ABC_GP_IPM:A Python package for applying GP and ABC to Integral Projection Models, with a Soay sheep case study

Abstract The integration of Gaussian Process (GP) models with Approximate Bayesian Com-putation (ABC) has been explored as a flexible framework for constructing Integral Projection Models (IPMs), enabling non-parametric modelling of demographic rela-tionships and the incorporation of population-level information without explicit like-lihoods. However, the practical implementation of this framework – particularly the selection of ABC summary statistics and the execution of ABC-PMC samplers – re-mains non-trivial and can limit its broader adoption. To address this gap, we introduce ABC_GP_IPM , a Python package that provides a streamlined and user-friendly interface for constructing GP- and ABC-based IPMs. We demonstrate the utility and performance of the package through a real-world case study of Soay sheep ( Ovis aries ), illustrating how the software simplifies complex modelling workflows and data management in IPMs.