Sur la région de la Crau, le système l’irrigation gravitaire appliqué aux prairies de foin joue un rôle important dans le maintien du cycle hydrologique en étant le principal contributeur à la recharge de la nappe souterraine de la Crau (70 à 80% de la recharge). Dans le futur, des pressions sur la ressource en eau alimentée ce système d’irrigation risquent de s’accroître du fait des changements climatiques et de l’augmentation des autres usages de l’eau (domestiques,industriels, ..) et induisent la nécessité de l’optimisation de ce système afin de maintenir l’état des ressources en eau souterraines. Cette optimisation nécessite la connaissance des vrais quantités d’eau apportées à la parcelle qui sont mal connues, et de spatialiser ces quantités à l’échelle du territoire. Ainsi, l’objectif de ce travail est de développer une méthodologie permettant d’améliorer la quantification de ces volumes à deux échelle (parcellaire et régionale). Un système numérique incluant un modèle d’irrigation gravitaire dénommé ‘Calhy’ a été développé au cours de ce travail. Ce système permet de caractériser le fonctionnement des principaux processus intervenant dans ce système d’irrigation (infiltration de l’eau dans le sol et la propagation à la surface de la parcelle). Mais l’estimation de ces processus est limité par la connaissance de certains paramètres non mesurables, tels que la conductivité hydraulique à saturation du sol et la rugosité hydraulique de la surface parcellaire. Une analyse de sensibilité AS a été menée dans un premier temps au cours de ce travail afin de définir la contribution de la variation de chaque paramètre non mesurable sur la variance des variables de sortie. Les résultats montrent la possibilité d’estimation ces paramètres à partir des accessibles variables auxquels ils sont sensibles. Ainsi, une méthode d’inversion s’est basé sur les résultats d’AS, combine le modèle Calhy et un dispositif expérimental a été appliqué dans un second temps pour l’estimation paramètres et valider l’approche proposée. Les résultats montrent que cette approche est robuste et efficace pour estimer ces paramètres. A la fois les paramètres ont été issus pour démarrer le système (estimés ou mesurés), nous avons étudié différentes modifications du système d’irrigation actuel (changement de la pente de la parcelle, du sens d’irrigation, l’apport de l’eau en différents points de la parcelle, ….), et leurs impacts sur l’homogénéité de l’infiltration et la durée de l’irrigation.296En parallèle, Nous avons établi un modèle empirique de dose d’irrigation basé sur l’analyse des pratiques d’irrigation investigués auprès des enquêtes procédés chez les agriculteurs. Différents modèles empiriques ont été développés en basant sur des régressions calculant la dose d’irrigation et la durée en fonction du débit disponible et des paramètres parcellaires caractérisant les conditions d’irrigation tels que la géométrie de la parcelle (longueur, largeur et surface). Le modèle de dose empirique investigué au cours de ce travail permet de fournir une estimation de la dose distribué sur tout le territoire de la Crau en intégrant ce modèle dans le simulateur de laCrau. / Worldwide, the irrigation accounts for 70% of all water consumption: understanding therelationship between irrigation and ecosystems and optimizing the irrigation practices cancontribute to the sustainability management of water resources. In the Crau region (southern ofFrance), the flooding irrigation system used for irrigating the hay fields plays an important role inwater cycle: in this system, considerable amounts of water are brought to the hay fields (about 20000 m3/ha/year i.e. 2 000 mm), which participate strongly to the recharge of the Crau aquifer(between 66% and 80% of the recharge). In the future, the pressures on the availability of waterresources that feed this irrigation system (the reservoir of Serre Ponçon) may increase because ofthe climate change and the increase in the another water uses. Thus, it becomes necessary tooptimize the irrigation practices in order to conserve the water and ensure a sufficient rechargefor aquifer of the Crau. This optimization requires i) the knowledge of the amount of waterbrought to the plot that are not currently known, ii) spatialize these amounts over the regionalscale. This work aims to develop a methodology to improve the quantification of these volumesat the field and regional scales. A numerical system that includes a flooding irrigation modelcalled 'Calhy' was developed, takes into account the main processes involved in this irrigationsystem (water infiltration into the soil and the runoff of water slide over the plot surface). Firstly,a sensitivity analysis was conducted in order to classify the Calhy’s parameters according to theirimportance and define an optimal experimental apparatus allowing to estimate them using aninverse approach. Secondly, an inversion procedure based on the proposed experimental297apparatus and the previous model was implemented on several plots in the study area. The resultsshow that the important parameters can be estimated and then Calhy can be used to analyse andoptimize irrigation practices. Then, different optimization scenarios were identified. In parallel,we developed an empirical model of irrigation dose based on the analysis of irrigation practicesin a group of exploitations in the study region. Different empirical models were developed;regressions were used to compute the irrigation dose and duration from geometricalcharacteristics of the borders (length, width and surface) and available water inflow rate. Theempirical model of irrigation dose developed in this work would provide a spatial estimation ofirrigation doses overall plots in the study region, and would contribute to a better quantificationof water recharge of the Crau aquifer and its locations.

Les interactions entre la nappe phréatique peu profonde et les réseaux d’assainissement sont en général très complexes. Dans le but d’étudier l’impact des variations du niveau de la nappe phréatique sur les dégradations des conduites, un modèle hydrogéologique numérique tridimensionnel (3D) de la nappe phréatique ayant un aquifère multicouche au sein d’une zone de 83 km2 de l’agglomération bordelaise (Mérignac, Pessac, Talence) a été développé, calibré et validé. Pour ce faire, un modèle conceptuel de cet aquifère multicouche a dans un premier temps été défini sur la base des descriptions lithostratigraphiques des forages. Puis un travail géostatistique a été entrepris pour construire les limites des formations géologiques et les cartes piézométriques issues de campagnes de mesures de hauteurs d’eau de la nappe phréatique. Nous avons par la suite développé une méthodologie originale de transformation de l’information lithologique au droit des sondages en perméabilités, puis en champs de perméabilités en utilisant des méthodes géostatistiques. Dans un second temps, les conditions hydrogéologiques ont été simulées en utilisant le logiciel Visual Modflow ©. Les paramètres hydrogéologiques ont été calibrés manuellement et automatiquement sur la base des niveaux d’eau mesurés lors des campagnes de juillet 2010, octobre 2010 et mars 2011. La validation du modèle qui s’est effectuée en simulant les cotes piézométriques au pas mi-décadaire allant de juin 1993 à mai 2012 montre que les chroniques piézométriques observées sur les 35 piézomètres du domaine étudié sont correctement restituées et la dynamique de la nappe phréatique est reproduite. Dans un dernier temps, les distributions spatiales et temporelles des cotes piézométriques simulées ont permis d’identifier les zones d’ennoiement des réseaux d’assainissements en comparant les cotes piézométriques aux cotes des réseaux. Ces zones d’ennoiement ont été comparées aux principaux secteurs de casses de canalisations. Les secteurs de casses de canalisations sont localisés dans les zones à faible probabilité d’ennoiement des conduites. L’environnement naturel hydrogéologique pris comme paramètre isolé ne constitue donc pas un facteur explicatif au phénomène de casses des canalisations. / The interactions between shallow groundwater and sewer network are generally complex. The aim of this research is to study the impact of the water table level variation on the damages of the buried pipes. For this purpose, a three-dimensional (3D) numerical groundwater flow model of the shallow multi-aquifer of the Bordeaux urban area, within a 83 km2 area (Mérignac, Pessac, Talence), is developed, calibrated and validated. In order to do this, in a first step, a conceptual model was built based on stratigraphic descriptions of the boreholes. A geostastistical study was carried out to build the limits of the geological layers and the piezometrical maps from the measurement campaigns of water table levels. Then, an original methodology to transform the description of lithological units to local hydraulic conductivity values is proposed and the reconstruction of hydraulic conductivity fields at the urban scale was performed based on geostatistical methods. In a second step, hydrogeological conditions were simulated using the finite-difference groundwater flow model MODFLOW-2005 (Visual Modflow ©). The hydrogeological parameters were then calibrated manually and automatically based on water table elevation data measured in July 2010, October 2010 and March 2010. The model validation covered 6890 days, from June 1993 to may 2012. The results showed that the simulated heads are quite accurate and reproduce the main dynamics of the system at 35 piezometers. The spatial and temporal distribution of the heads of the piezometers simulated permitted to identify the flooding zones of the sewer network by comparing the piezometric head with the altitude of the networks. The sectors where the pipes are damaged were localized in the zone wherein there is low probability of flooding. Additionally, the natural hydrologic environment, taken as a sole factor, cannot explain the pipe breakage phenomena.

Le sujet de l’adaptation aux changements climatiques est devenu l’un des sujets contemporains les plus importants dans la vigne. Une grande focalisation a été mise sur la compréhension des effets du porte-greffe sur la croissance du scion, l’absorption des nutriments, et la tolérance au stress, dans l’objectif final de développer de nouveaux porte-greffes qui facilitent l’adaptation au changement climatique. L’objectif de cette thèse est d’examiner comment les différences dans la résistance à la sécheresse entre les génotypes peut résulter en de grandes différences dans leur capacité à maintenir leur croissance racinaire en situation de stress. Une meilleure compréhension sur la manière dont la structure, la croissance racinaire et l’absorption d’eau répondent au stress nous permettra de mieux comprendre quels sont les aspects de la physiologie racinaire qui contribuent à la tolérance face à la sécheresse. Des recherches précédentes qui s’étaient focalisées sur l’absorption d’eau racinaire chez la vigne ont suggéré que l’absorption d’eau racinaire pouvait être fortement liée à la vitesse de croissance racinaire instantanée (voir Gambetta et al. 2013). Cette observation implique que des différences entre les génotypes dans la résistance face à la sécheresse pourrait largement résulter de leur capacité à maintenir la croissance racinaire en conditions de stress. Deux porte-greffes de vigne avec des capacités contrastées en matière de résistance à la sécheresse, le Riparia Gloire de Montpellier (RGM) et le 110 Richter (110R) ont été sélectionnés pour étudier dans cette thèse. RGM est considéré comme sensible à la sécheresse, tandis que 110R est fortement résistant à la sécheresse (Carbonneau 1985). La thèse a examiné la relation entre la croissance racinaire et la capacité de résistance à la sécheresse en évaluant la vitesse de croissance racinaire, la conductivité hydraulique à travers deux variétés de porte-greffe en conditions de déficit en eau. Le niveau de l’expression des gènes d’aquaporines (via la qPCR et l’ARNseq) et leur contribution à la conductivité hydraulique racinaire ont été analysés dans les radicelles afin d’obtenir une meilleure compréhension sur les mécanismes impliqués dans la régulation de l’absorption de l’eau racinaire et la conductivité hydraulique au cours du développement et en réponse à un manque d’eau.Le traitement de stress d’eau prolongé a diminué le potentiel hydraulique de la plante. La croissance racinaire individuelle est très hétérogène : bien que le traitement de sécheresse réduise l’élongation racinaire en moyenne, la vitesse de croissance racinaire varie tout de même énormément. Un haut niveau de stress hydrique a réduit significativement la vitesse de croissance racinaire moyenne à la fois pour RGM et 110R. Globalement, la vitesse de croissance racinaire moyenne a montré une tendance réduite au cours du développement de la plante. La température du sol est aussi un facteur qui affecte la croissance racinaire. Pour RGM et 110R, en conditionsIIde bon arrosage et de stress hydrique, la vitesse de croissance quotidienne moyenne a été positivement corrélée avec la température du sol quotidienne moyenne. En conditions de bon arrosage, des vitesses de croissance racinaires plus importantes ont été constamment observées chez 110R par rapport à RGM, ce qui pourrait être une explication possible de sa meilleure résistance à la sécheresse par rapport à 110R. [...] / The subject of adaptation to climate change has become one of the most important contemporary topics in grapevine. Much focus has been placed on the understanding of rootstocks effects on scion growth, nutrient uptake, and tolerance to stress, with the ultimate goal of developing novel rootstocks that facilitate adaptation to a changing climate. The purpose of this thesis is to examine how differences in drought resistance between genotypes could result largely from differences in their ability to maintain root growth under stress. A better understanding of how root structure, growth, and water uptake respond to stress will allow us to better understand what aspects of root physiology contribute to drought tolerance. Previous research focused on root water uptake in grapevine suggested that root water uptake could be tightly coupled to a root’s instantaneous rate of growth (see Gambetta et al. 2013). This observation implies that differences in drought resistance between genotypes could result largely from their ability to maintain root growth under stress. Two grapevine rootstocks with contrasting drought resistance capacity, Riparia Gloire de Montpellier (RGM) and 110 Richter (110R), were selected to study in this thesis. RGM is considered as sensitive to drought, while 110R is highly resistant to drought (Carbonneau 1985). The thesis examined the relationship between root growth and drought resistant capacity by assessing root growth rate, hydraulic conductivity across two rootstock varieties subjected to water deficit. The role of aquaporin gene expression (via qPCR and RNAseq) and their contribution to root hydraulic conductivity were analyzed in fine roots in order to obtain a better understanding on the mechanisms involved in the regulation of root water uptake and hydraulic conductivity across development and in response to water deficit.Prolonged water stress treatment decreased plant water potential. Individual root growth is very heterogeneous, although drought treatment reduces root elongation on average, individual root growth rate still varies enormously. High level of water stress significantly reduced average root growth rate for both RGM and 110R. Globally, average root growth rate showed a decreased trend over plant development. Soil temperature is also a factor that affects root growth. For both RGM and 110R, under both well-watered and water-stressed conditions, average daily root growth rate was positively correlated with average daily soil temperature. Under well-watered conditions, higher root growth rates were constantly observed in 110R compared to RGM, which could be one possible explanation for the higher capacity in drought resistance of 110R.Root hydraulic conductivity (Lpr) was influenced by both water stress treatment and plant developmental stage. Generally, for both RGM and 110R, Lpr was significantly reduced under water stress in early stage. In mid and late stages, no significant differences in Lpr were observedIVbetween well-watered and water-stressed plants. Changes in individual root Lpr in response to pre-dawn leaf water potential (ᴪpredawn) were investigated as well. Lpr showed a fast drop in the beginning of water stress treatment when ᴪpredawn was higher than -0.5 MPa. However, with ᴪpredawn getting more negative, e.g. from -0.4 MPa to -2.0 MPa, the range of Lpr values measured in our study maintained constant. Lpr of well-watered plants decreased as well even though their ᴪpredawn was maintained at a high level (< 0.1 MPa) during the period of the experiment. [...]

Pour beaucoup de cultures, un sol non travaillé est une alternative de plus en plus crédible au travail intensif du sol, en particulier pour des raisons environnementales. Mais l’ananas (ananas comosus) présente un enracinement fragile très sensible à la structure, ce qui motive souvent un travail important du sol avant plantation. L’alternative d’une plantation sans travail du sol ne va donc pas de soi. L’objet de cette thèse est de répondre à la question de la faisabilité d’une culture d’ananas sans travail du sol, en remplaçant ce dernier par une plante de service « décompactante » dont les traits racinaires seraient favorables à la (re)structuration d’un sol compact. Le travail de cette thèse a donc porté sur l’évaluation de plusieurs espèces candidates, puis sur l’étude, au champ, des effets sur le sol de la plus prometteuse (Stylosanthes guianensis ). Enfin, dans le cadre d’un essai au champ comparant un système de culture innovant ananas sans travail du sol, on a étudié l’effet de la plante de service sur le fonctionnement de la culture de l’ananas. Dans un premier essai, c’est le Stylosanthes guianensis qui, comparé à huit autres espèces (Arachis pintoï, Brachiaria decumbens, Cajanus cajan, Crotalaria juncea, Cynodon dactylon, Eleusine coracana, Pueraria phaseoloides, et le maïs), a montré les traits racinaires les plus favorables à la structuration d’un sol compact. Les valeurs supérieures du diamètre racinaire moyen et de la densité de longueur racinaire (DRL) caractérisent les principaux traits impliqués. Dans un deuxième essai, les mesures de conductivité hydraulique, d’indice des vides du sol et d’analyse d’images de blocs de sol imprégné sur la répartition surfacique des différents types de porosité ont montré que la culture du Stylosanthes guianensis avait augmenté l’indice des vides du sol et provoqué l’apparition d’une porosité fissurale de grande taille, contribuant ainsi à l’amélioration de la structure du sol. Enfin, un troisième essai mettant en comparaison (i) un système de culture innovant où la culture d’ananas est implantée sans travail du sol après une culture de S. guianensis restructurante et (ii) un système conventionnel comportant un travail profond du sol (et pas de plante de service) a montré que le rendement en fruit est similaire dans les deux systèmes. Cet essai a permis de vérifier que l’enracinement de l’ananas en condition de sol non travaillé bénéficiait du précédent S. guianensis. / For many crops, direct drilling is a well-tried alternative facing the damaging effects of intensive tillage, mainly for environmental causes. But, pineapple (ananas comosus), presents a fragile rooting system which is very sensitive to soil structure. This leads frequently to intensive soil tillage before planting. Direct drilling is not so evident. The aim of this thesis is to give an answer to the feasibility of a no till system for pineapple cultivation, by using a plant with favorable roots traits for compacted soil (re)structuration. This work consisted in evaluating several candidate species, followed by the study, on the field, of the effects the most promising on soil (Stylosanthes guianensis). Finally, through a field experiment, comparing an innovating no till pineapple cultivation system we studied the use effects of Stylosanthes guianensis on the pineapple crop functioning. In a first experiment, Stylosanthes guianensis compared with eight other species (Arachis pintoï, Brachiaria decumbens, Cajanus cajan, Crotalaria juncea, Cynodon dactylon, Eleusine coracana, Pueraria phaseoloides and corn) showed better roots traits for structuring a compacted soil. Measures of average root diameter and root length density are the main implicated roots traits. In a second experiment, the measures of hydraulic conductivity, of the soil void ratio and the analysis of blocks of resin-impregnated soil on the surface distribution of the different type of porosity, all of this showed that Stylosanthes guianensis had increased the soil void ratio and had caused the creation of large-sized cracked porosity, thus contributing to the improvement of the soil structure. Finally, a third experiment involving a comparison between (i) an innovating cultivation system where pineapple is growing in a no till soil after a structuring crop of S. guianensis and (ii) a conventional system with deep tillage (without structuring crop), showed similar fruit yield. This experiment showed evidence that the rooting of pineapple in no till soil benefited from the previous Stylosanthes.

Conselho Nacional de Desenvolvimento Científico e Tecnológico / The social and economical importance of onion in Uruguai is very relevant. The soil mobilization and ridge till are crucial to reduce bulk density and increase the macro/micropores relation for Argiudoll, to induce conditions for a normal growth and development for most crops. The objectives of this study were to measure the effects of soil and plant management on aggregate stability, organic matter and some soil physico-hydric properties as porosity, bulk density, water retention curve, air permeability and water conductivity. Two experiments from National Institute of agriculture investigation (INIA) initiated in 1995 were used. Experiment one used the following treatment: 1) alfalfa; 2) festuca; 3) birdfoot trefoil, alfalfa and white clover mixture; 4) onion after alfalfa; 5) onion after festuca; 6) onion after plant mixture. The experiment two used the following treatments: Summer season - 1) fallow; 2) feijão; 3) corn; 4) setaria. During winter season onion was cultivated at all plots. In both experiments the plots were split to receive 0, 80 and 120 kg de N ha-1. To measure aggregate stability and organic matter preserved three soil samples from each plot were taken from 0-0.05 m layer. To measure soil physico-hydric properties samples were taken using metalic rings from 0-0.01 and 0.1-0.2 m layers. The variation of water stable aggregates were higher than of dry stable aggregates. The stability index was higher for pastures in the experiment I and for fallow in the experiment II, implying lower aggregates impact due tillage or higher organic matter addition by plants on soil surface. However, there is clear indication that tillage was necessary, in order to improve macroporosity, hydraulic conductivity and air permeability. / A importância social e econômica da cebola é muito grande no Uruguai. O revolvimento do solo e a construção de camalhões são essenciais para diminuir a densidade e aumentar a relação macro/microporos em solos com características vérticas para que ocorra o bom desenvolvimento da cultura. O objetivo deste trabalho foi analisar o efeito dos diferentes tratamentos, tendo como parâmetros a alteração da estabilidade de agregados, teor de matéria orgânica do solo (MO) e algumas propriedades físico-hídricas do solo como porosidades, densidade do solo, curva característica de retenção de água no solo, permeabilidade ao ar e condutividade hidráulica. Foram analisados dois experimentos pertencentes ao Instituto Nacional de Investigação Agropecuária (INIA) instalados no ano de 1995. No experimento I foram cultivadas as culturas de alfafa, festuca, consórcio (cornichão, alfafa e trevo-branco), cebola-pós-alfafa, cebola-pós-festuca e cebola-pós-consórcio, resultando em seis tratamentos. No experimento II no verão, as parcelas são ocupadas com pousio estival, feijão, milho e moha e, no inverno, cultiva-se cebola. Em ambos os experimentos as culturas foram submetidas a três distintas doses de nitrogênio (0, 80 e 120 kg de N ha-1). Para análise de agregados e MO, coletaram-se amostras com estrutura preservada na camada de 0-0,05 m, utilizando três subamostras de campo; para determinar as propriedades físico-hídricas do solo coletaram-se amostras utilizando cilindros volumétricos em duas camadas, 0,00-0,10 e 0,10-0,20 m de profundidade. Os resultados demonstraram uma maior variação do diâmetro geométrico úmido (DMGu) em relação ao peneiramento a seco. O índice de estabilidade (IE) foi maior nas parcelas cultivadas com pastagem no experimento I e na parcela que permaneceu em pousio no experimento II, resultando em uma maior preservação dos agregados ocasionado pela ausência de preparo ou adubação orgânica, a qual ocasionou um incremento de MO. Analisando as propriedades físico-hídricas, constatou-se que o preparo convencional é necessário neste tipo de solo, pois aumenta o volume de macroporos ocasionando redução na densidade do solo e incrementos na condutividade hidráulica saturada e permeabilidade ao ar.

KBS-3 and very deep borehole (VDH) concepts are two major types of long-term geologicaldisposal methods for high-level radioactive waste (HLW) isolating from the biosphere. TheKBS-3V concept for isolating the HLW at the depth of 400-500 m, is the officially proposedoption in Sweden and has been the subject of considerable research in the past few decades,while the VDH concept was considered as an option in the 1950s but later became discouragedbecause of insufficient experience in drilling technology. The greatest merit of the VDHconcept is that the almost stagnant groundwater in the deep boreholes prevents the transport ofthe possible release of radionuclides into the rock or up to the ground level. Since variousdisadvantages of the KBS-3V concept were found in previous research, the superiority of VDHconcept attracted the researchers to continue studying it into the late 1980s.The geological repositories of both of KBS-3V and VDH types primarily consist of a naturalbarrier (host rock) and of an engineering barrier (also known as a buffer/backfill barrier).According to the principle of IAEA and national relative research organizations, thebuffer/backfill material should have low permeability and good expandability, as well assuitable physical and sealing properties.The thesis concerns the VDH concept and is focused on the construction and performance ofthose parts of the sealed repository that are not affected by high temperature or gamma radiation.In the lower part of a VDH repository, the clay packages containing HLW will be exposed tohigh temperature (100-150 􀄇 ) in the borehole and to highly saline groundwater. In theinstallation phase of HLW, the groundwater will be pumped out and replaced by medium-softsmectite clay mud in which the HLW packages are installed vertically. During the hydrationand maturation of the clay components, the microstructural reorganization, water transport,migration of clay particles and redistribution of the density of the components take place. Thematuration determines the transient evolution of the clay seals and influences the rheologicaland soil mechanical behavior in the installation phase. The maturation of clay system alsodetermines their ultimate sealing potential of VDH repositories.This study presents the work carried out for investigating the maturation of the buffer-backfillclay in the HLW deep borehole. Initially in the study three types of clays, the Namontmorillonite,magnesium-rich and illite-smectite mixed layer clays, were examined for estimating their performance as the barrier candidate material. This is mainly presented in theliterature review. The experimental study was conducted on montmorillonite GMZ clays andI/S mixed-layer Holmehus clay. The expandability and permeability tests were carried out forinterpretation of the recorded swelling development and assessment of the effect of the salineconditions, with the goal of deriving a relationship between swelling pressure and hydraulicconductivity for different dry densities. The maturation tests of initially fully-saturatedHolmehus clay and partly saturated GMZ clay were performed. During the tests, the shearstrength mobilised by the relative movement of densified mud and migrated dense clay -contained in a perforated central tube - were determined. According to the results of shearstrength tests, the maximum operation time or the number of clay packages to be placed in asingle operation was evaluated, whilst the suitable saturation degree of the dense clay wasdiscussed as well.A model of the maturation of initially water-saturated clay seals based on Darcy’s law wasworked out and the evolution of the clay components in a lab-scale borehole using Holmehusclay were performed and compared with the experimental recordings. Good agreementsbetween the physical behaviors of the theoretical simulations and the measurements wasachieved by which the validity of the model was verified. Using the results, the hydration andsoil migration in the entire maturation process were presented in diagram. The model was alsoused for preliminary evaluation of the maturation products in real boreholes by assuming thesame Holmehus clay as used in the tests. Two constellation of borehole and dense clay withdifferent diameters, 80 cm borehole /60 cm clay and 80cm/50cm, were assumed. The resultsrespecting dry density and hydraulic conductivity of the ultimate maturation products, and thedegree of homogeneous of the buffer and backfill clay system in the assumed boreholes, arepresented and discussed. The options of different mineral types and initial physical propertiesof the candidate buffer clays provide a reference for engineering barrier design of HLW disposalin VDH.

The ensemble Kalman filter (EnKF) is nowadays recognized as an excellent inverse method for hydraulic conductivity characterization using transient piezometric head data. and it is proved that the EnKF is computationally efficient and capable of handling large fields compared to other inverse methods. However, it is needed a large ensemble size (Chen and Zhang, 2006) to get a high quality estimation, which means a lots of computation time. Parallel computing is an efficient alterative method to reduce the commutation time. Besides, although the EnKF is good accounting for the non linearities of the state equation, it fails when dealing with non-Gaussian distribution fields. Recently, many methods are developed trying to adapt the EnKF to non-Gaussian distributions(detailed in the History and present state chapter). Zhou et al. (2011, 2012) have proposed a Normal-Score Ensemble Kalman Filter (NS-EnKF) to character the non-Gaussian distributed conductivity fields, and already showed that transient piezometric head was enough for hydraulic conductivity characterization if a training image for the hydraulic conductivity was available. Then in this work, we will show that, when without such a training image but with enough transient piezometric head information, the performance of the updated ensemble of realizations in the characterization of the non-Gaussian reference field. In the end, we will introduce a new method for parameterizing geostatistical models coupling with the NS-EnKF in the characterization of a Heterogenous non-Gaussian hydraulic conductivity field. So, this doctor thesis is mainly including three parts, and the name of the parts as below. 1, Parallelized Ensemble Kalman Filter for Hydraulic Conductivity Characterization. 2, The Power of Transient Piezometric Head Data in Inverse Modeling: An Application of the Localized Normal-score EnKF with Covariance Inflation in a Heterogenous Bimodal Hydraulic Conductivity Field. 3, Parameterizing geostatistical models coupling with the NS-EnKF for Heterogenous Bimodal Hydraulic Conductivity characterization. / Xu, T. (2014). Ensemble Kalman filtering for hydraulic conductivity characterization: Parallelization and non-Gaussianity [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/43769 / TESIS

Embankment dams are made from soil materials of varying sizes and widely used all over the world.When constructing these, knowing the hydraulic conductivity (K) of the soil materials is a keyparameter in order to construct safe embankment dams. A knowledge gap regarding K measurementsin coarse grained soils has been identified. This thesis aims to provide a theoretical overview ofpresent day state-of-the-art methods for measuring hydraulic conductivity and the controllingcharacteristics for K. Coarse grained soils refers to a soil with the coarsest grain fraction being &gt; 20mm and/or have a K &gt; 10-4m/s.   It was found that the fixed wall permeameter is the most suitable laboratory method. In the field, itis possible to estimate K using tracer methods, these however show more potential for leakagepathway detection. Common for all K measurement methods are the controlling characteristics of K,grain size distribution, pore geometry, degree of compaction, particle movement and flow regime.These need to be considered when testing to produce useful measurements. If the relationshipbetween flow velocity and hydraulic head is non-linear, Darcy's law is not valid for calculating K. / Fyllnadsdammar är uppbyggda av jord och sprängsten av olika storlekar och finns över hela världen.Att känna till den hydrauliska konduktiviteten (K) av de olika lagren är viktigt för att kunna byggadessa på ett säkert och hållbart sätt. Det har identifierats en bristande kunskap angående K mätningar igrovkorniga jord- och stenmaterial. Målet med denna uppsats är att presentera en teoretisk översikt avden senaste kunskapen inom K mätningar i grovkorniga jord- och stenmaterial och vilka egenskapersom avgör ett materials K. Grovkorniga jord- och stenmaterial syftar till material där den grövstakornstorleken är &gt; 20 mm och/eller har ett K &gt; 10-4m/s.   För laboratorie mätningar är en permeameter med en solid vägg den mest lämpliga metoden. Förfältmätningar är det möjligt att mäta K med hjälp av spårämnen, men dessa har mer potential för attupptäcka läckage vägar i fyllnadsdammar. De faktorer som avgör ett materials K ärkornstorleksfördelningen, geometrin av porerna, graden av kompaktering, partikelrörelse ochflödestyp. För att producera mätningar som är användbara behöver dessa faktorer kontrolleras. Omsambandet mellan hydrauliskt huvud och flödeshastighet är icke linjärt kan K inte beräknas genomDarcy´s lag.

Masters of Science / The thesis investigates the potential of the Rietfontein and Spitskop aquifer to meet a demand of 1000 m3/d (12 Lis) as an alternative water resource for the Rietfontein and Spitskop Community. Increasing demand for clean and hygienic drinking water puts more pressure on one of our most valuable resources and supplying all communities with surface water is an extremely difficult and costly task in rural areas like Rietfontein and Spitskop in South Africa. Therefore it is necessary that interim water supplies be found from local aquifers and be utilized to address water supply challenges. Groundwater may serve as a short-term and an interim water supply which may be useful during future dry periods. Abstraction of groundwater is sensitive to recharge. Due to semi-arid conditions in Rietfontein and Spitskop area, there is high rainfall variation and disparity each year. During the research, hydrocensus was carried out. Water samples for chemistry analysis were taken. Literature review and pumping test data was utilized from the previous studies done by different consultants (Botha, 2000, Vivier and Pretorius, 2003). A numerical groundwater flow model for the local aquifers in the area and surroundings was constructed focusing on recharge and abstraction scenarios for the water supply from the local aquifers. For Rietfontein and Spitskop area, the mean annual precipitation (MAP) is 617mm/year. To be assured and rational in determining aquifers ability to meet the required demand a recharge with ninety-five (95th ) percentile was recommended, which estimates the MAP of 308mm/year which is 50% lower than the average MAP of 617mm/year. For a period of ninety six (96) years, the data indicates a severe drought occurred four (4) times where the rainfall was even lower than 95th percentile level of assurance of recharge estimated. This gives a comprehensible indication that average mean is not ideal or steadfast stature when building a water supply numerical groundwater flow modelling. These aquifers can only be exploited and managed if a reliable method can be obtained to estimate their long-term sustainable potential, since the sustainable potential of these aquifers to supply the communities is dependent upon the recharge from rainfall. The results from a numerical groundwater flow model indicated that a combined potential from the local aquifer from eighteen (18) boreholes is sufficient to meet the required demand and a total of 2600m3 Id can potentially be abstracted from the aquifer. With ninety-five (95th ) percentile recharge rate of 308mm/year a numerical groundwater flow model shows that the rate of abstraction is 80% far less than recharge, which gives high level of assurance in terms of local aquifer water supply demand. The abstraction of the boreholes confirmed by a numerical groundwater flow model shows the least impact on the surrounding aquifer for an extended period of time. In the event of drought, the boreholes will see a decline in water level after two (2) months of pumping local aquifer. The water level will decline steadily from two (2) months to two (2) years with a change in water levels of up to 40m. The impact of the drought is minimal compared to recharge rate, which verifies less depletion of the aquifer. The local aquifer shows the potential of 3MLld can be supplied to the communities with an assurance level of 95th percentile of rainfall. Reliable quantification of groundwater recharge rate remains the main challenges the hydrologist experienced and further research is essential for improvement of groundwater management for the area concerned.

An investigation of the variability of engineering properties of municipal solid waste as a function of placement conditions was conducted. Limited data have been reported for the engineering properties of municipal solid waste (MSW) as a function of placement conditions. Wastes have high variability of engineering properties due to heterogeneity in composition and component size; influence from time based effects; and presence of compressible solids. Control of moisture content of MSW at the time of waste placement provides opportunity for increased capacity at a given landfill site due to higher compacted unit weight as well as for control of other geotechnical properties. A laboratory experimental test program was conducted on manufactured municipal solid waste (MMSW) that was representative of waste stream in the United States. Large scale test equipment was used to minimize the effects of scaling on results. The experimental program included compaction, compressibility, hydraulic conductivity, and shear strength testing over moisture contents ranging from 11% to 110%. Baseline compaction curves were developed for different compactive efforts. Similar to soils, the MMSW had bell shaped compaction curves that peaked at a maximum dry unit weight and associated optimum moisture content. The compaction curve generated at modified compactive effort had a maximum dry unit weight of 5.1 kN/m3 and optimum moisture content of 66%. Four times modified compactive effort testing resulted in a maximum dry unit weight of 5.9 kN/m3 and corresponding optimum moisture content of 56%. The compaction curve generated for four times modified compactive effort was used as a baseline for subsequent testing. Compression index was calculated from the strain-log stress curves for total stress conditions and is referred to as apparent compression index. Apparent compression index decreased from 1.1 to 0.34 with increasing moisture content. Secant modulus of elasticity was calculated between 1% and 25% strain and ranged from approximately 200 kPa to 4,800 kPa over the range of tested moisture contents. Tangent modulus ranged from 400 kPa to 6,200 kPa between 1% and 25% strain. Both the secant and tangent modulus peaked between 30% and 56% moisture content. Wet of optimum, the moduli of elasticity decreased with increasing moisture content. The hydraulic conductivity was measured under constant head at a hydraulic gradient of 1 and decreased asymptotically from approximately 1.3x10-2 cm/s to 8x10-5 cm/s as the moisture content was increased to optimum. The hydraulic conductivity of the MMSW increased slightly wet of optimum. The internal angle of friction of the MMSW was measured at 15% shear strain and decreased from approximately 40° to 30° with increasing moisture content. Test results demonstrated that both the molding moisture content and dry unit weight have significant impact on the MMSW geotechnical properties, although it appears that molding moisture content ultimately controls the behavior. Based on the results of the tests it was speculated that, similar to clay soils, increases in moisture content allowed for breakdown of the fabric and physical rearrangement of waste components which in turn controlled geotechnical behavior. Overall trends were comparable for MMSW and soil and included: increased dry density and increased stiffness to optimum moisture content; decreased hydraulic conductivity with increased compaction moisture content; and decreased shear strength with increased compaction moisture content. The results of the test program have environmental and economic implications for design and operation of landfills as well as post closure use.