Skip to main content
Dryad

Impact of infectious diseases on wild bovidae populations in Thailand: Insights from population modelling and disease dynamics

Cite this dataset

Horpiencharoen, Wantida et al. (2024). Impact of infectious diseases on wild bovidae populations in Thailand: Insights from population modelling and disease dynamics [Dataset]. Dryad. https://doi.org/10.5061/dryad.kwh70rz6k

Abstract

The wildlife and livestock interface is vital for wildlife conservation and habitat management. Infectious diseases maintained by domestic species may impact threatened species such as Asian bovids, as they share natural resources and habitats. To predict the population impact of infectious diseases with different traits, we used stochastic mathematical models to simulate the population dynamics over 100 years for 100 times a model gaur (Bos gaurus) population with and without disease. We simulated repeated introductions from a reservoir, such as domestic cattle. We selected six bovine infectious diseases; anthrax, bovine tuberculosis, hemorrhagic septicaemia, lumpy skin disease, foot and mouth disease and brucellosis, all of which have caused outbreaks in wildlife populations. From a starting population of 300, the disease-free population increased by an average of 228% over 100 years. Brucellosis with frequency-dependent transmission showed the highest average population declines (-97%), with population extinction occurring 16% of the time. Foot and mouth disease with frequency-dependent transmission showed the lowest impact, with an average population increase of 200%. Overall, acute infections with very high or low fatality had the lowest impact, whereas chronic infections produced the greatest population decline. These results may help disease management and surveillance strategies support wildlife conservation.

README: InfectiousModel

Population dynamics and infectious disease modelling of wild bovids.
Explore the population impact of infectious diseases with different traits.

Description of the data and file structure

There are three R scripts:

  1. script_models_&_plotting_thisone.R: for model simulations and plotting population trends, the main parts contain:
    • Non-infectious disease model: includes birth rate (mu_b), natural death rate (mu_), and ageing rate (delta_) with 3 age classes: calf (c), subadult (sa), adult (a).
    • Six Bovine infectious disease models with five different model structures:
      • Anthrax, SI
      • Bovine tuberculosis, SEI
      • Haemorrhagic septicaemia, SIRS
      • Lumpy skin disease, SEIRS
      • Foot and mouth disease, SEIRMS/E
      • Brucellosis, SEIRMS/E 
      • Note: S = Susceptible animals, E = Exposed animals; I = Infected animal; R = Recovered animal; M = Calves with maternally derived immunity. Two modes of disease transmission were used in the script: frequency-dependent (FD) and density-dependent (DD).
    • Plotting the figures for all the model results and comparing the the population changes using the boxplots.
  2. script_population_dynamic_5sp.R: includes:
    • Non-infectious disease model for five bovid species (gaur, banteng, water buffalo, mainland serow, and Chinese goral), includes birth rate (mu_b), natural death rate (mu_), aging rate (delta_)  within 3 age classes.
    • Each species has different demographic parameters.
  3. PCA.R: for creating PCA biplots using the results from pca_table.xlsx. PCA is use here to identify which diseases showed similar traits group.

The result files from our models can be downloaded here:

  1. df_ndiff_allmodels.csv: for generating summary statistics and box plots.
    • Ndiff: Population change at the end of the simulation (100 years); negative values indicate a population reduction.
    • run: Number of model simulations.
    • model: Model name.
    • model2: Disease name used in modeling.
    • transmission: Mode of disease transmission; no = non-infectious disease model, FD = frequency-dependent, DD = density-dependent.
  2. ndiff_mean_all.csv: the mean population change for each model.
    • Model name: Model name.
    • Model code: Code for each model.
    • Average of the total population changes (%): Average percentage change in the total population, calculated from the starting population minus the ending population, divided by the number of simulations.
  3. pca_table.xlsx: for generating the PCA biplot, includes four important disease parameters.
    • model: Model name.
    • beta: Disease transmission rate.
    • incubation: Disease incubation rate.
    • infectious: Infectious rate.
    • fatality: Disease fatality rate.
    • Nchange_p: Average percentage of the total population changes.
    • disease: Disease name used in modeling. 
    • type: Modes of disease transmission; FD = frequency-dependent, DD = density-dependent.
Sharing/Access information

This data also accessible via https://github.com/Wantidah/InfectiousModel 

Code/Software

The R scripts were developed using R version 4.2.1 (2022-06-23). R software is required, and the necessary packages are provided within the scripts.

Methods

The R codes were developed to simulate the population dynamics of the gaur with and without disease infection. The disease and animal biological parameters were collected from peer-reviewed literature. We built the model structures from basic SIR models, with compartments varying based on disease traits. In the scripts, we used the Poisson distribution to calculate the probability of events and then calculated the average population changes (see equations 2.4 and 2.5 in the manuscript) to identify which diseases have the most impact on population changes. The population change results can be found in the supplementary file (ndiff_mean_all.csv).   

Funding

Education New Zealand, Manaaki New Zealand Scholarship

Royal Society Te Apārangi, Award: RDF-MAU1701

Percival Carmine Chair in Epidemiology and Public Health