How do forest buffers help control the spread of bovine tuberculosis?
Removing possums from strips of forest (buffers) adjacent to farmland is the main strategy for reducing the spread of bovine tuberculosis (TB) to livestock from uncontrolled possum populations in large areas of forest hinterland (Fig. 1). These buffers are created by aerial baiting followed by ground-based control, and they are usually about 5 km wide. They result in gradients in possum population density: high in untreated forest, medium along the forest boundary of the buffer (where some possums with home ranges overlapping the buffer are removed), and low inside the buffer. Such a gradient might influence the direction of movements of possums in forest immediately adjacent to buffer areas, and their movements may also be influenced by topographic features such as rivers and ridgelines.
Do possums cross buffers?
To assess the effectiveness of buffers in preventing the spread of TB to farmland, Andrea Byrom, with colleagues Roger Pech, Dean Anderson, Caroline Thomson and Morgan Coleman, investigated the movements of possums in podocarp-broadleaved forest immediately adjacent to buffers on five sites on the West Coast. The team set out to answer a series of questions, including (a) were there any consistent types of movement exhibited by possums, (b) are movement types associated with particular age/sex classes of possums, and (c) are movements influenced positively or negatively by major topographic features such as rivers and hills? To help TBfree New Zealand choose optimal buffer widths, field observations were modelled to determine the probability of infected possums crossing buffers of varying widths and arriving on farmland over the 6-month period of highest known dispersal (January to June). Also the model was used to predict how the likelihood of an infected possum arriving on farmland might vary if the density of possums is very low or very high in the untreated forest, and whether that risk increases if such populations have a high prevalence of TB.
Observations of possum movements
Each study site had a large river passing through the buffer and steep ridges in the untreated forest. In January, just before the peak dispersal period for subadult animals, possums were captured in untreated forest close to the buffer. GPS collars – placed on a mix of subadult, adult, male and female possums – were configured to provide 2–3 locations for each possum each night. In total, 79 possums yielded useful data for up to 4½ months.
Possums had four types of movement (Fig. 2): long-distance dispersal, exploratory moves, home-range displacement, and settled home range. The results suggest that long-distance dispersal is a relatively rare event and involved only 3 of 29 subadult possums. Two subadult males and 2 subadult females displayed exploratory movements and 6 possums (1 adult male, 2 adult females and 3 subadult males) were displaced. The remainder (83.5%) had settled home ranges. Dispersal and exploratory movements were not biased toward buffers even though these had few residual possums and the habitat was just as suitable as untreated forest. Instead, the possums settled in river valleys near waterways. There was no evidence that forested ridges changed movement patterns of possums or that possums crossed major rivers.
Predictive model for buffers
Computer simulations showed that the probability of an infected possum moving across a buffer to farmland was influenced by the width of the buffer (500–3000 m in the model), possum population density, and disease prevalence. For example, the predicted probability of an infected possum reaching farmland was 0.28 (or just over a quarter of all infected possums) with a buffer width of 500 m, population density of 9 possums/ha and disease prevalence of 0.10 (10%). The lowest probability of arrival was 0.0001 (or 1 in 10,000) and occurred with a buffer width of 3000 m, and low population density and disease prevalence (Fig. 3).
Implications for management
Buffers can provide short-term protection from disease incursions but the risk of a diseased possum moving to farmland will increase over time as (potentially infected) possums invade the buffer. The current 5-km buffer width generally applied to containment areas may be overly conservative, given the maximum observed dispersal distance of subadult possums of ~2.5 km. Buffer widths of just 2–3 km may be sufficient to prevent dispersing possums coming into contact with livestock, although the most cost-effective width will also depend on other factors such as the frequency of control. The results confirm that large rivers are barriers to possum movement and that it would be advantageous to concentrate follow-up possum control in riverine habitat.
This research was undertaken under contract to TBfree New Zealand.
Andrea Byrom, Roger Pech, Dean Anderson, Morgan Coleman & Caroline Thomson
Fig. 1 Conceptual illustration of the probability of an infected possum arriving on farmland (white) by traversing a poisoned buffer from uncontrolled forest (green). Possum silhouettes represent randomly located possums, and red silhouettes represent infected possums in an uncontrolled forest block (green). Arrows represent examples of dispersal and exploratory movements by possums.
Fig. 2 Examples of movements by (A) a subadult male dispersing long distance, (B) an exploratory subadult male, (C) an adult female displacing home range, and (D) an adult male showing a settled home range. The arrows indicate the major types of movement.
Fig. 3 Predicted probability of an infected possum crossing a forest buff er to farmland. Simulations were run with buff er widths ranging from 500 to 3000 m and with the following combinations of population density (number per hectare) and disease prevalence in untreated forest (in brackets): (A) 9 (0.1); (B) 2 (0.1); (C) 9 (0.02); and (D) 2 (0.02).
