Source-linked AI summary
Infomap Bioregions: Interactive mapping of biogeographical regions from species distributions
Daniel Edler, Thaís Guedes, Alexander Zizka, Martin Rosvall, Alexandre Antonelli
TL;DR
Existing bioregion identification is often subjective or technically cumbersome, despite its importance for biogeography and conservation. Infomap Bioregions provides an interactive tool that clusters species distributions into bioregions, and its validation identified biologically meaningful regions across global amphibian and mammal datasets.
Problem
Bioregion identification remains subjective or technically cumbersome, leading many studies to use arbitrarily defined areas.
Method
Infomap Bioregions clusters bipartite networks of species and grid cells with Infomap to map bioregions from point occurrences or range maps.
Results
The tool identified biologically meaningful bioregions in global amphibian range maps and mammal point occurrences, including 62 mammal bioregions.
Takeaways & Limitations
Infomap Bioregions supports data-driven bioregion mapping and applications such as conservation targeting and ancestral-range reconstruction.
Takeaways & Limitations
Mammal results may be affected by sampling biases, inaccurate georeferencing, and incorrect species identifications.
Abstract
from arXiv · showhide
Biogeographical regions (bioregions) reveal how different sets of species are spatially grouped and therefore are important units for conservation, historical biogeography, ecology and evolution. Several methods have been developed to identify bioregions based on species distribution data rather than expert opinion. One approach successfully applies network theory to simplify and highlight the underlying structure in species distributions. However, this method lacks tools for simple and efficient analysis. Here we present Infomap Bioregions, an interactive web application that inputs species distribution data and generates bioregion maps. Species distributions may be provided as georeferenced point occurrences or range maps, and can be of local, regional or global scale. The application uses a novel adaptive resolution method to make best use of often incomplete species distribution data. The results can be downloaded as vector graphics, shapefiles or in table format. We validate the tool by processing large datasets of publicly available species distribution data of the world's amphibians using species ranges, and mammals using point occurrences. We then calculate the fit between the inferred bioregions and WWF ecoregions. As examples of applications, researchers can reconstruct ancestral ranges in historical biogeography or identify indicator species for targeted conservation.
INTRODUCTION
Bioregions capture spatial patterns in how species are grouped, but selecting areas has often been subjective and existing algorithms overly technical. Infomap Bioregions addresses this need with an interactive web-based tool that clusters species–grid-cell bipartite networks to map bioregions from species distribution data.
- Background: Species are spatially grouped into recognizable biogeographical regions, termed bioregions here.These patterns occur at both small and large scales.
- Problem: Choosing biogeographical areas has often been subjective because predefined-area coding does not determine the areas themselves.Researchers developed algorithms to map grid cells into biologically relevant regions, but area selection lacked quantitative support.
- Problem: Existing bioregion-mapping algorithms often require multiple, overly technical steps, leaving many studies reliant on arbitrarily defined areas.This motivates tools that are simple, effective, and flexible for mapping species distribution data.
- Contribution: Infomap Bioregions is a web-based interactive tool designed to make bioregion identification simple and effective for any species distribution data.The tool maps relevant distribution data into bioregions.
- Contribution: The method clusters bipartite networks containing species and grid cells, and it has outperformed unipartite approaches based on species similarities between grid cells.The comparison concerns methods that abstract away species into species similarities between grid cells.
DESCRIPTION
Infomap Bioregions is an interactive application that identifies taxon-specific bioregions from species distribution data through adaptive spatial binning and network clustering. It provides interactive maps, supporting tables, species indicators, and optional phylogenetic ancestral-range reconstruction.
- Infomap Bioregions identifies taxon-specific bioregions from user-provided species distribution data through an interactive web application.
- Adaptive spatial binning uses larger grid cells for sparse data and smaller cells for dense data, addressing uneven biodiversity sampling.This approach offers an advantage over conventional uniform binning when distribution data are unevenly distributed.
- The binned data form a species–grid-cell bipartite network that Infomap clusters into bioregions, with common and indicative species identified for each unit.Results are presented as an interactive map with supporting tables describing each bioregion.
- The application supports loading time-calibrated or uncalibrated phylogenetic trees and uses Fitch’s maximum parsimony method to estimate ancestral ranges.Species absent from the distribution dataset are ignored during ancestral-range reconstruction.
Input data
Infomap Bioregions accepts species distributions as point occurrences or range maps, along with phylogenetic data in standard tree formats.
- Species distribution data: Species distributions can be supplied as point occurrences in CSV or TSV files, with user-specified name, latitude, and longitude columns.Range maps are provided as shapefiles, which include multiple files.
- Phylogenetic data: Phylogenetic data are supported in NEXUS and Newick tree formats.
Output data
Infomap Bioregions adaptively converts species distribution records into grid-cell clusters and bioregion maps, which can be exported in formats supporting visualization, spatial analysis, and ancestral-range reconstruction.
- Export formats: Bioregion maps can be exported as .svg or .png files, while bioregion shapes can be exported as .geojson or shapefiles.These outputs support graphical and spatial-data workflows.
- Workflow: The application adaptively bins species records into geographical grid cells, extracts a species–grid-cell bipartite network, clusters it with Infomap, and visualises the resulting bioregions.The workflow uses data density to determine spatial resolution.
- Export formats: Species presence/absence matrices can be exported in .nexus format for further analyses in BioGeoBEARS, BayArea, and other compatible ancestral reconstruction tools.The .nexus output is intended for downstream ancestral-range analyses.
- Export formats: Summary tables of the most common and indicative species for each bioregion can be exported as .csv files, and the tree can be exported as .svg.These outputs summarize species associations and the inferred tree.
Adaptive resolution
Infomap Bioregions adapts grid resolution to the amount and spatial distribution of input data using a quadtree, while preventing regions from being supported by too few records. This produces higher-resolution bioregions where data are abundant and lower-resolution bioregions where data are sparse.
- Adaptive resolution: A quadtree hierarchically subdivides geographical space into four increasingly smaller quadrants until reaching the user-provided maximum cell size, default 4°.The algorithm then aggregates species into the resulting grid cells.
- Adaptive resolution: Cells exceeding the maximum capacity of 100 species occurrence records are recursively subdivided until reaching the minimum cell size, default 1°.The procedure adapts resolution to data density by subdividing crowded cells.
- Adaptive resolution: If subdivision leaves fewer than 10 species in a cell, the algorithm reverses the latest subdivision to avoid regions with too few data points.The minimum cell capacity defaults to 10 species.
- Adaptive resolution: This approach identifies high-resolution bioregions where data are abundant and low-resolution bioregions where data are sparse, avoiding overfitting and underfitting.It automatically adapts grid size to the amount and spatial distribution of input data.
- Adaptive resolution: For range maps, the application records each species in every minimum-size cell intersecting its range polygon before applying adaptive binning.Point occurrence data proceed directly through the adaptive-resolution criteria.
- Adaptive resolution: Users can tune bioregion size by adjusting Markov time, directing Infomap to search for larger or smaller regions supported by the data.This provides interactive control over clustering resolution.
Bipartite network
Infomap Bioregions represents species distributions as a bipartite network linking species to the geographic grid cells where they occur. Links are unweighted to avoid sensitivity to spatially biased sampling, while record density increases spatial resolution.
- The network has species and geographic grid cells as its two node types, with each species linked to every cell where it is present.
- Links are unweighted by record counts so results are not driven by spatially biased sampling.Instead, denser species records increase spatial resolution, producing larger networks when data are dense.
Bioregions and indicator species
Infomap Bioregions defines bioregions by clustering grid cells and species in a bipartite network, then presents them through maps, tables, and species summaries. It also identifies common and most indicative species for grid cells and bioregions.
- Bioregions: Infomap clusters grid cells and species in a bipartite network, with the resulting clusters defining bioregions.The application displays bioregions in different colours on a map and provides summary statistics and species lists for each bioregion.
- Indicator species: The application lists the most common and most indicative species for both grid cells and bioregions.
RESULTS AND DISCUSSION
Infomap Bioregions was validated using global amphibian range maps and terrestrial mammal point occurrences, with adaptive grid resolution reflecting differences in data density. The mammal dataset was linked to a phylogeny for 4,426 matched species.
- Validation datasets: 6,069 amphibian species were analyzed using global IUCN range polygons, alongside terrestrial mammal georeferenced point occurrences.The validation used range maps for amphibians and point occurrences for terrestrial mammals.
- Adaptive resolution: Grid cells varied from 4° to 2° according to spatial data density, with maximum capacity 100 and minimum capacity 5.The adaptive resolution was configured to accommodate differences in data density.
- Phylogenetic matching: 4,426 terrestrial mammal species were matched between the georeferenced dataset and a phylogeny compiled from 1,000 species-level mammal phylogenies.The maximum clade credibility tree was calculated using TreeAnnotator in BEAST v.1.8.2.
Amphibians
Infomap Bioregions identified 87 amphibian bioregions, including large species-rich regions and small regions associated with island endemism and high species turnover. The inferred regions largely coincided with previous biogeographical regionalizations while resolving small bioregions in areas such as Hispaniola.
- Amphibians: 87 amphibian bioregions were identified from IUCN species range maps.White areas with insufficient data were excluded from the analysis.
- Amphibians: Most species belonged to relatively large bioregions, but small regions in the Caribbean and tropical Andes reflected island endemism and high species turnover.The tropical Andes also contained many species concentrated in only a few cells.
- Amphibians: The identified bioregions largely coincided with those reported by Vilhena and Antonelli, with differences attributed to adaptive resolution and its settings.In the Neotropics, clustering also reflected regionalizations proposed for several subregions and provinces.
- Amphibians: Infomap Bioregions successfully identified small bioregions, including examples on the island of Hispaniola.The world map also depicted many small bioregions in Central America, the West Indies, and northwestern South America.
Mammals
The mammal analysis identified 62 bioregions spanning scales from continental-wide areas to patches of only a few square degrees. Interpretation is constrained by sampling biases, inaccurate georeferencing, and incorrect identifications, particularly in poorly sampled regions.
- Mammals: 62 mammal bioregions were identified, alongside phylogenetic trees and ancestral range reconstructions.The results are presented in Figure 3, with detailed results available in supplementary material.
- Mammals: The bioregions ranged from major continental-wide areas to regions covering only a few square degrees.More than 10 bioregions were identified for Australia.
- Mammals: Sampling biases, inaccurate georeferencing, and incorrect identifications may affect the results and obscure whether small bioregions in Russia are biological patterns or data artefacts.Improving reliability requires careful taxonomic revision and increased spatial sampling.
Validation
Infomap Bioregions was validated against WWF ecoregions and Holt et al.’s zoogeographic regions, showing generally good agreement. Fit varied by dataset and comparison, with amphibians generally receiving higher global GOF scores than mammals.
- WWF ecoregion comparison: The globally best GOF map can have finer-resolution areas with poorer local GOF than the reference map’s bioregions.This reflects differences between the chosen reference-map resolution and local resolution.
- WWF ecoregion comparison: Infomap Bioregions showed a generally good fit with WWF ecoregions, with spatially very good agreement across most areas.Differences were mainly associated with a number of very small bioregions.
- WWF ecoregion comparison: 0.65 global GOF for amphibians exceeded 0.54 for mammals in the WWF ecoregion comparison.The difference might be related to using range polygons for amphibians and point occurrences for mammals.
- Zoogeographic-region comparison: 0.77 global GOF for amphibians and 0.52 for mammals was obtained when comparing Infomap Bioregions with Holt et al.’s zoogeographic regions.The amphibian regions used approximately the same data, while Holt et al.’s method included phylogenetic information.
CONCLUSIONS
Infomap Bioregions is a flexible web application for simple, effective, data-driven identification of bioregions from point occurrences or range polygons. It supports direct parameter modification, adaptive spatial resolution, rapid processing, and exports for printing or further analyses.
- Infomap Bioregions identifies bioregions from point-occurrence data and range polygons through a web application.Users can load both species-data formats and modify parameters directly in the web interface.
- The application uses adaptive spatial resolution and can process millions of records in a few minutes.
- Results can be exported in various formats for high-quality printing or further biogeographical analyses.
AVAILABILITY AND FORTHCOMING EXTENSIONS
Infomap Bioregions is an open-source, client-side application available online and through its freely available source code. Planned extensions include batch processing, additional discovery methods, hierarchical clustering, phylogenetic integration, and bootstrap-based significance analysis.
- Availability: As a pure client-side application, Infomap Bioregions keeps data and calculations on the user’s computer while running heavy calculations in a background thread.The authors associate this architecture with improved privacy and performance.
- Availability: Infomap Bioregions is open source under the GNU AGPL v3+ license and is available online with freely available source code.The application is available at http://bioregions.mapequation.org, and its source code is hosted at http://github.com/mapequation/bioregions.
- Forthcoming extensions: Forthcoming extensions include batch runs, additional methods for finding indicator species and bioregions, hierarchical bioregion clustering, and deeper phylogenetic integration.The authors also propose significance clustering with bootstrap to identify statistically significant bioregional boundaries.