Choice of landscape discretisation method affects the inferred rate of spread in wildlife disease spread models.

Autor: Reimert MM; Section for Animal Welfare and Disease Control, Department of Veterinary and Animal Sciences, University of Copenhagen,. Electronic address: mossa@sund.ku.dk., Gussmann MK; Section for Animal Welfare and Disease Control, Department of Veterinary and Animal Sciences, University of Copenhagen., Boklund AE; Section for Animal Welfare and Disease Control, Department of Veterinary and Animal Sciences, University of Copenhagen., Denwood M; Section for Animal Welfare and Disease Control, Department of Veterinary and Animal Sciences, University of Copenhagen.
Jazyk: angličtina
Zdroj: Journal of theoretical biology [J Theor Biol] 2025 Jan 07; Vol. 596, pp. 111963. Date of Electronic Publication: 2024 Oct 09.
DOI: 10.1016/j.jtbi.2024.111963
Abstrakt: Disease modelling at the livestock-wildlife interface is an important topic for which discrete-space models are used for the wildlife component. One prominent example is African Swine Fever, where wild boar play an influential role as reservoirs of disease spillover into domestic pig farms. In this paper, we present a simulation study that demonstrates the impact of seemingly arbitrary choices of landscape discretisation method on the inferred rate of spread within the model. We use an ordinary differential equation model to implement a simplified model of disease transmission between discrete groups of wild boar with spillover into domestic pig farms contained within a homogeneous landscape. We examine a range of scenarios whereby the landscape is discretised into wild boar patches of varying size and shape, and compare the rate of spread between domestic pig farms placed at fixed points on the landscape. Our results demonstrate a non-monotonic relationship between patch size and rate of spread, which is particularly unstable and unpredictable for square and triangular shaped patches. Discretisation of the landscape into hexagons appears to produce a more stable relationship between patch size and rate of spread for the three types of transmission kernel we investigated. Although the rate of disease spread does converge to a stable value, this occurs at patch sizes that are much smaller than would be used in practice for wild boar. We conclude that outputs of disease models containing a wildlife component should not be considered to be robust to arbitrary choices for patch size and placement, but rather as a source of uncertainty to be examined using sensitivity analysis. Furthermore, we strongly recommend the use of hexagons rather than squares or right triangles for landscape discretisation.
Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
(Copyright © 2024 The Authors. Published by Elsevier Ltd.. All rights reserved.)
Databáze: MEDLINE
Externí odkaz: