Workshop 1 (January 8)
Towards the ‘next generation’ of species distribution modeling:
emerging themes and methods
Full day - 8:30 AM to 5:00 PM, Room S58, building RW I [link to map]
- up to 40 participants -
- Anna Cord, Helmholtz Centre for Environmental Research – UFZ, Germany
- Tomáš Václavík, Helmholtz Centre for Environmental Research – UFZ, Germany & Palacký University Olomouc, Czech Republic
- Dennis Rödder, Zoological Research Museum Alexander Koenig (ZFMK), Germany
- Jan O. Engler, Zoological Research Museum Alexander Koenig (ZFMK), Germany
- Joseph D. Chipperfield, Biogeography Department, Trier University, Germany
Aim and Content:
A predictive understanding of how organisms are distributed in space is a central challenge in biogeography. Over the last decades, species distribution models (SDMs) emerged as the most widely used approach for describing the spatial patterns of relationships between species occurrence and environmental conditions. Their increasing applications for theoretical and applied purposes have made SDMs one of the hottest and most rapidly growing fields in ecology and biogeography. Many novel aspects of SDMs have been thoroughly investigated and discussed in the scientific arena. However, novel themes and methods are emerging that present new challenges for SDM applications and may form the basis for their ‘next generation’.
This workshop will provide an introduction to four emerging themes that represent some of the current frontiers in SDM research. We will present a theoretical overview for each topic and use practical examples from published studies to illustrate new methods and their relevance for biogeographers. A discussion period will follow after the presentations bringing the different topics together. Two exercises, each of about 2-hours, will allow hands-on experience with data and software and provide useful R scripts for future use.
If possible, the participants should bring their own laptop computer and have R installed. Some prior knowledge of species distribution modeling and how to use R is helpful but not mandatory.
- Modeling species distributions with remote sensing data
The remote sensing section will focus on possible applications, useful techniques and potential pitfalls when using remote sensing data in SDMs. Remote sensing has produced an incredible amount of data across multiple spatio-temporal scales, providing biogeographers with information on the distributions of organisms, habitats and environmental gradients. This section tackles both conceptual and practical issues, including an overview of useful remote sensing data sources and indicators, the impact of spatial and temporal non-stationarity, and the role of remote sensing data for predicting potential versus actual species distributions.
- Incorporating niche evolution and phylogenetic regression in dynamic range estimation
The section on niche evolution will provide an overview of the current available tools that combine phylogenetic information with both current and paleoclimatic datasets to evaluate signals of niche conservatism or niche shift. Although nowadays frequently applied, correlative SDMs may provide reliable results only if specific criteria are met, concerning both the types of input data and the application of appropriate techniques. Therefore, special focus will be on the interpretation of results considering both the species realized niche, potential niche and fundamental niche.
- Integrating SDMs with models of landscape genetics
We will introduce novel techniques to combine SDMs with models of gene flow to jointly parameterize surfaces of habitat availability and landscape resistance. By incorporating genetic information, these techniques can provide better estimations of species ranges than using occurrence data alone. We will discuss the advantages and drawbacks for the use of these techniques in the emerging field of landscape genetics. The dual nature of this analysis allows the functional evaluation of SDM performance in terms of model fit and algorithm efficiency. By applying range-wide genetic datasets for different species, we will exemplify issues of over-parameterization, quantify the usage of the AUC and related model fit metrics, and evaluate existing SDM algorithms compared to newly available tools for conducting biologically more meaningful predictions of species distributions.
- Landscape epidemiology as a framework to study the biogeography of diseases
This section will focus on the burgeoning discipline of landscape epidemiology that integrates concepts of disease ecology with the macroscale lens of biogeography in order to examine the interactions between landscape heterogeneity and disease spread. Concepts fundamental to practicing landscape epidemiology will be discusses, including spatial scale, static versus dynamic modeling, spatially implicit and explicit approaches, inference versus prediction and landscape connectivity.