Fire regimes, fire-sensitive vegetation and fire management of the sandstone Arnhem Plateau, monsoonal northern Australia

Citation
J. Russell-smith et al., Fire regimes, fire-sensitive vegetation and fire management of the sandstone Arnhem Plateau, monsoonal northern Australia, J APPL ECOL, 35(6), 1998, pp. 829-846
Citations number
73
Categorie Soggetti
Environment/Ecology
Journal title
JOURNAL OF APPLIED ECOLOGY
ISSN journal
00218901 → ACNP
Volume
35
Issue
6
Year of publication
1998
Pages
829 - 846
Database
ISI
SICI code
0021-8901(199812)35:6<829:FRFVAF>2.0.ZU;2-Q
Abstract
1. A fire history (1980 to the present) developed for Kakadu National Park, derived mostly from manual interpretation of LANDSAT MultiSpectral Scanner (MSS) imagery, was used as the temporal basis for undertaking rapid assess ment of the effects of fire regimes on floristically diverse vegetation con taining many regionally endemic species, occupying sandstone formations of the Arnhem Plateau, monsoonal northern Australia. 2. Three broad vegetation types were identified through TWINSPAN classifica tion of floristic composition at 108 sample plots: Allosyncarpia-dominated closed forest; open forest/woodland savanna with an annual grass understore y; and open shrubby heath interspersed with perennial hummock grasses. All sampled vegetation occurred on freely draining, oligotrophic substrates der ived from sandstone parent materials; grazing and browsing of sampled veget ation were not significant factors. 3. The three fire frequency parameters derived from the assembled fire hist ory ('years unburnt', 'fire frequency', 'shortest interfire interval') were autocorrelated. Correlative analyses between vegetation structure and soil nutrient variables, and 'years unburnt', demonstrated generally increasing shrub densities, fuel loads and soil fertilities with increasing time sinc e last burnt. Identified limitations in assembled data included: substantia l between-plot structural variation for respective vegetation types, reflec ting diverse ecological and fire histories of individual plots; and inabili ty to describe accurately the important fire regime variable 'fire intensit y' from the imagery used. 4. Assessment of the reproductive maturity of obligate seeder shrub species indicated that, whereas vegetation at some sites could tolerate burning af ter 3-year intervals without loss of floristic diversity, many species requ ired fire-free intervals of more than 5 years, particularly those occupying harsher, rocky sites. In contrast, fine grass and litter fuels were suffic ient to support intense fires in all representative situations under late d ry season climatic conditions within 1-3 years of having been burnt previou sly. 5. Published fire history data for the period 1980-94 show that 40% of vege tation occupying sandstone-derived landforms in Kakadu National Park have b een burnt at frequencies of at least 1 in 3 years; such fire frequencies ev idently cannot be sustained without substantial loss of obligate seeder spe cies, comprising 54% of the sampled shrubby heath flora. The data support o ther observations concerning the catastrophic impact of contemporary lire r egimes on fire-sensitive vegetation types in other sandstone regions of nor thern and north-western Australia. 6. Based on these insights, an established programme for strategic fire man agement of the Arnhem Plateau is outlined.