| Abstract: |
Prolonged unsustainable grazing regimes in the past, combined with the invasion of introduced herbivores and predators, has degraded communities in the semi-arid central and northern Flinders Ranges, South Australia. To improve the capacity of the landscape to support native species, multiple, integrated pest control and monitoring programs were activated under 'Bounceback'. To track the response of small mammals and reptiles to the removal of a regulated grazing regime (cattle and sheep) and rabbit control, eight monitoring webs were surveyed between 1997 and 2011.
This program represents a successful long-term monitoring program (i.e. a lot of people finding funding and providing support > 15 years) and rare opportunity to learn and use this experience to adapt monitoring programs in future.
In 2013, analyses indicated monitoring at a ~3.5 ha scale, and repeated sampling all species over time, you need a lot of data to detect any difference in species metrics. This is most likely due to:
-High natural variability of semi arid to arid environments (i.e. rainfall) driving species distribution and abundance
- additional species contributing to total grazing pressure not being included in the monitoring program, and finally
-Recovery being a long process (e.g. this program set out to detect a localised response to management programs in a broadly degraded landscape with reduced potential to supply a source of native animals or plants within close proximity)
We found a stronger response to the removal of livestock by focusing on microhabitats around pit-fall traps (4m radius) and specific species.
So in future, a useful monitoring program may:
-Target sensitive or indicator species, within environments these species are likely to occur, at a suitable scale to detect change.
-Measure impact, and
-Increase in coverage to match how the Bounceback program has expanded over 15 years
Pre-analysis.
-Insufficient captures were obtained per web to enable accurate density estimates (>60 captures required: Anderson et al. 1983 Ecology, 64: 674-680). Assume 120 pitfalls / 3.5 ha will catch all resident animals, and compare relative differences between metrics in area with and without livestock, using new captures only.
-Inherent differences in the landscapes in the central and northern Flinders Ranges require data from each location to be analysed separately.
-Bounceback includes a range of pest control programs between 1997 and 2011. However , the only treatments that could be considered (i.e. two treatment webs and two control webs in both the central and northern Flinders Ranges) were: i) livestock / no livestock between 1997 - 2008, ii) rabbit control 1998-2000 (note RHD reached the monitoring sites in 1996 precluding a before / after investigation).
Key results.
-Small vertebrate abundance or richness did not differ between areas where livestock remained or had been removed, in the central Flinders Ranges or northern Flinders Ranges (RMANOVA)
-Community composition differed significantly in the central Flinders Ranges (p<0.05), and close to being significant in the northern Flinders Ranges (P<0.05; nMDS), and potential indicator species were identified in each case
-Focusing on abundant species, capture rates were very low, providing little confidence in the significant difference detected for the beaked gecko being more abundant in areas of no livestock in the central Flinders Ranges vs areas of livestock in the northern Flinders Ranges (ANOVA).
-Focusing on species already known to be sensitive to grazing impacts, capture rates were very low. Captures of the Bolom's mouse were slightly higher in no livestock areas (P=0.08; ANOVA).........
See Reference for more;
REFERENCE:
Source: Haby, N.A. & Brandle, R. 2013. Small mammal and reptile response to a 15-year old integrated management program in a semi-arid environment. Oral presentation at Ecotas13 Auckland 24-29 November 2013 |