Sri Lanka Bird Diversity Dataset

Dataset Summary

The Sri Lanka Bird Diversity Dataset with Environmental and Climate Features is a large-scale biodiversity dataset containing 1,552,048 bird occurrence records covering 429 bird species across all 25 provinces of Sri Lanka.

The dataset integrates bird observation records with spatial, climatic, environmental, and land-cover variables extracted at each observation location. It provides a comprehensive resource for machine learning, ecological modeling, and biodiversity research.

The dataset spans observations collected between 2014 and 2024, enabling studies on:

• Species distribution modeling (SDM)
• Climate–biodiversity relationships
• Habitat suitability prediction
• Bird species classification
• Environmental impact analysis
• Ecological machine learning applications

---

Dataset Details

Dataset Description

Each record represents a bird observation associated with geographic coordinates, observation metadata, and environmental conditions at the observation location and time.

The dataset contains:

• 1,552,048 observation records
• 429 unique bird species
• 25 provinces of Sri Lanka
• 10 years of observations (2014–2024)
• 24 feature variables

The dataset combines biodiversity observations with environmental variables derived from satellite imagery, climate models, and atmospheric datasets.

---

Dataset Creators

Developed by:

• Dilusha Chandrasiri
• Maneesha Herath
• Muditha Herath
• Yasith Hewarathna
• Gishan Bandara

---

Data Sources

The dataset integrates information from multiple publicly available sources:

Bird Occurrence Data

• Global Biodiversity Information Facility (GBIF)
  Bird observation records with species taxonomy and geographic information.

Environmental Variables

Variable Group	Source
Vegetation indices	Satellite remote sensing data (MODIS/Sentinel-derived products)
Land cover	ESA WorldCover / land-use datasets
Elevation	SRTM Digital Elevation Model
Climate variables	ERA5 climate reanalysis
Atmospheric variables	NASA MERRA-2 aerosol reanalysis
Geographic information	Sri Lankan administrative boundaries

Users should acknowledge and cite the original data providers when using this dataset.

---

Dataset Structure

The dataset is provided as:

The file contains 24 columns.

---

Features

Feature	Type	Description
index	Integer	Unique row identifier
verbatimScientificName	String	Scientific name of observed bird species
stateProvince	String	Province where observation occurred
individualCount	Float	Number of observed individuals
decimalLatitude	Float	Latitude coordinate (WGS84)
decimalLongitude	Float	Longitude coordinate (WGS84)
eventDate	Date	Observation date
avg_rad	Float	Average surface radiation
NDVI_raw	Float	Raw vegetation index
NDVI	Float	Normalized vegetation index
LandCover_Class	Integer	Land cover category identifier
elevation_meters	Integer	Elevation above sea level
Carbon_Mass	Float	Atmospheric carbon aerosol mass
Dust_Mass	Float	Atmospheric dust aerosol mass
SO2_Mass	Float	Sulfur dioxide aerosol concentration
Sulfate_Mass	Float	Sulfate aerosol concentration
Sea_Salt_Mass	Float	Sea salt aerosol concentration
Total_Aerosol_Extinction	Float	Aerosol optical extinction
temp_mean	Float	Mean temperature
rainfall	Float	Rainfall measurement
wind_mean	Float	Average wind speed
humid_mean	Float	Relative humidity
shortwave_radiation	Float	Solar shortwave radiation
lka_general_2020	Float	Sri Lankan environmental baseline indicator

---

Dataset Statistics

Statistic	Value
Total records	1,552,048
Bird species	429
Geographic coverage	Sri Lanka
Provinces covered	25
Observation period	2014–2024
Features	24
Missing values	None after preprocessing

---

Geographic Coverage

The dataset covers all provinces of Sri Lanka:

• Ampara
• Anuradhapura
• Badulla
• Batticaloa
• Colombo
• Galle
• Gampaha
• Hambantota
• Jaffna
• Kalutara
• Kandy
• Kegalle
• Kilinochchi
• Kurunegala
• Mannar
• Matale
• Matara
• Monaragala
• Mullaittivu
• Nuwara Eliya
• Polonnaruwa
• Puttalam
• Ratnapura
• Trincomalee
• Vavuniya

---

Data Processing Pipeline

The dataset was generated through the following processing workflow:

1. Bird occurrence records were collected and filtered for Sri Lanka.
2. Records with invalid geographic coordinates were removed.
3. Species names were standardized using taxonomic information.
4. Environmental raster datasets were spatially sampled at each observation location.
5. Climate variables were matched based on observation date and location.
6. Land-cover and elevation information were extracted.
7. Environmental features were normalized where required.
8. Missing records were removed.
9. Final feature vectors were generated for machine learning applications.

---

Intended Uses

This dataset is suitable for:

Ecological Applications

• Species distribution modeling
• Habitat suitability analysis
• Biodiversity assessment
• Environmental impact studies
• Climate change research

Machine Learning Applications

• Multi-class species classification
• Regression-based abundance prediction
• Geospatial prediction models
• Feature importance analysis
• Explainable AI studies in ecology

---

Example Machine Learning Tasks

Task	Target Variable	Input Features
Species Classification	verbatimScientificName	Environmental + geographic features
Bird Abundance Prediction	individualCount	Climate + habitat variables
Habitat Modeling	Species presence	Location + environmental variables
Regional Biodiversity Analysis	Province/species distribution	Full feature set

---

Dataset Limitations and Biases

Sampling Bias

Bird occurrence datasets collected from citizen science platforms may contain:

• Uneven geographic coverage
• Higher observation density near accessible locations
• Seasonal observation biases

Environmental Resolution

Environmental variables are derived from remote sensing and climate products. Their spatial resolution may not perfectly represent local habitat conditions.

Taxonomic Limitations

Species identification accuracy depends on the quality of original observation records.

Recommended Use

Researchers should consider sampling bias correction and ecological validation before deploying models for conservation decision-making.

---

Ethical Considerations

This dataset is intended for scientific research and educational purposes.

Users should avoid:

• Using predictions without ecological validation
• Drawing conservation conclusions from biased samples
• Misinterpreting correlations as causal relationships

---

License

This dataset is released under the:

Creative Commons Attribution 4.0 International (CC BY 4.0)

Users are free to share and adapt the dataset with appropriate attribution.

---

Related Paper

This dataset accompanies the following research paper:

How Environment and Urbanization Shape Bird Diversity in Sri Lanka

arXiv: https://arxiv.org/abs/2607.00582

The paper describes the methodology for constructing this dataset, including data collection, environmental feature extraction, preprocessing, and the machine learning analyses performed using the dataset. Readers interested in the complete methodology and experimental results are encouraged to refer to the paper.

If you use this dataset in your research, please consider citing both the dataset and the accompanying paper.

Citation

If you use this dataset, please cite:

@dataset{sri_lanka_bird_diversity_2026,
  title        = {Sri Lanka Bird Diversity Dataset with Environmental and Climate Features},
  author       = {Chandrasiri, Dilusha and Herath, Maneesha and Herath, Muditha and Hewarathna, Yasith and Bandara, Gishan},
  year         = {2026},
  publisher    = {Hugging Face},
  version      = {1.0},
  license      = {CC-BY-4.0}
}