National Woodland Bat Survey: Unvalidated Acoustic Classifications

Unvalidated automated species classifications from an England-wide passive acoustic bat survey across National Forest Inventory squares. Data are georeferenced using a delocalised grid, with bounding coordinates provided for each square. These sample data are intended for applications where classification uncertainty can be modelled into analyses at regional or national scales.

Use Cases

• Modeling bat species distribution uncertainty based on automated, unvalidated acoustic classifications.
• Assessing bat biodiversity trends across English woodlands using georeferenced, delocalised grid data.
• Developing or benchmarking machine-learning classifiers for bat echolocation call detection.
• Informing forest management and policy through large-scale, passive acoustic biological monitoring data.

Strengths

• Covers a national-scale survey (England-wide) conducted across a structured sampling frame (National Forest Inventory squares).
• Includes geospatial references with bounding coordinates for each surveyed grid square.
• Project is part-funded by a UK government programme (Defra's Natural Capital and Ecosystem Assessment), indicating institutional support.

Limitations

• Data are explicitly unvalidated and will contain incorrect species classifications.
• The example dataset is limited to a sample of only three squares, with data from additional squares available only on request.
• Key metadata such as row count, specific column names, and precise license terms are not provided in the available sources.

Provenance

Source

Forestry Commission & Bat Conservation Trust, collected as part of the National Woodland Bat Survey.

Collection Method

Passive acoustic recordings analyzed with automated machine-learning classifiers for bat echolocation call detection and species classification.

Freshness

Last updated on 2026-04-07 according to two platform entries.

Geography

England, specifically National Forest Inventory squares.