Els estudis de supervivència s'interessen pel temps que passa des de l'inici de l'estudi (diagnòstic de la malaltia, inici del tractament,...) fins que es produeix l'esdeveniment d'interès (mort, curació, millora,...). No obstant això, moltes vegades aquest esdeveniment s'observa més d'una vegada en un mateix individu durant el període de seguiment (dades de supervivència multivariant). En aquest cas, és necessari utilitzar una metodologia diferent a la utilitzada en l'anàlisi de supervivència estàndard. El principal problema que l'estudi d'aquest tipus de dades comporta és que les observacions poden no ser independents. Fins ara, aquest problema s'ha solucionat de dues maneres diferents en funció de la variable dependent. Si aquesta variable segueix una distribució de la família exponencial s'utilitzen els models lineals generalitzats mixtes (GLMM); i si aquesta variable és el temps, variable amb una distribució de probabilitat no pertanyent a aquesta família, s'utilitza l'anàlisi de supervivència multivariant. El que es pretén en aquesta tesis és unificar aquests dos enfocs, és a dir, utilitzar una variable dependent que sigui el temps amb agrupacions d'individus o d'observacions, a partir d'un GLMM, amb la finalitat d'introduir nous mètodes pel tractament d'aquest tipus de dades. / Survival research is interested in the time that passes from the beginning of the study until the event of interest occurs. However, it is very common to find individuals who experience this event more than once during the period of study. In this case, a different methodology needs to be used to that of the standard univariate survival analysis.In this case, the duration between recurrences could be correlated due to the presence of unobserved individual factors. This type of event is normally dealt with by introducing individual random effects in the model, resulting in a multivariate model. The random effects represent the individual "frailty" and the variance of these effects measures the unobserved heterogeneity between individuals. Until recently, the most common way of dealing with this type of situation in survival analysis was by using marginal models such as the robust covariance matrix estimation in the Andersen-Gill approximation; the Wei, Lin and Weissfeld method or the Prentice, Williams and Peterson method; or using the conditional models such as the frailty models (EM algorhthym). The aim of this study is to model multivariate survival data, based on generalised linear mixed models (GLMM).

Cette thèse cherche à déterminer l’influence de l’environnement abiotique sur la structure des assemblages de poissons dans les estuaires européens tidaux à partir d’une approche macroécologique. L’environnement abiotique de 135 estuaires, du Portugal à l’Ecosse, est caractérisé par une quinzaine de descripteurs en utilisant une approche écohydrologique. Les assemblages de poissons d’une centaine d’estuaires sont caractérisés par les données de pêche acquises au cours de campagnes scientifiques conduites dans le cadre de la Directive-Cadre européenne sur l’Eau (DCE). Néanmoins, ces données sont souvent hétérogènes du fait des différences entre les protocoles d’échantillonnage utilisés. Afin de limiter cette hétérogénéité, une sélection rigoureuse et une procédure de standardisation des données ont été effectuées. Les assemblages de poissons sont décrits à l’aide d’indices globaux ou fonctionnels relatifs à la richesse spécifique et à l’abondance. A l’aide de modèles linéaires généralisés, des relations sont établies entre des attributs de l’ichtyofaune et des gradients abiotiques à large échelle et au sein de l’estuaire. La richesse spécifique totale, et en particulier celle des espèces marines et migratrices amphihalines, augmente avec la taille de l’estuaire. De plus, elle apparaît plus élevée dans les estuaires associés à un large plateau continental. Les plus fortes densités totales et, en particulier, celles des espèces résidentes et marines, sont associées aux estuaires présentant une grande proportion en zones intertidales. Les assemblages de poissons estuariens apparaissent fortement structurés par le gradient de salinité à la fois en termes de richesse spécifique et de densité. En parallèle, cette thèse apporte des éléments témoignant d’un décalage vers le Nord de plusieurs espèces de poissons estuariens dans le contexte du réchauffement climatique global. Les résultats de cette thèse contribueront à l’amélioration des indicateurs biotiques basés sur l’ichtyofaune qui sont actuellement développés dans le contexte de la DCE. / Based on a macroecological approach, this thesis aims at determining the influence of the abiotic environment on the structure of fish assemblages among European tidal estuaries. The abiotic environment of 135 North-Eastern Atlantic estuaries from Portugal to Scotland was characterised by fifteen descriptors using an ecohydrological approach. The fish assemblages of about a hundred estuaries were characterised by fish data collected during scientific surveys conducted in the context of the European Water Framework Directive (WFD). Nonetheless, differences among sampling protocols resulted in highly heterogeneous datasets. To limit this heterogeneity, a rigorous selections and standardisation processes were carried out. Fish assemblages were described by total or functional indices related to species richness or abundance. Relationships were identified between large-scale and intra-estuarine abiotic gradients and fish attributes by fitting generalised linear models. Results showed that the total number of species, and more especially of marine and diadromous species, increased with the estuary size. Moreover, the total species richness appeared higher in estuaries associated to a wide continental shelf. The greatest total densities, and more particularly total densities of resident and marine species, were associated to estuaries with a great proportion of intertidal areas. Fish assemblages appeared also strongly structured by the salinity gradient in terms of both species richness and density. Furthermore, this thesis brought some evidence of northward migration of estuarine fish species in the context of the global warming. The results of this thesis will contribute to improve the fish indicators that are currently developed in the context of the European WFD.

The term aggregation refers to overdispersion and both are used interchangeably in this thesis. In addressing the problem of prevalence of infectious parasite species faced by most rural livestock farmers, we model the distribution of faecal egg counts of 15 parasite species (13 internal parasites and 2 ticks) common in sheep and goats. Aggregation and excess zeroes is addressed through the use of generalised linear models. The abundance of each species was modelled using six different distributions: the Poisson, negative binomial (NB), zero-inflated Poisson (ZIP), zero-inflated negative binomial (ZINB), zero-altered Poisson (ZAP) and zero-altered negative binomial (ZANB) and their fit was later compared. Excess zero models (ZIP, ZINB, ZAP and ZANB) were found to be a better fit compared to standard count models (Poisson and negative binomial) in all 15 cases. We further investigated how distributional assumption a↵ects aggregation and zero inflation. Aggregation and zero inflation (measured by the dispersion parameter k and the zero inflation probability) were found to vary greatly with distributional assumption; this in turn changed the fixed-effects structure. Serial autocorrelation between adjacent observations was later taken into account by fitting observation driven time series models to the data. Simultaneously taking into account autocorrelation, overdispersion and zero inflation proved to be successful as zero inflated autoregressive models performed better than zero inflated models in most cases. Apart from contribution to the knowledge of science, predictability of parasite burden will help farmers with effective disease management interventions. Researchers confronted with the task of analysing count data with excess zeroes can use the findings of this illustrative study as a guideline irrespective of their research discipline. Statistical methods from model selection, quantifying of zero inflation through to accounting for serial autocorrelation are described and illustrated. / Statistics / M.Sc. (Statistics)