Fig 1.
Study areas in Guinea, Congo and Gabon.
Background: Natural Earth / with Shaded Relief and Water (https://www.naturalearthdata.com).
Table 1.
Factors associated to the risk of Ebolavirus spillover used in the GIS-MCE.
Similar factors (in parenthesis) were considered as one factor for the classification. Some studies did not provide a significance value for the factors (NA). Significance of the risk factor is shown as NS: non-significant /*: p < 0.05 or ≤ 33% of variation explained /**: p < 0.01 or 33% to 66% of variation / ***: 0.001 or > 66% of variation explained.
Fig 2.
Schematic diagram of the GIS-MCE procedure followed to create the suitability maps for EBOV spillover.
Table 2.
Weights for risk factors resulting from the AHP.
Fig 3.
Maps for the four categories of risk factors in study areas in A) Guinea (December 2013), B) Congo (December 2001) and C) Gabon (August2001) (dataset available at https://doi:10.18167/DVN1/FZANMS).
Fig 4.
Suitability maps for EBOV spillover in areas in Guinea (A), Congo (B-D) and Gabon (E) (dataset available at https://doi:10.18167/DVN1/FZANMS).
Fig 5.
Suitability maps for EBOV spillover in forested Guinea in 2013 (dataset available at https://doi:10.18167/DVN1/FZANMS).
Fig 6.
Pixel values distribution in suitability maps for EBOV spillover in Guinea, Congo and Gabon.
Red lines show pixel values of sites where EBOV spillover were located.
Table 3.
Ebolavirus spillover suitability index of sites where spillover events have been previously reported compared to the values (mean and standard deviation) of the whole study area.
The percentile of suitability index for the spillover site is shown.
Fig 7.
Suitability maps of Ebolavirus maintenance in fruit bats (A) and insectivorous bats (B) in Guinée forestière. The size of the pie charts is scaled to sampling effort in each site (i.e. n = 1 for the smallest pie chart; n = 21 for the largest pie chart) (dataset available at https://doi:10.18167/DVN1/FZANMS).
Fig 8.
Pixel values distribution in suitability maps for Ebolavirus circulation in fruit bats (A) and insectivorous bats (B). Lines show pixel values of sampling sites where at least one bat tested positive for Ebolavirus antibodies (red lines) or where all bats tested negative (blue lines).
Table 4.
Mean ± standard deviation of suitability index in sampling sites where at least one bat tested positive for Ebolavirus antibodies and where all bats tested negative.
* Significance at p < 0.1 for t-test.
Fig 9.
Average relative change and standard deviation of suitability maps for EBOV spillover in the study areas in Guinea, Congo (December 2001) and Gabon (dataset available at https://doi:10.18167/DVN1/FZANMS).
Fig 10.
Average relative change in pixel value due to a ± 25% change introduced to the weight of risk factors in each category of risk factors (i.e. environmental, climatic, associated with bushmeat consumption and associated with potential reservoir/host species) in the three studied areas.
Colors correspond to risk factors in the four categories.