Botanic Gardens Trust, Sydney, Australia

Monitoring change in the woodland

Findings from the research plots 1988-2005

The quest for knowledge begins with a blank page. Our woodland remnant was like the blank white page of a book waiting to be written. There was virtually no ecological research on western Sydney’s bushland and very little on the ecology of its individual species. We wanted to know what species made up our woodland, how they interacted and how the processes worked. We wanted to help the native species to flourish and make the weeds disappear. We set as our aim to understand the composition of the woodland and the patterns of change as the native species responded to the removal of stock grazing and mowing in 1988, and began to flourish and expand in what we hoped were more natural conditions. In the first instance we hoped that the native species would expand and overwhelm the exotics, perhaps with some assistance from us.

Our research aim at the Australian Botanic Garden, Mount Annan has been to understand the ecology of the woodland as it recovers from nearly two centuries of rural use, and apply this knowledge to conservation management. Our main performance indicator has been the ongoing survival and expansion of all the native plant species present in the woodland, while reducing the abundance of exotic species. Treatments used have been weed eradication and burning. No plantings of any species have been made.

The changing woodland 1988-2005

Ecology of Cumberland Plain Woodland

Ecologists at work
Doug Benson and Jocelyn Howell recording in the woodland.

Our long-term study methods

'Long-term data sets … have a tendency to surprise you.' David Tilman speaking about the results of monitoring British grassland experiments at Rothamsted since 1843.

To observe changes over time we set up permanently marked plots that we could periodically record. Initially we set up eight permanent square plots, each 5 m x 5 m, marked with steel posts and coloured tape. The Woodland Conservation Area included an area with large trees which appeared to have likely to have a good number of native species amongst the weeds, as well as areas with smaller trees and more open areas that had evidently been cultivated or possibly ploughed and pasture improved. When we first visited it had been grazed by stock for many years and had been recently slashed.

Historical accounts give the impression that Themeda australis, Kangaroo Grass was a major groundcover component of the woodland at the time of European settlement, but that this was readily eaten out by stock. Patches of Themeda australis were relatively common in our woodland and were relatively weed free. We regarded these as the type of groundcover we hoped would expand into the more weedy areas in the absence of grazing and mowing. We placed two sites in grassland mainly of Themeda (sites 3 & 4), and two sites in Themeda australis-Bursaria spinosa grassy-shrubby areas (sites 1 & 6). To sample the weedy areas we placed two sites in *Paspalum dilatatum-*Sida rhombifolia dominated areas (sites 2 & 5). Because of our particular interest in the endangered species Pimelea spicata we located one site to include part of its population, (site 7), and another, (site 8), amongst Chloris ventricosa-*Sida rhombifolia-Bursaria spinosa, with a similar shrubby structure but with no Pimelea spicata.

Initial field recording was done in December 1988 as we thought that most plants would be flowering late spring-early summer and most readily identified. We recorded the species present within each of the 25 1 m x1 m subplots. This gave us a frequency out of 25 for each plot, and a total frequency out of 200 subplots, a measure of abundance that we could remeasure, to observe increases or decreases. Our 8 plots set up in 1988 included occurrences of about 50 % of the plant species (69 out of the 134) recorded in the woodland at that time.

Some difficulties often encountered in monitoring Cumberland Plain  Woodland

  • Dry conditions seasonally and even daily
  • Tiny species make location and identification difficult, particularly if they are also short lived e.g. Phyllanthus virgatus
  • Similar closely-related species make identification difficult e.g. Austrodanthonia species

Rabbit exclosure plots

By 1991 rabbits were becoming common in the woodland and despite the assurances of the Gardens management that a control program would eliminate them within 6 months, evidence of their activities became increasingly obvious. We had not considered the impact of rabbits during the initial setting up of the plots as there was no evidence of their presence, and the open conditions of the grazed landscape and presence of dogs presumably kept them in check. However the taller growing groundcover and the fencing of the Gardens obviously provided them with suitable conditions to flourish. In 1991 we set up three fenced exclosure plots, Plot 9 paired with Plot 6, Plot 10 paired with Plot 8, and Plot 11 in open grassy area at the top of the woodland, paired with a new unfenced plot 12. We particularly wanted to see whether natural recruitment of native seedlings was retarded by rabbit grazing. These plots were recorded at the same time as Plots 1-8.

Subsequent recording of all the plots was done on an ad hoc basis as time allowed. The plots were recorded in May 1991, November 1991, December 1992, December 1993 and December 1996.

After the 1996 recording we began to question our expectations of the change in the woodland. There was not the great expansion of native species we had expected. Instead change was very slow. What was beginning to happen however was that some of the perennial weed species were beginning to expand. This was obvious in observations in the woodland, and was confirmed in the plots. In particular the weedy shrub *Olea europea subsp. cuspidata, African Olive, was clearly becoming more common in the woodland, particularly under large Eucalyptus trees. The time spent on the sample plots seemed disproportionate to the questions being raised, which related to rare species, that were mostly not represented in the plots. Other work commitments reduced our available time for the project, allowing only occasional visits to the woodland and not allowing for the week needed every year to record all 12 plots plus the time to consider the data. No recording was made from 1997 to 2000 though we visited the woodland sporadically during this time.

Following changes in work programs at the end of 2000, we were able to allocate more time to the woodland allowing us to record our sites in January 2001, December 2001, December 2002, December 2003, December 2004 and December 2005.

Monthly recording

In 2000, after reviewing the woodland plot data from the previous 10 years, during which relatively little seemed to have happened, we became concerned that our narrow window onto the woodland, our December recording, meant that we were missing important changes in the vegetation that occurred in other months. We had initially chosen December for annual recording to maximise our level of identification, based on our wider experience on sandstone areas where major flowering is in spring, and modified by knowledge that grasses are more readily identified by their fruiting in summer. We realized that many of the plant species were opportunists and that major ecological events such as seedling recruitment, flowering and fruiting, were being influenced by rainfall and temperature events which occurred irregularly over the year.

We therefore selected two of our plots, Plots 4 and 9, for regular monthly recording. Plot 4 was an unfenced grassy plot which was left undisturbed without any treatment. Plot 9 was a fenced Themeda-Bursaria plot set up in 1991. Neither plot has been burnt in the 20 years to 2006. These two plots contained a good variety of woodland species, 67 natives, (50% of our total of 134) and 34 exotics.

Our more frequent recording subsequently has shown that overall there is more growth and flowering after rains in late summer and autumn than there is in spring, although a handful of species grow more in spring.

Seasonality recording

To explore the issue of seasonality further we have been recording the flowering and fruiting times or seasons of the woodland plants. This is known as phenology. Our initial view that flowering and fruiting timing were likely to be similar to the patterns observed on the familiar Hawkesbury sandstone areas was changing, and we wanted to observe changes in our woodland more closely.

Fire trials

We carried out a trial ecological burn in part of the woodland in 1991.

We carried out more experimental burns in parts of the woodland in September 2001 when we again burned the area previously burnt in 1991. The adjacent area to the west was burned in September 2002, the area to the north, including adjacent grassland, in September 2005.

^ back to top

Fenced plot with quadrant
Fencing to exclude rabbits to measure seedling recruitment after fire

Jocelyn Howell recording seedling recruitment in the woodland

The changing woodland 1988-2005

Results - changes in the tree stratum 1988-2005

We originally chose the Conservation Research Woodland for a study site because of its large trees. In some parts the trees are up to 25 m high with diameters of up to 1 m.These are some of the largest trees we have seen in Cumberland Plain Woodland in western Sydney and are likely to be at least 100-200 years old. In some parts of the woodland however the trees are much smaller, evidently where the original trees were cleared during the farming period and young trees have grown since then.

Eucalypt seedling recruitment is generally confined to areas around the canopy of mature seed trees, or within 20 to 30 m of tree canopies This pattern is commonly observed in woodland at Mount Annan. This is because eucalypt seed is small and has no special dispersal mechanism apart from being shed 10-20 m above the ground to be blown a few metres away and may be washed a further few metres across a bare soil surface in rainwater. A lot of seedlings came up following the cessation of grazing in the 1980s and are now trees up to 8 m high.

Interestingly there has been very little subsequent recruitment in the woodland since the 1980s. A few seedling have been recorded in the sample plots but none of these have survived. The only obvious recent successful recruitment has been near the road along a line dug for a water pipe where backfilled soil provided bare soil for the seedling recruitment, trees overhead provided seed, and there was enough rainfall.

Most of the sapling recruitment has been of Grey Box Eucalyptus moluccana and Forest Red Gum Eucalyptus tereticornis, with a small amount of Narrow-leaved Ironbark. However during the 1990s many adult ironbarks, Eucalyptus crebra, gradually died despite relatively favourable rainfall condition. We consider that these deaths were a delayed long-term result of the 1980s drought, and the impacts of the increased nutrients and compaction of the soil by stock grazing.

Flowering and fruiting by the woodland eucalypts has very sporadic. At no time could we say that all mature trees of a particular species were in flower, though we would note occasional individual trees flowering. Nor could we note any clear seasonal trends in those individuals that did flower. Fruiting occurs a considerable time after flowering, but again we have no noted any distinctly recurring patterns.

Halo effects around trees

In some areas we noted that the understorey under large trees was different to that nearby. There were more Einadia plants, and weeds such as *Sida rhombifolia. We put this down to these being the site of old stock camps, resting places for animals during hot weather, where nutrients had been locally increased, conditions the saltbushes would like. However these conditions seem more widespread than we had thought, and a listing of species within about 5 to 10 m of large trees includes the following - 4 species of Einadia, 2 Solanums, *Olea europea subsp. cuspidata and *Sida rhombifolia, grasses rare. These species all have fleshy fruits and are likely to be spread by birds, therefore being concentrated below canopies of large and particularly remnant trees. These effects are most conspicuous under old well-established trees, and may take at least 20 years to become noticeable. First established under isolated trees are the *Olea eurapaea subsp cuspidata, followed by Einadia species It is likely that *Olea europaea subsp cuspidata will crowd out other species by its dense canopy and the growth of its own seedlings.

^ back to top 


Sapling Eucalypts germinated in bare soil exposed by trench for pipeline.

Bark from Eucalyptus crebra
Dead Ironbark. Bark on ground provides habitat for woodland invertebrates.


Results - the shrub stratum appears

Shrub species that occur occasionally in our Mount Annan woodland are Acacia implexa, Dodonaea viscosa, Myoporum montanum and Daviesia ulicifolia. However, the most common shrub species is Blackthorn, Bursaria spinosa.

Prior to 1988 patches of Bursaria in the woodland were generally associated with tree bases or steeper slopes which were not slashed. Our 1988 woodland photos showed an open grassy ground cover with previously slashed Bursaria rootstocks beginning to regrow. These Bursaria spinosa shrubs grew very vigorously to form dense thickets within about 5-7 years. Lateral spread, apart from longer spreading lateral branches as the plants increase in size, however after the initial regrowth was slow, with seedlings mainly recruiting under parent plants. As a result, in 2006, many of the areas between Bursaria spinosa clumps are still relatively open.

Bursaria spinosa can resprout from its base after cutting and fire. Where rootstocks have persisted in the soil after grazing or slashing vigorous recover takes place resulting in plants up to 2-3 m high in say 5-7 years. Total rooted frequency in our 8 measured plots (200 1 m x 1 m subplots) increased from 25% in 1988 to a peak of 47% 12 years later, though it subsequently declined to 38% in the following 3 (dry) years. This measurement includes both mature 2-3 m high plants, as well as smaller plants and seedlings 5-30 cm high.

Our interpretation is that growth of Bursaria  spinosa rapidly recovers where it has previously been suppressed, but that invasion into new areas is a slow process. We have previously thought that lateral spread was by rootsuckers, but extensive digging around established plants has not verified this as a possibility. Spread does appear to occur mainly adjacent to established plants, though occasional plants appear at distances of 10-20 m from the main edge by seedling recruitment. How successfully Bursaria recruits from seed, is not clear. Viable seed is produced each year, and there is no dormancy or soil-stored seedbank. We have only observed occasional seedlings, i.e. that can be identified from presence of cotyledons, but these have all subsequently disappeared. Nursery-grown plants and small plants dug up show a tapering carrot-like rootstock develops early in the establishment phase. There may also be some suppression of young plants by established adults.

While Bursaria spinosa generally resprouts vigorously after fire, regrowth of some plants burnt in the 2002 and 2005 burns has been very slow, presumably because of the low rainfall over this period.

Bursaria spinosa is a very common in many Cumberland Plain Woodland remnants and there is the perception that it will crowd out other species. On the other hand its dense prickly canopy may protect associated species from stock grazing and damage from trampling, and with its relatively open canopy allows light for other species to grow. As a major component of Cumberland Plain Woodland and because of its importance in remnant sites, more research on the ecology of Bursaria spinosa is needed.

The paired pictures below show some of the major changes in woodland structure at the Australian Botanic Garden, Mount Annan since 1988.

^ back to top

Tree and shrub layers
Flowering Bursaria spinosa

Woodland view Dec1988 
December 1988              
Woodland view Mar 2006
  March 2006
Woodland view 2 Dec 1988
December 1988

Woodland view 2 Mar 2006
March 2006



Results - groundcover species changes in the woodland

In 1988 our initial expectation for the Woodland Conservation Area was that, with the cessation of grazing and its management for conservation, that the native plant species would increase in abundance. Certainly tree and shrub cover increased noticeably over a 10 year period, but this only involved a few tree species and the shrub species, Bursaria spinosa which was mainly from previously suppressed individuals - areas that were previously devoid of shrubs are still relatively shrub free nearly 20 years later.

Changes in the relative abundance of native groundcover species (and these make up over 80 % of the plant species) have been much less obvious. Native species that were regarded as Frequent species in 1988 were still Frequent species in 2000 (an exception is the grass Bothriochloa macra which appears to change in response to changing structure), while a few less common species have increased in abundance e.g. the fern Cheilanthes sieberi.

Undisturbed woodland plots 1988-2004

Five of our 5 x 5 m monitoring plots (four plots from September 2002) were left essentially undisturbed i.e. over the period 16 year period from December 1988 to November 2004. Undisturbed plots were those kept unslashed, unburnt, ungrazed by domestic stock, and unfenced- allowing access by rabbits from 1991, and kangaroos from 2000. These are essentially control plots and represent the basic management treatment for the conservation area.

The figure below shows the mean species richness measures for these plots. Native species richness was high initially with 32 species per plot, but fluctuated over the 14 year period, generally dropping no more than 25% of the initial measure, and returning to similar high levels after 15 and 16 years. The exception was a major slump to 16 species (a 50% drop on the initial figure) in November 2002 during an exceptionally very hot dry period. Native species richness the following year recovered to 32 species per plot.

Exotic species richness declined over the 16 year period from 16 species to 7 species, and dropped to 3 species in the very dry 2002.

Gaps in recording years however could mask fluctuations but there does appear to be a trend in decrease of exotic species over the period. Many of these species would remain in the soil seed bank or as rootstocks as is indicated by the immediate response in 2003 after the dry conditions of 2002. 

Monitoring plot 4

The first five years - will the weeds go?

With the grazing or slashing regime lifted, we expected the native groundcover plants to increase in abundance and begin to dominate. While some native species were widespread and common in 1988, many native species had very localised populations often of only a few plants occupying an area of a few square metres. Species such as Pimelea spicata, Daviesia ulicifolia, Sorghum leiocladum, Vernonia cinerea, Calotis lappaceus, Ranunculus lappaceus and Ranunculus sessiliflorus were all very restricted and we expected to see an increase in their abundance and distribution.

During the five years from 1988 however the main changes in species abundance were among the exotic species and many of these weedy species, after an initial increase, decreased or virtually disappeared. These included *Petrorhagia nanteillei, *Vulpia bromoides *Silene gallica, *Lepidium species [they reappeared after drought], *Polycarpon tetraphyllum, , *Soliva sessilis, *Stachys arvensis, *Linum trigynum, *Centaurium tenuiflorum and *Verbascum virgatum. These species were generally short-lived species, with many of the exotics being weedy annual species and are part of a group of species that we called Short-term (<5 years) responders. Data from our plots showed that species of this group decreased in frequency during the first 5 years (between 1988 and 1993), and continued to decrease further in the following 10 years.

Many are common weeds in rural sites. Presumably these species benefited from the open conditions created by slashing and grazing, and particularly in the years immediately following but were unable to establish or persist as longer-lived herb species became more abundant, and the perennial grasses grew taller.

Short-term (< 5 years) responders also include some short-lived ephemeral native species e.g. Daucus glochidiatus, Fimbristylis dichotoma and Ranunculus sessiliflorus, and short-lived perennials e.g. Geranium homeanum, Oxalis perennans , as would be expected, but also perennial species with rootstocks e.g. Sida corrugata.

Most of these native species have long-lived soil seedbanks and reappear periodically during wet seasonal conditions, particularly where groundcover has been depleted by drought or disturbance.

The next ten years 1993-2002

During the first 5 years some species increased in frequency, but subsequently, between 1993 and 2002, decreased in frequency. These species appeared to respond best to conditions 5-10 years after grazing/slashing ceased. Medium term (5-10years) responders are herbs and monocots, including a particularly high component of exotic herbs. They also include annuals e.g. Wahlenbergia gracilis, *Sonchus oleraceus, *Conyza sumatrensis, and short-lived species e.g. Senecio quadridentatus, *Senecio madagascariensis, *Anagallis arvensis and perennial species with rootstocks.e.g. Asperula conferta, Pimelea spicata These species appear to be taller growing species that are susceptible to regular slashing or grazing but respond to open conditions. We have called these Medium-term (5-10 years) responders.

A third group of species either increased in frequency or at least remained relatively stable over the 14 year time span. These Long-term (> 10 years) responders are mostly native species, herbs and monocots but also includes the tree and shrub component of the woodland. Perennial species with rootstocks are a major component e.g. Dichondra repens, Brunoniella australis, Aristida ramosa, Cheilanthes sieberi, but there are also some relatively shortlived species that establish in the absence of disturbance, e.g. Solanum prinophyllum, Oplismenus imbecillis, Crassula sieberiana. 



Species persistence - changes in species present over 14 years in 8 permanent plots (total area 0.2 ha)
( ) = inconclusive due to ID problems or seasonal effects
Disappeared species not necessarily gone from woodland


Native species

Exotic species

Species disappearing within first 5 years



Species disappearing within 8 years



Species still present after 14 years

54 = 78%

19 = 48%

New species appearing during 14 yars

4 (4)







Species persistence – changes in species present over 12 years (1988-2000) in 8 permanent plots (total area 0.2 ha)
( ) = inconclusive


Native species

Exotic species

Species that disappeared within first 5 years


Daucus glochidiatus
Fimbristylis dichotoma
(Polygala japonica
Solenogyne bellioides


*Leontodon taraxicoides
*Petrorhagia nanteuilii
*Plantago myosuros
*Polycarpon tetraphyllum
*Silene gallica
*Soliva sessilis
*Stachys arvensis
*Verbascum virgatum
*Vulpia bromoides

Species disappearing within 8 years


Juncus subsecundus
Wahlenbergia communis
(Solenogyne dominii)


*Centaurium tenuiflorum
*Ciclospermum leptophyllum
*Medicago lupulina
*Paronychia brasiliana
*Romulea rosea
*Taraxicum officinale

Species still present after 12 years

54 = 78%

19 = 48%

New species appearing during 12 years

4 (4)

Eremophila debile
Oplismenus imbecillus
Rumex brownii
Cotula australis
(Lachnagrostis filiformis
Carex breviculmis
Dodonaea viscosa
Stackhousia viminea)


*Briza subaristata
*Hypericum perforatum
*Modiola caroliniana
*Solanum nigrum




Asterisk * indicates exotic species naturalised at the Australian Botanic Garden, Mount Annan