A avaliação de recursos minerais necessita da delimitação prévia de domínios estacionários para cálculo dos volumes do depósito mineral. O conhecimento a respeito dos processos geológicos de formação do depósito são traduzidos em modelos geológicos, essenciais para o planejamento da produção e à tomada de decisões a respeito da tecnologia e os recursos a empregados no empreendimento. A indústria mineral, normalmente, considera a incerteza dos teores na avaliação de recursos, entretanto a incerteza relacionada ao modelo geológico geralmente não é avaliada. Essa incerteza está relacionada a localização do limite entre domínios geológicos, e é uma da principal fonte de incerteza do empreendimento mineral. A incerteza do modelo geológico deve ser avaliada devido o potencial impacto no volume de recursos do depósito, e consequentemente na lucratividade do empreendimento. Esta dissertação busca avaliar a incerteza associada ao modelo geológico, através de três metodologias que possibilitam construir múltiplas realizações para domínios geológicos. Um banco de dados real com alta complexidade é utilizado na construção dos estudos de caso de cada metodologias e os resultados comparados com o modelo determinístico utilizado como referência. Dentre os métodos, a metodologia do parâmetro da incerteza C representou melhor a incerteza volumétrica, e portanto, foi selecionado para gerar o modelo geológico final, e demonstrar a importância da avaliação da incerteza na construção de modelos geológicos. / The evaluation of mineral resources requires the prior delimitation of a stationary domain geologically controlled. The knowledge about the ore genesis and geological processes involved are translated into geological models, essential for planning the production and decision-making regarding the technology and resources applied in the enterprise. The mineral industry usually considers the grade uncertainty in the evaluation of resources, however the uncertainty related to the geological model is generally under evaluated. This uncertainty related to the location of the boundary between geological domains is the great source of uncertainty in the mineral enterprise. The geological model uncertainty should be assessed due to the potential impact on the volume of deposit, and consequently the profitability of the enterprise. This dissertation evaluates the uncertainty associated to the geological model, through three methodologies that generate multiple realizations for geological domains. A real database with high geological complexity is used in the construction of the case studies for each methodology and the results compared to the deterministic model used as a benchmark. Among the methods, the uncertainty calculation methodology C was selected to generate the final geological model, and to demonstrate the importance of uncertainty evaluation in the construction of geological models.

A área de estudo está localizada na parte sudeste do estado do Paraná, município de Carambeí, PR. Nessa região, há uma ocorrência de ouro conhecida desde a metade do século 18, que foi explotada por garimpeiros. Esse estudo tem por objetivo a modelagem geológica e avaliação de recursos minerais por técnicas geoestatísticas. A mineralização de ouro é encontrada nas rochas hidrotermalizadas do Grupo Castro. A pesquisa mineral foi desenvolvdia pela Terra Pura, uma empresa brasileira. Todos os dados da área foram liberados para este estudo. Nesse estudo, foram considerados os dados de sondagens por circulação reversa e diamantada. A primeira questão que surge do uso de dados de diferentes métodos de sondagem se refere ao suporte amostral e consequentemente sua validação. Geralmente, a validação de tais dados é feita usando a aproximação tradicional de furos gêmeos, onde uma é a circulação reversa e outra a diamantada. Isso é muito difícil porque há variação geológica mesmo para pequenas distancias. Além disso, quando esses dados são comparados em um diagrama de dispersão, o coeficiente de correlação é geralmente baixo. Assim, nesse trabalho foi usada uma nova aproximação estatística para comparar as distribuições de frequências das amostras coletadas nas sondagens por circulação reversa e a diamante. O resultado foi muito bom mostrando uma aderência de mais que 90%, que implica em uma possível mudança dos métodos de exploração, pois o custo da circulação reversa é muito menor que da sondagem diamantada. Essa conclusão permitiu a utilização de ambos os conjuntos de dados para desenvolver as etapas posteriores desta pesquisa objetivando a modelagem geológica e avaliação de recursos minerais usando métodos geoestatísticos. Nesta área, as rochas hidrotermalizadas ocorrem conforme um trend local com direção N45E e mergulhando 30 para SE. Essa informação estrutural foi usada tanto para a modelagem geológica e geoestatística dos recursos minerais resultando em modelos compatíveis e úteis para o entendimento da distribuição de ouro. / The study area is located in southeastern of Paraná state, municipality of Carambeí. In this region there is an occurrence of gold known since middle of 18th century, which was exploted by prospectors. This study concerns geological modeling and ore resource estimation by geostatistical techniques. The gold mineralization is found within hydrothermalized rocks of the Castro Group. The mineral exploration was developed by Terra Pura, a Brazilian company. All data regarding this area was released for this research. For this study we considered data from reverse circulation and diamond drill holes. The first question that arises from using data coming from different drilling methods is the sampling support and consequently its validation. Usually, validation of such data is made using the traditional approach of comparing twin drill holes, in which one is reverse circulation and the other is diamond drill. This is very difficult because we have geological variation even for short distances. Moreover, when both data are compared in a scattergram, the correlation coefficient is usually low. Thus, in this work we used a new statistical approach to compare frequency distributions of samples collected from reverse circulation and diamond drill holes. The result was very good showing a match of more than 90% that implies in a possible change of exploration methods, because reverse circulation is much cheaper than diamond drill holes. This result allowed us to use both data sets to develop further steps of this research aiming the geological modeling and ore resource estimation using geostatistical methods. In this area hydrothermal rocks occur according to a local trend striking N45E and dipping 30 SE. The structural information was used for both geological and geostatistical modeling of mineral resources resulting and compatible and useful models for understanding the distribution of gold.

A área de estudo está localizada na parte sudeste do estado do Paraná, município de Carambeí, PR. Nessa região, há uma ocorrência de ouro conhecida desde a metade do século 18, que foi explotada por garimpeiros. Esse estudo tem por objetivo a modelagem geológica e avaliação de recursos minerais por técnicas geoestatísticas. A mineralização de ouro é encontrada nas rochas hidrotermalizadas do Grupo Castro. A pesquisa mineral foi desenvolvdia pela Terra Pura, uma empresa brasileira. Todos os dados da área foram liberados para este estudo. Nesse estudo, foram considerados os dados de sondagens por circulação reversa e diamantada. A primeira questão que surge do uso de dados de diferentes métodos de sondagem se refere ao suporte amostral e consequentemente sua validação. Geralmente, a validação de tais dados é feita usando a aproximação tradicional de furos gêmeos, onde uma é a circulação reversa e outra a diamantada. Isso é muito difícil porque há variação geológica mesmo para pequenas distancias. Além disso, quando esses dados são comparados em um diagrama de dispersão, o coeficiente de correlação é geralmente baixo. Assim, nesse trabalho foi usada uma nova aproximação estatística para comparar as distribuições de frequências das amostras coletadas nas sondagens por circulação reversa e a diamante. O resultado foi muito bom mostrando uma aderência de mais que 90%, que implica em uma possível mudança dos métodos de exploração, pois o custo da circulação reversa é muito menor que da sondagem diamantada. Essa conclusão permitiu a utilização de ambos os conjuntos de dados para desenvolver as etapas posteriores desta pesquisa objetivando a modelagem geológica e avaliação de recursos minerais usando métodos geoestatísticos. Nesta área, as rochas hidrotermalizadas ocorrem conforme um trend local com direção N45E e mergulhando 30 para SE. Essa informação estrutural foi usada tanto para a modelagem geológica e geoestatística dos recursos minerais resultando em modelos compatíveis e úteis para o entendimento da distribuição de ouro. / The study area is located in southeastern of Paraná state, municipality of Carambeí. In this region there is an occurrence of gold known since middle of 18th century, which was exploted by prospectors. This study concerns geological modeling and ore resource estimation by geostatistical techniques. The gold mineralization is found within hydrothermalized rocks of the Castro Group. The mineral exploration was developed by Terra Pura, a Brazilian company. All data regarding this area was released for this research. For this study we considered data from reverse circulation and diamond drill holes. The first question that arises from using data coming from different drilling methods is the sampling support and consequently its validation. Usually, validation of such data is made using the traditional approach of comparing twin drill holes, in which one is reverse circulation and the other is diamond drill. This is very difficult because we have geological variation even for short distances. Moreover, when both data are compared in a scattergram, the correlation coefficient is usually low. Thus, in this work we used a new statistical approach to compare frequency distributions of samples collected from reverse circulation and diamond drill holes. The result was very good showing a match of more than 90% that implies in a possible change of exploration methods, because reverse circulation is much cheaper than diamond drill holes. This result allowed us to use both data sets to develop further steps of this research aiming the geological modeling and ore resource estimation using geostatistical methods. In this area hydrothermal rocks occur according to a local trend striking N45E and dipping 30 SE. The structural information was used for both geological and geostatistical modeling of mineral resources resulting and compatible and useful models for understanding the distribution of gold.

In this study, a three-dimensional, three-phase dynamic simulation model based on geological investigations of Korolev oilfield in Kazakhstan was used as a development planning tool in order to improve performance of three new wells. The model, developed previously by means of a seismic study, well log and core data, incorporating with characteristics of oilfield productivity, properties of reservoir, liquids and gases that are saturating the hydrocarbon-bearing horizon can be used to calculate development parameters for Korolev field, including production well locations, drilling schedules, and to facilitate both long-term and short-term forecasting for the purposes of optimizing the hydrocarbon recovery from the field. The objective of this work is to assess the impact of adding 3 extra producing wells and find ways to optimize cumulative production with the least impact on the existing development plan by means of deeper understanding subsurface dynamic processes of the Korolev field. The challenge is a high degree of connectivity between wells in the productive formation throughout the field so that any change of production parameters affects the whole field&rsquo / s cumulative production. Trying to find a solution to optimum production of the reservoir forecast studies were carried out, the impact of each new well on development parameters was defined, sub-surface processes changes due to extra producers lead-in were explained and as a result of this thesis two optimization models were proposed, one of which will bring nearly 9.7 million barrels more oil.

Thesis (MSc)--Stellenbosch University, 2011. / ENGLISH ABSTRACT: This study presents an outcrop characterisation and modelling of the excellently exposed Permian Kookfontein Formation of the Ecca Group in the Tanqua-Karoo sub-Basin. The sedimentary modelling (i.e. facies architecture and geometry) and petrophysical characterisation followed a hierarchical and deterministic approach. Quantitative outcrop data were based on the thirteen sedimentary cycles that characterise this stratigraphic succession at the Pienaarsfontein se Berg locality; and these data were analysed using a combination of detailed sedimentary log, gamma ray log and photopanel analysis, as well as petrographic thin-section and grain size-based petrophysical analysis. Based on texture and sedimentary structures, twelve depofacies are recognised which are broadly grouped into four lithofacies associations i.e. sandstone facies, heterolithic facies, mudstone facies and soft-sediment deformation facies; these depofacies and lithofacies form the basic building blocks for the flooding surface-bounded facies succession (i.e. cycle). Also, based on sediment stacking and cycle thickness patterns as well as relative position to the shelf break, the succession is sub-divided into: (1) the lower Kookfontein member (i.e. cycles 1 to 5) exhibiting overall upward thickening and coarsening succession with progradational stacking pattern; representing deposition of mid-slope to top-slope/shelf-margin succession, and (2) the upper Kookfontein member (i.e. cycles 6 to 13) exhibiting overall upward thickening and coarsening succession with aggradational stacking pattern; representing deposition of top-slope/shelf-margin to outer shelf succession. Lateral juxtaposition of observed vertical facies variations across each cycle in an inferably basinwards direction exhibits upward change in features, i.e. decrease in gravity effects, increase in waves and decrease in slope gradient of subsequent cycles. This systematic upward transition in features, grading vertically from distal to proximal, with an overall upward thickening and coarsening progradational to aggradational stacking pattern indicates a normal regressive prograding delta. However, in detail, cycles 1-3 show some anomalies from a purely thickening and coarsening upward succession. Deposition of each cycle is believed to result from: (1) primary deposition by periodic and probably sporadic mouthbar events governed by stream flow dynamics, and (2) secondary remobilisation of sediments under gravity. The facies distribution, architecture and geometry which governs the sedimentary heterogeneity within the deltaic succession is therefore mainly a consequence of the series of mouthbar flooding events governed by sediment supply and base-level changes. These series of flooding events resulted in the delineation of the studied stratigraphic interval into two main parasequence sets, i.e. transgressive sequence set and the overlying regressive sequence set. This delineation was aided through the identification of a maximum flooding surface (i.e. maximum landwards shift in facies) above Cycle 3 in the field. The architecture and geometry of the ensuing deposystem is interpreted to have been a river-dominated, gravitationally reworked and waveinfluenced shelf edge Gilbert-type delta. Widespread distribution of soft-sediment deformation structures, their growth-style and morphology within the studied succession are empirically related to progradation of Gilbert-type mouthbars over the shelf break as well as the slope gradients of the Kookfontein deltaic clinoformal geometry. Analysis of hypothetical facies stacking and geometrical models suggests that the Kookfontein sedimentary cyclicity might not be accommodation-driven but rather sediment supply-driven. The workflow employed for petrophysical evaluation reveals that the distribution of reservoir properties within the Kookfontein deltaic sandbody geometries is strongly influenced both by depositional processes and by diagenetic factors, the latter being more important with increased burial depth. The reservoir quality of the studied sandstones decreases from proximal mouthbar sands, intermediate delta front to distal delta front facies. The major diagenetic factors influencing the reservoir quality of the studied sandstones are mechanical compaction, chemical compaction (pressure solution) and authigenic pore-filling cements (quartz cement, feldspar alteration and replacement, calcite cement, chlorite and illite). Mechanical compaction was a significant porosity reducing agent while cementation by authigenic quartz and clay minerals (i.e. illite and chlorite) might play a major role in permeability distribution. The porosity-permeability relationship trends obtained for the studied sandstones show that there is a linear relationship between porosity and permeability. The relative timing of diagenetic events as well as the percentages of porosity reduction by compaction and cementation indicates that compaction is much more responsible for porosity reduction than cementation. The described internal heterogeneity in this work is below the resolution (i.e. mm-scale) of most conventional well-logs, and therefore could supplement well-log data especially where there is no borehole image and core data. The combination of ‗descriptive‘ facies model and schematic geological model for this specific delta, and petrophysical characterisation make the results of this study applicable to any other similar ancient deposystem and particularly subsurface reservoir analogue. / AFRIKAANSE OPSOMMING: Hierdie studie bied ‘n dagsoomkarakterisering en -modellering van die duidelik blootgelegde Permiese Kookfontein-formasie van die Ecca-groep in die Tankwa-Karoo-subkom. Wat die sedimentêre modellering (d.w.s. fasiesargitektuur en -geometrie) en petrofisiese karakterisering betref is ‘n hiërargiese en deterministiese benadering gevolg. Kwantitatiewe dagsoomdata is gebaseer op dertien sedimentêre siklusse wat hierdie stratigrafiese opeenvolging in die Pienaarsfontein se Berg-lokaliteit kenmerk; en die data is geanaliseer met behulp van ‘n kombinasie van gedetailleerde sedimentêre seksie, gammastraal-profiel en fotopaneelanalises, asook petrografiese slypplaatjie- en korrelgrootte-gebaseerde petrofisiese analises. Op grond van tekstuur en sedimentêre strukture is twaalf afsettingsfasies onderskei wat rofweg in vier assosiasies van litofasies gegroepeer kan word: sandsteenfasies, heterolitiese fasies, moddersteenfasies en sagtesediment-deformasiefasies. Hierdie afsettingsfasies en litofasies vorm die basiese boustene vir die fasiesopeenvolging (d.w.s. siklus) wat oorstromingsoppervlakgebonde is. Verder word die opeenvolging aan die hand van sedimentstapeling en skilusdiktepatrone, asook relatiewe posisie tot die rakbreuk, in die volgende onderverdeel: (1) die benede-Kookfontein-deel (d.w.s. siklus 1 tot 5), wat in die geheel ‘n opwaartse verdikkings- en vergrowwingsopeenvolging met ‘n progradasiestapelpatroon vertoon en die afsetting van middelhelling-tot-boonstehelling- of rakrand-opeenvolging verteenwoordig, en (2) die benede-Kookfontein-deel (d.w.s. siklus 6 tot 13) wat in die geheel ‘n opwaartse verdikkings- en vergrowwingsopeenvolging met ‘n aggradasiestapelpatroon vertoon en die afsetting van boonste helling- of rakrand-tot-buiterakopeenvolging verteenwoordig. Die laterale jukstaposisie van waargenome vertikale fasiesvariasies oor elke siklus heen, in ‘n afleibare komwaartse rigting, vertoon opwaartse verandering wat kenmerke betref, naamlik afname in gravitasiegevolge, toename in golwe en afname in die hellinggradiënt van daaropvolgende siklusse. Hierdie stelselmatige opwaartse oorgang van kenmerke, wat vertikaal van distaal tot proksimaal gradiënteer en in die geheel opwaartse verdikking en vergrowwing in ‘n progradasie-tot-aggradasie-stapelpatroon vertoon, dui op ‘n normale regressiewe progradasiedelta. Van naby beskou, vertoon siklus 1-3 egter bepaalde afwykings van ‘n suiwer opwaartse verdikkings- en vergrowwingsopeenvolging. Die afsettings van elke siklus is vermoedelik die gevolg van: (1) primêre afsetting deur periodieke en waarskynlik sporadiese mondversperringsgebeure wat deur stroomvloeidinamika beheer word, en (2) sekondêre hermobilisering van sedimente deur gravitasie. Die fasiesverspreiding, -argitektuur en -geometrie wat die sedimentêre heterogeniteit in die deltaïese opeenvolging beheer, is dus hoofsaaklik ‘n gevolg van die reeks oorstromingsgebeure by die mondversperring, wat deur sedimentvoorsiening en basisvlakveranderings beheer word. Hierdie reeks oorstromingsgebeure het gelei tot die delineasie van die bestudeerde stratigrafiese interval volgens twee hoofparasekwensie stelle, naamlik die transgressiewe opeenvolgings- en die oordekkende, regressiewe opeenvolgingsgroep. Dié delineasie word ondersteun deur die feit dat ‘n maksimum oorstromingsoppervlak (d.w.s. maksimum landwaartse verskuiwing in fasies) bo siklus 3 in die veld uitgeken is. Die argitektuur en geometrie van die daaropvolgende afsettingstelsel word geïnterpreteer as behorende tot ‘n Gilbert-rakranddelta wat deur ‘n rivier gedomineer, deur gravitasie herbewerk en deur golfwerking beïnvloed is. Die wye verspreiding van sagtesediment-deformasiestrukture, en die groeiwyse en morfologie daarvan binne die bestudeerde opeenvolging, is empiries verwant aan die progradasie van Gilbertmondversperrings oor die rakbreuk heen, asook aan die hellinggradiënte van die Kookfontein-deltaïese, klinoformele geometrie. Die analise van hipotetiese fasiesstapeling en geometriese modelle dui daarop dat die Kookfontein-sedimentêre siklisiteit dalk nie deur akkommodasieruimte gedryf word nie, maar deur sedimentvoorsiening. Die werkvloei wat vir petrofisiese evaluering gebruik is dui daarop dat die verspreiding van reservoir-eienskappe in die Kookfontein- deltaïese sandliggaam geometries sterk beïnvloed word deur afsettingsprosesse en diagenetiese faktore. Die diagenetiese faktore word belangriker op groter begrawing diepte. Die reservoir-aard van die bestudeerde sandgesteentes neem algaande af van proksimale mondversperring-sandsoorte tot intermediêre deltafront tot distale deltafrontfasies. Die hoof-diagenetiese faktore wat die reservoir-kenmerke van die bestudeerde sandsteensoorte beïnvloed is meganiese verdigting, chemiese verdigting (oplossingsdruk) en outigeniese porievullingsement (kwartssement, veldspaatomsetting en -vervanging, kalsietsement, chloriet en illiet). Meganiese verdigting is ‘n beduidende poreusheidreduseermiddel, terwyl sementering deur outigeniese kwarts- en kleiminerale (d.w.s. illiet en chloriet) moontlik ‘n belangrike rol by permeabiliteitsverspreiding kan speel. Die poreusheid-permeabiliteit-verhoudingstendense wat bekom is vir die bestudeerde sandsteensoorte dui daarop dat daar ‘n lineêre verhouding tussen poreusheid en permeabiliteit bestaan. Die relatiewe tydberekening van diagenetiese gebeure, asook die persentasie poreusheidvermindering deur verdigting en sementering, dui daarop dat verdigting baie meer as sementering tot poreusheidvermindering bydra. Die interne heterogeniteit wat in hierdie werk beskryf word, is onder die resolusie (d.w.s. mm-skaal) van die meeste konvensionele boorgatopnames, en kan dus boorgatopnamedata aanvul, veral waar daar geen boorgatafbeelding en kerndata bestaan nie. Die kombinasie van die 'deskriptiewe‘ fasiesmodel en skematiese geologiese model vir hierdie spesifieke delta, asook petrofisiese karakterisering, beteken dat die resultate van hierdie studie op enige ander soortgelyke antieke afsettingstelsels toegepas kan word, maar veral op suboppervlakreservoir-analoogstelsels.

Les objectifs principaux de la géomodélisation sont la représentation et la compréhension du sous-sol. Les structures géologiques ont un rôle important pour comprendre et prédire son comportement physique. Nous avons défini un modèle géologique comme étant composé d'un ensemble de structures et de leurs connexions. Les maillages sont des supports numériques servant à résoudre les équations modélisant la physique du sous-sol. Il est donc important de construire un maillage représentant un modèle géologique afin de prendre en compte l'impact de ces structures dans les phénomènes du sous-sol. L'objectif de cette thèse est de développer des méthodes de maillage volumique pour les modèles géologiques. Nous proposons une méthode de génération de maillages non structurés volumiques permettant de construire deux types de maillages : un maillage tétraédrique et un maillage hex-dominant (c'est-à-dire composé de tétraèdres, prismes à base triangulaire, pyramides à base quadrilatérale et hexaèdres). Cette méthode génère dans un premier temps un maillage tétraédrique qui peut respecter différents types de données : (1) un modèle géologique défini par frontières afin de capturer les structures dans le maillage volumique, (2) des trajectoires de puits représentées par un ensemble de segments, (3) une propriété de taille d'éléments afin de contrôler la longueur des arêtes des éléments et (4) un champ de directions pour contrôler des alignements de sommets/éléments dans le maillage afin de favoriser certaines caractéristiques comme des éléments possédant des angles droits. Dans un deuxième temps, ce maillage tétraédrique peut être transformé en un maillage multi-éléments. La méthode reconnaît des relations combinatoires entre tétraèdres permettant l'identification de nouveaux éléments comme les prismes, les pyramides et les hexaèdres. Cette méthode est ensuite utilisée pour générer des maillages aux caractéristiques spécifiques correspondant à une application donnée afin de limiter les erreurs lors du calcul numérique. Plusieurs domaines d'applications sont considérés tels que les simulations géomécaniques, d'écoulements et de propagation d'ondes sismiques. / The geomodeling main goals are to represent and understand the subsurface. The geological structures have an important role for understanding and predicting its physical behavior. We defined a geological model as a set of structures and their connections. The meshes are numerical supports to solve the equations modeling the subsurface physics. So it is important to build a mesh representing a geological model to take into account the impact of these structures on the subsurface phenomena. The objective of this thesis is to develop volumetric meshing methods for geological models. We propose a volumetric unstructured meshing method to build two mesh types: an adaptive tetrahedral mesh and an hex-dominant mesh (i.e. made of tetrahedra, triangular prisms, quadrilateral pyramids and hexahedra). This method generates first a tetrahedral mesh that can respect different types of data: (1) a geological model defined by its boundaries to capture the structures in the volumetric mesh, (2) well paths represented as a set of segments, (3) a mesh size property to control the mesh element edge length and (4) a direction field to control vertex/element alignments inside the mesh to increase some features such as elements with right angles. Then, this tetrahedral mesh can be transformed in a mixed-element mesh. The method recognizes combinatorial relationships between tetrahedra to identify new elements such as prisms, pyramids and hexahedra. This method is then used to generate meshes whose features correspond to a given application in order to reduce errors during numerical computation. Several application domains are considered such as geomechanical, ow and wave propagation simulations.

Dès le début du 20ème siècle, les eaux souterraines de l’aquifère transfrontalier Milk River (MRA) ont constitué une importante ressource dans le sud de l’Alberta (Canada) et le nord du Montana (USA). L’utilisation intensive de cette ressource sous un climat semi-aride a provoqué une baisse importante des niveaux d’eau localement, si bien que des inquiétudes concernant la pérennité du MRA sont apparues dès les années 60. Les études précédentes du MRA étaient limitées par les frontières nationales, empêchant ainsi une compréhension complète de la dynamique de l’aquifère. L’objectif de cette thèse était de réaliser une étude transfrontalière du MRA afin de caractériser cette ressource régionale d’eau souterraine selon ses limites naturelles. À cette fin, trois modèles transfrontaliers ont été réalisés: un modèle géologique, un modèle conceptuel hydrogéologique et un modèle numérique d’écoulement souterrain. Des travaux de terrain de part et d’autre de la frontière Canada/USA et une revue de littérature exhaustive des études précédentes ont supporté la réalisation de ces modèles. De plus, une première évaluation des niveaux d’exploitation historiques du MRA a été réalisée en Alberta.Le modèle géologique 3D (50 000 km2) représente le MRA (Membre Virgelle de la Formation Milk River/Eagle) et ses unités confinantes continûment à travers la frontière internationale. Le développement de ce modèle a requis une harmonisation des nombreuses nomenclatures stratigraphiques de la région et la délimitation transfrontalière du MRA.Le modèle conceptuel hydrogéologique du MRA a montré deux flux d’écoulement souterrains transfrontaliers, dirigés du Montana vers l’Alberta. La Milk River intercepte la majeure partie du flux souterrain venant du sud, si bien que l’écoulement au nord de la rivière est très faible. Les analyses isotopiques confirment que les eaux du MRA sont principalement fossiles à l’exception de la zone de recharge située le long de la zone d’affleurement du MRA. Les zones du MRA avec une forte conductivité hydraulique et de faibles concentrations en chlorures correspondent à des voies d’écoulement préférentielles. L’émergence des eaux du MRA a lieu via la drainance à travers les unités confinantes notamment le long des vallées enfouies.Le modèle numérique (26 000 km2) d’écoulement souterrain du MRA en régime permanent est une transposition des modèles géologique et conceptuel. Il montre que le modèle conceptuel précédemment développé est hydrauliquement plausible. Le modèle d’écoulement donne une meilleure compréhension du système aquifère en représentant la dynamique de l’écoulement souterrain dans la situation pré-exploitation. Le traçage de particules indique des temps de résidence advectifs de près de 750 000 ans à la limite nord du MRA, ce qui est inférieur aux âges obtenus par les analyses isotopiques (2 Ma).Les bilans en eau des modèles conceptuels et numériques montrent tous les deux que l'extraction de l'eau souterraine dépasse de loin la recharge au nord de la Milk River.Un modèle d'écoulement transitoire serait requis pour définir le volume d'eau provenant de l'emmagasinement dans le MRA et préciser le rôle des aquitards durant le pompage. Par ailleurs, en tant que ressource internationale partagée, une gestion transfrontalière des eaux du MRA serait justifiée dans la région comprise entre la zone de recharge au Montana et le sud de la Milk River en Alberta. Ainsi, ces trois modèles transfrontaliers du MRA forment une base commune internationale de connaissances scientifiques à l'échelle de l'aquifère et pourraient supporter l'évaluation future du meilleur usage possible de cette ressource partagée et limitée. De futurs travaux peuvent inclure l'effet des champs de gaz localisés aux limites du MRA et les conditions paléo-hydrogéologiques liées à l'évolution géochimiqie de l'au souterraine. / Since the beginning of the 20th century, groundwater from the Milk River Aquifer (MRA) has been an important water resource in southern Alberta (Canada) and northern Montana (USA). The intensive use of this resource under a semi-arid climate has locally led to important drops in waters levels, thus raising concerns on the MRA sustainability since the 1960’s. Previous MRA studies were limited by the national boundaries, thus preventing a full understanding of the aquifer dynamics. The objective of this thesis was to carry out a transboundary study of the MRA to define this regional groundwater resource within its natural boundaries. For this purpose, three cross-border models were developed: a geological model, a conceptual hydrogeological model and a numerical groundwater flow model. These developments were supported by focused field work on both sides of the Canada/USA border and a comprehensive review of previous studies. Furthermore, a first assessment of the historical exploitation levels of the MRA was done for southern Alberta.The 3D geological model (50,000 km2) continuously represents the MRA (Virgelle Member of the Milk River/Eagle Formation) and confining units through the international border. Development of this model required harmonization of various stratigraphic nomenclatures in the study area and the transboundary delineation of the MRA extent.The hydrogeological conceptual model of the MRA indicated two transboundary groundwater fluxes from Montana to Alberta. The Milk River intercepts most of the groundwater flux incoming from the south, thus leading to limited groundwater flow north of the Milk River. Isotopic analyses confirm that the MRA contains mostly fossil groundwater, with the exception of the recharge area located along the outcrop area of the aquifer. Areas within the MRA with a high hydraulic conductivity and low chloride concentrations correspond to preferential groundwater flowpaths. Discharge from the aquifer occurs as vertical leakage through the overlying confining units, especially along bedrock valleys.The numerical steady-state groundwater flow model of the aquifer system (26,000 km2) is a transposition of the geological and conceptual models, and it shows that the previously developed conceptual model is hydraulically plausible. The groundwater flow model provides a better understanding of the aquifer system by representing the dynamics of groundwater flow under pre-development conditions. Particle tracking indicates advective residence times of up to 750,000 years in the northern limit of the aquifer, which is lower than ages obtained from isotopic analyses (2 Ma).The area south of the Milk River benefits from all of the transboundary groundwater flux from Montana whereas north of the Milk River, only a portion of the recharge flux coming from the south is transmitted due to the interception of the Milk River.Groundwater budgets from both the conceptual and numerical models show that groundwater extraction far exceeds recharge north of the Milk River. A transient model would be required to determine how much water is derived from storage in the MRA and to assess the role of the confining units during pumping. Furthermore, as an internationally shared resource, a transboundary management of the MRA would be appropriate, especially in the area comprised between the recharge area in Montana and the Canadian reach of the Milk River.Together, these transboundary models of the MRA constitute a common international basis of scientific knowledge at the aquifer scale and could support further assessments of the best possible exploitation of the shared but limited resource.Future works could include the study of the effects of the gas fields located at the limits of the MRA as well as paleo-hydrogeological conditions relative to the geochemical evolution of groundwater.

A caracterização geológica e geotécnica dos locais de implantação de aterros sanitários é fundamental, tanto para fornecer subsídios adequados ao projeto de instalação das células de deposição dos resíduos sólidos urbanos (RSU), quanto para que se conheçam as características da área antes de ser modificada pela obra. Neste sentido, o presente trabalho visou mostrar a contribuição que a geofísica, com o uso da técnica da eletrorresistividade e do potencial espontâneo, associado a informações de investigação geotécnica direta pode dar para o estabelecimento de um background da área de instalação do aterro sanitário da cidade de São Carlos-SP. Foram realizados 7 caminhamentos elétricos (CE) com o arranjo dipolo-dipolo com eletrodos espaçados de 10 metros e três sondagens elétricas verticais (SEV), utilizando o arranjo Schlumberger, com profundidade de investigação de até 100 metros. Para processamento dos modelos bidimensionais dos CE utilizou-se o programa RES2DINV, e enquanto que para as SEV foi utilizado o modelo unidimensional. A interpretação dos modelos geoelétricos foi aperfeiçoada pela confrontação com informações geológico-geotécnicas prévias da área, obtidas nos estudos de impacto ambiental (EIA/RIMA) e resultados de sondagens de simples reconhecimento (SPT), realizadas na fase de estudo para implantação do aterro sanitário. Com isso foi possível definir com clareza a profundidade do topo rochoso e das zonas de maior umidade nos CE e SEV. O conjunto de seções de CE foi georeferenciada com o auxílio de GPS de precisão, o que permitiu o tratamento espacial das informações e o estabelecimento dos mapas da superfície do topo rochoso, das zonas de fluxo de água subterrânea e da espessura dos materiais inconsolidados. O conjunto de informações obtido permitiu identificar locais possíveis para a instalação futura de poços de monitoramento, assim como os dados geoelétricos das várias camadas servirá como padrão para identificar eventuais contaminações do solo ou da água subterrânea no futuro. A identificação da presença do topo rochoso em pequena profundidade, em algumas áreas do aterro, permite o melhor planejamento das escavações para instalação das células de deposição do RSU. / The geological and geotechnical characterization of landfill implantation areas is crucial; even to provide appropriate subsidies of urban solid residues (USR) installation project of deposition cells and for knowing the characteristics of the area before being modified by the work. In this sense, the present work aimed to show the contribution that the geophysics, with the use of electrical resistivity technique and the spontaneous potential, associated with direct geotechnical investigation information can give to the establishment of a background of landfill installation area from São Carlos-SP city. Were performed 7 electrical imaging (EI) with the dipole-dipole array with electrodes spaced by 10 meters and three vertical electrical sounding (VES), using the Schlumberger array with depth research up to 100 meters. For processing of two-dimensional models of the EI was used the RES2DINV program and while for the vertical electrical sounding (VES)-was used one-dimensional model. Interpretation of geoelectric models was enhanced by the confrontation with geological and geotechnical information of prior area, obtained in the environmental impact assessment (EIA) and results of a simple reconnaissance probes, performed in the study phase for the landfill implementation. It was possible to clearly define the depth of the bedrock and the higher humidity areas in the EI and VES. The set of sections electrical Imaging (EI) was georeferenced with the aid of GPS accuracy that allowed the spatial treatment of Information and the setting maps of the bedrock surface, the groundwater flow zones and the thickness of the unconsolidated materials. The obtained set of information enabling identification of possible locations for the wells future monitoring installation, as well as the geoelectrical data from several layers will be use as a standard to identify possible contamination of the soil or groundwater in the future. The identification of bedrock presence in small depth, in some areas of landfill, allows better planning of excavation for installation of MSW deposition cells.

[pt] O principal objetivo deste trabalho é preencher canais turbidíticos com alguma propriedade petrofísica, como a permeabilidade. Estes canais são geometricamente limitados por lobos turbidíticos, gerando a simulação de um sistema deposicional. Simulações numéricas são usadas para tentar ajustar a permeabilidade a um caso de referência por meio de uma função objetivo. Um simulador convencional de diferenças finitas foi usado para comparar os dados de referência com as simulações, obtendo resultados próximos. / [en] The main objective of this work is to populate turbidite channels with some petrophysical property such as permeability. These channels are geometrically constrained by turbidite lobes creating a simulated depositional system. Numerical simulations are used to try to fit the permeability field to a reference case through an objective function. A conventional finite difference simulator was used to compare the reference data to the simulations, obtaining close results.

A caracterização geológica e geotécnica dos locais de implantação de aterros sanitários é fundamental, tanto para fornecer subsídios adequados ao projeto de instalação das células de deposição dos resíduos sólidos urbanos (RSU), quanto para que se conheçam as características da área antes de ser modificada pela obra. Neste sentido, o presente trabalho visou mostrar a contribuição que a geofísica, com o uso da técnica da eletrorresistividade e do potencial espontâneo, associado a informações de investigação geotécnica direta pode dar para o estabelecimento de um background da área de instalação do aterro sanitário da cidade de São Carlos-SP. Foram realizados 7 caminhamentos elétricos (CE) com o arranjo dipolo-dipolo com eletrodos espaçados de 10 metros e três sondagens elétricas verticais (SEV), utilizando o arranjo Schlumberger, com profundidade de investigação de até 100 metros. Para processamento dos modelos bidimensionais dos CE utilizou-se o programa RES2DINV, e enquanto que para as SEV foi utilizado o modelo unidimensional. A interpretação dos modelos geoelétricos foi aperfeiçoada pela confrontação com informações geológico-geotécnicas prévias da área, obtidas nos estudos de impacto ambiental (EIA/RIMA) e resultados de sondagens de simples reconhecimento (SPT), realizadas na fase de estudo para implantação do aterro sanitário. Com isso foi possível definir com clareza a profundidade do topo rochoso e das zonas de maior umidade nos CE e SEV. O conjunto de seções de CE foi georeferenciada com o auxílio de GPS de precisão, o que permitiu o tratamento espacial das informações e o estabelecimento dos mapas da superfície do topo rochoso, das zonas de fluxo de água subterrânea e da espessura dos materiais inconsolidados. O conjunto de informações obtido permitiu identificar locais possíveis para a instalação futura de poços de monitoramento, assim como os dados geoelétricos das várias camadas servirá como padrão para identificar eventuais contaminações do solo ou da água subterrânea no futuro. A identificação da presença do topo rochoso em pequena profundidade, em algumas áreas do aterro, permite o melhor planejamento das escavações para instalação das células de deposição do RSU. / The geological and geotechnical characterization of landfill implantation areas is crucial; even to provide appropriate subsidies of urban solid residues (USR) installation project of deposition cells and for knowing the characteristics of the area before being modified by the work. In this sense, the present work aimed to show the contribution that the geophysics, with the use of electrical resistivity technique and the spontaneous potential, associated with direct geotechnical investigation information can give to the establishment of a background of landfill installation area from São Carlos-SP city. Were performed 7 electrical imaging (EI) with the dipole-dipole array with electrodes spaced by 10 meters and three vertical electrical sounding (VES), using the Schlumberger array with depth research up to 100 meters. For processing of two-dimensional models of the EI was used the RES2DINV program and while for the vertical electrical sounding (VES)-was used one-dimensional model. Interpretation of geoelectric models was enhanced by the confrontation with geological and geotechnical information of prior area, obtained in the environmental impact assessment (EIA) and results of a simple reconnaissance probes, performed in the study phase for the landfill implementation. It was possible to clearly define the depth of the bedrock and the higher humidity areas in the EI and VES. The set of sections electrical Imaging (EI) was georeferenced with the aid of GPS accuracy that allowed the spatial treatment of Information and the setting maps of the bedrock surface, the groundwater flow zones and the thickness of the unconsolidated materials. The obtained set of information enabling identification of possible locations for the wells future monitoring installation, as well as the geoelectrical data from several layers will be use as a standard to identify possible contamination of the soil or groundwater in the future. The identification of bedrock presence in small depth, in some areas of landfill, allows better planning of excavation for installation of MSW deposition cells.