Shrub recruitment response to intensity and season of fire in a semi-arid woodland

Summary

Prescribed fire is a management strategy recommended for control of shrubs where they reach undesirably high densities in semi-arid woodlands of Australia. Here, the results of an experiment designed to establish the effect of fire on the pattern and level of shrub seedling recruitment in a semi-arid woodland are summarised. The specific aims were (i) to determine which species demonstrated increased germination following fire, (ii) to quantify relationships between fire-line intensity and the density and frequency of shrub recruits, and (iii) to determine the influence of season of fire on these relationships.

Study site The study was undertaken in semi-arid woodland on the 'Oakvale' pastoral property 'Oakvale' (146º30'E, 30°55'S), 70 km northwest of Nyngan, New South Wales, eastern Australia.

Vegetation: The woodland was dominated by two eucalypt tree species, gum collibah Eucalyptus intertexta (on ridges) and bimble box E. populnea (on slopes and flats). Average tree density was 35/ha with a canopy covering 13% of the area. Beneath the tress were shrub 'thickets' dominated by Eremophila mitchellii and Geijera parvzjlora, with 'inter-thicket' areas dominated by Acacia aneura, A.excelsa and Cassia eremophila. Eremophila sturtii and Doclonaea viscose were also common throughout. Shrub density was approximately 4,800 shrubs/ha. Ground flora comprised annual grasses and forbs.

Experimental fires: An area of 108 ha was fenced to exclude sheep, goats and pigs, kangaroos could still enter. Two areas of E.intertexta and two of E.populnea were delineated, within each, 24, 15 x 21 m plots were established. A plot included a tree at one end with associated 'thicket' and an adjacent 'inter-thicket' at the other end. Selection of trees and plot treatment allocation was random. Treatments were: burnt with a low or high amount of fuel or not burnt, on one of eight occasions (approximately 3 months apart) beginning in October 1977. Before each fire, the amount of litter was estimated within plots using a dry-weight rank method.

Wheat straw was spread over each plot and to 3 m outside, at 300 or 800 g/m², to supplement the low quantity of natural plant litter fuel. Prior to burning, c. 250 g of the surface 0-1 cm of soil (in which most shrub seeds are located) was removed from four random locations within each 'thicket' and 'inter-thicket' zone, bulked separately, mixed, and water content determined. Plots were ignited at one end with a drip torch so that the prevailing wind carried a head-fire through each plot. During fires the following were recorded: fire velocities; temperatures before and after; and wind speed during fires. The fire-line intensity for each plot was calculated

Post-fire measurements: At 3-6 month intervals after the fires, the number of shrub seedlings of each species were counted in plots. In plots burnt in February 1978, all C.eremophila seedlings that had survived 2 years post-burn were numbered with aluminium tags on wire stakes next to the seedlings. Their survival was periodically recorded. Seedling mortality is high initially, especially during summer months, therefore the number that survived through to the winter or spring in the year following fire were compared for analysis of treatment effects. Daily rainfall was recorded during the study.

Fire velocity and intensity: Fire-line velocities and intensities differed widely between occasions and sometimes between plots on the same burn date. The effect of fuel load on velocity of fires differed according to season. Fires in cool conditions had similar velocities in both low and high fuel plots. Velocities were doubled by the addition of extra fuel in fires lit in warm to hot conditions (both summers, one spring and one autumn)

There was a 17-fold difference in average fire-line intensities of the plots overall and a 36-fold difference between the lowest (341 kW/m) and highest (12300 kW/m) intensities.

During winter, and sometimes during spring and autumn months when air temperatures were lower and relative humidity higher, fire-line velocities were lower and similar for both low and high-fuel levels. Peak velocities occurred in summer (January 1979) when air temperatures were 31-37 ºC, relative humidities 20-32% with wind speeds of 3-9 km/h.

Species response: Shrub recruitment was strongly influenced by fire-line intensity; the proportion of 9-m² quadrats containing seedlings increased linearly with fire-line intensity from 0% for some controls to 100% when intensity reached c. 3,000 kW/m (depending on season). Seeds of only two of the seven dominant shrub species showed enhanced germination after fire: C.eremophila and D.viscose. C.eremophila was the most common species comprising 66-100% of seedlings in plots. The mortality of a fire-enhanced cohort of C.eremophila decreased exponentially over time.

Of the total of 21,048 seedlings initially recorded, numbers of seedlings of the five other species were very low: E.longijolia (10); M.deserti (46); A.aneura (5), E.mitchelli (10) and E.sturtii (5).

Rainfall: Post-fire rainfall was highly variable from month to month, varying from 0 to 158 mm/month. Less rain fell in 1979 and 1980 than in 1978.

Conclusions: Spring, summer and autumn fires enhanced germination more than winter ones, with the amount and time of post-fire rainfall influencing shrub seedling mortality and establishment. Results from this study suggest that the density and frequency of fire-recruited shrubs could be managed by selecting the season of, and environmental conditions for, prescribed fires.


Note: If using or referring to this published study, please read and quote the original paper, this can be viewed at:

http://links.jstor.org/sici?sici=0021-8901%28199104%2928%3A1%3C60%3ASRRTIA%3E2.0.CO%3B2-R