Search

Climate change threatens biodiversity and species distribution all over the world at unprecedented rates. Human induced changes to landscape structure and habitat are redefining the relation between species and their environment. Understanding, characterizing and modeling the relation between complex landscape features and geographical species presence, the subject of this thesis, have thus gained paramount importance in order to make informed decisions on near- and far-term management strategies for species and landscape protection. The present thesis pays particular attention to the understanding of how mountain landscapes, described as heterogeneous habitat matrices, govern the presence of a species in space and time. To this end, throughout this thesis a metapopulation model is considered, a spatially-explicit model based on the balance between colonization and extinction events in areas of different suitability. The classical patch-based modelling approach is extended here in various ways in order to incorporate landscape-explicit information about the habitat matrix. A first theoretical study is performed on geometric, realistic, and real landscapes which highlights how different levels of connectivity, the degree to which different areas of similar habitat are actually linked together in the landscape (say, derived from geomorphic structures, such as valleys and peaks or emerging structures shaped by fluvial erosion and tectonic uplift) impact the dynamic of mountain species under climate change. A second, applied, study investigates how different landscape descriptors derived from Earth Observation data can be used in order to dynamically model the spatial, landscape-explicit, presence of two carabid species (Pterostichus flavofemoratus and Carabus depressus) in the Gran Paradiso National Park. In a final study, the consistency of the metapopulation model is investigated when considering different resolutions of the landscape matrix, i.e. different levels of coarse-graining. This last part is fundamental to the understanding of how conclusions drawn from local studies can be made general and extrapolated to larger regions. Overall, this thesis bridges the fields of population and landscape ecology, focusing on modelling metapopulation dynamics in complex landscapes seen as the substrate for ecological interactions. The result from the three studies show that metapopulation ecology can profit from the insights of landscape ecology. The combination of the two fields can be a useful tool in furthering the understanding and monitoring of mountain species.

This work aims to forecast rain locally in Tambarga, Burkina Faso, to be able to fight against a worm inducing the disease called schistosomiasis. The chosen approach relies on a machine-leaning technique called Artificial Neural Networks, which simulates the synapses of a brain, with climatic parameters as inputs, activation functions and outputs in the form of rain prediction. A special case of Neural Networks using Bayesian Computations is used, along with as a transform allowing to capture the changes in climatic conditions, called Wavelet Transform. The precipitation is forecasted in different manners: binary forecast on the presence or absence of rain, linear forecast on the daily and weekly intensity, as well as a rain-class forecast. The most successful predictions have been found to be the binary forecast, as well as the weekly windowed cumulative rain forecast. The daily cumulative rain, as well has the classes forecast have not produced satisfying results, mainly because of the high temporal variability of the observations, as well as the very unequal distribution of observations in the different rain classes. In the end, it has been shown that it is possible to use Bayesian Networks to forecast precipitation in some extent, and that the wavelet transform of the inputs has a positive impact on the accuracy of the prediction.

The large number of potential applications for robotic odor source localization has motivated the development of a variety of plume tracking algorithms, the majority of which work in restricted two-dimensional scenarios. In this paper, we introduce a distributed algorithm for 3-D plume tracking using a system of ground and aerial robots in formation. We propose an algorithm that takes advantage of spatially distributed measurements to track the plume in 3-D and lead the robots to the source by integrating three behaviors -- upwind movement, plume centering, and Laplacian feedback formation control. We evaluate this strategy in simulation and with real robots in a wind tunnel. For a source close to the ground, results show that a team of robots running our algorithm reaches the source with low lateral error while also tracing the horizontal and vertical plume shape.

The Species-Area Relation (SAR), which describes the increase in the number of species S with increasing area A, is under intense scrutiny in contemporary ecology, in particular to probe its reliability in predicting the number of species going extinct as a direct result of habitat loss. Here, we focus on the island SAR, which is measured across a set of disjoint habitat patches, and we argue that the SAR portrays an average trend around which fluctuations are to be expected due to the stochasticity of community dynamics within the patches, external perturbations, and habitat heterogeneity across different patches. This probabilistic interpretation of the SAR, though already implicit in the theory of island biogeography and manifest in the scatter of data points in plots of empirical SAR curves, has not been investigated systematically from the theoretical point of view. Here, we show that the two main contributions to SAR fluctuations, which are due to community dynamics within the patches and to habitat heterogeneity between different patches, can be decoupled and analyzed independently. To investigate the community dynamics contribution to SAR fluctuations, we explore a suite of theoretical models of community dynamics where the number of species S inhabiting a patch emerges from diverse ecological and evolutionary processes, and we compare stationary predictions for the coefficient of variation of S, i.e. the fluctuations of S with respect to the mean. We find that different community dynamics models diverge radically in their predictions. In island biogeography and in neutral frameworks, where fluctuations are only driven by the stochasticity of diversification and extinction events, relative fluctuations decay when the mean increases. Computational evidence suggests that this result is robust in the presence of competition for space or resources. When species compete for finite resources, and mass is introduced as a trait determining species' birth, death and resource consumption rates based on empirical allometric scalings, relative fluctuations do not decay with increasing mean S due to the occasional introduction of new species with large resource demands causing mass extinctions in the community. Given this observation, we also investigate the contribution of external disturbance events to fluctuations of S in neutral community dynamics models and compare this scenario with the community dynamics in undisturbed non-neutral models. Habitat heterogeneity within a single patch, in the context of metapopulation models, causes variability in the number of coexisting species which proves negligible with respect to that caused by the stochasticity of the community dynamics. The second contribution to SAR fluctuations, which is due to habitat heterogeneity among different patches, introduces corrections to the coefficient of variation of S. Most importantly, inter-patches heterogeneity introduces a constant, lower bound on the relative fluctuations of S equal to the coefficient of variation of a habitat variable describing the heterogeneity among patches. Because heterogeneity across patches is inevitably present in natural ecosystems, we expect that the relative fluctuations of S always tend to a constant in the limit of large mean S or large patch area A, with contributions from community dynamics, inter-patches heterogeneity or both. We provide a theoretical framework for modelling these two contributions and we show that both can affect significantly the fluctuations of the SAR. (C) 2018 The Authors. Published by Elsevier Ltd.