Fire regimes and the conservation of sandstone heath in monsoonal northern Australia: frequency, interval, patchiness

Summary

Sandstone heaths are scattered across monsoonal northern Australia and they contain a number of fire-sensitive, obligate-seeder shrub species. The slow developmental time (3-5 years) required for post-fire germination to reproductive maturity, and therefore replenishment of the seed-bank, means that obligate-seeder shrubs are susceptible to frequent fire events caused by anthroprogenic factors. Worryingly, in the Nitmiluk National Park, over 40% of heaths have been burned on three or more occasions between 1989 and 1997. In this study, the effects of a fire lit three years after the previous fire on a mature heath community is investigated.

Study site: A site within Nitmiluk National Park on the Arnhem Plateau was selected that had a floristically rich mature heath vegetation, characteristic of sandstone substrates of the Arnhem Plateau, and that had a documented fire history. The vegetation at the chosen site thus comprised relatively species-rich shrubby heath (including a number of rapidly maturing obligate-seeder species) interspersed with spinifex hummock grass Triodia microstachya, and scattered scarlet gum Eucalyptus phoenicea and E. dunlopiana trees (< 8 m high). The site had burned by wildfire in October 1994 and in late 1989.

Transects: Five 50 x 10 m parallel transects were laid 20 m apart in the site. Each transect was split into ten 10 x 5 m plots.

Burning: The edges of the site were secured with small prescribed burns prior to the main burns. In early June, the site was burned in the late afternoon under mild fire-weather conditions. The fire was lit on the up-wind flank, 10 m outside of, but parallel to, the first transect. It was re-lit several times at its from edge as it moved across the plot. Afterwards, an estimate was made of the proportion of each plot that was burnt.

Monitoring: To ascertain the pre- (April 1997) and post- (April 1998, June 2000) burn communities, counts of the number of individuals and their reproductive state (none, flowering, fruiting/seeding) were made for each woody species per plot. Additionally, the presence/absence of each herbaceous species was recorded, and for spinifex the core diameter of each individual was measured. Finally, the percentage bare rock, leaf litter, bare soil, and grass (excluding spinifex) were estimated within ten 1 x 1 m quadrats for each plot.

Fire intensity: The fire was low intensity and patchy, with an estimated 49% of plots being burned. The area of a plot burnt was positively related to leaf litter and grass cover. The fire did not burn effectively in areas dominated by spinifex.

Herbaceous community: The majority of herbaceous species, both obligate-seeders and resprouters, increased after the fire. Indeed, of the 22 species that occurred in at least ten plots, fifteen obligate seeders and one resprouter increased significantly, while only one obligate-seeder decreased significantly (see Table 1, numbers estimated from original figure). Furthermore, while spinifex did not change in occurrence (98% plot occupancy) the mean number of individuals increased significantly from pre- (20 clumps) to post-fire (35 clumps). This was due to a seedling recruitment, illustrated by the significant decrease in core diameter from 1536 cm² ± 85 S.E. pre-fire to 822 cm² ± 79 post-fire.

Shrub community: Pre-fire, there was a total of 4099 woody plant individuals, of which 3607 (88%) were obligate-seeder shrubs, 306 (7.5%) resprouter shrubs, and 186 (4.5%) juvenile trees. In the post-fire community there was a decline of woody plant individuals of 48%, but the compositions of obligate-seeder (66.1% seedlings + 23% unburnt = 89.1%) and resprouter (10.9%) shrubs remained very similar. However, two thirds (n = 21) of obligate-seeder shrubs declined in frequency of occurrence and density post-fire, some by more than 50%. Indeed, of the 18 woody species that occurred in at least ten plots, mean densities of eight obligate-seeder and one resprouter species decreased significantly (see Table 1, numbers estimated from original figure).

By June 2000, the total numbers of individuals of woody obligate-seeders were similar to, and woody resprouters less than, April 1998 post-fire levels. Substantial seedling recruitment occurred in six obligate-seeder species (Boronia grandisepala [Rutaceae], B. lanuginose, Dryander's grevillea Grevillea dryandri, Homalocalyx ericaeus, Templetonia hookeri, and Tephrosia DNA1257) and two resprouter (Acacia gonocarpa and many-seeded wattle A. multisiliqua) species. However, populations of all but two (B. grandisepala and A. multisiliqua) of these species remained at least 50% below pre-fire levels.

Conclusions: The continued reduction in populations of obligate-seeders in June 2000, suggests that frequent fire has prevented the heath ecosystem from regenerating to pre-fire levels, and that further fire might have even more dramatic effects. Therefore, the authors suggest that individual fires for management purposes on sandstone heaths should be small, patchy and at intervals of at least 4-5 years apart.


Note: If using or referring to this published study please read and quote the original paper. This is available from http://www.environmental-expert.com/magazine/elsevier/biocon/index.htm. Please do not quote as a www.conservationevidence.com case as this is for previously unpublished work only.