La connaissance de la dynamique hydrosédimentaire des rivières avant et pendant l’installation des nombreux aménagements hydrauliques depuis l’Antiquité restaient très fragmentaire en Basse-Normandie et plus largement en Europe de l’Ouest. On savait encore peu de choses sur les processus d’érosion et de sédimentation qui se sont succédés au sein des rivières normandes au cours des deux derniers millénaires et la part des forçages anthropiques et climatiques qui ont influencé ces processus. L’objectif de la thèse est donc de comprendre et de mesurer la part des héritages dans le fonctionnement actuel des rivières afin de contribuer à leur gestion contemporaine. A cette fin, des recherches géomorphologiques et géoarchéologiques ont été menées dans plusieurs vallées bas-normandes. Ces études ont mis en évidence le poids des installations hydrauliques et en particulier le développement des moulins à eau dans la modification des chenaux d’écoulement mais aussi la construction des plaines alluviales. En effet, la structuration complète du cours d’eau depuis le Moyen Âge a fortement artificialisé les formes en plan et les pentes des cours d’eau et complètement cloisonné leur linéaire. Ces transformations ont favorisé la stabilité latérale des cours d’eau et la sédimentation limoneuse par débordement au sein des plaines alluviales. L’installation de moulins s’accompagne très souvent d’une simplification du tracé en plan de la rivière. Lors de l’équipement maximal des cours d’eau normands on trouvait un moulin tous les 2500 mètres de linéaires hydrographiques. Ces transformations signifient qu’une grande partie des cours actuels sont artificiels. En effet, ces aménagements et l’accélération de la sédimentation limoneuse depuis 1000 ans sont à l’origine de la mise en place d’un équilibre dynamique aboutissant aux formes hydrosédimentaires des rivières actuelles. L’abandon de la gestion des ouvrages hydrauliques et leur destruction au cours des cinquante dernières années viennent remettre en cause cet équilibre. / In Lower Normandy, knowledge of the hydrosedimentary dynamics of rivers before and during the installation of the numerous hydraulic installations since the Roman period remains very fragmentary. Little was known about the rates of erosion and sedimentation that have occurred over the last two millennia and how climatic and anthropogenic controls have influenced these processes. Thus, the challenge of this phD is to understand and measure the role of the inherited structures and operating in the current functioning of the rivers of Lower Normandy in order to help with their contemporary management. To provide answers, geomorphological and geoarchaeological researches has been carried out in several valleys. These studies make it possible to highlight the weight of the hydraulic installations and in particular the developments related to the mills in the setting up of floodplains. Indeed, the complete structuring and management of the watercourse since the Middle Ages artificialized riverbed forms and slopes, and partitioned rivers favoring lateral stability and overflow sedimentation in the floodplain conducting in the development of completely regulated rivers. The installation of mills is accompanied locally by a simplification of the river course. At the acme of the equipment there was a mill every 2,500 meters along the rivers of Lower Normandy. These transformations mean that a large part of current courses are artificial. These structures and the increase of the overbank silt sedimentation for 1000 years are at the origin of the establishment of a dynamic equilibrium resulting in the hydrosedimentary forms of the present rivers which are largely inherited. The abandonment of the management of the hydraulic structures and their destruction during the last fifty years comes to question this balance.

This dissertation addresses how Paleoindians used the karst drainage of the Aucilla River in northwestern Florida during the Pleistocene/Holocene transition (approximately 15-10,000 14C yr B.P.). I take a geoarchaeological approach to discuss Paleoindian land use by first defining the Late Pleistocene and Holocene geological record, and then by creating a model of site formation processes in the Aucilla River. Both underwater and terrestrial fieldwork were performed. Underwater fieldwork consisted of hand-driven cores and surface survey, vibrocoring, underwater 1 x 1 m unit excavation, and controlled surface collection. Terrestrial fieldwork consisted of shovel and auger test pits. Seventeen cores were collected from five different submerged sinkhole sites, which were used to select two sites for further study: Sloth Hole (8JE121), which had been previously excavated, and Wayne's Sink (8JE1508/TA280), which was recorded but not formally investigated. Five vibrocores and two 1 x 1m units were used, with previous research, to define the geological and geoarchaeological context of Sloth Hole. Fifteen vibrocores, six 1 x 1 m excavation units, and ten 1 x 1 m surface collection units were used to define the geological, geoarchaeological, and archaeological context of Wayne's Sink. A combination of 130 shovel and auger test pits was used to define the geological, geoarchaeological, and archaeological potential of the terrestrial landscape. Five new Holocene-aged terrestrial sites were recorded. All of these data were evaluated with archival data from previously-excavated sites to create models of site formation and Paleoindian land use in the lower Aucilla Basin. This research shows that there have been four major periods of sinkhole infill in the lower Aucilla basin. The first occurred prior to the Last Glacial Maximum, with each sinkhole containing peat deposits that date in excess of 21,000 calendar years ago (cal B.P.). These peats are overlain by sandy colluvium that dates to approximately 14,500 cal B.P. The colluvium is overlain by clays that contain evidence for soil formation. These soils vary in age, with radiocarbon dates of approximately 14,500-10,000 cal B.P. These clays are directly overlain by peats dating to 5,000-3,500 cal B.P., which are overlain by peats and clays that date to 2,500-0 cal B.P. Intact Paleoindian and Early Archaic deposits are possible in the late Pleistocene soils.

This thesis deals with questions on various applications using soils and sediments as sources of information in archaeological research. Human environmental impact on soils and sediments, in terms of pollution, is a well known phenomenon as the industrialisation intensified during historical times and onwards and has left strong pollutive marks. However, humans have always accumulated or emitted matter and various compounds in connection to their habitats for subsistence, but these earlier traces are not always detectable, depending on soil and sediment state of preservation. Bioessential elements are intimately linked to humans and their dwellings and especially phosphate has been evident in this respect. It was established already in the 1930s, that even Stone Age settlements could be located through elevated phosphate content in extensive soil phosphate mappings. This thesis is a compilation on results from several sites and excavations from the southern to the northern parts of Sweden. There is a wide variety of soil types and chronological setting in the material, from highly acid podzols to calcareous soils, and sediments dated to Younger Dryas to current top soils. Sites from the Mesolithic, Neolithic, Bronze Age, Early and Late Iron Age and Medieval are all represented. Methods that have been implemented are firstly various forms of analysis in regards to soil phosphate, magnetic susceptibility and organic matter. Furthermore, metal and non-metal elements have been considered, as well as lead isotopes. A multiproxy approach is applied in some examples where biological and chemical data is combined to interpret past events. In this thesis there are also five papers presented. The first paper deals with methodological issues concerning multi-element analyses of various soil samples (off-site to feature) from an archaeological excavation. The second paper is about the possibilities that may be used when analysing the soil organic phosphate in relation to prehistoric agriculture. Paper three and fourth are compilations of large scale contract archaeological project. These papers deal with theoretical, methodological and practical issues concerning environmental archaeology in relation to contract archaeology. Studies on landscape development and erosion are among the cases presented. The last paper deals with a late Mesolithic - early Neolithic settlement in Vuollerim, N. Sweden, and spatial dimensions on the human use of settlement (off-site to on-site) and house floors (intra-site), are discussed.

Les perturbations environnementales enregistrées aujourd’hui dans le cadre du changementclimatique global exacerbent les craintes des sociétés quant au développement durable. Si l’impact desactions anthropiques sur les changements contemporains est désormais reconnu scientifiquement, lesspécialistes des sociétés anciennes et des géosciences se demandent si les changementsenvironnementaux d’origine naturelle ou anthropique n’ont pas déjà et à plusieurs reprises, affecté ledéveloppement des sociétés au cours des derniers millénaires. Les investigations menées, dans le cadrede cette thèse, proposent de contribuer à ce débat à partir de l’étude sur le temps long de la basse valléedu Strymon (Grèce du Nord) depuis le début de l’Holocène. L’enjeu est de comprendre les relationsqu’ont entretenues les sociétés avec leur environnement afin d’examiner les questions d’éventuelseffondrements culturels d’origine environnementale comme le suggèrent certains chercheurs pour la findu Néolithique ou l’Âge du Bronze, et celles des conséquences des actions des sociétés (néolithisation,conquête grecque, etc.) sur l’environnement et la durabilité de son exploitation. Pour dépasser les étudesde cas, nous proposons une approche géographique multiscalaire et géoarchéologique. Elle s’appuie surune étude des archives paléoenvironnementales combinant (1) une analyse géomorphologique et (2) uneanalyse palynologique des cours d’eau et des zones humides afin de renseigner les changementsenvironnementaux au plus près des lieux de vie des populations et de leurs pratiques agropastorales.Les résultats obtenus sont ensuite confrontés aux recherches archéologiques et historiques pourcomprendre les interactions Sociétés/Environnement de l’échelle de l’habitat à celle de la région étudiéeà différentes époques, et participer activement aux discussions multidisciplinaires sur la question. / Environmental disturbances recorded nowadays in global changes rise up society fears aboutsustainable development. Even if the responsibility of activities on contemporary changes is now wellknown,the specialists of past societies are still wondering if such environmental changes, either fromnatural or from anthropogenic origin, have not already affected societies development in last thousandsof years. This research project contributes to this debate on the base of long time study of Strymona lowvalley (northern Greece), from Early Holocene. The aim is to understand the relationships betweensocieties and environment by questioning the possibility of cultural collapses from environmentalorigin, and to examine the consequences of human activities (Neolithisation, Greek conquest, ...) onenvironment and his sustainability. To overcome case studies, we suggest a geographical andgeoarchaeological approach. Based on study of palaeoenvironmental archives which combinegeomorphological and palynological analyses of streams and wetlands. Our aim is to analyze the role ofagropastoral practices and climate events and oscillations on environmental changes, and on thedynamics of mosaic landscapes. Results will be compared with archaeological and historical researchesto understand the interactions between societies and their environment from sites to regional scale, fromthe Neolithic to Ottoman period and to take part actively to multidisciplinary discussions on this debate.

Ce travail de recherche a mis en œuvre une approche géoarchéologique multi-indicateurs pour analyser les impacts et les forçages anthropo-climatiques des paysages du plateau de Corent et de ses marges de la fin du Néolithique à la période romaine. Une cartographie géomorphologique détaillée du secteur d’étude a été réalisée en s’appuyant sur l’analyse de données LiDAR et la prospection de terrain. Deux secteurs ont été sélectionnés pour conduire des prospections géoarchéologiques approfondies : i) le Lac du Puy, un petit bassin à proximité immédiate du site archéologique sur le plateau volcanique de Corent, et ii) la Narse de la Sauvetat, un grand bassin hydromorphe dans la basse-plaine marno-calcaire située au pied du versant sud-ouest du plateau.Des sondages géoarchéologiques ont été réalisés dans la cuvette du Lac du Puy, conduisant à la découverte d’une batterie d’un millier de silos de stockage du premier âge du Fer. Les résultats de cette campagne de terrain ont également permis de reconstruire le cadre chrono-stratigraphique et les grands traits de l’évolution hydro-morpho-sédimentaire et pédologique de la cuvette. L’analyse géoarchéologique multi-proxys de plusieurs carottes sédimentaires a fourni davantage de précisions sur l’évolution environnementale du Lac du Puy. L’impact anthropique y est nettement perceptible au long de cinq millénaires et apparaît depuis le Néolithique comme le principal facteur de contrôle de l’évolution paléoenvironnementale du bassin jusqu’à son anthropisation complète.En contrebas de ce site de plateau, dans le bassin de la Narse de la Sauvetat, l’étude géomorphologique a révélé que la cuvette hydromorphe dérive de la fermeture par un glissement de terrain d’une paléo-vallée qui était située au pied du versant sud-ouest du puy de Corent. Cet évènement s’est probablement produit au début du Subboréal. Les grandes phases de l’évolution de la cuvette ont pu être reconstruites, d’un fonctionnement alluvial au Pléistocène Supérieur jusqu’à des dynamiques hydromorphes et palustres, sous l’effet du contrôle aval exercé par le barrage de la vallée à partir du Néolithique Final, mettant en évidence le fort potentiel paléoenvironnemental de la cuvette. L’analyse géoarchéologique multi-proxys de trois carottes sédimentaires, combinée aux résultats d’analyses malacologiques externes à cette thèse, a permis de retracer l’évolution environnementale de l’ensemble de la partie basse de la cuvette hydromorphe entre le Néolithique Final et le début du Moyen-âge.Les résultats obtenus dans le Lac du Puy et la Narse de la Sauvetat ont été intégrés dans le système morpho-sédimentaire plus large de l’ensemble du plateau et de ses marges entre 3000 BC et 0 AD, puis mis en perspective géographique et temporelle à l’échelle de l’Europe occidentale nord-alpine. L’évolution du système socio-environnemental de Corent et de Limagne méridionale au premier millénaire BC a été analysée en détail, en discutant le rôle des phénomènes de proto-urbanisation dans le forçage des systèmes morpho-sédimentaires qui marque l’entrée dans l’anthropocène. / This thesis presents the results of a multi-proxy geoarchaeological study of how anthropogenic forces and climate have shaped the evolution of the landscapes of the plateau of Corent (Auvergne, France) from the Neolithic to Roman period. A detailed geomorphological map of the study area was constructed based on LiDAR data and fieldwork. Two areas were selected for geoarchaeological analysis: i) the Lac du Puy, a small pond located within the archaeological site of the plateau of Corent, and ii) la Narse de la Sauvetat, a larger hydromorphic basin in the calcareous lowlands, at the foot of the southwestern slopes of the plateau.Geoarchaeological survey in the Lac du Puy revealed an expansive group of over one thousand storage pits, dating from the Hallstatt period. The results of fieldwork permitted the construction of a chrono-stratigraphic framework and allowed a reconstruction of the main phases of the hydro-morpho-sedimentary and pedological evolution of the basin. Multi-proxy geoarchaeological analysis of sedimentary cores provided additional information concerning palaeoenvironmental evolution of the Lac du Puy. Anthropogenic impact is clearly discernable throughout five millennia starting in the Neolithic, and is identified as the main driver of the palaeoenvironmental evolution of the basin until its final destruction by human activity.In the lowlands, geomorphological analysis of la Narse de la Sauvetat revealed that the hydromorphic basin is the result of the closure of a palaeovalley by a massive landslide issuing from the southwestern slopes of the Puy de Corent. Geomorphological and stratigraphic evidence indicates that the isolation of the basin likely occurred at the beginning of the Subboreal. Lithostratigraphic analysis was used to reconstruct the main phases of the evolution of the basin from its initial alluvial phase in the late Pleistocene, through to its hydromorphic and palustrine dynamics from the late Neolithic, highlighting the strong palaeoenvironmental potential of its sedimentary record. Multi-proxy geoarchaeological analysis of three sedimentary cores from the basin, combined with results of external malacological analysis, allowed a reconstruction of the palaeoenvironmental evolution of the basin from the late Neolithic to the early Middle Ages.Results obtained from the Lac du Puy and La Narse de la Sauvetat were integrated into the broader morpho-sedimentary system of the plateau and its margins for the period between 3000 BC and 1 AD. The geographic and temporal perspective was then further enlarged through comparison with temperate Western Europe data. The evolution of the socio-environmental system of Corent and southern Limagne was analysed in detail for the first millennium BC. Finally the role of proto-urbanization processes in the forcing of morpho-sedimentary systems is examined, marking their entry into the anthropocene.

Nesta dissertação, apresentamos os procedimentos adotados para a elaboração de um modelo geoarqueológico de localização de material arqueológico relacionado aos sítios pré-coloniais. A área de estudo abrangeu os limites da carta topográfica do município de Pindamonhangaba, situado na região do Vale do Paraíba no estado de São Paulo. Para a construção desse modelo, realizamos o mapeamento morfológico de feições que julgamos ser propícias ao assentamento humano, por meio da estereoscopia de fotos aéreas e uso de técnicas de geoprocessamento. Com base na carta topográfica, também se produziu uma carta hipsométrica e perfis topográficos. Após a confecção de um mapa prévio de potencial arqueológico, com destaque para as áreas das grandes depressões, colinas e do terraço fluvial do rio Paraíba do Sul, foram elencados 3 setores para a realização de prospecção arqueológica. A execução das etapas de controle de campo resultou na coleta de um fragmento de cerâmica, encontrado a 3 metros de profundidade, próximo a uma depressão, bem como de alguns fragmentos de quartzo, em relevo escarpado que, após análise em laboratório, foram classificados como duvidosos quanto ao potencial arqueológico. Conclui-se que o emprego da Geoarqueologia no estudo de grupos pré-coloniais do Vale do Paraíba Paulista constituiu-se de uma abordagem eficiente para o entendimento do padrão de assentamento em função da distribuição dos recursos naturais. / In this dissertation, we present the procedures used for the creation of an archaeogeological model employed for the localization of archeological material related to precolonial sites. This study encompassed the city limits of Pindamonhangaba, in the Vale do Paraiba region, located in the state of Sao Paulo. For the purposes of this project, we used morphological mapping we deemed suitable for human settlement, through stereocopy of aerial photographs and geoprocessing techniques. Based on the topographic map, we also produced a hypsometric map and topographic profiles. After the compilation of a previous map with archeological potential, highlighting areas of great depressions, hills and the fluvial terrain of the Paraiba do Sul River, three sectors were selected for the archeological prospection. The field work yielded the discovery of a ceramic fragment, found in a excavation of 3 meters deep into the ground, close to a depression, as well as some quartz fragment. After laboratorial testing, these were found to have no archeological value. It is concluded that employing techniques from Geoarcheology to study precolonial groups in Paraíba Paulista Valley emerges as an efficient approach to the knowledgle of natural resources-oriented settlement patterns.

Abstract. The goal of this case study is to further the understanding of the social and economic structure, such as trade routes and/or prehistoric man’s movement, during prehistory in the inland of Västerbotten, Sweden. This is achieved by studying the sets of lithic tools found in six archaeological sites (RAÄ 977:1 Vilhelmina, RAÄ 553:1 Vilhelmina, RAÄ 132:1 Vilhelmina, RAÄ 519:1 Vilhelmina, RAÄ 399:1 Vilhelmina och RAÄ 129:1 Åsele) using NIR-spectroscopic (Near InfraRed-spectroscopy), statistical and archaeological methods. By using PCA-models (Principal Component Analysis-models) and the classification method SIMCA (Soft Independent Modelling of Class Analogies) on NIR-spectroscopic data collected over the course of this study, it was possible to show signs that prehistoric man in the studied area, might have deposited quartzite materials, not naturally occurring at the RAÄ 519:1 Vilhelmina, Sweden, site. Four geographic areas (the vicinity of the sea Vojm, the North and South part of the sea Malgomaj and the vicinity of the Southwest part of the Ångerman river in the studied area) could also be shown to display distinct patterns in the PCA-models, related to the use of particular combinations of quartzites. These findings lead to the conclusion that prehistoric man in this area, typically used locally available materials for toolmaking. Furthermore, this study resulted in a large NIR-spectroscopic dataset from the archaeological sites that makes up the main material for this study, that may be beneficial to future NIR-spectroscopic studies in archaeology and/or further studies of NIR-spectroscopy applied to lithic materials.

L’approche géoarchéologique a pour objectif de discuter l’intérêt des archives sédimentaires des rammadiyat pour appréhender les dynamiques d’occupation des sols au Capsien et Néolithique. Les rammadiyat se présentent sous forme de monticules dont la stratigraphie est peu exprimée. Les fouilles récentes ont mis en évidence des sols d’occupation renouvelant les questionnements sur les modes d’occupation de ces populations capsiennes. L’approche géoarchéologique repose sur la micromorphologie des sols. Elle a été menée sur 5 sites capsiens de Tunisie et d’Algérie. Elle a permis de caractériser les modes de formation de ces rammadiyat. Les informations recueillies montrent que les rammadiyat relèvent principalement des processus liés aux activités humaines, qui ont interagi à des degrés variables avec les processus naturels. Elles ont permis de construire un modèle qualitatif de la dynamique de formation des rammadiyat, exprimé par une classification des faciès sédimentaires identifiés, selon les pratiques d’aménagement, selon les activités et selon le degré de fréquentation. Ce modèle met en évidence un premier schéma de l’organisation spatiale des sols d’occupation, matérialisant trois types espaces : un espace construit (bâtiment), un espace de cour segmenté avec des aires domestiques, voies de passage et zones de rejets et un espace peu fréquenté, en marge de l’occupation principale. L’enregistrement sédimentaire témoigne d’un mode d’occupation continu, localement rythmé par des variations dans le degré de fréquentation. Des épisodes d’abandon temporaire ou de très faible fréquentation sont ainsi observés. Cependant, leur généralisation à l’ensemble du site reste à préciser. / The aim of the geoarchaeological approach is to discuss the interest of sedimentary archives of rammadiyat, in order to understand the dynamics of soil occupancy at the Capsian and Neolithic period. The rammadiyat present the shape of a small mound, whose stratigraphy is quite undifferentiated. Recent excavations show the presence of occupation surfaces, renewing the questions about the settlement strategies and of these capsian populations. The geoarchaeological approach is based on soil micromorphology .This research was conducted on 5 capsian sites in Tunisia and Algeria and allowed us to characterize the formation processes of these rammadiyat. The collected data show that rammadiyatis mainly formed by anthropic processes related to human activity interacting with natural processes on varying degrees. It’s then possible to construct a qualitative model relating to the formation dynamics of rammadiyat, expressed by a classification of the identified sedimentary facies, according to the building practices, the activities and the intensity of human occupation. This model highlights a first framework of the spatial organization of occupation surfaces, materializing three types of spaces: a built space (dwelling), a segmented yard space with domestic areas, passageways and dump areas, and a marginal area, showing a sporadic occupation. The sedimentary record shows a continuous occupancy, locally punctuated by variations in the intensity of human occupation. Episodes of temporary abandonment or very low occupation are observed. However, their generalization to the whole site remains to be clarified.

Long-term human-environmental interactions in naturally fragile drylands are a focus of geomorphological and geoarchaeological research. Furthermore, many dryland societies were also affected by seismic activity. The semi-arid Shiraki Plain in the tectonically active southeastern Caucasus is currently covered by steppe and largely devoid of settlements. However, numerous Late Bronze to Early Iron Age city-type settlements suggest early state formation between ca. 3.2-2.5 ka that abruptly ended after that time. A paleolake was postulated for the lowest plain, and nearby pollen records suggest forest clearcutting of the upper altitudes under a more humid climate during the Late Bronze/Early Iron Ages. Furthermore, also an impact of earthquakes on regional Early Iron Age settlements was suggested. However, regional paleoenvironmental changes and paleoseismicity were not systematically studied so far. We combined geomorphological, sedimentological, chronological and paleoecological data with hydrological modelling to reconstruct regional Holocene paleoenvironmental changes, to identify natural and human causes and to study possible seismic events during the Late Bronze/Early Iron Ages. Our results show a balanced to negative Early to Mid-Holocene water balance probably caused by forested upper slopes. Hence, no lake but a pellic Vertisol developed in the lowest plain. Following, Late Bronze/Early Iron Age forest clear-cutting caused lake formation and the deposition of lacustrine sediments derived from soil erosion. Subsequently, regional aridification caused slow lake desiccation. Remains of freshwater fishes indicate that the lake potentially offered valuable ecosystem services for regional prehistoric societies even during the desiccation period. Finally, colluvial coverage of the lake sediments during the last centuries could have been linked with hydrological extremes during the Little Ice Age. Our study demonstrates that the Holocene hydrological balance of the Shiraki Plain was and is situated near a major hydrological threshold, making the landscape very sensitive to small-scale human or natural influences with severe consequences for local societies. Furthermore, seismites in the studied sediments do not indicate an influence of earthquakes on the main and late phases of Late Bronze/Early Iron Age settlement. Altogether, our study underlines the high value of multi-disciplinary approaches to investigate human-environmental interactions and paleoseismicity in drylands on millennial to centennial time scales.

O estudo de oito sambaquis do litoral sul do Estado de Santa Catarina e de um concheiro da Terra do Fogo permitiu entender a dinâmica dos processos de formação de diferentes tipos de concheiros. No caso dos sambaquis catarinenses, o objetivo principal foi compreender os processos de formação culturais(atividades humanas) e naturais (tafonômicos) e suas mudanças ao longo do tempo, a partir da análise estratigráfica de sítios cuja cronologia envolve todo o período de ocupação humana pré-histórica da região (c. 7400-1000 anos AP). No caso fueguino, realizou-se o estudo micro-estratigráfico de um concheiro etno-histórico, para servir como modelo interpretativo na formulação de hipóteses sobre a formação dos concheiros. Os métodos utilizadospara caracterização dos sedimentos arqueológicos incluíram: granulometria, zooarqueologia, isotopia de C e N ( \'? POT.13\'C e \'? POT.15\' N), micromorfologia e microscopia eletrônica de varredura. Utilizou-se também a analogia experimental com fogueiras acesas em diversos contextos conhecidos e queima controlada de moluscos. Três tipos de sambaquis foram analisados: quatro de padrão estratigráfico conchífero; dois de núcleo arenoso; e dois em montículo ictiológico. A formação do primeiro e terceiro tipos seguiu um padrão recorrente e contínuo que envolve retrabalhamento de elementos depositados e queimados em localdiferente do final. Estes elementos incluem resíduos alimentares, como conchas, restos de peixe (ossos e tecidos) e material vegetais (carvões e plantas de ciclo fotossintético \'C IND.3\'), assim como componentes terrígenos relacionados com o substrato sedimentar nos arredores do sambaqui e com sedimentos provenientes dos bancos de moluscos explorados. Os sambaquis de núcleo arenoso representam ocupações efêmeras, mas planejadas. A sua formação envolveu levantamento de montículos de areia e/ou aproveitamento de dunas eólicas, com posterior deposição de moluscos e resíduos de fogueiras. Desta análise, extraem-se duas implicações substanciais: 1) o hábito de retrabalhamento de resíduos pode ter incluído a destruição dos locais de moradia, nunca achados até o momento, associados aos grandes sambaquis; 2) a correspondência entre o processo de formação identificado nos montículos ictiológicos e nos sambaquis conchíferos sugere continuidade na atividade deposicional, apesar da mudança dematéria prima (substituição das conchas por restos de peixe) e da adoção da tecnologia cerâmica em tempos recentes. O concheiro etno-histórico analisado paracomparação geo-etnoarqueológica foi formado a partir de ocupações domésticas, recorrentes durante pelo menos um século. A análise do anel de conchas periférico possibilitou caracterizar micromorfologicamente os episódios de deposição massiva de conchas, pisoteamentoe abandono do sítio. A análise das fogueiras localizadas na área central do concheiro permitiuidentificar estruturas de combustão de temperatura alta e moderada (superior e inferior a 500° C, respectivamente). A comparação das microfácies do concheiro fueguino com as microfácies identificadas nos sambaquis catarinenses mostra diferentes trajetórias pré-deposicionais, relacionadas, no primeiro caso, com a deposição imediata e secundária de detritos de alimentação nos arredores do local de moradia, e, no caso dos sambaquis, com a sucessão intrincada de ações de deposição, queima e transporte, associada à formação de depósitos terciários. Estas observações corroboram a maior complexidade no processo de formação dos sambaquis. / To understand site formation processes in shell sites, eight sambaquis(shell mounds) from the southern coast of Santa Catarina and one shell midden (conchero) from Tierra del Fuego were studied. For the sambaquis of Santa Catarina, the aim was to understand the cultural and natural formation processes (human activities and taphonomy) and the way they changed through time through the whole period of prehistoric human occupation in the region (c. 7400-1000 years BP). In Tierra del Fuego, the micro-stratgraphic study of an ethnohistoric shell midden was done to serve as model for hypothesis on formation processes of shell sites. The methods used for characterization of archaeological sediments included: grain-size analyses, zooarchaeology, C and N isotopy (\'? POT.13\'C e \'? POT.15\' N), micromorphology and scanning electron microscopy. Experimental archaeology was done using different hearths lit on known contexts and by controlled burning of mollusk shell. Three types of sambaquis were analyzed: four shell mounds; two sand mounds; and two fish mounds. Formation of the first and third group followed a recurrent and continuous pattern of reworking of items, accumulated and burnt in a different location than the final. These items included food refuse, like shell, fish remains (bone and tissue) and plant material (charcoal and residues of \'C IND.3\' plants), as well as terrigenous components from the sedimentary substrate of the surroundings of the site and from the exploredshell beds. Sand mounds are ephemeral but planned occupations. Their formation involved rising of a sand mound and/or occupation over sand dunes, with deposition of shells and plant remains on top. Two substantial implications are extracted from this analyses: 1) the reworking of food residues may be destructing the remains of daily occupations associated with the large sambaquis, that have never been found to this moment; 2) the correspondence of formation process in shell mounds and fish mounds suggests continuity in the depositional activities, besides the change in material (substitution of shell by fish remains) and adoption of ceramic technology in recent times. The ethnohistorical site analyzed for geo-ethnoarchaeological comparison was formed by recurrent domestic occupations that lasted for a century. Analyses of the peripheral shell ring allowed the micromorphological characterization of episodes of massive shell deposition, trampling and site abandonment. Analyses of hearths located in the central habitation area showed micromorphological indicators of high temperature and moderate temperature combustion structures (over and under 500° C, respectivley). Comparison of microfacies from the fueguian shell midden and microfacies from the sambaquis shows different pre-depositional trajectories. In the first case, deposition is related with the immediate and secondary discard of food refuse around the living space. In the second case, formation is related with an intricate succession of deposition, burning and transport of items that resulted in a tertiary deposit. These observations corroborate the great complexity in the formation processes of sambaquis.