# Quantifying Invasive Pest Dynamics through Inference of a Two-Node Epidemic Network Model

^{1}

^{2}

^{3}

^{4}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Materials and Methods

#### 2.1. The Data

#### 2.2. The Two-Node Model

#### 2.3. Linear Noise Approximation

#### 2.4. Bayesian Inference

## 3. Results

## 4. Discussion

## Author Contributions

## Funding

## Data Availability Statement

## Acknowledgments

## Conflicts of Interest

## Abbreviations

LNA | linear noise approximation |

MJP | Markov jump process |

ODE | ordinary differential equation |

OPM | Oak processionary moth |

SDE | stochastic differential equation |

SIR | susceptible, infected, removed |

## References

- Manchester, S.J.; Bullock, J.M. The impacts of non-native species on UK biodiversity and the effectiveness of control. J. Appl. Ecol.
**2000**, 37, 845–864. [Google Scholar] [CrossRef] - Kenis, M.; Auger-Rozenberg, M.; Roques, A.; Timms, L.; Péré, C.; Cock, M.J.W.; Settele, J.; Augustin, S.; Lopez-Vaamonde, C. Ecological effects of invasive alien insects. Biol. Invasions
**2009**, 11, 21–45. [Google Scholar] [CrossRef] - Freer-Smith, P.H.; Webber, J.F. Tree pests and diseases: The threat to biodiversity and the delivery of ecosystem services. Biodivers. Conserv.
**2017**, 26, 3167–3181. [Google Scholar] [CrossRef] - Cuthbert, R.N.; Bartlett, A.C.; Turbelin, A.J.; Haubrock, P.J.; Diagne, C.; Pattison, Z.; Courchamp, F.; Catford, J.A. Economic costs of biological invasions in the United Kingdom. NeoBiota
**2021**, 67, 299–328. [Google Scholar] [CrossRef] - Department for Environment Food Rural Affairs. A Plant Biosecurity Strategy for Great Britain; Department for Environment Food Rural Affairs: London, UK, 2021.
- Eiswerth, M.E.; Johnson, W.S. Managing nonindigenous invasive species: Insights from dynamic analysis. Environ. Resour. Econ.
**2002**, 23, 319–342. [Google Scholar] [CrossRef] - Allen, L.J.S. A primer on stochastic epidemic models: Formulation, numerical simulation, and analysis. Infect. Dis. Model.
**2017**, 2, 128–142. [Google Scholar] [CrossRef] - Hulme, P.E.; Baker, R.; Freckleton, R.; Hails, R.S.; Hartley, M.; Harwood, J.; Marion, G.; Smith, G.C.; Williamson, M. The Epidemiological Framework for Biological Invasions (EFBI): An interdisciplinary foundation for the assessment of biosecurity threats. NeoBiota
**2020**, 62, 161–192. [Google Scholar] [CrossRef] - Wadkin, L.E.; Branson, J.; Hoppit, A.; Parker, N.G.; Golightly, A.; Baggaley, A.W. Inference for epidemic models with time-varying infection rates: Tracking the dynamics of oak processionary moth in the UK. Ecol. Evol.
**2022**, 12, e8871. [Google Scholar] [CrossRef] - Wagenhoff, E.; Blum, R.; Engel, K.; Veit, H.; Delb, H. Temporal synchrony of Thaumetopoea processionea egg hatch and Quercus robur budburst. J. Pest Sci.
**2013**, 86, 193–202. [Google Scholar] [CrossRef] - Groenen, F.; Meurisse, N. Historical distribution of the oak processionary moth Thaumetopoea processionea in Europe suggests recolonization instead of expansion. Agric. For. Entomol.
**2012**, 14, 147–155. [Google Scholar] [CrossRef] - Wagenhoff, E.; Veit, H. Five years of continuous Thaumetopoea processionea monitoring: Tracing population dynamics in an arable landscape of South-Western Germany. Gesunde Pflanz.
**2011**, 63, 51–61. [Google Scholar] [CrossRef] - Suprunenko, Y.F.; Castle, M.D.; Webb, C.R.; Branson, J.; Hoppit, A.; Gilligan, C.A. Estimating expansion of the range of oak processionary moth (thaumetopoea processionea) in the UK from 2006 to 2019. Agric. For. Entomol.
**2021**, 24, 53–62. [Google Scholar] [CrossRef] - Mitchell, R.; Bellamy, P.; Ellis, C.; Hewison, R.; Hodgetts, N.; Iason, G.; Littlewood, N.; Newey, S.; Stockan, J.; Taylor, A. Collapsing foundations: The ecology of the British oak, implications of its decline and mitigation options. Biol. Conserv.
**2019**, 233, 316–327. [Google Scholar] [CrossRef] - Mitchell, R.; Bellamy, P.; Ellis, C.; Hewison, R.; Hodgetts, N.; Iason, G.; Littlewood, N.; Newey, S.; Stockan, J.; Taylor, A. OakEcol: A database of Oak-associated biodiversity within the UK. Data Brief
**2019**, 25, 104120. [Google Scholar] [CrossRef] [PubMed] - Mindlin, M.J.; Le Polain de Waroux, O.; Case, S.; Walsh, B. The arrival of oak processionary moth, a novel cause of itchy dermatitis, in the UK: Experience, lessons and recommendations. Public Health
**2012**, 126, 778–781. [Google Scholar] [CrossRef] [PubMed] - Tomlinson, I.; Potter, C.; Bayliss, H. Managing tree pests and diseases in urban settings: The case of Oak Processionary Moth in London, 2006–2012. Urban For. Urban Green.
**2015**, 14, 286–292. [Google Scholar] [CrossRef][Green Version] - Godefroid, M.; Meurisse, N.; Groenen, F.; Kerdelhué, C.; Rossi, J.P. Current and future distribution of the invasive oak processionary moth. Biol. Invasions
**2020**, 22, 523–534. [Google Scholar] [CrossRef] - Cowley, D.J.; Johnson, O.; Pocock, M.J.O. Using electric network theory to model the spread of oak processionary moth, Thaumetopoea processionea, in urban woodland patches. Landsc. Ecol.
**2015**, 30, 905–918. [Google Scholar] [CrossRef][Green Version] - Edina Digimap. High Resolution (25 cm) Vertical Aerial Imagery [JPG Geospatial Data], Scale 1:500. Available online: https://digimap.edina.ac.uk (accessed on 12 May 2022).
- Ordnance Survey. 1:250000 Scale Colour Raster™. 2021. Available online: https://osdatahub.os.uk/downloads/open (accessed on 21 September 2022).
- Kermack, W.O.; McKendrick, A.G. A contribution to the mathematical theory of epidemics. Proc. Maths. Phys.
**1927**, 115, 700–721. [Google Scholar] - Andersson, H.K.; Britton, T. Stochastic Epidemic Models and Their Statistical Analysis; Lecture Notes in Statistics; Springer: New York, NY, USA, 2000; pp. 11–18. [Google Scholar]
- Ho, L.S.T.; Xu, J.; Crawford, F.W.; Minin, V.N.; Suchard, M.A. Birth/birth-death processes and their computable transition probabilities with biological applications. J. Math. Biol.
**2018**, 76, 911–944. [Google Scholar] [CrossRef] - Fintzi, J.; Wakefield, J.; Minin, V.N. A linear noise approximation for stochastic epidemic models fit to partially observed incidence counts. Biometrics
**2021**, 78, 1530–1541. [Google Scholar] [CrossRef] [PubMed] - Golightly, A.; Wilkinson, D.J. Bayesian parameter inference for stochastic biochemical network models using particle Markov chain Monte Carlo. Interface Focus
**2011**, 1, 807–820. [Google Scholar] [CrossRef] [PubMed][Green Version] - Fuchs, C. Inference for Diffusion Processes: With Applications in Life Sciences; Springer Science & Business Media: Berlin/Heidelberg, Germany, 2013. [Google Scholar]
- Kurtz, T.G. The relationship between stochastic and deterministic models for chemical reactions. J. Chem. Phys.
**1972**, 57, 2976–2978. [Google Scholar] [CrossRef] - van Kampen, N.G. Stochastic Processes in Physics and Chemistry; North-Holland Personal Library; Elsevier: Amsterdam, The Netherlands, 2001. [Google Scholar]
- Komorowski, M.; Finkenstadt, B.; Harper, C.; Rand, D. Bayesian inference of biochemical kinetic parameters using the linear noise approximation. BMC Bioinform.
**2009**, 10, 343. [Google Scholar] [CrossRef][Green Version] - Fearnhead, P.; Giagos, V.; Sherlock, C. Inference for reaction networks using the Linear Noise Approximation. Biometrics
**2014**, 70, 457–466. [Google Scholar] [CrossRef][Green Version] - Golightly, A.; Henderson, D.A.; Sherlock, C. Delayed acceptance particle MCMC for exact inference in stochastic kinetic models. Stat. Comput.
**2015**, 25, 1039–1055. [Google Scholar] [CrossRef][Green Version] - Townsend, M. Oak processionary moth in the United Kingdom. Outlooks Pest Manag.
**2013**, 24, 32–38. [Google Scholar] [CrossRef] - Straw, N.A.; Forster, J. The effectiveness of ground-based applications of Bacillus thuringiensis var. kurstaki for controlling oak processionary moth Thaumetopoea processionea (Lepidoptera: Thaumetopoeidae). Ann. Appl. Biol.
**2022**, 181, 48–57. [Google Scholar] [CrossRef] - Forestry Commision, Oak Processionary Moth (Thaumetopoea Processionea) Contingency Plan. 2021. Available online: https://planthealthportal.defra.gov.uk/assets/uploads/PPM-CPv2022.pdf (accessed on 28 October 2021).

**Figure 1.**Satellite images with eastings and northings (obtained from EDINA Digimap Aerial © Getmapping Plc [20]) of (

**a**) Bushy Park and (

**b**) Richmond Park with locations of nest removals between 2013 and 2021. (

**c**) The ‘Removed’ prevalence time series, $R\left(t\right)\equiv {R}_{t}$ (the cumulative numbers of tree locations where nest removal has taken place) is used as our observational data set. Bushy Park is shown in blue with open diamonds and Richmond park is shown in orange with filled circles. (

**d**) A subsection of the Ordnance Survey map OS Open Data [21]) with eastings and northings showing the location of the two parks.

**Figure 2.**Illustration of the two-node epidemic network model. In each node, the tree population transitions through the available states: Susceptible (S), infested (I), and removed (R). The transition between I and R is governed by the removal rate parameter, $\gamma $, which (in this case) is chosen to be identical within each node. For the transition between S and I, there is the standard infestation rate parameter, $\beta $, plus an additional infestation pressure resulting from the other node, with ${\alpha}_{12}$ corresponding to the infestation pressure from node 1 on node 2, and ${\alpha}_{21}$ the infestation pressure from node 2 on node 1. The resulting stochastic differential equation model for this scenario is given in Equations (1)–(5).

**Figure 3.**Inference results for the two-node network model applied to Bushy and Richmond Park. The within-sample posteriors for (

**a**) the susceptible tree time series for (

**i**) Bushy Park, ${S}_{1,t}$, and (

**ii**) Richmond Park, ${S}_{2,t}$, (

**b**) the infested tree time series, ${I}_{1,t}$ and ${I}_{2,t}$, and (

**c**) the removed tree time series, ${R}_{1,t}$ and ${R}_{2,t}$. The median of the within-sample posteriors is shown for Bushy Park with blue diamonds and Richmond Park with orange circles. In all cases, the shaded error bars represent the 95% credible interval. In (

**c**), the observed time series for ${R}_{1,t}$ and ${R}_{2,t}$ are shown as black dashed lines. In (

**d**) (

**i**–

**v**) the posterior parameter distributions are shown for $\beta ,\gamma ,\sigma ,{\alpha}_{12}$ and ${\alpha}_{21}$, respectively.

**Figure 4.**The estimated intra-park ($\beta {I}_{i}{S}_{i}$) and inter-park ($\beta {\alpha}_{ij}{I}_{i}{S}_{j}$) infestation components (see (2)) using the median posterior estimations for I and S in each year for (

**a**,

**b**) Bushy Park, and (

**c**,

**d**) Richmond Park.

**Figure 5.**An example scenario for possible new infestations occurring in (

**a**) Bushy and (

**b**) Richmond Park, averaged over 50 simulations of the stochastic two-node epidemic model with the median parameters estimated through the inference scheme (shown in Table 2). In both panels, the solid lines show the new infestations resulting from within the park (intra-park), with dashed lines showing new infestations resulting from the neighbouring park (inter-park). Error bars show the 95% credible intervals over the 50 simulations.

**Figure 6.**Forward simulations for the removal prevalence in (

**a**) Bushy and (

**b**) Richmond Park from the stochastic two-node epidemic model with the median parameters estimated through the inference scheme (shown in Table 2). In both panels, the solid lines show the mean removal prevalence with error bars showing the 95% credible intervals over 50 simulations. Black dashed lines show the observed data.

**Figure 7.**OPM nest density. (

**a**) Probability density of the number of recorded removed nests per tree in Bushy Park (blue, open diamonds) and Richmond Park (orange, filled circles). For visualisation purposes, this shows nest numbers up to 20 per tree only. (

**b**) Box plots of the number of logged nests per tree on a logarithmic scale, showing all recorded data.

**Table 1.**The input parameters used in the inference scheme for the two-node model detailed in Section 2.2, Section 2.3 and Section 2.4.

Input Parameters | |
---|---|

Total population | ${N}_{1}=5000$, ${N}_{2}$ = 40,000 |

Initial infested | ${I}_{1,0}=240$, ${I}_{2,0}=1400$ |

Initial model parameters | ${\theta}_{0}=({\beta}_{0},{\gamma}_{0},{\alpha}_{{12}_{0}},{\alpha}_{{21}_{0}},{\sigma}_{0})=(0.00001,0.8,0.3,2.5,1)$ |

Prior distributions | $log{\theta}_{i}\sim $N(0, 1${}^{2}$), $i=1,\dots ,4$ |

Tuning parameter | $\Sigma =\left(\begin{array}{ccccc}0.009& 0.003& -0.059& -0.009& 0.002\\ 0.003& 0.006& 0.011& 0.001& 0.001\\ -0.059& 0.011& 0.712& 0.072& -0.006\\ -0.009& 0.001& 0.072& 0.013& -0.002\\ 0.002& 0.001& -0.006& -0.002& 0.052\end{array}\right)$ |

Initial ODE mean | ${x}_{0}={\eta}_{0}=({N}_{1}-{I}_{1,0}-{R}_{1,0},{I}_{1,0},{N}_{2}-{I}_{2,0}-{R}_{2,0},{I}_{2,0})$ |

Initial ODE variance | ${V}_{0}={0}_{4,4}$ |

Observation matrix | $P=\left(\begin{array}{cccc}1& 1& 0& 0\\ 0& 0& 1& 1\end{array}\right)$ |

**Table 2.**The posterior means, standard deviations, medians, and 95% credible intervals for the inferred model parameters.

Parameter | Mean | Standard Deviation | Median | 95% Credible Interval |
---|---|---|---|---|

$\beta $ | $1.74\times {10}^{-5}$ | $1.40\times {10}^{-6}$ | $1.75\times {10}^{-5}$ | $(1.44,2.02)\times {10}^{-5}$ |

$\gamma $ | $0.78$ | $0.056$ | $0.77$ | $(0.67,0.90)$ |

$\sigma $ | $1.32$ | $0.289$ | $1.28$ | $(0.88,1.99)$ |

${\alpha}_{12}$ | $0.38$ | $0.268$ | $0.32$ | $(0.06,1.05)$ |

${\alpha}_{21}$ | $3.12$ | $0.312$ | $3.08$ | $(2.66,3.83)$ |

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |

© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Wadkin, L.E.; Golightly, A.; Branson, J.; Hoppit, A.; Parker, N.G.; Baggaley, A.W. Quantifying Invasive Pest Dynamics through Inference of a Two-Node Epidemic Network Model. *Diversity* **2023**, *15*, 496.
https://doi.org/10.3390/d15040496

**AMA Style**

Wadkin LE, Golightly A, Branson J, Hoppit A, Parker NG, Baggaley AW. Quantifying Invasive Pest Dynamics through Inference of a Two-Node Epidemic Network Model. *Diversity*. 2023; 15(4):496.
https://doi.org/10.3390/d15040496

**Chicago/Turabian Style**

Wadkin, Laura E., Andrew Golightly, Julia Branson, Andrew Hoppit, Nick G. Parker, and Andrew W. Baggaley. 2023. "Quantifying Invasive Pest Dynamics through Inference of a Two-Node Epidemic Network Model" *Diversity* 15, no. 4: 496.
https://doi.org/10.3390/d15040496