Estimation of Parameters in Inhomogeneous Neyman-Scott Processes Using Presence/Absence Data

Environmental monitoring is of particular importance for studying biodiversity and ecosystems. Many environmental monitoring programs emphasize plant registrations as part of their inventory responsibilities. Among the various methods available for surveying plant communities, we focus on presence/absence (P/A) sampling due to its underutilized potential. P/A sampling offers several advantages over other methods, particularly its efficiency in terms of time and cost. However, interpreting direct information from this type of data can be challenging, as the results are heavily dependent on plot size and species distribution patterns. To overcome these difficulties, model-based assumptions are necessary. In this article, we propose a method for estimating parameters of an inhomogeneous Neyman-Scott point process, specifically a Matérn cluster process, using P/A data. The inhomogeneity is modeled by allowing the offspring process intensity to vary with environmental covariates. The proposed estimators and their corresponding confidence intervals are evaluated through Monte Carlo simulations and empirical data (P/A registrations for three plant species) collected by surveyors in Northern Sweden. The results indicate that the method generally produces nearly unbiased estimators, particularly when the sample size is sufficiently large. These parameter estimates from the underlying inhomogeneous Neyman-Scott point process can subsequently be used to compute local estimates of expected plant density.