A técnica de mineração de regras de associação surgiu com o objetivo de encontrar conhecimento novo, útil e previamente desconhecido em bancos de dados transacionais, e uma grande quantidade de algoritmos de mineração de regras de associação tem sido proposta na última década. O maior e mais bem conhecido problema destes algoritmos é a geração de grandes quantidades de conjuntos freqüentes e regras de associação. Em bancos de dados geográficos o problema de mineração de regras de associação espacial aumenta significativamente. Além da grande quantidade de regras e padrões gerados a maioria são associações do domínio geográfico, e são bem conhecidas, normalmente explicitamente representadas no esquema do banco de dados. A maioria dos algoritmos de mineração de regras de associação não garantem a eliminação de dependências geográficas conhecidas a priori. O resultado é que as mesmas associações representadas nos esquemas do banco de dados são extraídas pelos algoritmos de mineração de regras de associação e apresentadas ao usuário. O problema de mineração de regras de associação espacial pode ser dividido em três etapas principais: extração dos relacionamentos espaciais, geração dos conjuntos freqüentes e geração das regras de associação. A primeira etapa é a mais custosa tanto em tempo de processamento quanto pelo esforço requerido do usuário. A segunda e terceira etapas têm sido consideradas o maior problema na mineração de regras de associação em bancos de dados transacionais e tem sido abordadas como dois problemas diferentes: “frequent pattern mining” e “association rule mining”. Dependências geográficas bem conhecidas aparecem nas três etapas do processo. Tendo como objetivo a eliminação dessas dependências na mineração de regras de associação espacial essa tese apresenta um framework com três novos métodos para mineração de regras de associação utilizando restrições semânticas como conhecimento a priori. O primeiro método reduz os dados de entrada do algoritmo, e dependências geográficas são eliminadas parcialmente sem que haja perda de informação. O segundo método elimina combinações de pares de objetos geográficos com dependências durante a geração dos conjuntos freqüentes. O terceiro método é uma nova abordagem para gerar conjuntos freqüentes não redundantes e sem dependências, gerando conjuntos freqüentes máximos. Esse método reduz consideravelmente o número final de conjuntos freqüentes, e como conseqüência, reduz o número de regras de associação espacial. / The association rule mining technique emerged with the objective to find novel, useful, and previously unknown associations from transactional databases, and a large amount of association rule mining algorithms have been proposed in the last decade. Their main drawback, which is a well known problem, is the generation of large amounts of frequent patterns and association rules. In geographic databases the problem of mining spatial association rules increases significantly. Besides the large amount of generated patterns and rules, many patterns are well known geographic domain associations, normally explicitly represented in geographic database schemas. The majority of existing algorithms do not warrant the elimination of all well known geographic dependences. The result is that the same associations represented in geographic database schemas are extracted by spatial association rule mining algorithms and presented to the user. The problem of mining spatial association rules from geographic databases requires at least three main steps: compute spatial relationships, generate frequent patterns, and extract association rules. The first step is the most effort demanding and time consuming task in the rule mining process, but has received little attention in the literature. The second and third steps have been considered the main problem in transactional association rule mining and have been addressed as two different problems: frequent pattern mining and association rule mining. Well known geographic dependences which generate well known patterns may appear in the three main steps of the spatial association rule mining process. Aiming to eliminate well known dependences and generate more interesting patterns, this thesis presents a framework with three main methods for mining frequent geographic patterns using knowledge constraints. Semantic knowledge is used to avoid the generation of patterns that are previously known as non-interesting. The first method reduces the input problem, and all well known dependences that can be eliminated without loosing information are removed in data preprocessing. The second method eliminates combinations of pairs of geographic objects with dependences, during the frequent set generation. A third method presents a new approach to generate non-redundant frequent sets, the maximal generalized frequent sets without dependences. This method reduces the number of frequent patterns very significantly, and by consequence, the number of association rules.

Les Infrastructures de Données Géographiques (IDG) ont été mises en place depuis une vingtaine d’années afin notamment de contribuer aux politiques publiques environnementales en facilitant la mise à disposition et le partage des informations géographiques. Au vu des dynamiques en cours relatives à la fois à la production et à la circulation de l’information géographique environnementale, le questionnement à l’origine de cette thèse s’inscrit dans le cadre général d’une meilleure compréhension des pratiques de gestion de l’environnement. Il concerne l’étude de la contribution des IDG à la Gestion Intégrée des Zones Côtières (GIZC) française, gouvernance territoriale spécifique et complexe pour laquelle le recueil et le partage, notamment via les IDG, d’informations géographiques ont été identifiés depuis les années 2000 comme des composantes majeures pour fonder la politique de la mer et du littoral. Néanmoins, la contribution effective des IDG dépend des réponses qu’elles apportent à leurs usagers. Du point de vue méthodologique, ce travail de recherche met en oeuvre une stratégie exploratoire se focalisant sur les usages et les besoins des acteurs côtiers ainsi que sur la nature complexe et dynamique des IDG : - Une enquête nationale basée sur un questionnaire en ligne et exploitée par une triple analyse : statistique, par réseaux sociaux (Social Network Analysis) et thématique ; - Une étude de cas infranationale, reposant sur des entretiens semi-directifs auprès de personnes-ressources et analysée thématiquement puis par diagrammes de flux de données (DFD). Ces techniques de recueil et d’analyse ont été choisies et mises en oeuvre de manière intégrée et complémentaire pour permettre une compréhension, une explication et une généralisation du système IDG dans le contexte de la GIZC. Pour chacune des phases de la recherche, trois types de résultats sont dégagés : le profil des usagers, les IDG jouant un rôle dans la gestion des littoraux français, et leur contribution à la gestion de ces territoires. Il ressort de cette série de synthèses, d’abstractions et de modélisations, une contribution à la connaissance des phénomènes d’appropriation des IDG et des flux informationnels associés. / Spatial Data Infrastructures (SDIs) have been implemented over the last twenty years in particular in order to contribute to Environmental Policy by improving geographic information access and sharing processes. Given the current dynamics of both production and circulation of environmental geographic information, the general aim underlying this research is part of the general framework for a better understanding of environmental management practices. This research focuses specifically on French SDIs and their contribution to Integrated Coastal Zone Management (ICZM), a complex form of territorial governance for which the collection and sharing, particularly via SDIs, of geographic information have been identified, since the 2000s, as a major component on which to found coastal and marine public policies. However, the real contribution of SDIs depends on the answers they provide to their users. From a methodological perspective, this PhD is implemented by an exploratory research approach focusing on the usages and needs of coastal stakeholders as well as on the complex and dynamic nature of SDIs. The research approach is divided into two phases: - A national online survey examined by a triple analysis: Statistical analysis, Social Network Analysis (SNA) and thematic analysis;- A sub-national case study based on semi-structured interviews with key informants and analyzed thematically and by Data Flow Diagrams (DFD). These data collection and analysis techniques have been selected and implemented in an integrated and complementary approach to enable understanding, explanation and generalization of the "SDI system" in the context of ICZM. For each phase of the research, three types of results emerged: the profile of users, the type of SDIs playing a role in the management of French coasts, and their contribution to the management of these areas. The above-described research has resulted in a contribution to the knowledge of SDI appropriation and the associated information flows.

A técnica de mineração de regras de associação surgiu com o objetivo de encontrar conhecimento novo, útil e previamente desconhecido em bancos de dados transacionais, e uma grande quantidade de algoritmos de mineração de regras de associação tem sido proposta na última década. O maior e mais bem conhecido problema destes algoritmos é a geração de grandes quantidades de conjuntos freqüentes e regras de associação. Em bancos de dados geográficos o problema de mineração de regras de associação espacial aumenta significativamente. Além da grande quantidade de regras e padrões gerados a maioria são associações do domínio geográfico, e são bem conhecidas, normalmente explicitamente representadas no esquema do banco de dados. A maioria dos algoritmos de mineração de regras de associação não garantem a eliminação de dependências geográficas conhecidas a priori. O resultado é que as mesmas associações representadas nos esquemas do banco de dados são extraídas pelos algoritmos de mineração de regras de associação e apresentadas ao usuário. O problema de mineração de regras de associação espacial pode ser dividido em três etapas principais: extração dos relacionamentos espaciais, geração dos conjuntos freqüentes e geração das regras de associação. A primeira etapa é a mais custosa tanto em tempo de processamento quanto pelo esforço requerido do usuário. A segunda e terceira etapas têm sido consideradas o maior problema na mineração de regras de associação em bancos de dados transacionais e tem sido abordadas como dois problemas diferentes: “frequent pattern mining” e “association rule mining”. Dependências geográficas bem conhecidas aparecem nas três etapas do processo. Tendo como objetivo a eliminação dessas dependências na mineração de regras de associação espacial essa tese apresenta um framework com três novos métodos para mineração de regras de associação utilizando restrições semânticas como conhecimento a priori. O primeiro método reduz os dados de entrada do algoritmo, e dependências geográficas são eliminadas parcialmente sem que haja perda de informação. O segundo método elimina combinações de pares de objetos geográficos com dependências durante a geração dos conjuntos freqüentes. O terceiro método é uma nova abordagem para gerar conjuntos freqüentes não redundantes e sem dependências, gerando conjuntos freqüentes máximos. Esse método reduz consideravelmente o número final de conjuntos freqüentes, e como conseqüência, reduz o número de regras de associação espacial. / The association rule mining technique emerged with the objective to find novel, useful, and previously unknown associations from transactional databases, and a large amount of association rule mining algorithms have been proposed in the last decade. Their main drawback, which is a well known problem, is the generation of large amounts of frequent patterns and association rules. In geographic databases the problem of mining spatial association rules increases significantly. Besides the large amount of generated patterns and rules, many patterns are well known geographic domain associations, normally explicitly represented in geographic database schemas. The majority of existing algorithms do not warrant the elimination of all well known geographic dependences. The result is that the same associations represented in geographic database schemas are extracted by spatial association rule mining algorithms and presented to the user. The problem of mining spatial association rules from geographic databases requires at least three main steps: compute spatial relationships, generate frequent patterns, and extract association rules. The first step is the most effort demanding and time consuming task in the rule mining process, but has received little attention in the literature. The second and third steps have been considered the main problem in transactional association rule mining and have been addressed as two different problems: frequent pattern mining and association rule mining. Well known geographic dependences which generate well known patterns may appear in the three main steps of the spatial association rule mining process. Aiming to eliminate well known dependences and generate more interesting patterns, this thesis presents a framework with three main methods for mining frequent geographic patterns using knowledge constraints. Semantic knowledge is used to avoid the generation of patterns that are previously known as non-interesting. The first method reduces the input problem, and all well known dependences that can be eliminated without loosing information are removed in data preprocessing. The second method eliminates combinations of pairs of geographic objects with dependences, during the frequent set generation. A third method presents a new approach to generate non-redundant frequent sets, the maximal generalized frequent sets without dependences. This method reduces the number of frequent patterns very significantly, and by consequence, the number of association rules.

Le fonctionnement hydrologique des petits bassins versants ruraux de quelques km² à régime intermittent est complexe car de nombreux processus affectent le cheminement des eaux de surface. Il en résulte une variabilité de la densité de drainage et de la dynamique de l'écoulement au sein du réseau hydrographique. Cette dynamique de fonctionnement est peu prise en considération du fait des difficultés d'ordres techniques et économiques pour suffisamment observer les mouvements de l'eau à la surface. Or, c'est une information essentielle pour décrire la connectivité hydrologique du bassin qui représente la distribution spatiale et temporelle des surfaces contributives à l'écoulement par leur connexion au réseau de drainage. De telles connaissances seraient utiles pour traiter la question de la séparation entre production et transfert effectuée dans les modèles hydrologiques, avec la perspective de proposer des simulations plus justes physiquement. L'objectif de la thèse est de proposer des approches spatiales pour mieux intégrer la dynamique de fonctionnement du réseau de drainage en lien étroit avec la réponse hydrologique du bassin. Le Mercier (7 km²) est le site expérimental situé en tête du bassin versant de l'Yzeron (142 km²) localisé à l'ouest de l'agglomération lyonnaise. Ce bassin sur socle cristallin est composé essentiellement de surfaces agricoles et de forêts. Son fonctionnement hydrologique est affecté par l'existence de zones humides contributives. Un réseau de routes et de fossés anthropiques s'ajoute aux talwegs naturels. Les méthodes développées relèvent de deux approches : (1) la microtopographie issue d'un MNT LiDAR (Light Detection and Ranging) permet d'identifier et décrire des extensions fonctionnelles du réseau de drainage d'une part au niveau des linéaires artificiels avec un apport minimal de données externes, et d'autre part au niveau des talwegs naturels en distinguant la présence ou l'absence d'un chenal de drainage, signe d'un potentiel d'écoulement concentré. (2) Un dispositif de 18 capteurs limnimétriques est mis en œuvre pour assurer un suivi permanent de la réponse hydrologique par emboitement de stations au sein du réseau hydrographique. Cette observation « multi-locale » permet de mesurer l'évolution de la densité de drainage, d'identifier localement la hiérarchie des facteurs qui affectent la réponse et de distinguer différentes dynamiques de transfert dans le réseau de drainage. Les résultats des approches par la microtopographie et par observations multi-locales aident à identifier des régions du bassin au comportement différent. Ils permettent notamment de mieux comprendre les interrelations entre occupation du sol et processus hydrologiques, voire géomorphologie et processus. Ces résultats valident donc l'intérêt du capteur LiDAR et celui d'un dispositif in situ souple et adaptable pour proposer un « pattern de drainage » réaliste en limitant le recours au terrain. Enfin, ce pattern décrivant la tendance d'organisation spatiale des écoulements, est paramétré dans une fonction de transfert géomorphologique calculée sur la base des cheminements fournis par un MNT. L'usage de cet outil constitue une ébauche mais conforte l'idée d'un couplage entre le pattern de drainage et la réponse du bassin versant dans des conditions d'intensité pluvieuse soutenue et d'humidité modérée pour expliquer la réponse rapide du bassin. L'ensemble des résultats justifie donc la mise en avant de la nature transitoire du réseau de drainage pour paramétrer des modèles spatialisés avec la perspective d'améliorer leurs capacités prédictives. / Hydrological functioning of small temporary catchments depends on several processes governing transfer from surface water paths. As a result flow dynamics and drainage density are highly variable in space and time. But this complex dynamic is not enough taken into account because of technical and economical limitations. However, it is essential to describe hydrological connectivity as a spatial and temporal pattern of contributive areas to the drainage network. Get this pattern may facilitate the conceptual distinction between production and transfer functions to improve spatially distributed models. The aim of the study is to develop methods to describe spatial and temporal patterns of the drainage network in relation to catchment hydrological responses. The Mercier headwater catchment (7 km²) is located into the Yzeron catchment near Lyon (France). The land use is principally composed of agricultural plots and forested areas. The hydrographical network consists of natural thalwegs and many roadside ditches and agrarian ditches. Both approaches are developed for this purpose: first, the micro-topography from a LiDAR DEM helps to describe potential flow lengths from drainage network extensions during rainfall events. On the one hand, main artificial ditches are mapped from the DEM with minimal corrections from ancillary data. On the other hand, channelized or unchannelized reaches are located from the DEM into the natural thalwegs. Second, a water level sensor device is set up to record hydrological response from 18 stations located in nested sub-catchments into the hydrological network. These synoptic measurements are used to estimate temporal changes in drainage density, to analyze local hydrological functioning, or to describe flood propagation to the outlet. Results from both approaches lead to the identification of specific behaviors inside the hydrological network controlled by functional thresholds. These patterns help to better understand the relationship between land use and hydrological processes. The results also show the interest of LiDAR DEM and the suitability of adaptable distributed measurements as a substitute to heavy in situ studies for the identification of drainage patterns. Finally, to test the assumption of a gradual drainage network extension during a rainfall event, a simplified drainage pattern is processed into a geomorphological transfer function. This tool is fairly easy to set and is used to initiate an association between local hydrological knowledge and global catchment response. Both approaches lead to conclude that various drainage pattern must be integrated into spatially distributed models according to hydrological conditions, rather than a single hydrological network.

A técnica de mineração de regras de associação surgiu com o objetivo de encontrar conhecimento novo, útil e previamente desconhecido em bancos de dados transacionais, e uma grande quantidade de algoritmos de mineração de regras de associação tem sido proposta na última década. O maior e mais bem conhecido problema destes algoritmos é a geração de grandes quantidades de conjuntos freqüentes e regras de associação. Em bancos de dados geográficos o problema de mineração de regras de associação espacial aumenta significativamente. Além da grande quantidade de regras e padrões gerados a maioria são associações do domínio geográfico, e são bem conhecidas, normalmente explicitamente representadas no esquema do banco de dados. A maioria dos algoritmos de mineração de regras de associação não garantem a eliminação de dependências geográficas conhecidas a priori. O resultado é que as mesmas associações representadas nos esquemas do banco de dados são extraídas pelos algoritmos de mineração de regras de associação e apresentadas ao usuário. O problema de mineração de regras de associação espacial pode ser dividido em três etapas principais: extração dos relacionamentos espaciais, geração dos conjuntos freqüentes e geração das regras de associação. A primeira etapa é a mais custosa tanto em tempo de processamento quanto pelo esforço requerido do usuário. A segunda e terceira etapas têm sido consideradas o maior problema na mineração de regras de associação em bancos de dados transacionais e tem sido abordadas como dois problemas diferentes: “frequent pattern mining” e “association rule mining”. Dependências geográficas bem conhecidas aparecem nas três etapas do processo. Tendo como objetivo a eliminação dessas dependências na mineração de regras de associação espacial essa tese apresenta um framework com três novos métodos para mineração de regras de associação utilizando restrições semânticas como conhecimento a priori. O primeiro método reduz os dados de entrada do algoritmo, e dependências geográficas são eliminadas parcialmente sem que haja perda de informação. O segundo método elimina combinações de pares de objetos geográficos com dependências durante a geração dos conjuntos freqüentes. O terceiro método é uma nova abordagem para gerar conjuntos freqüentes não redundantes e sem dependências, gerando conjuntos freqüentes máximos. Esse método reduz consideravelmente o número final de conjuntos freqüentes, e como conseqüência, reduz o número de regras de associação espacial. / The association rule mining technique emerged with the objective to find novel, useful, and previously unknown associations from transactional databases, and a large amount of association rule mining algorithms have been proposed in the last decade. Their main drawback, which is a well known problem, is the generation of large amounts of frequent patterns and association rules. In geographic databases the problem of mining spatial association rules increases significantly. Besides the large amount of generated patterns and rules, many patterns are well known geographic domain associations, normally explicitly represented in geographic database schemas. The majority of existing algorithms do not warrant the elimination of all well known geographic dependences. The result is that the same associations represented in geographic database schemas are extracted by spatial association rule mining algorithms and presented to the user. The problem of mining spatial association rules from geographic databases requires at least three main steps: compute spatial relationships, generate frequent patterns, and extract association rules. The first step is the most effort demanding and time consuming task in the rule mining process, but has received little attention in the literature. The second and third steps have been considered the main problem in transactional association rule mining and have been addressed as two different problems: frequent pattern mining and association rule mining. Well known geographic dependences which generate well known patterns may appear in the three main steps of the spatial association rule mining process. Aiming to eliminate well known dependences and generate more interesting patterns, this thesis presents a framework with three main methods for mining frequent geographic patterns using knowledge constraints. Semantic knowledge is used to avoid the generation of patterns that are previously known as non-interesting. The first method reduces the input problem, and all well known dependences that can be eliminated without loosing information are removed in data preprocessing. The second method eliminates combinations of pairs of geographic objects with dependences, during the frequent set generation. A third method presents a new approach to generate non-redundant frequent sets, the maximal generalized frequent sets without dependences. This method reduces the number of frequent patterns very significantly, and by consequence, the number of association rules.

Les champs continus sont des types de représentations spatiales utilisées pour modéliser des phénomènes tels que la température, la pollution ou l’altitude. Ils sont définis selon une fonction de mapping f qui affecte une valeur du phénomène étudié à chaque localisation p du domaine d’étude. Par ailleurs, la représentation des champs continus à différentes échelles ou résolutions est souvent essentielle pour une analyse spatiale efficace. L’avantage des champs continus réside dans le niveau de détails généré par la continuité, ainsi que la qualité de l’analyse spatiale fournie par la multi-résolution. L’inconvénient de ce type de représentations dans l’analyse spatio-multidimensionnelle est le coût des performances d’analyse et de stockage. Par ailleurs, les entrepôts de données spatiaux et les systèmes OLAP spatiaux (EDS et SOLAP) sont des systèmes d’aide à la décision qui permettent l’analyse spatio-multidimensionnelle de grands volumes de données spatiales et non spatiales. L’analyse des champs continus dans l’architecture SOLAP représente un défi de recherche intéressant. Différents travaux se sont intéressés à l’intégration de ce type de représentations dans le système SOLAP. Cependant, celle-ci est toujours au stade embryonnaire. Cette thèse s’intéresse à l’intégration des champs continus incomplets représentés par une grille régulière de points dans l’analyse spatio-multidimensionnelle. Cette intégration dans le système SOLAP implique que l’analyse des champs continus doit supporter : (i) les opérateurs OLAP classiques, (ii) la vue continue des données spatiales, (iii) les opérateurs spatiaux (slice spatial) et (iv) l’interrogation des données à différentes résolutions prédéfinies. Dans cette thèse nous proposons différentes approches pour l’analyse des champs continus dans le SOLAP à différents niveaux de l’architecture relationnelle, de la modélisation conceptuelle à l’optimisation des performances de calcul. Nous proposons un modèle logique FISS qui permet d’optimiser les performances d’analyse à multi-résolution en se basant sur des méthodes d’interpolation. Puis, nous exposons une méthodologie basée sur la méthode d’échantillonnage du Clustering, qui permet d’optimiser les opérations d’agrégation des grilles régulières de points dans l’architecture SOLAP relationnelle en effectuant une estimation des résultats. / Continuous fields are types of spatial representations used to model phenomena such as temperature, pollution or altitude. They are defined according to a mapping function f that assigns a value of the studied phenomenon to each p location of the studied area. Moreover, the representation of continuous fields at different scales or resolutions is often essential for effective spatial analysis. The advantage of continuous fields is the level of details generated by the continuity of the spatial data, and the quality of the spatial analysis provided by the multi-resolution. The downside of this type of spatial representations in the multidimensionnal analysis is the high cost of analysis and storage performances. Moreover, spatial data warehouses and spatial OLAP systems (EDS and SOLAP) are decision support systems that enable multidimensional spatial analysis of large volumes of spatial and non-spatial data. The analysis of continuous fields in SOLAP architecture represents an interesting research challenge. Various studies have focused on the integration of such representations in SOLAP system. However, this integration still at an early stage. Thus, this thesis focuses on the integration of incomplete continuous fields represented by a regular grid of points in the spatio-multidimensional analysis. This integration in the SOLAP system involves that the analysis of continuous fields must support:(i) conventional OLAP operators, (ii) Continuous spatial data, (iii) spatial operators (spatial slice), and (iv) querying data at different predefined levels of resolutions. In this thesis we propose differents approaches for the analysis of continuous fields in SOLAP system at different levels of the relational architecture (from the conceptual modeling to the optimization of computing performance). We propose a logical model FISS to optimize the performances of the multi-resolution analysis, based on interpolation methods. Then, we present a new methodology based on the Clustering sampling method, to optimize aggregation operations on regular grids of points in the relational SOLAP architecture.

Le palmier babaçu (Attalea speciosa Mart. ex Spreng.) est une plante native de la forêt amazonienne. La déforestation a pour impact qu'il apparaît dans des milieux ouverts désormais anthropisés (pâturages, et champs cultivés). Le babaçu fait partie des ressources " extractivistes " du Brésil : activité de cueillette suivie de commercialisation de produits non ligneux. Cette activité concerne des personnes parmi les plus démunies du pays d'où l'importance de gérer au mieux sa durabilité dans un contexte conflictuel. Force est de constater que les connaissances du fonctionnement durable de l'espèce au sein de ces milieux anthropisés font cruellement défaut: son cycle de vie est peu connu et sa dynamique de population non étudiée.Le travail de thèse est donc pionnier et s'inscrit au sein d'un projet interdisciplinaire Open Science de la fondation Agropolis. L'objectif est double : il consiste à produire un modèle de la dynamique de population du babaçu validé par une analyse in situ et à capitaliser les connaissances issues de divers milieux scientifiques (biologistes, écologues, mathématiciens, informaticiens et économistes français et brésiliens). Le modèle, à plus long terme, doit étayer les recommandations relatives à la gestion durable de l'espèce.L'approche retenue a consisté à expliciter au sein de divers modèles, les connaissances partagées sur l'espèce et les modèles mathématiques potentiels ainsi que sur le contexte sociétal. Ceci nous a permis de proposer un modèle d'observation conforme aux préconisations d'OBOE (Ontologie relative a la sémantique des observations scientifiques) permettant la mise en place du protocole d'acquisition des données. L'acquisition in situ a été réalisée par le partenariat franco-brésilien entre 2013 et 2016 au sein du territoire de la communauté de Benfica (Pará, Brésil). La pérennisation des données s'est faite après conception, au sein d'une base de données spatiales prenant en compte l'aspect socio-économique.Enfin, nous proposons un modèle matriciel aléatoire ayant pour entrée des variables agrégées au niveau des stades biologiques de l'espèce. Les probabilités de la matrice de transition entre les stades sont modélisées selon un modèle hiérarchique Dirichlet-Multinomial. L'intégration des informations a priori a été formalisée grâce à une approche bayésienne. L'estimation et la validation du modèle ont été effectuées avec des critères bayésiens.Des simulations basées sur les paramètres estimés constituent une première ébauche d'étude du comportement de l'espèce. / The babassu palm tree (Attalea speciosa Mart. ex Spreng.) is an endemic species of the amazonian forests. The deforestation highlights henceforth this palm tree in the anthropogenic open areas (pastures and cultivated fields). The babassu is one of the "extractive" resources in Brazil: gathering activity followed by marketing of non-timber products. This activity involves people among the most disadvantaged in the country thus the importance to better manage its sustainability in a context of conflict. It is clear that knowledge of sustainable functioning of the species within these manmade environments is sorely lacking: its life cycle is not well known and its population dynamics unstudied.This PhD work is pioneer and is a part of an Open Science interdisciplinary project of the Agropolis foundation. The objective is twofold it is to generate a model of population dynamics of the babassu palm tree validated by in situ analysis and capitalizing knowledge from various scientific communities (biologists, ecologists, mathematicians, computer scientists and economists from France and Brazil). In the long-term the model should support the recommendations for the sustainable management of the species.Our approach explicit within different models the shared knowledge about the species and the potential mathematical models as well as the societal context. This has allowed us to provide an observation model compliant with the recommendations ofOBOE (Ontology on a semantic scientific observations) for the development of data acquisition protocol. The acquisition was made in situ by the French-Brazilian partnership between 2013 and 2016 in the community of Benfica (Pará, Brazil). The perpetuation of the data was made after conception, within a spatial database taking into account the socio-economic aspect.Finally, we proposed a random matrix model having as input aggregated variables based on the biological stages of the species. The probabilities of the between stages transition matrix are modeled using a hierarchical Dirichlet-Multinomial model. The integration of prior information was formulated through a Bayesian approach. The estimation and model validation were performed with Bayesian criteria.Simulations based on the estimated parameters represent a first outline of the study of the species behavior.

This thesis describes and analyzes the current state of environmental management of geographic data within geoinformation infrastructure of the Czech Republic in response to European legislation. The work is divided into four main parts. The first part deals with geoinformation infrastructure of the Czech Republic in its basic components due to the efficiency of the management of geographic data. Another part deals with the general characteristics of geospatial data underlying the description methods, organization and retrieval in the next section. The final part of its qualitative research describes in detail the management of geospatial data in the environmental sector, which is affected by the phenomenon of the implementation of INSPIRE and draws conclusions and recommendations. The work also supports the analysis of the availability of data sources for the implementation of the INSPIRE directive, the questionnaire used as a guide for research purposes, to illustrate the services Geoportal three pictures. Powered by TCPDF (www.tcpdf.org)

Nowadays, geoprocessing has moved in large parts from isolated desktop usage into the Web. Thereby, the overall availability of geoprocessing functionality has theoretically improved. Nevertheless, web-based geoprocessing functionality is still not readily available and usable as means to find and subsequently compare functionality are yet missing. Discoverability and exchangeability of geoprocessing functionality are limited, and the fundamental benefits of online usage are not fully exploited. To close this gap, this thesis defines a conceptualization with geooperators representing well-defined geoprocessing functionality, and categories representing distinct geooperator attributes as a starting point. Geooperators and categories are connected by associative and hierarchical links forming an interlinked network. The conceptualization serves as basis for two purposes: On the one hand, discovery of geooperators has to be improved. There is a multitude of different GIS users from varying backgrounds, all having a slightly different view on geoprocessing functionality. To embrace these varying views, several actually existing and established categorizations for geooperators are integrated into the conceptualization and structured hierarchically. Thus, multiple different perspectives on geooperators are enabled. The perspectives and the underlying categorizations are integrated into a geooperator browser serving as client that users can employ for discovery. The geooperator browser offers a faceted browsing interface based on the derived perspectives and categories. Several different search modes are offered and can be used simultaneously, thereby facilitating an improved discovery. On the other hand, to establish comparability and subsequently semantically interoperable exchangeability of geooperators, respective geooperator attributes are defined. Thereby, backend and provider independence of geoprocessing services is achieved. The conceptualization is formalized to allow for machine readability and processing which is required for usage in Spatial Data Infrastructures and the Semantic Web. The formalized conceptualization is labeled geooperator thesaurus. It is encoded by Semantic Web standards and offered in a Semantic Web compliant manner. The integration into geoprocessing service metadata is enabled by injecting semantic annotations that link to the respective concepts in the geooperator thesaurus. The thesaurus serves as the underlying data model for the geooperator browser. The thesis concludes with an outlook and discussions of future work. Foremost, the content of the geooperator thesaurus needs to be extended as it currently only comprises a representative subset of geooperators. An approach is suggested to involve the geoprocessing community as an important source of geoprocessing expertise in the maintenance and further development of the thesaurus. On a conceptual level, the thesaurus needs to be extended to a fully-fledged ontology including a formal geoprocessing algebra to support the creation of geooperator workflows. From a discovery point of view, the concept of geoprocessing patterns is outlined. Geoprocessing patterns list geooperators that are commonly used for certain geoprocessing tasks, and provide best practices about applying them in a meaningful manner and sequence. / An Stelle von isolierter Desktop-basierter Nutzung von Geoprozessierung werden heutzutage mehr und mehr webbasierte Angebote zur Verfügung gestellt. Dadurch hat sich die generelle Verfügbarkeit von Geoprozessierungsfunktionalität theoretisch verbessert. Da jedoch das Auffinden und anschließend oft notwendige Vergleichen von Funktionalität nur eingeschränkt möglich ist, können die Vorteile einer webbasierten Verfügbarkeit nicht vollständig genutzt werden. Um dieser Problematik zu begegnen, wird im Rahmen der vorliegenden Arbeit eine Konzeptualisierung entwickelt, die wohldefinierte Geoprozessierungsfunktionalität als Geooperatoren und Kategorien als Unterscheidungsmerkmale für Geooperatoren zur Verfügung stellt. Geooperatoren und Kategorien sind über assoziative und hierarchische Links zu einem Netzwerk verknüpft. Diese Konzeptualisierung erfüllt zwei grundlegende Aufgaben: Einerseits muss das Auffinden von Geooperatoren für Nutzer mit unterschiedlichsten Anwendungshintergründen möglich sein. Diese Nutzer haben unterschiedlichste Sichten auf Geooperatoren, die berücksichtigt und unterstützt werden sollen. Dazu werden etablierte Kategorisierungen aus Literatur und Praxis in die Konzeptualisierung integriert, entsprechend hierarchisch strukturiert und als Perspektiven auf Geooperatoren für den Zugriff nach Außen zur Verfügung gestellt. Diese Perspektiven und die darunterliegenden Kategorien werden als Facetten in einen Geooperatorbrowser integriert, der als webbasierter Client von den Nutzern für das Auffinden von Geooperatoren verwendet werden kann. Die explorative Suche über Facetten nach Geoprozessierungsfunktionalität wird durch eine Schlüsselwortsuche und einen geführten Suchmodus ergänzt und damit insgesamt die Auffindbarkeit von Geooperatoren verbessert. Andererseits sollen die Vergleichbarkeit und der anschließend semantisch interoperable Austausch von Geooperatoren ermöglicht werden. Dazu werden durch einen Vergleich von Geooperatorattributen Ähnlichkeiten von Geooperatoren definiert und zusammen mit der Konzeptualisierung als Geooperatorthesaurus formalisiert. Durch die Formalisierung wird eine Maschinenlesbarkeit und -prozessierbarkeit erreicht. Nur so kann ein anbieter- und backendunabhängiger Austausch von Geoprozessierungsdiensten in Geodateninfrastrukturen ermöglicht werden. Der Thesaurus nutzt Semantic-Web-Standards und wird in einer Semantic-Web-kompatiblen Art und Weise im Web publiziert. Die Integration in Metadaten von Geoprozessierungsdiensten wird durch semantische Annotationen erreicht, die auf die entsprechenden Konzepte im Thesaurus verlinken. Der Thesaurus dient als Datenbasis für den Geooperatorenbrowser. Die Arbeit schließt mit einem Ausblick auf und einer Diskussion von zukünftigen Forschungsarbeiten ab. Da der Thesaurus zurzeit nur eine repräsentative Menge an Beispieldaten enthält, muss als erster wichtiger Schritt der Inhalt des Thesaurus erweitert werden. Für die mittel- und langfristige Instandhaltung und Weiterentwicklung des Thesaurus wird ein Konzept zur Einbindung der GIS-Community vorgeschlagen, da hier gebündelte Geoprozessierungsexpertise vorhanden ist. Auf einer konzeptuellen Ebene muss der Thesaurus für eine semantisch und technisch verbesserte Verkettung von Geooperatoren zu einer vollständigen Ontologie mit einer formalen Geoprozessierungsalgebra weiterentwickelt werden. Für eine weitere Verbesserung der Auffindbarkeit von Geooperatoren wird das Konzept von Geoprozessierungspatterns skizziert, die Geooperatoren zusammenfassen, die oft im Kontext einer bestimmten Aufgabe zusammen genutzt werden. Zusätzlich werden über die Patterns Best Practices zur sinnvollen Anwendung und Verkettung der enthaltenen Geooperatoren zur Verfügung gestellt.

With sharp climate goals to be climate neutral by 2045, Sweden needs to rapidly change the present energy system. The Swedish government has assigned Gotland as a pilot area for the energy transition, due to its geographical constrains as an island, and because of the major structural changes planned for Gotland’s energy supply and industrial establishment. These changes create the opportunity to build a more flexible, efficient, and robust energy system. To support the transition, IVL Swedish Environmental Research institute has initiated the GOT Heat project. This thesis contributes to the GOT Heat project by developing a GIS model that represent the heat sector of Gotland. The model has been developed in parallel with a TIMES model. The purpose of the GIS model is to be used as a heat atlas and to support the TIMES model of Gotland’s energy system by spatial visualization and knowledge. The GIS model was successfully developed and utilized to enhance the representation of Gotland’s energy system within the TIMES model. Data from Boverket, Lantmäteriet, Energimyndigheten and companies has been collected, processed for the development of the GIS model. The incorporation of spatial knowledge and visualization of the heat sector through the GIS model is expected to improve the overall outcomes of the TIMES model. The heat atlas also proved to be a valuable support tool, enabling the visualization of excess heat potentials within the heat sector. This visualization provides insights into potential opportunities for industrial and sector coupling, which would lead to more efficient utilization of excess heat.