Assessing Catharus bicknelli (Bicknell’s Thrush) habitat dynamics: A high-resolution model based on LiDAR metrics

Assessing species occurrences can be challenging due to issues such as species detectability and difficulties in accessing habitats for surveys. These limitations hinder the ability to adequately inform conservation or management decisions, particularly for cryptic species, which can be partially addressed with high-resolution, spatially explicit models. However, high-resolution, spatially explicit models projecting species can help resolve these challenges. We modeled the probability of occurrence of Catharus bicknelli (Bicknell’s Thrush), a globally vulnerable species, in southern Québec (Canada) using LiDAR-derived metrics. The model was calibrated with 139 occurrences (with 10-m spatial accuracy), each paired with 10 random locations within a 523-1,569 m buffer, and validated with an independent dataset of 3,928 point counts. Based on model output, we applied two statistical thresholds to guide decision-making. The probability of occurrence of the species was highest in dense, low-c..., , # Assessing Catharus bicknelli (Bicknell’s Thrush) habitat dynamics: A high-resolution model based on LiDAR metrics

Dataset DOI: [10.5061/dryad.m0cfxppgc](10.5061/dryad.m0cfxppgc)

## Description of the data and file structure

#### File: BITH\_Threshold\_2025.zip

**Description:** Shapefile of the statistical thresholds from the LiDAR model on the probability of habitat occupancy of the Bicknell's Thrush

We provide the shapefile of the model (BITH_Threshold_2024.shp) on the probability of occurrence of Bicknell’s Thrush (*Catharus bicknelli*) in the massif de la Jacques-Cartier in southern Québec (Canada) using LiDAR-derived metrics and the R script (R_script_Tremblay-et-al-2025.R for for bird habitat modeling). This model was calibrated with 139 occurrences (with 10-m spatial accuracy), each paired with 10 random locations within a 523-1,569 m buffer, and validated with an independent dataset of 3,928 point counts. Based on model output, we applied two statistical thresholds to guid...,