Allocating ranger patrol effort

Using conservation planning tools

Andy Plumptre

Andy Plumptre received a Fairfield Osborne Memorial Fund grant from Wildlife Conservation Society to spend a three month sabbatical at University of Queensland to learn Marxan and to analyse data that had been collected in his eleven years in the Albertine Rift region of central Africa. Supervised by Richard Fuller and Hugh Possingham he used a novel approach of using Marxan to assess ways of improving ranger patrolling strategies to maximise their conservation impact in the Greater Virunga Landscape.

Ranger patrolling is one of the key tools used by conservation managers, particularly in tropical countries, to ensure law enforcement occurs in protected areas and that poaching is minimized. Several recent assessments of the effectiveness of protected areas have shown that supporting law enforcement at a site is one of the most important factors for ensuring the maintenance of the integrity of the site. Yet very little research has looked at how best to allocate patrol effort to deter illegal activities within a protected area and to maximize efficiency of law enforcement. Law enforcement using ranger patrols is effective where the level of deterrence is sufficient to make it difficult for poachers (of wildlife, timber and non-timber products) to gain from illegal activities. In the case of ivory from elephants and rhino horn, the incentives to poach can be huge and providing sufficient deterrence is very difficult. However for the other species (and products) the incentives are a lot lower and ranger patrolling can be an effective deterrence if planned well. But how much is enough? There is a need to better understand what level of re-visitation to an area provides an effective deterrence. For the most part ranger patrols are focussed on zones within protected areas where it is believed illegal activities are most abundant. This is very much influenced by the ability to detect illegal activities, and also the fact that as patrolling effort in one area increases then poachers adapt and move to other areas. Frequently, managers do not have the time to coordinate all the daily patrols from several ranger stations. Consequently, the patrolling effort is sometimes haphazard and often re-visits areas near the posts while leaving other areas rarely visited. The Greater Virunga Landscape (GVL) is one of the most biodiverse pieces of real estate in the world. Straddling the borders of Uganda, Rwanda and the Democratic Republic of Congo, it is made up of 13 protected areas including three World Heritage sites, one Ramsar site and one Man and Biosphere reserve. Since the early 2000s the Wildlife Conservation Society ( WCS) has been supporting the protected area authorities of these three countries to implement a law enforcement monitoring program using the software MIST and more recently SMART (www.smartconservationsoftware.org).

Figure 1: Frequency of patrolling within the GVL. Number of locations records per month are plotted per km2 and four locations (2 hrs) were deemed necessary to sufficiently search a cell. Therefore observations with more than an average of 48 locations per month were deemed to be searched sufficiently intensively to act as a deterrence.

Figure 1: Frequency of patrolling within the GVL. Number of locations records per month are plotted per km2 and four locations (2 hrs) were deemed necessary to sufficiently search a cell. Therefore observations with more than an average of 48 locations per month were deemed to be searched sufficiently intensively to act as a deterrence.

  “We developed a novel way of using Marxan to more efficiently target patrol effort to ensure the conservation of the key species and habitats in the GVL.”

We examined the existing data on patrol effort in the parks of the Greater Virunga Landscape and showed that while about 60% of the landscape was visited between 2000-2010, only 22% had been visited at a frequency of at least once each month over this period (figure 1). Assuming that a visitation rate of once per month provides sufficient deterrence to poachers, this meant that poaching was not being effectively tackled.

Analysis of the distances that ranger patrols travel from the 125 patrol posts in the landscape also show that most patrol effort was within 5 km of a patrol post (75% of patrol locations) and 50% were within 3 km of the patrol post. As the area that requires patrolling increases with the radial distance from a patrol post this means that the funding for law enforcement per unit area of the landscape is mostly expended within 2-3 km of a patrol post (figure 2). These scarce resources are therefore being used very inefficiently at present and we therefore investigated whether patrolling could be better planned and implemented for similar costs. Conservation planning methods generally target where to conserve a suite of conservation targets. In this study we developed a novel way of using Marxan to more efficiently target patrol effort to ensure the conservation of the key species and habitats in the GVL (Plumptre et al., 2014).

Figure 2. The current expenditure of park law enforcement budgets in relation to distance from the 125 patrol posts in the landscape.

Figure 2. The current expenditure of park law enforcement budgets in relation to distance from the 125 patrol posts in the landscape.

Using a habitat map developed by WCS and a suite of landscape species for the GVL, we set target levels for each habitat based on their size and conservation importance. We also set targets for each species based on its population size. A cost layer was developed using the predicted costs of patrolling in the GVL correcting for the time it takes to travel through different habitats and terrain and incorporating additional costs for overnight patrolling. Costs were based on the current expenditures for these activities made by the Uganda Wildlife Authority and Institut Congolais pour la Conservation de la Nature. A separate cost layer was similarly developed for mobile patrolling where a vehicle can transport a team of rangers to a site before they move off on foot. A third cost layer identified which of these two patrol methods was cheapest for any site within the landscape. Using these three cost layers we were able to show that the cost of patrolling the whole landscape at a frequency of once per month would absorb the total budgets for park operations. If however, patrols targeted those areas where illegal activities were most common (77% of all activities) then the costs could be brought down to about 57% of these total costs. But if patrolling targets those areas which are critical for the conservation of the key species and habitats identified, rather than trying to patrol the whole landscape, costs could be brought down to 38% of these total costs and as low as 16% of the total costs if patrolling focused in the areas where threats were most common within the Marxan selected areas ( Table 1).

table1

Table 1. Costs of patrolling a) whole landscape effectively, b) areas where threats are abundant in landscape, c) area selected under the Marxan best scenario to minimize costs to protect conservation features and d) areas where threats are abundant within the Marxan best scenario for conservation targets. The percentage of the total costs for patrolling the whole landscape is given in brackets.

Combining a conservation planning approach that selects areas in the landscape critical for the conservation of the key species and habitats and focusing patrol effort in these areas will ensure the protection of the conservation features of a site provided the target levels selected are appropriate. The method accepts that it is not possible to fully patrol the whole landscape effectively but at present this is clearly the case for the GVL where illegal activities remain high despite large resources being invested in ranger patrolling. The method requires testing but we believe that it has the potential to significantly reduce the impacts of illegal activities on the key species and habitats in the GVL.


More info: Andy Plumptre aplumptre@wcs.org References: Plumptre AJ, RA Fuller, A Rwetsiba, F Wanyama, D Kujirakwinja, M Driciru, G Nangendo, JEM Watson & HP Possingham (2014). Efficiently targeting resources to deter illegal activities in protected areas. Journal of Applied Ecology doi: 10.1111/1365-2664.12227

Leave a Reply