Brisbane Rainforest Action & Information Network

The vegetation and flora of Brisbane Forest Park - Part 5

Bruce Noble
Field Survey
Sampling Strategy
Data Collection
Bitterlich's Variable Radius Method
Coming Soon...
1a. Hellhole Scrub - Brisbane
1b. Darlington Range - Upper Ormeau
Combined Sites TA02 and TA02a
1c. Bahrs Scrub - Beenleigh
Combined Sites TA05 and TA05A
Table 2. 'Nearest Neighbour' survey of Hellhole Scrub - Brisbane Forest Park
Site 18.4 - Co-ords 27026'40", 1520 52' 25"

Bruce NobleBack to top

This issue firstly sees Peter Young talking about the method used to survey the vegetation of BFP

Field SurveyBack to top

Two important questions require resolution prior to beginning field work. These are:

Type of sampling strategy? As a corollary; what is an adequate sample of the vegetation given a certain survey area and the usual constraints of time, access, disturbed areas?
What data should be collected at each site?

To date, many vegetation surveys of larger areas (e.g. geographic regions) in Australia have relied upon ad hoc choice of sites, with data collection at each site limited to compilation of a species list. This method, referred to by Austin and Basinski (1 978) as 'purposive' is sufficient to provide a description of vegetation, as the field worker usually selects sites that he considers both 'typical' and `unique'. Beyond providing a description this technique fails: as Kellman (1980) notes, neither a species list nor a (structural) description permits any exploration of the functioning of vegetation, a central objective of any analysis. Fulfilment of this objective requires the collection of qualitative (i.e. presence or absence of species) or quantitative data at sites that constitute a representative sample of the survey area.

Sampling StrategyBack to top

The large proportion of relatively undisturbed or intact vegetation in Brisbane Forest Park and adequate time allotted to field survey (at least 12 weeks) made it possible to systematically sample vegetation using a grid. Size of grid squares was determined by the target number of sites (see below). Within each grid square, sites were placed along north-south traverses, so orientated as to capture variation in aspect. Number of sites within each grid square varied, though ideally the aim was two traverses each with two to three sites. In addition, extra sites were sometimes located along traverses where warranted by local variation or 'uniqueness'. Although subjective, this was viewed as substantially contributing to knowledge and more thorough collection of the flora. The ultimate location of traverses was equally subjective, as road access, lantana thickets, recent disturbance etc. all had to be taken into consideration.

Data CollectionBack to top

Multiple nearest neighbour (Williams et al. 1969 a.) was used for sampling vegetation. The method is fast and free from the practical difficulties and debate associated with use of quadrants for sampling trees (see Greig - Smith 1964, for example). Williams et al. (1973), sampling twenty-five contiguous trees (one tree and its twenty-four nearest neighbours) found the method successful and recommended the technique for sampling stands of rainforest. Following discussion with M. Dale and D. Ross of C.S.I.R.O. Division of Computing Research it was decided to increase the sample to thirty-three contiguous trees. This is considered more satisfactory for the relatively species-poor Eucalyptus forests.
Vegetation data recorded at each site included:

Canopy cover
Height of canopy
Cover and structural description of lower strata or understorey.

Sampling the flora included:

Identification of a tree and its thirty-two nearest neighbours
Distance of each tree from the 'central' tree. This was measured using a 'Rangefinder' optical device
Girth breast height (g.b.h.) of each tree
Species list of understorey plants, epiphytes, lianas etc.

Species that could not be identified in the field were collected and subsequently identified at the Queensland Herbarium. Identification of rainforest trees often required sampling of canopy foliage using a slingshot or wherever possible a shotgun. However Eucalyptus, unlike rainforest trees cannot be identified on foliage alone and infertile trees were sometimes troublesome. Problems arose with:

Grey gums E. propinqua and E. punctata var. didyma (now E. biturbinata) growing together.
Ironbarks E. crebra and E. decepta (now E. siderophloia) growing together.
Small sterile stringybarks with opposite leaves persisting beyond the juvenile stage that were observed on Mt. Coot-tha and at Ironbark Gully. A large stringybark in the vicinity of Northbrook Mountain.

Where no capsules could be found under individual grey gums in mixed stands, reference to bark characteristics was used for species identification. E. punctata var. didyma (E. biturbinata) appeared to have a whiter trunk than E. propinqua during the survey period. The problem with the ironbarks was confined to smaller subcanopy trees at harsher sites. Caution was required with the opposite-leaved stringybarks, as smudgee (Angophora woodsiana) is known from some northern Brisbane suburbs. However fruit of yellow stringybark (E. acmenoides) were consistently collected on the ground beneath the trees, and foliage material matched sterile herbarium material of coppice and 'sucker leaves' collected on Mt. Coot-tha (White 1917). The stringybark on Northbrook Mountain is referred to here as E. eugenioides. Material collected from this locality matched herbarium material and the trees were shown to Mr. W.J.F. McDonald of the Queensland Herbarium, for a second opinion. He concurred with the tentative identification. To avoid confusion it is also necessary to point out that naming of Queensland Grey lronbark E. drepanophylla follows Blakely and this species is called E. decepta (now E. siderophloia) here.
Most field survey was undertaken in the winter months. Consequently some annual grasses, forbs, filmy ferns etc. have not been collected and recorded.
Environmental data recorded at each site included:

Geology (this often entailed collection of samples)
Evidence of recent burning
Evidence of burning that is not recent
Presence of rock on ground surface (if present, an estimate of cover)

In addition hand-augered soil profiles to depth one metre were made at twenty sites. Methods and soil descriptions followed Northcote (1979). Identification of geological specimens was kindly undertaken by Dr. A. Ewart of the Department of Geology and Mineralogy, University of Queensland.
A pilot survey indicated that four to six sites could be completed per day in Eucalyptus open forest and woodland and two to three in rainforest. This resulted in a target of 140 sites overall and was used to determine grid square sizes. In all, 124 sites were completed."
It would be opportune to discuss and compare another method of survey used on a recent field trip with Bill McDonald of the Queensland Herbarium. The site surveyed was chosen within a patch of Dry rainforest known as 'the Hellhole Scrub'. The name belies the magnificence of this relatively large intact piece of rainforest spanning both sides of Enoggera creek, upstream of the Enoggera reservoir. The following extract from 'Aims and Methods of Vegetation Ecology' (Mueller Dombois & Ellenberg, published by John Wiley & sons, New York, 1984) explains the technique used on the day

Bitterlich's Variable Radius MethodBack to top

BITTERLICH (1948) discovered a remarkably efficient way to measure stem cover in tree stands by applying the point-frequency principle. Since stem cover is the same as tree basal area, and since basal area is one of the basic units for tree volume determination, the method is of great value to forestry inventory work (GROSENBAUGH 1952). But also, since stem cover by species is defined as their dominance, and since cover or dominance is one of the most important quantitative parameters in vegetation ecology, the BITTERLICH method has become an important quantitative method, particularly in North American vegetation ecology.

The technique; Trees are counted in a circle from a central sampling point with an angle-gauge. Only trees that are larger in diameter than a specified angle are included in the count. The others are ignored. Therefore, the circular plot around the central sampling point has no fixed radius; instead, the radius varies with the diameter of each tree counted. This also renders the method plotless, because no fixed area-sample is involved.
When trees are counted in this manner with an angle-gauge, their number is proportional to their stem or basal area per unit ground area.
The standard North American angle-gauge is usually made of a 33-inch long stick. Mounted at one end is a 1-inch wide cardboard, plastic, or metal crosspiece and at the other end a similar piece with a notch or peephole.
The angle-gauge is held with the peephole or notch at the eye like an Abney level and pointed with the 1-inch wide crosspiece horizontally at each tree surrounding the sampling point. The point aimed at on each tree must be at a fixed height, usually breast height. The same ratio or angle of 145'can be obtained with a 33 cm long gauge and a 1 cm wide crosspiece. When using a round stick, a peephole or notch is not even necessary.
Only those trees are counted whose diameter exceeds the crosspiece. Therefore, small diameter trees are included in the count only if they are close to the observer, while large diameter trees are included at greater distances away from the observer. With the 1:33 gauge, trees counted will not be further away from the sampling point or observer than 33 times their diameter. Thus, a tree with a 4-inch (1 0 cm) diameter must be within 4 times 33 = 132 inches (3.35 m) of the sampling point, while a 20-inch tree will be included if it is within 20 times 33 = 660 inches (16.8m). The selection of the 1:33 ratio for the construction of the gauge, or the equivalent sighting angle of 145', was recommended by GROSENBAUGH (1952), because the tree count at this angle permits immediate calculation of the basal area in square feet per acre. This is done by multiplying the count by 10."

As we live in the metric age we need to calculate the basal area in square metres per hectare. The author explains:

"BITTERLICH (1948) recommended a gauge ratio of 1.41 cm to 100 cm giving a much narrower sighting angle (50') and more than twice the tree count per sampling point. At this ratio, the tree count divided by 2 results in the basal area in square metres per hectare. A still simpler ratio for calculation is 2 cm to 100 cm or 50:1, which is equivalent to a sighting angle of 110'. The resulting count is directly equal to the basal area in square metres per hectare.

This last ratio was in fact the one used on the day. The gauge used was 50 cm long with a l cm wide crosspiece. As I sighted the trees, Bill identified them and included them in the count. The whole process of sighting the trees took about 40 minutes.

Coming Soon...Back to top

The survey method at work. Bill (right) identifies
the trees while I use the gauge (centre). Note
Kenneth (thermos in hand) and Klaus
discussing where we should have smoko!
To finish the survey, Bill noted the surrounding shrubs, trees, epiphytes, and ground layer species not included in the tree count, as well as other relevant vegetation and environmental data. The end result was a simplified version of the method explained above. Although basal area was not calculated (as additional data needs to be included to calculate basal area), the relative dominance of species could be determined. The results of the relative dominance are given included in the following set of tables.

Bill McDonald using binoculars to identify the tree species.
(In all the following tables, the abbreviations used are: E1- Emergent tree; T1- Canopy tree; T2 - Understorey tree; S1 - Shrub. Rel. Dom is short for relative dominance. The botanical names of some species have been changed since the survey was done. Ed.)
Table 1. Comparison of basal areas of rainforest derived by Bitterlich Method
W. McDonald, Queensland Herbarium, Feb. 1999

1a. Hellhole Scrub - BrisbaneBack to top

Site Species E1 T1 T2 S1 Total Rel. Dom Notes
BF46 Araucaria cunninghamii 3 1     4 6.5%  
BF46 Austromyrtus bidwillii     3 2 5 8.1%  
BF46 Beilschmiedia obtusifolia       3 4.8%  
BF46 Brachychiton discolor   1     1 1.6%  
BF46 Diospyros pentamera   2     2 3.2%  
BF46 Dissilliaria baloghioides   9 9   18 29.0%  
BF46 Drypetes deplanchii   3 1   4 6..5%  
BF46 Flindersia bennettiana   3 1   4 6.5%  
BF46 Flindersia schottiana 2 2     4 6.5%  
BF46 Hodgkinsonia ovatiflora     1   1 1.6%  
BF46 Jagera pseudorhus     1 1 2 3.2%  
BF46 Medicosma cunninghamii     3 1 4 6.5%  
BF46 Podocarpus elatus 1 1 3   5 8.1%  
BF46 Premna now(vitex) lignum-vitae    2     2 3.2% 1 with strangler fig
BF46 Pseudoweinmannia lach. 1       1 1.6% With Ficus watkin.
BF46 Stenocarpus sinuatus   1     1 1.6%  
BF46 Syzygium francisii   1     1 1.6%  
BF46 TOTAL 7 29 22 4 62    

1b. Darlington Range - Upper OrmeauBack to top

Site Species E1 T1 T2 S1 Total
TA02 Acacia bakeri   2 1   3
TA02 Alchornea ilicifolia       1 1
TA02 Aphananthe philippinensis   1     1
TA02 Arytera distylis       1 1
TA02 Capparis arborea     1   1
TA02 Cassine australis     1   1
TA02 Croton achronychioides     2   2
TA02 Dendrocnide excelsa   5 1   6
TA02 Dendrocnide photinophylla   3 2   5
TA02 Dissilliaria baloghioides   8 8 2 18
TA02 Excoecaria dallachyana   1 2   3
TA02 Pouteria australis     1   1
TA02 Premna lignum-vitae   3 1   4
TA02 TOTAL   23 20 4 47

Site Species E1 T1 T2 S1 Total
TA02a Acacia bakeri   7 2   9
TA02a Arytera divaricata     1   1
TA02a Capparis arborea     1 1 2
TA02a Dendrocnide photinophylla   1     1
TA02a Dissilliaria baloghioides   7 9 2 18
TA02a Drypetes deplanchii     1   1
TA02a Pouteria australis     1   1
TA02a Pouteria eerwah   1     1
TA02a Premna lignum-vitae   5 1   6
TA02a TOTAL   21 16 3 40

Combined Sites TA02 and TA02a Back to top

Species No. Rel. Dom
Acacia bakeri 12 13.8%
Alchornea ilicifolia 1 1.1%
Aphananthe philippinensis 1 1.1%
Arytera distylis 1 1.1%
Arytera divaricata 1 1.1%
Capparis arborea 3 3.4%
Cassine australis 1 1.1%
Croton achronychioides 2 2.3%
Dendrocnide excelsa 6 6.9%
Dendrocnide photinophylla 6 6.9%
Dissilliaria baloghioides 36 41.4%
Drypetes deplanchii 1 1.1%
Excoecaria dallachyana 3 3.4%
Pouteria australis 2 2.3%
Pouteria eerwah 1 1.1%
Premna lignum-vitae 10 11.5%
TOTAL 87  

1c. Bahrs Scrub - BeenleighBack to top

Site Species E1 T1 T2 S1 Total
TA05 Acacia bakeri 4     1 5
TA05 Arytera divaricata       1 1
TA05 Dissilliaria baloghioides   18 11 3 32
TA05 Flindersia schottiana   1     1
TA05 Flindersia xanthoxyla 1   1   2
TA05 Pouteria eerwah 1 1     2
TA05 Premna lignum-vitae   1     1
TA05 Malaisia scandens (liane)   1     1
TA05 TOTAL 6 22 12 5 45

Site Species E1 T1 T2 S1 Total
TA05A Argyrodendron trifoliolatum   1     1
TA05A Arytera distylis   1     1
TA05A Capparis arborea       1 1
TA05A Dendrocnide photinophylla   1     1
TA05A Dissilliaria baloghioides   18 6 1 25
TA05A Ficus superba (on Premna) 1       1
TA05A Macadamia integrifolia   1     1
TA05A Premna lignum-vitae 5 2     7
TA05A Malaisia scandens (liane)   1     1
TA05A TOTAL 6 25 6 2 39

Combined Sites TA05 and TA05ABack to top

Species No. Rel. Dom
Acacia bakeri 5 6.0%
Argyrodendron trifoliolatum 1 1.2%
Arytera distylis 1 1.2%
Arytera divaricata 1 1.2%
Capparis arborea 1 1.2%
Dendrocnide photinophylla 1 1.2%
Dissilliaria baloghioides 57 67.9%
Ficus superba (on Premna) 1 1.2%
Flindersia schottiana 1 1.2%
Flindersia xanthoxyla 2 2.4%
Macadamia integrifolia 1 1.2%
Pouteria eerwah 2 2.4%
Premna lignum-vitae 8 9.5%
Malaisia scandens (liane) 2 2.4%
TOTAL 84  

Comparing the surveys, one can note greater species diversity between the Hellhole and the other 2 scrubs. Even the combined species count from the 2 separate surveys taken both at the Darlington Range Scrub (16) and Bahrs Scrub (14) falls short of the single survey taken in the Hellhole (17). This contrast can be explained by a difference in the environmental features of each site. The geology at all three sites is derived from metasediments, but the Hellhole site (situated on a flood plain of Enoggera creek) has been enriched by alluvial deposits. Bahrs scrub and the Darlington Range receive less rainfall and being situated on ridges are more exposed than the Hellhole. Other notable points include a lack of emergent species at Darlington Range and the dominance of one species in the canopy and understorey at all 3 sites, that species being Dissilliaria baloghioides (lancewood).

A much broader comparison could be made using the additional vegetation and environmental data taken with each survey (not included with this article).
The following is a transcript of the original field notes taken by Peter Young using the 'nearest neighbour' method at a site in the Hellhole. Although no direct comparison should be made as the survey was taken across the creek from that of the BITTERLICH survey, it is interesting to note the species mix and dominance is still relatively similar. Also included is the vegetation and environmental data from the site.

Table 2. 'Nearest Neighbour' survey of Hellhole Scrub - Brisbane Forest Park Back to top

Peter Young, November, 1982.

Site 18.4 - Co-ords 27026'40", 1520 52' 25" Back to top

Canopy Trees
Beilschmiedia obtusifolia   6
Austromyrtus acmenoides 2
Grevillea robusta  1
Dissiliaria baloghioides  12
Argyrodendron trifoliolatum  2
Podocarpus elatus  2
Dendrocnide photinophylla  1
Jagera pseudorhus  1
Hodgkinsonia ovatiflora 1
Ficus fraseri 1
Diospyros pentamera  1
Bouchardatia neurococca  1
Pseudoweinmannia lachnocarpa  1
Beilschmiedia elliptica  1

Leaf Size: Notophyll & some microphyll
Structure:  Complex
 Geology: Metamorphics locally enriched by alluvium & andesite
 Aspect:   0
 Topography:  Valley floor
 Rainfall: 1250mm
Altitude:  125m
Slope: 10

Shrub/low tree: Amorphospermum antilogum, Ailanthus triphysa, Stenocarpus sinuatus, Mischocarpus pyriformus, Wilkiea macrophylla, Cryptocarya laevigata var. bowiei, Cleistanthus cunninghamii, Actephila moorei.

Ground stratum/vines: Coelospermum paniculatum, Doodia aspera, Arthropteris tenella, Lastreopsis marginans, Pellea paradoxa, Rauwenhoffia(now Melodorum) leichhardtii, Tetrastigma nitens, Cissus antarctica, Melodonis australis, Platycerium bifurcatum, Platycerium superbum, Dendrobium speciosum,Pyrrosia confluens.

Vicinity trees (outside plot): Choricarpia subargentea, Rhodomyrtus psidioides, Rhodamnia argentea.

Peter Young notes in his discussion that:
"Similarly, distribution records of species such as Dissiliaria baloghioides, Atalaya multiflora, Medicosma cunninghamii, Choricarpia leptopetala and Amorphospermum antilogum provide insight into the affinities of the Enoggera creek rainforests. D. baloghioides has an extremely disjunct distribution. To the north it is recorded from Imbil and Kin Kin in the Gympie district, Baffle creek north of Bundaberg and the Sarina and Proserpine districts. The species, though common where it grows, is restricted to several stream catchments in central and southern Queensland. Atalaya multiflora was also recorded from the Samford Valley in 1925. It is extremely uncommon in southern Queensland, known from only the vicinity of Brisbane Forest Park, Beechmont and Imbil. Its chief area of occurrence in Queensland, is in the complex notophyll-microphyll closed forests of the Whitsunday coastline in northern-central Queensland. It has also been recorded from a few sites in northern New South Wales.

Medicosma cunninghamii was recorded only from two alluvial sites along Enoggera Creek. Such sites are characterised by marara Pseudoweinmannia lachnocarpa, Capparis sp. nov. (hairy form) and brown pearwood Amorphospermum antilogum, species typical of coastal/subcoastal lowland rainforests e.g. Mary Valley. Furthermore, brown pearwood has not been recorded on the other side of the watershed in Gold Creek/Brookfield areas, although this species and Choricarpia leptopetala, a species that is rare in Queensland, are shared by Enoggera Creek and Cabbage Tree Creek (Lake Manchester) catchments.

Rainforests of the mid Mary Valley (Imbil area) and lower Brisbane River and its tributaries therefore, share many species and constitute a distinctive regional 'type' flora. While most of these areas are now cleared or extensively modified, some fine stands remain along Enoggera Creek and these represent a truly magnificent resource: they are probably little different from the old 'Three Mile Scrub' now part of the city at Newmarket, a popular collection site for Queensland's famous early botanists."

The information provided by these types of surveys gives an insight into the structure of the vegetation type concerned. They greatly enhance the value of a species list for an area of forest and should be used as a tool by those interested in forest ecology and regeneration. I wish to acknowledge Bill McDonald (Queensland Herbarium) and Peter Young (Dept. of Environment & Heritage) for their help in preparing this article.