Epidemics, disorder, and percolation

Spatial models for spread of an epidemic may be mapped onto bond percolation. We point out that with disorder in the strength of contacts between individuals patchiness in the spread of the epidemic is very likely, and the criterion for epidemic outbreak depends strongly on the disorder because the critical region of the corresponding percolation model is broadened. In some networks the percolation threshold is zero if another kind of disorder is present, namely divergent fluctuations in the number of contacts. We give an example, a network with a well-defined geography, where this is not necessarily so, and discuss whether real infection networks are likely to have this property.