• carbon addition;
  • fire;
  • grassy woodland;
  • seed addition;
  • soil nitrate


  1. Top of page
  2. Abstract
  9. Supporting Information

Cumberland Plain grassy woodland in western Sydney has been reduced to less than 12% of its pre-settlement distribution; efforts to restore it on cleared and grazed sites within its former distribution have met with mixed success. Elevated soil nitrate levels, coupled with propagule and establishment limitation, have been identified as barriers to restoration in other grasslands. Our study used a factorial combination of carbon addition, fire and native seed addition to test whether these barriers operated on a former Cumberland plain woodland site dominated by exotic perennial grasses. Replicate field plots were established in November 2004; fire plots were burnt in December 2004; carbon was then added as sugar every 3 months until September 2005; and seeds of five native grasses were added in January 2005. Carbon addition significantly reduced soil nitrate, the effect appearing in October 2005. Carbon addition and fire each reduced the total abundance of exotics; when combined, they halved the abundance of the two dominant exotic grasses. Total abundance of native species responded positively to carbon and seed addition, but significant responses to carbon were not detected for individual species. Abundance of two native grasses responded positively to fire; after treatment the native proportion of total abundance rose from 26% on controls to 44–65% on carbon and/or fire plots. Exotic species richness was decreased independently by carbon addition and fire. Native species richness was increased independently by fire and seed addition. All five native grasses established sporadically, but only on carbon and/or fire plots. The three treatments each significantly and independently affected species composition, which showed the greatest change when all three were applied. The three treatments collectively increased the proportion of natives in measures of both plant abundance and species richness. The study confirmed that elevated soil nitrate, plus propagule and recruitment limitation are barriers to restoring this grassy woodland on cleared and grazed sites.


  1. Top of page
  2. Abstract
  9. Supporting Information

In ecosystems that have been used for agricultural activities for prolonged periods, restoration is often undertaken to reverse the effects of land degradation. A soundly based ecosystem model can be critical to the success of these efforts allowing correct decisions to be made about whether the degradation can be reversed naturally by the system without intervention, or whether intervention is required to overcome barriers to restoration. If barriers to restoration are operating, identifying and then overcoming them are prerequisites for success (Suding & Hobbs 2009).

Efforts to restore grasslands and grassy woodlands have been common in a number of continents, but have met with mixed success (Bartolome et al. 2009; Prober et al. 2009; Standish et al. 2009; Bond & Parr 2010). A barrier to restoration often encountered has been altered nutrient cycles: degradation and invasion of exotics has been accompanied by elevated levels of nutrients, commonly soil nitrate (Perry et al. 2010) and also phosphorus (Standish et al. 2009; Lindsay & Cunningham 2011).

Addition of carbon to the soil has been used to reduce soil nitrate (Perry et al. 2010). The additional carbon stimulated growth of soil microbes, leading to measureable increases in microbial biomass or activity and decreases in nitrogen mineralisation (Zink & Allen 1998; Baer et al. 2003; Corbin & D'Antonio 2004; Brunson et al. 2010). Soil nitrate levels declined on carbon-treated plots (Perry et al. 2010). This phenomenon has been used in a number of trials to test whether using carbon addition to reduce soil nitrate will decrease growth of exotics and favour native species. The underlying model being tested is that carbon addition, by lowering seasonally high available soil nitrate that currently favours exotics, reduces soil nitrate to its former levels, at which native species are competitively superior (Wedin & Tilman 1990; Prober et al. 2009).

Experimental trials of soil carbon addition have not always favoured natives over exotics. In some cases, no effect on either exotics or natives has been reported (Wilson & Gerry 1995; Cione et al. 2002; Corbin & D'Antonio 2004; Thomsen et al. 2006; Kardol et al. 2008; Doll et al. 2011). Others have reported a negative effect on exotic species, with little detectable effect on natives (Reever Morghan & Seastedt 1999; Alpert & Maron 2000; Paschke et al. 2000; Averett et al. 2002; Bleier & Jackson 2007; Brunson et al. 2010). Positive effects of carbon addition on natives have been reported in a more limited number of cases (Zink & Allen 1998; Perry et al. 2004).

Mixed results can occur, suggesting that species respond individually to nitrate reduction (Eschen et al. 2006). Blumenthal et al. (2003) found positive effects of carbon addition on the growth of six prairie species (but also one annual and three perennial weeds), and negative effects on six annual weeds. Baer et al. (2004) found positive effects of carbon addition on three native prairie grasses, but a negative effect on the dominant native grass Panicum virgatum. Spiegelberger et al. (2008) found that carbon addition reduced biomass of grasses and forbs in a mountain grassland, but not that of a major forb weed.

In Australia, positive effects of carbon addition on natives, and negative effects on exotics have been observed by Prober et al. (2005) and Prober and Lunt (2009) in degraded remnants of grassy White Box (Eucalyptus albens) woodland, which formerly occurred extensively in a band on the western slopes of the ranges, from southern Queensland to central Victoria. Remnants are small, highly fragmented, and annual exotics have invaded the grassy understory, formerly dominated by perennial native grasses (Prober et al. 2002a,b). This invasion has been accompanied by a change in soil nitrogen dynamics, with higher concentrations of nitrate found in late summer/autumn in soil of badly invaded remnants than in the remnants with good cover of perennial native grasses (Prober et al. 2002a, 2005). Addition of carbon in two degraded remnants strongly reduced the seasonal peak in soil nitrate to levels found under intact native grass swards, resulting in a reduction in the growth of exotic annuals and an increase in that of native grasses (Prober et al. 2005). Once carbon addition ceased, soil nitrate reverted to pre-treatment levels; however, Prober and Lunt (2009) demonstrated that Themeda australis swards, once established, maintained soil nitrate levels at low values similar to those observed at reference sites dominated by Themeda.

Grassy woodlands in eastern Australia occur also in a disjunct distribution in coastal valleys (Keith 2004). Cumberland Plain woodland (Benson 1992) occurs on soils derived from Wianamatta shales west and south-west of Sydney. Termed Shale Hills Woodland and Shale Plains woodland by Tozer (2003), it has been reduced to 7.7–11.3% of its previous distribution (Tozer 2003). The characteristic canopy trees are Eucalyptus tereticornis and E. moluccana; layers of smaller trees, some shrubs and an understory of native grasses and forbs are present (Benson 1992; French et al. 2000; Tozer 2003). Areas of former Cumberland Plain woodland that were cleared and grazed currently have ground layers where exotic species make up two-thirds of canopy cover and up to half of the species richness (Hill et al. 2005). However the invasive ground-layer species are mainly exotic perennial grasses (Hill et al. 2005), in contrast to the exotic annuals that invade White Box woodland, and the response of these grasses to nutrient manipulations is unknown. Few data exist comparing soils under pastures with those under remnant woodland in this system. A comparison of nutrient levels in the surface layer of pastures that formerly supported Cumberland Plain woodland, with three patch types (open, shrub or tree) within remnant woodland showed little difference for most nutrients, except nitrate and ammonium. Concentrations of nitrate and ammonium were highest in the pasture soils at three of the five sites surveyed, and second highest (after trees) at the remaining two sites (Fitzgerald 2009). A time-course study at one of these latter sites showed that over 9 months, the highest nitrate concentrations were found in the pasture soil surface layer on three of the four sampling occasions (Fitzgerald 2009).

While the effectiveness of soil carbon addition in favouring native ground layer species has been demonstrated in the white box woodlands, its usefulness in degraded remnants of coastal valley grassy woodlands is unknown. Other potential barriers to restoration of grassy woodlands on former agricultural land are propagule and recruitment limitation (Lunt & Morgan 2002; Clark & Davison 2001, 2004; Dickson & Foster 2008). The former can be tested for by providing propagules (Stevens et al. 2004; Sheley et al. 2006). Lack of recruitment opportunities have been tested for by removing above-ground biomass (clearing, grazing, burning) and/or disturbing the soil surface (Clark & Davison 2001, 2004; Prober et al. 2005; Lindsay & Cunningham 2011). The aim of this study was to trial carbon addition, fire and seed addition in an abandoned pasture that formerly supported Cumberland Plain Woodland, to test whether elevated soil nitrate, together with propagule and recruitment limitation act as potential barriers to restoration.


  1. Top of page
  2. Abstract
  9. Supporting Information

The study area

The study area was located at the Eastern Creek demonstration site, Doonside, which is part of the Western Sydney Parklands. The precinct, at 33°47′15S, 150°52′E, contains 137 hectares of former grazing land and small pockets of remnant woodland and is now reserved for conservation. The area receives an average annual rainfall of 871.3 mm with a higher proportion of the total annual rainfall occurring in the warmer months of the year. Summer rainfall is usually less variable than winter rainfall. Mean maximum temperature for January is 28°C, while the mean July minimum temperature is 6°C. Rainfall over the experimental period was below the monthly average rainfall for most months with the annual rainfall for 2005 only 626.7 mm. Soil at the site belongs to the Blacktown group; these are residual soils developed on Wianamatta Group shales (Bannerman & Hazelton 1990). Livestock grazing occurred on the study area in the past, but had not occurred in the last 10 years. Most of the area was cleared and previously used for grazing, and is now dominated by exotic grasses as described by Hill et al. (2005).

Experimental design and field sampling

Carbon addition (control and sugar addition), fire (unburnt and burnt) and seed addition (unseeded and seeded) were combined factorially to give eight treatments. Three blocks, about 50 m apart were established in November 2004 in treeless areas where exotic grasses dominated the vegetation. Eight plots (5 × 5 m) were established with a 1-m buffer zone between plots, in a 2 × 4 arrangement within each block. Treatments were randomly assigned to plots within blocks.

Carbon in the form of white sugar was applied at a rate of 0.5 kg m−2 (0.21 kg C per m−2) every 3 months from January to September 2005. Sugar was applied to the surface of the soil on days when there was a likelihood of rain within the next 48 h. A single fire application was undertaken at the beginning of January 2005.

Seed of five native grasses was applied to the soil surface at the end of January 2005 after the fire application. Seed-bearing material was collected from local areas by Greening Australia. These species were selected as they were native to the area and were available from Greening Australia in large enough quantities for experimental use. Two of the seeded species were present in the experimental plots at the start of the experiment, but in low abundance with uneven distribution (Dichelachne micrantha, T. australis); three were not present (Capillipedium parviflorum, Poa labillardieri, Sorghum leiocladum). The application rate of the material ranged from 0.0076 g m−2 for Poa to 0.1–0.4 g m−2 for the remaining species.

Floristic monitoring

The vegetation in each plot was surveyed in December 2004 (prior to any treatments) and in November 2005 (11 months after treatments began) using a point-intercept technique (modified from Everson & Clarke 1987). Following Prober et al. (2005), an 8-mm dowel was placed vertically at each of 50 random points on a grid located on the central 4 × 4 m of each plot; the relative abundance (points) for any species was the number of points at which any of its leaves, stems or inflorescences intercepted the dowel. Species that were present but did not intercept the dowel at any point were allocated a nominal abundance of 0.5. Species nomenclature follows Harden (1990–1993).

Species recorded from the vegetation were grouped into exotic or native categories based on their origin and grasses were classified as C3 or C4 (Downton 1975; Raghavendra & Das 1978). Further grouping of species based on growth form (forbs, graminoids, shrubs) and life history (perennial/annual) followed descriptions in Harden (1990–1993).

Soil sampling and analyses

Soil was sampled in December 2004 before treatments began, then in March, June and November 2005. At each sampling event, soil was collected from five points along a diagonal transect in each plot using a 4-cm-diameter soil corer inserted to a depth of approx. 10 cm. The soil from each plot was bulked together, creating one sample per plot, and kept cool until receipt in the laboratory, within 24 h. Soil samples were then air-dried at 32°C for at least 24 h. When samples were thoroughly air-dry, they were passed through a 2-mm and then a 1-mm sieve. The less than 1-mm fractions were then stored in sealed plastic bags until chemical analyses could be performed. Two soil subsamples per plot were analysed for extractable nitrate, total soil organic nitrogen, total soil organic carbon and pH.

For each chemical analysis, a duplicate sample was used on every eighth sample to check the consistency of results. The average of the two replicates per plot was used for statistical analyses. A soil standard was also analysed for every chemical analysis to check calibration of methodology. All soil standard results were within the given limits for the sample, and close to the median of results.

Extractable nitrate was measured on a 3.0-g subsample of soil using the mineral nitrogen with 2 mol L−1 KCl automated colour method (ASPAC method code (7C2)) (Rayment & Higginson 1992). Soil was extracted with 30 mL 2.0 mol L−1 KCl. Nitrate concentrations were measured using a Lachat QC-8000 Flow Injection Analyser at UWS Hawkesbury, and then converted to microgram of NO3- kg−1 soil using original soil weights. Total soil organic nitrogen was measured using the total nitrogen semi-micro Kjeldahl steam distillation method (ASPAC method code (7A1)) (Rayment & Higginson 1992). Two grams of soil was digested with 12 mL 98% sulphuric acid and heated. Cooled samples were then distilled in a Tecator 1026 distillation unit and titrated with 0.1 mol L−1 HCl, then converted to percentage nitrogen using original soil weights. Total soil organic carbon was measured on a 0.2-g subsample of soil using the total organic carbon Heanes wet oxidation method (ASPAC method code (6B1)) (Rayment & Higginson 1992). Soil was oxidised with 10 mL 0.0167 mol L−1 potassium dichromate and 20 mL 18 mol L−1 sulphuric acid and heated. Deionised water was then added to cooled samples, samples were then centrifuged to clear the supernatant and the absorbance was measured at 600 nm using a Helios UV-Vis spectrophotometer and then converted to percentage carbon using original soil weights. The C/N ratio was determined using the results from the percentage carbon and nitrogen values. The pH of the samples was measured using the 1:5 soil/0.01 mol L−1 CaCl2 extract direct method (ASPAC method code (4B1)) (Rayment & Higginson 1992). Thirty millilitres of 0.01 mol L−1 CaCl2 was added to 6.00 g of soil and mechanically shaken. Samples were allowed to settle, then the electrodes of a Hach sensION4 pH/ISE meter were positioned in the unstirred supernatant and the pH value recorded when the meter reading appeared steady.

Statistical analyses

Data for soil variables, plant relative abundance and species richness were analysed by Repeated Measures anova. Treatments (carbon addition, fire, seed addition) were the between-plots fixed factors, and time the within-plots factor. Preliminary analyses detected no effect of seed addition for the analyses of soil variables and exotic plant relative abundance and species richness, and so data from this treatment were pooled with their corresponding carbon and fire treatments. Data were analysed using spss 19; the assumption of sphericity was examined by Mauchly's W (if there were more than two times), the equality of covariances by Box's test, and homogeneity of variances by Levene's test. Where data violated assumptions, data were transformed. Where Mauchly's W was still significant after transformation the probability of the F-tests after the Greenhouse-Geisser adjustment was reported. Back-transformed means are presented where necessary. Treatment effects if present should be detected as a time × treatment(s) interaction. For post hoc comparisons of means, if the assumption of sphericity was violated, the Bonferroni method was used as it controls α in this situation. If the assumptions of sphericity and equality of covariances were met, it is safe to use other comparison methods with more power and so the Student Newman Keuls test was used (Stevens 2002).

Analyses of plant relative abundance involved sharing of data across multiple analyses. Within each of the exotic and native species groups, analyses were conducted for total abundance, subgroups of total (native grasses, native and exotic forbs), and for individual species with sufficient data. Data were shared across two analyses for exotics and three for natives; adjusting alpha downwards to 0.025 for exotics and 0.0167 for natives to control experiment-wide error resulted in 4 out of 16 P-values significant at 0.05 being regarded as non-significant. As this exceeded the expected Type I error rate, an alpha value of 0.05 was retained for analyses of abundance.

Comparison of treatment effects on species composition was examined by the permanova routine in PRIMER (Anderson et al. 2008). Data were square root transformed; the dispersion of data within treatments was compared by the PERMDisp routine. The repeated measures design already described for univariate analysis was used; pair-wise comparisons between treatments within each time was carried out for significant terms in the analysis. A non-metric multi-dimensional scaling (MDS) of the mean species composition data per treatment was carried out after calculation of Bray-Curtis similarity of the transformed data (Clarke & Gorley 2006).


  1. Top of page
  2. Abstract
  9. Supporting Information

Soil responses to plot treatments

Soil nitrate on control plots fluctuated seasonally, with a higher value for the first summer (mean of 2.25 mg kg−1) followed by significantly lower values over autumn and winter (0.29–0.48 mg kg−1) returning to higher values the next spring (1.79 mg kg−1; Fig. 1; raw data range 0.3–3 mg kg−1). Carbon addition significantly altered this seasonal pattern (time × carbon significant, Table 1). Treating soil with carbon initially had little effect, with soil nitrate levels on the carbon-addition plots closely tracking those on the control plots in autumn and winter (Fig. 1). However carbon addition significantly reduced the concentration of soil nitrate compared with the control over the following spring, when concentrations remained at 0.54 mg kg−1 on carbon addition plots (Fig. 1). Fire had no detectable effect on soil nitrate levels (fire × time not significant (NS), Table 1).


Figure 1. Nitrate concentration in plots with no added carbon (no C) and with carbon added (C added) from pre-treatment (December 2004) through post-treatment (January to November 2005) at Doonside. Means (log back-transformed) are shown. Timing of fire treatment (F), seed application (+S) and carbon addition (+C) shown by arrows on x-axis. Result for comparison of treatment means at each time is shown above means; *P < 0.05; ns, not significant. Within each treatment, means followed by different letters are significantly different at P < 0.05 (upper case = added C, lower case = no added C).

Download figure to PowerPoint

Table 1. Repeated measures anova for soil variables at the Doonside site
Sourced.f. d.f.(GG)NO3TSOCTONC : N
  • Greenhouse-Geisser (GG) P-value. Degrees of freedom (d.f.) and P of F-ratios are shown (P < 0.05 bolded). Data for seed treatments pooled within carbon × fire treatments. P of Cox's test for equality of covariance matrices and Mauchly's test of sphericity is also shown for each analysis. Analyses are shown for NO3- (nitrate), TSOC (total soil organic carbon), TON (total organic nitrate) and C : N (carbon : nitrogen ratio). c., carbon; fi, fire.

Between plots
Carbon1  0.026 0.1170.7950.055
Fire1 0.4380.069 0.035 0.312
Carbon * fire1 0.5010.6050.4460.734
Within plots
Time (ti)32.107 0.000 0.030 0.000 0.520
ti × c.32.107 0.0016 0.5750.9240.126
ti × fi32.1070.7510.2920.1600.970
ti × c. × fi32.1070.6190.2140.2130.975
Cox's test  0.0780.3830.6060.864
Mauchly's W   0.030 0.003 0.2470.187

Total soil organic carbon varied significantly only with time (Table 1), being higher in the warmer months on control plots (3.29–3.51%), declining to lower values (3.12%) in autumn and winter (Table 2). The same pattern was apparent on carbon addition plots, with higher values particularly after carbon addition commenced; however, the increase was not significant (time × carbon NS, Table 1). Total organic nitrogen also only varied significantly with time (Table 1), with higher warm season values (0.19 –0.21%) on control plots falling to 0.17–0.18% in cooler months (Table 2). While total organic nitrogen concentrations were higher on burnt plots, the difference was not significant (time × fire NS, Table 1).

Table 2. Treatment means for total soil organic carbon (TSOC), total organic nitrogen (TON) and C : N ratio over time
No CC added
TSOCDecember 20043.513.560.13
 March 20053.123.420.12
 June 20053.123.300.09
 November 20053.293.480.12
TONDecember 20040.2070.2130.008
 March 20050.1780.2070.009
 June 20050.1810.1970.006
 November 20050.1940.2150.006
  No CC added 
  1. SEs are shown for means in each row.

C : NDecember 200416.916.80.49
 March 200516.218.20.56
 June 200516.617.50.39
 November 200516.216.90.40

The soil C : N ratio did not differ significantly with any treatment over time (range 16.2–18.0, Table 1). Soil pH fluctuated seasonally over the range 4.7–5.03 with higher values in winter, but there was no significant treatment effect on pH at any measurement date (data not shown).

Floristic responses to treatments

Plant abundance

At the pre-treatment survey, total plant relative abundance varied little across all plots; the proportion of total relative abundance contributed by exotic species ranged from 63% to 69% except on the fire plots, where it was 50–53% (Fig. 2a). The exotic grasses Briza subaristata (C3) and Axonopus affinis (C4) combined made up over half of total relative abundance, with other exotic C4 grasses and forbs present at low relative abundances (Appendix S1). For native species, native grasses as a group made the largest contribution to total relative abundance (24–29%, rising to 39% on fire plots) but this was spread across two C3 and seven C4 species (Appendix S1). Other native species were present at low relative abundances including one shrub, one subshrub, graminoids and perennial forbs (Appendix S1).


Figure 2. Mean total relative abundance by treatment for (a) pre-treatment (2004) and (b) post-treatment (2005) survey. Means (square root back-transformed ± SEs) shown for exotic and native mean relative abundance.

Download figure to PowerPoint

At the post-treatment survey, total plant relative abundance, and the proportion as exotics had changed little on the control plots, whether seeded or not (Figs 2b,3a). Total plant relative abundance decreased on all treated plots (Fig. 2b); large decreases in exotic plant relative abundance in response to carbon and fire were only partially offset by smaller increases in native relative abundance in response to carbon and seed treatments (Fig. 2b). The combined effect of these treatments meant that post-treatment, the proportion of total relative abundance as natives rose to 44–49% on carbon plots, 64–65% on fire plots and 55–57% on carbon + fire plots (Fig. 2b).


Figure 3. Mean relative abundance of (a) total exotics and (b) Briza subaristata at the pre-treatment (2004) and post-treatment (2005) surveys. Interaction means (square root back-transformed) for time × carbon × fire are shown. Between surveys, the result for comparison of treatment means is shown above each treatment; *P < 0.05; ns, not significant. Within each survey, treatment means followed by different letters are significantly different at P < 0.05 (upper case = 2004 survey; lower case = 2005 survey).

Download figure to PowerPoint

Carbon and fire combined interactively to reduce total exotic relative abundance (time × carbon × fire, and both first-order interactions significant, Table 3; Fig. 3a). Carbon addition alone significantly decreased total exotic plant relative abundance (Fig. 3a) and fire alone gave a further significant decrease; combining the treatments did not result in any further reduction (Fig. 3a). The decrease in total exotic relative abundance on plots treated with carbon and/or fire was due to the response of the two dominant species, B. subaristata and A. affinis. The response of Briza followed the same pattern as for total exotic abundance (time × carbon × fire, and both first-order interactions significant, Table 3); carbon and fire approximately halved relative abundance when applied singly, but combining the treatments gave no further decrease (Fig. 3b). Axonopus relative abundance was halved independently by either treatment (time × carbon, time × fire significant, Table 3; Fig. 4a). Setaria gracilis also showed a strong negative effect of carbon (time ×  carbon significant, Table 3) and a lesser negative effect of fire (time × fire significant, Table 3); relative abundances were low, and comparison of means in the time × fire interaction failed to detect differences (Table 3; Fig. 4b). Paspalum dilatatum and exotic forbs showed no significant response to any treatment (data and analyses not shown).

Table 3. Repeated measures anova for exotic species abundance (total, individual species) and species richness
Sourced.f.AbundanceSpecies richness
Total exotic Briza Axonopus Setaria
  1. P of F-ratios are shown (P < 0.05 in bold). Data for seed treatments pooled within carbon × fire treatments. P of Cox's test for equality of covariance matrices is also shown for each analysis.

Between plots       
Carbon (c.)10.5610.5590.691 0.014 0.042
Fire (fi)1 0.0001 0.006 0.002 0.8320.443
c. × fi1 0.017 0.047 0.6130.5450.272
Within plots
Time (ti)1 0.00000 0.00000 0.0000 0.803490.04557
ti × c.1 0.00002 0.00012 0.0012 0.00002 0.01784
ti × fi1 0.00002 0.00004 0.0215 0.02727 0.00141
ti × c. × fi1 0.01002 0.00480 0.33790.391320.73987
Cox's test 0.3120.8220.3900.753300.86562

Figure 4. Mean relative abundance of (a) Axonopus affinis and (b) Setaria gracilis at the pre-treatment (2004) and post-treatment (2005) surveys. Interaction means (square root back-transformed) for time × carbon and time × fire are shown. Between surveys, the result for comparison of treatment means is shown above each treatment; *P < 0.05; ns, not significant. Within each survey, treatment means followed by different letters are significantly different at P < 0.05 (upper case = 2004 survey; lower case = 2005 survey).

Download figure to PowerPoint

Total relative abundance of natives showed a small but significant positive response to carbon, and to seed addition (time × carbon, time × seed significant, Table 4; Fig. 5). No significant response to carbon addition was detected for native grasses or forbs, or another eight individual native species with sufficient data (analyses not shown).

Table 4. Repeated measures anova for native species abundance (total, individual species) and species richness
Sourced.f.Abundance Dichelachne Species richness
Total native Themeda
  1. P of F-ratios are shown (P < 0.05 in bold). P of Cox's test for equality of covariance matrices is also shown for each analysis.

Between plots
Carbon (c.)10.28000.8550.5210.426
Fire (fi)10.25540.8940.2930.309
Seed (se)10.56260.5690.4420.130
c. × fi10.27300.7390.3600.217
c. × se10.97900.7650.1760.426
fi × se10.78620.6300.1780.789
c. × fi × se10.72040.7860.8240.789
Within plots
Time (ti)1 0.010 0.007 0.000 0.000
ti × c.1 0.030 0.0770.8150.085
ti × fi10.249 0.018 0.015 0.028
ti × se1 0.014 0.2100.509 0.037
ti × c. × fi10.9780.7170.7300.615
ti × c. × se10.5190.8580.5760.615
ti × fi × se10.3430.8670.0750.183
ti × c. × fi × se10.7600.3770.2310.519
Cox's test 0.0980.4510.5800.077

Figure 5. Mean relative total abundance of natives (square root back-transformed) for (a) time × carbon and (b) time ×  seed interactions at the pre-treatment (2004) and post-treatment (2005) surveys. Between surveys, the result for comparison of treatment means is shown above each treatment; *P < 0.05; ns, not significant. Within each survey, treatment means followed by different letters are significantly different at P < 0.05 (upper case = 2004 survey; lower case = 2005 survey).

Download figure to PowerPoint

While total native relative abundance did not respond to fire (time × fire not significant, Table 4), two individual species did (time × fire significant for T. australis and D. micrantha, Table 4). Relative abundance of Themeda increased threefold, and that of Dichelachne sevenfold, from low initial abundances on fire plots (Fig. 6). Other native species did not respond to treatments, but increased in abundance over time (time significant for Aristida ramosa, A. vagans, native forbs and Wahlenbergia gracilis as a species, analyses not shown).


Figure 6. Mean relative abundance of (a) Themeda australis and (b) Dichelachne micrantha at the pre-treatment (2004) and post-treatment (2005) surveys; time × fire interaction means (square root back-transformed) are shown. Between surveys, the result for comparison of treatment means is shown above each treatment; *P < 0.05; ns, not significant. Within each survey, treatment means followed by different letters are significantly different at P < 0.05 (upper case =  2004 survey; lower case = 2005 survey).

Download figure to PowerPoint

Species richness

A total of 47 species (18 exotic, 29 native) were present over the two surveys. At the post-treatment survey, exotic species richness was reduced independently both by carbon addition and by fire (time × carbon, time × seed significant, Table 3). There was a decrease of 1.8 species per plot on carbon addition plots, and 2.2 species per plot on burnt plots (Fig. 7). Examination of individual exotic species gained or lost identified three forb species absent on carbon addition plots, and five forb species absent on fire plots after treatment. The combination of the negative effects of carbon and fire resulted in six exotic forb species and one exotic grass being absent on carbon + fire plots at the post-treatment survey (Appendix S3).


Figure 7. Exotic species richness: means for (a) time ×  carbon and (b) time × fire interactions are shown at the pre-treatment (2004) and post-treatment (2005) surveys. *P < 0.05; ns, not significant. Within each survey, treatment means followed by different letters are significantly different at P < 0.05 (upper case = 2004 survey; lower case = 2005 survey).

Download figure to PowerPoint

Native species richness was increased independently by both fire and seed addition (time × fire, time × seed addition significant, Table 4). There was an increase of 3.9 species per plot if burnt, and 3.8 species per plot if seeded (Fig. 8). Non-seeded native species that were detected as gains on burnt plots post-treatment included one grass (A. ramosa) and two forbs (Wahlenbergia spp., Appendix S3). None of the five seeded grass species established on control plots; all five established in low numbers on treated plots, and three were new species for the site. While carbon addition led to a net gain of some native species (3.6 species per plot), the effect was not significant (Table 4).


Figure 8. Native species richness: means for (a) time × fire and (b) time × seed interactions are shown at the pre-treatment (2004) and post-treatment (2005) surveys. *P < 0.05; ns, not significant. Within each survey, treatment means followed by different letters are significantly different at P < 0.05 (upper case = 2004 survey; lower case = 2005 survey).

Download figure to PowerPoint

Species composition

Fire, carbon and seed addition each interacted independently with time to affect total species composition (first-order interactions of each treatment with time significant, second- and third-order interactions with time non-significant, Table 5). Within each treatment, differences between untreated and treated plots, which were not significant at the pre-treatment survey, became significant at the post-treatment survey (pair-wise comparisons, Table 6). The P(perm) was the lowest for fire out of the pair-wise comparisons, indicating the strongest effect for this treatment, with P(perm) for the carbon and seed addition treatments being higher, but still significant (Table 6).

Table 5. Repeated measures permanova of species composition by time of survey and treatment; 999 permutations; significant P-values shown in bold
Sourced.f.SSMSPseudo-F P(perm)Unique perms
  1. PERMdisp test of dispersal F1,15 = 0.762, P(perm) = 0.941.

Between plots
Carbon (c.)11363.81363.81.00790.393999
Fire (fi)11622.71622.71.19920.306999
Seed (se)1870.17870.170.643070.689999
c. × fi11147.41147.40.847940.499999
c. × se1884.72884.720.653820.685999
fi × se11122.21122.20.829350.55999
c. × fi × se1597.65597.650.441670.896999
Plot (fi × ca × se)16216501353.218.1670.001999
Within plots
Time (ti)12735.22735.236.723 0.001 999
ti × c.1744.01744.019.989 0.001 999
ti × fi1812.68812.6810.911 0.001 998
ti × se1515.7515.76.9238 0.004 999
ti × c. × fi1112.16112.161.50590.306998
ti × c. × se1179.79179.792.41380.136998
ti × fi × se1204.75204.752.7490.104998
ti × c. × fi × se1116.5116.51.56420.301999
Residual error161191.774.482   
Table 6. permanova pair-wise comparison of treatments: values for t, P(perm), number of unique terms and similarity between treatments shown for each year of survey
t P(perm)Unique termsSimilarity (%) t P(perm)Unique termsSimilarity (%)
No carbon vs. carbon0.8850.55199865.11.522 0.032 99860.2
Unburnt vs. burnt0.6940.82099965.31.785 0.005 99859.4
No seed vs. seed65.81.471 0.028 99860.3

On an MDS plot of treatment means, seven of the eight treatments from the pre-treatment survey lay within a cluster of at least 80% similarity; the fire treatment was the least similar to other treatments (less than 75%, Fig. 9), and this treatment differed from other plots in exotic/native abundance as already noted (Fig. 2a). Application of single treatments decreased similarity, apparent as shifts in position of and greater scattering among post-treatment survey points on the MDS plot (Fig. 9). Combining pairs of treatments led to greater decreases in similarity, e.g. for carbon + fire, and carbon + seed treatments, pre- and post-treatment plots were only at least 75% similar (Fig. 9). Combination of all three treatments resulted in the greatest decrease in similarity, so that the carbon + fire + seed treatment was less than 75% similar to pre-survey plots (Fig. 9), conforming with the independent combination of treatments effects indicated by the permanova.


Figure 9. Non-metric multi-dimensional scaling plot of treatment × survey (trt × sur) means. C, control; S, control + seed; F, fire; FS, fire + seed; Ca, carbon; CaS, carbon + seed; CaF, carbon + fire; CaFS, carbon + fire + seed. 1 = pre-treatment survey; 2 = post-treatment survey. Cluster lines of equal similarity are shown for 75%, 80% and 85% similarities.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  9. Supporting Information

While our study was limited in duration, the three treatments demonstrated the presence of the postulated barriers to restoration at the former Cumberland Plain woodland site. Carbon addition reduced soil nitrate, and carbon addition and/or fire, combined with native seed addition, favoured native species over exotics by adversely affecting exotic abundance and species richness, and increasing native species abundance and richness.

Carbon addition

The summer concentrations of soil nitrate from the former Cumberland Plain woodland site at Doonside approached the upper limit for Themeda-dominated remnants of White Box woodland (less than 2.5 mg kg−1, Prober et al. 2009) but did not reach the values of up to 30 mg kg−1 (Prober et al. 2005) observed for White Box sites dominated by annual exotics, nor the highest value of 17.4 mg kg−1 recorded under former Cumberland Plain woodland sites by Fitzgerald (2009). The Cumberland Plain woodland site differed from the White Box woodland sites in two respects: the peak in nitrate concentration was earlier (December) than in the White Box soils (May), and annuals were a very minor component of the vegetation at Doonside. Mid-summer seasonal peaks in nitrate concentration were observed under a pasture of perennial C3 grass (Phalaris aquatica) on the southern tablelands of New South Wales by Simpson (1962). Phalaris aquatica survived the summer dry period by senescence of the bulk of its above-ground growth (McWilliam 1968), which Briza may do at the Cumberland Plain site. Because the time between carbon addition and soil sampling was longer for the cool season samples (10 weeks) than for the spring sample (6 weeks) in our study, it is possible that short-term depression of nitrate concentrations occurred on carbon addition plots in the cool season, and was missed by the sampling regime. More detailed sampling of the time-course of soil nitrate after carbon addition would be required to fully characterise the response to carbon addition.

Current grassland models link low mineralization of nitrogen in the soil with positive feedbacks from warm season perennial C4 grasses, which have high values for nitrogen use efficiency, allocation of biomass to root and C : N ratio of litter (Wedin & Tilman 1990; Tilman & Wedin 1991; Prober et al. 2009). Maintenance of low levels of soil nitrate by native C4 summer grasses has been observed in prairie systems (Wedin & Tilman 1990) and by Themeda in White Box woodland (Prober et al. 2005; Prober & Lunt 2009). The reduction in abundance of the exotic cool-season C3 grass Briza under carbon addition fits this model; however, the reduction in abundance of the warm-season C4 Axonopus does not. The negative effect on the warm-season C4 exotic grass is promising, as a number of other exotic C4 grasses are dominant weeds in Cumberland Plain woodland. Our results highlight the importance of quantifying the species-specific response of vegetation to carbon addition (Eschen et al. 2006) in order to determine the exact nature of the plant-soil feedbacks operating.

Effects of added carbon on species richness have been less commonly reported than effects on growth. Such an effect could be achieved if either the competitive advantage conferred on native species by low nitrate levels was sufficient for them to competitively eliminate exotic species, or the lowered nitrate levels were insufficient to sustain nitrophilic exotics. While the negative effects on exotic species richness observed in our study are promising, it is not known how permanent they would be. A short-term negative effect of carbon addition on native species richness in northern Californian grassland disappeared in subsequent years (Alpert & Maron 2000), as did a 1-year negative effect on species richness of non-native prairie species (Baer et al. 2004). However native species richness increased over 3 years, was highest and was inversely related to productivity under the reduced-N treatment in a study on prairie restoration in a former cultivated field (Baer et al. 2003, 2004). While the small positive effect of carbon addition on native species richness in our study was insufficient to be significant, the evidence from the prairie study suggests that reducing soil nitrate can have long-term benefits for native species richness.

If carbon addition is to be used in restoring Cumberland Plain woodland at larger scales, further research is needed on what forms of carbon work (Zink & Allen 1998; Kardol et al. 2008) and what dose to use. The dose used in our study (0.21 kg m−2) was effective in the White Box woodland studies (Prober et al. 2005; Prober & Lunt 2009). A range of carbon addition levels have been tested, up to 0.24 kg m−2 by Brunson et al. (2010) and up to 3.1 kg m−2 by Blumenthal et al. (2003). Weed biomass continued to decline over the whole range used in both studies, but a positive response by natives required 1 kg m−2 of carbon in the latter case. Once soil nitrate concentrations have been reduced by carbon addition, the best long-term method for maintaining the suppression is growth of native species that will do so; otherwise, the effect disappears once carbon addition ceases (Perry et al. 2010). Stands of other native grasses were not as effective as Themeda in this role (Prober et al. 2002b; Prober & Lunt 2009), leading Prober and Lunt (2009) to propose that Themeda was a keystone species in the White Box woodland. Whether Themeda can play this role in Cumberland Plain woodland remains to be tested.


Fire had a number of desirable effects on the native/exotic balance, for both abundance and species richness. Similar responses to burning were observed in the White Box woodland, with decreases in abundance of exotic annual grasses and increases in that of native grasses occurring (Prober et al. 2005; Smallbone et al. 2007).

While recruitment processes were not observed directly in our study, a large body of evidence points to the effects of fire on plant demographic processes, more than on soil nutrients, as driving the immediate responses of grasslands to fire. Recruitment in grasslands with dense canopies is very limited; flowering and seed set are poor, and recruitment of seedlings under the canopy does not occur. Many grassland species lack a persistent seedbank. This pattern has been observed in demographic studies in temperate Themeda grasslands in Victoria (Morgan 1999; Lunt & Morgan 2002), for warm and cool season grasslands on the north-west slopes of New South Wales (Lodge 1981; Lodge & Whalley 1981), and for tallgrass prairies in North America (Benson & Hartnett 2006; Dalgleish & Hartnett 2009). Many grassland species recover from vegetative structures after fire, resulting in little short-term change in species composition after a single fire (Lunt & Morgan 2002), a result observed for the ground layer species in a high-quality remnant of burnt Cumberland Plain woodland (Hill & French 2004). Dalgleish and Hartnett (2009) demonstrated that fire effects on below-ground vegetative bud demography determined the response to fire in tallgrass prairie. A model proposed for grasslands that lack a persistent native seed bank is that fire does not lead directly to germination and recruitment; rather, it facilitates recruitment in subsequent years by promoting flowering in the first year, and opening up the canopy to provide microsites for germination and establishment in the second year (Lunt & Morgan 2002; Bond & Parr 2010).

However studies in the temperate Themeda grasslands have shown that repeated burning over longer periods favours native species cover and richness (Lunt & Morgan 1999; Prober & Lunt 2009). Frequent burning in tallgrass prairie favours warm season native perennial grasses (Benson & Hartnett 2006). In addition to the beneficial effects of repeated frequent fires on native plant demography already noted, fire can have negative effects on exotic species. The effect of repeated spring burns on exotic annual grasses observed by Prober and Lunt (2009) was attributed to negative demographic effects on seed banks. Soil nitrate concentrations were unaffected by the burn treatment in our study, as observed in the second burn of Prober et al. (2005), and in Prober and Lunt (2009). Soil nitrate concentrations can be higher after an individual fire (Ojima et al. 1994) as mineralisation shows a short-term increase; over the longer term, frequent fires can reduce soil nitrogen (Johnson & Matchett 2001; Reich et al. 2001).

The effects of fire on exotics in our study were similar to those of carbon; both strongly reduced the abundance of the dominant exotic grasses, and they both reduced exotic species richness. The mechanisms differed however; carbon addition acted via plant-soil feedback, and fire via biomass removal and lethal effects of heat. Fire effects on natives differed somewhat from those of carbon, being observed mainly as an increase in native species richness, with increases in native abundance restricted to two species, present at low abundances only. The evidence from other Australian temperate grasslands suggests that repeated burning would be required to maintain this mix of negative effect on exotics, and positive effects on natives (Lunt & Morgan 2002) and that these effects would be mediated via demographic processes, rather than via plant-soil feedback.

Seed addition

While germination and establishment from added seed were limited in our study, effects on native species abundance and richness were detectable. Merely supplying native plant propagules was insufficient by itself; the absence of recruitment of seeded species on the control plots in our study shows the role of the intact grass canopy in restricting native recruitment in these grasslands (Lodge 1981; Lunt & Morgan 2002; Smallbone et al. 2007). Reduction of the canopy by the carbon treatment, or removal by burning, were additionally required to overcome this limitation to establishment.


Suding and Gross (2006) proposed that increasing the native-to-exotic species ratio should be a goal in restoration studies. The three treatments used increased this ratio, from less than 1.5 without treatment to values approaching or surpassing 2.0 with treatment. The treatments generally combined independently in their effect for the duration of this study, reflecting differing mechanisms of operation via plant-soil feedbacks, or demographic effects.

The results demonstrate that elevated soil nitrate levels, and propagule/recruitment limitation operate in cleared and grazed Cumberland Plain woodland and are likely to be barriers to restoration.


  1. Top of page
  2. Abstract
  9. Supporting Information

This work was supported by an Australian Research Council Linkage Grant with New South Wales National Parks and Wildlife Service and Greening Australia (Western Sydney) as Industry partners. M. de Barse was supported by an Australian Postgraduate Award in Industry. Mark Emanuel conducted the soil chemical analyses.


  1. Top of page
  2. Abstract
  9. Supporting Information
  • Alpert P. & Maron J. L. (2000) Carbon addition as a countermeasure against biological invasion by plants. Biol. Invasions 2, 3340.
  • Anderson M. J., Gorley R. N. & Clarke K. R. (2008) PERMANOVA+ for PRIMER: Guide to Software and Statistical Methods. PRIMER-E Ltd, Plymouth.
  • Averett J. M., Klips R. A. & Curtis P. S. (2002) Reverse fertilization reduces cover of weedy species in tallgrass prairie restoration (Ohio). Ecol. Restor. 20, 1278.
  • Baer S. G., Blair J. M., Collins S. L. & Knapp A. K. (2003) Soil resources regulate productivity and diversity in newly established tallgrass prairie. Ecology 84, 72435.
  • Baer S. G., Blair J. M., Collins S. L. & Knapp A. K. (2004) Plant community responses to resource availability and heterogeneity during restoration. Oecologia 139, 61729.
  • Bannerman S. M. & Hazelton P. A. (1990) Soil Landscapes of the Penrith 1:100 000 Sheet. Soil Conservation Service of NSW, Sydney.
  • Bartolome J. W., Jackson R. Z. D. & Allen-Diaz B. H. (2009) Developing data-driven descriptive models for Californian grasslands. In: New Models for Ecosystem Dynamics and Restoration (eds R. J. Hobbs & K. N. Suding) pp. 12435. Island Press, London.
  • Benson D. H. (1992) The natural vegetation of the Penrith 1:100 000 map sheet. Cunninghamia 2, 54196.
  • Benson E. J. & Hartnett D. C. (2006) The role of seed and vegetative reproduction in plant recruitment and demography in tallgrass prairie. Plant. Ecol. 187, 16377.
  • Bleier J. S. & Jackson R. D. (2007) Manipulating the quantity, quality and manner of C addition to reduce soil inorganic N and increase C3:C4 grass biomass. Restor. Ecol. 15, 68895.
  • Blumenthal D. M., Jordan N. R. & Russelle M. P. (2003) Soil and carbon addition controls weeds and facilitates prairie restoration. Ecol. Applic. 13, 60515.
  • Bond W. J. & Parr C. L. (2010) Beyond the forest edge: ecology, diversity and conservation of the grassy biomes. Biol. Conserv. 143, 2395404.
  • Brunson J. L., Pyke D. A. & Perakis S. S. (2010) Yield responses of ruderal plants to sucrose in invasive-dominated sagebrush steppe of the Northern Great Basin. Restor. Ecol. 18, 30412.
  • Cione N. J., Padgett P. E. & Allen E. B. (2002) Restoration of a native shrubland impacted by exotic grasses, frequent fire, and nitrogen deposition in southern California. Restor. Ecol. 10, 37684.
  • Clark P. J. & Davison E. A. (2001) Experiments on the mechanism of tree and shrub establishment in temperate grassy woodlands: seedling emergence. Austral Ecol. 26, 40012.
  • Clark P. J. & Davison E. A. (2004) Emergence and survival of herbaceous seedlings in temperate grassy woodlands: recruitment limitations and regeneration niche. Austral Ecol. 29, 32031.
  • Clarke K. R. & Gorley R. N. (2006) PRIMER V.6: User Manual/Tutorial. PRIMER-E Ltd, Plymouth.
  • Corbin J. D. & D'Antonio C. M. (2004) Can carbon addition increase competitiveness of native grasses? A case study from California. Restor. Ecol. 12, 3643.
  • Dalgleish H. J. & Hartnett D. C. (2009) The effects of fire frequency and grazing on tallgrass prairie productivity and plant composition are mediated through bud bank demography. Plant Ecol. 201, 41120.
  • Dickson T. L. & Foster B. L. (2008) The relative importance of the species pool, productivity and disturbance in regulating grassland plant species richness: a field experiment. J. Ecol. 96, 93746.
  • Doll J. E., Haubensak K. A., Bouressa E. L. & Jackson R. D. (2011) Testing disturbance, seeding time, and soil amendments for establishing native warm-season grasses in non-native cool-season pasture. Restor. Ecol. 19, 18.
  • Downton W. J. S. (1975) The occurrence of C4 photosynthesis among plants. Photosynthetica 9, 96105.
  • Eschen R., Müller-Schärer H. & Schaffner U. (2006) Soil carbon addition affects plant growth in a species-specific way. J. Appl. Ecol. 43, 3542.
  • Everson C. S. & Clarke G. P. Y. (1987) A comparison of six methods of botanical analysis in the montane grasslands of Natal. Vegetatio 73, 4751.
  • Fitzgerald J. K. (2009) The soil of abandoned farmland, Cumberland Plain woodland and restored vegetation: implications for the restoration of an endangered ecological community (PhD Thesis). University of Western Sydney, Sydney.
  • French K., Callaghan B. & Hill S. (2000) Classifying endangered vegetation communities: a case study of Cumberland Plain woodlands. Pac. Conserv. Biol. 6, 1209.
  • Harden G. J., ed. (19901993) Flora of NSW, Vols 1–4. University of NSW Press, Sydney.
  • Hill S. J. & French K. (2004) Potential impacts of fire and grazing in an endangered ecological community; plant composition and shrub and eucalypt regeneration in Cumberland Plain woodland. Aust. J. Bot. 52, 239.
  • Hill S. J., Tung P. J. & Leishmann M. R. (2005) Relationships between anthropogenic disturbance, soil properties and plant invasion in endangered Cumberland Plain Woodland, Australia. Austral Ecol. 30, 77588.
  • Johnson L. C. & Matchett J. R. (2001) Fire and grazing regulate belowground processes in tallgrass prairie. Ecology 82, 337789.
  • Kardol P., Van der Wal A., Bezemer T. M. et al. (2008) Restoration of species-rich grasslands on ex-arable land: seed addition outweighs soil fertility reduction. Biol. Conserv. 141, 220817.
  • Keith D. (2004) Ocean Shores to Desert Dunes: The Native Vegetation of New South Wales and the ACT. Department of Environment and Conservation (NSW), Hurstville.
  • Lindsay E. A. & Cunningham S. A. (2011) Native grass establishment in grassy woodlands with nutrient enriched soil and exotic grass invasion. Restor. Ecol. 19, 13140.
  • Lodge G. M. (1981) Establishment of warm- and cool-season native perennial grasses on the north-west slopes of New South Wales: II. Establishment and seedling survival in the field. Aust. J. Bot. 29, 12134.
  • Lodge G. M. & Whalley R. D. B. (1981) Establishment of warm- and cool-season native perennial grasses on the north-west slopes of New South Wales: I. Dormancy and germination. Aust. J. Bot. 29, 11119.
  • Lunt I. D. & Morgan J. W. (1999) Effect of fire frequency on plant composition at the Laverton North Grassland Reserve, Victoria. Vic. Nat. 116, 849.
  • Lunt I. D. & Morgan J. W. (2002) The role of fire regimes in temperate lowland grasslands of south-eastern Australia. In: Flammable Australia: The Fire Regimes and Biodiversity of a Continent (eds R. A. Bradstock, J. E. Williams & M. A. Gill) pp. 17798. Cambridge University Press, Port Melbourne.
  • McWilliam J. R. (1968) The nature of the perennial response in Mediterranean grasses. II. Senescence, summer dormancy and survival in Phalaris. Aust. J. Agric. Res. 19, 397409.
  • Morgan J. W. (1999) Defining grassland fire events and the response of perennial plants to annual fire in temperate grasslands of south-eastern Australia. Plant Ecol. 144, 12744.
  • Ojima D. S., Schimel D. S., Parton W. J. & Owensby C. E. (1994) Long- and short-term effects of fire on nitrogen cycling in tallgrass prairie. Biogeochemistry 24, 6784.
  • Paschke M. W., McLendon T. & Redente E. F. (2000) Nitrogen availability and old-field succession in a shortgrass steppe. Ecosystems 3, 14458.
  • Perry L. G., Blumenthal D. M., Monaco T. A., Paschke M. W. & Redente E. F. (2010) Immobilizing nitrogen to control plant invasion. Oecologia 163, 1324.
  • Perry L. G., Galatowitsch S. M. & Rosen C. J. (2004) Competitive control of invasive vegetation: a native sedge suppresses Phalaris arundinacea in carbon-enriched soil. J. Appl. Ecol. 41, 15162.
  • Prober S. M. & Lunt I. D. (2009) Restoration of Themeda australis swards suppresses soil nitrate and enhances ecological resistance to invasion by exotic annuals. Biol. Invasions 11, 17181.
  • Prober S. M., Lunt I. D. & Morgan J. W. (2009) Rapid internal plant-soil feedbacks lead to alternative stable states in temperate Australian grassy woodlands. In: New Models for Ecosystem Dynamics and Restoration (eds R. J. Hobbs & K. N. Suding) pp. 15688. Island Press, London.
  • Prober S. M., Lunt I. D. & Thiele K. R. (2002a) Determining reference conditions for management and restoration of temperate grassy woodlands: relationships among trees, topsoils and understorey flora in little-grazed remnants. Aust. J. Bot. 50, 68797.
  • Prober S. M., Thiele K. R. & Lunt I. D. (2002b) Identifying ecological barriers to restoration in temperate grassy woodlands: soil changes associated with different degradation states. Aust. J. Bot. 50, 699712.
  • Prober S. M., Thiele K. R., Lunt I. D. & Koen T. B. (2005) Restoring ecological function in temperate grassy woodlands: manipulating soil nutrients, exotic annuals and native perennial grasses through carbon supplements and spring burns. J. Appl. Ecol. 42, 107385.
  • Raghavendra A. S. & Das V. S. R. (1978) The occurrence of C4 photosynthesis: a supplementary list of C4 plants reported in during late 1974–mid 1977. Photosynthetica 12, 2008.
  • Rayment G. E. & Higginson F. R. (1992) Australian Laboratory Handbook of Soil and Water Chemical Methods. Inkata Press, Melbourne.
  • Reever Morghan K. J. & Seastedt T. R. (1999) Effects of soil nitrogen reduction on non-native plants in restored grasslands. Restor. Ecol. 7, 515.
  • Reich P. B., Peterson D. W., Wedin D. A. & Wrage K. (2001) Fire and vegetation effects on productivity and nitrogen cycling across a forest-grassland continuum. Ecology 82, 170319.
  • Sheley R. L., Mangold J. M. & Anderson J. L. (2006) Potential for successional theory to guide restoration of invasive-dominated rangeland. Ecol. Monogr. 76, 36579.
  • Simpson J. R. (1962) Mineral nitrogen fluctuations in soils under improved pasture in southern New South Wales. Aust. J. Agric. Res. 13, 105972.
  • Smallbone L. T., Prober S. M. & Lunt I. D. (2007) Restoration treatments enhance early establishment of native forbs in a degraded temperate grassy woodland. Aust. J. Bot. 55, 81830.
  • Spiegelberger T., Müller-Schärer H., Matthies D. & Schnaffer U. (2008) Sawdust addition reduces the productivity of nitrogen-enriched mountain grasses. Restor. Ecol. 17, 86872.
  • Standish R. J., Cramer V. A. & Yates C. J. (2009) A revised state-and-transition model for the restoration of woodlands in Western Australia. In: New Models for Ecosystem Dynamics and Restoration (eds R. J. Hobbs & K. N. Suding) pp. 16888. Island Press, London.
  • Stevens J. (2002) Applied Multivariate Statistics for the Social Sciences. Lawrence Erblaum Associates, London.
  • Stevens M. H., Bunker D. E., Schnitzer S. A. & Carson W. P. (2004) Establishment limitation reduces species richness as soil resources rise. J. Ecol. 92, 33947.
  • Suding K. N. & Gross K. L. (2006) Modifying native and exotic species richness correlations: the influence of fire and seed addition. Ecol. Applic. 16, 131926.
  • Suding K. N. & Hobbs R. J. (2009) Models of ecosystem dynamics as frameworks for restoration ecology. In: New Models for Ecosystem Dynamics and Restoration (eds R. J. Hobbs & K. N. Suding) pp. 321. Island Press, London.
  • Thomsen M. A., Corbin J. D. & D'Antonio C. M. (2006) The effect of soil nitrogen on competition between native and exotic perennial grasses from northern coastal California. Plant Ecol. 186, 2335.
  • Tilman D. & Wedin D. A. (1991) Plant traits and resource allocation for five grasses growing on a nitrogen gradient. Ecology 72, 685700.
  • Tozer M. (2003) The native vegetation of the Cumberland Plain, western Sydney: systematic classification and field identification of communities. Cunninghamia 8, 175.
  • Wedin D. A. & Tilman D. (1990) Nitrogen cycling, plant competition, and the stability of tallgrass prairie. In: Proceedings of the Twelfth North American Prairie Conference – Recapturing a Vanishing Heritage (eds D. D. Smith & C. A. Jacobs) pp. 58. University of Northern Iowa, Cedar Falls.
  • Wilson S. D. & Gerry A. K. (1995) Strategies for mixed-grass prairie restoration: herbicide, tilling and nitrogen manipulation. Restor. Ecol. 3, 2908.
  • Zink T. A. & Allen M. F. (1998) The effects of organic amendments on the restoration of disturbed coastal sage scrub habitat. Restor. Ecol. 6, 528.

Supporting Information

  1. Top of page
  2. Abstract
  9. Supporting Information

Appendix S1. Exotic species: treatment means of relative abundance raw data for surveys in 2004 (pre-treatment) and 2005 (post-treatment) at Doonside, New South Wales.

Appendix S2. Native species: treatment means of relative abundance raw data for surveys in 2004 (pre-treatment) and 2005 (post-treatment).

Appendix S3. Exotic and native species that were recorded only at the (a) second survey (gains); (b) first survey (losses).

aec2426_sm_Appendxi_A-C.docx35KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.