Abstract: Survival probability within a certain time horizon T is a common measure of population viability. The choice of T implicitly involves a time preference, similar to economic discounting: Conservation success is evaluated at the time horizon T, while all effects that occur later than T are not considered. Despite the obvious relevance of the time horizon, ecological studies seldom analyze its impact on the evaluation of conservation options. In this paper, we show that, while the choice of T does not change the ranking of conservation options for single species under stationary conditions, it may substantially change conservation decisions for multiple species. We conclude that it is of crucial importance to investigate the sensitivity of model results to the choice of the time horizon or other measures of time preference when prioritizing biodiversity conservation efforts.
Post-publication notes (will be updated sporadically)
Although this wasn’t our starting point, it soon turned out that this study inherited a lot of ideas from Emily Nicholson et al.’s excellent paper on “Objectives for Multiple-Species Conservation Planning”. Basically, we take a similar multi-species objective functions, and test for sensitivity of optimal conservation actions to a change of the time horizon. Turns out that most of them can be highly sensitive when time horizons are changed within the typical range of 30 to a few hundred years.
I found then, and still think, that this is a quite important and really under-researched topic – imo, already the question of conservation preferences in general is quite underrepresented in the literature, but questions related to time preferences are discussed to an even lesser extent. Unfortunately, this paper hasn’t really changed that, at least judging from the number of citation attracted so far, and I fear that this won’t change much in the future. It was my first paper, and in retrospective I think I presented the subject way to technical for the target audience. I hope a recent paper by Howard B. Wilson et al. that examines the same question will do a better job – I discuss this paper in a recent blog post