Reconstructing large scale pattern of biodiversity using mechanistic simulation models ¶

Department of Ecology
Federal University of Goiás
Goiânia, GO, Brazil

Fellowship Period:  06.2019-10.2019

Research interests and main activities

I’m a macroecologist with general interest in the processes that create and maintain organic and cultural diversity. The spatial patterns of diversity at large scales, such as the latitudinal gradient in species richness, are among the most conspicuous natural phenomena, yet they are not fully understood. These patterns emerge from the interaction among agents (e.g. species or populations) and the dynamic environment, forming a highly complex ecological system. In addition, the deep historical origins of these patterns, and contingent events during their evolution, prevent the use of standard scientific experimentation to identify causal processes.

My contribution to the study of diversity patterns at broad scales is to design mechanistic simulation models to evaluate proposed hypotheses for the mechanisms that generate and maintain these patterns, such as niche conservatism, habitat heterogeneity and habitat connectivity. In a virtual realm it is possible to perform in silico evolutionary experiments to evaluate these hypotheses, as the contrast between simulated and real-world patterns of diversity indicates the relative contribution of natural processes behind the hypotheses.

Working at WSL could not be more exciting and enjoyable! Here I’m collaborating closely with Dr. Löic Pellissier and his team, at the Land Change Science and Landscape Ecology unit, designing and evaluating cutting-edge simulation model for the origin of biodiversity patterns at the global scale. I’m also collaborating with Dr. Catherine Graham and her team, at the Biodiversity and Conservation Biology unit, applying novel statistical methods to study regional effects of environmental factors on patterns of species richness of tetrapods.

Project Description
Ecological and evolutionary processes that drive patterns of biodiversity, such as species richness and phylogenetic diversity, are emergent properties of processes operating at lower biological level. A true mechanistic understanding of how these processes originated and maintain biodiversity over time requires the integration of knowledge across multiple disciplines, such as macroevolution, paleo-biology, biogeography, community ecology. My main research interest is to combine and evaluate existing causal hypotheses of biodiversity phenomena, across multiple branches of knowledge, using a single, integrative analytical framework.

Despite substantial theoretical achievements in the field of macroecology during the last three decades, the macroecological research agenda on the drivers of biodiversity has now stalled. The statistical description of the biodiversity patterns of large-bodied organisms has reached a global extent, and it is unlikely that novel insights about the drivers of biodiversity patterns will emerge from either mapping the geographic distribution of additional species, or from employing the traditional correlative analyses to describe the relationship between biodiversity and environmental factors. Indeed, macroecology is now shifting focus to employing realistic process-based models to causally predict and explain observed patterns of biodiversity from first ecological and evolutionary principles (Leprieur et al. 2016, Rangel et al. 2018).

For over fifteen years I have been employing computer simulations to robustly evaluate causal mechanisms of biodiversity patterns (Rangel and Diniz-Filho 2003, Rangel et al. 2007, Rangel et al. 2018). With the rise of computing power and more accurate reconstructions of paleo-environments, both the development and the use of spatial mechanistic models of diversification to study the emergence of biodiversity gradients has rapidly increased in ecology and evolution. Spatial models of diversification are process-based implementation of causal mechanisms, and they are used to infer the dynamics of species over a dynamic landscape. Those models may include mechanisms such as allopatric speciation, dispersal and ecological interactions. The philosophy and reasoning behind those models move beyond verbal and vague description of theories in ecology and evolution, towards an explicit test of theoretical expectation using simulation of complex systems.

Mechanistic dynamic simulation models of diversification draw from widely used niche-based species distribution models (Zimmermann et al. 2010), which are within the expertise of Drs. Graham and Zimmermann. However, these models are not applied to a single point in time (e.g. the present), but continuously iterated over the paleo-climate record, across a spatially heterogeneous and temporally dynamic landscape. As climatic factors change over time and space, species may respond by shifting their geographic ranges (i.e. dispersal), by adapting to novel climatic conditions (e.g. climatic niche evolution), or by going extinct locally or globally. Dynamic simulation models can realistically incorporate these ecological and evolutionary processes, generating a plethora of emergent biodiversity patterns that can be described and contrasted against real-world patterns (Leprieur et al. 2016, Rangel et al. 2018). Dynamic simulation models are within my expertise and that of Dr. Pellissier. The study of spatial patterns in phylogenetic diversity is within the expertise of Drs. Graham and Zimmerman, whereas the study of spatial patterns in species richness and phylogenetic diversity is within the expertise of all four of us.

Given our common research interests I propose to work with Drs. Graham, Pellissier and Zimmermann on five research projects:

• Developing the epistemological distinction between correlative and mechanistic models of biodiversity, as well as the statistical implications to evaluate, test and compare these two sorts of models. A manuscript in preparation by Dr. Pellissier and me has already been produced under this project, and it may become a review paper targeted at a medium-impact journal, or a conceptual paper targeted at a more philosophically-oriented journal.
• Comparing, and potentially integrating, simulations models designed and implemented independently in my lab (e.g. Rangel et al. 2018) and in Dr. Pellissier’s lab (e.g. Leprieur et al. 2016). Dr. Pellissier and I have already started discussing the methodological details of this project during my visit to WSL last fall, and we are confident that a comparative analysis of the two models have the potential to reach high-impact journal within macroecology (e.g. Ecology Letters).
• In my recent publication I employed a novel mechanistic simulation model (Rangel et al. 2018) to successful predict emergent patterns of species richness in South America. In in collaboration with Drs. Graham, Pellissier and Zimmermann I now plan to study how (i) simulated evolutionary events match real-world phylogenetic patterns of South American terrestrial vertebrates, and (ii) how existing phylogenetic methods may be used to reconstruct simulated biogeographical events using only the simulated phylogeny.
• Employing the deep-time mechanistic simulation model under development in Dr. Pellissier’s lab (GaSM - General Allele Simulation Model) to causally reconstruct Wallace’s zoogeographic regions of the world (Holt et al. 2013; on which Graham is an author). This project is being developed by Oskar Hagen, a PhD candidate in Dr. Pellissier’s lab, as a result of a suggestion I gave during my short visit to WSL during the summer of 2018. We expect that this project will yield a hyper-impact publication (e.g. Nature, Science or PNAS).

Employing a high-resolution medium-time simulation model to reconstruct the effects of the Great American Biotic Interchange to the patterns of biodiversity of the New World. This is a project that I will lead and would substantially benefit from inputs from Drs. Graham, Pellissier and Zimmermann. We will also aim at a hyper-impact journal (e.g. Nature, Science or PNAS).