As an abundant sedimentary rock, shale is widely used as construction material around the world. However, shale is a fissile and laminated material and is therefore subject to deterioration due to environmental and chemical forces (i.e., weathering), which is possible to cause high maintenance cost on associated structures and failures of earth slopes and embankments. However, currently, there is lack of efficient method to monitor the weathering process of shale. This thesis uses several shale samples collected from the commonwealth of Kentucky to study the deformation and shear behaviors of weathered compacted shale. A new electrical approach was developed to access the deformation behavior of shale. The long term deformation behaviors, such as collapse and swell can be predicted from specific electrical parameters. The critical state theory was used to describe the shear behavior of weathered compacted shale. Some findings observed by previous researchers were confirmed, and new empirical equations were provided to estimate the shear strength parameters of weathered compacted shale.

Le fonctionnement et la structure hydrogéologiques des aquifères de socle des granites et des gneiss sont aujourd'hui relativement bien compris. En revanche, dans cet ensemble, les aquifères dans les péridotites sont très peu étudiés et mal compris. Dans ce contexte, les massifs obductés de Nouvelle-Calédonie présentent un laboratoire naturel exceptionnel pour améliorer la connaissance de cet hydrosystème original en contexte tropical. Ainsi l'objectif du présent travail de thèse est d'approfondir la connaissance de la structure et du fonctionnement hydrogéologiques de ces massifs. L'analyse porte d’abord sur le manteau d'altération constitué de la cuirasse, des latérites, des saprolites grossières et du saprock. Plus de 60 essais hydrauliques sont menés sur les massifs de péridotites et les résultats sont compilés aux données existantes. La conductivité hydraulique moyenne des latérites est évaluée à 1.10-7 m/s et celle des saprolites grossières et du saprock à 8.10-7 m/s. L'hétérogénéité de cet horizon altéré est marquée par une gamme de variation de la conductivité hydraulique sur six ordres de grandeur et l'analyse piézométrique met en évidence des connexions hydrauliques avec le substratum fracturé profond. Le substratum est ensuite considéré. L'étude de la fracturation est réalisée à partir de mesures structurales sur affleurement et de la description de près de 1000 m de carottes de forages. L'analyse de la fracturation met en évidence l'importance du réseau serpentineux par sa densité d’une part, et par son lien avec l'altération supergène d'autre part. De plus, il est vérifié que la conductivité hydraulique du substratum diminue avec la profondeur. Cette variation est liée à la diminution de la densité de fractures altérées. Ainsi, à l'issue de ces analyses, la structure des massifs de péridotites est définie. Un réseau primaire de fractures d’espacement décimétrique lié au réseau serpentineux préstructure les péridotites. Sur ce réseau se surimpose un réseau de fractures dont l'espacement est décamétrique et caractérisé par une altération supergène. Les fractures altérées présentent localement de fortes conductivités hydrauliques, de l'ordre de 10-5 m/s. En profondeur l'espacement des fractures est hectométrique et les fractures sont majoritairement fermées, scellées par les minéraux néoformés ou par l'effet de la pression lithostatique. Les réseaux de fractures déca et hectométriques, visibles également sur l'effet d’échelle de la conductivité hydraulique, sont majoritairement verticaux, développés par instabilité de dissolution lors des processus d’altération. Cependant, des structures à faible pendage existent également et permettent la percolation du réseau. Enfin, à partir de ces nouveaux résultats et de l’intégration de l'ensemble des données acquises sur les différents massifs, un modèle de structure et de fonctionnement hydrogéologiques est proposé à l'échelle du massif. Ce modèle comprend l'horizon des latérites qui constitue un aquitard homogène sous lequel se développe l'aquifère dont l'épaisseur est de l'ordre de cinquante mètres. Le substratum est discrétisé en trois couches dont la conductivité hydraulique décroît de 2.10-7 à 2.10-8 m/s entre 50 m et 250 m environ sous le mur de l'aquifère. Les modèles numériques construits permettent de valider le modèle conceptuel unitaire et montrent que l'état de saturation des massifs est contraint par leur géomorphologie. Au terme de ce travail, plusieurs aspects doivent encore être approfondis. Le rôle hydrogéologique de la cuirasse doit être précisé et considéré dans le modèle hydrogéologique. Enfin, compte tenu du développement possible de structures très perméables, voire pseudo-karstiques, la connaissance de la distribution des structures drainantes doit être améliorée. Les résultats appliqués de ce travail de recherche sont exposés dans un rapport final et un guide méthodologique livrés dans le cadre du projet CNRT « HYPERK ». / Water resources of hard-rock (gneisses or granites) aquifers have significantly been studied in the past two decades. The hydrogeological behavior and structure of these aquifers are thus relatively well understood. On the other hand, aquifers in mantle-type basements, such as peridotites, are poorly studied and understood, mainly because they are not common and of limited extent. In this context, New Caledonia is a great laboratory offering unique opportunity to improve the knowledge of these original types of hydrosystems in tropical climate. Thus, the objective of this thesis is to improve the knowledge of these aquifer systems within weathered peridotites. Firstly, the study focuses on the characterization of the weathered layers of the peridotites composed of, from top to bottom, iron oxides/ferricrete, laterite, coarse saprolite and saprock (ie. top of the bedrock, with up to 20 % of weathered material). More than 60 hydraulic tests are performed and results were compiled with existing data. Mean hydraulic conductivity (K) of laterites is estimated around 1.10-7 m/s while mean value in coarse saprolites and saprock is around 8.10-7 m/s. Heterogeneity of this altered layer is high; K varies between six orders of magnitude and hydraulic head data analysis reveals a hydraulic connection with the deep fractured bedrock. Secondly, the fresh rock part of peridotites is studied. Fracture network analysis is derived from outcrop structural measurements and from the description of about 1000 m of cumulated borehole cores. This work highlights the importance of serpentine network, because of its high density and its critical impact on weathering. Moreover, the observations reveal that hydraulic conductivity decreases with depth within the substratum, due to the vertical decrease of weathered fractures density. These new results allow defining a structural framework of the massifs. It is characterized by a primary decimetrical fracture network closely related to the serpentine network. This network is overprinted by a secondary weathering network which reveals decametric spacing (ca. 30 m) and in places K values of 10-5 m/s. At depth, spacing is hectometrical and fractures are sealed by lithostatic pressure and/or subsequent mineral precipitations. These deca- and hectometric fracture networks, which are also visible on the scale effect of hydraulic conductivity, are primarily vertical and are the result of dissolution instabilities occurring during weathering processes. However, low- angle fractures do occur and allow the percolation of the network. Finally, on the basis of these new results and the integration of all existing data from different massifs a new hydrogeological conceptual model is proposed at the scale of a massif. The model includes a homogeneous lateritic aquitard and a coarse saprolite and saprock aquifer which is about 50 m thick. The bedrock is subdivided into three layers whose hydraulic conductivity decreases from 2.10-7 m/s to 2.10-8 m/s, respectively 50 and 250 m below the aquifer base. Numerical modelling validates this unitary conceptual model and reveals that the saturation of the massifs depends on their morphology. At last, several aspects require further research. The role of the ferricrete layer must be specified and considered in the hydrological model. Moreover, distribution of the fracture network remains to be fully addressed and should be studied with care given the potential development of highly permeable structures that could conform to pseudo-karstic drains. The applied results of this work are available in a “Technical guide” and a “Technical report” of the “HYPERK” CNRT Project.

A interação de elementos potencialmente tóxicos com solos e a caracterização quanto à mobilidade e potencial de liberação ao ambiente são importantes para avaliação de risco ambiental. No caso de solos altamente intemperizados, como os Latossolos, a interação pode ser diferenciada devido à elevada acidez, baixa densidade de cargas negativas, maiores quantidades de minerais de argila 1:1 e de (hidr)óxidos de Fe, Al e Mn, em que predominam cargas negativas variáveis com o pH. O objetivo nesta tese foi caracterizar a capacidade de sorção, a distribuição entre as frações do solo, a especiação temporal, a predição da capacidade de sorção e a dessorção com método dinâmico de extração de Cd(II) em seis solos tropicais, incluindo três Latossolos. Amostras dos seis solos foram utilizadas para avaliação da interação com Cd(II), utilizando extração sequencial e especiação por espectroscopia de absorção de raios-X próximo à estrutura da borda (XANES) para avaliação da cinética de sorção. Dados de 29 solos foram utilizados para desenvolvimento de modelo de regressão linear para predição das capacidades de sorção de Cd(II). A caracterização da dessorção de Cd por extração convencional e dinâmica foi feita em solos de textura médio-arenosa, argilosa e em um solo de referência certificado. Os resultados de sorção foram ajustados ao modelo de Langmuir. As capacidades de sorção e as energias livres padrão de Gibbs variaram de 37 à 1296 mg kg-1 e de -16,6 até -27,0 kJ mol-1, respectivamente. A sorção foi fraca e reversível, e mais de 90% do Cd estava sorvido como espécies disponíveis. A especiação temporal indicou evidências fracas e variáveis para as alterações químicas do Cd no solo, sugerindo que o elemento liberado nessas amostras se ligou à matéria orgânica do solo e aos óxidos minerais ou permaneceu dissolvido, com pequenas alterações na especiação nos meses seguintes. O modelo linear representou 98% dos resultados empíricos apenas em função de uma variável, que foi o ensaio de sorção simplificado com apenas uma concentração de Cd(II). Apesar de a capacidade de sorção prevista com o modelo ter apresentado variação de cerca de 20% em relação ao empírico, o modelo de predição apresenta potencialidade de aplicação para avaliações iniciais e rápidas. A dessorção avaliada em batelada e em fluxo indicou dessorção próxima de 100% nas frações que representam disponibilidade no ambiente, enquanto o sistema desenvolvido para extração dinâmica foi adequado na mistura das soluções com a amostra de solo e propiciou rápida troca de extratores. De modo geral os atributos do solo, bem como a classe, influenciaram, mas não definiram a interação do Cd(II) com o solo. Por sua vez, o Cd(II) incorporado ao solo apresenta grande risco ambiental e de interação com a biota / The interaction of potentially toxic elements in soils and characterization as mobility and potential of environmental release are important for environmental risk assessment. In the case of highly weathered soils, as Oxisols, interaction can be differentiated, due to the high acidity, low density of negative charges, and higher amounts of 1:1 clay minerals and Fe, Al and Mn (hidr)oxides, in which negative charges are variables with pH. The objectives in this thesis was to characterize the Cd(II) sorption capacity, distribution between soil fractions, temporal speciation, prediction of sorption capacity and desorption with dynamic extraction method in six tropical soils, including three Oxisols. Samples of the six soils were used to assess the interaction with Cd(II) using sequential extraction and the speciation with X-ray absorption near edge structure spectroscopy (XANES) to evaluate the adsorption kinetics. Data of 29 soils were used to develop a linear regression model for prediction of Cd(II) sorption capacity. The Cd desorption characterization trough conventional and dynamics extractions was performed in a medium-sandy, a clayey and a certified reference soils. The sorption results were fitted to the Langmuir model. Sorption capacities and standard Gibbs free energy ranged from 37 to 1296 mg kg-1 and from -16.6 to -27.0 kJ mol-1, respectively. Sorption was weak and reversible, and more 90% Cd was sorbed as available species. The temporal speciation indicated weak and variables evidence of chemical changes of Cd in the soil, suggesting that this element released in these soils bound to the soil organic matter and to mineral oxides or remained dissolved, with minor changes in speciation in the following months. The linear model accounted for 98% of empirical results only on the basis of a single variable, which was the sorption simplified experiment with only one Cd(II) concentration. Although the sorption capacity predicted with the model ranged about 20% of the empirical, prediction model has potential of application for initial and rapid assessments. The desorption evaluated in batch and flow indicated desorption of about 100% in the fractions representing availability in the environment, meanwhile the system developed for dynamic extraction was adequate to mix the solutions with the soil sample and provided rapid exchange of extractants. Overall, the soil characteristics and the class had influence, but did not define the interaction of Cd(II) with the soil. On the other hand, the Cd(II) incorporated into the soil has a great environmental risk and to interact with the biota

Les facteurs qui régissent l'intensité et la répartition de la recharge naturelle et artificielle dans les aquifères cristallins altérés et fracturés sont mal connus. Ce sont cependant les caractéristiques déterminantes de ce type de roche—ces roches sont très hétérogènes—qui rendent difficile l’estimation des flux dans ces milieux ainsi que des propriétés hydrauliques qui les contrôlent. La première partie de ce manuscrit fournit un état des connaissances sur la recharge des eaux souterraines et ses méthodes d’estimation, permettant de comprendre les défis scientifiques et sociétaux abordés dans cette thèse. La deuxième partie présente les travaux numériques et expérimentaux menés pour approfondir notre compréhension de la dynamique des flux d’eaux souterraines dans ces milieux hétérogènes à plusieurs échelles. Le premier axe de recherche porte sur les processus de recharge naturelle à l’échelle du bassin versant. La recharge diffuse a été modélisée avec un modèle physique simple d’infiltration et comparée à des estimations préalables de recharge totale. Nos résultats illustrent la forte dépendance de la recharge aux précipitations et à l’irrigation, et l’importance de la recharge focalisée. Les facteurs responsables de la distribution spatiale de la recharge sont aussi étudiés. Le deuxième axe est basé sur le suivi de la mise en eau d’un bassin de recharge artificielle dans un site hautement monitoré et bien équipé. Ces observations ont été interprétées avec des modèles analytiques et numériques. Ces modèles ont mis en évidence l’existence de flux préférentiels horizontaux, mais aussi d’une compartimentation latérale qui entrave la propagation des intrants de recharge. / The factors governing the intensity and distribution of natural and artificial recharge in weathered and fractured crystalline aquifers are poorly understood. However, it is the defining characteristics of this type of rock—these rocks are very heterogeneous—that make the estimation of fluxes and the hydraulic properties controlling them difficult. The first of its two parts provides the theoretical framework on groundwater recharge processes and its estimation methods for comprehending the scientific and societal challenges discussed in this thesis. The second part presents the numerical and experimental work carried out to deepen our understanding of the dynamics of groundwater flows in these heterogeneous underground environments at several scales. The first line of research focuses on natural recharge processes at the watershed scale. Diffuse recharge was modeled with a simple physical infiltration model and compared to previous estimates of total recharge. Our results highlight the strong dependence of recharge on rainfall and irrigation, and the importance of focused recharge. The factors responsible for the spatial distribution of recharge are also studied. The second axis is based on the monitoring of the filling of an artificial recharge basin at a highly monitored and well-equipped site. These observations were interpreted with analytical and numerical models to improve our knowledge of flow dynamics in fractured crystalline rocks at the medium scale. These models illustrated the existence of preferential horizontal flows, but also of a lateral compartmentalization that hinders the propagation of recharge inputs.

Phosphorus is the main limiting nutrient for plant growth in large areas of the world and the availability of phosphorus to plants and microbes can be strongly affected by soil properties. Even though the phosphorus cycle has been studied extensively, much remains unknown about the key processes governing phosphorus availability in different environments. In this thesis the complex dynamics of soil phosphorus and its availability were studied by relating various phosphorus fractions and soil characteristics to microbial respiration kinetics. The soils used represent a range of aluminium, iron, carbon and total phosphorus content, and were located in four different biomes: subtropical forest, warm temperate forest, boreal forest and tundra. The results showed that NaOH extractable phosphorus, a fraction previously considered to be available to plants only over long time scales, can be accessed by microbes in days or weeks. Microbial phosphorus availability was not related to aluminium or iron content in any of the studied systems, not even in highly weathered soils with high aluminium and iron content. This is in contrast with other studies of soils with high sorption capacity and shows the variability of factors that govern phosphorus availability in different environments. In the boreal forest chronosequence, no difference could be seen with age in total phosphorus content or concentrations of occluded phosphorus forms. However, there were lower concentrations of labile phosphorus forms in older systems, which were correlated with a decrease in microbial respiration. This was most likely related to organic matter quality in the system, and not to geochemical factors. Phosphorus availability was linked to differences in topography (water regime) and vegetation in the tundra ecosystems. The results suggest that the availability of phosphorus, both for microbes and plants, was lower on the meadow vegetation sites compared to the two types of heath vegetation. Many factors are important for phosphorus availability in soils, but these results suggest that microbes can access less available phosphorus if not restricted by carbon, and this may be important in regard to forest management practices as well as effects of environmental change.

Philosophiae Doctor - PhD / Coal combustion accounts for 95% of electricity generation in South Africa while globally coal combustion for energy generation stands at 42%. It has been predicted that coal utilization for energy generation will continue due to its low cost and availability in huge quantities in different parts of the world. Additionally brine and gaseous emissions are produced in the power generation and coal combustion processes. In fact, it has been established that CO2 emissions from power plants are the main cause of the green-house effect leading to global warming. Mitigation of the effects of disposal of fly ash, brine and CO2 emissions is critical for sustainable energy generation from coal and environmental protection. The study investigated whether South African coal fly ash could be used for brine remediation and CO2 capture using fly ash based hydrotalcites and zeolites. Four main objectives were investigated. These were; firstly, to compare the natural CO2 capture potential of a power station ash dam with an accelerated ex-situ mineral carbonation process. Secondly, to probe the effect of accelerated ex-situ mineral carbonation on brine quality with regards to major, minor and trace elements concentration. Furthermore, the study investigated the feasibility of synthesizing hydrotalcites from fly ash by optimizing the synthesis parameters such as acid concentration, aging time, aging temperature, pH during aging, crystallization time and crystallization temperature. Finally the study compared the CO2 adsorption capacities of the fly ash based hydrotalcites with fly ash based zeolites NaA, and NaX. The natural carbonation potential of the wet disposed ash dam at Secunda was investigated by coring a 20 year old dam. Three cores (SI, S2 and S3) were obtained by air flush coring the dam along a geophysical line and establishing the geophysical profile of the three cores. The surface of the three cores was of medium resistivity with values between 9.3 and 12.2 nm while the midsections were of low resistivity with values ranging between 4 and 7 nm. The base section of core SI had a resistivity of 28.3 nm, that of S2 was between 16.2 and 21.4 nm and that of S3 between 12.2 and 16.2 nm; implying that SI had the lowest salt load while S3 had the highest salt content. Moisture content was observed to be high deeper down the profiles of S2 and S3 with samples appearing water logged while SI had the highest moisture content at the surface showing the inhomogeneity of the ash dam. The morphology of fresh fly ash taken from the ash collection hoppers at Secunda was observed to be spherical. Weathered ash from the ash dam showed irregularly agglomerated particles while accelerated ex-situ mineral carbonation resulted in the formation of acicular particles of calcite. Fresh ash, weathered ash and the accelerated carbonated ash were all class F with a sum total of silica, alumina and iron oxide totaling more than 70%. A reduction in silica and alumina content with instability of fly ash. Dumping of spent iron catalyst (resulting from the petrochemical operations as Sasol) on the ash dam led to an increase in Fe203 content of the weathered ash. Enrichment of Nb, Sr, Y, Th, Na, Cl, S04, K and S with natural carbonation as well as during accelerated ex-situ mineral carbonation was observed and was due to the contact of ash with brine during these two processes. Reduction of Zr, Rb, Pb, Ni, Co and V content of ash was observed with weathering. Mineralogically, all the ash samples had main phases of mullite, quartz, magnetite and hematite, with weathered and accelerated carbonated ash having additional phases of calcite. The aluminosilicious nature of the three ashes was identified by structural evaluation using Fourier transform infrared analysis which revealed that, bands associated with C-O in-plane and out of plane bending of carbonates was only visible in weathered and carbonated ash.

The use of coarse-grained fill materials has grown significantly in recent years especially on account of their use in dams and transportation networks. This study investigates compressibility of various coarse-grained fill materials in dry and wet loading conditions in oedometer test. Four materials were used in the experiments, which falls into GP, GW, GM and GC categories respectively. GP material is a weathered rock obtained from Eymir Lake region. This material was chosen especially to be able to investigate degradation and particle breakage due to compaction and compression. GW, GM and GC materials were obtained by using the material called &lsquo / bypass&rsquo / which is a fill material used in the construction of metro of Eryaman. Using these four materials, large-scale double oedometer tests were carried out to investigate compressibility in both dry and wet conditions. The double oedometer testing technique is used to investigate the effect of soaking on compressibility behaviour of compacted fill materials. Various compactive efforts were used in the compaction stage to investigate the effect of compactive effort on compressibility and degradation of the four gravelly materials. Gradations of the post-test samples were obtained and particle breakage due to compaction using various compactive efforts and particle breakage due to compression were determined. It is found that amount of compression does not necessarily depend on the dry density of the material and fine fraction is also a dominating property regarding the compressibility in coarse-grained fill materials. The vertical strains induced by soaking are on the order of 12% - 20% of the compression measured in dry loading case for the well-graded coarse-grained fill materials tested. Besides, there is significant particle breakage in the compaction process and no further particle breakage in the oedometer test for GP material.

A landslide occurred during the construction of the Giresun &ndash / Espiye road between Km: 1+030 &ndash / 1+170 in April 2003. Investigating the causes and mechanism of this slope failure along with suggesting a proper stabilization technique is aimed in this study. For that purpose, a detailed site investigation study, including engineering geological mapping, drilling work, in situ and laboratory tests, was performed. Weathered tuffite, tuffite, flysch and dacitic tuffite, from top to bottom, are the major units in the study area. A &ldquo / translational slide&rdquo / occurred in completely weathered tuffite owing to the disturbance of the stability of the slope by the excavations performed at the toe of the slope / particularly the foundation excavation for the restaurant building and for the road cut for the Giresun &ndash / Espiye road. After establishing the model of the landslide in detail, shear strength parameters of the failure surface were determined by the back analysis method as &quot / cohesion&quot / =2.5 kN/m2 and &quot / friction angle&quot / =9&deg / . Toe buttressing, ground water and surface water drainage options were considered for stabilizing the slope. For the back analysis calculations, the Morgenstern-Price and Spencer methods were used with the aid of the SLOPE/W computer program.

Les aquifères fracturés sont un enjeu majeur de l’hydrogéologie actuelle. Ils constituent une ressource essentielle pour de nombreuses populations. Le massif de l’Ursuya (France, 64), en est une illustration. Constitué de formations métamorphiques fracturées, il est intensément exploité pour l’alimentation en eau potable. Une approche multidisciplinaire a permis la compréhension du fonctionnement de ce système. Les résultats obtenus constituent une avancée dans la connaissance des aquifères discontinus et offrent des pistes pour une gestion raisonnée de cette ressource.Le signal d’entrée est caractérisé dans sa composante quantitative et qualitative. Le suivi des paramètres climatiques permet d’estimer la lame d’eau participant à la recharge de l’aquifère. La caractérisation isotopique de l’eau précipitée met ensuite en évidence une origine majoritairement atlantique des masses d’air, responsables des précipitations sur le nord-ouest du Pays Basque. Des circulations atmosphériques sur la péninsule Ibérique, l’Europe du Nord et la région méditerranéenne sont cependant responsables de certains événements pluvieux. La composition chimique de l’eau de pluie, résultant de ces origines, est caractérisée par une faible acidité et par des concentrations en éléments anthropiques parfois élevées.L’hydrochimie des eaux souterraines est étudiée conjointement avec des données de temps de séjour acquises par l’interprétation des concentrations en 3H, CFC et SF6. Les caractéristiques physico-chimiques, les temps de séjours mesurés (moins de 10 ans à plus de 50 ans) et les phénomènes de mélanges associés permettent de proposer un modèle conceptuel des écoulements souterrains. Celui-ci met en exergue le rôle prépondérant du profil d’altération développé dans les milieux cristallins (de la surface vers la profondeur : altérites, roche fissurée et roche saine), du point de vue de la minéralisation de l’eau comme de celui des modalités d’écoulement.Une approche quantitative est finalement proposée. Du point de vue hydrodynamique, l’hétérogénéité spatiale est importante (10–4 m s–1 < K < 10–8 m s–1). Les altérites offrent une capacité de stockage et de régulation de la recharge. Un modèle numérique synthétise et valide ces résultats. Les simulations montrent de fortes interactions entre les réseaux d’écoulements superficiel et souterrain et de faibles impacts de l’exploitation actuelle sur les flux d’eau souterraine. Les évolutions climatiques ne modifieront pas significativement les écoulements durant les prochaines décennies. Une diminution des débits des sources et des cours d’eau est toutefois probable. Ce modèle numérique et l’ensemble des résultats obtenus serviront de base pour une gestion raisonnée de la ressource en eau de l’aquifère de l’Ursuya. / Hard-rock aquifers are one of the major challenges for the current hydrogeology. These fractured formations are intensively exploited for drinking water supply becoming an essential resource for many people. The metamorphic massif of Ursuya (France, 64) is one of this strategic aquifer. A multidisciplinary approach has led to understand the functioning of this system. The results help to progress in the knowledge of discontinuous media and they provide supports for a rational management of this resource.The input signal is characterized by both quantitative and qualitative methods. Monitoring of climate parameters is used to estimate the amount of aquifer recharge. Isotopic characterization of rainwater highlights the principal Atlantic origin of air masses which are responsible of precipitation on the north-western Basque Country. Some rainfall events are also due to atmospheric circulation over the Iberian Peninsula, the Northern Europe and the Mediterranean region. The chemical composition of rainwater, resulting from these origins, is characterized by a low acidity and sometimes by high concentrations of anthropogenic elements.The hydrochemistry of groundwater is studied in conjunction with residence time data. These are acquired by the interpretation of 3H, CFCs and SF6 concentrations. The chemical characteristics, the measured residence times (less than 10 years to over than 50 years) and the associated mixing processes allow the achievement of a conceptual model of groundwater flow. It points out the role of the weathering profile (from the surface to the depth : weathered layer, fractured rock and fresh bedrock) from the point of view of the mineralization as well as the flowpaths.A quantitative approach is finally proposed. Boreholes studies show highly heterogeneous hydrodynamic properties (10-4 m s-1 < K <10-8 m s-1). Weathered materials provide a storage capacity and a smoothing of the recharge variations, these results are synthesized and validated by a numerical model. The simulations show strong interactions between the surface flow network and the water table. It also highlights the low impact of the current groundwater exploitation on the aquifer and that climate change will not significantly modify the groundwater flows in the coming decades. Nevertheless, a decrease of the springs and streams flow is probable. This numerical model and all these results constitute the basis for a rational management of water resources from the Ursuya aquifer.

O fósforo (P) é um elemento essencial para o desenvolvimento das plantas e precisa ser suprido em quantidades adequadas para obtenção de altas produtividades das culturas agrícolas. O conteúdo de P no solo e consequente perda deste para as águas superficiais são importantes fatores relacionados a eutrofização de ambientes aquáticos. Nesse trabalho foi obtido o Grau de saturação de P (GSP) em 29 solos tropicais altamente intemperizados e seus valores foram correlacionados com atributos químicos, físicos e mineralógicos dos solos. Antes de calcular o GSP, foram determinados parâmetros como Índice de sorção de P (ISP), Capacidade de sorção de P (CSP) e fator de escala ?. Os solos foram divididos em grupos, e os solos de cada grupo receberam doses diferenciadas de P para obtenção do ISP devido a variação na capacidade de adsorção,. As doses utilizadas foram 200, 500, 1000, 1500 e 3000 mg L-1 de P. Os períodos de reação avaliados foram 1, 3, 7, 21, 42, e 84 d. A porcentagem média de P adsorvida no final do período de contato (84 d) variou de 23% a 49% do P adicionado inicialmente. Os Latossolos, os Gleissolos e os Nitossolos apresentaram os maiores valores de ISP, enquanto os Argissolos e os Neossolos tiveram os menores valores. Os maiores valores foram encontrados nos solos com elevados teores de argila, matéria orgânica (MO) e óxidos cristalinos e mal cristalizados de Fe e de Al. O ISP foi maior no final do período de contato (84 d), o que realça a influência do tempo na adsorção do P. O fator de escala ? e o GSP foram calculados em todos os períodos de contato. Os valores de ? aumentaram e os de GSP diminuíram em função do tempo devido à maior interação do P com o solo. Os menores valores de GSP foram obtidos nos solos com alta capacidade de adsorção de P, ao passo que os maiores valores de GSP foram dos solos com menor capacidade de adsorção. O maior GSP (31%) foi obtido no Neossolo Litólico. O valor limite de GSP considerado nesse trabalho foi 23%. As relações entre GSP, CSP, ISP e alguns atributos dos solos foram abordadas por meio da análise de componentes principais. Com o agrupamento de todos os solos, os teores de óxido mal cristalizado de alumínio (Alox), argila, MO e de óxido cristalino de alumínio (AlDCB) foram os que melhor se correlacionaram com o ISP e a CSP. Avaliando somente os Latossolos, os teores de argila e de óxidos cristalinos e mal cristalizados de ferro e de alumínio (Feox, FeDCB, AlDCB e Alox) representaram os componentes principais ligados à adsorção de P. Com a adição dos demais solos, os teores de fósforo extraídos por solução ácida de oxalato de amônio (Pox), Alox, argila e MO formaram os componentes principais para ISP e CSP. O GSP não correlacionou significativamente com nenhum dos atributos. A partir do dendograma, os solos foram separados em dois grupos. O grupo I foi composto por solos pouco suscetíveis a perdas de P em função dos valores de CSP e de GSP. Já o grupo II reuniu os solos mais suscetíveis. Equações de regressão linear múltipla foram utilizadas na predição do ISP e CSP e duas delas obtiveram um alto coeficiente de determinação (R2 = 0,94 para ambas), podendo ser utilizadas para estimar a CSP e o ISP nos solos. / Phosphorus (P) is an essential element for plant growth and must be supplied in adequate quantities to the growers obtain high yields of agricultural crops. The content of P in the soil and its loss to surface water are important factors related to eutrophication of aquatic environments. In this study, we obtained the degree of P saturation (DPS) in 29 highly weathered tropical soils, and their values were correlated with chemical, physical and mineralogical soil attributes. Before calculating the DPS, parameters such as P sorption index (PSI), P sorption capacity (PSC) and the scaling factor ? were obtained. Because of the variation in the soils` ability to adsorb P, they were divided into groups, and the soils from each group received different P rates for obtaining the PSI. The rates were 200, 500, 1000, 1500 and 3000 mg L-1 of P. The periods of reaction evaluated were 1, 3, 7, 21, 42, and 84 d. The average percentage of P adsorbed at the end of the contact period (84 d) ranged from 23% to 49% of P added initially. The Oxisols, the Gleysols and the Alfisols showed the highest values of PSI, while the Ultisols and Entisols had the lowest values. The highest values were found in soils with high clay contents of organic matter (OM) and crystalline and poorly crystalline Fe and Al oxides. ISP was higher at the end of the contact period (84 d), which highlights the influence of weathering on P adsorption by soils. The scale factor ? and the DPS were calculated for all periods of contact. The values of ? increased while DPS decreased as a function of time because of the greater interaction of P with the soil. The lowest values of GSP were obtained in soils with high P adsorption capacity, whereas the highest values of DPS were obtained in soils with lower adsorption capacity. The largest DPS (31%) was obtained in the Typic Udorthent. The limit value of DPS considered in this study was 23%. Correlations between DPS, PSC, PSI and some soil attributes were addressed by principal component analysis. By grouping all soils, the levels of poorly crystalline aluminum (Alox), clay, OM and crystalline aluminum oxide (AlDCB) were the atributes best correlated with both PSI and PSC. When evaluating only the Oxisols, it was observed that the contents of clay and crystalline and poorly crystalline iron and aluminum oxides (Feox, FeDCB, AlDCB and Alox) represented the main components related to the adsorption of P. When the other soils were added, the contents of P extracted by an acid ammonium oxalate solution (Pox), Alox, MO and clay composed the main components for PSC and PSI. DPS did not correlate significantly with any of the attributes. From the dendrogram, the soils were separated into two groups: group I, consisted of soils with little susceptiblity to P loss based on the PSC and DPS, and group II, that met the more susceptible soils for P losses. Multiple linear regression equations were used to predict PSI and PSC, and two of them had a high determination coefficient (R2 = 0.94 for both). So, they can be used to estimate the PSC and PSI at the soils.