Technology Category
Machine Learning: Calibration & Uncertainty QuantificationReasoning under Uncertainty: Relational Probabilistic ModelsData Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
Modelling benthic animals in space and time using Bayesian Point Process with cross validation: the case of Holoturians
🤖 AI Summary
Modeling the spatial distribution of sea cucumbers faces challenges including sparse presence-only data, uneven sampling effort, and habitat heterogeneity–induced detection bias. To address these, we develop a Bayesian spatiotemporal model based on the Log-Gaussian Cox Process (LGCP), integrating dive-based visual surveys and high-resolution photogrammetric data from nine field campaigns conducted on Giglio Island, Italy, between 2022 and 2024. The model incorporates a shared spatial Gaussian process component to jointly capture habitat structure (e.g., seagrass meadows), environmental covariates, and temporal variation. We propose a novel k-fold cross-validation strategy tailored for point processes and—uniquely for such models—evaluate predictive performance using the Continuous Ranked Probability Score (CRPS). Inference is performed efficiently via Integrated Nested Laplace Approximation (INLA). Our approach substantially improves predictive accuracy and provides a generalizable statistical framework for marine biodiversity monitoring in the Mediterranean Sea.
Technology Category
Application Category
📝 Abstract
Understanding the spatial distribution of Holothurians is an essential task for ecosystem monitoring and sustainable management, particularly in the Mediterranean habitats. However, species distribution modeling is often complicated by the presence-only nature of the data and heterogeneous sampling designs. This study develops a spatio-temporal framework based on Log-Gaussian Cox Processes to analyze Holothurians' positions collected across nine survey campaigns conducted from 2022 to 2024 near Giglio Island, Italy. The surveys combined high-resolution photogrammetry with diver-based visual censuses, leading to varying detection probabilities across habitats, especially within Posidonia oceanica meadows. We adopt a model with a shared spatial Gaussian process component to accommodate this complexity, accounting for habitat structure, environmental covariates, and temporal variability. Model estimation is performed using Integrated Nested Laplace Approximation. We evaluate the predictive performances of alternative model specifications through a novel k-fold cross-validation strategy for point processes, using the Continuous Ranked Probability Score. Our approach provides a flexible and computationally efficient framework for integrating heterogeneous presence-only data in marine ecology and comparing the predictive ability of alternative models.
Problem
Research questions and friction points this paper is trying to address.
Modeling Holothurians' spatial distribution with Bayesian Point Process
Addressing data challenges in presence-only species distribution modeling
Integrating heterogeneous marine data for ecosystem monitoring and management
Innovation
Methods, ideas, or system contributions that make the work stand out.
Bayesian Point Process for spatio-temporal modeling
Integrated Nested Laplace Approximation for estimation
k-fold cross-validation with Continuous Ranked Probability Score
🔎 Similar Papers
No similar papers found.
