The UK government's plans to license badger culling for the control of tuberculosis (TB) in cattle are controversial; by contrast, the Welsh Assembly has decided to vaccinate rather than cull badgers.

Extensive badger culls may reduce cattle TB (C. A. Donnelly et al. Nature 439, 843–846; 2006), but complex disease dynamics mean that killing too few animals can actually increase it (C. A. Donnelly et al. Nature 426, 834–837; 2003). However, culling too many badgers risks local extinction, contravening the Bern Convention on the Conservation of European Wildlife and Natural Habitats. Natural England, the agency monitoring the cull, will therefore be required to set minimum and maximum cull numbers for each licence. But the effects are difficult to predict.

Targets for licences will draw on regional estimates of badger abundance, but badger densities are uncertain, owing to their secretive behaviour. Surveys of TB-affected areas in Gloucestershire, where one of two pilot culls is planned, indicate a mean density of 3.3 badgers per square kilometre, with a 95% confidence interval of 2.4–4.6 and substantial local variation (D. Parrott et al. Eur. J. Wildl. Res. 58, 23–33; 2012).

As well as measurement uncertainty, there will be random (Poisson) variation about mean densities, and binomial variation around mean capture probabilities. These three sources of uncertainty together mean that licensed culling of 344 badgers — intended to represent 70% of badgers within a 150-km2 area — could eradicate anywhere between 51% of the resident badger population (risking an increase in cattle TB) and 100% (risking a breach of the Bern Convention).

This uncertainty cannot be eliminated, but could be reduced by detailed badger surveys before and after each cull. This would increase culling costs, which are already projected to exceed the financial benefits for farmers.