Predicting Tree-related Microhabitats abundance and richness using Terrestrial Laser Scanner data in an old-growth forest: a comparison between parametric linear models and machine learning
Identification
- Clé publique
- radar:r-739ef27e711f9a01a9f02abdccbfa65c
- Auteurs
- di Biase, Antonio · Alvites Diaz, Cesar Ivan · Spina, Pierdomenico · Motta, Renzo · Santopuoli, Giovanni
- Année déclarée
- 2026
- Langue déclarée
- und
- Type source
- Software
- Première détection
- 2026-11-13T08:39:53.335854+00:00
- Dernière observation des métadonnées
- 2026-11-13T08:39:53.335854+00:00
- Dernière vérification HTTP
- Non effectuée
Institutions déclarées par la source
University of Molise · University of Florida · University of Turin
Résumé documentaire — langue originale
Tree-related Microhabitats (TreMs) are essential biodiversity indicators, yet large-scale temporal monitoring remains a challenge. While Terrestrial Laser Scanner (TLS) provides high-resolution data, its predictive capacity for TreMs abundance and richness is underexplored. This study evaluates TLS-derived metrics as TreMs proxies at the single-tree level in a Dinaric Alps old-growth forest. Twenty-one independent tree-level morphological (e.g. crown shape, asymmetry) and allometric (e.g. DBH, tree height) metrics were extracted from segmented point clouds to evaluate their predictive performance across 35 TreMs target. To apply a feature selection, a preliminary Random Forest (RF) algorithm isolated the five most relevant predictors per target based on their importance. Subsequently, parametric approaches, as Multiple Linear regression (LM) and Zero-Inflated Negative Binomial (ZINB), were compared against non-parametric Machine Learning (ML) models, specifically RF and eXtreme Gradient Boosting (XGB). Results demonstrate that Machine Learning algorithms, particularly RF, provided the highest predictive accuracy for most targets, excelling in the detection of macroscopic architectural alterations such as growth deformations (R² up to 0.361). Furthermore, ZINB regression proved critical for handling the zero-inflated nature of TreMs counts. Functional clustering identified allometric variables as primary ecological drivers of TreMs occurrence, with geometric irregularity providing target-specific fine-tuning. Despite robust overall performance (RMSE of 1.77 for total TreMs abundance), predictive accuracy dropped sharply for superficial TreMs like epiphytes and bark loss, highlighting the volumetric limitations of pure point cloud geometry. These findings confirm that TLS-derived metrics, coupled with an advanced ML framework, offer an innovative tool for biodiversity monitoring and sustainable forest management.
Identifiants
- DOI
- 10.5281/zenodo.23262474
Repères documentaires
- Famille documentaire
- unknown
- Types déclarés par les sources
- Software
- Éditeurs
- Zenodo
Les mentions d’entités ne constituent pas une preuve de reconnaissance juridique. La juridiction étudiée est distincte du pays d’affiliation des auteurs.
Sources
DataCite · Document 1
Citer cette référence
di Biase, Antonio; Alvites Diaz, Cesar Ivan; Spina, Pierdomenico; Motta, Renzo; Santopuoli, Giovanni. Predicting Tree-related Microhabitats abundance and richness using Terrestrial Laser Scanner data in an old-growth forest: a comparison between parametric linear models and machine learning. 2026. https://doi.org/10.5281/zenodo.23262474
JSON · CSL-JSON · BibTeX · RIS · CSV · JSON-LD ·
APA, Chicago, ISO 690 : formats bibliographiques fondés sur les champs disponibles. OSCOLA / Bluebook : données structurées uniquement ; aucune conformité juridique complète n’est revendiquée sans vérification des éléments nécessaires.
- APA
- di Biase, Antonio; Alvites Diaz, Cesar Ivan; Spina, Pierdomenico; Motta, Renzo; Santopuoli, Giovanni. (2026). Predicting Tree-related Microhabitats abundance and richness using Terrestrial Laser Scanner data in an old-growth forest: a comparison between parametric linear models and machine learning. https://doi.org/10.5281/zenodo.23262474
- Chicago
- di Biase, Antonio; Alvites Diaz, Cesar Ivan; Spina, Pierdomenico; Motta, Renzo; Santopuoli, Giovanni. Predicting Tree-related Microhabitats abundance and richness using Terrestrial Laser Scanner data in an old-growth forest: a comparison between parametric linear models and machine learning. 2026. https://doi.org/10.5281/zenodo.23262474
- ISO 690
- di Biase, Antonio; Alvites Diaz, Cesar Ivan; Spina, Pierdomenico; Motta, Renzo; Santopuoli, Giovanni. Predicting Tree-related Microhabitats abundance and richness using Terrestrial Laser Scanner data in an old-growth forest: a comparison between parametric linear models and machine learning. 2026. https://doi.org/10.5281/zenodo.23262474
Warning: Undefined array key "last_link_verified" in /home/clients/e27d81e12f04d2bfd747aa264540ac56/sites/_UDTR_APP/legal_library_reference.php on line 31
UDTR — Bibliothèque juridique mondiale des droits de la Nature. Predicting Tree-related Microhabitats abundance and richness using Terrestrial Laser Scanner data in an old-growth forest: a comparison between parametric linear models and machine learning. Notice documentaire: https://www.declarationuniverselledesdroitsdelarbre.org/global-rights-of-nature-radar/reference/r-739ef27e711f9a01a9f02abdccbfa65c/ · Dernière vérification HTTP : non effectuée
Données de cette version figée
Dossiers et relations documentaires
Rôles des sources
Source bibliographique / publication d’origine · https://zenodo.org/doi/10.5281/zenodo.23262474
Un lien vers une autorité ne prouve pas à lui seul une affirmation juridique. Les résumés et les textes intégraux restent soumis aux conditions de leur source.
Statut documentaire
Pertinence à examiner
Bibliothèque
Cette identité n’est pas encore intégrée à la Bibliothèque juridique.