To obtain a better understanding of the pharmacokinetic and/or pharmacodynamic characteristics of an investigated treatment, clinical data is often analysed with nonlinear mixed effects modelling. The developed models can be used to design future clinical trials or to guide individualised drug treatment. Missing data is a frequently encountered problem in analyses of clinical data, and to not venture the predictability of the developed model, it is of great importance that the method chosen to handle the missing data is adequate for its purpose. The overall aim of this thesis was to develop methods for handling missing data in the context of nonlinear mixed effects models and to compare strategies for handling missing data in order to provide guidance for efficient handling and consequences of inappropriate handling of missing data. In accordance with missing data theory, all missing data can be divided into three categories; missing completely at random (MCAR), missing at random (MAR) and missing not at random (MNAR). When data are MCAR, the underlying missing data mechanism does not depend on any observed or unobserved data; when data are MAR, the underlying missing data mechanism depends on observed data but not on unobserved data; when data are MNAR, the underlying missing data mechanism depends on the unobserved data itself. Strategies and methods for handling missing observation data and missing covariate data were evaluated. These evaluations showed that the most frequently used estimation algorithm in nonlinear mixed effects modelling (first-order conditional estimation), resulted in biased parameter estimates independent on missing data mechanism. However, expectation maximization (EM) algorithms (e.g. importance sampling) resulted in unbiased and precise parameter estimates as long as data were MCAR or MAR. When the observation data are MNAR, a proper method for handling the missing data has to be applied to obtain unbiased and precise parameter estimates, independent on estimation algorithm. The evaluation of different methods for handling missing covariate data showed that a correctly implemented multiple imputations method and full maximum likelihood modelling methods resulted in unbiased and precise parameter estimates when covariate data were MCAR or MAR. When the covariate data were MNAR, the only method resulting in unbiased and precise parameter estimates was a full maximum likelihood modelling method where an extra parameter was estimated, correcting for the unknown missing data mechanism's dependence on the missing data. This thesis presents new insight to the dynamics of missing data in nonlinear mixed effects modelling. Strategies for handling different types of missing data have been developed and compared in order to provide guidance for efficient handling and consequences of inappropriate handling of missing data.

A juçara (Euterpe edulis Mart.) é uma palmeira importante ecologicamente e considerada fonte de produtos valiosos culturalmente e economicamente. Há mais de duas décadas tem sido considerada ameaçada de extinção pelo corte de palmito, e atualmente surge uma nova oportunidade para reduzir sua vulnerabilidade por meio da valorização de seus frutos. Nesta pesquisa investigamos se o manejo de juçara para produção de polpa de frutos pode funcionar como catalisador da conservação da biodiversidade e da melhoria dos meios de vida das populações humanas locais. Algumas questões foram elaboradas e organizadas em quatro eixos convergentes a questão central: (I) Quais os principais aspectos socioeconômicos e práticas de manejo de E. edulis para produção de polpa nas áreas de estudo? Qual o potencial da atividade para o fortalecimento local e conservação da espécie? (II) Como varia a composição e a estrutura florestal ao longo de gradientes de manejo e altitude? Quais as espécies vegetais nativas mais valorizadas por estas comunidades locais? (III) Como a produção e a colheita de frutos variam através de diferentes sistemas de manejo ao longo dos anos? Quais os principais fatores que explicam essa variação? (IV) De que maneira o manejo de frutos de juçara afeta sua dinâmica populacional? Qual a taxa máxima sustentável de remoção de frutos? Quais as estratégias de manejo sustentável recomendada para diferentes condições? Para isso, foi realizado o monitoramento de populações de E. edulis em áreas manejadas por quatro populações humanas que habitam o entorno e interior do Parque Estadual da Serra do Mar (PESM), Núcleos Santa Virgínia e Picinguaba, nos municípios de Natividade da Serra e Ubatuba, SP. Também amostramos populações de E. edulis em áreas de proteção integral no interior do PESM, entrevistamos os moradores de cada comunidade local, assim como usamos dados de fontes secundárias sobre a vegetação do PESM e sobre a atividade de manejo da juçara nos últimos anos. Verificamos que existe uma nova possibilidade de trabalho e renda nas comunidades, inclusive para mulheres e jovens, possibilitando diversificação da economia local. As práticas não destrutivas de colheita dos frutos e as características do sistema reprodutivo da espécie não implicam em severas consequências ecológicas adversas, permitindo considerar o potencial de conservação circa situm, principalmente diante as mudanças climáticas e pressão do mercado de palmito. A diversidade arbórea reduziu localmente com o aumento da intensidade de manejo agroflorestal, mas permite a manutenção de espécies arbóreas de interesse para a conservação, além de proporcionar sustento às famílias dos produtores. As práticas de manejo realizadas nos sistemas agroflorestais tradicionais podem favorecer a manutenção do conhecimento ecológico local, além de aumenta a produtividade de frutos e reduzir a variação interanual da frutificação. A colheita de frutos de juçara em si não representa prejuízos em termos populacionais. Sendo o novo manejo de juçara uma das estratégias recomendadas para manter a viabilidade das populações em longo prazo, desde que o corte de palmito seja controlado e a regeneração seja promovida. / Juçara (Euterpe edulis Mart.) is a palm source of various culturally and economically valued products. It has been considered threatened of extinction for more than two decades. The simple opposition between strategies to obtain its different products arises a new opportunity to reduce its vulnerability. In this research we investigated if the management for fruit production could act as a catalyst to biodiversity conservation and local livelihood improvement. Some questions were elaborated and organized into four axes convergent to the central question: (I) How do the management systems of E. edulis for fruit production are characterized at the study area? Which is the activity potential to improve local livelihoods? (II) How does the forest structure and composition vary along management and altitude gradients? What native plant species are most valued by these communities? (III) How do fruit production and harvest vary through different management systems along different years? What are the main factors explaining such variation? (IV) How do juçara management for fruit production affect its population dynamics? What is the maximum sustainable rate of fruit removal? Which sustainable management strategies are recommended for different conditions? We built empirical evidences through monitoring E. edulis populations managed by four local human populations living around and inside Parque Estadual da Serra do Mar (PESM), Nuclei Santa Virgínia and Picinguaba. We also sampled E. edulis populations in strictly protected areas inside PESM, as well as used secondary data about juçara management activities in the last years. The opportunity for work and income, inclusive for women and young people, the stimulus for social organization, production diversification, among other aspects recorded in this study promote community resilience in the face of political and economic change. Juçara fruit harvest allows frugivorous foraging and indicates the potential for in situ and circa situm conservation approach, mainly in the face of climate change and market pressure for palm heart. Arboreal diversity reduced locally with increasing agroforesty management intensity, which in turn raised diversity at the landscape level. Management practices, performed at the traditional agroforestry systems, could strengthen local cognitive systems, enhance fruit productivity and reduce fruiting inter annual variation. Juçara fruit harvest benefits recruitment and could even induce population increase at the long term. Neverheless, other factors affected negatively affected the population negatively. The protection of adult palms for fruit harvest is one of the recommended strategies to keep population viability in the long term and could function as in situ and circa situm conservation strategies.

Pour limiter l'impact des activités humaines sur la biodiversité, les populations animales sont gérées dans de nombreux contextes. La gestion des populations animales comprend la conservation des espèces menacées, comme l’exploitation/régulation des espèces gibiers ou considérées nuisibles. Elle consiste à modifier la dynamique des populations soit en favorisant leur croissance, soit en réduisant leur abondance. Face à l’urgence d’agir dans un contexte de ressources financières limitées, une gestion efficiente des populations animales requiert une bonne compréhension de leur dynamique en réponse aux actions mises en œuvre. Durant mon doctorat, j’ai travaillé en partenariat avec des gestionnaires agissant sur deux systèmes biologiques : la conservation de rapaces méditerranéens (aigles de Bonelli et vautours percnoptères) et la régulation des densités de renards en paysage rural. Pour chaque système, mon travail a consisté i) en l’analyse des suivis de populations qui sont réalisés pour ii) évaluer l’effet de la gestion sur la dynamique des espèces gérées. Dans les deux contextes, j’ai pu iii) mettre en évidence la contribution de l’immigration à la croissance des populations menacées, comme à l’atténuation des effets de la régulation. Après iv) avoir extrait des recommandations pouvant améliorer l’efficience de la gestion en tenant compte de ces processus d’immigration, mon travail s’est conclu par v) une réflexion en retour sur l’optimisation des suivis de populations, afin qu’ils génèrent le plus d’informations pour un investissement moindre. Dans l’ensemble, ce travail aboutit à une réflexion sur les moyens favorisant une gestion efficiente des populations animales. / Due to the multiple interactions linking species together, human activities and animal species influence each other. Animal populations are therefore managed to favour long-term cohabitation. Wildlife management includes the conservation of endangered species, the harvest of game species and the control of species considered as pest. It consists in impacting population dynamics (density variation in a territory) either by favouring or limiting population growth. Faced with the complexity of ecological processes and the urgent need for acting in a context of decreasing allocated resources, an efficient management requires a precious understanding of population dynamics in response to actions. During my PhD, I collaborated with managers supervising two contrasted biological systems: the conservation of endangered Mediterranean raptors (Bonelli’s eagles and Egyptian vultures) and the control of fox densities in French rural landscapes. For each case of study, my work consisted in i) analysing data from population monitoring designed to ii) evaluate the management impact on population dynamics. In both systems, I highlighted iii) the crucial importance of immigration either in boosting endangered population or compensating for fox regulation. After having iv) derived concrete guidelines to improve management facing with immigration. I concluded my PhD by v) searching in turn for cost-effective designs of population monitoring. Overall, I questioned the contrasted systems I studied to understand pitfalls and solutions favouring an efficient management of animal populations.

No description available.

There is anecdotal evidence that populations of the New Zealand endemic red admiral butterfly Bassaris gonerilla (F.) have declined since the early 1900s. This decline has been associated with the introduction of the generalist pupal parasitoids Pteromalus puparum (L.) and Echthromorpha intricatoria (F.). The former was deliberately introduced for the biological control of the cabbage white butterfly (Pieris rapae (L.)); the latter is an adventitious arrival from Australia. The objective of this thesis was to quantify, using population models, the effect that P. puparum is having on B. gonerilla abundance. Population monitoring and a phenology model (based on temperature-related development rates) indicated that B. gonerilla has two full generations and one partial generation per summer in the Banks Peninsula region of New Zealand. B. gonerilla abundance was greatly reduced in drought summers, which was probably due to the negative effects of drought on the quality and quantity of the larval host plant Urtica ferox Forst. A life table study showed that egg parasitism by the unidentified scelionid Telenomus sp. was the largest mortality factor for the pre-imaginal stages of B. gonerilla, followed by "disappearance" mortality (predation and dispersal) in the larval stages. Pupal mortality due to P. puparum was lower compared with that caused by E. intricatoria, with 1-19% and 20-30% of pupae being parasitised by P. puparum and E. intricatoria, respectively. Collection of B. gonerilla pupae from the Christchurch, Dunedin and Wellington areas confirmed higher rates of percentage parasitism by E. intricatoria. B. gonerilla collected from the Banks Peninsula had a 50: 50 sex ratio and lifetime fecundity was estimated in the laboratory as 312 eggs per female. There was no evidence of density-dependent parasitism of B. gonerilla pupae by P. puparum in the field, although there was a significant positive relationship between life table estimates of E. intricatoria parasitism and B. gonerilla pupal abundance. Larval dispersal from the host plant showed a positive relationship with larval instar but no relationship with larval density. Rates of change in B. gonerilla adult abundance between generations within a year showed evidence of density dependence, and this negative feedback was stronger in a drought year. A discrete-time model for B. gonerilla population dynamics was constructed which had two summer generations per year and a partial overwintering generation. The model showed that the presence of this overwintering generation provides a temporal refuge from high levels of E. intricatoria parasitism. Removal of parasitoid mortality from the model suggested that P. puparum was suppressing B. Gonerilla populations on the Banks Peninsula by 5% and E. intricatoria by 30%. An important assumption of the model was that parasitism rates were independent of B. gonerilla density. This assumption appears valid for P. puparum parasitism, but may not be valid for E. intricatoria; therefore the estimated suppression levels due to this adventive parasitoid should be viewed with some caution. It is too soon to generalise on what determines the magnitude of non-target effects by arthropod biocontrol agents, this being only the second study to quantify effects at a population level. However, in this case retrospective analysis has shown that the impact of non-target parasitism by P. puparum on B. gonerilla abundance has been small. There is anecdotal evidence that populations of the New Zealand endemic red admiral butterfly Bassaris gonerilla (F.) have declined since the early 1900s. This decline has been associated with the introduction of the generalist pupal parasitoids Pteromalus puparum (L.) and Echthromorpha intricatoria (F.). The former was deliberately introduced for the biological control of the cabbage white butterfly (Pieris rapae (L.)); the latter is an adventitious arrival from Australia. The objective of this thesis was to quantify, using population models, the effect that P. puparum is having on B. gonerilla abundance. Population monitoring and a phenology model (based on temperature-related development rates) indicated that B. gonerilla has two full generations and one partial generation per summer in the Banks Peninsula region of New Zealand. B. gonerilla abundance was greatly reduced in drought summers, which was probably due to the negative effects of drought on the quality and quantity of the larval host plant Urtica ferox Forst.. A life table study showed that egg parasitism by the unidentified scelionid Telenomus sp. was the largest mortality factor for the pre-imaginal stages of B. gonerilla, followed by "disappearance" mortality (predation and dispersal) in the larval stages. Pupal mortality due to P. puparum was lower compared with that caused by E. intricatoria, with 1-19% and 20-30% of pupae being parasitised by P. puparum and E. intricatoria, respectively. Collection of B. gonerilla pupae from the Christchurch, Dunedin and Wellington areas confirmed higher rates of percentage parasitism by E. intricatoria. B. gonerilla collected from the Banks Peninsula had a 50: 50 sex ratio and lifetime fecundity was estimated in the laboratory as 312 eggs per female. There was no evidence of density-dependent parasitism of B. gonerilla pupae by P. puparum in the field, although there was a significant positive relationship between life table estimates of E. intricatoria parasitism and B. gonerilla pupal abundance. Larval dispersal from the host plant showed a positive relationship with larval instar but no relationship with larval density. Rates of change in B. gonerilla adult abundance between generations within a year showed evidence of density dependence, and this negative feedback was stronger in a drought year. A discrete-time model for B. gonerilla population dynamics was constructed which had two summer generations per year and a partial overwintering generation. The model showed that the presence of this overwintering generation provides a temporal refuge from high levels of E. intricatoria parasitism. Removal of parasitoid mortality from the model suggested that P. puparum was suppressing B. Gonerilla populations on the Banks Peninsula by 5% and E. intricatoria by 30%. An important assumption of the model was that parasitism rates were independent of B. gonerilla density. This assumption appears valid for P. puparum parasitism, but may not be valid for E. intricatoria; therefore the estimated suppression levels due to this adventive parasitoid should be viewed with some caution. It is too soon to generalise on what determines the magnitude of non-target effects by arthropod biocontrol agents, this being only the second study to quantify effects at a population level. However, in this case retrospective analysis has shown that the impact of non-target parasitism by P. puparum on B. gonerilla abundance has been small.

Increased temperatures due to anthropogenic-induced climate change may raise the threat of extinction for taxa with sessile life histories (e.g., plants) in the near future. Linking climate change models to demographic models may provide useful insights into the potential effects of environmental changes on rare plants, and therefore aid in their current and future conservation. Population demographers generally agree that mechanistic models from a reductionist perspective are necessary to test assumptions in population drivers. For the first study, I assessed the climate vulnerability of a rare plant species, Pyrrocoma radiata, with a mechanistic model of four climatically-similar populations. I used environmentally-driven demographic models to estimate vital rates and population sizes from a nonlinear, nonparametric regression with local climate variables. I assessed the utility of this environmentally-correlated, stage-structured population matrix model compared to a stationary model of independent and identically-distributed environmental stochasticity. I then simulated future population projections based on climate conditions predicted by General Circulation Models (GCMs) under opposing emission scenarios. The second study hopes to answer population-level questions using a traditionally community-level method, non-metric multidimensional scaling, which considers correlation structure between response variables and can be used to find environmental correlates of the ordination axes. Demographic data on a threatened perennial, Astragalus tyghensis, were collected from five sites in the Tygh Valley, OR. I considered correlation structure between demographic vital rates to find environmental correlates of the ordination axes. The search for an environmental driver of population vital rates was successful for the two study species. Previous year dry dormant season precipitation likely affects the fertility rates a year later in P. radiata populations, and dry growing season reference evapotranspiration rates positively correlated with a growth gradient in A. tyghensis. Based on predicted precipitation, P. radiata is expected to rapidly decline by 2050, but this may be due to biases in the two GCMs and reliance on only one environmental factor. The NMS ordination adequately captured most of the variation in transition elements for the years and populations from A. tyghensis demographics. I provided support to the claim that model predictions can improve with the inclusion of mechanistic relationships. The inclusion of abiotic drivers in models used to predict population trends is supported by our study and may enhance predictive power in population viability assessments under changing climates. / Graduation date: 2013