Individual-based distribution modelling with RangeShifter
Submitted by editor on 28 October 2021.
Figure 1. A dispersal heatmap showing how often each cell was part of an individual dispersal trajectory, overlaid on a patch-based landscape. All dispersers started from the large patch on the center left. Find out how to generate results like this with our RangeShifter tutorials on <https://rangeshifter.github.io/RangeshiftR-tutorials/>.
by Greta Bocedi and Anne-Kathleen Malchow
As we are faced with an accelerating global biodiversity crisis that is caused by multiple interacting and often anthropogenic environmental changes (Ceballos et al. 2015, IPBES 2019), it becomes an increasingly urgent challenge to understand and predict how species will respond to these threats in both ecological and evolutionary terms (Ehrlén and Morris 2015, Sih et al. 2010, Urban 2019). Such predictions are needed to design appropriate management interventions for protecting threatened species and limiting invasive species as well as to target habitat restoration efforts that can both enhance in-situ conservation and promote range shifting.
In order to generate useful predictions that accomodate rapid global changes we need species distribution models that adequately reflect the eco-evolutionary mechanisms which mediate species’ responses to environmental drivers. These mechanisms include the species’ physiology and demography, their capacity for dispersal and (genetic) adaptation as well as thier interactions with other species and the habitat (Urban et al. 2019). Such mechanistic models, also called process-based models, can help us to move beyond correlative modelling approaches whose underlying assumptions are often not met due to the dynamic nature of Earth’s current environmental conditions.
Mechanistic models, on the other hand, are often criticised as hard to parametrise, especially compared to their correlative counterparts. The required information to specify the different process rates and functional relationships that describe a mechanism can be manifold and their recording is rarely part of monitoring programs. However, promising avenues for obtaining this data exist in conducting targeted experiments, deriving knowledge from first principles like allometric scalings and energy balances, estimating parameters inversely from abundance or occurrence data, and gathering big data from remote sensing and citizen science projects.
Considering the high complexity of the problem of understanding population responses, the advancement that mechnistic models can offer has already been recognised. Several simulation models and platforms are actively being developed, among others there are Ramas, Vortex, Hexsim, Nemo-age, Niche Mapper, Steps and RangeShifter (for more examples see Lurgi et al. 2015).
RangeShifter is a mechanistic model that we initially developed (Bocedi et al. 2014b) in response to the many calls for moving towards integrated dynamic modelling approaches. The main objective was to provide an individual-based, spatially explicit modelling platform that integrated population dynamics with sophisticated dispersal behaviour, and that could be used for a variety of applications, from theory development to in-silico testing of management interventions. Indeed, since its release, RangeShifter has been used in studies addressing a range of issues, including testing the effectiveness of alternative management interventions to improve connectivity and population persistence (Aben et al. 2016, Henry et al. 2017, Bleyhl et al. 2021), facilitating range expansion (Synes et al. 2015, 2020), improving reintroduction success (Heikkinen et al. 2015, Ovenden et al. 2019), and investigating range dynamics of invasive (Fraser et al. 2015, Dominguez Almela et al. 2020) and recovering species (Sun et al. 2016).
Motivated by these and other successful applications, we are now happy to be able to provide a new and enhanced version, RangeShifter 2.0, along with the package RangeShiftR that interfaces the modelling platform with R. Two major novelties are the option for implementing temporally dynamic landscapes and a module for the explicit modelling of neutral and adaptive genetics controlling the dispersal traits. RangeShifter is entirely open source as published under the public licence GPLv3 and hence may be used, modified and shared under their terms. In order to provide easy access for all users, the RangeShifter website https://rangeshifter.github.io/ presents elaborate tutorials and a comprehensive user manual to offer learning resources at all levels.
The RangeShifter project constitutes an important step towards making frameworks for modelling range dynamics under global change accessible to a wider audience. We hope that the new possibilities and resources offered by RangeShifter 2.0 and the RangeShiftR package will inspire a more widespread use of mechanistic distribution models for further exciting applied and theoretical research.
References:
Aben, J. et al. 2016. The importance of realistic dispersal models in conservation planning: application of a novel modelling platform to evaluate management scenarios in an Afrotropical biodiversity hotspot. – J. Appl. Ecol. 53: 1055– 1065.
Bleyhl, B. et al. 2021. Reducing persecution is more effective for restoring large carnivores than restoring their prey. – Ecol. Appl. 31: e2338.
Bocedi, G. et al. 2014b. RangeShifter: a platform for modelling spatial eco-evolutionary dynamics and species' responses to environmental changes. – Methods Ecol. Evol. 5: 388– 396.
Ceballos, G. et al. 2015. Accelerated modern human–induced species losses: entering the sixth mass extinction. – Sci. Adv. 1: e1400253.
Dominguez Almela, V. et al. 2020. Integrating an individual-based model with approximate Bayesian computation to predict the invasion of a freshwater fish provides insights into dispersal and range expansion dynamics. – Biol. Invas. 22: 1461– 1480.
Ehrlén, J. and Morris, W. F. 2015. Predicting changes in the distribution and abundance of species under environmental change. – Ecology Lett. 18: 303-314.
Fraser, E. J. et al. 2015. Range expansion of an invasive species through a heterogeneous landscape – the case of American mink in Scotland. – Divers. Distrib. 21: 888– 900.
Heikkinen, R. K. et al. 2015. Modelling potential success of conservation translocations of a specialist grassland butterfly. – Biol. Conserv. 192: 200– 206.
Henry, R. C. 2017. Tree loss impacts on ecological connectivity: developing models for assessment. – Ecol. Inform. 42: 90– 99.
IPBES 2019. Global assessment report on biodiversity and ecosystem services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services. – IPBES secretariat, Bonn, Germany, 1148 pp., <https://doi.org/10.5281/zenodo.3831673>.
Lurgi, M. et al. 2015. Modelling range dynamics under global change: which framework and why? – Methods Ecol. Evol. 6: 247– 256.
Ovenden, T. S. et al. 2019. Improving reintroduction success in large carnivores through individual-based modelling: how to reintroduce Eurasian lynx Lynx lynx to Scotland. – Biol. Conserv. 234: 140– 153.
Sih, A. et al. 2011. Evolution and behavioural responses to human‐induced rapid environmental change. – Evol. Appl. 4: 367-387.
Sun, Y. et al. 2016. Predicting and understanding spatio-temporal dynamics of species recovery: implications for Asian crested ibis Nipponia nippon conservation in China. – Divers. Distrib. 22: 893– 904.
Synes, N. W. et al. 2015. A multi-species modelling approach to examine the impact of alternative climate change adaptation strategies on range shifting ability in a fragmented landscape. – Ecol. Inform. 30: 222– 229.
Synes, N. W. et al. 2020. Prioritising conservation actions for biodiversity: lessening the impact from habitat fragmentation and climate change. – Biol. Conserv. 252: 108819.
Urban, M. C. et al. 2016. Improving the forecast for biodiversity under climate change. – Science 353: aad8466.
Urban, M. C. 2019. Projecting biological impacts from climate change like a climate scientist. – WIREs Clim. Change 10: e585.