Dataset: AI-powered underwater imaging reveals multiscale patterns of presence and volume-based abundance in coastal fishes

Le résumé fourni par la source

Data and code associated with the manuscript submitted to a Scientific Journal ___________________________________________________________________ Title of the article: AI-powered underwater imaging reveals multiscale patterns of presence and volume-based abundance in Mediterranean coastal fishes Authors: Ignacio A. Catalán, Amaya Álvarez-Ellacuría, Jose-Luis Lisani, Josep Sánchez, Amalia Grau, Benjamin Casas, Miquel Palmer. Journal: Ecological indicators --------------------------------------------------------------------------- GENERAL DESCRIPTION --------------------------------------------------------------------------- This repository contains the data and analysis scripts used in the study by Catalán et al., 2026, published in Ecological Indicators. The materials provided here allow full reproducibility of the analyses presented in the manuscript, including modelling of fish presence–absence and volume-based abundance based on long-term underwater imaging and environmental covariates. The repository includes three data files and two analysis scripts, described below. --------------------------------------------------------------------------- DATA FILES --------------------------------------------------------------------------- 1) env_daily (csv). Daily environmental dataset used in the analyses. This file contains daily aggregated environmental variables (e.g. means, totals, or derived metrics) relevant to fish occurrence and abundance. These data are used mainly in daily-scale modelling. 2) env_hourly (csv). Hourly environmental dataset matching the temporal resolution of the underwater image acquisition. This file includes high-frequency environmental covariates used to characterize diel patterns and short-term environmental variability, and to link image-based observations with contemporaneous environmental conditions. 3) images_detections_master (csv). Master dataset of fish detections derived from automated underwater imaging. This file contains timestamps, taxonomic classifications, and associated metadata for all detections used in the study. It represents the biological response data for both presence–absence and volume-related abundance analyses. --------------------------------------------------------------------------- ANALYSIS SCRIPTS. --------------------------------------------------------------------------- 1) Presence_models R Script (txt) implementing the statistical models used to analyse fish presence–absence patterns. It includes data preprocessing, construction of environmental predictors, model fitting, and generation of outputs related to species occurrence. Running this script reproduces all presence–absence results reported in the manuscript. The user has to specify the species name (one is set as default). 2) Abundance_models R Script (txt) implementing the statistical models used to analyse volume-based abundance patterns. It includes data filtering, aggregation, model fitting, and visualization steps. Running this script reproduces all abundance results presented in the manuscript. The user has to specify the species name (one is set as default). Note: each species run can take 10-30 mins Note: outputs are saved into the working directory --------------------------------------------------------------------------- REPRODUCIBILITY NOTES --------------------------------------------------------------------------- The datasets provided correspond exactly to those used in the submitted version of the manuscript. File names and variable structures are consistent with those referenced in the Methods section. Analyses were conducted using reproducible scripts without manual intervention. --------------------------------------------------------------------------- ACCESS AND AVAILABILITY --------------------------------------------------------------------------- The contents of this repository are currently under embargo. All data and scripts will be made publicly available upon acceptance of the manuscript.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.