Contact rates in wild boar populations: Implications for disease transmission

Abstract

Inter‐individual contacts in wildlife populations are usually highly heterogeneous. This variation translates into differential disease transmission rates between individuals, which have vital consequences for the spread, persistence, and control of infectious diseases. Wild boar (Sus scrofa) is an abundant game species across Europe that poses serious health threats to wildlife, livestock, and humans. However, factors shaping contact rates and structure in wild boar populations, key parameters in disease ecology, remain poorly studied. We quantified dyadic association rates, as a proxy of contacts, and individual‐based network centrality measures using telemetry data from 3 wild boar populations across Europe. Next, we examined the effect of sex, age, group membership, and space use on association rates and individual centrality. Contact rates depended strongly on the distance between individual home ranges; the most frequent associations occurred at distances of 0–1 km (mostly within groups), less frequent at 1–3 km (mostly between groups), and sporadic at >4 km. Association rates were an order of magnitude higher within social groups than between them. Between‐group association rates were only dependent on the distance between groups, with no apparent effect of animal sex or age. At the social network level, young animals (0.5–2 yr) showed greater between‐group connectivity and a more central position in the network than adults. Our results highlight substantial contact heterogeneities in wild boar populations, which should be considered in epidemiological modeling and disease control actions. First, wild boar contact rates are strongly constrained socially and spatially. Hence, management measures reducing these constraints, such as supplementary feeding and intensive hunting, may lead to increased disease transmission rates. Second, young wild boars show exceptional connectivity within the population, highlighting their high capacity for disease transmission. Therefore, targeted removal of yearlings should be considered to optimize disease control efforts. © 2018 The Wildlife Society.

Ecological drivers of African swine fever virus persistence in wild boar populations: Insight for control

“…Disease transmission was modeled using the force of infection equation (FOI; rate at which susceptible individuals become infected) outlined in Equation , where x i,j is the distance between the home range centroid of infectious individual i ( I i ) and susceptible individual j (S j ) (as defined by their x and y coordinates), a denotes alive individuals, b denotes infectious carcasses, d denotes direct transmission, c denotes carcass‐based transmission, β is the transmission probability that is specific to the transmission mechanism ( d or c ). To account for spatial contact behavior in wild boar (Podgórski, Apollonio, & Keuling, ), transmission probabilities were assumed to decay exponentially with distance according to the rate parameter λ (Table ). Additionally because wild boar exhibit heterogeneous contact structure due to family grouping (Pepin et al, ; Podgórski et al, ), probabilities of direct transmission and carcass‐based transmission were assumed to be more likely if contact occurred within the same family group ( β wd and β wc ).…”

“…To account for spatial contact behavior in wild boar (Podgórski, Apollonio, & Keuling, ), transmission probabilities were assumed to decay exponentially with distance according to the rate parameter λ (Table ). Additionally because wild boar exhibit heterogeneous contact structure due to family grouping (Pepin et al, ; Podgórski et al, ), probabilities of direct transmission and carcass‐based transmission were assumed to be more likely if contact occurred within the same family group ( β wd and β wc ). In Equation , w denotes I i ‐S j contacts that are within the same family group.…”

“…Each parameter had a uniform prior distribution as follows: frequency of introduction ϕ ~Unif(0,60), β d ~Unif(0.0001,1), β c ~Unif(0.0001,0.99), ρ h ~Unif(0.0005,0.1), ρ c ~Unif(0.0005,0.8), π ~Unif(0.1,1.5), θ ~Unif(0.5,6), λ ~Unif(0.1, 2.5), β wd ~Unif(0.01,1), and β wc ~Unif(0.001,1). Prior distribution ranges were informed by movement and contact data (Kay et al, ; Pepin et al, ; Podgórski et al, , ). As part of the parameter generation process we implemented the following constraints for each parameter set: β d > β c , β wd > β d , β wc > β c ; to further inform prior distributions with biologically realistic knowledge.…”

Detection error influences both temporal seroprevalence predictions and risk factors associations in wildlife disease models

“…While demographic risk factors may influence exposure risk at a local scale (Figure ), spatial heterogeneity of these effects might cause them to diminish at the national scale. Previous studies have found that wild pig contact is strongly influenced by local scale factors of animal age, group membership, and distance between group home ranges (Pepin et al, ; Podgórski, Apollonio, & Keuling, ). Additionally, it is possible that strong spatial factors (e.g., climatic or ecological effects) might overshadow any demographic effects that would otherwise be observed.…”

“…While demographic risk factors may influence exposure risk at a local scale (Figure 4), spatial heterogeneity of these effects might cause them to diminish at the national scale. Previous studies have found that wild pig contact is strongly influenced by local scale factors of animal age, group membership, and distance between group home ranges (Pepin et al, 2016;Podgórski, Apollonio, & Keuling, 2018 and have home ranges that are 160% smaller than uninfected deer (Edmunds et al, 2018). This reduced activity can result in increased sampling of uninfected individuals, leading to lower estimates of prevalence (Nusser, Clark, Otis, & Huang, 2008).…”