Case study: Bounceback — rabbit control in the Flinders Ranges

The European rabbit is a major environmental pest in semi-arid and arid environments across Australia, including the Flinders Ranges in South Australia. In the past, rabbit numbers in the Flinders Ranges National Park (FRNP) have been among the highest in the region due to suitable terrain and soil types. Warren densities of 80—100 warrens/km2 were common across the park, with some higher-density infestations of up to 150/km2 in localised areas.

The Bounceback recovery program was established in response to the continued decline in habitat quality from overgrazing by introduced herbivores including rabbits, as well as other ecological issues in the Flinders region.

Rabbit control began in FRNP in 1994, as part of the early stages of the Bounceback program. In 1995, rabbit haemorrhagic disease (RHD) significantly reduced rabbit numbers and allowed the Bounceback team to take advantage of the low-density population.
Bounceback currently operates across a range of landholdings including pastoral properties, private sanctuaries and national park reserves. Land in the region is used mainly for conservation, ecotourism and sheep production, and also includes Aboriginal-protected areas. Rabbit control has been focused on areas that have suffered from high grazing pressure; both from domestic stock before the park was established and from introduced and native grazing animals (including rabbits, feral goats, euros, and red and western grey kangaroos). To date, warren ripping and blasting have been applied to more than 200 km2 of the northeast part of FRNP. Other pest control has also been carried out as part of Bounceback, including fox and goat control. The arid and rugged nature of the landscape continues to present a challenge for rabbit management.

Aim

  • The aims of ongoing rabbit control, as part of the Bounceback program, are to:
  • control the rabbit population to remove/reduce the impact of their grazing
  • reduce the abundance of rabbits
  • treat each management block within the larger treatment area on a rotational basis within a four-year cycle
  • monitor the resultant regeneration and condition of native vegetation.

Partners and management

Many partnerships have been developed through the broader Bounceback project including: SA Arid Lands Natural Resource Management (NRM) Board, Northern and Yorke NRM Board, Biosecurity SA, Sporting Shooters Association of Australia, Baker’s Vertebrate Pest Control, Conservation Volunteers Australia, Adnyamathanha community, Greening Australia, Arkaroola Sanctuary, Warraweena Private Conservation Park (Wetlands and Wildlife), Zoos South Australia, Flinders University, University of Adelaide, University of South Australia, Nature Foundation SA, Scientific Expedition Group, Rural Solutions SA, Yellow-footed Rock-wallaby Preservation Association, Australian Wildlife Conservancy, Bush Heritage Australia, Forestry SA and SA Water.

Process

Four phases of the rabbit control program have been completed so far:

  1. warren destruction by deep ripping using heavy machinery
  2. warren destruction by hand (using explosives and fumigation) — in areas inaccessible to machinery
  3. ongoing follow-up treatment of ripped warrens and new holes to ensure warrens are completely destroyed
  4. trial of 1080-poison oat baits as a supplementary control method (2011).

Following Aboriginal heritage site clearance a broadscale ripping program began in 1996 after an initial, smaller-scale round of ripping near Wilkawillina Gorge. The area was divided into a number of treatment blocks to be treated one by one. Before treatment began, warrens were GPS mapped and allocated a unique number.

Additional details recorded for each warren included the number of entrance holes, activity status and whether the warren was new or reopened. This information was used to make sure every marked warren was destroyed and to help park managers plan future control activities.

A bulldozer fitted with three winged tines was used to rip all accessible warrens across all treatment blocks to a depth of 90 cm. The winged tines helped to completely destroy warren structures and avoided the need to cross-rip.

Warren ripping was completed in 2002 and all follow-up work has since been done manually. This involves searching for and treating warren re-openings and new holes with explosives (also known as ‘blasting’). This destroys inaccessible warrens in difficult or remote areas and where ripped warrens have been re-opened or new holes are found. Because there are only a small number of re-opened and new holes apparent over a wide area, it is not cost-effective to use a bulldozer. A bulldozer also causes extra damage to the land — some tracks are still visible more than a decade after ripping.

Rabbit impacts on native vegetation were measured before and after warren ripping at a number of sites within the northern part of FRNP and nearby section of Gum Creek Station. Monitoring includes noting the number of perennial plants and the type of species along unfenced vegetation transects as well as fenced exclosures.

Herbivore spotlight counts and dung transect counts are also done by parks staff four times a year. Rabbit population density is estimated from the weight of dung collected on the transect each year, assuming 1 g/m2 equates to 0.8 rabbits/ha. Spotlighting is done from a vehicle that travels along three 30 km-long tracks in the treated and untreated areas.

Contract staff also spotlight opportunistically from quad bikes in the rabbit treatment area to help with ongoing control efforts. When any rabbit is spotted, it is shot, a GPS location is recorded and a marker peg is placed at the site. This allows the team to return to the area in daylight to search for possible areas of reinfestation or missed rabbit holes.

Features of the study

  • The main features of the Bounceback rabbit control program are:
    major warren ripping and blasting works
  • thorough follow-up treatment to ensure all warrens are fully destroyed
  • ongoing monitoring.

Results

Over 140,000 rabbit holes have been permanently destroyed by explosives since 2002. Spotlight observations have shown significant differences in rabbit abundance between treated and untreated areas over time. In the untreated section of FRNP, spotlight counts carried out 1996—2004 ranged from 80—470 rabbits/100 km of spotlight transect. Counts have increased considerably since, to around 4000 rabbits/100 km in 2010; this is equivalent to levels before RHD was introduced. In contrast, spotlight counts in treated areas from 1996—2008 were maintained below ~60 rabbits/100 km, sometimes dropping below the target level of 20 rabbits/100 km. However, recent good seasons have seen an increase in rabbit numbers to over 100/100 km in 2009-10.

Increased vegetation growth due to effective control and good seasons has made it harder to locate new warrens. In 2008-10, only 1 or 2 treatment blocks per year were treated in the rabbit control area. Contractors found that while few warrens were reopened after treatment (2% over five years), many ‘new’ holes were located during follow-up activities. Although these may include old holes that were initially missed, the high number of new holes suggests that follow-up needs to include the whole area and not just investigation of previously treated warrens. Rabbit densities also varied across the landscape, depending on the soil type and vegetation. Contractors found several hotspots of high rabbit numbers, which need more frequent follow-up treatment than the four-year cycle of retreatment initially allocated.

Monitoring of vegetation has shown that rabbit control has improved perennial seedling recruitment and allowed recovery from previous overgrazing. Significant increases in numbers of juvenile plants of elegant wattle (Acacia victoriae) and bitter saltbush (Atriplex stipitata) were observed on the warren-ripped areas. This suggests that young plants survived and grew better in areas subject to rabbit control. However, increased compensatory grazing by red kangaroos, euros, and sheep (where rabbit numbers have been reduced) may subsequently prevent such vegetation recovery. Ongoing monitoring is needed to determine the long-term response.

Cost

The cost of the first round of rabbit warren ripping was about $800/km2, depending on terrain, warren density and warren size. Cost of labour for ongoing treatment of warrens has become unsustainable due to the huge effort required to locate new holes. Regardless, where many rabbits appear to be living on the surface, warren destruction alone cannot reduce rabbit numbers to the required level6.

Mapping out the warrens across each treatment block is the most time-consuming part of the program. The effort put into mapping is on average just over two and a half times more than the actual blasting or treating of warrens across the same area. As an example, it would take an estimated 528 person days for contractors to map and blast an 88 km2 section of the 200 km2 treated area. However, this could vary depending on seasonal conditions, rabbit reinfestation rates, and potentially increased difficulty of mapping as vegetation growth (and cover) reduces the visibility of warrens. This shows the extraordinary effort and commitment needed to effectively control rabbits in such an environment.

What worked

  • Monitoring has confirmed that treated areas have significantly lower rabbit densities than nearby untreated areas.
  • Long-term monitoring has helped the project team evaluate program success and adapt management accordingly, eg current direction towards 1080 oat baiting.    
  • A high level of consistent rabbit control has consolidated the benefits of RHD and maintained reduced rabbit numbers with subsequent regeneration of vegetation, except for most palatable species.

What didn’t work

  • Rabbit control in the Flinders Ranges has been very labour intensive and the effort required for follow-up manual search and destruction of warrens has proved to be unsustainable.
  • The four-year treatment cycle has not been achievable in the last few years, due to increased vegetation growth in the control area.
  • Rabbits appear to be living above-ground in some areas, due in part to the impacts of systematic fox baiting which commenced in 1995.
  • Warren blasting has become more difficult and on its own cannot maintain reduced rabbit densities. To compensate for this, techniques such as 1080-poisoned oat baits are being investigated for use with existing methods.

Conclusion

Bounceback has achieved success in reducing rabbit impacts, primarily using warren destruction. However, long-term vegetation recovery in FRNP will be a slow process, even if rabbit numbers remain low. Ongoing monitoring and retreatment remain a vital part of the rabbit control program. Results have shown that an adaptive management approach is needed to guarantee continued success.

Links

PestSmart: Rabbits

Author Invasive Animals CRC
Year 2011
Publisher Invasive Animals CRC
Pages 4
ISBN/ISSN PestSmart code: RABCS2
Control method Warren Ripping
Region SA
Documents

PestSmart Case Study: Bounceback – rabbit control in the Flinders Ranges (650 kb PDF)