Actuellement, la Loire, voit sa géodynamique de plus en plus détériorée. Le déficit du transport solide provoque un enfoncement du lit mineur sur plusieurs secteurs du bassin Amont. Ceci a pour conséquence une déstabilisation progressive des ouvrages (digues, ponts) et l'érosion ou le colmatage des berges ainsi qu'une baisse du niveau piézométrique des nappes alluviales. Le traitement de ces problèmes passe par une meilleure connaissance des processus d'érosion et d'incision fluviale, le premier laissant place au second une fois que la couche « alluviale » a été totalement emportée, laissant apparaître un affleurement rocheux dans le lit du cours d'eau. Ce phénomène complexe est actuellement constaté en divers endroits de la Loire, dans la plaine du Forez. Nous proposons trois approches pour l'analyser.La première approche est fondée sur la modélisation numérique utilisant les équations de Barré-Saint-Venant, pour l'écoulement, couplées aux équations d'Exner et de Meyer-Peter Müller pour le transport solide. Ce modèle monodimensionnel permet d'obtenir la côte du fond du lit de la rivière et le flux solide au droit de chaque point de calcul.La seconde approche, mécaniste, consiste à déterminer de façon semi-empirique le taux d'incision du substratum marneux à partir de deux variables majeures: la puissance hydraulique totale et le coefficient d'abrasion en un point donné. La puissance hydraulique est calculée à l'aide d'un modèle hydraulique tandis que le nombre d'abrasion est une propriété mécanique de la marne qui est déterminée à partir d'essais en laboratoire. La troisième approche relève de la géochimie. Elle consiste à déterminer les provenances des matériaux contribuant à la recharge latérale des sédiments, à partir de leurs signatures géochimiques. Ceci a fait l'objet d'analyses en laboratoire sur des échantillons prélevés sur l'ensemble du linéaire entre Grangent et Balbigny. Les résultats obtenus montrent que ces approches indépendantes sont complémentaires et permettent une description à la fois qualitative et quantitative de l'enfoncement du lit de la Loire dans le secteur d'étude. / Currently, the Loire river, sees its geodynamic increasingly deteriorated. The deficit of sediment transport causes erosion of the bed on several areas of the basin. This results in a gradual destabilization of structures (dams, bridges), erosion or clogging of banks or a decline in piezometric level alluvial.Treating these problems requires a better understanding of the processes of erosion and river incision, the second succeeding the first, once the alluvial material of the bottom was completely removed, revealing a bed outcrop.This complex phenomenon is currently found in various parts of the Loire river, in the plain of Forez.We propose three approaches for this analysis.The first approach is based on numerical modeling using the equations of Barre-de-Saint-Venant, for flow, coupled with the equations of Exner and Meyer-Peter Müller for sediment transport. This monodimensional model allows the simulation of riverbed changement and sediment discharge, right each calculation point of grid mesh.The second approach is mechanistic and consists of determining the rate of marly bedrock incision by a semi-empirical method by the use of two major variables: the total hydraulic power and an abrasion coefficient. The hydraulic power is calculated using a hydraulic model, while the abrasion coefficient is a mechanical property of the marl which is determined from laboratory tests.The third approach is the geochemistry. It consists in determining the provenance of the materials from tributaries and in the main chanel by analyzing their geochemical signatures. This has been the subject of laboratory tests on samples taken across the linear from Grangent to Balbigny.The results obtained show that these independent approaches are complementary and provide both a qualitative and quantitative description of the incision of the Loire river in the study area.

In the James Bay Lowlands of northern Ontario, kimberlite pipes are concealed by peat, thick layers of till, and Tyrell sea sediments. Studies have shown that buried ore bodies produce geochemical signatures in surface media. This thesis explores the geochemistry of near-surface groundwater above concealed kimberlite pipes using factor analysis to determine whether (1) a factor analysis can reveal an underlying structure (factors) in a multivariate groundwater geochemical dataset, and whether (2) those factors are related to the presence of concealed kimberlite. Factor analysis was performed on two datasets of nearsurface groundwater, collected at 0.2 m and 1.1 m below ground surface in peat. Results revealed that (1) there is a significant difference in the behaviour of elements in groundwater near the surface compared to those in deeper groundwater, which is sheltered from the effects of the atmosphere; (2) for both datasets, the first factor is dominated by elements known to be enriched in kimberlite, notably rare earth elements (REE), U, Th, Ti – the composition of factor one is consistent with their derivation from kimberlite in a limestone background where such elements are in very low concentration; (3) high-valence and lowvalence kimberlite indicator elements (KIE) are found separated into distinct factors suggesting that once released from the kimberlite after weathering, KIE are subjected to various geochemical processes to be differentiated as they migrate upward to the surface; and (4) Fe and Mn load on a factor distinct from other metals, suggesting that in this environment Fe-Mn-O-OH is not a significant controller of metal mobility in groundwater. Overall, this research has further highlighted the multivariate nature of geochemical processes in groundwater. Compared with previous work in geochemical exploration where often only univariate or bivariate statistics or single element profiles over concealed ore bodies were used, this thesis has shown that factor analysis, as a multivariate data analysis technique, is a robust exploration tool, able to shed light on relevant geochemical processes hidden within geochemical datasets. This thesis shows that high-valence KIE, notably U,V, Th, Ti and the REE, as a group, are better indicators of the presence of kimberlites than other well-known KIE. Single element concentration profiles such as Ni or Cr (known KIE) show similar anomalies over a concealed kimberlite as a factor score profile for factor one (U, V, Th, Ti, REE, Ni) would; however, it is the peculiar assemblage of elements in factor one that makes it unique to kimberlites, a feature that can be used in future exploration work for concealed kimberlites in similar surficial environments, such as the Siberian wetlands. The results suggest that future geochemical exploration work involving groundwater should focus on the more stable groundwater located below the zone of oxidation, sheltered from the effects of the atmosphere.

In the James Bay Lowlands of northern Ontario, kimberlite pipes are concealed by peat, thick layers of till, and Tyrell sea sediments. Studies have shown that buried ore bodies produce geochemical signatures in surface media. This thesis explores the geochemistry of near-surface groundwater above concealed kimberlite pipes using factor analysis to determine whether (1) a factor analysis can reveal an underlying structure (factors) in a multivariate groundwater geochemical dataset, and whether (2) those factors are related to the presence of concealed kimberlite. Factor analysis was performed on two datasets of nearsurface groundwater, collected at 0.2 m and 1.1 m below ground surface in peat. Results revealed that (1) there is a significant difference in the behaviour of elements in groundwater near the surface compared to those in deeper groundwater, which is sheltered from the effects of the atmosphere; (2) for both datasets, the first factor is dominated by elements known to be enriched in kimberlite, notably rare earth elements (REE), U, Th, Ti – the composition of factor one is consistent with their derivation from kimberlite in a limestone background where such elements are in very low concentration; (3) high-valence and lowvalence kimberlite indicator elements (KIE) are found separated into distinct factors suggesting that once released from the kimberlite after weathering, KIE are subjected to various geochemical processes to be differentiated as they migrate upward to the surface; and (4) Fe and Mn load on a factor distinct from other metals, suggesting that in this environment Fe-Mn-O-OH is not a significant controller of metal mobility in groundwater. Overall, this research has further highlighted the multivariate nature of geochemical processes in groundwater. Compared with previous work in geochemical exploration where often only univariate or bivariate statistics or single element profiles over concealed ore bodies were used, this thesis has shown that factor analysis, as a multivariate data analysis technique, is a robust exploration tool, able to shed light on relevant geochemical processes hidden within geochemical datasets. This thesis shows that high-valence KIE, notably U,V, Th, Ti and the REE, as a group, are better indicators of the presence of kimberlites than other well-known KIE. Single element concentration profiles such as Ni or Cr (known KIE) show similar anomalies over a concealed kimberlite as a factor score profile for factor one (U, V, Th, Ti, REE, Ni) would; however, it is the peculiar assemblage of elements in factor one that makes it unique to kimberlites, a feature that can be used in future exploration work for concealed kimberlites in similar surficial environments, such as the Siberian wetlands. The results suggest that future geochemical exploration work involving groundwater should focus on the more stable groundwater located below the zone of oxidation, sheltered from the effects of the atmosphere.

In the James Bay Lowlands of northern Ontario, kimberlite pipes are concealed by peat, thick layers of till, and Tyrell sea sediments. Studies have shown that buried ore bodies produce geochemical signatures in surface media. This thesis explores the geochemistry of near-surface groundwater above concealed kimberlite pipes using factor analysis to determine whether (1) a factor analysis can reveal an underlying structure (factors) in a multivariate groundwater geochemical dataset, and whether (2) those factors are related to the presence of concealed kimberlite. Factor analysis was performed on two datasets of nearsurface groundwater, collected at 0.2 m and 1.1 m below ground surface in peat. Results revealed that (1) there is a significant difference in the behaviour of elements in groundwater near the surface compared to those in deeper groundwater, which is sheltered from the effects of the atmosphere; (2) for both datasets, the first factor is dominated by elements known to be enriched in kimberlite, notably rare earth elements (REE), U, Th, Ti – the composition of factor one is consistent with their derivation from kimberlite in a limestone background where such elements are in very low concentration; (3) high-valence and lowvalence kimberlite indicator elements (KIE) are found separated into distinct factors suggesting that once released from the kimberlite after weathering, KIE are subjected to various geochemical processes to be differentiated as they migrate upward to the surface; and (4) Fe and Mn load on a factor distinct from other metals, suggesting that in this environment Fe-Mn-O-OH is not a significant controller of metal mobility in groundwater. Overall, this research has further highlighted the multivariate nature of geochemical processes in groundwater. Compared with previous work in geochemical exploration where often only univariate or bivariate statistics or single element profiles over concealed ore bodies were used, this thesis has shown that factor analysis, as a multivariate data analysis technique, is a robust exploration tool, able to shed light on relevant geochemical processes hidden within geochemical datasets. This thesis shows that high-valence KIE, notably U,V, Th, Ti and the REE, as a group, are better indicators of the presence of kimberlites than other well-known KIE. Single element concentration profiles such as Ni or Cr (known KIE) show similar anomalies over a concealed kimberlite as a factor score profile for factor one (U, V, Th, Ti, REE, Ni) would; however, it is the peculiar assemblage of elements in factor one that makes it unique to kimberlites, a feature that can be used in future exploration work for concealed kimberlites in similar surficial environments, such as the Siberian wetlands. The results suggest that future geochemical exploration work involving groundwater should focus on the more stable groundwater located below the zone of oxidation, sheltered from the effects of the atmosphere.