in situ probes for the detection and management of cyanobacterial breakthrough in high and low-risk DWTPs. In situ phycocyanin YSI EXO2 probes were devised for remote control and data logging to monitor the cyanobacteria in raw, clarified, filtered, and treated water in three full-scale DWTPs. An additional probe was installed inside the sludge holding tank to measure the water quality of the surface of the sludge storage tank in a high-risk DWTP. Simultaneous grab samplings were carried out for taxonomic cell counts and toxin analysis. A total of 23, 9, and 4 field visits were conducted at the three DWTPs. Phycocyanin readings showed a 93-fold fluctuation within 24 h in the raw water of the high cyanobacterial risk plant, with higher phycocyanin levels during the afternoon period. These data provide new information on the limitations of weekly or daily grab sampling. Also, different moving averages for the phycocyanin probe readings can be used to improve the interpretation of phycocyanin signal trends. The in situ probe successfully detected high cyanobacterial biovolumes entering the clarification process in the high-risk plant. Grab sampling results revealed high cyanobacterial biovolumes in the sludge for both high and low-risk plants.