Do the boundaries of protected areas keep invasive species out?

Overland trails along the KNP boundary

Overland trails travelling along the KNP boundary, showing an area of natural vegetation adjacent to the park boundary. © KNP

Do boundaries act as barriers to invasions? Staring at a map of the Kruger National Park with thousands of small dots on it, this is what C·I·B core team members Llewellyn Foxcroft and Dave Richardson, and collaborators Vojtěch Jarošík, Petr Pyšek and Mathieu Rouget asked themselves at a recent workshop. They then asked – if any “barrier effect” could be detected, could the features that make this a barrier be quantified. The unique dataset of invasive alien plants for Kruger seemed to offer a superb opportunity to test this idea. Kruger covers 20,000km² (equivalent to the area of Wales), extends about 360 km north to south, and has a boundary perimeter of about 1020 km. More than 27,000 spatially-explicit alien plant records are available – the thousands of dots on the map they were staring at.

Using a PDA (personal digital assistant) with integrated GPS, and loaded with customised CyberTracker software, rangers in Kruger collect data on features they observe in the field during their daily patrols. They not only record alien plants but many other observations such as rare animals, surface water availability, carcasses, fence breaks and many others. A major strength of the dataset is that these other (non-alien plant) observations can be used as absence records, since alien plants would be recorded at these sites if they were present. The total dataset (with these “absence” records) amounted to about 2 million records.

Fence along the KNP boundary

Fence running along the KNP boundary. The ‘hard’ boundary separates the park from areas being grazed by livestock or used for subsistence farming. © KNP

Plotting and analysing the distance of the plant records from the boundary, the researchers were amazed to find a clear pattern. At almost exactly 1500 m from the boundary there was a sharp drop in the numbers of alien plants. Concerned that this might be an artefact of the data and collection methods, they again analysed the data, this time using the ratio of alien plants to absence records in bands radiating into the park. The same pattern emerged again- clearly this was a robust result.

Using this knowledge the next step was to see whether the factors responsible for this effect could be identified. Outside the park, starting at the boundary and extending away from the park, they determined the density of major (national) roads and all (including secondary and gravel) roads, land use outside the park boundary, mean annual water runoff, presence of protected or natural areas, primary productivity of vegetation and others; in total they had 37 variables to test. The statistical wizardry of teammate Vojtěch Jarošík was harnessed to show that the average annual water runoff from the catchment outside the park was the most important variable for explaining the presence of alien plant records. It appears that there is a quantifiable threshold value of water runoff from surrounding areas below which invasion of alien plants is less likely. However, in these areas of lower runoff the chances of having alien plants present was only likely in areas with high road density within 10 km outside the park boundary. They also found that where rivers flowed into the park the presence of alien plants was unlikely only in areas where natural vegetation (although also possibly also grazed by livestock) comprised over 90% of the land use in a 5 km radius outside the boundary, and where there were no roads along the boundary inside the park.

Construction of Giriyondo entrance gate

Construction of the Giriyondo entrance gate, between Kruger National Park and Limpopo National Park in Mozambique. Areas of high road density adjacent to the park boundary should be prioritised for control and surveillance due to their higher potential for invasion. © KNP

Well, although very interesting, what does this mean for Kruger and other protected areas? First, the fact that the boundary acts as a natural filter means that maintaining a corridor along the boundary free of alien plants will significantly reduce the introduction of alien plants into the park across the landscape. Of course park managers cannot control alien plant control activities in the broader catchments outside the park (other than collaborating on landscape wide water and conservation committees), but they should actively engage in protecting the areas along the rivers outside the park in as undisturbed state as possible. Secondly, clearing should be prioritised along areas in the park that adjoin areas outside the park with a high road density, as this should also minimise the invasion of plants. Although the work also showed that it is ideal to have no roads along the park boundary to help limit invasions, and can provide insight into placement of new roads, work should also be carried out along those areas where there are roads in the park.

The researchers caution that this is only one aspect of preventing alien plant invasions in the park. For example, previous work showed that the introduction of plants for use in gardens was responsible for many of what are now the most important plant invasions in the park. Thus a multi-pronged approach to alien plant management requires clear strategies to prevent introductions and to prioritise control work.

Read the paper:

Foxcroft, L.C., Jarošík, V, Pyšek, P., Richardson, D.M. & Rouget, M. (2011). Protected-area boundaries as filters of plant invasions. Conservation Biology 25: 400405

Aerial photograph showing the boundary of Kruger National Park, with a number of different land uses adjacent to the park border. The different land use practices include urban areas, subsistence farming, commercial farming, roads and railways. (Courtesy of Google Earth).

Click on image to view enlargement.