It appears to be unanimous that alien species in island environments tend to cause considerably more negative than positive impacts. To assess the potential level of threat aliens may pose to the native environment, understanding a species’ population structure and dynamics is of ultimate importance. Assessing both impacts and consequences of management interventions to alien species is likewise only possible through the comprehension of its population structure and dynamics. This can be achieved by estimating the number of individuals in the study site, as well as other population parameters through time, applying population models such as capture-recapture to the collected datasets. Nonetheless, alien species that have low capture rates, such as small mammals, might present a considerable obstacle for conservation, as available capturerecapture models need a relatively large dataset to precisely and accurately estimate population parameters. To improve accuracy and precision of estimates that use sparse datasets, the present study developed an integrated concurrent marking-observation capture-recapture model (C-MOM). The model proposed here, contrary to the commonly available mark-recapture and mark-resight models, allows for two different datasets (i.e. a capture-recapture and a population count) to be integrated, as well as for marking and observation (recapture) data to be collected simultaneously. While few models can integrate different datasets, no model is known to allow for concomitant capture-markobservation activities. To assess the performance of the C-MOM when estimating population parameters for sparse datasets, a virtual ecology study was carried out. The population dynamics of a small rodent, the rock cavy (Kerodon rupestris), as well as capture-recapture and population count datasets, were simulated under different scenarios. The sampled datasets were then analyzed by the C-MOM, and by two other established statistical models: a classical mark recapture (CMR) (based on the Jolly-Seber model), and a zero-truncated Poisson log-normal mixed effects (ZPNE), the only integrated mark-resight model that allows for recapture sampling with replacement. Estimates of population parameters provided by the three models were then compared in terms of bias, precision and accuracy. C-MOM and ZPNE models were afterwards applied to real data collected on a rock cavy colony in the island of Fernando de Noronha. The estimated parameters were used to extrapolate the number of individuals in the rock cavy colony to the whole population in the island. Subsequently, these results were used to develop a risk assessment for the species by modelling historical and management scenarios, simulating both the establishment of the species in the island, and the consequences of different management interventions applied to it. The virtual ecology study showed that, in comparison to the CMR and the ZPNE, the C-MOM presented improved accuracy without overestimating the precision of population parameter’s estimates. The last also presented reduced amplitude of the calculated credible interval at 95% when applied to real data in comparison to the ZPNE. While the extrapolation of C-MOM estimates suggests that the rock cavy population in Fernando de Noronha is 6,652 ± 1,587, ZPNE estimates are of 5,854 ± 3,269 individuals. In the risk assessment, historical simulation models demonstrated that even though different combinations of uncertainty in reproductive parameters of the rock cavy might be possible for the species, these did not interfere significantly in either establishment or spread of the rock cavy population in the island. Moreover, historical yearly mortality has most likely been under 30%. Regarding the species’ management simulations, the most effective management interventions to achieve population extinction were spaying and neutering of both sexes, although harvest effort presented the highest influence on this populations’ extirpation. Nonetheless, the relative influence of female and both sexes’ based interventions did not differ significantly regarding the frequency of extinction of stochastic replicates’. Moreover, none of the management interventions guaranteed the population extinction within the time span and harvest effort proposed for the management program. Neutering of both sexes was most inversely influential on time to extinction of this population, followed by removal of both sexes. Briefly, the C-MOM has proven to be a resourceful and precise model to estimate population parameters when low capture rates result in sparse datasets. Moreover, the rock cavy is well established in the island and likely at carrying capacity. In general, the risk assessment showed that the management interventions in the time span and harvest effort simulated in the present study were ineffective to extinguish the rock cavy population in Fernando de Noronha. Considering this, as well as the importance of investigating other vital factors to decide in favour of or contrary to the management of this species, it is recommended that both an impact assessment of the rock cavy and a cost-effectiveness analysis of the management interventions should be performed to complement the current study.:Acknowledgement III Abstract IV Zusammenfassung VI Resumen IX Table of Contents XII List of Tables and Figures XIV List of Abbreviations XIX 1. Introduction 1 1.1. Invasive alien species and their consequences 1 1.2. Population dynamics analysis 2 Capture-recapture models 3 Observation models 4 Integrated population models 5 Software 7 Model analysis 8 1.3. Fernando de Noronha and the rock cavy 10 1.4. Objectives 12 Overall Objectives 12 Specific Objectives 13 2. Study Framework 15 3. Methods 19 3.1. Study area 19 3.2. Study case species 21 3.3. Research Steps 24 RESEARCH STEP I: Comparing the C-MOM to established models – does this concurrent marking-observation model produces accurate estimates of population parameters for sparse datasets? 24 RESEARCH STEP II: C-MOM application to a real case study 40 RESEARCH STEP III: The rock cavy population in Fernando de Noronha 45 RESEARCH STEP IV: The colonization and eradication of the rock cavy in Fernando de Noronha 47 4. Results 63 4.1. RESEARCH STEP I: Comparing the C-MOM to established models – does this concurrent marking-observation model produces accurate estimates of population parameters for sparse datasets? 63 4.2. RESEARCH STEP II: C-MOM application to a real case study 72 4.3. RESEARCH STEP III: The rock cavy population in Fernando de Noronha 73 4.4. RESEARCH STEP IV: The colonization and eradication of the rock cavy in Fernando de Noronha 74 Sensitivity analysis 74 Simulation experiments 80 5. Discussion 83 5.1. Bias, precision and accuracy of population dynamic models for sparse datasets 85 Simulated data 85 Study case 90 5.2. Advantages and disadvantages of the C-MOM approach 93 5.3. Development and applications of the integrated models and the C-MOM 96 5.4. The reversed use of the PVA software Vortex to simulate AS and IAS populations’ extinction 97 5.5. Status of the rock cavy population in the island of Fernando de Noronha 100 The colonization of the rock cavy in Fernando de Noronha 101 Management of the rock cavy in Fernando de Noronha 104 Study case limitations and future researches 112 6. Conclusion 116 References 118 Appendices 124 APPENDIX I – Assessment of biological invasions 124 APPENDIX II – Population dynamics simulation and dataset sampling 125 APPENDIX III – CMR and C-MOM model codes in R 134 APPENDIX IV – ZPNE model code in R 138 APPENDIX V – C-MOM model used for real datasets 143 APPENDIX VI – Rock cavy colony sizes and number of individuals in Fernando de Noronha 145 APPENDIX VII – Parameter’s ranking of C-MOM, CMR and ZPNE models 148 APPENDIX VIII – Bias, precision and accuracy table 149

Diese Dissertation beschäftigt sich mit regionaler Klimamodellierung über Südamerika, der Analyse von Modellsensitivitäten bezüglich Wolkenparametrisierungen und der Entwicklung neuer Methoden zur Modellevaluierung mithilfe von Klimanetzwerken. Im ersten Teil untersuchen wir Simulationen mit dem COnsortium for Small scale MOdeling model in CLimate Mode (COSMO-CLM) und stellen die erste umfassende Evaluierung dieses dynamischen regionalen Klimamodells über Südamerika vor. Dabei untersuchen wir insbesondere die Abhängigkeit simulierter tropischer Niederschläge von Parametrisierungen subgitterskaliger cumuliformer und stratiformer Wolken und finden starke Sensitivitäten bezüglich beider Wolkenparametrisierungen über Land. Durch einen simultanen Austausch der entsprechenden Schemata gelingt uns eine beträchtliche Reduzierung von Fehlern in klimatologischen Niederschlags- und Strahlungsmitteln, die das COSMO-CLM über tropischen Regionen für lange Zeit charakterisierten. Im zweiten Teil führen wir neue Metriken für die Evaluierung von Klimamodellen bezüglich räumlicher Kovariabilitäten ein. Im Kern bestehen diese Metriken aus Unähnlichkeitsmaßen für den Vergleich von simulierten mit beobachteten Klimanetzwerken. Wir entwickeln lokale und globale Unähnlichkeitsmaße zum Zwecke der Darstellung lokaler Unähnlichkeiten in Form von Fehlerkarten sowie der Rangordnung von Modellen durch Zusammenfassung lokaler zu globalen Unähnlichkeiten. Die neuen Maße werden dann für eine vergleichende Evaluierung regionaler Klimasimulationen mit COSMO-CLM und dem Statistical Analogue Resampling Scheme über Südamerika verwendet. Dabei vergleichen wir die sich ergebenden Modellrangfolgen mit solchen basierend auf mittleren quadratischen Abweichungen klimatologischer Mittelwerte und Varianzen und untersuchen die Abhängigkeit dieser Rangfolgen von der betrachteten Jahreszeit, Variable, dem verwendeten Referenzdatensatz und Klimanetzwerktyp. / This dissertation is about regional climate modeling over South America, the analysis of model sensitivities to cloud parameterizations, and the development of novel model evaluation techniques based on climate networks. In the first part we examine simulations with the COnsortium for Small scale MOdeling weather prediction model in CLimate Mode (COSMO-CLM) and provide the first thorough evaluation of this dynamical regional climate model over South America. We focus our analysis on the sensitivity of simulated tropical precipitation to the parameterizations of subgrid-scale cumuliform and stratiform clouds. It is shown that COSMO-CLM is strongly sensitive to both cloud parameterizations over tropical land. Using nondefault cumulus and stratus parameterization schemes we are able to considerably reduce long-standing precipitation and radiation biases that have plagued COSMO-CLM across tropical domains. In the second part we introduce new performance metrics for climate model evaluation with respect to spatial covariabilities. In essence, these metrics consist of dissimilarity measures for climate networks constructed from simulations and observations. We develop both local and global dissimilarity measures to facilitate the depiction of local dissimilarities in the form of bias maps as well as the aggregation of those local to global dissimilarities for the purposes of climate model intercomparison and ranking. The new measures are then applied for a comparative evaluation of regional climate simulations with COSMO-CLM and the STatistical Analogue Resampling Scheme (STARS) over South America. We compare model rankings obtained with our new performance metrics to those obtained with conventional root-mean-square errors of climatological mean values and variances, and analyze how these rankings depend on season, variable, reference data set, and climate network type.