Table 1.
Definition of parameters used in the model.
Fig 1.
(a): Biting behaviour of mosquitoes, when (i) imprinted on an amplifying host A, and (ii) on a dead-end host D.
A mosquito’s first bite is determined by host density and the host preference (modulated by the parameters ,
). All subsequent bites are determined by a mixture of fidelity (modulated by the parameter f), and host preference if the mosquito is not loyal to its imprinted host species. (b): Compartmental models. (i) Mosquitoes are introduced nulliparous (Sm), and imprinted onto the host species they first bite (
,
), and later may become infected by biting infected amplifying hosts (
,
, with potential for
if infected on their imprinting bite). Mosquitoes can die in any compartment. (ii), (iii) Amplifying and dead-end hosts respectively. Hosts can become infected if bitten by an infected mosquito and later go on to recover. Hosts are born susceptible and can die in any compartment. Cow and pig images obtained from https://www.phylopic.org, licensed under CC0 1.0.
Table 2.
Estimates and range of parameter values used in model analysis and global sensitivity analysis.
Fig 2.
Variation of feeding patterns of mosquitoes with fidelity, in the instance where cows (D) and pigs (A) are available in equal proportions.
The height of the bars indicates the proportion of mosquitoes that are imprinted on that host species (x-axis). The fill indicates the proportion of subsequent bites by those imprinted mosquitoes. With an increasing preference for pigs (left-to-right), a higher proportion of imprinting bites are on pigs. With increasing fidelity (top-to-bottom), the proportion of subsequent bites remaining loyal to the imprinted host increases. In the case of no fidelity (f = 0, top row), the proportion of subsequent bites replicates the proportion of initial bites. While the overall proportions feeding on D or A does not change with fidelity — as fidelity increases the population feeding on cattle vs pigs becomes increasingly isolated from each other.
Fig 3.
Basic reproduction number R0 values based on Eq B in S1 Text and values in Table 2.
The top panel represents the initial preference for amplifying hosts, and the right panel is the mosquito-to-host ratio. Blue indicates regions where R0<1, i.e. there is no outbreak.
Fig 4.
Dynamics of infected amplifying and dead-end host assuming 5 mosquitoes per host and fixed parameter values in Table 2.
The right panel indicates host composition and initial preference for amplifying hosts, while the top panel indicates different fidelity values. While outbreak size in amplifying hosts increases with fidelity, less than perfect fidelity leads to larger outbreaks in the dead-end host.
Fig 5.
Basic reproduction number (R0) for Cx. tritaeniorhynchus, Cx. gelidus, and Cx. vishnui based on varying host composition .
The initial preference for amplifying host species () and fidelity values (f) are provided in Table A in S1 Text, along with their corresponding 95% confidence intervals. Other parameters held constant: transmission probability—1, biting rate—1/3 day−1, mosquito mortality and birth rate—1/30 day−1, host recovery rate—1/5 day−1, host mortality, birth rate—1/365 day−1, and mosquitoes-to-host ratio—5.