RESEARCH REPORT
doi: 10.1111/emr.12139
Recovery after African Olive invasion: can a ‘bottom-up’ approach to ecological restoration work? By Peter Cuneo and Michelle R. Leishman
Summary Peter Cuneo is Manager – Natural Heritage at the Australian PlantBank, Royal Botanic Gardens and Domain Trust, (The Australian Botanic Garden, Mount Annan, Narellan Rd, Mount Annan, NSW 2567, Australia; Email: peter.
[email protected]; Tel: +61 2 46347915 +61 2 46476035). Michelle R. Leishman is a professor in the Department of Biological Sciences, Macquarie University; (Email: michelle.
[email protected]). This collaborative research project is part of a large-scale African Olive control programme at the Australian Botanic Garden, Mount Annan, NSW 2109, Australia.
Introduction
E
ucalypt woodlands are the most extensively cleared and modified vegetation type in Australia, particularly in the main eastern and south-western agricultural zones (Lindenmayer et al. 2005). Consequently, many of these woodland communities are now listed as threatened under state and national legislation, and the restoration of the remaining widely scattered and degraded remnants is a key component of strategies to conserve them (Prober & Thiele 2005). Cumberland Plain Woodland (CPW) is a grassy eucalypt woodland restricted to the Sydney Basin Bioregion, in an area known as the Cumberland Plain to the immediate west of Sydney, one of the earliest Australian regions utilised for agriculture and subsequent urban development. This vegetation type has a characteristic woodland structure with grassy understorey and a canopy height up to 30 m dominated by several eucalypt species. CPW is now listed as a critically endangered eco-
African Olive (Olea europaea ssp. cuspidata) is a densely crowned evergreen small tree, native to eastern Africa that is highly invasive in areas where it has been introduced, including Hawaii and Australia. Invasion by African Olive threatens Cumberland Plain Woodland, a critically endangered grassy eucalypt woodland from western Sydney, Australia, through the formation of a dense mid-canopy excluding the regeneration of native species. We established a 3-year field experiment to determine the effectiveness of direct seeding and fire, as techniques for early stage restoration of a 2 ha historically cleared and degraded Cumberland Plain Woodland site after the removal of African Olive. Direct seeding was able to re-establish a native perennial grass cover which was resistant to subsequent weed invasion and could be managed as an important first stage in woodland restoration with fire and selective herbicide. Fire was able to stimulate some germination of colonising native species from the soil seed bank after 15 years of African Olive invasion; however, germination and establishment of native shrubs from the applied seed mix was poor. We propose a ‘bottom-up’ model of ecological restoration in such highly degraded sites that uses a combination of direct seeding and stimulation of the soil seed bank by fire, which could be applicable to other degraded grassy woodland sites and plant communities. Key words: Cumberland Plain Woodland, direct seeding, ecological restoration, grassy eucalypt woodland, invasive Olea, soil seed bank.
logical community under NSW State and Australian Federal legislation, and remaining areas face further significant pressure from urban development. Land clearing remains the most significant threat to CPW, and the small remnants have a high edge-to-area ratio which makes them particularly vulnerable to ‘edge effects’ such as weed invasion. African Olive impacts on C u m b e r l a n d P l a i n Wo o d l a n d The dramatic recent expansion of the invasive exotic tree African Olive (Olea europaea L.subsp. cuspidata Wall ex G.Don Ciferri) poses the greatest invasive threat to CPW (New South Wales Scientific Committee 2009). African Olive is a small dense-crowned tree with a centre of natural distribution in eastern Africa. It was introduced into Australia in the midnineteenth century for horticultural purposes, and is now well established as a woody invasive plant in the Cumberland Plain and Hunter Valley regions of New South Wales (Cuneo & Leishman 2006).
African Olive is regarded as an environmental weed in Australia and Hawaii (Muyt 2001; Starr et al. 2003) where it forms a dense mid-storey, preventing native shrub and herb regeneration (Benson & Howell 1990). The progressive development of an African Olive canopy over 10+ years effectively limits the growth and recruitment of high light requiring understorey species, such as grasses and forbs, as well as the eucalypt canopy trees (Cuneo & Leishman 2013). Dense African Olive infestations are now established at a landscape scale throughout the Cumberland Plain region (Cuneo et al. 2009), particularly in hilly sites, historically cleared for grazing. Remnant CPW trees and understorey have persisted at many of these sites, extensively cleared in the late 1800s – early 1900s; however, the invasion of African Olive in recent decades has effectively prevented CPW regeneration. There is now strong interest in the control and removal of mature (15+ years) African Olive stands in such sites using techniques such as
ª 2014 The Authors ECOLOGICAL MANAGEMENT & RESTORATION VOL 16 NO 1 JANUARY 2015 Ecological Management & Restoration ª 2014 Ecological Society of Australia and Wiley Publishing Asia Pty Ltd
33
Ecological Society of Australia
RESEARCH REPORT
mechanical mulching/processing, and subsequent restoration interventions to re-establish woodland plant diversity lost through African Olive invasion. For grassy woodland communities such as CPW, the re-establishment of early successional understorey species would be a highly desirable first step in the restoration of these ecosystems. In this study, we investigated options for ‘bottom-up’ ecological restoration for highly degraded sites, focusing on the reestablishment of CPW understorey and mid-canopy species. Specifically, we conducted a 3-year factorial field experiment to address the following questions: (i) What is the vegetation response following the removal of well-established dense African Olive? (ii) Can fire be used to stimulate the germination of native species from the soil seed bank, following the removal of the African Olive canopy? and (iii) Does direct seeding of native grasses and shrubs promote native species abundance and diversity after the removal of African Olive canopy?
Methods Soil testing Comparative soil tests between African Olive invaded sites and native woodland were taken at the commencement of the experiment, to determine whether there were any significant differences in soil pH or key soil elements such as phosphorus, potassium and calcium which may impact on restoration success. Soil sampling was conducted at three site types: (i) native Cumberland Plain vegetation (woodland and grassland), (ii) dense African olive site (15+ years age) and (iii) dense African olive site (15+ years age) after clearing and mulching of all biomass. Sampling sites were located within the Australian Botanic Garden, Mount Annan, and an adjacent property of similar geology and vegetation where the land-use history was known. At each site type, six soil samples from random locations were taken from the top ~100 mm of the soil profile after the removal of surface organic matter. Samples were then combined to give a com34
posite sample for each site. Soil samples were analysed for pH, salinity (electrical conductivity), nitrogen (nitrate), total phosphorus, potassium and calcium. All analyses were conducted following the methods of Rayment and Higginson (1992) by Sydney Environmental and Soil Laboratory (Thornleigh, NSW). Fi e l d e x p e r i m e n t The study site was located at The Australian Botanic Garden, Mount Annan, in south-west Sydney, a region which has a temperate climate and an annual rainfall ranging from 759 to 922 mm. Dense stands of African Olive have established throughout the region, including many of the CPW remnants within the Australian Botanic Garden during the past 25 years. Also located within the botanic garden are high-quality examples of CPW where African Olive invasion has been excluded. These were used as reference sites for plant diversity. An experimental site measuring 18 m 9 30 m was located in a dense 15-year-old stand of African Olive established beneath several emergent large remnant Grey Box trees with an east-facing aspect (see Fig. 1). Historical records and archival aerial photography dating back to 1961 indicate that the experimental site had been cleared for grazing with some remnant native trees and shrubs retained. Vegetation consisted of dense African Olive canopy to ~13 m
canopy height comprised of trees up to a maximum trunk diameter of 30 cm (see Fig. 2). There are typically no understorey plants beneath the dense Olive canopy, providing an initially bare site for the experiment following African Olive removal. Site preparation and clearing commenced in November 2005, with African Olive trees cut with a chainsaw to ground level and cut stumps treated with undiluted glyphosate (360 g/L). All cut tree material was stacked onsite and later mechanically chipped in February 2006 using a Bandit 65 wood chipper (Bandit Industries, Inc., Remus, MI, USA), with all Olive biomass subsequently spread evenly across the experimental site at an average depth of 50 mm. The experimental site soil was a heavy clay loam characteristic of the shale-derived soils of the Cumberland Plain region. A factorial experiment was established at the site consisting of twenty 2 m 9 2 m plots, allowing five replicates of the four treatments across the site. Due to the high cost of mechanical clearing, establishment costs and the labourintensive nature of a long-term field experiment, the treatments were not replicated at other sites, with the study being designed as an exploratory analysis of potential restoration techniques. The following treatments were randomly allocated to plots: (i) control – no treatment,
Figure 1. View of experimental site prior to treatment with typical African olive invasion of Cumberland Plain Woodland understorey.
ECOLOGICAL MANAGEMENT & RESTORATION VOL 16 NO 1 JANUARY 2015 ª 2014 The Authors Ecological Management & Restoration ª 2014 Ecological Society of Australia and Wiley Publishing Asia Pty Ltd
RESEARCH REPORT
Figure 2.
Interior view of the experimental site, prior to clearing.
plots covered in mulched Olive material; (ii) burn – mulched Olive material burnt and plots left to allow regeneration; (iii) sow seed – native seed mix applied over mulched Olive material; and (iv) burn and sow seed – mulched Olive material burnt and native seed mix applied to bare soil (hard-seeded legume seeds for treatments 3 and 4 were incorporated into the soil to a depth of ~20 mm prior to burning). Fire treatments were carried out on 9th March 2006 (air temperature 33°C) after several weeks of dry weather to ensure dry fuel conditions. A medium-hot intensity fire was achieved (max flame height ~1 m) (See Fig. 3) with ignition at plot edges and even combustion across the
plots, assisted by aspiration with a leaf blower. All plots were allowed to burn for 2.5 hours, to ensure soil heat penetration and total fuel consumption. African Olive produces a dense, hard wood, which burns with a very high calorific value that is comparable to or greater than most eucalypt species (Eberhard 1990), resulting in a high likelihood that temperatures required to break dormancy of hard-seeded species (Auld & O’Connell 1991) were delivered at the required duration to the top 5 cm of the soil surface. The addition of native seed was a key component of the field experiment. The native seed mix included the following native shrub and grass species that occur in CPW and are known colonisers: Hick-
ory Wattle (Acacia implexa), Blackthorn (Bursaria spinosa), Native Indigo (Indigofera australis), Weeping Grass (Microlaena stipoides), Common Wheatgrass (Elymus scaber), Tall Windmill Grass (Chloris ventricosa) and Kangaroo Grass (Themeda triandra). All seed was wild harvested during summer 2005/2006 at the Australian Botanic Garden, Mount Annan, and processed using a zigzag gravity seed aspirator (Selecta machinefabriek BV, The Netherlands). to eliminate nonfilled seed and checked using a cut test method to ensure >70% viable seed. Eucalypt seed was not added to the seed mix, as the retention of several mature overhanging Grey Box (Eucalyptus moluccana) provided consistent natural deposition of branches and seed capsules across the experiment site. Details of the native seed mix are provided in Table 1. Plots were monitored three times per year for 3 years. All plants in the plots, including exotic species, were identified to species level using Harden (1990– 1993), and their abundance was assessed visually and recorded as Braun–Blanquet abundance scores. Exotic species were not removed from the plots, but they were controlled in the areas surrounding and between the plots, by either hand or herbicides. Statistical analyses Percentage plot cover for native and exotic species was analysed using a linear mixed model incorporating a REML (restricted maximum likelihood) procedure. The mid-points of the Braun–Blanquet cover classes were used to calculate mean percentage cover prior to analysis. Species richness (count type) data were analysed using a Poisson generalised linear mixed model (GLMM). Both factors (burn, seeded) were treated as fixed. All analyses were completed using Genstat version 12 (VSN International 2010).
Results Soil chemistry Figure 3.
Burn treatments being carried out at the establishment of the experimental plots.
Comparisons between African olive forest sites and native woodland sites indicated
ª 2014 The Authors ECOLOGICAL MANAGEMENT & RESTORATION VOL 16 NO 1 JANUARY 2015 Ecological Management & Restoration ª 2014 Ecological Society of Australia and Wiley Publishing Asia Pty Ltd
35
RESEARCH REPORT
Table 1. Species used in the native seed mix that was applied to the ‘seeded’ plots in the restoration experiment
Species Hickory Wattle Acacia implexa (Fabaceae, Mimosoideae) Native Indigo Indigofera australis (Fabaceae) Blackthorn Bursaria spinosa (Pittosporaceae) Tall Windmill Grass Chloris ventricosa (Poaceae) Weeping Grass Microlaena stipoides (Poaceae) Kangaroo Grass Themeda triandra (Poaceae) Common Wheatgrass Elymus scaber (Poaceae)
Ecology/comments
Total number of seeds sown/4 m² plot*
Small tree with green falcate leaves, pale yellow flowers and fibrous trunk. Hard-seeded legume, likely to form persistent soil seed bank with germination stimulated by disturbance such as fire Spreading shrub to 2.5 m occurs in understorey and woodland edges. Pink flowers in early spring. Hard-seeded legume, likely to form persistent soil seed bank and germination stimulated by disturbance such as fire Spreading native shrub to 5 m with thorny branches and white flowers during summer. Key habitat and understorey shrub in the Cumberland Plain
125
Erect, stoloniferous native perennial grass to 1 m with branched stems. Colonising grass
2000
Slender tufted grass to 70 cm. Partial shade area coloniser which can form monocultures. Responsive to good soil moisture. Cool season grower
300
Tufted perennial grass to 1 m, often forming dense long-lived stands which produce abundant seed in summer. Bronze/red foliage during winter
8400 (combination seed and floret used)
Loosely tufted perennial grass to 1 m, forming occasional patches in woodland areas. Cool season grower
Unknown, small amount present in mechanically harvested Themeda triandra seed/floret material
200
1000
No seed pretreatments were used.
Native species cover and richness
36
Control Burn Burn/seed Seeded
0
4
8
12
7
16
20
24
28
32
Control Burn Burn/seed Seeded
6 5 4 3 2 1 0
36
0
4
8
Time (months)
12
16
20
24
28
32
36
Time (months)
(c)
(d) Exotic species richness
70
% Exotic cover
There were significant differences between the treatments in native species cover. Overall, the seeded treatments produced greater native species cover. For native cover (Fig. 4a), there was a significant treatment 9 time interaction (F = 2.80; df = 24, 128; P < 0.001), indicating the time profiles differed across the treatments. ‘Control’ always had the lowest native cover, and ‘burn’ was never significantly different from the ‘control’, despite being slightly higher. ‘Seeded’ and ‘burn/seeded’ had significantly greater native cover than the control from 16 months, but by 36 months, there were no significant differences. The native cover established in the ‘seeded’ and ‘burn/seeded’ treatments consisted predominantly of native grasses, with Common Wheatgrass establishing in the first few months (autumn), followed by Weeping Grass which eventually dominated
(b) 90 80 70 60 50 40 30 20 10 0
Native species richness
(a)
% Native cover
no clear differences in the soil chemistry between the sites (see Table 3). The only clear trend was that N concentration was higher at the experimental site, as a result of the early decomposition phase of the mulched olive biomass.
Control Burn Burn/seed Seeded
60 50 40 30 20 10 0 0
4
8
12
16
20
24
Time (months)
28
32
36
10 9 8 7 6 5 4 3 2 1 0
Control Burn Burn/seed Seeded
0
4
8
12
16
20
24
28
32
36
Time (months)
Figure 4. (a) Mean percentage cover of native species for 4 m² experimental plots over 3 years following removal of African Olive and application of four experimental treatments: no treatment; burning only; burning plus direct seeding; direct seeding only. Values are mean standard error. (b) Mean native species richness for 4 m² experimental plots over 3 years following the removal of African Olive and application of four experimental treatments: no treatment; burning only; burning plus direct seeding; direct seeding only. Values are mean standard error. (c) Mean percentage cover of exotic species for 4 m² experimental plots over 3 years following the removal of African Olive and application of four experimental treatments: no treatment; burning only; burning plus direct seeding; direct seeding only. Values are mean standard error. (d) Mean exotic species richness for 4 m² experimental plots over 3 years following the removal of African Olive and application of four experimental treatments: no treatment; burning only; burning plus direct seeding; direct seeding only. Values are mean standard error.
ECOLOGICAL MANAGEMENT & RESTORATION VOL 16 NO 1 JANUARY 2015 ª 2014 The Authors Ecological Management & Restoration ª 2014 Ecological Society of Australia and Wiley Publishing Asia Pty Ltd
RESEARCH REPORT
seeded treatment plots. Germination of other native species in these seeded plots was low, with Native Indigo and Tall Windmill Grass producing no seedlings, and only a few scattered plants established for Hickory Wattle and Blackthorn. Germination of native species from control plots was low; however, the ‘burn’ only treatment resulted in the highest germination and establishment of the mid-canopy species Hickory Wattle and understorey forb Kidney Weed (Dichondra repens) from the soil seed bank. The highest germination of Grey Box was observed in the ‘burn/seeded’ plots after 16 months (20 seedlings) during moist autumn/winter conditions; however, these did not survive through the subsequent hot dry spring and only one Grey Box seedling was alive after 36 months. Trends in native species richness during the experiment were similar to percentage cover with a significant treatment 9 time interaction (F = 1.81; df = 24, 126; P = 0.019). The ‘control’ treatment consistently had the lowest native species richness across months (see Fig. 4b). ‘Seeded’ and ‘burn/seeded’ most frequently had higher native species richness compared with the ‘control’; however, these differences were not significant at the conclusion of the experiment as perennial grasses became dominant.
Exotic species cover and richness Overall, control plots had the greatest exotic cover (maximum at 24 months), with ‘seeded’ and ‘burn/seeded’ plots frequently having a significantly lower exotic cover. At 32 and 36 months, there were no treatments significantly different to control and the trend suggested that this would continue (see Fig. 4c). There was a significant treatment 9 time interaction (F = 2.09; df = 24, 128; P = 0.005), indicating different time profiles across the treatments. Statistically significant differences between control and all treatments occurred between 16 and 28 months. For control plots, the exotic species cover was dominated by the short-lived perennial shrub Blackberry Nightshade (Solanum nigrum) individuals of which reached full development at 20– 24 months, (see Fig. 5) and then senesced towards the end of the experiment. Exotic species richness followed a similar pattern to percentage cover across the treatments (see Fig. 4d), and there was a significant treatment 9 time interaction (F = 2.93; df = 24, 127; P < 0.001); however, the ‘seeded’ treatment produced a greater exotic species richness than the control, while ‘burn’ had the lowest exotic species richness. A significant proportion of the exotic species richness in the
Figure 5. View of experimental site after 2 years, showing typical native grass establishment in direct seeded plots (centre foreground) and establishment of Solanum nigrum in control plots (left foreground). Olive mulch areas surrounding plots have been spot herbicide treated for weeds.
seeded plots was introduced grasses (contamination from mechanically harvested seed) such as Quaking Grass (Briza maxima), Perennial Rye Grass (Lolium perenne), Prairie Grass (Bromus catharticus) and Bearded Oats (Avena barbata). By contrast, ‘burn’ plots with the lowest exotic species richness had no grasses present, with annual/biennial weeds such as Spear Thistle (Cirsium vulgare), Tall Fleabane (Conyza sumatrensis) and Fireweed (Senecio madagascariensis) being the predominant species (Table 2). African Olive establishment Overall, African Olive germination and plant establishment were low. Some germination was recorded at 8 months, during favourable moist autumn conditions in 2007 (see Fig. 6). The more open soil surface of the ‘burn’ plots supported the highest African Olive seedling establishment at 36 months, followed by ‘control’ plots. The highest number of African Olive seedlings (10) was observed in the ‘seeded’ treatment plots. However, at 36 months, the ‘seeded’ and ‘burn/ seeded’ plots were observed to have the lowest number of established African Olives.
Discussion Revegetation of sites degraded by threats including clearing, weed invasion and overgrazing is critical for promoting native plant diversity, controlling weeds and achieving sustainable woodland landscapes (Prober & Thiele 2005). However, lack of knowledge on how best to achieve successful restoration hinders many projects. In this study, we investigated the vegetation response following control of longestablished and dense African Olive in Cumberland Plain Woodland and assessed the potential of burning and direct seeding to assist in the reinstatement of ecological processes that are favourable to the regeneration and reintroduction of native vegetation. Our initial comparison of soil properties between native woodland areas and African Olive invasion sites (Table 3) indicated no major differences for soil pH or key soil elements such as phosphorus, potassium and calcium; thus, ecological
ª 2014 The Authors ECOLOGICAL MANAGEMENT & RESTORATION VOL 16 NO 1 JANUARY 2015 Ecological Management & Restoration ª 2014 Ecological Society of Australia and Wiley Publishing Asia Pty Ltd
37
RESEARCH REPORT
Table 2. Maximum cover abundance for native and exotic species recorded in the four experimental treatments (mean percentage cover and standard errors are indicated). Species that were added as seed are indicated by *. Species in bold achieved >25% cover abundance in any one plot during the experiment
Species Native species Acacia implexa* Aristida ramosa Austrodanthonia sp Bursaria spinosa* Cayratia clematidea Cyperus sp. Dichelachne micrantha Dichondra repens Einadia hastata Einadia nutans ssp linifolia Einadia polygonoides Elymus scaber Eucalyptus sp. Geranium homeanum Microlaena stipoides* Themeda triandra* Exotic species Arauja sericifera Asparagus asparagoides Avena barbata Briza maxima Briza minor Bromus catharticus Bromus mollifolius Cirsium vulgare Conyza sumatrensis Cyclospermum leptophyllum Gaemochaeta americana Hypocaris radicata Lactuca saligna Lollium perenne Lycium ferocissimum Mellilotus indica Nasella neesiana Olea europaea ssp cuspidata Paspalum dilitatum Plantago lanceolata Senecio madagascariensis Silybum marianum Solanum nigrum Sonchus oleraceus
Control
Burn
Seeded
3.0 2.9
14.0 6.6
0.1 0.1 0.12 0.09
13.0 12.3 0.1 0.1 0.02 0.02 20.0 12.8 0.1 0.1 0.1 0.1 0.1 0.1
10.5 0.02 0.02 18.0
7.3 0.02 0.02 8.4
0.5 0.5 0.02 0.02 0.02 0.02 0.22 0.11
0.02 0.02 3.0 3.0 1.6 0.5 3.1 2.9 0.02 0.02
0.4 0.4 30.5 18.5 0.02 0.02
0.02 0.02 9.5 3.3 0.12 0.89
3.0 3.0
0.02 0.02
0.12 0.12
0.04 0.02
3.0 3.0
42.5 19.6 0.12 0.09
3.1 2.9 0.2 0.1
0.12 0.09 3.1 2.97
Burn/seed 0.12 0.12 0.12 0.22 3.02
0.09 0.09 0.09 0.11 2.99
0.04 0.02
0.4 0.1
58.0 14.1 0.1 0.1
53.0 14.6 0.3 0.12
62.5 7.9 3.72 2.85
48.5 16.4 23.5 11.7
0.02 0.02
0.5 0.02 0.12 0.1 0.24 4.2 0.32 3.06
0.5 0.02 0.09 0.1 0.1 2.73 0.11 2.98
3.0 0.1 0.1 9.0 1.2 0.3 0.04
3.0 0.1 0.1 3.67 0.53 0.12 0.02
0.22 0.11 0.6 0.11
0.22 0.04 0.2 0.12 0.02 3.22 0.14
0.11 0.02 0.12 0.09 0.02 2.94 0.09
2.1 0.4 0.02 0.02 0.02 0.02 0.1 0.1 0.32 0.11 0.12 0.09 7.5 7.4 0.32 0.11
Acacia implexa seed was not added to ‘burn’ treatments.
restoration had a good chance of long-term success. Known phosphorus-/pH-sensitive native species such as Acacias grew strongly at the cleared olive experimental site without any signs of high nutrient chlorosis/toxicity. A key observation from the experimental site was the very low germination and establishment of African Olive, particularly considering the abundant seed crop that is commonly produced in its invasive 38
range (von Richter et al. 2005). The overall low density of African Olive seedlings at other mechanically cleared sites at Mt Annan, combined with the relatively short soil seed bank persistence in the soil (~2.4 years) determined in previous studies (Cuneo et al. 2010), indicates that seedlings occur at a level that could be readily controlled by spot herbiciding or hand removal depending on the size of the site.
Secondly, the presence of many annual and biennial exotic species was seen to be transient and did not significantly impact on the establishment of most native species that germinated at the same time. Thirdly, the response from control plots indicated that, despite significant decomposition of the mulch layer and soil exposure by the third year, none of the original woodland grasses and few of the forbs germinated. When considered in the context of what is known about the longevity of perennial herbs, it suggests that these herbs had largely been lost from the soil seed bank. The plant life cycles and soil seed banks in CPW are distinct from nearby fire-prone Sydney sandstone communities (von Richter et al. 2004) which have a high proportion of obligate seeders and persistent soil seed banks. Lunt (1997) found that many perennial native grassy woodland species found in Victorian grasslands (comparable to CPW) did not form a persistent soil seed bank. When considering this tendency of grassland herbs to have a relatively short-term soil seed bank, the limited dispersal ability of many CPW species – particularly in combination with a history of grazing/clearing followed by olive invasion, a depauperate soil seed bank would be expected. The burn treatment was included in the experiment to test for the presence of any persistent hard-seeded species. Burning proved to be successful in stimulating the germination of hard-seeded mid-canopy CPW species such as the tree Hickory Wattle and the perennial herb Kidney Weed which had persisted for over 15 years in the soil seed bank. This is consistent with other studies of fire-prone woodland vegetation in Australia (Auld & O’Connell 1991; Bell et al. 1991; Hill & French 2003), where burning is able to break the physical dormancy of hardseeded species such as Acacia which typically form persistent soil seed banks (Richardson & Kluge 2008). Mulching is likely to have played some role in reducing native germination in the unburnt plots, as it was deliberately retained as a product of olive removal/processing. In cases where soil seed banks are likely to include high levels of weed seed,
ECOLOGICAL MANAGEMENT & RESTORATION VOL 16 NO 1 JANUARY 2015 ª 2014 The Authors Ecological Management & Restoration ª 2014 Ecological Society of Australia and Wiley Publishing Asia Pty Ltd
RESEARCH REPORT
Number of African olive seedlings
3.5
Seeded
3
Control
2.5
Burn
2
Burn/seeded
1.5 1 0.5 0
4
8
12
16 20 24 28 Time (months)
32
36
Figure 6. Mean number of African Olive seedlings observed for 4 m² experimental plots over 3 years following removal of African Olive canopy and application of four experimental treatments: no treatment; burning only; burning plus direct seeding; direct seeding only. Values are mean standard error.
a typical restoration approach to provide an improved starting point for revegetation is to apply repeated herbicide treatments over time or to scalp weedcontaminated topsoil (Gibson-Roy et al. 2010). At this typically hilly and sloping African Olive invasion site, topsoil scalping on dispersive sodic clay soils was not considered appropriate; however, an alternative option of covering cleared sites with the large volume of wood chip generated from the mechanical removal of the Olive biomass was considered to be the most practical. This mulch layer was used as a fuel for burning or a weed suppressant and substrate for seeding with native understorey species. Management actions that address multiple constraints (e.g. burning and adding native seeds) often prove more successful than would either approach alone (Holmes 2001), and a key objective of the study was to also assess the effectiveness of direct seeding native grasses and shrubs to promote native species cover and diversity. Direct seeding proved to
be effective, with seed application treatments achieving the highest native species cover. Native grasses were most successfully established, in particular Weeping Grass which eventually dominated seeded treatment plots. Importantly, the ‘seeded’ and ‘burn/seeded’ plot treatments consistently had the lowest cover of exotic species, indicating the degree of weed resistance provided by the dense native grass cover. It was apparent, however, that long-term restoration success would require a balance between competitive native grass swards and the creation of interstitial gaps (e.g. through use of fire) for progressive recruitment of a wider range of native species. Direct seeding did not, however, result in significant establishment of shrub and tree species, with no germination of Native Indigo, and only a few scattered plants established for Hickory Wattle and Blackthorn. Given the expense of the seed of less common species, it may be more efficient to trial appropriate pre-treatments for this species and/or add shrub and tree tubestock at
later stages as restoration proceeds. Restoration efforts at degraded sites commonly rely on the replanting of locally occurring trees and shrubs; however, this ‘topdown’ approach does not necessarily result in a long-term improvement of understorey plant diversity in grassy woodland ecosystems (Wilkins et al. 2003; Munro et al. 2009; Nichols et al. 2010). As an alternative, based on the results and observations from the field experiment, we propose a ‘bottom-up’ restoration approach for CPW sites degraded by African Olive invasion. In this technique, priority is given to the quick establishment of understorey species to form an early successional grassland stage through the addition of propagules via direct seeding, and potential exists for later establishment of trees if they do not already occur onsite. This native grass cover has a number of advantages to land managers, including protection from soil erosion, competitive resistance to weed re-invasion and the opportunity to manage broad leaf weeds through the use of selective herbicide or fire. Importantly, the re-establishment of this grassy understorey, as well as colonising acacias from the soil seed bank, is a positive first step in restoring a degraded CPW site and commencement of a trajectory towards native woodland. We conclude that historically cleared CPW degraded by African Olive invasion is not resistant to restoration and that this restoration approach is worthy of testing in similar situations in other temperate grassy woodland sites. Further research and refinement are needed in the techniques of direct seeding to achieve greater germination and establishment of shrubs and canopy tree species. Some studies suggest that seedbed preparation, seed
Table 3. Soil test results for the major chemical properties for African olive sites and Native woodland
Test site African olive sites Dense African Olive, cleared and mulched experiment site Dense African Olive undisturbed Native vegetation sites Cumberland Plain Woodland Native grassland with shrubs Native grassland with scattered trees
pH (in H2O)
EC (1:5) dS/m
6.7
0.13
7.1
0.23
6.0 7.1 7.0
0.05 0.09 0.13
Nitrogen (nitrate) mg/kg
Phosphorus (total P) mg/kg
Potassium mg/kg
Calcium mg/kg
685
598
3166
5.9
598
1126
5375
8 4 8.7
425 597 448
203 301 822
1489 3206 5033
38
ª 2014 The Authors ECOLOGICAL MANAGEMENT & RESTORATION VOL 16 NO 1 JANUARY 2015 Ecological Management & Restoration ª 2014 Ecological Society of Australia and Wiley Publishing Asia Pty Ltd
39
RESEARCH REPORT
soil contact and soil moisture are critical areas requiring further research attention (Gibson-Roy et al. 2007) to optimise the use of a scarce and valuable seed resource. For land managers dealing with the impacts of invasive woody species on temperate grassy woodlands, we have presented field data showing how a combination of native soil seed bank management, fire and direct seeding could provide a restoration method with potential for testing in similar vegetation communities.
Acknowledgements We would like to thank staff at the Australian Botanic Garden, Mount Annan, for their support and assistance; Geoff Goodwin, Sarah Dempster and Craig Ward for clearing and wood chipping the research site; and Lotte von-Richter for assistance with the identification of seedlings. We also thank Simon Leake and Tiffany Carroll-MacDonald of Sydney Environmental and Soil Laboratory for soil testing and interpretation of results. Peter Thompson and Louise Helby from University of Sydney provided valuable assistance with statistical analyses.
References Auld T. D. and O’Connell M. A. (1991) Predicting patterns of post-fire germination in 35 eastern Australian Fabaceae. Australian Journal of Ecology 16, 53–70. Bell D. T., Plummer J. A. and Taylor S. K. (1991) Seed germination ecology in southwestern western Australia. Botanical Review 59, 24– 73. Benson D. H. and Howell J. (1990) Taken for Granted: The Bushland of Sydney and its Suburbs. Kangaroo Press, Kenthurst. Cuneo P. and Leishman M. R. (2006) African Olive (Olea europaea subsp. cuspidata) as an environmental weed in eastern Australia: a review. Cunninghamia 9, 545–577.
40
Cuneo P. and Leishman M. R. (2013) Ecological impacts of invasive African olive (Olea europaea ssp.cuspidata) in Cumberland Plain Woodland, Sydney, Australia. Austral Ecology 38, 103–110. Cuneo P., Jacobson C. R. and Leishman M. R. (2009) Landscape-scale detection and mapping of invasive African olive (Olea europaea L. subsp.cuspidata Wall ex G. Don Ciferri) in SW Sydney, Australia using satellite remote sensing. Applied Vegetation Science 12, 145–154. Cuneo P., Offord C. A. and Leishman M. R. (2010) Seed ecology of the invasive woody plant African olive (Olea europaea subsp. cuspidata): implications for management and restoration. Australian Journal of Botany 58, 342– 348. Eberhard A. A. (1990) Fuelwood calorific values in South Africa. South African Forestry Journal 152, 17–22. Gibson-Roy P., Delpratt J. and Moore G. (2007) Restoring Western (Basalt) Plains grassland. 2. Field emergence, establishment and recruitment following direct seeding. Ecological Management & Restoration 8, 123–132. Gibson-Roy P., Moore G. and Delpratt J. (2010) Testing methods for reducing weed loads in preparation for reconstructing species-rich native grassland by direct seeding. Ecological Management & Restoration 11, 135– 139. Harden G. (1990–1993) Flora of New South Wales. NSW University Press, Kensington. Hill S. J. and French K. (2003) Response of the soil seed-bank of Cumberland Plain Woodland to heating. Austral Ecology 28, 14–22. Holmes P. M. (2001) Shrubland restoration following woody alien invasion and mining: effects of topsoil depth, seed source, and fertilizer addition. Restoration Ecology 9, 71–84. Lindenmayer D., Crane M. and Michael D. (2005) Woodlands: a disappearing landscape. CSIRO Publishing, Collingwood, Vic. Lunt I. D. (1997) Germinable soil seed banks of anthropogenic native grasslands and grassy forest remnants in temperate south-eastern Australia. Plant Ecology 130, 21–34. Munro N. T., Fischer J., Wood J. and Lindenmayer D. B. (2009) Revegetation in agricultural areas: the development of structural complexity and floristic diversity. Ecological Applications 19, 1197–1210. Muyt A. (2001) Bush Invaders of South East Australia – A Guide to the Identification and Control of Environmental Weeds in South Eastern Australia. R.G. & F.J. Richardson, Meredith, Vic.
New South Wales Scientific Committee (2009) Cumberland Plain Woodland in the Sydney Basin Bioregion – critically endangered ecological community listing, final determination. [Accessed 9 June 2014.] Available from URL: http://www.environment.nsw.gov.au/ determinations/cumberlandwoodlandsFD. htm. Nichols P. W. B., Morris E. C. and Keith D. A. (2010) Testing a facilitation model for ecosystem restoration: does tree planting restore ground layer species in a grassy woodland? Austral Ecology 35, 888–897. Prober S. and Thiele K. R. (2005) Restoring Australia’s temperate grasslands and grassy woodlands: integrating function and diversity. Ecological Management & Restoration 6, 16– 27. Rayment G. E. and Higginson F. R. (1992) Laboratory Handbook of Soil and Water Chemical Methods. Inkata Press, Melbourne, Vic. Richardson D. M. and Kluge R. L. (2008) Seed banks of invasive Australian Acacia species in South Africa: role in invasiveness and options for management. Perspectives in Plant Ecology, Evolution and Systematics 10, 161–177. von Richter L., Little D. and Howell J. (2004) Firerelated responses of seeds in a grassy woodland in western Sydney. Proceedings of Fifth Australian Workshop on Native Seed Biology, Brisbane, Queensland 21–23 June 2004. (eds. S. W. Adkins, P. J. Ainsley, S. M. Bellairs, D. J. Coates, L. C. Bell) Australian Centre for Minerals Extension and Research, Brisbane, pp. 105–115. von Richter L., Little D. and Benson D. H. (2005) Effects of low intensity fire on the resprouting of the weed African Olive (Olea europaea subsp. cuspidata) in Cumberland Plain Woodland, Western Sydney. Ecological Management & Restoration 6, 230–232. Starr F., Starr K. and Loope L. (2003) Olea europaea subsp. cuspidata. United States Geological Survey, Biological Resources Division, Haleakala Field Station, Maui, Hawai’i. [Accessed 9 June 2014.] Available from URL: http://www.starrenvironmental. com/publications/species_reports/pdf/olea_ europaea_subsp_cuspidata.pdf. VSN International (2010) Genstat for Windows, 12th edn. VSN International, Hemel Hempstead, UK. Wilkins S., Keith D. A. and Adam P. (2003) Measuring success: evaluating the restoration of a grassy eucalypt woodland on the Cumberland Plain, Sydney, Australia. Restoration Ecology 11, 489–503.
ECOLOGICAL MANAGEMENT & RESTORATION VOL 16 NO 1 JANUARY 2015 ª 2014 The Authors Ecological Management & Restoration ª 2014 Ecological Society of Australia and Wiley Publishing Asia Pty Ltd
