Inhomogeneous Spatial Point Process Models for Species Distribution Analysis: A Systematic Review

Abstract

This study aims to systematically review the application of inhomogeneous spatial point process models (ISPPMs) for species distribution analysis. The review focused on (i) the trend in the use of ISPPMs, (ii) the general characteristics of the studies reviewed, and (iii) the practice of inhomogeneous spatial point process modeling. Based on specific criteria, a search using Publish or Perish (PoP) software and Google Scholar databases was performed for published papers on ISPPMs from 2006 to 2020. The study revealed a significant evolution in the use of ISPPMs. Most of the studies were conducted at regional, national, and continental scales. More than 60% of the papers used presence-only data. The linear model was the most used (47.12%). Maximum likelihood (21%) and minimum contrast estimation (19%) were the primary methods for estimating the fitted model parameters. The goodness of fit, performance analysis and model comparison guided fitting model validation. Moreover, many of these studies (56.91%) did not explicitly address the issues of model specification and spatial dependence. Furthermore, 47% of the articles considered did not clarify the estimation method used. New challenges and perspectives are to be explored.

Keywords

• cox point process
• inhomogeneous point process
• inhomogeneous poisson process
• markov point process
• spatial point process models