An exploration of the capacity of a database generated by post-fire vegetation monitoring to yield useful information about responses of individual species to fire.
There have been two major fires in Wilsons Promontory National Park (the Prom) this century. The first was in 2005 after an ecological burn near Tidal River escaped its planned boundaries. The fire was fanned by a hot, northerly wind and burned 6 500 ha of the southern half of the Prom in less than 24 hours, extending from Tidal River to the Lighthouse. The fire formed a mosaic of intensities, ranging from low to high with occasional areas remaining unburnt. The second fire occurred in 2009 and was started by a lightning strike in the Cathedral Range. Fanned by easterly winds, it travelled north-west over four weeks and finally died out about 7 km from the Yanakie Entrance to the Prom. This fire was less intense than that of 2005, but also formed a mosaic of intensities.
Following the first fire, a group of volunteers, the Prom'n'aides, worked under supervision of Parks Victoria rangers to monitor recovery of the Coastal Grassy Woodland, previously identified as being in moderate to poor condition (Parks Victoria 2017). Two Ecological Vegetation Classes or EVCs (Davies et al. 2002) from this ecosystem type were selected for study due to their generally poor condition and the invasion of Coast Tea-tree Leptospermum laevigatum (Burrows 2007; Ellis 2013). These EVCs were Coast Banksia Woodland (EVC 2) and Coastal Dune Scrub Mosaic (EVC 1). All data were entered on a Parks Victoria database (PROM_VegPlot Data).
In following years, the vegetation monitoring was extended to include small mammal trapping sites located throughout the Prom in other EVCs. Coincidentally, the 2009 fire affected some of these sites, providing the opportunity to add pre- and post-fire data from these EVCs to the study. The other area added to the study was the site of a controlled ecological burn in 2002, burnt again in the 2009 fire. This was a small area north of Darby Saddle where Saw Banksia B. serrata was the canopy species in Banksia Woodland (EVC 14).
Factors that affected the variability of the data were (1) the time it took to collect data at four quadrats each visit; (2) the distance to travel to sites, resulting in work on quadrats in the same area (it was not possible to visit each area every year nor always in the same season); and (3) the addition of new sites as the project progressed without continuous scientific oversight, making the size of the total project unwieldy.
After 10 years, monitoring was discontinued because many of the sites had become impenetrable thickets of understorey species such as Coast Tea-tree and Coast Wattle Acacia longifolia subsp. sophorae. It was intended that the information in the database, which included data on all species in the locations studied, would be useful for the future management of the Prom with respect to the appropriate frequency of controlled burns.
To explore the usefulness of the database, data were extracted for six species to examine their response to fire: Coast Banksia Banksia integrifolia; Saw Banksia B. serrata; Large-leaf Bushpea Pultenaea daphnoides; Soft Bush-pea P. mollis; Cut-leaf Xanthosia Xanthosia dissecta and Hill Xanthosia X. tridentata.
Data were extracted for two species within each of the three genera to enable comparison within each genus. Coast Banksia was of interest as it was in serious decline for unknown reasons, while Saw Banksia was doing well. Our field observations and botanical knowledge of the Bush-peas and Xanthosias suggested they were significant plants. They were widespread throughout the Prom, resulting in reasonable sample sizes, and they represented understorey and ground cover plants respectively.
The database was generated from field data recorded between 2006 and 2015. More than 100 permanent 20 m x 20 m quadrats were established in various locations in the Prom. In each of these, five 1 m x 1 m quadrats were established, one in each corner and one in the centre, and all species present within the smaller quadrats were recorded. During the initial data collection, post-fire height was measured for both living and dead canopy plants in the 20 m x 20 m quadrats. Dead plants were not measured in subsequent years. The data recorded for both understorey and ground cover species in 1 m x 1 m quadrats were: number of plants/stems; per cent cover; and minimum and maximum height (Burrows 2007).
Variation occurred in the measure of per cent cover for plants. From 2006 to July 2012, per cent cover was recorded using the Braun-Blanquet scale (Table 1). For a brief time, from August 2012 to January 2013, it was recorded as a percentage. Then the Domin scale (Table 1) was used until the work was discontinued in January 2015. The earlier data used in this study was adjusted so that per cent cover was converted to the Domin scale. In order to under- rather than over-estimate per cent cover, all conversions to the Domin scale were to the lower value shown in Table 1.
Tolsma and Cheal (2012) recommended recording per cent cover only for ground cover species. This practice was implemented after 2012.
The database did not indicate the absence of canopy in quadrats where there was no canopy. These quadrats were either on heathland, where the vegetation did not grow above 1.5 m, or where the canopy had been destroyed by fire. This absence should have been recorded as zero. After 2012, the cover provided by the understorey and canopy was measured as a single measurement of cover and zero was recorded where appropriate.
Habitats and data used
Coast Banksia occurs in the EVCs Coast Dune Scrub Mosaic (EVC 1), Coastal Banksia Woodland (EVC 2) and Riparian Scrub Complex (EVC 17). Saw Banksia occurs in the EVCs Banksia Woodland (EVC 14), Sand Heathland (EVC 6), Sand/Wet Heathland Mosaic (EVC 307) and Heathy Woodland (EVC 48). The Banksias are the dominant canopy species in these EVCs. For the Banksias in the 20 m x 20 m quadrats, the proportion of dead plants and live plants, the per cent cover and minimum and maximum heights were used to show the quality of the canopy. For the canopy of Coast Banksia, where some trees were dead, the proportion of dead trees in unburnt 20 m x 20 m quadrats indicated the amount of stress a species was under pre-fire and could be compared with the proportion that were dead after fire. Data from the 1 m x 1 m quadrats were used to capture any regeneration by seedlings or root suckers.
The EVCs where Bush-peas and Xanthosias occur are more widespread and variable; they include Damp Forest (EVC 29) and Lowland Forest (EVC 16), Sand (EVC 6) and Wet (EVC 7) Heathlands and are found in the same EVCs as the Banksias. For these species, the number of plants, per cent cover and minimum and maximum heights were used to determine the response of seedling regeneration post-fire and the longer-term recovery in size to maturity.
Data gathered for Coast Banksia (Fig. 1) in 2006/7, the first season after the 2005 fire, compared burnt and unburnt quadrats (Table 2). Fourteen of 20 (70%) trees were dead in burnt quadrats, while 32 of 59 (54%) were dead in unburnt quadrats. The proportion of dead compared to living trees in unburnt quadrats suggests this species was under threat before the fire. Not surprisingly, fire increased the proportion of dead trees. In Tables 3 to 8, data for burnt and unburnt quadrats were combined.
Table 3 shows that in 2006/7 60 per cent of the mature Coast Banksias were dead. Over the next few years, many fell. By 2014, the regrowth of the understorey in Coast Banksia quadrats had made the sites impenetrable and unsafe for volunteers.
Within six of the 1 m x 1 m quadrats (Table 4), a dead trunk occupied the whole quadrat. For this reason, the data from those six quadrats are shown separately in the top line of the table. In all other quadrats, living Coastal Banksia plants were seedlings. There was little recruitment in the four to eight years after the fire and comparatively little change in size of plants or percent cover, suggesting that the decline of Coast Banksia was continuing.
Saw Banksia (Fig. 2) occurs in three EVCs in the northern section of the Prom: Sand Heath-land, Sand Heathland/Wet Heathland Mosaic and Heathy Woodland. It was not recorded after the 2005 fire in the southern section of the Prom. The canopy data (Table 5) in 2008 represent the sixth year post the 2002 controlled burn and the data in 2015 represents six years post the 2009 fire. The data from both are comparable. Unlike Coast Banksia, Saw Banksia regenerated well. There were no dead trees.
Table 6 shows there was an increase in recruitment three or four years after the fire, possibly a response to the flood in 2011. From the height data, it is assumed that many of the seedlings did not survive the hot dry summers in the following years.
For shrubs and ground cover species, only data from 1 m x 1 m quadrats were considered.
The data for Bush-pea species is shown in Tables 7 and 8. (See also Figs 3 and 4 for illustrations of the species). In the first year post-fire, there was recruitment of seedlings, but the number of stems decreased thereafter. By five years after fire, both species had reached their full height (1-3 m) and there was 5-10 percent cover (D=4 or 5). Survival of seedlings may have been affected by competition, hot dry summers and grazing of these species.
The number of plants, per cent cover and growth recorded for Cut-leaf Xanthosia in burnt areas (Table 9) appear to be very similar to those that were in unburnt areas (Table 10), although there were fewer quadrats in the latter.
Hill Xanthosia occurred much more frequently and in more EVCs than Cut-leaf Xanthosia, but like Cut-leaf Xanthosia showed no particular effects due to fire (Tables 11 and 12), although recruitment was greatest three to five years after the fire.
Post-fire response of species
The Coast Banksias around Oberon Bay were severely diminished by the 2005 fire. This whole area had been heavily invaded by Coast Tea-tree with the result that the fire was particularly hot at canopy height. Had it been a grassland fire, the Banksias would have stood a greater chance of surviving (Jim Whelan, pers. comm., October 2017). Nearly 500 mature Coast Banksias were burnt and many died (Table 2). The size of the dead trunks, all of which were measured in the first twelve months after the fire, gives an indication of how many large trees a healthy ecosystem had supported. Regeneration was slow and limited, with few seedlings.
It is probable that the trees were already under stress before the fire. Approximately 54% of the trees were dead in the unburnt quadrats while 70% were dead in the burnt quadrats (Table 2). Populations of Coast Banksia have been subject to die-back in many locations from the Mornington Peninsula to the Prom. Among possible causes are borer attack, drought stress and pollution (Barraclough 2009). Further studies are underway at the Prom under the supervision of Jim Whelan, Ecological Restoration Project Officer Yanakie Isthmus Grasslands Restoration. The Prom'n'aides are working with him on this project where appropriate.
By contrast, after the 2009 fire, Saw Banksias regenerated well, and the EVCs where they occur appear to be in vigorous health with mature trees and a dense species-rich understorey, including Soft Bush-pea. This fire was not as severe as the 2005 fire, being slow burning from the Cathedral Range to the north of the Prom over a period of four weeks. The 2005 fire burnt from Tidal River to the Lighthouse in a day.
Bush-pea species are prominent in most habitats in southern Australia with reliable rainfall (Elliot and Jones 2002). They can dominate the vegetation as they germinate en masse following a fire.
This was seen dramatically in the spring following the fires, when the southern slopes of Darby Saddle Track were coloured reddish-brown with the flowers of Large-leaf Bush-pea, and in late winter the edge of Five Mile Track was yellow with the flowers of Soft Bush-pea.
The role of Xanthosias in stabilising the soil could not be determined in this study of revegetation after fire, but the two species included in this report recovered within two years and were still present up to nine years post-fire. They do not appear to rely on fire for germination (Tables 9, 10, 11 and 12).
Usefulness of the database
There are only a few long-term studies of regeneration of vegetation after fires. One that was organised and sustained by volunteers was conducted after the 1983 wildfire in Anglesea by Wark (2000). The methodology and analysis were different from those used in the Prom project and they benefited from having one person managing the collection of data, analysis and reporting. The Anglesea project confirmed that the vegetation of the local area was both extremely diverse and resilient. Wark (2000) reported similar key findings to the Prom project; most plants regenerated by regrowth, some from seed only and some by both regrowth and from seed, and structural recovery occurred in seven to ten years.
The Prom project methodology was designed by a visiting Canadian Ranger. After his return to Canada, a number of local rangers assisted, with one taking charge for several years. For the final two years there were again several rangers involved, highlighting the lack of continuous scientific supervision. The implementation of recommendations made in 2012 (Tolsma and Cheal 2012) following the analysis of part of the data created inconsistencies in the database. Some of their other recommendations were impractical to implement because of the factors affecting variability in data collection mentioned earlier. This highlighted the disconnection between the field and the desk. Long-term projects need scientists to design the collection and analysis of data, funding, expert local naturalists and a pool of volunteers.
However, a database such as the one that eventuated at the Prom is valuable. It forms not only a baseline record of the state of a species but also could be used to provide information about vegetation communities, species composition following fire, the likely long-term effects of a fire and as a comparison or support of the work of others. For example, Morgan and Nield (2011: 59) investigated the effects of fire (2005) severity on B. integrifolia in coastal plant communities at the Prom and found:
The widespread senescence of Banksia integrifolia stands, the high levels of fire-induced tree mortality, low and patchy seedling recruitment after fire, as well as high density of Leptospermum laevigatum seedlings in woodlands after fire, all point towards a complete state change of burnt Banksia woodlands at Wilsons Promontory.
Our data showed that B. integrifolia was under stress before the fire (54% dead trees) (Table 2); that post-fire the majority of trees were dead (70%) (Table 2); and that surviving trees were deteriorating, e.g. a general decrease in minimum and maximum height (Table 3), that seedling regeneration was not high and that seedlings were not doing well (Table 4). Thus, in spite of changes in supervision and methodology of the project and variation in quadrat numbers, our longer term data parallels the findings of Morgan and Nield (2011) and supports their prediction that stand-level changes in dominance are likely to occur in B. integrifolia woodlands.
This database recorded response to two fires. It highlighted the vulnerability of one species, Coast Banksia. It may be useful to correlate vegetation composition with small mammal populations. It may provide information for use in future management including controlled burns. The results presented here represent a minute portion of the data held in the database. Further analysis of this extensive database may yield more information about vegetation communities and species composition following fire.
We thank Parks Victoria for supporting this long-term project, especially the rangers who took us out in the field.
The Prom'n'aides volunteered six hours a day, once a month for ten years, generally 20-25 people each day totalling more than 15 000 hours. There were more than 60 volunteers over that time, all of whom felt they were privileged participants. This report was written on their behalf.
We are very grateful to Jim Whelan, Terri Allen, Jenny Lawson and Ros Gibson for their helpful comments on drafts of this manuscript. We had a very insightful discussion with an editor of The Victorian Naturalist, and thank her for her input.
Barraclough L (2009) What's Happened to Our Banksias? The Creek 13, 3.
Burrows FG (2007) Wilson's Promontory National Park: Post-fire Integrated Monitoring Vegetation Protocol. Final Report to Parks Victoria.
Davies JB, Oates AM and Trumball-Ward AV (2002) Ecological Vegetation Class Mapping at 1:25 000 in Gippsland. Final Report. Department of Natural Resources and Environment, Victoria.
Elliot WR and Jones DL (2002, 2010) Encyclopaedia of Australian plants suitable for cultivation. Volumes Eight and Nine. (Lothian Books: South Melbourne)
Ellis M (2013) Pre- and post-fire vegetation monitoring Wilsons Promontory National Park from 2006 to 2013 by volunteers known as the Prom'n'aides. Final Report to Parks Victoria.
Morgan JW and Nield C (2011) Contrasting effects of fire severity on regeneration of the dominant woody species in two coastal plant communities at Wilsons Promontory, Victoria. Cunninghamia 12, 53-60.
Parks Victoria (2017) Conservation Action Plan for Parks and Reserves Managed by Parks Victoria. Wilsons Promontory. Parks Victoria, Melbourne.
Tolsma A and Cheal D (2012) An analysis of post-fire monitoring--Wilsons Promontory National Park. Arthur Rylah Institute for Environmental Research. Department of Sustainability and Environment, Melbourne.
Wark MC (2000) After the 1983 Wildfire: the Anglesea Vegetation Regeneration Project--How it Grew. The Victorian Naturalist 117, 96-106.
Received 19 October 2017; accepted 19 April 2019
Mary Ellis (*), Lorraine Norden and Margaret Rowe
(*) Corresponding author: email@example.com
Table 1. Per cent cover (%) as Braun-Blanquet or Domin scales. Braun-Blanquet % Domin % + 1 +,1,2 1 1 2-5 3 1-4 4 5-9 2 6-24 5 10-24 6 25-32 3 25-49 7 33-49 4 50-74 8 50-74 5 75-100 9 75-94 10 >95 Table 2. Data for canopy of Coast Banksia in burnt and unburnt 20 m x 20 m quadrats where this species was recorded (n) and the mean of each measurement (N: number of plants; D: Domin scale; Min: minimum height; Max: maximum height). Per cent cover of dead trees was not recorded. Year n N D Min (m) Max (m) Burnt 2006/7--dead trees 18 14 4.5 9.3 2006/7--living trees 7 6 4 6.3 13 Unburnt 2006/7--dead trees 5 32 3.7 11.2 2006/7--living trees 6 27 5 5.7 14.2 Table 3. Data for canopy of Coast Banksia showing number of 20 m x 20 m quadrats where this species was recorded (n) and the mean of each measurement (N: number of trees; D: Domin scale; Min: minimum height; Max: maximum height) and the number of years post-fire. The number and heights of dead trees were recorded, but not per cent cover. Year n N D Min (m) Max (m) Years post-fire 2006/7--dead trees 23 20 5 11.3 1 2006/7--living trees 13 13 5 6 13.5 1 2008--living trees 2 2 1 1.75 2.25 3 2009--living trees 7 8 3 8.7 13 4 2010--dead trees 2 32 5 2010--living trees 2 4 1 3.6 4.4 5 2011--living trees 8 11 3 5.3 10.7 6 2012--living trees 4 3 2 3.2 4.7 7 2013--dead trees 17 12 3.9 11.4 8 2013--living trees 7 9 3 5.4 11.4 8 Table 4. Data for 1 m x 1 m quadrats of Coast Banksia showing the number of quadrats where this species was recorded (n) and the mean of each measurement (N: number of plants; D: Domin scale; Min: minimum height; Max: maximum height) and the number of years post-fire. Year n N D Min (mm) Max (mm) Years post-fire 2006/7--dead trees 6 1 9 1240 1240 1 post 2005 2006/7--living trees 11 2 3 55 61 1 post 2005 2009--living trees 4 1 1 55 55 4 post 2005 2010--living trees 2 1 1 165 165 5 post 2005 2011--living trees 3 1 1 160 160 6 post 2005 2013--living trees 2 1 1 60 60 8 post 2005 Table 5. Data for canopy of living Saw Banksia showing the number of quadrats (n) where this species was recorded and the mean of each measurement (N number of plants, D Domin scale, Min minimum height, Max maximum height) and the number of years post-fire. Year n N D Min (m) Max (m) Years post-fire 2008 7 7 3 3.6 5.3 6 post 2002 2010 2 7 1 3.3 5.5 1 post 2009 2011 3 3 4 4.3 5.6 2 post 2009 2012 1 35 5 0.5 6 3 post 2009 2013 2 48 4 1 5.5 4 post 2009 2014 4 4 2 1.5 5.6 5 post 2009 2015 2 7 2 0.4 5 6 post 2009 Table 6. Data for 1 m x 1 m quadrats for Saw Banksia showing the number of quadrats where this species was recorded (n) and the mean of each measurement (N: number of plants; D: Domin scale; Min: minimum height; Max: maximum height) and the number of years post-fire. Year n N D Min (mm) Max (mm) Years post-fire 2008 4 2 2 175 225 6 post 2002 2010 7 2 2 256 864 1 post 2009 2011 3 1 1 167 167 2 post 2009 2012 9 1 1 178 181 3 post 2009 2013 5 1 4 1590 1590 4 post 2009 2014 3 1 2 300 310 5 post 2009 2015 4 1 6 2627 2627 6 post 2009 Table 7. Data for Large-leaf Bush-pea showing the number of quadrats where this species was recorded (n) and the mean of each measurement (N: number of plants; D: Domin scale; Min: minimum height; Max: maximum height) and the number of years post-fire. Year n N D Min (mm) Max (mm) Years post-fire 2008 18 4 3 626 984 3 post 2005 2010 22 13 3 200 449 1 post 2009 2012 1 4 1 230 710 3 post 2009 2013 16 9 4 1081 1851 4 post 2009 2014 1 2 1 340 440 5 post 2009 2015 7 1 3 1510 1510 6 post 2009 Table 8. Data for Soft Bush-pea showing the number of quadrats where this species was recorded (n) and the mean of each measurement (N: number of plants; D: Domin scale; Min: minimum height; Max: maximum height) and the number of years post-fire. Year n N D Min (mm) Max (mm) Years post-fire 2010 2 11 1 840 1325 5 post 2005 2008 1 1 + 280 280 pre 2009 fire 2010 23 17 2 156 399 1 post 2009 2012 17 12 3 424 1079 3 post 2009 2014 17 10 4 519 1188 5 post 2009 2015 2 3 5 1500 2250 6 post 2009 Table 9. Data for Cut-leaf Xanthosia in quadrats burnt in either the 2005 or the 2009 fire showing the number of quadrats where this species was recorded (n), the mean of each measurement (N: number of plants; D: Domin scale; Min: minimum height; Max: maximum height) and the years post-fire. Year n N D Min (mm) Max (mm) Years post-fire 2006 1 1 + 60 60 1 post 2005 2007 8 8 2 33 55 2 post 2005 2009 1 2 1 20 20 4 post 2005 2010 18 7 1 40 61 1 post 2009 2011 7 2 1 46 56 2 post 2009 2012 4 1 3 post 2009 2013 3 1 4 post 2009 2014 6 1 5 post 2009 Table 10. Data for Cut-leaf Xanthosia in unburnt quadrats showing the number of quadrats where recorded (n), the mean of each measurement (N: number of plants; D; Domin scale; Min: minimum height and Max maximum height). Year n N D Min (mm) Max (mm) 2006 1 1 1 60 60 2007 1 1 1 50 50 2008 2 2 1 35 40 2009 3 2 1 35 37 Table 11. Data for Hill Xanthosia in quadrats burnt in 2005 showing the number of quadrats where recorded (n), the mean of each measurement (N: number of stems; D: Domin scale; Min: minimum height; Max: maximum height) and the years post-fire. Year n N D Min (mm) Max (mm) Years post-fire 2006 2 2 1 35 35 1 2007 8 3 1 37 106 2 2008 4 3 1 95 105 3 2009 1 2 1 100 120 4 2010 4 18 2 73 99 5 2011 10 12 1 119 202 6 2012 64 3 7 2013 2 1 8 2014 44 2 9 Table 12. Data for Hill Xanthosia in quadrats unburnt in 2005 showing the number of quadrats where recorded (n), the mean of each measurement (N: number of stems; D: Domin scale; Min: minimum height; Max: maximum height). Year n N D Min (mm) Max (mm) 2007 1 1 1 30 30 2008 17 7 2 72 199 2010 2 1 1 50 50
|Printer friendly Cite/link Email Feedback|
|Author:||Ellis, Mary; Norden, Lorraine; Rowe, Margaret|
|Publication:||The Victorian Naturalist|
|Date:||Jun 1, 2019|
|Previous Article:||Higher rates of non-breeding territory occupancy of urban compared with rural Masked Lapwings Vanellus miles on Phillip Island, Victoria.|
|Next Article:||Ninety-nine Years Ago: THE BLACK SNAIL.|