Soil seed bank dynamics in post-fire heathland ... - Springer Link

2 downloads 0 Views 204KB Size Report
Jul 21, 2006 - capacity to persist in the seed bank over time, leading to changes in seed bank composition and propagules available for post-fire colonisation.
Plant Ecol (2007) 190:1–12 DOI 10.1007/s11258-006-9186-4

ORIGINAL PAPER

Soil seed bank dynamics in post-fire heathland succession in south-eastern Australia T. J. Wills Æ J. Read

Received: 23 November 2005 / Accepted: 20 June 2006 / Published online: 21 July 2006  Springer Science + Business Media B.V. 2006

Abstract Soil seed banks can exert a strong influence on the path of vegetation succession following fire, with species varying in their capacity to persist in the seed bank over time, leading to changes in seed bank composition and propagules available for post-fire colonisation. This study examined the effect of time since fire on soil seed bank dynamics in a chronosequence of seven sites spanning 26 years in a south-eastern Australian sand heathland. No significant change was evident in the species richness and density of the germinable soil seed bank, but species composition differed significantly among young (0–6 years since fire), intermediate (10–17 years since fire) and old-aged (24–26 years since fire) sites (using presence/absence data). No significant trend was observed in the similarity between the extant vegetation and the soil seed bank with time since fire. A total of 32% of the species recorded in the soil seed bank were not present in the above-ground vegetation at the same site, which suggests that species requiring fire for germination may be present in the seed bank. Most species present in the extant vegetation were not recorded (63%) or were in very low abundances in the soil seed bank (29%). The mode of T. J. Wills (&) Æ J. Read School of Biological Sciences, Monash University, Clayton, VA 3800, Australia e-mail: [email protected]

regeneration appears to be the major determinant of species absence in the soil seed bank, as 66% of species occurring in the extant vegetation but not in the seed bank have the capacity to regenerate by resprouting. This study shows that a major shift in the successional pathway after fire due to altered seed bank composition is unlikely in this vegetation; most species not recorded in the seed bank are either resprouters (obligate or facultative) or serotinous, suggesting that they will readily regenerate following fire. Unless fire frequencies are high and kill fire-sensitive obligate seeders before they reach maturity, the chance that the soil seed bank could substantially alter vegetation composition within the study area after fire is low. However, it is unclear how successional pathways may alter in response to severe fires with the potential to kill both seeders and resprouters. Keywords Chronosequence Æ Fire Æ Germination trial Æ Resprouting Æ Species composition Æ Species richness

Introduction Soil seed banks are a highly important functional element of almost all ecosystems. Along with resprouting (Pausas 2001), propagule dispersal (Whelan 1986) and canopy seed storage (Lamont

123

2

Plant Ecol (2007) 190:1–12

et al. 1991), soil seed banks provide a means by which species can reproduce and maintain populations at a given site (Parker et al. 1989). However, the persistence of individual species may be strongly influenced by variability in disturbance regimes, which often leads to changes in community composition (Keith 1996). This may subsequently influence the path of vegetation succession. Pattern and change in the composition of soil seed banks over successional time has been well documented, with Milberg (1995) citing 25 papers between 1893 and 1993 that address this topic using the chronosequence approach. A general finding is that seed bank density and species richness usually decrease or remain relatively unchanged as time since disturbance (or abandonment) increases (Milberg 1995; Bakker et al. 1996; Dalling and Denslow 1998; Ne’eman and Izhaki 1999). However, this pattern may not be universal (Bekker et al. 1999). Because the majority of studies examining soil seed bank successions are from oldfields in the Northern Hemisphere or relatively fertile and/or infrequently disturbed environments, the general trends reported may not be widely applicable to the fire-prone, mediterranean-type heathlands of southern Australia and South Africa, where disturbance may have a marked effect on species composition (Groves and Specht 1981; Kruger 1983). Therefore, this study was undertaken to examine the soil seed bank dynamics of a southeastern Australian heathland, using the chronosequence approach at seven sites of varying time since fire. The aim was to determine the role of the soil seed bank in vegetation succession, and whether differences occurred among young, intermediate and old-aged sites in compositional similarity between the soil seed bank and the extant vegetation. Three predictions were made relating to effects of time since fire on the soil seed bank: 1.

Overall species richness of the soil seed bank was predicted to increase and then decrease with time since fire. This prediction is based on the assumption that richness will be relatively low in the first few years after fire, as

123

2.

3.

most soil-stored seed should be either killed or stimulated to germinate by the fire, and seed production will be low as most species will still be reproductively immature. As species reach reproductive maturity, seed input to the soil increases, and consequently, richness is predicted to increase. As species with transient (short-lived) seed banks die out, or become competitively excluded, seed bank richness is predicted to gradually decline with time since fire. Total density of soil-stored seed was predicted to increase and then decrease with time since fire, as suggested by Whelan et al. (2002). An initial increase was predicted, as the number of reproductively mature species and the number of years they have been reproductively mature increases. The longevity of reproductive maturity is particularly important as this hypothetically facilitates increased soil seed densities, especially in species with long-lived (persistent) seed banks, which appear to be relatively common in heathlands (Keeley 1995). As total vegetative cover begins to decrease ca. 25 years after fire (Wills 2002), total soil seed bank density is also predicted to decline (although possibly with a delay), given the potentially reduced seed source. Species composition of the soil seed bank was predicted to differ between young (0–6 years), intermediate (10–17 years) and old (24–26 years) sites. However, there may be less variation in the below-ground (soil seed bank) component of the vegetation, as viable seeds may remain in the soil well past the lifespan of parent plants.

Materials and methods Study area The study area is situated within the Gippsland Lakes Coastal Park in south-eastern Australia, and comprises approximately 3500 ha of heathland on a level sandplain (ca. 5–10 m a.s.l.), with a small area of stabilised dunefields toward the

Plant Ecol (2007) 190:1–12

eastern end (ca. 15–20 m a.s.l.). The soils are deep, uniform, well-drained, siliceous sands that are highly acidic and extremely low in nutrients (Wills 2002). The climate is mediterranean, with warm summers, cool to mild winters and a mean annual rainfall of 616 mm (Bureau of Meteorology 2000). The heath is dominated by a range of species from Myrtaceae, Fabaceae, Epacridaceae and Proteaceae. Leptospermum myrsinoides is the dominant shrub, in association with Monotoca scoparia, Banksia marginata, Allocasuarina misera and Epacris impressa. Caustis pentandra (Cyperaceae) and Hypolaena fastigiata (Restionaceae) dominate the ground stratum, while the small tree, Banksia serrata, and the ‘mallee-form’ eucalypts, Eucalyptus aff. willisii (Gippsland Lakes) and E. viminalis subsp. pryoriana, are emergent and scattered throughout. Experimental procedure Seven study sites were chosen according to three criteria. First, only sites on sandy soils and comprising ‘heath’ vegetation, or a potentially related successional stage, were included. Heath was defined as vegetation less than 2 m in height, with a projective foliage cover of the tallest stratum greater than 30%, and might include scattered emergent trees (following Specht 1970). All sites fitted this definition except SS-26 with an average vegetation height (excluding emergent trees) of 2.3 m (Wills 2002). Second, sites were required to be at least 4 ha, to allow a comprehensive overview of composition and richness for a particular post-fire age. Finally, each site had to be burnt by a separate fire to ensure site independence. Site age was determined using two methods: (a) fire history maps of the Department of Natural Resources and Environment were used to date sites burnt after 1980; and (b) whorls of Banksia leaf nodes were counted to estimate the age of sites burnt before 1981 (Wills 2003). A 4 ha plot was randomly located within each site, with the only condition being that one of its boundaries was within 50 m of an existing vehicle track, for ease of access. For this study, each 4 ha plot is considered representative of a particular post-fire age, and site fire histories are unknown

3

and assumed to be random. The validity of the chronosequence used in this study has previously been assessed by Wills (2002), with the uniformity of climate, topography, soil profiles, soil saturation capacity and nutrient availability suggesting that the fundamental assumptions of the chronosequence (Milberg 1995) have been met. To obtain a representative range of site ages, two sites were randomly chosen from each of three age classes (young, 0–6 years since fire; intermediate, 10–17 years; old, 24–26 years), with an extra site randomly selected from the ‘old’ age class. The young age class was designed to incorporate any post-fire ephemerals that may be present, while the old age class was chosen (within the confines of the age range available) to be representative of sites where heath senescence had already begun. The sites/ages chosen were as follows: young, BS-3 (3 years since fire) and SS-6 (6 years since fire); intermediate, TT-10 (10 years) and SS-17 (17 years); old, GP-24 (24 years), T4–26 and SS-26 (26 years). At each site, 88 soil cores were taken at random locations to a depth of 5 cm using an 8 cm diameter auger (total surface area of cores sampled per site: 0.442 m2; total volume per site: 0.022 m3). Additional seed may occur below 5 cm; however, seed numbers are likely to be lower (e.g. Carroll and Ashton 1965). In a tradeoff between sampling depth and area, it was decided that it was important to sample as much of the surface area as possible. In addition, sampling to a depth of 5 cm is a common benchmark in seed bank studies in mediterranean systems. Surface litter was included, as it is a potentially important source of seed, particularly in relatively old heath with high litter cover. Sampling was undertaken on March 20–22, 2000 after most species had set and released seed. This also allowed comparison of results with a previous soil seed bank study at one of the study sites, ‘GP’ (Wills and Read 2002). The previous study at GP used an application of smoked water and a heat treatment of 80–100C to promote seed germination from soil cores (Wills and Read 2002). The results indicated no difference in mean species richness of the soil seed bank of smoketreated relative to control samples, while the heat treatment promoted germination in a significantly

123

4

higher number of species. Furthermore, only one of the 21 species identified in the previous study occurred solely in the smoked water treatment, while six occurred solely in the heat-treated soil and three were restricted to controls. Therefore, to promote maximum germination of individual species, efforts were concentrated on the heat and control treatments. Heat treatment of seeds was undertaken a few days after soil collection and prior to potting, with 44 randomly selected cores from each site spread on trays to a thickness of ca. 1 cm and heated in ovens set at 80C for 15 min. Cores were randomised across three ovens. The remaining 44 cores from each site were left untreated to determine which species possessed non-dormant (transient) seed. Four cores from each site-treatment combination were then combined and spread to a depth of 1 cm, over 2.5 cm of sterilised river sand in germination trays. This gave 11 trays per site for each treatment. All cores were ‘potted’ on March 25, 2000 and randomly placed on benches within a fine-mesh enclosure, open to sunlight, precipitation and wind, while minimising external wind-borne seed and seedling herbivores. The experiment ran for 14 months. Watering was controlled using an electronically timed mistsprayer system. Mean monthly minimum and maximum temperatures in the germination enclosure were 12 and 24C, respectively at the commencement of the experiment. Temperatures dropped to a low of 9–16C in May 2000, and subsequently peaked in February 2001 (17–31C). Seedling emergence was recorded on a weekly basis for the first 3 months of the germination trial and thereafter on a fortnightly basis. Monitoring of punnets containing only sterilised sand indicated minor input of extraneous seed during the experiment; however, these species were easily identifiable and were subsequently removed. Nomenclature follows Ross (2000). Species composition (presence/absence) of the standing vegetation at each study site was sampled up to 12 months prior to collection of the soil cores, as part of a study examining post-fire vegetation succession at different sampling grains (Wills 2002). The samples comprised 41 quadrats from five spatial grains (1 m2: 20 quadrats; 10 m2: 10 quadrats; 100 m2: 7 quadrats; 900 m2: 3 quad-

123

Plant Ecol (2007) 190:1–12

rats; 1 ha: 1 quadrat), together with general reconnaissance across each site. Sampling efficiency was estimated at each grain using three different measures, with results consistently indicating that observed species richness was an excellent estimator of actual richness at each site. Sites were re-surveyed in spring to detect the presence of any annuals or ephemerals. Data analysis Non-metric Multidimensional Scaling (NMDS) was used to ordinate site data for standardised density and presence/absence data, using the Bray–Curtis similarity index. Standardised density was calculated as the percentage of the total density in that sample. All cores from a given site were combined for the analysis, to enable largescale differences in composition among the three age classes to be assessed. Differences in species composition among age classes were tested by one-way Analysis of Similarity (ANOSIM) (Clarke and Warwick 1994). Pairwise comparisons were unable to be assessed at the 0.05 significance level, as there were only 10 possible permutations. Therefore, pairwise R values were used to give an indication of the degree of separation between two age classes. On a scale of zero to one, R > 0.75 suggests that age classes are well separated, R = 0.5–0.75 suggests that groups are overlapping but clearly different, while R < 0.25 implies that groups are barely separable (Clarke and Gorley 2001). Mean similarity between treatments was calculated using the Similarity Percentage (SIMPER) procedure in PRIMER (Clarke and Warwick 1994). The effect of time since fire on total soil seed bank species richness and seedling density was analysed using linear and quadratic regression, with square-root and log10 transformations where appropriate. Quadratic regression was used to determine if there was an initial increase in species richness and seedling density followed by a decrease, as predicted. Paired t-tests were used to determine the effect of treatment (control versus heat) on seedling density and species richness of the seed bank, using sites as replicates. No data transformations were necessary. A critical value of a = 0.05 was used for all hypothesis testing.

Plant Ecol (2007) 190:1–12

5

Compositional similarity between the soil seed bank and the standing vegetation was calculated using both Bray–Curtis and Jaccard similarity indices (Kent and Coker 1992), to enable comparison with other studies. Pearson correlation was used to determine whether there was a significant relationship between these similarity indices and time since fire.

Results Species richness, density and composition of the germinable soil seed bank The germinable soil seed bank comprised 34 identifiable species (all native), from 31 genera and 20 families (Appendix 1). A small percentage of seedlings were unable to be identified ( < 1%), or were only identified to genus, owing to premature death (2%). Most (>99%) seedlings germinated within the first 7 months. Total species richness at each site ranged from 13 to 22 species; however, it did not vary significantly with increasing time since fire, using either linear (log10 transformed data: F = 1.33, P = 0.301) or quadratic regression (log10 transformed data: F = 1.67, P = 0.297) (Fig. 1a). Mean species richness was significantly higher in heattreated soils compared with controls (t = – 6.57, P = 0.001) (Fig. 1b). Epacris impressa accounted for 79% of all seedlings, with Leucopogon ericoides (4%), Leptospermum myrsinoides (4%) and Calytrix tetragona (3%) accounting for a further 11% of seedlings (Appendix 1). The remaining 10% of seedlings comprise 30 species. Total seedling density did not vary significantly with time since fire, using both linear (log10 transformed data: F = 0.75, P = 0.426) and quadratic regression (log10 transformed data: F = 0.51, P = 0.637). The data were re-analysed after excluding the dominant species (E. impressa), but conclusions were not altered. L. myrsinoides was the only species that increased significantly in seedling density with increasing time since fire (square-root transformed data: F = 11.9, P = 0.018). Seedling densities of heat-treated soil yielded 2005 ± 932 seedlings m–2 for young sites (n = 2),

Fig. 1 The relationship between soil seed bank species richness at each study site and time since fire. (a) Total germinable species richness (sample area per site of 0.442 m2; sample volume per site of 0.022 m3), and (b) mean germinable species richness ( ± 1 SE) per germination tray (four soil cores totalling 0.020 m2 and 0.001 m3).

•, control; m, heat

5213 ± 1154 seedlings m–2 for intermediate-aged sites (n = 2) and 3713 ± 419 seedlings m–2 for old sites (n = 3). Control soil yielded 1109 ± 330 seedlings m–2 for young sites (n = 2), 1538 ± 77 seedlings m–2 for intermediate-aged sites (n = 2) and 1379 ± 151 seedlings m–2 for old sites (n = 3). Mean seedling density was significantly higher in heat-treated soils compared with controls (t = – 4.45, P = 0.004) (Fig. 2).

123

6

Plant Ecol (2007) 190:1–12

site were excluded from the analysis, or when E. impressa was excluded. Similarity between the germinable soil seed bank and vegetation

Fig. 2 The relationship between mean germinable seedling density per germination tray at each study site ( ± 1 SE) and time since fire (four soil cores totalling 0.020 m2 and 0.001 m3). •, control; m, heat

Ordination by NMDS using presence/absence data showed significant differences in composition among sites (R = 0.55, P = 0.029) (Fig. 3). Based on the guidelines of Clarke and Gorley (2001), young and old sites are well separated (R = 0.83), intermediate and old sites are overlapping but clearly different (R = 0.50), while young and intermediate sites are not separable (R < 0.01). NMDS of standardised density data yielded no significant difference among age classes (R = 0.15, P = 0.200), which was most likely due to the large variation in density in some species within age classes. Conclusions were unchanged when the seven species found at only one

Fig. 3 Ordination by NMDS of soil seed bank species composition (presence–absence data) at each site. Sites were grouped a priori into young (0–6 years), intermediate (10–17 years) and old (24–26 years) age classes

123

The Bray–Curtis similarity between the composition of the soil seed bank and the extant vegetation ranged from a low of 37% at the youngest site to a high of 58% at the oldest site (F = 0.73, P = 0.063). Jaccard similarities exhibited a similar trend and ranged from 16–22% (F = 0.72, P = 0.067). There was no significant relationship between either of the similarity indices and time since fire. All species found in the soil seed bank over the seven sites were also present in the extant vegetation of at least one site. However, 11 of the 34 seed bank species were not found in the extant vegetation at the same site. Of these species, eight (73%) were herbaceous, two were shrubs and one was a subshrub.

Discussion Species richness, density and composition of the germinable soil seed bank Species richness and seedling density of the germinable soil seed bank were not significantly correlated with time since disturbance (fire), which concurs with other studies undertaken in a range of systems and disturbance types (Zammit and Zedler 1994; Milberg 1995; Dalling and Denslow 1998). However, this finding is contrary to our predictions. Although the youngest site (BS-3) followed predictions by having the lowest number of species, a significant trend of increasing richness after fire followed by a decrease some years later was not recorded. Additional sites at the ‘young’ end of the time since fire spectrum may have provided a greater insight into the extent of a post-fire ‘carry-over’ seed bank, which may have had an influence on the significance of trends with time since fire; however, such sites were not available at the time this study was undertaken. In the context of other studies that have examined soil seed bank chronosequences

Plant Ecol (2007) 190:1–12

(Granstro¨m 1988; Zammit and Zedler 1988; Roberts and Vankat 1991; Zammit and Zedler 1994; Milberg 1995; Dalling and Denslow 1998; Bekker et al. 1999; Ne’eman and Izhaki 1999; Bekker et al. 2000; Izhaki et al. 2000), the seven sites studied here constitute a comparatively large sample size. However, it is still a relatively small number of sites and may be insufficient to detect any real trend in species richness or seedling density in relation to time since fire. The time span used in this study (26 years) is short in comparison to forest chronosequences, but is consistent with the life spans of the component species, and with changes recorded in the extant vegetation. Furthermore, in south-eastern Australian heathlands it is very difficult to find sites older than 25 years (Specht et al. 1958; Siddiqi et al. 1976; Russell and Parsons 1978; Wark et al. 1987; Enright et al. 1994), because current fire frequencies inhibit the vegetation from reaching ages much older than this (see Cheal (2000) and McMahon (1984) for exceptions from semi-arid heathland). Within the timeframe examined, we conclude that time since fire is not a major determinant of seed bank richness or density. However, even though richness and density did not change, the species composition of the soil seed bank differed significantly between age classes (based on presence/absence data), with the greatest compositional difference between young and old sites. The main driver contributing to the dissimilarity between young and old sites appears to be herbaceous species, with 11 out of 14 species contributing most to the variation between young and old sites being herbaceous species or subshrubs. Heat-treated soil consistently yielded higher germination levels than controls, in terms of both mean species richness and mean seedling density, indicating that heat had a promotive effect on seedling germination. The seedling density of heat-treated soil from the previous study at GP yielded 4575 seedlings m–2 (Wills and Read 2002), which was comparable with the 3111 seedlings m–2 from GP-24 in this study, and values from other heat-treated soil in Australian heathlands (Enright et al. 1997; Marsden-Smedley et al. 1997). However, in this study, seedling densities in control soils were substantially higher at GP-24

7

(1148 m–2) compared to the previous study (485 m–2) (Wills and Read 2002). The study by Wills (2002) indicated that seedling densities in heat-treated soil were almost ten times higher than those in the controls; however, in this study they were only ca. three times higher, even though densities in heat-treated soil were relatively comparable. Therefore, the reduced effect of the heat-treatment in this study appears to have arisen not from a deficiency in the heat treatment, but from a relative increase in the germination of seed from control soils. This increase may have been caused by increased seed density and viability in the season’s seed crop (particularly in E. impressa), and/or better climatic conditions during the germination trial for breaking dormancy and inducing germination. As estimates of seed densities and richness are dependent on the intensity and duration of the treatments used to break dormancy and induce germination (Warcup 1980; Auld and O’Connell 1991; Musil 1991; Baldwin et al. 1994; Roche et al. 1997), the results of this study are only applicable to the particular seed treatment used. Estimates of soil seed bank species richness, density and composition are also dependent on a range of other factors, including seed longevity (Auld et al. 2000), dormancy (Baskin and Baskin 1989; Bell 1999), seed production and viability (Auld 1986; Bell et al. 1987), seed predation (Andersen and Ashton 1985; Louda 1989; Clarke et al. 1996), seed bank heterogeneity (Thompson 1986; Ne’eman and Izhaki 1999), seasonal variation in seed banks (Thompson and Grime 1979; Grant and Koch 1997; Ward et al. 1997) and sampling intensity (Dessaint et al. 1996). In this study, the minimum detectable density of soilstored seed with a 95% confidence level and assuming a Poisson distribution, was 14 seeds m–2 (per treatment), according to the formula of Thompson et al. (1997). Species with seed abundances below this level, and/or with very clumped distributions, were likely to be overlooked using the sampling regime employed in this study. Epacris impressa was by far the most abundant species, accounting for 79 ± 10% of all seedlings. This dominance of the seed bank by a single species is not unusual in heath, e.g. Calluna vulgaris in Europe (Granstro¨m 1988; Mitchell et al. 1998)

123

8

and Adenostoma fasciculatum in California (Zammit and Zedler 1988). However, species that dominate seed banks are also usually dominant in the above-ground vegetation, unlike E. impressa in this study, which had a mean vegetative cover of only 2 ± 0%. In contrast, the dominant shrub at every site, Leptospermum myrsinoides, contributed only 4 ± 1% of all seedlings emerging from the soil seed bank, while its mean vegetative cover across all sites was 22 ± 3% (Wills 2002). This was the only species to increase in seedling density with increasing time since fire, suggesting that L. myrsinoides will recruit more seedlings following fire after long fire-free periods, therefore increasing plant density, and consequently facilitating even greater dominance of the heath. In addition, L. myrsinoides can regenerate from basal sprouts during the inter-fire period, which further enables this species to maintain and increase its dominance under relatively long firefree periods. The seven most common species in the soil seed bank were long-lived (>25 years) shrubs. Conversely, seeds of post-fire ephemeral and short-lived species were uncommon in the soil seed bank, with only three post-fire ephemeral species, Euchiton sphaericus, Isolepis marginata and Laxmannia orientalis, germinating from soilstored seed. This is consistent with the uncommon nature of post-fire ephemeral and short-lived species in the above-ground component of the vegetation. Similarity between the germinable soil seed bank and vegetation The similarity between the composition of the seed bank and the extant vegetation was not significantly related to time since fire. Studies in tropical forest (Dalling and Denslow 1998), Mediterranean pine forest (Ne’eman and Izhaki 1999), Dutch dune slacks (Bekker et al. 1999) and hayfield successions (Bekker et al. 2000) also suggest that similarities between the soil seed bank and the extant vegetation are not significantly related to site age, whether it be time since the last fire or agricultural abandonment. However, the low number of sites used in most studies (usually less than six), the heterogeneity of soil

123

Plant Ecol (2007) 190:1–12

seed banks (Thompson 1986; Ne’eman and Izhaki 1999), and the inherent variability between sites in a chronosequence (Pickett 1989), act together to mitigate against a significant trend, even if one exists. For example, when the composition of the seed bank at GP is compared over the 2 years, differences in composition of the soil seed bank are evident, with seven species recorded from this study that were not recorded during the previous study (Wills and Read 2002). This is a substantial figure given that Wills and Read (2002) recorded only 20 species from the germinable soil seed bank, and suggests that sampling over multiple years (and/or seasons) may substantially increase the accuracy of soil seed bank density and richness estimates. Most species present in the extant vegetation were not recorded (63%) or were in very low abundances (i.e. < 1% of all seedlings recorded) in the soil seed bank (29%), a finding often reported elsewhere (Zammit and Zedler 1988; Molnar et al. 1989; Dalling and Denslow 1998; Morgan 1998). The mode of regeneration or method of seed storage (serotiny) appears to be the major determinant of species absence in the soil seed bank, as 66% of all species occurring in the extant vegetation had the capacity to resprout either obligately or facultatively (Wills 2002). Once these species are taken into account, there were usually only six or seven obligate seeders at each site that were not present in the germinable soil seed bank, which probably reflected unusually low seed abundances at the time of sampling (although little is known of seed production rates and viability in heath species), or the inability of the heat treatment to break seed dormancy and induce germination. Obligate resprouting reduces extinction risks associated with both abnormally high and low fire frequencies, so long as fire occurs within the life span of the plant, and is not unusually severe (Bellingham and Sparrow 2000). Conversely, obligate seeders risk local extinction if fire frequencies are too high or low (Keith 1996) relative to the fire regime to which they are adapted. There is no evidence to suggest that the proportion of seeders to resprouters is changing over time (Wills 2002). Species found in the soil seed bank that are not present in the extant vegetation are particularly

Plant Ecol (2007) 190:1–12

9

most species not recorded from the soil seed bank are either resprouters (obligate or facultative) or serotinous, suggesting that they will readily regenerate following fire. Unless fire frequencies are very high and kill fire-sensitive obligate seeders before they reach reproductive maturity, the chance that the soil seed bank (or lack thereof) could substantially alter vegetation composition within the study area after fire is low. However, it is unclear how vegetation succession may alter in response to very severe fires that have the potential to kill both seeders and resprouters.

important in that they give an insight into the changes in vegetation composition that may occur in the post-fire environment. Eleven such species were recorded during the study, with seven (64%) of these species being obligate seeders, in contrast to the lower proportion of obligate seeders (34%) in the extant vegetation (Wills 2002). Three of these species appeared to be post-fire ephemerals that regenerate soon after fire and then quickly die off, leaving their seed remaining in the soil. Only one of the 11 species was classified as an obligate resprouter. In conclusion, although there was no change in species richness and seedling density of the germinable soil seed bank with increasing time since fire, there was a significant difference in species composition among age classes. Despite this difference, it is unlikely to contribute to a major shift in the post-fire successional pathway, given that

Acknowledgements We thank Parks Victoria for granting permission to work in the Gippsland Lakes Coastal Park (Permit No. 10000248); Chad McLean, John Thompson and Derek van Netten for field assistance, and two anonymous referees for helpful comments on the manuscript. This work was undertaken with the support of an Australian Postgraduate Award to TJW.

Appendix Appendix 1 Total density of seedlings recorded in the germinable soil seed bank for each treatment at the seven sites Time since fire (years)

Life-form

MR

Species/treatment Monocotyledons Cyperaceae Caustis pentandra Isolepis marginata Liliaceae Laxmannia orientalis Orchidaceae Caleana major Genoplesium sp. aff. rufum Pyrorchis nigricans Poaceae Poa sieberiana var. sieberiana Restionaceae Hypolaena fastigiata Xanthorrhoeaceae Lomandra glauca Dicotyledons Apiaceae Platysace ericoides Xanthosia pilosa Asteraceae Euchiton sphaericus Casuarinaceae Allocasuarina sp.

BS-3

SS-6

TT-10

SS-17

GP-24

T4-26

SS-26

3

6

10

17

24

26

26

C

H

C

H

C

H

C

H

C

H

C

H

C

H

Total

Gr Gr

OS OS

1 0

0 0

1 0

0 3

0 0

0 1

0 0

1 0

0 0

0 0

0 2

0 0

2 1

0 0

5 7

FG

OS

0

0

5

9

0

4

0

0

1

7

0

0

0

0

26

FG FG FG

OR OR FR

0 0 0

0 0 1

0 0 0

0 2 0

0 0 0

0 0 2

1 0 1

0 0 0

0 0 2

0 0 1

0 0 0

0 0 0

0 0 0

0 0 0

1 2 7

Gr

FR

0

0

0

0

10

0

3

1

3

2

0

0

0

0

19

Gr

OR

1

1

0

3

1

16

0

2

9

7

4

5

0

1

50

Gr

FR

0

0

0

0

0

0

1

0

0

0

0

0

0

0

1

S Ss

OS FR

0 0

0 0

0 0

0 0

0 0

0 0

1 0

5 0

0 1

0 0

0 0

1 5

0 1

0 0

7 7

FG

OS

1

0

1

2

0

0

0

0

0

0

0

0

0

0

4

S

FR

0

0

0

0

0

0

0

1

0

0

0

0

0

0

1

123

10

Plant Ecol (2007) 190:1–12

Appendix 1 continued Time since fire (years)

Life-form MR BS-3 3

Species/treatment Dilleniaceae Hibbertia acicularis Hibbertia virgata Droseraceae Drosera peltata Epacridaceae Astroloma pinifolium Epacris impressa Leucopogon ericoides Leucopogon virgatus var. virgatus Monotoca scoparia Euphorbiaceae Amperea xiphoclada var. xiphoclada Fabaceae Bossiaea cinerea Bossiaea heterophylla Dillwynia spp. Gompholobium huegelii Goodeniaceae Dampiera stricta Haloragaceae Gonocarpus tetragynus Lauraceae Cassytha glabella Mimosaceae Acacia oxycedrus Myrtaceae Calytrix tetragona Leptospermum myrsinoides Thryptomene micrantha Thymelaeaceae Pimelea linifolia ssp. linifolia Unidentified seedlings Total No. seedlings Seedlings m–2

C

SS-6

TT-10

SS-17

GP-24

T4-26

SS-26

6

10

17

24

26

26

H

C

H

C

H

C

H

C

H

C

H

C

H

Total

S S

FR 0 FR 0

0 0

0 0

0 2

0 0

0 0

0 0

0 2

1 0

0 0

0 0

0 0

0 0

0 0

1 4

FG

FR 2

1

19 7

0

0

0

0

1

0

0

0

0

0

30

S S S S

OS FR OS OR

0 615 6 0

1 36 23 0

2 42 37 4

0 253 34 0

0 1296 24 1

0 283 5 1

0 804 24 0

2 146 15 0

0 567 27 0

3 212 13 0

0 792 19 0

23 182 38 0

31 599 57 0

62 6122 327 6

S

FR 0

0

0

0

0

0

0

0

0

0

0

0

2

1

3

Ss

FR 0

0

0

0

0

0

1

0

0

1

0

0

2

1

5

S S S S

FR FR FR FR

0 6 10 0

0 8 0 0

0 28 7 0

0 3 1 0

0 0 29 0

0 3 3 1

1 10 25 8

0 5 0 0

0 18 4 1

0 5 1 0

0 52 17 1

1 0 0 0

3 1 15 1

5 141 112 12

Ss

FR 0

0

0

3

1

0

2

1

0

0

0

1

0

1

9

FG

OS 1

0

0

0

11

17

0

2

0

0

12

9

23

36

111

Tw

OS 0

0

0

0

0

0

0

0

0

1

0

1

0

0

2

S

OS 0

0

0

1

0

0

1

0

0

2

0

0

0

1

5

S S

OS 0 FR 6

1 2

26 40 8 13

20 21

5 10

0 15

2 7

0 66

9 37

1 28

0 41

46 24

44 24

194 302

S

OS 4

2

43 30

1

0

0

0

0

0

31

5

1

3

120

S

OS 0

0

0

0

0

0

0

0

0

0

0

0

0

1

1

?

?

4

1

2

1

2

1

1

2

4

1

1

1

4

25

0 295 5 0

0 2 0 0

0

318 649 172 237 357 1407 323 897 254 688 313 950 347 824 7736 1439 2937 778 1072 1615 6367 1462 4059 1149 3113 1416 4299 1570 3729 2500

FG, Forb/geophyte; Gr, graminoid; S, shrub; Ss, subshrub; Tw, twiner; C, control; H, heat; MR, mode of regeneration (T. Wills, unpublished data); OS, obligate seeder (>91% seed germination); OR, obligate resprouter (0–10% seed germination); FR, facultative resprouter (11–90% seed germination)

References Andersen AN, Ashton DH (1985) Rates of seed removal by ants at heath and woodland sites in southeastern Australia. Austr J Ecol 10:381–390

123

Auld TD (1986) Dormancy and viability in Acacia suaveolens (Sm.) Willd. Austr J Bot 34:463–472 Auld TD, O’Connell MA (1991) Predicting patterns of post-fire germination in 35 eastern Australian Fabaceae. Austr J Ecol 16:53–70

Plant Ecol (2007) 190:1–12 Auld TD, Keith DA, Bradstock RA (2000) Patterns in longevity of soil seedbanks in fire-prone communities of south-eastern Australia. Austr J Bot 48:539– 548 Bakker JP, Bakker ES, Rose´n E, Verweij GL, Bekker RM (1996) Soil seed bank composition along a gradient from dry alvar grassland to Juniperus shrubland. J Vege Sci 7:165–176 Baldwin IT, Staszak-Kozinski L, Davidson R (1994) Up in smoke: I. Smoke-derived germination cues for postfire annual, Nicotiana attenuata Torr. Ex. Watson. J Chem Ecol 20:2345–2371 Baskin JM, Baskin CC (1989) Physiology of dormancy and germination in relation to seed bank ecology. In: Leck MA, Parker VT, Simpson RL (eds) Ecology of soil seed banks. Academic Press, San Diego, California, pp 53–66 Bekker RM, Lammerts EJ, Schutter A, Grootjans AP (1999) Vegetation development in dune slacks: the role of persistent seed banks. J Vege Sci 10:745–754 Bekker RM, Verweij GL, Bakker JP, Fresco LFM (2000) Soil seed bank dynamics in hayfield succession. J Ecol 88:594–607 Bell DT (1999) The process of germination in Australian species. Austr J Bot 47:475–517 Bell DT, Vlahos S, Watson LE (1987) Stimulation of seed germination of understorey species of the northern Jarrah forest of Western Australia. Austr J Bot 35:593–599 Bellingham PJ, Sparrow AD (2000) Resprouting as a life history strategy in woody plant communities. Oikos 89:409–415 Bureau of Meteorology (2000) Climate Averages: East Sale. Commonwealth of Australia, Canberra Carrol EJ, Ashton DH (1965) Seed storage in soils of several Victorian plant communities. Victorian Natural 82:102–110 Cheal D (2000) Seed regeneration in long-unburnt and recently-burnt heathland at Wyperfeld National Park. Proc Roy Soc Victoria 112:111–118 Clarke KR, Gorley RN (2001) PRIMER v5: User manual/ tutorial. PRIMER-E, Plymouth Clarke KR, Warwick RM (1994) Change in marine communities: an approach to statistical analysis and interpretation. Plymouth Marine Laboratory, Plymouth Clarke PJ, Myerscough PJ, Skelton NJ (1996) Plant coexistence in coastal heaths: between- and withinhabitat effects of competition, disturbance and predation in the post-fire environment. Austr J Ecol 21:55–63 Dalling JW, Denslow JS (1998) Soil seed bank composition along a forest chronosequence in seasonally moist tropical forest, Panama. J Vege Sci 9:669–678 Dessaint F, Barralis G, Caixinhas ML, Mayor J-P, Recasens J, Zanin G (1996) Precision of soil seedbank sampling: how many soil cores? Weed Res 36:143–151 Enright NJ, Goldblum D, Ata P, Ashton DH (1997) The independent effects of heat, smoke and ash on emergence of seedlings from the soil seed bank of a

11 heathy Eucalyptus woodland in Grampians (Gariwerd) National Park, western Victoria. Austr J Ecol 22:81–88 Enright NJ, Miller BP, Crawford A (1994) Environmental correlates of vegetation patterns and species richness in the northern Grampians, Victoria. Austr J Ecol 19:159–168 Granstro¨m A (1988) Seed banks at six open and afforested heathland sites in southern Sweden. J Appl Ecol 25:297–306 Grant CD, Koch JM (1997) Ecological aspects of soil seedbanks in relation to bauxite mining. II. Twelve year old rehabilitated mines. Austr J Ecol 22:177–184 Groves RH, Specht RL (1981) Seral considerations in heathland. In: Gillison AN, Anderson DJ (eds) Vegetation classification in Australia. Australian National University Press, Canberra, pp 78–85 Izhaki I, Henig-Sever N, Ne’eman G (2000) Soil seed banks in Mediterranean Aleppo pine forests: the effect of heat, cover and ash on seedling emergence. J Ecol 88:667–675 Keeley JE (1995) Seed-germination patterns in fire-prone Mediterranean-climate regions. In: Arroyo MTK, Zedler PH, Fox MD (eds) Ecology and biogeography of Mediterranean ecosystems in Chile, California and Australia. Springer-Verlag, New York, pp 239–273 Keith D (1996) Fire-driven extinction of plant populations: a synthesis of theory and review of evidence from Australian vegetation. Proc Linnaean Soc NSW 116:37–78 Kent M, Coker P (1992) Vegetation description and analysis. CRC Press, Boca Raton Kruger FJ (1983) Plant community diversity and dynamics in relation to fire. In: Kruger FJ, Mitchell DT, Jarvis JVM (eds) Mediterranean-type ecosystems: the role of nutrients. Springer-Verlag, Berlin, pp 446–472 Lamont BB, Le Maitre DC, Cowling RM, Enright NJ (1991) Canopy seed storage in woody plants. Bot Rev 57:277–317 Louda SM (1989) Predation in the dynamics of seed regeneration. In: Leck MA, Parker VT, Simpson RL (eds) Ecology of soil seed banks. Academic Press, San Diego, California, pp 25–51 Marsden-Smedley JB, Appleby MWA, Pyrke AF, Battaglia M (1997) Soil seed bank of Tasmanian grasslands, grassy woodlands and heathy forests: methodology for optimising the germination of native species. Parks and Wildlife Service, Department of Environment and Land Management, Hobart McMahon A (1984) The effects of time since fire on heathlands in the Little Desert, N.W. Victoria, Australia. In: Dell B (ed) MEDECOS IV, Proceedings 4th international conference on Mediterranean ecosystems. The Botany Department, University of Western Australia, Nedlands, Western Australia, pp 99–100 Milberg P (1995) Soil seed bank after eighteen years of succession from grassland to forest. Oikos 72:3–13 Mitchell RJ, Marrs RH, Auld MHD (1998) A comparative study of the seedbanks of heathland and successional

123

12 habitats in Dorset, Southern England. J Ecol 86:588– 596 Molnar CD, Fletcher D, Parson RF (1989) Relationships between heath and Leptospermum laevigatum scrub at Sandringham, Victoria. Proc Roy Soc Victoria 101:77–87 Morgan JW (1998) Composition and seasonal flux of the soil seed bank of species-rich Themeda triandra grasslands in relation to burning history. J Vege Sci 9:145–156 Musil CF (1991) Seed bank dynamics in sand plain lowland fynbos. S Afr J Bot 57:131–142 Ne’eman G, Izhaki I (1999) The effect of stand age and microhabitat on soil seed banks in Mediterranean Aleppo pine forests after fire. Plant Ecol 144:115–125 Parker VT, Simpson RL, Leck MA (1989) Pattern and process in the dynamics of seed banks. In: Leck MA, Parker VT, Simpson RL (eds) Ecology of soil seed banks. Academic Press, San Diego, California, pp 367–384 Pausas JG (2001) Resprouting vs seeding – a Mediterranean perspective. Oikos 94:193–194 Pickett STA (1989) Space-for-time substitution as an alternative to long-term studies. In: Likens GE (ed) Long-term studies in ecology: approaches and alternatives. Springer-Verlag, New York, pp 110–135 Roberts TL, Vankat JL (1991) Floristics of a chronosequence corresponding to old field-deciduous forest succession in southwestern Ohio. II. Seed banks. Bull Torrey Bot Club 118:377–384 Roche S, Koch JM, Dixon KW (1997) Smoke enhanced seed germination for mine rehabilitation in the southwest of Western Australia. Restor Ecol 5:191– 203 Ross JH (2000) A census of the vascular plants of Victoria. National Herbarium of Victoria, Royal Botanic Gardens, South Yarra Russell RP, Parsons RF (1978) Effects of time since fire on heath floristics at Wilson’s Promontory, Southern Australia. Austr J Bot 26:53–61 Siddiqi MY, Carolin RC, Myerscough PJ (1976) Studies in the ecology of coastal heath in New South Wales. III. Regrowth of vegetation after fire. Proc Linnean Soc NSW 101:53–63 Specht RL (1970) Vegetation. In: Leeper GW (ed) The Australian environment. CSIRO and Melbourne University Press, pp 44–67

123

Plant Ecol (2007) 190:1–12 Specht RL, Rayson P, Jackman ME (1958) Dark Island Heath (Ninety-Mile Plain, South Australia). VI. Pyric succession: changes in composition, coverage, dry weight, and mineral nutrient status. Austr J Bot 6:59– 88 Thompson K (1986) Small-scale heterogeneity in the seed bank of an acidic grassland. J Ecol 74:733–738 Thompson K, Grime JP (1979) Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats. J Ecol 67:893–921 Thompson K, Bakker JP, Bekker RM (1997) The soil seed banks of North West Europe: methodology, density and longevity. Cambridge University Press, Cambridge Warcup JH (1980) Effect of heat treatment of forest soil on germination of buried seed. Austr J Bot 28:567–571 Ward SC, Koch JM, Grant CD (1997) Ecological aspects of soil seed-banks in relation to bauxite mining. I. Unmined jarrah forest. Austr J Ecol 22:169–176 Wark MC, White MD, Robertson DJ, Marriott PF (1987) Regeneration of heath and heath woodland in the north-eastern Otway Ranges following the wildfire of February 1983. Proc Roy Soc Victoria 99:51–88 Whelan RJ (1986) Seed dispersal in relation to fire. In: Murray DR (ed) Seed dispersal. Academic Press, Sydney, pp 237–271 Whelan RJ, Rodgerson L, Dickman CR, Sutherland EF (2002) Critical life cycles of plants and animals: developing a process-based understanding of population changes in fire-prone landscapes. In: Bradstock RA, Williams JE, Gill AM (eds) Flammable Australia: the fire regimes and biodiversity of a continent. Cambridge University Press, Melbourne, pp 94–124 Wills TJ (2002) Succession in sand heathland at Loch Sport, Victoria: changes in vegetation, soil seed banks and species traits. PhD Thesis, School of Biological Sciences, Monash University, Melbourne Wills TJ (2003) Using Banksia (Proteaceae) node counts to estimate time since fire. Austr J Bot 51:239–242 Wills TJ, Read J (2002) Effects of heat and smoke on germination of soil-stored seed in a south-eastern Australian sand heathland. Austr J Bot 50:197–206 Zammit CA, Zedler PH (1988) The influence of dominant shrubs, fire, and time since fire on soil seed banks in mixed chaparral. Vegetatio 75:175–187 Zammit C, Zedler PH (1994) Organisation of the soil seed bank in mixed chaparral. Vegetatio 111:1–16