Analyzing spatial patterns in biodiversity data

São Paulo School of Advanced Science (SPSAS)
Co-designing Biodiversity Assessments

Maurício Vancine

October 31, 2024

Access to slides

Maurício Vancine

Ecologist and PhD in Ecology
Post-Doc in Spatial Ecology (Prof. Mathias - Unicamp)
Spatial Ecology
Ecological Modeling
Ecological and Spatial Data Analysis
Amphibian Ecology and Conservation
Open source (R, QGIS, GNU/Linux)

Biodiversity concepts
Species occurrences
Spatial analysis
Application: Seleção Natural Platform

Install R and RStudio
Install packages
Computer testing
Basic R explanation (maybe)

Basic R explanation
Species occurrences
Spatial analysis

Objectives

Contribution of spatial data and analysis to biodiversity knowledge

Basic tools for downloading and cleaning species occurrence data
Basic tools for spatial analysis:
- Spatial units (grids, hexagons, convex hull)
- Point density (kernel maps)
- Species Distribution Modeling (very basic)

IMPORTANT!!!

We are in a safe and friendly space

Feel free to interrupt me and ask questions

IMPORTANT!!!

My English is a work in progress…

Biodiversity concepts

What is biodiversity?

Sustainable Ecology and Economic Development (SEED)

Biodiversity concepts

How to measure biodiversity?

Biodiversity concepts

How to measure biodiversity?

Biodiversity concepts

How to measure biodiversity?

Why do we want so much to measure biodiversity?

Biodiversity concepts

Carbon credit

Allows private companies to trade carbon emissions
Measured in tonnes of carbon dioxide equivalent (tCO2e)

Biodiversity concepts

Biodiversity credit

Allows private companies to finance activities (e.g. forest conservation or restoration) for positive biodiversity gains

Biodiversity concepts

Carbon credit vs Biodiversity credit

There is a small problem: we do not know all the biodiversity……

Biodiversity concepts

Seven biodiversity shortfalls

Species taxonomy (Linnean)
Species distribution (Wallacean)
Species abundance (Prestonian)
Species evolutionary patterns (Darwinian)
Species abiotic tolerances (Hutchinsonian)
Species traits (Raunkiæran)
Species biotic interactions (Eltonian)

Biodiversity concepts

Let’s select a biodiversity shortfall to explore

Species taxonomy (Linnean)
Species distribution (Wallacean)
Species abundance (Prestonian)
Species evolutionary patterns (Darwinian)
Species abiotic tolerances (Hutchinsonian)
Species traits (Raunkiæran)
Species biotic interactions (Eltonian)

Species occurrences

Format

Species occurrences

Sources

Field collections (field sampling)
Literature (articles, data papers, …)
Citizen science (e-Bird, iNaturalist, …)
Scientific collections and museums (National Museum of Brazil, Royal Botanic Gardens - Kew, …)
Databases (GBIF, speciesLink, VertNet, …)

Species occurrences

Databases

Species occurrences

R packages

Species occurrences

spocc

Global Biodiversity Information Facility (GBIF) (rgbif): Earth’s biodiversity
iNaturalist (inat): citizen science data on species observations
VertNet (rvertnet): vertebrate records from institutions and museums
eBird (rebird): citizen science data on bird abundance and distribution and checklist
iDigBio (ridigbio): biological and paleobiological specimens
Ocean Biogeographic Information System (OBIS) (obis): marine species data sets from all of the world’s oceans
Atlas of Living Australia (ALA)(ala): information about species in Australia

There is another huge problem: sampling bias…

Species occurrences

Sampling bias - Brazil

1 million (total) and 900 thousand (valid) occurrences for ~4,500 species
Groups: vertebrates, arthropods and angiosperms
All occurrences <1km from access routes (roads and rivers)

Species occurrences

Sampling bias - World

740 million occurrences of 375 thousand species
Representing only 6.7% of the sampled globe
At least 80% of records were within 2.5 km of roads

Species occurrences

CoordinateCleaner

Automated flagging of common spatial and temporal errors in species occurrences
Errors include:

Country and province centroids
Capital coordinates
Coordinates of biodiversity institutions
Duplicated coordinates per species
Urban areas
Seas
Equal longitude and latitude

Species occurrences

sampbias

A statistical method to evaluate and visualize geographic sampling biases
Biases include:

Cities
Airports
Roads
Rivers

How to analyze this data spatially?

Spatial analysis

Spatial units

Spatial analysis

Spatial units - Grids

Geographic area in regular cells, forming a grid
Each cell is a spatial unit that contains values

Spatial analysis

Spatial units - Hexagons

Geographic area in regular hexagons
Better for uniformity and connectivity, reduces edge effects

Spatial analysis

Convex hull

Smallest convex region that contains all the occurrences

Spatial analysis

Point density (kernel maps)

Represent the concentration of occurrences or events in a geographic area

Spatial analysis

Point density (kernel maps)

Kernel functions

Spatial analysis

Point density (kernel maps)

Bandwidth

Spatial analysis

Point density (kernel maps)

Kernel density estimation (KDE)

Spatial analysis

Species Distribution Modeling (SDMs)

Spatial analysis

Species Distribution Modeling (SDMs)

Generalized Linear Models (GLM) - binomial

Spatial analysis

Species Distribution Modeling (SDMs)

Suitability map | Potential distribution

Spatial analysis

Species Distribution Modeling (SDMs)

Application: Seleção Natural Platform

Minimum Viable Product

Application: Seleção Natural Platform

Aim

Prioritizes vegetation fragments that are most relevant for biodiversity conservation
Conservation actions:

Searching for new populations
Monitoring population dynamics
Creating of protected areas
Places for reintroduction of species

Application: Seleção Natural Platform

Methods

SDMs to predict the potential distribution of endangered animal species
Vegetation fragment prioritization based on species composition and landscape metrics