Proceedings of the conference held 8-9 October 1994, Footscray, Melbourne
Biodiversity Series, Paper No. 8
Department of the Environment, Sport and Territories, 1996
7. Effects of repeated fires on dry sclerophyll (E. sieberi) forests in eastern Tasmania
Mark Neyland and Michael Askey-Doran
Parks and Wildlife Service, Department of Environment and Land Management, Tasmania
Eucalyptus sieberi forests occur on infertile substrates in the low rainfall areas of north-eastern Tasmania. These forests have been subjected to a fire regime of regular repeated fires, at least since European settlement of Tasmania. Floristic, edaphic and fuel load data was collected from a range of sites across the range of E. sieberi.
The E. sieberi forests are currently subject to a fuel reduction burning program with a seven year return cycle. Analysis of the fuel load data suggests that fuels in these forests reach equilibrium between 15 and 20 years after fire. If nutrient cycling and invertebrate populations require similar time intervals to re-establish following fire then the fire return interval of seven years is likely to have detrimental effects on these forests. Unfortunately there are no sites of sufficient age (ie, unburnt for greater than 15 years) available for comparative studies.
In order to gain a better understanding of the long term dynamics of these forests, a series of permanent transects have been established.
Key words: repeated fires, Eucalyptus sieberi, sclerophyll forest, eastern Tasmania.
Since 1970, Forestry Tasmania has advocated the use of fuel reduction burning in the dry forests of eastern Tasmania, on a seven year cycle. In the case of the E. sieberi forests, this has resulted in a regime of frequent regular fires being applied to an area which had been subjected to a similar regime, largely through arson, for many years.
The E. sieberi forests have been intensively used by Europeans since the early days of settlement. The trees split readily and provide durable fenceposts, firewood with excellent burning qualities and high quality sawn timber. Minerals, predominantly gold and tin, were discovered in these areas very early in the history of European settlement and extensive water races were cut through these areas to provide sufficient water for alluvial extraction of the minerals.
This project was established to investigate the impact that repeated fuel reduction burning is having on the E. sieberi forests. In particular the project was designed to address four questions. Does fuel reduction burning reduce fuels and if so for how long? Does fuel reduction burning affect fire behaviour? Does fuel reduction burning change the floristics and if so, how, and does fuel reduction burning affect the structure of the forest and if so, how?
7.2.1 The study area
The Eucalyptus sieberi forests of eastern Tasmania occur in a broken band on the northern east coast (Figure 7.1). The E. sieberi forests occur on three different substrates; Jurassic dolerite, Devonian granite, and fine grained Ordovician sediments colloquially known as Mathinna beds. Only a small area of E. sieberi forest occurs on dolerite and these forests are not considered any further here. The soils derived from the granite and Mathinna bed parent materials within the range of E. sieberi are infertile (Davies & Nielsen 1987). On granite substrates the soils are largely gravel. On Mathinna bed substrates the soils are very fine grained loams with a high gravel content.
Topographic relief on the two different substrates is markedly different. In Mathinna bed areas ridges are sharp and slopes are steep, with deeply incised creeklines. Aspect in these areas has a marked effect on the nature of the vegetation. In the granite areas the relief is considerably more subdued. Slopes are gentle and drainage lines are commonly broad, although the upper reaches of some drainage lines around the few peaks in the area can be steep. In the granite country the effect of aspect is less significant than in the Mathinna bed areas.
The climate on the north east coast is mild with very little difference between the winter and summer maximum temperatures (Gentilli 1972). Rainfall is around 800mm per annum and is erratic. There is no pronounced winter or summer peak, however rainfall is often associated with intense low pressure systems which can produce large amounts of rain in quite short periods (up to 150mm in 48 hours). This is often followed by long periods in which there is very little rainfall.
7.3.1 Floristic sampling
Floristic lists and TASFORHAB structural profiles (Peters 1984) were collected during the establishment of the permanent transects and from all the sites from which fuel loads were sampled. Profiles were also collected opportunistically from other sites which were examined for their suitability for other studies but not used. A set of profiles was also collected from the coastal E. sieberi forests which were not otherwise examined.
Data from 255 TASFORHAB plots were entered into the ecological database, DECODA (Minchin 1990). This package was used to prepare outputs for classification and ordination. DECODA was also used to derive species richness and diversity.
7.3.2 Fuel load sampling
Fuel loads were assessed in detail at 54 sites. Fuels were assessed by taking 15, 0.1 m² samples of the fuel load (samples were collected using a 316mm x 316mm wooden square) at each site. A random start point was established by throwing a note book behind the recorder. From the start point a line transect was established parallel to the contour and fuel samples were taken every ten metres along the transect. The transect was laid out so as to remain wholly within a single vegetation type. If the transect ran into a different type a second line was established parallel to the first and either above or below the original line. At each point the wooden square was laid on the ground with one side parallel to the tape, and all the fine fuel contained within an imaginary column in and up to one metre above the square was collected. Fine fuels are defined as organic material less than 6mm in diameter (Luke & McArthur 1977). Fifteen samples were collected from each site. As fuels throughout the E. sieberi forests are very largely composed of dead material, live and dead fuels were not separated, although both (where both were present) were collected.
The fuel samples were oven dried at 80°C for twenty-four hours. Gravel, rocks and fine material below 0.7mm was removed from the sample which was then weighed.
The fire age of each site was determined both from Forestry Commission records and from on-site sampling. Banksia marginata is reputed to grow in annual increments and counting the nodes is believed to provide a reasonable estimate of the time since the last fire, where there are cohorts of Banksia marginata individuals all having the same node count (Brown & Podger 1982). Most of the study sites have been burnt to reduce fuel within the last ten years and the Banksia ages could be compared to the known last fire date.
The fuel load was modelled using Fensham's (1992) modified form of Walker's (1981) model.
Wt = Wss(1-e-kt) (Walker's (1981) model).
Fensham (1992) modified the model as below to allow for the fact that immediately following fire there is a certain amount of unburnt fuel left on
Wt = Wss(1-e-kt) + 1.92(e-kt) (Fensham's model)
Wt = weight at time t (years after fire)
Wss = a constant equal to the amount of fuel under steady state conditions (tonnes per ha)
k = a constant equal to the proportion of litter that decomposes
t = time since fire (years)
Fensham (1992) tested the goodness of fit of this model versus the unmodified model to his data and found the modified function to perform better for six of eight vegetation types.
Throughout the range of E. sieberi the forests are largely monospecific and remarkably similar. On the driest (north facing) slopes E. sieberi forms pure stands. E. viminalis may also be present on the lowest slopes and on more south facing slopes. E. amygdalina may be present as a subdominant on granite sites. E. sieberi is replaced by E. obliqua on more mesic sites.
The understorey is simple. Large shrubs are rare with the exception of Allocasuarina littoralis which can form dense thickets but which is more often found as scattered shrubs. Small shrubs are invariably sparse. On granite sites the dominant understorey species are Gonocarpus tetragynus, Pteridium esculentum, A. littoralis, Goodenia lanata, Hibbertia empetrifolia, Lepidosperma concavum, L. laterale, Lomatia tinctoria, Epacris impressa, Aotus ericoides and Banksia marginata. On Mathinna bed sites the dominant understorey species are Gonocarpus tetragynus, Pteridium esculentum, A. littoralis, Goodenia lanata, Tetratheca labillardieri, Pultenaea gunnii and Acacia dealbata.
On granite sites, there are 37 species which occur on more than 20 per cent of the sites. Of these species, 29 are vegetative resprouters, (and 20 of the commonest 23 are vegetative resprouters). On Mathinna bed sites, 36 species occur on more than 20 per cent of sites. Of these, 26 are vegetative reproducers, (and all of the ten most abundant species are vegetative resprouters).
7.4.2 Fuel loads
Fensham's model using the whole data set gave an estimate of Wss of 49.16 t/ha, with a k value of 0.033. (R2 of 0.577).
Removing the outlier (top right of figure above), gave an estimate of Wss of 15.17 t/ha, with a k value of 0.225. (R2 of 0.467).
Fox et al. (1979), recalculated the data of Van Loon (1977) and found the steady state fuel load in E. sieberi forests in the Blue Mountains of New South Wales to be 14.8 tonnes per ha which compares well with the second model above.
Analysis of the floristic data indicates that the species which occur in the understorey of E. sieberi forests are predominantly those which are able to reproduce vegetatively.
Species richness varied from 13 (driest north facing slopes on Mathinna bed substrates) to 26 (coastal sites). Similar communities in Victoria (Forbes et al. 1982) have species richness values which range from 33 to 57. Similar communities in Victoria are defined as those in which E. sieberi is present in at least 50 per cent of the quadrats for that community. There is too little information available to make a direct comparison of the proportion of vegetative reproducers to obligate seeders in the Victorian forests; it is possible that the greater floristic richness in the Victorian forests is determined by a number of factors apart from fire history. Nonetheless it is interesting that the Victorian E. sieberi forests show a consistently higher species richness than their Tasmanian counterparts.
7.5.2 Fuel loads
From Figure 7.2, it is apparent that the accumulation of fuel in E. sieberi forests is typical of eucalypt forests in that the fuel load shows a rapid early accumulation rate which then tends to level off. The notable exception, in this case, is the site with an exceptionally high fuel load at the top right of the figure. This site, which has not been burnt for at least fifteen years, has a near-closed understorey dominated by Allocasuarina littoralis. The fuel load is comprised very largely of Allocasuarina needles which in places formed a layer over 15 cm deep. This site was also interesting as it was the only site at which an active and diverse invertebrate fauna was observed.
If it weren't for this site (top right), it would appear from the graph to be a reasonable assumption that fuel loads in E. sieberi forests reach a steady state of around 15 to 20 tonnes/ha at 15 to 20 years after a fire. However the exceptional site also points to the possibility that sites carrying Allocasuarina littoralis can carry much heavier fuel loads. Allocasuarina littoralis is ubiquitous in E. sieberi forests and therefore the possibility of fuel loads exceeding 25 tonnes/ha (in the long term absence of fire) cannot be discounted.
Both Hutson and Veitch (1985) and Birk and Simpson (1980) indicate that deriving decomposition constants for sites which have yet to reach steady state conditions is statistically suspect. The model only derives the curved portion of the fuel accumulation curve, which is then extrapolated forward in time to indicate the steady state fuel load. Small variations in the litterfall rate can lead to larger variations in the steady state fuel load.
In the present case, where the model indicates that steady state fuels are reached between 15 and 20years after fire, but where the collected data is from sites which have predominantly been burnt within the last ten years, the results must be interpreted with caution.
We posed four questions at the start of this paper. Does fuel reduction burning reduce fuels and if so for how long? Does fuel reduction burning affect fire behaviour? Does fuel reduction burning change the floristics and if so, how, and does fuel reduction burning affect the structure of the forest and if so, how?
Fuel reduction burning does reduce fuels. Figure 2 above shows that sites recently burnt have significantly reduced fuels. The figure also shows that fuels can reach 'unmanageable' levels (ie, in excess of 12 tonnes/ha) in as little as five years. So the period of protection offered by fuel reduction burning can be as short as three years, although it is longer than this at many sites. Fuel reduction burning will therefore have an impact on fire behaviour but mainly on sites which have been fuel reduced within the last five years.
The impact of fuel reduction burning on the floristics and structure of these forests is more difficult to quantify. The understoreys throughout the range of E. sieberi are dominated by species which reproduce vegetatively after fire, and species which are obligate seed regenerators are less common and generally restricted to more mesic sites. With an average fire return period of less than ten years, a secondary shrub layer rarely develops, although Allocasuarina littoralis can be prominent in the understorey. There are four known sites all of less than 1 ha in extent which have a more diverse understorey which includes a number of less common species such as Eriostemon virgatus and Leucopogon ericoides, which are obligate seed regenerators. If it is the case that these sites are indicative of the less frequently burnt environment then the floristics and structure of the E. sieberi forests have been greatly simplified by the past fire regime. However it is not possible to accurately quantify the fire history of any of the sites in the northeast so such statements need to be read with caution.
The fuel load data indicate that steady state fuel loads are reached around 15 to 20 years after fire. If nutrient cycling and invertebrate populations require similar time intervals to re-establish after fire then a fire return interval of seven years is likely to have a detrimental effect on these forests. Studies on nutrient cycling in the E. sieberi forests have not been undertaken and studies on invertebrate populations are in their infancy. Further work in these areas will greatly benefit our understanding and management of these forests.
Because of the difficulty of assessing the impact of the past fire regime on the vegetation of the north-east, a series of permanent transects, recording the present nature of the vegetation, have been established in the northeast and will be monitored through time. These transects will hopefully demonstrate the future change to the vegetation of this area.
Hypothesising about the nature and impact of the past fire regime is interesting but will always be speculative, because we cannot be certain of the regimes which prevailed prior to European settlement. The more pertinent question that needs to be addressed is how do we wish to manage these forests now? Fuel loads, particularly around identified assets such as plantations and towns, clearly must be managed in such a way as to provide a point of control of wildfires. But the imposition of a blanket fire regime over the whole range of E. sieberi in Tasmania has led to the creation of a large area with very similar and simple vegetation. By applying a regime of frequent low intensity fires to specified areas for fire management purposes and also allowing other areas to remain unburnt for varying periods of time will create a more diverse regional vegetation pattern. Long term in situ monitoring is essential to follow changes in the vegetation and fire free control areas also need to be established.
This research was supported by a research grant from the Tasmanian Forest Research Council. Our thanks to Dr Mick Brown (Forestry Tasmania) and Jayne Balmer (Parks and Wildlife Service, Tasmania) for their comments on an earlier draft of this manuscript.
Birk, E. M. & Simpson, R. W. 1980, 'Steady state and the continuous input model of litter accumulation and decomposition in Australian eucalypt forests.', Forest Ecology and Management, vol. 29, pp. 59-79.
Brown, M. J. & Podger, F. D. 1982, 'Floristics and fire regimes of a vegetation sequence from sedgeland-heath to rainforest at Bathurst Harbour, Tasmania.', Australian Journal of Botany, vol. 30, pp. 659-676.
Davies, J. B. & Nielsen, W. A. 1987, 'Granite soils in State forests – N. E. Tasmania. A reconnaissance survey with particular reference to erosion problems associated with forestry operations,' Forestry Commission, Tasmania.
Fensham, R. J. 1992, 'The management implications of fine fuel dynamics in bushlands surrounding Hobart, Tasmania,' Journal of Environmental Management, vol. 36, pp. 301-320.
Forbes, S. J., Walsh, N. G. & Gullan, P. K. 1982, 'Vegetation of East Gippsland,' Muelleria, vol. 5, pp. 53-113.
Fox, B. J., Fox, M. D. & McKay, G. M. 1979, 'Litter accumulation after fire in a eucalypt forest,' Australian Journal of Botany, vol. 27, pp. 157-165.
Gentilli, J. 1972, Australian Climate Patterns, Nelson, Melbourne.
Hutson, B. R. & Veitch, L. G. 1985, 'Relationships between litterfall rate, litter mass and decomposition rate in Eucalyptus forests in south-eastern Australia', Australian Journal of Ecology, vol. 10, pp. 443-450.
Luke, R. H. & McArthur, A. G. 1977, Bushfires in Australia, Australian Government Publishing Service, Canberra.
Minchin, P. R. 1990, DECODA: Database for Ecological Community Data. Australian National University, Canberra.
Peters, D. G. 1984, 'TASFORHAB,' in Survey Methods for Nature Conservation, Proceedings of a workshop at Adelaide University 31 Aug. – 2 Sept. 1983. C.S.I.R.O. Div. of Water and Land Resources. Vol. 2, pp. 47-66.
Van Loon, A. P. 1977, 'Bushland fuel quantities in the Blue Mountains, litter and understorey', Forestry Commission N.S.W. Research Note No. 33.
Walker, J. 1981, 'Fuel dynamics in Australian vegetation', in Fire and the Australian Biota. Eds, A. M. Gill, R. H. Groves & I. R. Noble, Australian Academy of Science, Canberra.