ecoTolerance: An R package for Assessing Road and Human Footprint Tolerance in Wildlife Species

The paper introduces ecoTolerance, an open-source R package that calculates Road Tolerance Index (RTI) and Human-Footprint Tolerance Index (HFTI) metrics to automate the assessment of wildlife species' tolerance to human pressures, thereby aiding in the identification of vulnerable species for conservation planning.

The Gist

Imagine you are trying to figure out which animals are the "tough guys" of the wildlife world and which ones are the "delicate flowers" when it comes to living near humans.

For a long time, scientists had to do this math by hand, using complex maps and spreadsheets. It was like trying to bake a cake without a recipe, measuring every ingredient with a ruler instead of a cup. It was slow, prone to mistakes, and hard to share with others.

Enter ecoTolerance, a new digital "kitchen tool" (an R package) created by researchers Diego Fernandes Miranda and Lucas Rodriguez Forti. Think of it as an automated recipe book that instantly tells you how well different species can handle the noise, lights, and roads of human civilization.

Here is how it works, broken down into simple concepts:

1. The Two "Stress Tests"

The package runs two specific tests on every animal or plant record it finds:

• The "Roadside Test" (Road Tolerance Index - RTI):
  Imagine a road is a loud, dangerous river. Some animals love to swim right next to the banks; others are terrified and stay far away.

  • The tool measures how close a specific animal was found to the nearest road.
  • The Score: If an animal is found right next to the road, it gets a high score (close to 1), meaning it's a "road warrior." If it's found deep in the forest, far from any asphalt, it gets a low score (close to 0), meaning it's a "road shy."

• The "Human Footprint Test" (HFTI):
  This is a broader test. Imagine the Earth is covered in a giant, invisible blanket of human activity. Some parts of the blanket are light and thin (wilderness); others are heavy and thick (cities, farms, bright lights at night).

  • The tool checks how "heavy" the human blanket is at the exact spot where the animal was found.
  • The Score: A high score means the animal is comfortable living in a busy, modified world. A low score means it needs quiet, untouched nature to survive.

2. How the Tool Works (The Magic Machine)

Before ecoTolerance, a scientist might have to download a map of all the roads, download a map of human activity, find thousands of animal locations, and then manually calculate distances for each one. It was like trying to count every grain of sand on a beach one by one.

ecoTolerance is the machine that does the counting for you.

• It cleans the data: It throws out bad information (like animals recorded in the middle of the ocean or in city centers where they don't actually live).
• It thins the crowd: If someone recorded the same bird in the exact same spot 50 times, the tool says, "Okay, we only need one of those," so the math isn't skewed.
• It does the math: It instantly calculates the scores for thousands of animals in seconds.
• It draws the pictures: It automatically creates maps showing where the "tough" animals live versus the "sensitive" ones.

3. Real-World Examples: Who Passed the Test?

The researchers tested this tool on five different species to see if it worked:

• The "Road Warriors": They found a frog called Boana faber. This frog had a high tolerance score. It was found right next to busy roads and in areas with lots of human activity. In fact, there were many records of this frog getting hit by cars (roadkill), proving it really does hang out near traffic.
• The "Road Shy": They looked at another frog, Boana boans. This one had a very low score. It stayed far away from roads and preferred quiet, untouched forests.
• The "Tree-Huggers": They looked at a tree, Copaifera langsdorffii. Interestingly, the same tree species had different scores depending on where it was growing. In one region, it was more tolerant of roads; in another, it was more sensitive. This shows that even the same species can adapt differently depending on its local neighborhood.

Why Does This Matter?

Think of conservation like triage in a hospital. You have limited resources (money, time, land) and you need to know which patients (species) are in the most danger.

• If a species has a low tolerance score, it's like a patient with a broken leg in a busy emergency room. They need immediate protection, and we should avoid building new roads or cities near them.
• If a species has a high tolerance score, they are like a patient with a minor scrape. They can survive in a busy world, so we might focus our limited resources on the ones that are struggling.

The Bottom Line

ecoTolerance is a free, open-source tool that turns a complicated, hours-long math problem into a simple, automated process. It helps scientists quickly answer the question: "Who can live with us, and who needs us to stay away?"

By making this data easy to get and easy to understand, the tool helps governments and conservationists make smarter decisions to protect the planet's biodiversity in our increasingly human-dominated world.

Technical Summary

1. Problem Statement

Human activities, particularly urbanization and the expansion of linear infrastructure (roads, railways), are primary drivers of biodiversity loss and habitat fragmentation in the Anthropocene. While some species are sensitive to these changes and disappear, others persist or thrive in modified environments.

• The Gap: Existing methods to quantify species tolerance to anthropogenic pressure are often fragmented, taxon-specific, or require complex, non-standardized manual workflows involving multiple software tools.
• The Need: There is a critical demand for a streamlined, reproducible, and open-source tool that can systematically calculate tolerance metrics for roads and human footprints across diverse taxa and spatial scales (from local plots to global inventories) to inform conservation strategies.

2. Methodology

The authors developed ecoTolerance, an open-source R package designed to automate the calculation of two continuous, complementary metrics: the Road Tolerance Index (RTI) and the Human Footprint Tolerance Index (HFTI).

Core Workflow

The package follows a structured pipeline (Figure 1 in the paper) comprising the following steps:

1. Data Preprocessing:
   • process_occurrences: Cleans species occurrence data by removing duplicates and applying spatial thinning (default 1 km buffer) to prevent over-representation of clustered points.
2. Data Loading:
   • load_roads and load_human_footprint: Imports vector road networks and raster human footprint data (global or regional, e.g., Brazilian datasets).
3. Index Calculation:
   • Road Tolerance Index (RTI):
     • Calculates the distance ($D_i$) from each occurrence record to the nearest road.
     • Applies a reference distance ($ref\_dist$), representing a "zero-effect" threshold (default 3.5 km for animals, adjustable for taxa like plants).
     • Computes the Individual Road Index ($RI_i$) using the formula:
       $$RI_i = \frac{ref\_dist}{ref\_dist + D_i}$$
     • The species-level RTI is the median of all $RI_i$ values for that species.
   • Human Footprint Tolerance Index (HFTI):
     • Extracts the Human Footprint value (0–50 scale) within a 1-km buffer of each record using the global Human Footprint raster (Mu et al., 2022).
     • Normalizes the value to a 0–1 scale:
       $$Normalized\ Human\ Footprint_i = \frac{Human\ Footprint\ Value_i}{50}$$
     • The species-level HFTI is the median of these normalized values.
4. Output Generation:
   • generate_all_reports(): Produces spatial distribution maps, kernel density plots for both indices, and CSV summaries.

Technical Dependencies

The package relies on standard R spatial and data manipulation libraries: sf (vector data), raster (raster data), dplyr (data manipulation), and ggplot2 (visualization).

3. Key Contributions

• Standardization: Provides a unified framework for calculating tolerance indices, reducing inconsistencies caused by varying manual scripts across studies.
• Automation: Automates data cleaning, spatial thinning, distance extraction, and statistical summarization, significantly reducing coding time and user error.
• Flexibility:
  • Supports both global and local datasets (e.g., specific shapefiles).
  • Allows users to adjust parameters (e.g., reference distance for roads, buffer sizes) to match specific biological traits (e.g., dispersal capacity).
  • Generates multi-level outputs (individual, population, and community levels).
• Open Science: As an R package available on CRAN, it ensures reproducibility and transparency, allowing the scientific community to inspect and modify the code.

4. Results and Application

The authors validated ecoTolerance using 3,782 occurrence records across five species representing different taxonomic groups and life histories:

• Mammals: Bradypus variegatus (sloth) and Sylvilagus brasiliensis (rabbit) showed moderate tolerance (RTI: 0.520–0.599; HFTI: 0.279–0.341), suggesting that even slow-moving or cursorial species can persist near roads if habitat cover remains.
• Amphibians: A comparison between Boana faber (183 roadkill records) and Boana boans (<5 roadkill records) revealed a stark contrast. B. faber exhibited high tolerance (RTI = 0.654, HFTI = 0.392), while B. boans showed low tolerance (RTI = 0.183, HFTI = 0.111). This correlates the indices with actual roadkill vulnerability.
• Plants: Copaifera langsdorffii was analyzed across two biomes (Cerrado vs. Atlantic Forest). The indices varied significantly between populations (RTI: 0.321 vs. 0.419), demonstrating the package's ability to detect local adaptation or differences in land-use history.
• Performance: The package successfully processed datasets ranging from 250 to 35,000 occurrences efficiently, confirming its scalability.

5. Significance

• Conservation Utility: ecoTolerance enables rapid identification of vulnerable species and priority areas for mitigation (e.g., wildlife crossings, habitat protection). It helps distinguish between species that are truly tolerant of human pressure and those that are merely persisting in degraded habitats.
• Hypothesis Testing: The tool facilitates testing ecological hypotheses regarding functional redundancy, local adaptation to anthropogenic pressures, and the impact of specific land-use changes.
• Scalability: By bridging the gap between local community studies and continental-scale biodiversity assessments, it offers a versatile tool for researchers, policymakers, and conservation planners to make evidence-based decisions in the face of rapid environmental change.

Limitations & Future Directions: The accuracy of the indices depends heavily on the precision of georeferenced occurrence data. Future versions aim to incorporate automated data quality checks and integrate additional layers such as deforestation rates and fragmentation metrics.