Figure 3

Density of Infected bats and basic reproduction number. In all panels blue/red accounts for Meliandou/Bamako respectively. The time series include one year of predictions based on monthly averaged environmental parameters from 2000 to 2015. The density of infected bats, (A), indicates that typical values (for β = 0.5) at Meliandou are ≈2bats/km² but environmental factors can increase at times this value to ≈15bats/km², e.g. July at Bamako. The shaded areas show the bounds for the infected bats as a function of β. When the environmental pressure pushes K^* → μK₀, the infected bats density increases (upper bound of the shaded areas) and reduces when K^* → K₀ (lower bound). As expected the basic reproduction number, \({ {\mathcal R} }_{0}\), panel (B) correlates with the density of infected bats as shown in panel (C). The disease propagates among the bats population as long as \({ {\mathcal R} }_{0} > 1\). This requires more than 7.30bats/km² infected bats. This prediction is based on a regression of the correlation curve: \({ {\mathcal R} }_{0}=-\,0.14+0.67\cdot {B}_{I}^{0.26}\) (R² = 0.99), with B _I expressed in \(\frac{{\rm{bats}}}{{{\rm{km}}}^{2}}\). The two selected locations display a counterphase behavior due to distinct climate conditions for the same period of the year. Within the confidence bounds of our framework (the bar stands for the standard deviation for β = 0.5) our model predicts a peak of infection in December at Meliandou when the Ebola epidemic started.