Using Bayesian Belief Networks to assess how to manage commercial wildlife trade in a complex world

International commercial trade is one of the greatest threats to many wildlife species. The conservation community is often deeply divided on how to manage the trade to ensure the long-term sustainability of these species. Trade strategies range from a ban on all trade, to various forms of regulated trade. The impact a strategy will have on the demand for wildlife products and on reducing illegal trade in a particular species is often hotly debated.

For example by making trade legal some might fear that consumers will view the wildlife product more favourably leading to an increase in demand and a negative impact on the population. Others may counter that the income from legal trade will increase resources for management of sites where the species occurs. This will increase the changes of reducing illegal wildlife trade activities on that land impacting the population positively. Often the debate suggests that only one or other of these things would happen. In fact both may  occur and it is the relative importance of each of these on the species that is key. Without a clear evidence-based framework for exploring this it can be difficult to reach consensus or understand where the key disagreements lie.

In a paper I wrote, with Dr Liz Bennett of the Wildlife Conservation Society and Professor EJ Milner-Gulland of University of Oxford, we  developed an evidence-based modelling approach to tackle this problem. What I like about our approach is that we don’t set out to test whether these different scenarios could occur. Instead we make it possible to investigate what would be the effect on the sustainability of the species under the different scenarios. For example, we show how to explore whether a species remains sustainable if we switch from a ban to regulated trade under different scenarios – for example if (1) consumer preferences remain negative and there are few resources for management available or if (2) consumer preferences for the species become positive and (3) if more resources become available for managing source sites (4) both consumer preferences become positive and there are more resources for site management.

If modelling shows that the species always does best under one trade strategy irrespective of the scenario that arises, different groups of stakeholders might be able to agree on the trade strategy even though they disagree about which scenarios they think would occur. Alternatively, the modelling may show that the best trade strategy for the species depends on which scenario arises. This could suggest that further work and potentially a more refined modelling approach is needed to understand which scenarios are most likely to occur. And it might indicate the relative importance of these different scenarios in affecting species sustainability.

In our paper we used Bayesian Belief Networks (BBN). These are a type of mathematical model that describes causal relationships between variables and can be used to explore the probabilities of outcomes under different scenarios. BBNs are transparent and user-friendly so stakeholders can contribute to the development of all aspects of the model, explore for themselves the consequences of different decisions and examine the interplay and relative importance of different drivers of the trade on the sustainability of a species. We demonstrated the approach on a theoretical species but would love to try implementing this approach for a real species.