No description available.

No description available.

The radiance distribution provides complete information regarding the geometrical structure of the ambient light field within the ocean. A quantitative study of the radiance field in the dynamic ocean water is presented in this thesis work. The study starts with the development of a novel radiance camera for the measurement of the full spherical radiance distribution at the ocean surface and depth. Nonlinear response functions are designed and advanced radiometric calibrations are developed. The resulting camera measures the radiance distribution in absolute units over an extremely high dynamic range at fast rates. With the newly obtained radiance data, I have examined the fine structure of both the downwelling and upwelling radiance distribution and its variation with depth in optically diverse water types. The fully specified radiance distribution data are used to derive all apparent optical properties and some inherent optical properties including the absorption coefficient. With the camera fixed at shallow depths, I have observed and determined the sea surface wave disturbance of the radiance distribution. It is found that the radiance fluctuates anisotropically with regard to its amplitude and periodicity. Typical spatial structures of the dynamic radiance field are identified and shown relevant to the surface waves and the solar zenith angles. The variability in the radiance field also propagates to the irradiance field; the variability is pronounced in measured irradiance depth profiles in the upper layers of the ocean. The statistics of the irradiance fluctuations along the water depth, including the dominant frequency and coefficient of variation, are derived using wavelet techniques and fitted to novel analytic models. The results from the irradiance depth-profile decomposition are in agreement with theoretical models and other independent measurements. This thesis work represents the first attempt to quantify the full light field and its variability in dynamic ocean waters and is of significant relevance to many other optics-related applications.

Les contaminants, dissouts ou adsorbés sur les particules, sont principalement délivrés au milieu marin par les fleuves. La dynamique sédimentaire constitue alors un proxy de la dynamique de ces contaminants. Cette thèse s'inscrit dans le projet ANR AMORAD, et se focalise sur la dynamique du matériel particulaire délivré par le Rhône au Golfe du Lion (Méditerranée nord-occidentale), principal contributeur d'apports solides au Golfe (80 % des sédiments). Alors que des études antérieures ont permis de bien représenter les processus au niveau du fond, les processus régissant les comportements des matières en supension (MES), majoritairement rencontrées dans le panache turbide du Rhône, sont encore mal apréhendés. En vue de mieux décrire la dynamique de ces MES et d'améliorer les modèles hydrosédimentaires existants, l'objectif est de mieux caractériser ces particules. À cette fin, un vaste jeu de données issu de capteurs déployés in situ (données collectées pour 12 campagnes en mer, réalisées de 2011 à 2016) a été exploité, permettant d’obtenir une vision 2D verticale mais seulement ponctuelle (spatialement et temporellement). De façon complémentaire, une base de données d'images satellitaires (donnée couleur de l'eau du capteur MERIS-300m acquise entre 2002 et 2012), offrant une vue plus synoptique et long terme mais uniquement en surface, a été exploitée. Le jeu de données d'images satellitaire (plus de 800 images) a été traité de façon innovante par l'application d'un traitement semi-automatique permettant l'extraction de différentes métriques du panache turbide du Rhône (e.g. aire, limites d'extension, forme, centres géométriques, concentrations). La distribution spatiale et les caractéristiques physiques des MES telles que leur concentration dans l'eau, leur diamètre médian ou encore leur vitesse de chute ont été étudiées et estimées en fonction des différents forçages hydrométéorologiques actifs sur la zone d'étude (e.g. débit du Rhône, vents dominants). Un nouveau modèle hydrosédimentaire reposant sur le couplage du modèle hydrodynamique MARS-3D et du module sédimentaire multiclasse MIXSED a été configuré et les données in situ et satellitaires ont pu être mobilisées afin de contraindre la vitesse de chute des sédiments, paramètre clef de la modélisation de la dynamique hydrosédimentaire. / Contaminants, which can be dissolved in water or adsorbed on particles, are mainly delivered to the coastal environment by rivers. Thus, sediment dynamics reperesent a relevant proxy of contaminants dynamics. ThisPhD thesis is part of the ANR AMORAD project, of which one workpackage focuses on the fate of sediments in the coastal environment. This work focuses on the dynamics of Rhône River sediments in the Gulf of Lions (north-western mediterranean), this river delivering 80 % of the sediments of the Gulf. While previous studies over the area allowed a better understanding of physical processes at the water-sediment interface, processes driving suspended particulate matter (SPM) dynamics are still poorly understood. To better describe this SPM dynamics and improve hydrosedimentary models, the aim is to better characterize these particles. To this end, a large dataset collected from in situ deployed sensors (data collected for 12 field campaigns, conducted from 2011 to 2016) was analyzed to get a 2D vertical but ponctual view (both spatially and temporally). Complementary, a satellite images dataset (MERIS-300m ocean colour archive from 2002 to 2012) was built in order to get a long term and more synoptic view (but limited to surface).This dataset (more than 800 images) was originaly studied, applying a semi-empirical process to extract various Rhône River turbid plume metrics (e.g. area of extension, south-east-westernmost points, shape, centroids, SPM concentrations). Plume metrics and physical properties of SPM such as their concentration in water, their median diameter or their settling velocity were investigated regarding the different hydrometeorological forcings (e.g. Rhône River discharge, prevailing winds). A new hydrosedimentary model, based on the coupling of the 3D hydrodynamical model MARS-3D and the sedimentary module MIXSED, was set and ocean color and in situ data were used to constrain the settling velocity of particles, key parameter of hydrosedimentary modelling.