Votre recherche

RÉSUMÉ : Le fjord du Saguenay est une vallée glacière de 110 km de long et de 280 m de profondeur maximale qui relie la rivière Saguenay à sa tête à l'estuaire du Saint-Laurent à son embouchure. La bathymétrie du fjord est caractérisée par 3 seuils : le seuil peu profond (~ 20 m) à l'embouchure, un seuil intermédiaire (60 m) à 20 km en amont et un seuil profond (120 m) à 35 km en amont. Ces seuils séparent le fjord en 3 bassins : le bassin extérieur, le bassin intermédiaire et le bassin intérieur. La circulation dans le fjord est forcée par l'apport d'eau douce de la rivière Saguenay à sa tête, de grandes marées (jusqu'à 6 m de marnage) à son embouchure qui apportent de l'eau salée ainsi que par le vent. La circulation à grande échelle a été caractérisée par trois régimes saisonniers au cours desquels les eaux profondes, intermédiaires et de sous-surface du bassin intérieur sont renouvelées respectivement au début de l'hiver, en été et à la fin de l'hiver. Des indications indirectes suggèrent que ces régimes sont déterminés par des processus turbulents qui se produisent localement à chacun de ces trois seuils. Ici, nous présentons les résultats des expériences de terrain que nous avons menées qui visaient à étudier plus directement la dynamique des processus de seuil induits par les courants de marées ainsi que les modifications de masse d'eau qui surviennent aux seuils. À ce jour, nos mesures fournissent la description la plus précise et la plus complète des structures d'écoulement de marée stratifié autour des seuils internes du fjord du Saguenay. Nous avons également observé qu'un ressaut hydraulique interne semble se former lors de chaque marée descendante en aval du seuil intermédiaire, mais pas lors des marées montantes. Des recherches sont toujours en cours pour mieux comprendre cette asymétrie, mais notre hypothèse est que la présence d'une masse d'eau plus salée en amont du seuil intermédiaire empêche la formation d'un ressaut hydraulique, un processus qui pourrait être similaire à celui documenté à Knight Inlet (Colombie-Britannique, Canada). -- Mot(s) clé(s) en français : Fjord du Saguenay, Estuaire du St-Laurent, Processus de Seuil, Mélange Turbulent, Océanographique Côtière. -- ABSTRACT : The Saguenay Fjord is a 110 km long and 280 m deep (max depth) multi-silled glacial valley that connects the Saguenay River at its head with the St. Lawrence Estuary at its mouth. The bathymetry is characterized by 3 sills : a shallow 20-m deep sill at the mouth, an intermediate 60-m deep sill 20 km up-fjord and a deep 120-m sill 35 km up-fjord. These sills separate 3 basins, the outer, the intermediate and the inner basins. The circulation in the fjord is forced by the Saguenay River at its head that brings freshwater, large tides (up to 6 m range) at its mouth that brings salt water and by wind. The large-scale circulation has been characterized by four seasonally dependent regimes during which the deep, intermediate and subsurface waters of the inner basin are being renewed, respectively, during early winter, summer and late winter. There are indirect indications that those regimes are determined by turbulent processes occurring locally at each of these three sills. Here, we carried out field experiments to more directly investigate the detailed dynamics of tidally-driven sill processes and water mass modifications occurring across these three sills. Our measurements provide to date the most accurate and complete description of the stratified tidal flow structures around these sills. We also found that an internal hydraulic jump seems to form every ebb tide on the down-fjord side of the intermediate sill but not during flood tide on the up-fjord side. Research is ongoing to better understand this asymmetry but our hypothesis is that it is the presence of a salty pool up-fjord of the sill that prevents the formation of a hydraulic jump, a process that may be similar to that documented in Knight Inlet (British Columbia, Canada). -- Mot(s) clé(s) en anglais : Saguenay Fjord, St-Lawrence Estuary, Sill Processes, Turbulent Mixing, Coastal Oceanography.

While impressive results have been achieved in the well-known fields where Deep Learning allowed for breakthroughs such as computer vision, its impact on different older areas is still vastly unexplored. In Computational Fluid Dynamics and especially in Flood Modeling, many phenomena are very high-dimensional, and predictions require the use of numerical simulations, which can be, while very robust and tested, computationally heavy and may not prove useful in the context of real-time predictions. This issue led to various attempts at developing Reduced-Order Modeling techniques, both intrusive and non-intrusive. One recent relevant addition is a combination of Proper Orthogonal Decomposition with Deep Neural Networks (POD-NN). Yet, to our knowledge, little has been performed in implementing uncertainty-aware regression tools in the example of the POD-NN framework. In this work, we aim at comparing different novel methods addressing uncertainty quantification in Neural Networks, pushing forward the POD-NN concept with Deep Ensembles and Bayesian Neural Networks, which we first test on benchmark problems, and then apply to a real-life application: flooding predictions in the Mille-Iles river in Laval, QC, Canada. Building a non-intrusive surrogate model, able to know when it doesn’t know, is still an open research area as far as neural networks are concerned.

L’algue Didymosphenia geminata est une diatomée à potentiel envahissant. Cette algue microscopique, attachée au substrat, peut sécréter des quantités phénoménales de mucilage et créer de vaste amas fibreux tapissant le lit d’une rivière. Autrefois considérée très rare et représentative d’une qualité d’eau impeccable, elle est maintenant répertoriée omniprésente dans les rivières oligotrophes à travers le monde. Au cours des deux dernières décennies, son émergence a créé un intérêt marqué de la part des écologistes et gestionnaires de rivières puisque sa biologie et les facteurs régissant son occurrence et sa sévérité sont méconnus. Au Québec, c’est à l’été 2006 que les premières proliférations massives de l’algue didymo ont été observées dans la rivière Matapédia. Le manque de connaissances à son égard, plus précisément sur sa distribution, les causes de l’apparition des proliférations et de l’impact négatif possible de cet envahissement sur le réseau trophique mais surtout sur le cycle de vie du saumon atlantique en rivière a créé un état de panique chez les gestionnaires et les utilisateurs des rivières à saumon. Il existait dès lors un urgent besoin de combler ces lacunes. Afin de mieux comprendre les facteurs régissant tant la croissance cellulaire que la croissance des amas mucilagineux, un modèle conceptuel a été élaboré grâce à la collaboration de plusieurs chercheurs travaillant au Colorado, à l’île de Vancouver, en Nouvelle-Zélande et au Québec afin de colliger les observations et les données de terrain. Ce travail de synthèse a permis d’identifier les facteurs favorisant sa colonisation, sa croissance et sa persistance en rivière. La dynamique spatio-temporelle de l’algue didymo est dépendante de divers facteurs (ou conditions environnementales) dont les seuils critiques déterminent son comportement de prolifération. En rivière, la croissance d’algues benthiques est principalement déterminée par les variations saisonnières du débit, la température, la lumière et la chimie de l’eau. Ainsi, la connaissance des paramètres d’habitats favorables permet de déterminer si un bassin versant est favorable à la présence cellulaire de l’algue didymo et à la probabilité que celle-ci puisse y proliférer. Le modèle conceptuel permet d’établir la variabilité et la sévérité saisonnière du comportement envahissant de l’algue didymo en rivière. Afin d’appliquer le modèle conceptuel développé à l’échelle du bassin versant, nous avons mis sur pied un réseau de suivi volontaire des proliférations d’algue didymo dans le bassin versant de la rivière Restigouche. Vingt-deux organisations différentes totalisant 70 bénévoles ont été formés à identifier et quantifier les proliférations d’algue didymo. Entre 2010 et 2015, 1 228 observations ont été réalisées. L’analyse de cette base de données, nous a permis de déterminer que 71% de la variabilité de la sévérité des proliférations durant l’été est inversement proportionnelle à l’intensité de la crue printanière. À l’échelle du tronçon, l’analyse des proliférations pour différents types de faciès-substrat a permis de d’identifier une préférence accrue pour les seuils. Par la suite, afin d’appliquer le modèle conceptuel à l’échelle régionale, nous avons comparé la distribution de l’algue didymo dans trois régions de l’est du Québec (i.e. Gaspésie, Bas-St-Laurent et Côte-Nord) avec divers paramètres physico-chimiques de vingt-neuf rivières à saumon. Le principal facteur régissant la présence-absence de cellules est le pH. Les diatomées ont des optimums de pH très spécifiques et la géologie contrastante de la Côte-Nord (Bouclier canadien) par rapport aux basses terres du Saint-Laurent et les Appalaches en Gaspésie, a permis d’identifier que les eaux acides riches en tanins et lignines ne sont pas favorables à la survie et la croissance de l’algue didymo. Grâce à une meilleure connaissance des facteurs de contrôle de l’algue didymo à diverses échelles spatiales, nous pouvons déterminer les rivières à risque de proliférations massives. Au sein des rivières exhibant des proliférations, nous avons déterminé que les préférences d’habitat de l’algue didymo sont identiques à celles du saumon atlantique juvénile (i.e. eaux rapides, peu profondes, claires, froides, faible en nutriments avec un substrat grossier). Afin de déterminer l’ampleur de l’impact de la présence des proliférations sur le réseau trophique, nous avons vérifié son impact sur la communauté algale. La structure dense et fibreuse crée un environnement physique dans lequel plusieurs diatomées peuvent s’y loger. Nos résultats confirment que l’algue crée un habitat de choix pour les plus petites diatomées favorisant ainsi une plus grande diversité taxonomique. Malgré une augmentation de la complexité du biofilm suivant son épaississement, il n’y a pas d’impact supplémentaire sur la structure et la diversité taxonomique des échantillons. Puisque la structure de l’habitat est modifiée par les tapis et que ces derniers sont susceptibles d’altérer la structure et le fonctionnement de l’écosystème, nous avons évalué l’effet des proliférations sur le comportement alimentaire des saumons juvéniles. Grâce à l’observation en apnée du comportement, nous pouvons conclure que les saumons juvéniles effectuent une plus grande proportion de quêtes benthiques en fonction du pourcentage de recouvrement algal. Ce changement de comportement n’est pas attribuable à une diminution des proies disponibles au sein de la dérive. Notre étude confirme la grande plasticité comportementale des saumons juvéniles face aux modifications de l’habitat. Afin de vérifier l’impact des proliférations sur les ressources alimentaires et la diète des saumons juvéniles, nous avons utilisé l’approche par analyse d’isotopes stables. Cette approche permet d’établir un portrait intégré de l’utilisation des ressources par les poissons. Les signatures isotopiques divergentes entre les différents tissus des saumons confirment une modification de la diète saisonnière. Les saumons juvéniles en présence de l’algue didymo ont une diète moins diversifiée et appauvrie en carbone principalement composée de petits chironomides et de trichoptères. Malgré que l’indice de condition physique des saumons juvéniles soit similaire entre les deux sites, leur taille est plus petite dans le site avec proliférations. Les travaux futurs devront tenter de valider l’impact de la diminution de la qualité de la diète par les proliférations d’algue didymo sur le contenu en lipides des juvéniles. Suite aux diverses collaborations internationales et discussions avec les gestionnaires, nous nous sommes rendus à l’évidence qu’il devait y avoir une recommandation sur les mesures de gestion vis-à-vis cette espèce. En fonction des connaissances scientifiques développées au fil des ans et plusieurs cas de gestion dans divers pays, nous avons recommandé aux gestionnaires d’éduquer les utilisateurs à vérifier, nettoyer, sécher et congeler leurs équipements. Que l’algue didymo soit une espèce exotique ou indigène, elle peut être propagée. De plus, la mise en valeur des saines pratiques permettent de limiter la propagation d’autres organismes pouvant potentiellement être plus dévastateurs que l’algue didymo. Les résultats de cette thèse contribuent à l’avancement des connaissances sur les facteurs de contrôle de l’algue didymo tant à l’échelle microscopique que macroscopique. En plus d’avoir contribué à élaborer le modèle conceptuel, nous l’avons amélioré en l’appliquant à diverses échelles spatiales : à l’échelle du tronçon et du bassin versant ainsi qu’à l’échelle régionale. C’est d’ailleurs la première étude à élaborer un réseau de suivi des proliférations par l’implication des acteurs du milieu ainsi qu’une des seules études à avoir examiné la variabilité interannuelle sur une période de six ans. C’est également la première étude à avoir évalué l’impact de l’algue didymo sur les communautés périphytiques dans l’est du Canada. De surcroît, c’est la première étude exhaustive qui a évalué l’effet des proliférations sur l’écologie du saumon atlantique juvénile. Les recherches menées sur le comportement alimentaire est également novateur et contribue considérablement à la compréhension des mécanismes et les interactions entre les divers niveaux trophiques et l’impact sur les salmonidés.<br /><br />Didymosphenia geminata is mat-forming nuisance diatom. This epilithic microscopic alga can secrete copious amounts of mucilage creating thick and extensive mats covering the entire riverbed. Once considered very rare and representative of pristine water quality, it is now ubiquitous in rivers around the world. Over the past two decades, this alga has emerged as a nuisance diatom and generated much interest among freshwater ecologists and river managers alike. Nonetheless, controlling factors governing the occurrence and severity of D. geminata are still not well understood. In Quebec, the first massive proliferations of D. geminata were reported in 2006 in the Matapedia River. There was an urgency to fill this knowledge gap as managers and government agencies panicked at the lack of knowledge regarding its distribution, the causes of its onset and mostly, fear of that this alga would act as an additional stressor to Atlantic salmon populations. To better understand controlling factors of both cell division and mat formation, a conceptual model was developed with the collaboration of several researchers working in Colorado, Vancouver Island, New Zealand and Quebec to collate observations and field data. This synthesis work helped identify the factors favoring its colonization, growth and persistence in rivers. The spatiotemporal dynamics of D. geminata are dependent on various dynamic thresholds of flow, temperature, light and water chemistry within the habitat window. We can therefore test various arrays of these parameters to determine whether rivers are likely to present occasional to persistent D. geminata mats. To test and improve upon the conceptual model at various spatial scales, we developed a monitoring network for D. geminata mat presence. Twenty-two different organizations totaling 70 volunteers were trained to identify and quantify the proliferations of D. geminata. Between 2010 and 2015, 1 228 observations were made. The analysis of this database helped determine that 71% of the variability of D. geminata severity is inversely proportional to the intensity of the spring flood. At the reach scale, habitat type was determinant for presence-absence of mats with a strong significant preference for riffles. At the provincial scale, we compared the distribution of D. geminata in three regions of eastern Quebec (i.e. Gaspesie, Bas-St-Laurent and Cote-Nord) against numerous water chemistry parameters for twentynine rivers. At the broad regional scale, pH was the most important factor governing the presence-absence of cells. Diatoms have very specific pH optima and the contrasting geologies between regions confirmed that low pH and high tannins and lignin are not favorable to the survival and growth of D. geminata. Habitat preferences for D. geminata mats are identical to those of juvenile Atlantic salmon (i.e. shallow, fast, clear, cool, low nutrient and coarse substrate). To determine the extent of the impact of the presence of mats on the food web, we verified its impact on the algal community. The dense interwoven matrix creates a suitable physical environment for small diatoms, thus promoting greater taxonomic diversity. Despite an increase in the complexity of the biofilm with mat accrual, there is no additional impact on the structure and taxonomic diversity of the samples. D. geminata effects to higher trophic levels have been suspected since the structure of the habitat is modified and could alter the structure and functioning of the ecosystem. We studied the foraging behaviour of juvenile Atlantic salmon in contrasting D. geminata severities. With increasing percent cover of the alga, juvenile Atlantic salmon switch from a drift-foray to benthic-foray strategy. This change in behavior is not due to limited drifting prey availability. Our results reinforce the notion that juvenile Atlantic salmon have high behavioral plasticity in response to habitat changes. To test the impact of mats on juvenile salmon diet, we used a stable isotope analysis approach. This approach provides an integrated view of resource use by fish. Contrasting isotopic signatures between the liver and muscle tissues confirm a change in the seasonal diet. Juvenile salmon sampled in the D. geminata impacted site have a less diversified and carbon-depleted diet composed mainly of small chironomids and Trichoptera. Juvenile salmon condition factors and C:N ratios were not different between sites, though their size is smaller in the impacted site. Future work should focus on the effects of reduced prey quality on lipid content of fish. Over various international collaborations and discussions with managers, it became clear that there was a need for a clear and standardized recommendation on management measures for this nuisance species. Based on recent scientific knowledge and several management cases in various countries, we recommended that managers educate users to check, clean, dry and freeze their equipment. Regardless of whether D. geminata is an exotic or native species, it can be spread. In addition, the development of best practices limits the spread of aquatic organisms as others may be far more devastating than D. geminata. The results of this thesis contribute to the advancement of knowledge on the controlling factors of D. geminata for both cells and mats. In addition to contributing to the development of the conceptual model, we have tested and improved it by applying it to a variety of spatial scales: at the reach, watershed and regional scales. This is the first study to develop a monitoring network for mat severity with the involvement of local stakeholders. It is also the first study to evaluate the impact of D. geminata on periphytic communities in eastern Canada. Moreover, it is the first comprehensive study that has evaluated the effect D. geminata mats on juvenile Atlantic salmon ecology. Research on foraging behavior is also innovative and contributes significantly to the understanding of the mechanisms and interactions by which D. geminata impacts various trophic levels and salmonids.

Les inondations constituent l’un des risques naturels les plus fréquents. Une inondation historique survenue au Québec s’est produite aux abords de la rivière Richelieu entre avril et juin 2011. Suite aux dommages et aux frais occasionnés par cet épisode, l’INRS s’est engagé avec la ville de Saint-Jean-sur-Richelieu afin de la doter d’un système de Gestion et d’Analyse du Risque d’Inondation (GARI) : le niveau d’eau constitue une donnée essentielle pour l’évaluation de l’ampleur d’une inondation en temps réel. Ce travail de recherche a pour objectif de détecter automatiquement la ligne d’eau sur des images acquises par des caméras placées aux abords de la rivière Richelieu par des méthodes d’analyse d’images et de convertir cette dernière en niveau d’eau par stéréoscopie (transfert des valeurs en pixels en unité métrique). Le premier volet de ce travail de recherche s’intéresse à l’extraction automatique de la ligne d’eau à partir de données acquises par des caméras de surveillance. Afin de répondre à cette exigence, des séquences d’images de durée déterminée (une minute) ont été enregistrées et une image moyenne a été calculée pour chaque séquence. L’image moyenne est ensuite segmentée puis classifiée à l’aide d’une technique supervisée. Pour la première séquence d’images à traiter, des sites d’entraînement décrivant les parties immergées et libres sont créées afin d’obtenir deux classes de sortie correspondant aux surfaces d’eau et au reste de l’image. Une détection de contour permet d’isoler les pixels de la ligne d’eau. Pour les séquences suivantes, l’algorithme utilise automatiquement la ligne d’eau créée à la séquence précédente pour définir les sites d’entraînements à la séquence considérée. Le taux de réussite d’extraction de la ligne d’eau avoisine 90%. Le second volet consiste à convertir les coordonnées des pixels de la ligne d’eau du système bidimensionnel image vers un système tridimensionnel dans un référentiel connu. Les observations par stéréoscopie permettent cette transformation via les procédures d’orientation intérieure des caméras et d’orientation extérieure des couples stéréoscopiques. Pour le modèle de caméra considéré, les tests effectués montrent que les erreurs de positionnement planimétrique de la ligne d’eau sont inférieures au mètre et l’erreur de positionnement altimétrique (niveau d’eau) est inférieure à 10 cm lorsque le rapport entre la base stéréoscopique et la distance avec l’objet à modéliser est supérieur à 0.1.<br /><br />Flooding is one of the most common natural hazards. A historic flood occurred in the vicinity of the Richelieu River (Province of Quebec, Canada) between April and June 2011. After the damages and expenses incurred by this episode, INRS and the city of Saint-Jean-sur-Richelieu decided to create a flood risk management and analysis system : water level is an essential factor for assessing magnitude of a flood in real time. This research project aims to automatically detect the water line on images acquired by cameras located near the Richelieu River with image analysis methods and to convert it to water level by stereoscopy (transfer of values from pixel unit to metric unit). The first part of this research work focuses on automatic extraction of water line from data acquired by surveillance cameras. In order to meet this requirement, fixed-duration (one minute) image sequences were recorded and an average image was calculated for each sequence. Average image is then segmented and classified using a supervised technique. For the first sequence of images to be processed, training sites describing submerged and free portions are created to obtain two output classes corresponding to water and the rest of the image. Then, a contour detection is used to determine water line. For the following sequences, the algorithm automatically uses water line created in the previous sequence to define training sites for the present sequence. Success rate of water line extraction is around 90%. Second part consists in converting pixels coordinates of water line known in a two-dimensional image system to a three-dimensional system in a known reference frame. Stereoscopic observations allow this transformation via interior and exterior orientations procedures. For the camera model considered in this project, tests highlight that planimetric positioning errors of the water line are lower than 1 m and altimetric positioning error (water level) is lower than 10 cm when the ratio between stereoscopic basis and distance to the object to be modeled is higher than 0.1.

La fréquence et l'ampleur des dommages causés par les inondations au Canada augmentent chaque année, nécessitant des solutions préventives. Les étages inférieurs des bâtiments sont particulièrement vulnérables, justifiant la nécessité de détecter avec précision les ouvertures comme les portes et les fenêtres. Les nuages de points obtenus par télémétrie mobile sont bien adaptés pour identifier la présence de ces ouvertures sur les façades de bâtiments. Toutefois, la présence d'occlusions dans les nuages de points causées par des objets obstruant le faisceau LiDAR rend cette tâche complexe. Le présent projet de maîtrise vise à répondre à cette problématique en proposant une approche pour détecter et caractériser les occlusions dans les scènes acquises dans les milieux résidentiels et ruraux. La conception de la méthode est articulée autour de deux volets : la segmentation sémantique pour étiqueter le nuage de points et la détection d'occlusions par lancer de rayons. La méthode a été validée à partir de données simulées car elles permettent un contrôle total sur l'environnement et facilitent la visualisation des résultats. Cette analyse a ensuite été transposée sur les résultats obtenus à partir du nuage de points réel. Les résultats obtenus par la méthode proposée ont démontré une capacité à détecter la présence d'occlusions dans les milieux résidentiels et ruraux. La combinaison de la segmentation sémantique pour l'étiquetage du nuage de points avec la détection d'occlusions par lancers de rayons a permis une identification robuste des occlusions. Ces résultats soulignent l'efficacité de la solution dans des contextes réalistes, renforçant ainsi sa pertinence pour la détection précise des ouvertures sur les façades, une étape cruciale pour évaluer les risques liés aux inondations. La transposition réussie de l'analyse des données simulées aux résultats du nuage de points réel valide la robustesse de l'approche et son applicabilité dans des scénarios du monde réel.

Abstract : Flood forecasting plays a pivotal role in effective water resource management and flood risk mitigation. Despite recent advancements, existing ensemble forecasting systems often grapple with issues of unreliability and under-dispersion. The uncertainty in ensemble forecasts can emanate from various sources, including inputs like precipitation, temperature, and streamflow, as well as initial conditions, model structure and parameters, and boundary conditions. Over the past two decades, numerous endeavors have been made to address bias and under-dispersion through the introduction of various statistical post-processing methods in meteorological and hydrological forecasts. However, a significant challenge lies in selecting the appropriate method and strategy in the forecast chain. Limited research has addressed the integration of pre-processing and post-processing in streamflow ensemble forecasting. Moreover, existing studies have generally focused on lumped hydrological models, while the performance of this integration in distributed hydrological models remains scarcely examined. On the other hand, streamflow data, often indirectly measured through rating curves, are particularly susceptible to errors. While various methods have been proposed to estimate rating curve uncertainty (RCU), the impact of RCU on streamflow forecasting system performance remains underexplored. In addition, current post-processing methods utilized in streamflow forecasting systems have disregarded the inherent uncertainty in observational streamflow data. Thus, this thesis delves into the potential and hurdles associated with considering different sources of uncertainty in flood forecasting systems and their impact on system performance. The primary objectives entail 1) assessing and identifying optimal scenarios from pre-processing of both temperature and precipitation and post-processing of streamflow forecasts approaches, or a combination thereof, and 2) evaluating the consideration of rating curve uncertainty on flood forecasting system performance through incorporating into targeted post-processing methodologies. The thesis focuses on a short-range ensemble streamflow forecasting framework spanning lead times of 1 to 5 days for the au Saumon watershed in southern Quebec, Canada. This watershed, being flood-prone, encountered significant challenges during the 2019 and 2017 flood events, including widespread inundation, road closures, and damage to property and infrastructure. Enhancing streamflow ensemble forecasts and upgrading flood forecasting systems holds the potential to greatly benefit decision-makers and the local populace. Chapter 4 (Results part 1), presented as a scientific article, thus delves into assessing and employing various pre- and post-processing strategy scenarios within the flood forecasting framework employing a spatially distributed hydrological model. By applying different statistical processing and bias correction techniques, the performance and quality of ensemble streamflow forecasts were evaluated across different scenarios. The findings highlight biases and under-dispersion as significant factors affecting raw ensemble forecasts. Pre-processing partially improves raw forecasts but doesn't fully address bias in under-dispersed forecasts. Combining pre- and post-processing enhances forecast skill and reliability, albeit with some variations compared to post-processing alone. Integrating flood events into the training dataset and optimizing its length improves the effectiveness of processing methods, underscoring the critical role of data management strategies in enhancing streamflow forecasting systems. Subsequently, Chapter 5 (Results part 2) , evaluates the consideration of rating curve uncertainty, derived from the Voting Point Method (VPM) and Bayesian Rating curve (BaRatin) estimation methods, on the performance of streamflow ensemble forecasts by integrating into a targeted post-processing approach using Weighted Ensemble Dressing (WED) and Cumulative Distribution Function Matching (CDFM) post-processing methods. Post-processing without RCU enhances forecasting system skill and reliability but overlooks the inherent uncertainty in observational data, posing concerns about its effectiveness. Conversely, integrating RCU improves forecast skill, reliability, and effectively addresses observational data uncertainty, offering a notable advantage over traditional post-processing methods. Comparing WED and CDFM post-processing methods highlighted nuanced differences in forecast outcomes, influenced by uncertainty estimation techniques. Both methods showed favorable accuracy metrics, with RCU integration, especially from VPM, notably enhancing forecast quality. However, variations in RCU from different estimation methods may lead to forecast underestimation or overestimation, warranting careful consideration. These findings collectively shed light on the potential and challenges associated with incorporating different sources of uncertainty into flood forecasting systems.

Cette thèse vise à améliorer notre compréhension du modèle hédonique et de son application sur les données des biens immobiliers afin d'étudier l'impact d'un événement / externalité / environnementale liée à la présence d'inondation sur la valeur des propriétés résidentielles. Étant donné que les données immobilières sont réparties dans l'espace et dans le temps, des "corrections" temporelles et spatiales sont nécessaires dans le processus de modélisation économétrique. La recherche prend appui sur l’équation de prix hédonique. L’analyse empirique recours également à l’estimateur de type différence de différences spatio-temporelles (STDID) afin d’étudier l’effet d’une inondation survenue en 1998 sur le prix des résidences dans la ville de Laval au Canada entre 1995-2007. Les résultats suggèrent que l’utilisation des informations sur les zones inondables dans le but d’évaluer l’impact des inondations sur les valeurs résidentielles n’est pas une approche nécessairement appropriée. Les conclusions suggèrent que la grande hétérogénéité des résultats notés dans la littérature n’est probablement pas étrangère à la façon de définir les résidences touchées par les inondations. Cela signifie que les recherches empiriques sur les effets des inondations sur la valeur immobilière mesurent en réalité la valeur liée à la perception du risque d'inondation et non l’effet réel de l'inondation. Les résultats suggèrent que les applications futures dans la littérature devront porter une attention particulière à la manière de définir les zones inondables et d’identifier les résidences réellement touchées.

Abstract Fluvial biogeomorphology has proven to be efficient in understanding the evolution of rivers in terms of vegetation succession and channel adjustment. The role of floods as the primary disturbance regime factor has been widely studied, and our knowledge of their effects on vegetation and channel adjustment has grown significantly in the last two decades. However, cold rivers experiencing ice dynamics (e.g., ice jams and mechanical breakups) as an additional disturbance regime have not yet been studied within a biogeomorphological scope. This study investigated the long‐term effects of ice dynamics on channel adjustments and vegetation trajectories in two rivers with different geomorphological behaviours, one laterally confined (Matapédia River) and one mobile (Petite‐Cascapédia River), in Quebec, Canada. Using dendrochronological analysis, historical data and aerial photographs from 1963 to 2016, this study reconstructed ice jam chronologies, characterized flood regimes and analysed vegetation and channel changes through a photointerpretation approach. The main findings of this study indicate that geomorphological impacts of mechanical ice breakups are not significant at the decadal and reach scales and that they might not be the primary factors of long‐term geomorphological control. However, results have shown that vegetation was more sensitive to ice dynamics. Reaches presenting frequent ice jams depicted high regression rates and turnovers even during years with very low floods, suggesting that ice dynamics significantly increase shear stress on plant patches. This study also highlights the high resiliency of both rivers to ice jam disturbances, with vegetation communities and channel forms recovering within a decade. With the uncertainties following the reach/corridor and decadal scales, future research should focus on long‐term monitoring and refined spatial scales to better understand the mechanisms behind the complex interactions among ice dynamics, vegetation and hydrogeomorphological processes in cold rivers.

ABSTRACT Flood risk management (FRM) involves planning proactively for flooding in high‐risk areas to reduce its impacts on people and property. A key challenge for governments pursuing FRM is to pinpoint assets that are highly economically exposed and vulnerable to flood hazards in order to prioritize them in policy and planning. This paper presents a novel flood risk assessment, making use of a dataset that identifies the location, dwelling type, property characteristics, and potential economic losses of Canadian residential properties. The findings reveal that the average annual costs are $1.4B, but most of the risks are concentrated in high‐risk areas. Data gaps are uncovered that justify replication through local validation studies. The results provide a novel evidence base for specific reforms in Canada's approach to FRM, with a focus on insurance that improves both implementation and effectiveness.

Abstract The flood-prone Saint John River (SJR, Wolastoq), which lies within a drainage basin of 55 110 km 2 , flows a length of 673 km from its source in northern Maine, United States, to its mouth in southern New Brunswick, Canada. Major industries in the basin include forestry, agriculture, and hydroelectric power. During the 1991–2020 reference period, the SJR basin (SJRB) experienced major spring flood events in 2008, 2018, and 2019. As part of the Saint John River Experiment on Cold Season Storms, the objective of this research is to characterize and contrast these three major spring flood events. Given that the floods all occurred during spring, the hypothesis being tested is that rapid snowmelt alone is the dominant driver of flooding in the SJRB. There were commonalities and differences regarding the contributing factors of the three flood years. When averaged across the upper basin, they showed consistency in terms of positive winter and spring total precipitation anomalies, positive snow water equivalent anomalies, and steep increases in April cumulative runoff. Rain-on-snow events were a prominent feature of all three flood years. However, differences between flood years were also evident, including inconsistencies with respect to ice jams and high tides. Certain factors were present in only one or two of the three flood years, including positive total precipitation anomalies in spring, positive heavy liquid precipitation anomalies in spring, positive heavy solid precipitation anomalies in winter, and positive temperature anomalies in spring. The dominant factor contributing to peak water levels was rapid snowmelt.

Droughts are increasingly recognized as a significant global challenge, with severe impacts observed in Canada's Prairie provinces. While less frequent in Eastern Canada, prolonged precipitation deficits, particularly during summer, can lead to severe drought conditions. This study investigates the causes and consequences of droughts in New Brunswick (NB) by employing two drought indices: the Palmer Drought Severity Index (PDSI) and Standardized Evapotranspiration Deficit Index (SEDI)– at ten weather stations across NB from 1971 to 2020. Additionally, the Canadian Gridded Temperature and Precipitation Anomalies (CANGRD) dataset (1979–2014) was utilized to examine spatial and temporal drought variability and its alignment with station-based observations. Statistical analyses, including the Mann–Kendall test and Sen's slope estimator, were applied to assess trends in drought indices on annual and seasonal timescales using both station and gridded data. The results identified the most drought-vulnerable regions in NB and revealed significant spatial and temporal variability in drought severity over the 1971–2020 period. Trend analyses further highlighted the intensification of extreme drought events during specific years. Coastal areas in southern NB were found to be particularly susceptible to severe drought conditions compared to inland regions, consistent with observed declines in both the frequency of rainy days and daily precipitation amounts in these areas. These findings underscore the need for targeted drought mitigation strategies particularly in NB’s coastal zones, to address the region’s increasing vulnerability to extreme drought events.

Les changements climatiques anthropogéniques posent des défis énormes pour toutes les sociétés humaines. Ces défis majeurs mettront à l’épreuve les capacités d’adaptation des États et de ses institutions et des communautés partout dans le monde et devront se résoudre par un élan de solidarité humaine afin d’en atténuer les conséquences. Le Canada connaît déjà un réchauffement climatique important. Le pays a d’ailleurs récemment été touché par des événements climatiques extrêmes : des canicules, des feux de forêt, une sécheresse anormale et des inondations dont l’intensité est prévue d’augmenter avec les changements climatiques anthropogéniques. La province du Québec a quant à elle été touchée par de fortes inondations entre 2017 et 2019. L’objectif principal de la présente étude vise à discuter la manière dont le paradigme écosocial peut faire évoluer le travail social en tant que champ de savoir et d’intervention dans un contexte de changements climatiques. Cette étude s’est appuyée sur des données issues de groupes focus réalisés avec des intervenants suite aux inondations survenues au Québec (2017-2019). Notre analyse vise les interventions réalisées en contexte d’inondations, dans le sud de la province, mise en œuvre par le système de santé. Les données ont été collectées lors d’entrevues de groupe réalisées avec des intervenants psychosociaux et des gestionnaires de CI(U)SSS au courant des mois d’octobre et de novembre 2019. Les thèmes suivants ont émergé des analyses: les caractéristiques des inondations de 2019, les divergences d’opinions vis-à-vis des changements climatiques, l’aide et le soutien apportés durant les inondations et la participation citoyenne. J’insisterai également sur l’exacerbation possible des inégalités sociales dans ce contexte. D’autres thèmes se sont également révélés importants : l’engagement des intervenants psychosociaux, la participation et la décentralisation des décisions politiques. Enfin, mes réflexions porteront sur les conséquences sociales qu’entrainent les inondations et sur les types de pratiques sociales qui s’avèrent pertinentes à l’ère des changements climatiques et dans un contexte d’urgence.

Seasonal snowpack deeply influences the distribution of meltwater among watercourses and groundwater. During rain-on-snow (ROS) events, the structure and properties of the different snow and ice layers dictate the quantity and timing of water flowing out of the snowpack, increasing the risk of flooding and ice jams. With ongoing climate change, a better understanding of the processes and internal properties influencing snowpack outflows is needed to predict the hydrological consequences of winter melting episodes and increases in the frequency of ROS events. This study develops a multi-method approach to monitor the key snowpack properties in a non-mountainous environment in a repeated and non-destructive way. Snowpack evolution during the winter of 2020–2021 was evaluated using a drone-based, ground-penetrating radar (GPR) coupled with photogrammetry surveys conducted at the Ste-Marthe experimental watershed in Quebec, Canada. Drone-based surveys were performed over a 200 m2 area with a flat and a sloped section. In addition, time domain reflectometry (TDR) measurements were used to follow water flow through the snowpack and identify drivers of the changes in snowpack conditions, as observed in the drone-based surveys. The experimental watershed is equipped with state-of-the-art automatic weather stations that, together with weekly snow pit measurements over the ablation period, served as a reference for the multi-method monitoring approach. Drone surveys conducted on a weekly basis were used to generate georeferenced snow depth, density, snow water equivalent and bulk liquid water content maps. Despite some limitations, the results show that the combination of drone-based GPR, photogrammetric surveys and TDR is very promising for assessing the spatiotemporal evolution of the key hydrological characteristics of the snowpack. For instance, the tested method allowed for measuring marked differences in snow pack behaviour between the first and second weeks of the ablation period. A ROS event that occurred during the first week did not generate significant changes in snow pack density, liquid water content and water equivalent, while another one that happened in the second week of ablation generated changes in all three variables. After the second week of ablation, differences in density, liquid water content (LWC) and snow water equivalent (SWE) between the flat and the sloped sections of the study area were detected by the drone-based GPR measurements. Comparison between different events was made possible by the contact-free nature of the drone-based measurements.

Abstract This study aims to isolate and quantify the role of shrinking glaciers in recent hydrological changes in eight watersheds in the southwestern Yukon (Canada) by using an original dual approach that consists of (i) watershed hydrological regime identification, followed by a trend analysis of discharge time series, and (ii) a model‐based peak water (PW) analysis using glacier cover change measurements. A distinction between hydrological regimes is a necessary add‐up to commonly used trend attribution methods as the lake runoff regime shares common characteristics with the glacier regime. Results show a link between shrinking glaciers and hydrological changes in the region, but the link is complex, and glacier retreat does not explain all the observed changes. Model outputs show that the two watersheds with a glacierized area exceeding 30% and one watershed with 2.9% glacierized area have not reached PW, whereas a 9.2% glacierized watershed and another watershed with 2.1% glacierized area have already passed it. These results suggest that glacierized area alone cannot explain short‐term changes related to watershed current position in terms of PW, and the rate of glacier retreat must be considered. By contrast, the actual rate of glacier retreat does not influence long‐term changes, such as the magnitude of PW and of the consequent drop in discharge. Once glaciers will have retreated to a point close to extinction, declines in summer discharge from 10% to 70% and proportional to the actual glacier cover are anticipated at watersheds that are currently more than 9% glacierized. , Plain Language Summary In this study, we aim to understand how shrinking glacier cover affects river discharges. In conditions of continuous retreat, glaciers produce an initial increase in runoff as they lose mass. The discharge then reaches a turning point, a plateau called peak water, and subsequently declines as the volume of glacial ice continues to decrease. When analyzing eight watersheds with different glacier covers in the southwestern Yukon, we found that two watersheds that are 30% covered by glaciers have not yet reached this plateau, and therefore, the discharge will continue to increase. Several watersheds with smaller glacierized portions have passed peak water, which means that the discharge will now continue to decrease. We were also able to estimate the magnitudes of these changes in discharge. We show that two watersheds with 30% glacierized area can still experience a 1.5‐ to 2‐fold increase in discharge and that watersheds currently more than 9% glacierized are predicted to show noticeable changes after peak water, with the possibility of discharge decreasing by a factor of 3 to 5 by the time glaciers have retreated to a point when their hydrological influence at the watershed scale becomes insignificant. , Key Points Noticeable acceleration of glacier retreat occurred in southwestern Yukon since 1999 with measured consequences for the regional hydrology Various hydrological changes have been detected at the study watersheds. Glacier retreat explains many but not all of those changes Long‐term hydrological changes are glacier cover dependent while decadal‐scale changes are driven by glacier retreat rate

Abstract Temporal variations in concentrations of pathogenic microorganisms in surface waters are well known to be influenced by hydrometeorological events. Reasonable methods for accounting for microbial peaks in the quantification of drinking water treatment requirements need to be addressed. Here, we applied a novel method for data collection and model validation to explicitly account for weather events (rainfall, snowmelt) when concentrations of pathogens are estimated in source water. Online in situ β ‐ d ‐glucuronidase activity measurements were used to trigger sequential grab sampling of source water to quantify Cryptosporidium and Giardia concentrations during rainfall and snowmelt events at an urban and an agricultural drinking water treatment plant in Quebec, Canada. We then evaluate if mixed Poisson distributions fitted to monthly sampling data ( = 30 samples) could accurately predict daily mean concentrations during these events. We found that using the gamma distribution underestimated high Cryptosporidium and Giardia concentrations measured with routine or event‐based monitoring. However, the log‐normal distribution accurately predicted these high concentrations. The selection of a log‐normal distribution in preference to a gamma distribution increased the annual mean concentration by less than 0.1‐log but increased the upper bound of the 95% credibility interval on the annual mean by about 0.5‐log. Therefore, considering parametric uncertainty in an exposure assessment is essential to account for microbial peaks in risk assessment.

Abstract In northern hardwood forests, tree markers select the trees to be harvested during logging operations using classification systems that assign harvest priorities based on the presence of a wide range of individual defects. According to the most recent advances in our understanding of the impact of defects on both tree vigour (the risk of mortality or decline in growth) and quality (the potential for recovering valuable sawlogs), tree markers should adopt a simpler classification system that considers fewer defects than the current operational practice, and they should prioritize the removal of trees with crown dieback. Since the probability of developing defects and dying increases substantially with tree diameter, tree markers should also favour the removal of larger trees that have maintained their quality. However, these recommendations were developed based on tree-level analyses. To provide further validation at the stand scale, we compared stand improvement and value recovery under three tree marking regimes: a new, simplified regime based on the recommendations above, and two regimes used in the province of Quebec, Canada. To do so, we conducted tree marking simulations and value recovery assessments in 14 managed stands distributed across the northern hardwood range of Quebec. Our results confirmed that the simplified tree marking regime not only facilitated stand improvement by removing a greater proportion of low-vigour trees, but also recovered significantly more value (17% on average) at the stand scale. By prioritizing the removal of trees with crown dieback, the simplified regime was superior at salvaging the current value of low-vigour trees before they die or decline in quality. Based on our results, we propose simplified and empirically-validated tree marking guidelines for northern hardwood forests.

Multi-cohort forest management in northern hardwood stands may well be the best way to successfully regenerate tree species of intermediate shade tolerance, such as yellow birch (Betula alleghaniensis Britt.). The creation of large enough gaps in the canopy favors increased light availability within the opening, while soil scarification provides suitable germination seedbeds. Evidence of these methods’ success nonetheless remains mostly the purview of experimental studies rather than operational tests. In Quebec, Canada, the multi-cohort methods promoted include group selection cutting and patch cutting. The present study tested their implementation at an operational scale and over a large territory in both hardwood-dominated and mixedwood stands. We assessed their efficacy in promoting natural regeneration of commercial hardwood trees, notably yellow birch and sugar maple (Acer saccharum Marsh.). We conducted regeneration surveys at 2, 5, 10, and 15 years after harvest. Overall, group selection and patch cuttings were successful in regenerating the target species. Yellow birch, for instance, showed a mean stocking around 60% and a mean sapling density around 3400 stems ha−1 after 15 years. We compared several variables for measuring regeneration in early years, and found that the relative abundance, the stocking based on one stem per sampling unit, and the mean maximum height were good predictors of the relative presence of yellow birch and sugar maple in 15-year-old canopy openings. Using smaller sampling units (6.25 m2 rather than 25 m2) and waiting until year 5 may be more useful for making such predictions. In addition, there was an important turnover in vertical dominance in these openings. Non-commercial woody competitors were frequently dominant in early years but were often replaced by commercial hardwoods, notably yellow birch. We propose certain thresholds for assessing the success of post-harvest regeneration and for evaluating the need for a cleaning treatment.