Bjoern Reineking - Academia.edu (original) (raw)

Papers by Bjoern Reineking

OBJECTIFS - Le projet AdaMont, soutenu par le programme Gestion des Impacts du Changement Clima-t... more OBJECTIFS - Le projet AdaMont, soutenu par le programme Gestion des Impacts du Changement Clima-tique (GICC) et l’Observatoire National des Effets du Réchauffement Climatique (ONERC), a été réalisé de 2015 à 2017. Conformément aux attendus du programme GICC, le projet s’est intéressé à caractériser et à projeter les impacts du changement climatique pour un territoire, et à proposer une méthodologie d’approche intégrée de l’adaptation au changement climatique à l’échelle de ce territoire, dans une dé-marche participative et pluridisciplinaire.MONTAGNE – Le choix du terrain s’est porté sur les massifs et Parcs Naturels Régionaux (PNR) des Préalpes, des territoires de moyenne montagne très sensibles et déjà impactés par le réchauffement cli-matique. Ces massifs offrent une large palette de milieux et de climats le long de gradients d’altitude, d’exposition, de latitude et de pression urbaine. Ils sont également des terrains privilégiés d’observation et de recherche, en lien avec les PN...

Sciences Eaux & Territoires, 2019

AimProcesses driving current tree species distribution are still largely debated. Attempts to rel... more AimProcesses driving current tree species distribution are still largely debated. Attempts to relate species distribution and population demography metrics have shown mixed results. In this context, we would like to test the hypotheses that the metapopulation processes of colonization and extinction are linked to species distribution models.LocationEurope: Spain, France, Germany, Finland, and Sweden.TaxonAngiosperms and Gymnosperms.MethodsFor the 17 tree species analyzed we fitted species distribution model (SDM) relating environmental variables to presence absence data across Europe. Then using independent data from national forest inventories across Europe we tested whether colonization and extinction probabilities are related to occurrence probability estimated by the SDMs. Finally, we tested how colonization and extinction respectively drive probability of presence at the metapopulation equilibrium.ResultsWe found that for most species at least one process (colonization/extincti...

Ecosphere, 2019

Tree mortality submodels drive simulated long-term forest dynamics: assessing 15 models from the ... more Tree mortality submodels drive simulated long-term forest dynamics: assessing 15 models from the stand to global scale.

International Journal of Remote Sensing, 2014

ABSTRACT Light detection and ranging (lidar) is a useful tool for measuring three-dimensional hab... more ABSTRACT Light detection and ranging (lidar) is a useful tool for measuring three-dimensional habitat structure; hence, its use in habitat suitability models has been explored, both as a single resource and in combination with other remote-sensing techniques. Here, we evaluated the suitability of airborne lidar data in comparison with aerial photographs and field surveys for modelling the distribution of an endangered and cryptic forest species, the hazel grouse (Bonasa bonasia). The study was conducted in the Bavarian Forest National Park of southeast Germany. Subsequently, a prediction map for conservation planning was generated for a large area, which encompassed the National Park. We examined the utility of lidar data for generating a hazel grouse distribution model by using machine learning (boosted regression trees), and then compared the results to variables derived from field surveys and aerial photographs, both separately and in combination. The cross-validated discrimination ability of the model was slightly higher when using lidar data (area under the receiver operator characteristic curve (AUC), 0.79) compared to models using aerial photographs (AUC, 0.75) or field survey data (AUC, 0.78). The predictive performance consistently increased when combining the predictors from different sources, with an AUC of 0.86 being produced in the model combining all three data sources. The three data sources complemented one another, with each data source probably having an advantage at deriving one of three key aspects of the hazel grouse habitat, namely, vertically well-structured forest stands, horizontally mixed successional vegetation stages, and certain deciduous trees as food resources such as mountain ash (Sorbus aucuparia). In addition, the diverse lidar metrics might be applied to simultaneously characterize vertically and horizontally well-structured forest stands. We conclude that public available airborne lidar data are a viable source for creating habitat suitability maps for large areas and may have increased utility for detecting forest characteristics and valuable wildlife habitats.

$Research paper thumbnail of Adding structure to land cover – using fractional cover to study animal habitat use$

Movement Ecology, 2014

Background: Linking animal movements to landscape features is critical to identify factors that s... more Background: Linking animal movements to landscape features is critical to identify factors that shape the spatial behaviour of animals. Habitat selection is led by behavioural decisions and is shaped by the environment, therefore the landscape is crucial for the analysis. Land cover classification based on ground survey and remote sensing data sets are an established approach to define landscapes for habitat selection analysis. We investigate an approach for analysing habitat use using continuous land cover information and spatial metrics. This approach uses a continuous representation of the landscape using percentage cover of a chosen land cover type instead of discrete classes. This approach, fractional cover, captures spatial heterogeneity within classes and is therefore capable to provide a more distinct representation of the landscape. The variation in home range sizes is analysed using fractional cover and spatial metrics in conjunction with mixed effect models on red deer position data in the Bohemian Forest, compared over multiple spatio-temporal scales. Results: We analysed forest fractional cover and a texture metric within each home range showing that variance of fractional cover values and texture explain much of variation in home range sizes. The results show a hump-shaped relationship, leading to smaller home ranges when forest fractional cover is very homogeneous or highly heterogeneous, while intermediate stages lead to larger home ranges. Conclusion: The application of continuous land cover information in conjunction with spatial metrics proved to be valuable for the explanation of home-range sizes of red deer.

Trees, 2013

ABSTRACT Functional convergence in water use of trees across species from diverse geographic loca... more ABSTRACT Functional convergence in water use of trees across species from diverse geographic locations was examined using data on tree water use parameters, with the intention of gaining an understanding on the capacity for water transport for trees with varying structural and functional traits. Wood density (ρw), which is reported to have a negative exponential relation with sap flow density (SFD), showed a bell-shaped curve when the daily SFD data from 101 tree species belonging to 35 angiosperm and gymnosperm families were plotted. The species came from 23 different geographical locations representing all continents. Trees were most efficient in water transport when the ρw was between 0.51 and 0.65 g cm−3. When the ρw increased or decreased from this range, there was a gradual fall in their water transport rate as indicated by lower daily SFD. The unexpected reduction in SFD with decreasing ρw is explained in terms of reduced conductance in the transport pathway, which is a precaution taken by the tree for avoiding cavitation or implosion in larger conducting tubes, which is characteristic of low density wood. The development of severe leaf water potential variations, which is frequently reported in such trees, supports this notion. The SFD versus ρw relation has a potentially wide applicability in predicting water use by forest stands with varying ρw. In addition, the occurrence of a high number of tree species with ρw values in the range of 0.51–0.65 g cm−3 across all continents examined points towards the importance of ρw in the evolutionary process as related to efficient functioning of the water transport mechanism.

Theoretical Ecology, 2009

Understanding the impacts of environmental changes on species survival is a major challenge in ec... more Understanding the impacts of environmental changes on species survival is a major challenge in ecological research, especially when shifting from singleto multispecies foci. Here, we apply a spatially explicit twospecies simulation model to analyze the effects of geographic range shifting and habitat isolation on different coexistence mechanisms. The model explicitly considers dispersal, local competition, and growth on a single resource. Results highlight that both range shifting and habitat isolation severely impact coexistence. However, the strength of these impacts depends on the underlying coexistence mechanisms. Neutrally coexisting species are particularly sensitive to habitat isolation, while stabilized coexistence through overcompensatory density regulation is much more sensitive to range shifting. We conclude that, at the community level, the response to environmental change sensitively depends on the underlying coexistence mechanisms. This suggests that predictions and management recommendations should consider differences between neutral versus stabilized community structures whenever possible.

Oikos, 2010

Ecologists carry a well-stocked toolbox with a great variety of sampling methods, statistical ana... more Ecologists carry a well-stocked toolbox with a great variety of sampling methods, statistical analyses and modelling tools, and new methods are constantly appearing. Evaluation and optimisation of these methods is crucial to guide methodological choices. Simulating error-free data or taking high-quality data to qualify methods is common practice. Here, we emphasise the methodology of the 'virtual ecologist' (VE) approach where simulated data and observer models are used to mimic real species and how they are 'virtually' observed. This virtual data is then subjected to statistical analyses and modelling, and the results are evaluated against the 'true' simulated data. The VE approach is an intuitive and powerful evaluation framework that allows a quality assessment of sampling protocols, analyses and modelling tools. It works under controlled conditions as well as under consideration of confounding factors such as animal movement and biased observer behaviour. In this review, we promote the approach as a rigorous research tool, and demonstrate its capabilities and practical relevance. We explore past uses of VE in different ecological research fields, where it mainly has been used to test and improve sampling regimes as well as for testing and comparing models, for example species distribution models. We discuss its benefits as well as potential limitations, and provide some practical considerations for designing VE studies. Finally, research fields are identified for which the approach could be useful in the future. We conclude that VE could foster the integration of theoretical and empirical work and stimulate work that goes far beyond sampling methods, leading to new questions, theories, and better mechanistic understanding of ecological systems.

Ecology Letters, 2011

Statistical models are the traditional choice to test scientific theories when observations, proc... more Statistical models are the traditional choice to test scientific theories when observations, processes or boundary conditions are subject to stochasticity. Many important systems in ecology and biology, however, are difficult to capture with statistical models. Stochastic simulation models offer an alternative, but they were hitherto associated with a major disadvantage: their likelihood functions can usually not be calculated explicitly, and thus it is difficult to couple them to well-established statistical theory such as maximum likelihood and Bayesian statistics. A number of new methods, among them Approximate Bayesian Computing and Pattern-Oriented Modelling, bypass this limitation. These methods share three main principles: aggregation of simulated and observed data via summary statistics, likelihood approximation based on the summary statistics, and efficient sampling. We discuss principles as well as advantages and caveats of these methods, and demonstrate their potential for integrating stochastic simulation models into a unified framework for statistical modelling.

Basic and Applied Ecology, 2013

Improving our understanding of biodiversity and ecosystem functioning and our capacity to inform ... more Improving our understanding of biodiversity and ecosystem functioning and our capacity to inform ecosystem management requires an integrated framework for functional biodiversity research (FBR). However, adequate integration among empirical approaches (monitoring and experimental) and modelling has rarely been achieved in FBR. We offer an appraisal of the issues involved and chart a course towards enhanced integration. A major element of this path is the joint orientation towards the continuous refinement of a theoretical framework for FBR that links theory testing and generalization with applied research oriented towards the conservation of biodiversity and ecosystem functioning. We further emphasize existing decision-making frameworks as suitable instruments to practically merge these different aims of FBR and bring them into application. This integrated framework requires joint research planning, and should improve communication and stimulate collaboration between modellers and empiricists, thereby overcoming existing reservations and prejudices. The implementation of this integrative

Agriculture, Ecosystems & Environment, 2011

Ecography, 2012

Collinearity refers to the non independence of predictor variables, usually in a regression-type ... more Collinearity refers to the non independence of predictor variables, usually in a regression-type analysis. It is a common feature of any descriptive ecological data set and can be a problem for parameter estimation because it infl ates the variance of regression parameters and hence potentially leads to the wrong identifi cation of relevant predictors in a statistical model. Collinearity is a severe problem when a model is trained on data from one region or time, and predicted to another with a diff erent or unknown structure of collinearity. To demonstrate the reach of the problem of collinearity in ecology, we show how relationships among predictors diff er between biomes, change over spatial scales and through time. Across disciplines, diff erent approaches to addressing collinearity problems have been developed, ranging from clustering of predictors, threshold-based pre-selection, through latent variable methods, to shrinkage and regularisation. Using simulated data with fi ve predictor-response relationships of increasing complexity and eight levels of collinearity we compared ways to address collinearity with standard multiple regression and machine-learning approaches. We assessed the performance of each approach by testing its impact on prediction to new data. In the extreme, we tested whether the methods were able to identify the true underlying relationship in a training dataset with strong collinearity by evaluating its performance on a test dataset without any collinearity. We found that methods specifi cally designed for collinearity, such as latent variable methods and tree based models, did not outperform the traditional GLM and threshold-based pre-selection. Our results highlight the value of GLM in combination with penalised methods (particularly ridge) and threshold-based pre-selection when omitted variables are considered in the fi nal interpretation. However, all approaches tested yielded degraded predictions under change in collinearity structure and the ' folk lore '-thresholds of correlation coeffi cients between predictor variables of |r| Ͼ 0.7 was an appropriate indicator for when collinearity begins to severely distort model estimation and subsequent prediction. Th e use of ecological understanding of the system in pre-analysis variable selection and the choice of the least sensitive statistical approaches reduce the problems of collinearity, but cannot ultimately solve them.

AimLinking local population dynamics and species distributions is critical to predicting the impa... more AimLinking local population dynamics and species distributions is critical to predicting the impacts of climate change. While many studies focus on the mean fitness of populations, theory shows that species distributions can be shaped by demographic stochasticity or population resilience. Here we examine how mean fitness (measured by invasion rate), demographic stochasticity, and resilience (measured by the ability to recover from disturbance) constrain populations at the edges compared to the climatic center.LocationEurope: Spain, France, Germany, Finland, and Sweden.PeriodForest inventory data used for fitting the models cover the period from 1985 to 2013.Major taxaDominant European tree species; Angiosperms and Gymnosperms.MethodsWe developed dynamic population models covering the entire life cycle of 25 European tree species with climatically dependent recruitment models fitted to forest inventory data. We then ran simulations using integral projection and individual-based model...