TABMON_Smit_Internship

Description Project Title: Quantifying the Relationship Between Environmental Vairables and BirdNET Classification Confidence Using Earth Observation Data and GIS This dataset contains the processed data layers for the TABMON project, conducted during an internship at the Norwegian Institute for Nature Research (NINA), Trondheim The dataset includes three primary components: BirdNET Detections: Aggregated and cleaned acoustic detection data from the TABMON recorder network. This includes scientific nomenclature, timestamps (UTC and Local), and confidence scores, filtered and audited for taxonomic consistency. Earth Observation (EO) Variables: A longitudinal suite of 16 environmental covariates extracted via Google Earth Engine (GEE). Data includes Climate (ERA5-Land: Temperature, Humidity, Wind, Precipitation), Terrain (MERIT DEM: Elevation, Slope, TEI), Vegetation (GEDI/MODIS: Canopy Height, LAI), and Anthropogenic factors (GHSL/VIIRS: Built-up fraction, Night-lights). All variables are spatially linked to recorder locations using a 250m ecological buffer. Device Metadata: Detailed metadata for the acoustic recording units (deployment IDs, hardware versions, and site-specific coordinates) used to account for device-level variance in the statistical models. Technical Specifications: File Formats: Primary data is provided in .parquet format for high-performance I/O and reduced disk footprint. Temporal Coverage: January 2025 – November 2025. Geospatial Reference: TABMON station network (Norway). Associated Code: The analytical pipeline used to generate and analyze this data is available on GitHub at: https://github.com/smit16912/TABMON