Fitting individual-based models of spatial population dynamics to long-term monitoring data

Abstract

Generating spatial predictions of species distribution is a central task for research and policy. Among the currently most widely used tools for this purpose are correlative species distribution models (cSDMs). Their basic assumption of a species distribution in equilibrium with its environment, however, is rarely met in real data and prevents dynamic projections. Process-based, dynamic SDMs (dSDMs) promise to overcome these limitations as they explicitly represent transient dynamics and enhance spatio-temporal transferability. Software tools for implementing dSDMs become increasingly available, yet their parameterisation can be complex. Here, we test the feasibility of calibrating and validating a dSDM using long-term monitoring data of Swiss red kites Milvus milvus. This population has shown strong increases in abundance and a progressive range expansion over the last decades, indicating a non-equilibrium situation. We construct an individual-based model with the RangeShiftR modelling platform and calibrate it using Bayesian inference. This allows the integration of heterogeneous data sources, such as parameter estimates from published literature and observational data from monitoring schemes, and consistent quantification of parameter uncertainties. Our monitoring data encompass counts of breeding pairs at 267 sites across Switzerland over an annual time-series of 22 years. We validate our model using a spatial-block cross-validation scheme and assess predictive performance with a rank-correlation coefficient. Our model showed very good and excellent predictive accuracy of spatial and temporal projections, respectively, well representing the observed population dynamics over the last two decades. Results suggest that reproductive success was the most decisive factor driving the observed range expansion. According to our predictions, the Swiss red kite population fills most of its range but density will further increase before stabilising mid century. With our case study we demonstrate the practicality of data integration and validation for dSDMs using available tools. This approach can improve predictive performance compared to cSDMs. The workflow exemplified here can be adopted for any population for which prior knowledge on demographic and dispersal parameters as well as spatio-temporal observations of abundance or occupancy are available. The resulting calibrated model provides refined insights into the ecology of a species and its predictions can inform conservation and management.

Fitting individual‐based models of spatial population dynamics to long‐term monitoring data

“…Data and code (Malchow, Fandos, et al, 2023) to replicate the presented analyses are available in Zenodo at https://doi.org/10.5281/zenodo.10435418. A development version of the R package RangeShiftR was used and is available in Malchow et al (2022) in Zenodo at https://doi.org/10.5281/zenodo.10689677.…”