A decision theory approach to conservation fencing

"In a conservation funding climate that is characterized by insufficient resources, and where donors...are increasingly demanding that return on investment be demonstrated, conservation fencing decisions cannot continue to be based solely on ecological intuition." This is the assessment of a group of Australian collaborators seeking a practical method of identifying the best sites for erecting conservation fences. Fences may not sound like the most exciting or innovative management technique, but in fact they are often a vital component of reintroduction schemes, the success of which is often dependent on the exclusion of potential predators such as foxes (Vulpes vulpes), cats (Felix catus), feral dogs (Canis lupus familiaris), and dingoes (Canis lupus dingo); fences can also be used to limit the activity of exotic herbivores that compete with the natives for vegetative resources.

(An example of conservation fencing)

According to the Australian research group, there has been a fair amount of research on how to design useful fences, but not much work on how to choose places in which to erect them. A variety of local features can influence habitat suitability, including vegetation, management history, distance from infrastructure or boundaries, and cultural significance; expenses associated with construction and maintenance will also be important. This large number of variables can make fencing seem "a costly and uncertain action," which is why the researchers have used decision theory to develop a method of selecting--and defending, if necessary--fencing locations. They have also field-tested their recommendations in a conservation site in Western Australia in order to examine the efficacy of their plan.

The selection process involves four steps, the first of which is to state a specific conservation objective and then select a metric (a "benefit function") that can be used to measure how well a particular plan achieves this goal. For instance, the collaborators were tasked with creating an exclusion site (or sites) in the planned Lorna Glen Conservation Reserve; there, locally extinct mammal species could be reintroduced as part of the Operation Rangelands Restoration project. The failure of previous "hard introductions" indicated that many species would benefit from a "soft introduction" that would allow them to acclimate to their new habitat. Thus, the objective in Lorna Glen was to set aside as large a predator-free area as possible in which these reintroduced mammals could live and breed; size of site would therefore be an accurate indication of a plan's suitability.

(One area of the Lorna Glen Conservation Reserve, or Matuwa)

The second step is to collect data on spatial features that could impact the benefit function. These may be ecological, economic, social, or even political. In the case of Lorna Glen, for instance, the  managers collected information on the vegetation in, and underlying geology of, preserve land; land management history (particularly in association with pastoral activities), and the cultural significance of certain areas to local Aboriginal residents. Collection of these data should allow researchers to identify all potential fencing locations within a larger site. Once this list is compiled, researchers can move on to the third step: identifying constraints associated with the project. Cost is likely to be a major constraint, and managers should be sure to consider not only the initial outlay but also expenses that will be required for maintenance. Each project will likely also have a number of other constraints associated with the spatial feature data collection previously. For instance, conservationists in Lorna Glen wanted to avoid land degraded by cattle stockading, utilize existing clearings (in order to minimize costs associated with fence erection), incorporate a particular type of habitat (sand dunes/plains with spinifex) and underlying geology (calcrete) preferred by the mammals they were reintroducing, and stay away from areas of Aboriginal significance.

Each of these constraints can then be used as a filter to remove potential fencing sites from the list. If a site doesn't meet any one of the requirements, it cannot be included on the final consideration list, so the order in which the filters are applied is unimportant. However, progress will be much quicker if managers start with the filters they expect to be most stringent, as this will more speedily trim the list to a more workable number of options. The results of the filtering process can be dramatic; whereas the Lorna Glen project started with an infinite number of possibilities, the list was ultimately reduced to just 32 options. The final phase of the evaluation process involves assessing the benefit function associated with each potential choice in order to see which maximizes the desired output. The Lorna Glen managers used an equation that calculated the maximum amount of undegraded land that could be set aside by fencing in one or more of the areas that had made it through all of their filters. The calculation indicated one clear solution that was head and shoulders above the rest; this fencing scheme would protect 560 ha of non-culturally significant land, incorporate appropriate habitat features, and provide maintenance staff with easy access to its boundaries, all while staying within both short- and long-term budgetary constraints.

(A boodie, Bettongia lesueur, one of the native mammals that has been successfully introduced into the new exclosure at Lorna Glen)

The main goals of the decision theory method was to provide managers with a defensible way of choosing fence location. However, one beneficial side effect is that this process improves the ability of managers to logically consider what might otherwise be an overwhelming amount of information. According to the authors, people who are confronted with a large number of choices will often ignore many of them and then use oversimplified rules for making a selection; unsurprisingly, this method of choosing something that is "good enough," rather than "optimal," often does not facilitate the best conservation or management plans. By following a pre-established list of steps, such as those outlined here, managers are more likely to select an appropriate course of action. Further, by keeping detailed notes on each step of the process, they can make their methodology transparent--providing information not only to funding bodies, but also to other managers who may face similar issues.

The authors do admit, however, that their "flexible decision-support tool" is not perfect. During installation of the Lorna Glen fencing, for instance, they ran into several problems resulting from inaccurate or missing data; clearly, some amount of ground-truthing would have been useful to ensure that all ecological and spatial information used in the analyses was up-to-date. The collaborators are quick to point out that all methods are associated with uncertainty, however, and emphasize that their optimal site remained the first choice even after the updated data had been included in their analyses. Thus, the system appears to be robust, and may become even more so as future modifications are made to account for uncertainty more explicitly.

---
If you enjoy reading these posts, you should visit the Anthrophysis Facebook page!

Bode, M., Brennan, K.E.C., Morris, K., Burrows, N., and Hague, N. 2012. Choosing cost-effective locations for conservation fences in the local landscape. Wildlife Research 39:192-201.

Thanks to the following websites for providing the images used in this post:
http://www.conservation-contractors.co.uk/fencing-contractors.html
http://www.exploroz.com/Places/59059/WA/Lorna_Glen.aspx
http://www.milamba.com/australia/inhabit/animals/anim25.htm