Elsevier

Biological Conservation

Volume 142, Issue 2, February 2009, Pages 256-269
Biological Conservation

Soil seed bank compositional change constrains biodiversity in an invaded species-rich woodland

https://doi.org/10.1016/j.biocon.2008.10.019Get rights and content

Abstract

Relationships between plant invasion and the soil seed bank in highly diverse fire adapted mediterranean woodlands are poorly understood, yet critical for that ecosystem’s conservation. Within the biodiversity hotspot of southwest Australia we investigated the composition and diversity of the Banksia woodland soil seed bank in good condition (GC), medium condition (MC) and poor condition invaded by the South African perennial species Ehrharta calycina (PCe) and Pelargonium capitatum (PCp). The investigation assessed three questions: (1) Do soil seed banks of invaded sites have fewer germinants of native species and more germinants of introduced (non-native) species than sites with minimal invasion? (2) Do soil seed banks show shifts in ecological functional types with invasion? (3) Is the soil seed bank of introduced species persistent? Native species germinants, mainly shrubs and perennial herbs, were highest in GC sites and least in poorer condition sites suggesting a reduction in their numbers had occurred over time. Introduced germinants were dominated by perennial and annual grasses, and annual herbs. E. calycina had the greatest seed density (8328 germinants m−2). More introduced than native germinants occurred in the litter. Rapid germination of introduced species (30% in week 1) compared to native species (4% in week 1) provides the capacity for their early dominance. A limited native soil seed bank and dominant persistent introduced soil seed bank represent great challenges for the structural and functional conservation and restoration of woodland ecosystems. This study provides key new knowledge, applicable to a wide range of ecosystems, to help formulate conservation protocols to control dominant introduced species and conserve and restore biodiverse-rich woodlands.

Introduction

The soil seed bank, in which seeds enter through seed dispersal and leave through germination or death (Walck et al., 1996), provides a memory of past vegetation and represents the structure of future populations. A ‘healthy’ soil seed bank, in which all species are adequately represented, is vitally important for the conservation and long term survival of individual species as well as plant communities. The study of soil seed banks, therefore, has the potential to provide insights into community dynamics. As a significant tool in the conservation of ecosystems, they can provide an understanding of succession or the likely community impacts following natural or anthropogenic disturbances (Adams et al., 2005, Satterthwaite et al., 2007). Moreover, when combined with knowledge about the ecology of the species, soil seed banks can display shifts in functional groups or keystone species, which may be associated with shifts in ecosystem processes, understandings important for successful conservation management. In the present paper, we examine changes in soil seed banks of Banksia woodland, subject to different degrees of invasion and fire frequency. The Banksia woodland community under study is dominated by Banksia attenuata R. Br. and Banksia menziesii R.Br. (Proteaceae) (both up to 6–8 m tall), with scattered emergent Eucalyptus gomphocephala DC., over a low open understorey consisting mainly of shrubs and perennial herbs. This community is located within the southwest Australian biodiversity hotspot (Myers et al., 2000). Invasion poses one of the greatest threats to the biodiversity of this bioregion (Hopper and Gioia, 2004), comparable to threats anticipated for other mediterranean biomes (Sala et al., 2000).

Fire is a natural disturbance agent in Banksia woodlands and is important in determining floristic composition and structure; however, there is ongoing debate about optimal fire regimes. Species in Banksia woodlands possess a range of fire tolerance and avoidance mechanisms (Bell et al., 1984). One convenient classification of strategies is that of resprouters and seeders. Resprouter species rely predominantly on regrowth following fire, only requiring sufficient seed to offset the mortality of parent plants; however survival of seeder species is dependent entirely on seed produced between disturbance events (Meney et al., 1994). Fire has the potential in mediterranean systems to promote germination of the soil seed bank by heating the soil, inducing germination through smoke stimulus, changing soil structure and nutrient levels, providing greater access to water and light due to temporary removal of the aboveground vegetation, and reducing allelopathic influences (Snyman, 2005). Therefore, excessively frequent fire will result in a decline in native seeder species (Burrows and Wardell-Johnson, 2003), and will favour those introduced species, such as Ehrharta calycina Sm. (Poaceae), which resprout from rhizomes and produces large numbers of seeds soon after a fire event (Baird, 1977).

Invasion has potential impacts on ecosystem conservation through alteration of community composition and ecosystem function and processes (Loreau et al., 2001). The 437 ha urban Reserve in which the present study was conducted, is heavily invaded by introduced species, but still contains some vegetation patches that visually appear in near-pristine condition. The Reserve has 336 native plant species and 232 introduced species (Barrett and Tay, 2005) with many of the introduced species producing large numbers of seeds. Effective management of introduced species requires detailed knowledge of these species’ reproductive biology, including seed production, persistency in soil seed banks (Fenner and Thompson, 2005) and response to disturbance. Studies of the soil seed bank provide an opportunity to predict the restoration potential of invaded areas (Holmes, 2002) and are particularly important in developing conservation strategies to manipulate and deplete the introduced soil seed store to constrain invasive species while enhancing and ensuring dominance of the native soil seed bank (Bakker and Wilson, 2004, Adams et al., 2005).

Many species arrest seed development to enable the seed to persist into the future. This strategy is important for therophytes (annuals), particularly at times when the plants may not survive as adults during unfavourable conditions (Fenner and Thompson, 2005). Methods for predicting persistence of seed vary and may be based on germination timing (Thompson and Grime, 1979), seed size and shape (Thompson et al., 1993), depth distribution of seeds in the soil (Thompson et al., 1993, Thompson et al., 1998) and seasonal dormancy and germination patterns (Walck et al., 2005). While a relationship has been found between persistency and smaller and more rounded seeds in British species (Thompson et al., 1993) this is not the case for Australian species, probably due to ecological differences of burial and disturbance operating in the different habitats (Leishman and Westoby, 1998). Moreover, while seed persistence is chiefly a species trait, the potential exists for modification by environmental conditions indicating that as species move into a new habitat there may be changes in species persistence away from the home range (Fenner and Thompson, 2005). An understanding of the persistence of introduced species in their new ecosystems is one key to the conservation of these systems following invasion (Jalili et al., 2003, Adams et al., 2005).

The study aim was to conduct an audit of the spatial extent and diversity of the germinable soil seed bank to investigate if the soil seed bank is affected by site condition, and to assess if the soil seed bank is likely to indicate future shifts in species and ecology. Four vegetation states, i.e. good condition (minimal invasion), medium condition, poor condition invaded by the perennial grass E. calycina, and poor condition invaded by the perennial herb Pelargonium capitatum (L.) L’Her. (Geraniaceae) were utilised to investigate: (1) Do soil seed banks of invaded sites have fewer native and more introduced germinants than sites with minimal invasion? (2) Do soil seed banks show evidence of shifts in ecological functional types with invasion? (3) Are the species in the soil seed bank likely to germinate, i.e. do they require germination cues and is the soil seed bank persistent?

Section snippets

Location and site selection

The study site was in a Banksia woodland community located in Bold Park, a 437 hectare metropolitan reserve in Perth, Western Australia. Nine plant communities have been described for this Reserve, with the dominant Banksia woodland covering 58% of the Reserve (Botanic Gardens and Parks Authority, 2000). The area has a mediterranean climate of extended hot, dry summers and mild wet winters with an average rainfall (1993–1999) of 773 mm (Botanic Gardens and Parks Authority, 2000). The Banksia

Data analyses

Final (accumulated) numbers of seeds for all treatments were expressed per m2 by taking into account the surface area of soil sampled. Examination of the data before analysis indicated no gross violations of assumed normality (Kolmogorov–Smirnov method). Data were also tested for homogeneity of variance using Levene’s test (Sokal and Rohlf, 1981). Analysis of variance (ANOVA) was used to test for significant differences between treatments, depths and species categories. For all analyses

General trends in seed density and species richness

For the native species as a group greater number of germinants occurred in the smoke (S) (p < 0.001) (46% increase) and smoke-heat (SH) (p < 0.01) (41% increase) treatments than the heat (H) or control (C) treatments. In the field experiment the numbers of germinants were significantly more in smoked plots than control plots, being 1.4, 1.5, 4.8, 25 and 29 times greater for the introduced E. calycina, Hypochaeris glabra L., Gladiolus caryophyllaceus (Burm.f.) Poir., Ursinia anthemoides (L.) Poir.

Discussion

It is clear from this study that the deterioration in the vegetation condition of the studied Banksia woodland (Fisher et al., unpublished) is associated with significant changes in the soil seed bank, namely, a depletion of native species seeds, and a predominance of seeds of introduced species. These shifts in the soil seed bank are likely to be signs of shifts in Banksia woodland species and life form composition, and demonstrate the importance of considering the soil seed bank in

Conservation and management implications

The demonstrated shift in density and life form composition of the soil seed bank in Bold Park has changed the growth and regeneration of native species with alterations to ecosystem function (Fisher et al., unpublished) and processes (Fisher et al., 2006). The soil seed bank data demonstrate that native seed limitation in the poor condition Banksia woodland areas, PCe and PCp, will constrain recruitment of native annual grasses, annual herbs, perennial herbs, perennial sedges and shrubs as has

Acknowledgements

The authors acknowledge Deanna Rokich for her considerable advice and assistance and Ray Wills for his support with the in situ field experiment. Parts of this study were conducted while the first author was the recipient of a scholarship with the Botanic Gardens and Parks Authority.

References (58)

  • D.T. Bell et al.

    Effects of temperature on the germination of selected Australian native species used in the rehabilitation of bauxite mining disturbance in Western Australia

    Seed Science and Technology

    (1992)
  • D.T. Bell et al.

    Fire in the Kwongan

  • S.M. Bellairs et al.

    Canopy-borne seed store in three Western Australian plant communities

    Australian Journal of Ecology

    (1990)
  • P.J. Bellingham et al.

    Resprouting as a life history strategy in woody plant communities

    Oikos

    (2000)
  • D.L. Benoit et al.

    Factors influencing the precision of soil seed bank estimates

    Canadian Journal of Botany

    (1989)
  • D.W. Bigwood et al.

    Spatial pattern analysis of seed banks: an improved method of optimizing sampling

    Ecology

    (1988)
  • Botanic Gardens and Parks Authority

    Bold Park Environmental Management Plan 2000–2005

    (2000)
  • N. Burrows et al.

    Fire and plant interactions in forested ecosystems of south-west Western Australia

  • A.H. Cheam

    Longevity of Bromus diandrus Roth. Seed in soil at three sites in Western Australia

    Plant Protection Quarterly

    (1987)
  • A.J. De Villiers et al.

    Seed bank classification of the Strandveld Succulent Karoo, South Africa

    Seed Science Research

    (2002)
  • A.J. De Villiers et al.

    Germination strategies of Strandveld Succulent Karoo plant species for revegetation purposes: II. Dormancy-breaking treatments

    Seed Science Technology

    (2002)
  • A.J. De Villiers et al.

    Germination strategies of Strandveld Succulent Karoo plant species for revegetation purposes: I. Temperature and light requirements

    Seed Science and Technology

    (2002)
  • K.W. Dixon et al.

    The promotive effect of smoke derived from burnt native vegetation on seed germination of Western Australian plants

    Oecologia

    (1995)
  • M. Fenner et al.

    The Ecology of Seeds

    (2005)
  • Fisher, J.L., 1999. Revegetation Strategies within the Friends of Bold Park Bushland (Inc.) Natural Heritage Trust...
  • J.L. Fisher et al.

    Enhanced soil and nutrient status of a Western Australian Banksia woodland community invaded by Ehrharta calycina and Pelargonium capitatum

    Plant and Soil

    (2006)
  • A. Hansen et al.

    Growth and reproductive performance of a seeder and a resprouter species of Bossiaea as a function of plant age after fire

    Annals of Botany

    (1991)
  • R.J. Hobbs et al.

    Fire-related dynamics of a Banksia woodland in south-western Western Australia

    Australian Journal of Botany

    (1990)
  • R.J. Hobbs et al.

    Novel ecosystems: theoretical and management aspects of the new ecological world order

    Global Ecology and Biogeography

    (2006)
  • Cited by (67)

    • Active seed sowing can overcome constraints to passive restoration of a critically endangered vegetation type

      2021, South African Journal of Botany
      Citation Excerpt :

      Plant invasions are particularly problematic in Mediterranean-type ecosystems, which constitute global biodiversity hotspots. Effects have been studied in regions including Portugal (Marchante et al., 2003), California (D'Antonio et al., 2007) and Western Australia (Fisher et al., 2009), as well as South Africa (Van Wilgen et al. 1998) where the Working for Water Programme was initiated to deal with invasive species. Fynbos vegetation is a highly diverse sclerophyllous shrubland confined to the Mediterranean climate region in the southwest of South Africa, where many areas are invaded by alien tree species, particularly Australian acacias (Le Maitre et al. 2011).

    • Contrasting land use systems influence soil seed bank composition and density in a rural landscape mosaic in West Africa

      2019, Flora: Morphology, Distribution, Functional Ecology of Plants
      Citation Excerpt :

      The interaction between land use and soil depth did not have a significant effect on SSB richness, indicating that SSB richness varied less across the landscape with depth. This result is consistent with the findings of Fisher et al. (2009) who found no interaction between soil depth and site condition for species richness and added that site condition does not influence the depth of seed occurrence. There were low similarities in species composition of the SSBs from all of our studied land uses, suggesting this is possibly due to variation in the types and intensity of disturbances between the land uses (Savadogo et al., 2017).

    View all citing articles on Scopus
    View full text