A multi-stage Bayesian approach to fit spatial point process models

📅 2025-08-04
📈 Citations: 0
Influential: 0
📄 PDF

career value

202K/year
🤖 AI Summary
Bayesian spatial point process (SPP) modeling faces challenges including heavy reliance on numerical integration, high computational cost, and labor-intensive hyperparameter tuning. To address these, we propose a multi-stage recursive Bayesian framework that integrates parallel computation with recursive posterior updating. This enables efficient estimation of model coefficients and derived parameters within compact observation windows, while supporting posterior predictive inference for total abundance and point locations in unobserved regions. Compared to conventional approaches, our method substantially reduces dependence on numerical integration and manual tuning, enhancing both computational efficiency and scalability. We validate the framework through simulation studies and application to remote-sensing data of harbor seals in Johns Hopkins Inlet, Glacier Bay, Alaska. Results demonstrate superior accuracy and robustness, particularly under sparse observational regimes. The proposed method provides a novel, scalable tool for spatially explicit population modeling and conservation decision-making in ecology.

Technology Category

Application Category

📝 Abstract
Spatial point process (SPP) models are commonly used to analyze point pattern data, including presence-only data in ecology. Current methods for fitting these models are computationally expensive because they require numerical quadrature and algorithm supervision (i.e., tuning) in the Bayesian setting. We propose a flexible and efficient multi-stage recursive Bayesian approach to fitting SPP models that leverages parallel computing resources to estimate point process model coefficients and derived quantities. We show how this method can be extended to study designs with compact observation windows and allows for posterior prediction of total abundance and points in unobserved areas, which can be used for downstream analyses. We demonstrate this approach using a simulation study and analyze data from aerial imagery surveys to improve our understanding of spatially explicit abundance of harbor seals (Phoca vitulina) in Johns Hopkins Inlet, a protected tidewater glacial fjord in Glacier Bay National Park, Alaska.
Problem

Research questions and friction points this paper is trying to address.

Efficient Bayesian fitting for spatial point process models
Handling compact observation windows and unobserved areas
Improving abundance estimation in ecological presence-only data
Innovation

Methods, ideas, or system contributions that make the work stand out.

Multi-stage recursive Bayesian approach
Leverages parallel computing resources
Enables posterior prediction in unobserved areas
R
Rachael Ren
Department of Statistics and Data Sciences, The University of Texas at Austin, Austin, TX, USA
M
Mevin B. Hooten
Department of Statistics and Data Sciences, The University of Texas at Austin, Austin, TX, USA
T
Toryn L.J. Schafer
Department of Statistics, Texas A&M University, College Station, TX, USA
N
Nicholas M. Calzada
Department of Statistics and Data Sciences, The University of Texas at Austin, Austin, TX, USA
B
Benjamin Hoose
Department of Statistics, Texas A&M University, College Station, TX, USA
J
Jamie N. Womble
Glacier Bay National Park and Preserve and Southeast Alaska Network, National Park Service, Juneau, AK, USA
S
Scott Gende
Glacier Bay National Park and Preserve and Southeast Alaska Network, National Park Service, Juneau, AK, USA