A Zero-Inflated Spatio-Temporal Model for Integrating Fishery-Dependent and Independent Data under Preferential Sampling

2025-09-11

AI Summary

In fisheries stock assessment, integrating fishery-dependent data (FDD)—subject to preferential sampling—and fishery-independent data (FID)—collected via systematic surveys—is hindered by sampling bias and zero-inflation. To address this, we propose a zero-inflated mixed-effects spatiotemporal point process model with a six-layer hierarchical structure, jointly modeling presence/absence and biomass observations. The model explicitly corrects for vessel-level behavioral preferences in FDD and incorporates environmental covariates alongside vessel-specific random effects. Bayesian inference enables robust parameter estimation and preference signal detection. Simulation studies confirm accurate recovery of true parameters and reliable identification of sampling bias. Applied to sardine (*Sardina pilchardus*) in southern Portugal, the model substantially improves characterization of species distribution dynamics and stock assessment accuracy under multi-source data integration. This framework provides a generalizable, statistically principled approach for synthesizing heterogeneous fisheries data in management-relevant assessments.

Technology Category

• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Sustainable management of marine ecosystems is vital for maintaining healthy fishery resources, and benefits from advanced scientific tools to accurately assess species distribution patterns. In fisheries science, two primary data sources are used: fishery-independent data (FID), collected through systematic surveys, and fishery-dependent data (FDD), obtained from commercial fishing activities. While these sources provide complementary information, their distinct sampling schemes - systematic for FID and preferential for FDD - pose significant integration challenges. This study introduces a novel spatio-temporal model that integrates FID and FDD, addressing challenges associated with zero-inflation and preferential sampling (PS) common in ecological data. The model employs a six-layer structure to differentiate between presence-absence and biomass observations, offering a robust framework for ecological studies affected by PS biases. Simulation results demonstrate the model's accuracy in parameter estimation across diverse PS scenarios and its ability to detect preferential signals. Application to the study of the distribution patterns of the European sardine populations along the southern Portuguese continental shelf illustrates the model's effectiveness in integrating diverse data sources and incorporating environmental and vessel-specific covariates. The model reveals spatio-temporal variability in sardine presence and biomass, providing actionable insights for fisheries management. Beyond ecology, this framework offers broad applicability to data integration challenges in other disciplines.

Problem

• Integrating fishery-dependent and independent data with preferential sampling
• Addressing zero-inflation and sampling bias in ecological data analysis
• Modeling spatio-temporal species distribution patterns for fisheries management

Innovation

• Zero-inflated spatio-temporal model integrating fishery data
• Six-layer structure separates presence-absence and biomass observations
• Addresses preferential sampling biases in ecological data integration