On: 17 December 2011, At: 22:51Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Plant Biosystems - An International Journal Dealingwith all Aspects of Plant BiologyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/tplb20

Pre-logging carbon accounts in old-growth forests, viaallometry: An example of mixed-forest in Tasmania,AustraliaC. Dean a b , N. B. Fitzgerald c & G. W. Wardell-Johnson ba School of Biological, Earth and Environmental Sciences, University of NSW, Sydney, NSW,2052, Australiab Curtin Institute for Biodiversity and Climate, School of Science, Curtin University, GPO BoxU1987, Perth, WA, 6845, Australiac The Wilderness Society, GPO Box 716, Hobart, TAS, 7001, Australia

Available online: 10 Nov 2011

To cite this article: C. Dean, N. B. Fitzgerald & G. W. Wardell-Johnson (2011): Pre-logging carbon accounts in old-growthforests, via allometry: An example of mixed-forest in Tasmania, Australia, Plant Biosystems - An International Journal Dealingwith all Aspects of Plant Biology, DOI:10.1080/11263504.2011.638332

To link to this article: http://dx.doi.org/10.1080/11263504.2011.638332

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.

This article was downloaded

OLD GROWTH FORESTS

Pre-logging carbon accounts in old-growth forests, via allometry:An example of mixed-forest in Tasmania, Australia

C. DEAN1,2, N. B. FITZGERALD3, & G. W. WARDELL-JOHNSON2

1School of Biological, Earth and Environmental Sciences, University of NSW, Sydney, NSW 2052, Australia; 2Curtin

Institute for Biodiversity and Climate, School of Science, Curtin University, GPO Box U1987, Perth, WA 6845, Australia;3The Wilderness Society, GPO Box 716, Hobart, TAS 7001, Australia

AbstractUncertainty in past and future anthropogenic carbon emissions obscures climate change modelling. Available allometrics areinsufficient for regional-level accounting of old-growth, pre-logging carbon stocks. The project goal was to determine theaboveground carbon (biomass and necromass) for a typical old-growth Eucalyptus delegatensis-dominated mixed-forest inTasmania. Allometrics were developed for aboveground biomass of Eucalyptus delegatensis and generic rainforest understoreyspecies. A total of 207 eucalypts with DBH 0.21–4.5 m, and 897 rainforest understorey trees with DBH 0.01–2.52 m weremeasured across 7.7 ha. DBH frequency distribution of E. delegatensis showed at least two age cohorts and distinct positiveskew, whereas its DBH carbon distribution showed distinct negative skew. Half of the eucalypt biomass was from trees withDBH4 2.4(0.1) m, and 16% with DBH� 3.5 m (from *1.1 trees ha71) – indicating the importance of allometrics for highDBH. Aboveground carbon was 622(180) Mg ha71, with *20% from understorey and *25% from necromass. The carbonin aboveground biomass was above the median value for temperate forests. The long-term aboveground-carbon emissionsfrom clearfelling the same forest type from 1999 to 2009 is likely to be 2.9(+1.3) Tg, depending on the growth and seralstages of the forest logged.

Keywords: Allometric, carbon distribution, carbon emissions, logging, mixed-forest, old-growth

Introduction

Accounting of forestry’s industrial carbon emissions

is important not only to determine the carbon fluxes

accompanying present and future industrial activity

but also in climate change modelling. Uncertainty in

anthropogenic emissions contributes to *28C un-

certainty in the temperature effect of climate change

(Meir et al. 2006). Also, positive feedback from

climate change will increasingly detract from the

carbon stores of forests in some localities (Houghton

1997; Kashian et al. 2006) but the magnitude of

positive feedback is obscured by high uncertainty in

past and future emissions from land-use, which

includes forestry (Gloor et al. 2010).

Australian old-growth forests have amongst the

highest carbon stocks globally, particularly the tall

(434 m) wet-eucalypt forests of the temperate and

maritime-temperate south-east (e.g. Dean & Rox-

burgh 2006; Keith et al. 2009; Wood et al. 2010).

The wet-eucalypt forests are comprised of wet-

sclerophyll and mixed-forests (Hickey 1994).

Mixed-forests, in south-east Australia, are a compo-

site of an upper canopy of tall-eucalypt with a

[closed] cool-temperate rainforest understorey below

(Gilbert 1959). In the absence of stand-replacing

fire, wet-sclerophyll can lead to mixed-forests and

subsequently to rainforest (Gilbert 1959; Ashton &

Attiwill 1994). Mixed-forests represent the spatial

and/or temporal overlap, between two forest types

with different regeneration patterns. The interacting

effects of fire and logging influence the landscape

mosaic of these forests (Lindenmayer et al. 2009),

with spatio-temporal implications for determining

carbon dynamics. Determining stand-level carbon

stocks in forests of varying seral stages is a crucial

first step to investigating landscape-level, long-term

carbon dynamics.

Modelling of clearfell logging in old-growth, wet-

eucalypt forests of Tasmanian (Australia), has shown

Correspondence: C. Dean, PO Box 3520, Rundle Mall, Adelaide, SA 5000, Australia. Tel: þ61 8 8268 8230. Fax: þ61 2 9385 1558.

Email: cdeanspace@gmail.com, christopher.dean@unsw.edu.au, christopher.dean@curtin.edu.au

Plant Biosystems, 2011; 1–14, iFirst article

ISSN 1126-3504 print/ISSN 1724-5575 online ª 2011 Societa Botanica Italiana

http://dx.doi.org/10.1080/11263504.2011.638332

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

a long-term efflux of approximately half the original

carbon stock (Dean & Wardell-Johnson 2010). This

occurred when old-growth logging was followed by

an intense, eucalypt-regeneration burn and conver-

sion was to either plantation or to native eucalypt

forest by aerial sowing with eucalypt seed (i.e.

clearfell, burn and sow – CBS silviculture; Ellis

et al. 1982). The old-growth, wet-eucalypt forests,

have long-been a prime source of pulpwood and high

quality sawlogs (Felmingham et al. 2004; Elliot et al.

2008). Conversion of these old-growth forests

proceeded from the c. 1940s (Helms 1945), although

logging in the remaining old-growth is to be phased

out in State forests, by 2030 (Forestry Tasmania

2009). From the 1960s, production increased and

pulpwood export dominated (Florence 1993), multi-

plying at least fourfold from 1970–1980 to *2.2 Tg

year71 (dry weight) and remaining near that level to

2000 (Tasmanian Government 2006).

The magnitude of the long-term emissions (i.e.

2000 years, for stabilisation of wood-product carbon

and soil carbon levels) is logically dependent on the

original long-term mean of the pre-logged forest (and

the logging-cycle forest carbon plus wood-products).

Short-term emissions (i.e. over the first harvest

rotation) are more dependent on stocks immediately

prior to logging. In both cases carbon accounting of

commercial forestry activities relies on knowledge of

the pre-logging carbon stock. Developments in forest

inventory to-date in Australia have concentrated

mostly on post-harvest stands up to merchantable

age. Consequently, allometrics for advanced-mature

and old-growth trees are rare. Allometrics suitable

for carbon accounting were developed for all growth

stages of wet-Eucalyptus regnans F. Muell. (swamp

gum or mountain ash) forests with eucalypt DBH up

to 6.44 m. (Dean et al. 2003; Dean & Roxburgh

2006) (thus enabling both long-term and short-term

emission forecasts). This old-growth forest type was

a primary target in the Florentine and Styx valleys

(Tasmania), although the closely related wet E.

delegatensis R. Baker (white-top/gum-top stringy

bark, alpine ash)- and E. obliqua L’Herit. (brown-

top stringy bark, messmate) mixed-forests were also

logged (Australian Newsprint Mills 1960s coupe

inventories, unpublished, Forestry Tasmania; Walk-

er & Felton 2007). Logging of old-growth wet-

eucalypt forest for pulpwood and sawlogs is ongoing

on State and private land across Tasmania.

Here we aim to provide the first account of

aboveground (biomass and necromass) carbon for a

typical old-growth E. delegatensis mixed-forest stand,

prior to logging, based on field inventory. This

entails construction of applicable allometrics, devel-

oped from state-of-the-art allometrics and the

currently, publically available literature and software.

As error margins are important in interpreting

carbon accounts, we also estimate the errors in

aboveground carbon. Representativeness of likely

emissions is also presented, from a review of logging

records. Although soil carbon can notably decrease

after several logging cycles (e.g. Ferre et al. 2005;

Johnson et al. 2010) it was not examined as part of

this study due to resource constraints.

Materials and methods

Study site

The study site was located in the Florentine valley of

southern Tasmania. The geology, climate and forest

ecology of the Florentine valley are described in

Gilbert (1959). The valley is in the southern-central

distribution of wet-E. delegatensis forests in Tasma-

nia, with a high-rainfall region of rainforests to the

west, and agricultural areas and dryer forests to the

east (Figure 1). Although the wet-eucalypt forests in

the mid-elevation Florentine valley are dominated by

E. regnans (Gilbert 1959) the upper Florentine wet-

eucalypt forests are dominated by E. delegatensis, E.

obliqua and E. regnans (approximately 1500, 800 and

350 ha respectively). The study site was centred in

the upper Florentine Valley at 42 440 36.1000 S, 146

240 41.0600 E, within the planned logging area

(coupe) FO044A, of *60 ha. Mean annual pre-

cipitation at the study site was 1600 mm year71, and

elevation across the site was 448 m in the north to

505 m in the south. Ground slope within data

collection plots, averaged 5(2)8. The study site was

on Ordovician limestone, overlain in parts with

Holocene sediments.

The density of eucalypt trees in the old-growth E.

delegatensis mixed-forest at the study site varied from

under 5% to over 40% crown cover (Figure 2). At

the landscape-scale the mixed-forest was inter-

spersed with rainforest and the more-frequently

burnt, wet-sclerophyll forest (Figure 3). The rain-

forest understorey was dominated by Nothofagus

cunninghamii f. Hook. (myrtle beech, or myrtle),

Atherosperma moschatum Labill. (sassafras) and in

some locations Phyllocladus aspleniifolius Labill.

(celery-top pine) (Figure 4) – typical of mixed-forests

in Tasmania. There was a low abundance and cover

of other typical understorey species: [Acacia dealbata

Link. (silver wattle), Nematoleposis squamea Labill.

ssp. squamea (lancewood or satinwood) and Acacia

melanoxylon R. Br. (blackwood)]. Mean dominant

height of the mature eucalypts ranged from approxi-

mately 41–76 m (from Forestry Tasmania’s photo

interpretation), and myrtles reached up to *40 m.

The northern part of the coupe (wet-sclerophyll) had

been subject to moderate or low-intensity wildfire

in the past 100 years. The forest in the remainder

of the coupe (with myrtles present) appeared to be

2 C. Dean et al.

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

post-wildfire regeneration following a fire several 100

years ago and had not since been subject to fire. The

site had not been logged prior to 2009. Data were

collected from mid-2009 to 2010 and recent logging

had terminated when fieldwork began.

Field data collection and processing

The centres of 10 circular plots for data collection

were positioned randomly within the unlogged

portion of the operational boundary of the coupe.

Each plot comprised three nested sub-plots of 7, 30

and 60 m radius. The diameter at breast height

(1.3 m) (DBH) was the primary data type recorded.

Other data recorded were species, growth stage

(regrowth, mature or senescent), trunk hollow size,

burl size, evidence of fire or other feature likely to

impact on carbon accounting. DBH was corrected

for burls and adventitious roots. Trees with DBH

ranges from 0.01 to 50.2 m, 0.2 to 51.0 m and

41.0 m were recorded in the 7, 30 and 60 m sub-

plots respectively. Dimensions of necromass (snags,

and fallen logs and branches (coarse woody debris,

Figure 1. Distribution of E. delegatensis- wet-sclerophyll and mixed-forest from TASVEG, overlaid on mean annual precipitation. The colour

scheme for E. delegatensis was from the FullCAM layer of aboveground, potential biomass (Richards & Brack 2004).

Figure 2. Aerial photograph looking approximately north and

showing the upper Florentine Valley in the foreground; showing

both senescent and intact E. delegatensis over rainforest under-

storey in old-growth forest. Pale-green, conical-shaped trees are

sassafras. Photo corrected for perspective distortion. Original

photo: R. Blakers, 2010.

Old-growth pre-logging C, via allometry 3

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

CWD) with at least one diameter40.2 m were

recorded in the 30 m plots. Seven DBH-tree-height

pairs were measured using a clinometer and laser

rangefinder, for comparison with reported sizes

elsewhere.

Due to resource constraints four of the plots could

only be measured out to 30 m, and in one plot the

60 m circle partially overlaid a clearfelled area.

Spatial density data were corrected for these detrac-

tions during processing. Plot area was also corrected

for the average ground slope across each plot. The

total area sampled (after corrections) was 7.656 ha.

We used allometric equations to estimate the

biomass for each tree based on the measured DBH.

This value was then discounted to account for decay

in mature and senescent trees. Biomass of dead

wood was estimated by calculating the volume of the

log and discounting for decay.

For necromass, heights, lengths and diameters

were recorded. The size of hollows were also

measured, or estimated where out of reach. A decay

class was recorded, being one of H, M and S (hard,

medium and soft) corresponding to 100%, 66.66%

and 33.33% undecayed mass present for the

corresponding volume. That decay class was also

used in the field to represent voids, that were either

too numerous or irregular to be measured as hollows.

Allometrics

The two main approaches for deriving tree biomass

(or carbon) allometrics are destructive sampling

(including weighing and carbon assay) and taper

formulas. Our approach was to combine information

from published reports that had used both of these

methods. A commonly used approximation where

carbon assay is not performed (e.g. as part of

destructive sampling) is to assume that C is 50

wt% of dry vegetative matter (biomass or necro-

mass). For wet-sclerophyll E. obliqua in Tasmania,

Ximenes et al. (2008) found the fraction in stem

cross sections to be 0.497(0.004), being close to

0.5 – which was used throughout in the present work.

Allometrics for mainland populations of E. delega-

tensis ssp. delegatensis are not differentiated here from

those for the Tasmanian endemic subspecies

E. delegatensis ssp. tasmaniensis Boland, due to

similarity in growth habit. Previously (e.g. in the

1960s logging records) the Tasmanian variety was

known as E. gigantea f. Hook.

Figure 3. Landscape-level mosaic in the upper Florentine Valley:

cool temperate rainforest in riparian zones, juxtaposed with mixed-

forest – where the rainforest persists beneath emergent E.

delegatensis. Photo: N. Fitzgerald, 2010.

Figure 4. Mixed-forest in the southern part of the study area,

viewed from within recent clearing at the edge of the aggregate

retention operation; showing spatially dispersed, old-growth E.

delegatensis (both senescent and intact) over a dense rainforest

understorey of varied-sized myrtle (darker, mid-height trees) and

sassafras (lighter, shorter trees). Photo: N. Fitzgerald, 2010.

4 C. Dean et al.

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

The available biomass allometric for the dominant

canopy species (E. delegatensis) covered the DBH

range 0.119–0.832 m (Raison et al. unpublished, in

Keith et al. 2000), which was too limited, as the

mature trees in the Florentine reached over 4.5 m.

Merchantable volume allometrics covered the DBH

range 0.616–1.656 m (Wang & Hamilton 2003).

However, biomass allometrics for the related ash

species, E. obliqua and E. regnans, covered the DBH

ranges 0.262–2.84 m (Keith et al. 2000) and 0.01–

10.0 m (Dean et al. 2003; Dean & Roxburgh 2006)

respectively.

The relationships between biomass and between

stem volumes, for these three eucalypts, are known

within limited ranges of DBH. Their stem volumes

can be calculated using the Farm Forestry Toolbox

(FFT, Warner 2007). Within the FFT the majority

of stem taper formulas and their accommodated

range of DBH, for species encountered in this study,

were commercial-in-confidence and consequently

they could not be deployed in this project. Only

one of the formulas was publically available – for E.

obliqua (Goodwin 1992), which accommodated

DBH 0.1–2.8 m. However, from the FFT, it was

noted that for a given DBH and height, stem volume

was in the order wet E. obliqua4E. delegatensis4E.

regnans. Biomass allometrics based on DBH alone

for the three species are compared in Keith et al.

(2000) over the DBH range 0.1–1.0 m: the same

sequence applies to biomass as for volume (from the

FFT). Thus the sequence represents common trends

in both taper and basic density. Thus combining E.

obliqua and E. regnans would have provided a reliable

approximation for E. delegatensis.

The allometric for merchantable volume could not

be directly converted to one for biomass owing to the

location-specific criteria for merchantability and the

absence of an accompanying allometric for entire

stem volume. However in the same report there was a

merchantable volume allometric for E. regnans over a

comparable DBH range (Wang & Hamilton 2003).

We derived one new allometric using the approxima-

tion that merchantable volume was an equal fraction

of the whole stem volume for both species. When

combined with the allometric for E. regnans stem

volume from Dean and Roxburgh (2006) the ratio of

merchantable volumes for the two species yielded the

stem volume of E. delegatensis, for a given DBH. The

stem volume was converted to dry biomass using the

basic density of 524 kg m73 for E. delegatensis (Ilic

1997) and combined with other aboveground com-

ponents (e.g. branches and leaves) as for E. regnans

from Dean et al. (2003), to provide an allometric for

aboveground biomass for E. delegatensis. The corre-

sponding allometric was called Edel_by_ratio.

Thus, there were three possible biomass allo-

metrics: for E. delegatensis: E. obliqua, E. regnans and

Edel_by_ratio. As none of the three was over-

whelmingly more suited to E. delegatensis over the

sampled range of DBH, the average of the three

was used. That average was fitted by [non-linear]

regression to a logistic form. A range of equation

types was examined using [automated] symbolic

regression (‘‘Eureqa’’ software, Schmidt & Lipson,

2009), but the fit was only improved on the logistic

form when using an order five polynomial (i.e. six

parameters instead of the three for the logistic

function). The logistic function was therefore

retained, and it had the additional benefits of

being better-defined than the polynomial for high

DBH (similarly when compared with log/log

allometrics, which can increase exponentially) and

providing reasonable biomass for DBH below

0.1 m. Adoption of the logistic form was also

found suitable for juvenile to advanced-mature E.

regnans allometry (e.g. Dean et al. 2003, Dean &

Roxburgh 2006).

For the rainforest species, no species-specific

biomass allometrics were available. For some of the

observed species, stem volume could be calculated

using the FFT, however the formulas were not

available, their DBH range was unknown, and we

had not measured tree heights. Two generic biomass

allometrics for rainforest species were: (1) for

temperate rainforest species (Keith et al. 2000) and

(2) wet-sclerophyll and mixed-forest species (Dean

et al. 2003). The Keith et al. (2000) allometric was

comprised of two parts: generic rainforest and a

correction for temperate rainforest (in the form of

data points), both covering the DBH range 0.1–

1.0 m. We derived a single allometric for temperate

rainforest from those two components. The allo-

metric given in Dean et al. (2003) was intended for

use within the spatio-temporal carbon modelling

software, CAR4D. In applications of CAR4D how-

ever the formula was rarely used in its raw form, but

adjusted to suit environmental conditions through

scaling magnitude and growth rate. The most

common adjustment for mixed-forest was to scale

the magnitude by 0.5 (e.g. Dean et al. 2004; Dean &

Roxburgh 2006). That adjustment was applied here.

Thus there were two allometrics available for the

rainforest species: (1) temperate rainforest and (2)

Dean et al. 2003/2004. As neither of these two

allometrics was overwhelmingly more suited to the

understorey in E. delegatensis mixed-forest over

the sampled range of DBH, we used the average of

the two (which additionally provided a gauge of

uncertainty effects in understorey biomass). That

average was formulated as a logistic function, fitted by

[non-linear] regression. As for E. delegatensis, sym-

bolic regression revealed that the fit could only be

improved using a polynomial of order 5, and there-

fore the logistic function was retained.

Old-growth pre-logging C, via allometry 5

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

DBH size distributions of species (per hectare)

for the plots were tallied across the study area,

corrected for plot size (in projection), allocated to

DBH classes and graphed as histograms. DBH

class width was varied between species to provide

maximum clarity in the histograms, (e.g. class

width for E. delegatensis was 0.2 m). The above-

ground biomass allometrics were applied to the

categories, rather than to the trees separately prior

to tallying. The C in aboveground biomass per unit

area for E. delegatensis, in the sample plots, was

calculated using the three allometrics, and aver-

aged; the average was equivalent to using the

derived logistic function. The C in biomass for E.

delegatensis was multiplied by 0.7, 0.75, 0.8 and 1

to test the effect of accommodating senescence on

stand-level carbon stocks (i.e. the sensitivity of the

allometrics to senescence). Similarly to E. delega-

tensis, the C in aboveground understorey biomass

was calculated from both contributing allometrics,

and that average was equivalent to using the

derived logistic function.

The allometric used for determining the carbon in

snags was the average of the three eucalypt allo-

metrics described above for aboveground biomass.

Apart from height loss, to accommodate general loss

of branches, bark, and unrecorded internal voids, the

calculated aboveground mass was multiplied by

0.6666, in addition to the recorded decay class

factor. For example, a snag of decay class S (i.e.

0.33333) had the necromass of an equivalent-sized

living tree, multiplied by 0.2222. Loss of height was

interpreted by multiplying the necromass from the

allometric by the proportion of the original height

remaining. That, in effect, assumed a cylindrical

stem and was therefore more conservative than

volume based on stem taper, when more of the

upper stem had decomposed. The original height

was estimated from the height-DBH relationship for

E. regnans (Dean et al. 2003) but with an adjustment

to accommodate E. delegatensis possibly being of

shorter stature, and to blend with the measured

DBH-height pairs; i.e. a new allometric was derived.

Parameters in that height allometric were varied to

gauge its impact on calculated necromass. The

volume of hollows, fallen logs and branches was

calculated as a frustum of a cone. For snags and

stumps the volume was converted to mass using the

basic density of 524 kg m73 for E. delegatensis (Ilic

1997) and applying the same decay factors as for

the solid section (i.e. all necromass was calculated

as if it were E. delegatensis). For fallen logs and

branches the volume was converted to mass using a

conservative basic density of 400 kg m73, and the

recorded decay class. Necromass was pooled and

calculated on a per plot basis, then averaged over

the study area.

State-wide, wet-eucalypt data

Spatio-temporal, electronic records of forestry opera-

tions in wet-sclerophyll and mixed- E. delegatensis

forests from 1999 to 2009 were obtained from the

Forest Practices Authority (FPA) (the limit of electro-

nic records to-date). The data were examined for

trends in areas logged, and silvicultural prescriptions

over that period, for the forest type in our study area.

Potential biomasses for wet- E. delegatensis, E.

obliqua and E. regnans forests, can be calculated from

the continental potential biomass layers from Berry

and Roderick (2006) (termed B&R here onwards)

and FullCAM (Richards & Brack 2004), by over-

laying them (using GIS) on the vegetation commu-

nity layer (TASVEG, DPIW 2010). The FullCAM

and B&R potential biomass layers were derived by

remote sensing and ground-truthing but with mini-

mal calibration for mature wet-eucalypt (i.e. higher

biomass) forests, resulting in the biomass being

underestimated for that forest type (Richards &

Brack 2004; Berry & Roderick 2006; Berry personal

communication 2010; Keith et al. 2010). Thus those

layers are unsuitable for calculation of absolute

values, for mature wet-eucalypts in Tasmania but

nevertheless, they are satisfactory for ranking relative

potential biomass of similar wet-eucalypt forest types

within a region. We used them to compare relative

biomass for the three, major, ash wet-eucalypt forests

in Tasmania, and to compare the potential biomass

at the study site against that for the average wet-

eucalypt E. delegatensis forest in Tasmania.

The B&R layer quantified C in biomass (including

roots), whereas the FullCAM layer included only C

in aboveground biomass. To average the two layers,

15% (representing root biomass) was subtracted

from the B&R layer before averaging with that from

FullCAM (the figure of 15% was an average of

10.5% for ash eucalypts (Feller 1980) and 20% for

temperate, high-biomass eucalypt forests (Mokany

et al. 2006)). The B&R and FullCAM layers did not

have common medians across the range of vegetation

communities in Tasmania. Therefore, to compare

the relative potential biomass for forests dominated

by the three eucalypts used in allometrics here, the

potential biomasses for each vegetation community

in Tasmania were scaled from 0 to 1 (normalised),

separately for the two layers. Once normalised, the

average between the two layers for each forest-type

was calculated to yield a single ranking scheme.

Results

Allometrics

Comparison of existing stem volume allometrics for

E. obliqua, E. delegatensis and E. regnans (Figure 5)

6 C. Dean et al.

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

showed that extrapolation of the allometric for E.

delegatensis beyond its range to accommodate mature

trees would have produced unreasonably high

volumes (Figure 5). The stem volume of E.

delegatensis, from the ratio of merchantable volumes

of E. delegatensis and E. regnans derived from the

allometrics reported in Wang and Hamilton (2003)

was:

Edel by ratio ¼ Ereg vol

� ð0:62400 þ ð3:6428exp

� ð�0:3337DBH0:57072ÞÞÞð1Þ

where Edel_by_ratio is the stem volume of E.

delegatensis in m3, Ereg_vol is the stem volume of E.

regnans in m3 from Dean and Roxburgh (2006), and

DBH is in metres. The stem volumes from Equation

1 (‘‘Edel_by_ratio’’) were higher than for E. regnans

at small DBH (e.g. E. delegatensis volume was twice

that of E. regnans when DBH 0.08 m), but the ratio

decreased to 1 when DBH 0.285 m, after which it

yielded E delegatensis volumes less than those of E.

regnans (Figure 5).

The averaging of the allometrics for E. obliqua, E.

regnans and Edel_by_ratio produced volumes that

were below that for E. regnans for DBH from 0.15–

3.3 m but above, thereafter (Figure 5). It also

generally fitted the sequence of stem volumes

calculated for the species from the FFT but indicated

lower biomass for higher DBH (i.e. it was more

conservative). The logistic function for E. delegatensis

aboveground biomass, corresponding to the average

of the three contributing allometrics (E. obliqua, E.

regnans and Edel_by_ratio) was:

Edel AGB¼1612:4

� 1� 1

1þ DBHþ0:01ð Þ=12:714ð Þ2:2283

! !

ð2Þ

where Edel_AGB is in Mg of dry matter, and DBH is

in metres (R2¼ 1.00, N¼ 800, DBH¼ 0.01–8.0 m).

The forecast biomasses from Equation 2 (i.e. the

average of the three allometrics) were generally

between those of E. obliqua and Edel_by_ratio and

close to those for E. regnans, for most values of DBH

(similar to the corresponding relationships for stem

volumes, Figure 5).

Combining the biomass allometric for generic

rainforest species with the correction points for

temperate rainforest species from Keith et al.

(2000) gave the following biomass allometric for

temperate rainforest species:

Temp rain AGB ¼ 0:001exp

� ð2:5667 lnðDBHÞ þ 8:9133ÞÞðð3Þ

where Temp_rain_AGB is aboveground dry biomass

in Mg and DBH is in metres. The average of the two

allometrics for understorey biomass could be repre-

sented by the following logistic function:

und ave ¼ 209:72

� 1� 1

1þ DBHþ 0:01ð Þ=4:0551ð Þ2:5063

! !

ð4Þ

Figure 5. Stem volume versus DBH for different allometrics tested for E. delegatensis. Covering DBH range encountered on the study site

(left), and close-up at small DBH (right). Dotted lines are where an allometric from the literature was extrapolated beyond its originally

intended DBH range (and may therefore yield unrealistic volumes). The crosses show volumes calculated using the FFT, for individual trees

which had their DBH and height measured (DBH range accommodated in FFT is unknown).

Old-growth pre-logging C, via allometry 7

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

where und_ave is aboveground dry biomass in Mg,

and DBH is in metres (R2¼ 1.00, N¼ 300,

DBH¼ 0.01–3.0 m). Equation 4 (i.e. the average of

the two allometrics) produced biomasses in between

the two averaged allometrics (as expected) but did

not increase exponentially for high DBH, e.g. when

DBH was41 m (Figure 6).

Stand-level distributions

The stand density of rainforest understorey out-

numbered that of the eucalypt canopy by around

fourfold (Table I). Conversely, average DBH of the

eucalypts was fivefold that of the understorey.

Apart from annual growth increments and long-

evity, that difference is related to the continual

regeneration of the myrtles and sassafras without

fire (Read & Hill 1988, Ashton 2000), which was

indicated in the stand-density-DBH histograms,

showing marked positive skew (Figure 7). Celery-

top pine is more likely to require fire for

regeneration than do myrtle or sassafras – accord-

ingly the smallest specimens were noted in the

same plot where a eucalypts had notable fire scars.

A single age-cohort of ash eucalypts usually has a

DBH frequency distribution in the shape of a log-

normal distribution, i.e. showing a positive skew

(e.g. Ashton 2000). From Figure 7 this implies

Figure 6. Allometrics for carbon in biomass, for rainforest

understorey species. Dotted line indicates extrapolation. ‘‘Aver-

age’’ corresponds to Equation 4 (multiplied by 0.5 to convert from

biomass to carbon).

Table I. Numbers of species present and primary DBH statistics

(standard deviation in brackets).

Species N

DBH (m)

Minimum Maximum Average

E. delegatensis 207 0.21 4.50 1.63 (0.81)

Understorey total 897 0.01 2.52 0.31 (0.29)

Myrtle 353 0.01 2.52 0.41 (0.37)

Sassafras 332 0.01 2.05 0.30 (0.20)

Celery-top pine 83 0.01 0.51 0.17 (0.13)

Figure 7. DBH frequency histograms. Those for understorey

species were converted to natural logarithms (on the ordinate) for

clarity. Where the number of stems per hectare in any DBH class

was51 then the ordinate was shifted upwards for clarity (factor

noted on axis label). Error bars (positive) are the standard errors

corresponding to variability between the 10 sample plots.

8 C. Dean et al.

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

that there were at least two age cohorts of E.

delegatensis present, which was corroborated by the

presence of frequency peak where the youngest

cohort overlapped the next-oldest cohort (near

DBH¼ 1 m).

The E. delegatensis DBH frequency distribution

(which was a compound of separate cohorts) showed

a positive skew (sample skewness 1.6) (Figure 7) but

E. delegatensis DBH carbon distributions (Figure 8)

showed a negative skew (for the allometric average,

sample skewness was -0.41). That negative skew was

more pronounced using the extrapolated allometric

for E. obliqua (sample skewness 71.3) – correspond-

ing to its much higher biomass for high DBH when

compared to the other allometrics (Figure 5). With

the derived biomass allometric (Equation 2, i.e. the

average of the three contributing allometrics) being

conservative, it is noteworthy that the relatively small

percentage of high DBH trees (*1.1 tree per

hectare) from DBH 3.5–4.5 m contributed *16%

to the stand-level, E. delegatensis biomass. The

contribution of high DBH trees is also reflected in

the DBH cut-off that represents 50% of the carbon

distribution, being at 2.4(0.1) m (Table I). That

DBH value did not significantly alter with varying the

senescence.

The effect of eucalypt senescence on the DBH

carbon distribution was only marginally noticeable

in the histograms and was probably more related to

the difference between plots. For example the

maximum difference (between with- and without

senescence) for any DBH class was *13%,

(between DBH classes 2.1 and 2.9 m), but it was

only 3% for the DBH class of 2.7 m. Overall, the

effect of attributing 30% missing biomass to

senescent trees only altered the stand-level

E. delegatensis biomass by *8% (e.g. from the

average of the three allometrics, Table II).

The DBH carbon distributions for understorey

species, when pooled together, showed a marked

positively-skewed distribution (Figure 9). However,

for celery-top pine and sassafras, the higher

DBH classes (0.525 m and 2.05 m respectively)

contributed more to the stand-level carbon

stocks than did their minimum DBH classes, even

though those upper DBH classes were not near

the upper limit of the Dean et al. 2003/2004

allometric.

The understorey contributed 129(28) Mg ha71

(27%) to the stand-level carbon in biomass, with the

majority of that (66%), coming from myrtle (Table

III). The contribution of only 2% from celery-top

pine was due to its limited occurrence and smaller

average size.

For E. delegatensis the 25%-senescence scenario

was considered most likely and sufficiently conser-

vative for this site. Using the carbon in biomass for

that scenario from the E. delegatensis allometric-

average [466(132) Mg ha71] (Table II), and the

understorey carbon from the understorey allometric-

average [129(28) Mg ha71] (Table III), gave the

stand-level carbon in biomass of 469(132) Mg ha71.

Variation of parameters, within realistic ranges, for

the height allometric for snags only contributed to

*5% variation in the total necromass of 153(51) Mg

ha71, due to the relatively small contribution from

snags (Table IV). Combining the stand-level biomass

carbon with the necromass carbon, gave a stand-level

Figure 8. DBH carbon histograms for E. delegatensis using different

allometrics. Error bars (positive) are the [fractional] standard

errors corresponding to variability between the 10 sample plots

[multiplied by the category values].

Old-growth pre-logging C, via allometry 9

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

Table II. Carbon density for E. delegatensis in the study site using different allometrics and different percentage of missing biomass for trees

recorded as senescent; and DBH category above which lies 50% of the stand-level carbon. The allometrics used for ‘‘Eobl_equiv’’ and for

‘‘Ereg_equiv’’ were from Keith et al. 2000 (extrapolated) and Dean et al. 2003 and Dean & Roxburgh 2006. That for ‘‘Edel_by_ratio’’ was

from Equation 1. (Standard deviations in brackets are from the average of the three allometrics).

Allometric

Eobl_equiv Ereg_equiv Edel_by_ratio

Average

% of total

stand biomass

Stand-level C (Mg ha71)

0% Senescence 459 382 242 361 (110) 78

20% Senescence 438 364 231 344 (105) 74

25% Senescence 432 359 228 340 (104) 73

30% Senescence 427 355 225 336 (103) 72

DBH cut-off for

50% of stand C (m) 2.54 2.39 2.37 2.43 (0.09)

Figure 9. DBH carbon histograms for rainforest understorey using different allometrics (all-understorey and dominant species separately).

‘‘ctp’’ is celery-top pine. Error bars (positive) are the [fractional] standard errors corresponding to variability between the 10 sample plots

[multiplied by the category values].

10 C. Dean et al.

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

aboveground carbon of 619(180) Mg ha71. Thus the

understorey contributed *20% to the total, and

necromass contributed *25%, of which 83(13)%

was from fallen logs, i.e. fallen logs contributed

*21% to aboveground carbon, which was roughly

equivalent to the *20% contribution from the

understorey species in total.

State-wide data and logging emissions

The average carbon in biomass for the 10 plots, as

derived by GIS analysis from the FullCAM & B&R

layers, was 69.5(0.4) and 115.5 Mg ha71 respec-

tively, which was within one standard deviation of

the State-wide averages of 66(37) and 101(15) Mg

ha71 (Table V). Thus the site was reasonably

representative for potential carbon density in that

forest type, State-wide. Using the 50% long-term

decline in carbon stocks noted for conversion of

old-growth forests to a harvesting cycle (Dean &

Wardell-Johnson 2010), if the forest in the study

was clearfelled and converted to CBS logging

cycles then the forecasted, long-term above-

ground-carbon emission is 311(90) Mg ha71.

The logging records from the FPA revealed that

conversion of wet E. delegatensis forest to plantation

had declined overall from 1999 to 2009 at a rate of

*89 ha year71 (R2¼ 0.42), however CBS-type

logging showed a slight increase (*28 ha year71,

R2¼ 0.36, p¼ 0.053). The total area of wet E.

delegatensis forest logged in that period was 38225

ha (Table V), of which 7920 ha was converted to

plantation and 5523 ha was converted through

CBS. The long-term aboveground-carbon emission

from that clearfelling (using the 50% figure), using

the unit-area, long-term emission of 311(90) Mg

ha71, is likely to be as high as 4.2(1.2) Tg,

depending on the growth and seral stage at the

time of logging the primary forest. Additionally,

from State-wide inventory, Moroni et al. (2010)

found 232 Mg ha71 for mature (i.e.� 110 years)

wet-eucalypt forests (averaged over all eucalypt

types, mixed-forest and wet-sclerophyll forest).

However their carbon density for mature forest

was not significantly different to their value for

average-aged forest, and there was no differentia-

tion of carbon densities between old-growth and

early mature forest even though the former can be

50% higher (Dean 2011). Nevertheless, assuming

their value was applicable to mature, primary, wet

E. delegatensis forest then the long-term carbon

emission (50% of stock) would be 1.6 Tg. An

average of our carbon density (which appears

representative of potential biomass in E. delegatensis

State-wide) with that from Moronoi et al. (2010),

may well be representative of the growth and seral

stages of the forests logged from 1999 to 2009:

giving a long-term aboveground-carbon emission of

1.6-to-4.2 Tg, i.e. 2.9(+1.3) Tg.

Table III. Carbon in aboveground biomass for understorey species, from allometrics; and DBH category above- which is 50% of stand-level

carbon. The allometric ‘‘Keith et al. (2000) extrapolated’’ is Equation 3 (standard deviations in brackets are from the average of the two

allometrics).

Allometric

From Keith et al. 2000.

extrapolated Dean et al. 2003/2004

Average

% understorey % of total stand biomass

Stand-level C (Mg ha71)

All understorey 149 109 129(28) 100 27

Myrtle 95 69 82(18) 66 17

Sassafras 44 32 38(8) 29 8.1

Celery-top pine 4.2 3.4 3.8(0.5) 1.9 0.81

DBH cut-off for 50% of stand C (m)

All understorey 0.70 0.62 0.60(0.06)

Myrtle 0.82 0.77 0.81(0.04)

Sassafras 0.50 0.45 0.50(0.04)

Celery-top pine 0.22 0.23 0.23(0.01)

Table IV. Carbon in necromass (within the 30 m radius circle)

summed per plot, separated into snags and fallen logs (standard

deviation in brackets).

Plot

Snags

(Mg ha71)

Fallen logs

(Mg ha71)

Necromass

(Mg ha71)

Fallen logs

% of total

1 26.8 141.5 168.3 84.1

2 2.5 196.0 198.5 98.7

3 10.4 124.0 134.4 92.3

4 83.9 142.0 225.9 62.9

5 13.8 91.4 105.2 86.9

6 43.7 83.8 127.4 65.7

7 9.2 75.5 84.7 89.1

8 7.3 215.0 222.3 96.7

9 32.2 65.8 98.0 67.2

10 16.1 149.9 166.1 90.3

Average 25(24) 128(51) 153(51) 83(13)

Old-growth pre-logging C, via allometry 11

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

Discussion

Allometric derivation

The allometrics derived were stable over all of the

DBH values measured in the field, i.e. they did not

increase exponentially at high DBH and did not

produce negative values for DBH50.1 m. The

Edel_by_ratio may have produced E. delegatensis

values of biomass too high when DBH was below

0.29 m, but that effect would have been cancelled to

some degree by averaging with the other two

allometrics. Additionally, when the E. delegatensis

biomass for the study site was tallied then high

biomass at low DBH values was not observed

(Figure 8). For DBH4 0.53 m, stem volume from

Edel_by_ratio was the smallest of all allometrics,

which concords with the more-pronounced buttres-

sing of E. regnans (which produces faster tapering up

the length of the stem and hence lower volumes for

the same DBH). For the same reason, the similarity

of biomass calculated for E. delegatensis (from the

average allometric, Equation 2), and for E. regnans,

indicates that Equation 2 (i.e. the average of the three

allometrics) was conservative. That was confirmed

by comparison with the relative volumes and

biomasses for the three eucalypt species described

above, and by the volumes from the FFT being

higher (for high DBH) than the average allometric.

Thus it is unlikely that E. delegatensis biomass was

overestimated in this study.

The allometric derived for the understorey in the

mixed-forest, being of logistic form, was similarly

stable at high values of DBH. Until destructive

sampling is undertaken (such as during logging

operations, e.g. Ximenes et al. 2008), or accurate

taper formulas for the rainforest species become

publically available, the accuracy of their carbon

stocks cannot be increased. Although an average of

the allometrics for mixed-forest (Dean et al. 2003,

2004) and temperate rainforest (Equation 3) was

derived (Equation 4) and applied here, there was no

compelling reason to assume its effect was more

accurate than that from the original mixed-forest

allometric alone. In fact, as the temperate rainforest

allometric had to be extrapolated far beyond its DBH

range used in its construction, the original mixed-

forest allometric may be a more reliable representa-

tion. That would decrease the carbon contribution

from understorey species by 15%, and reduce

the total aboveground carbon by 3% to 602 (174)

Mg ha71.

Carbon distribution and significance

Three factors indicating the importance of reliable

allometrics for high DBH eucalypts, from the present

study are: (1) the negative skew on the DBH carbon

distribution, (2) the 50% DBH cut-off of 2.4(0.1) m

for E. delegatensis and (3) the 16% contribution to

eucalypt carbon stocks from the *1.1 trees ha71

with DBH� 3.5 m. If the study site had comprised

only a single age cohort then the high DBH trees

would have contributed a marginally higher percen-

tage. The E. obliqua biomass allometrics can accom-

modate up to DBH 2.8 m, but that is the highest in

the review of allometrics in Keith et al. (2000). The

E. regnans allometrics (Dean et al. 2003; Dean &

Roxburgh 2006) were calibrated with trees up to

*6.44 m DBH and corrected for cross-sectional

area deficit. Thus, in general, more work is required

on carbon accounting of old-growth wet-eucalypt

forests. The higher standard error on the highest

DBH category in the carbon DBH histograms of

some of the understorey trees indicates the more-

sporadic location of such large trees.

The necromass contribution of 25% to above-

ground carbon is above the mean value for the wider

category of Australian tall-open forests (Woldendorp

& Keenan 2005). The absolute value of the stock

however, agrees with other values for Tasmania,

which increase with more southern latitudes (i.e.

cooler temperatures), and the proportion of necro-

mass in fallen logs agrees with higher values noted

in old-growth forests (Woldendorp & Keenan

2005).

The high, measured biomass compared with the

potential value in the FullCAM and B&R layers

concurs with that found previously for wet-eucalypt

forest (e.g. Dean & Wardell-Johnson 2010). The

current, wide use of the FullCAM layer suggests that

Table V. Hectarages of main wet-eucalypt types logged from 1999 to 2009; and average potential carbon in biomass derived by overlaying

the FullCAM and B&R layers on vegetation mapping (TASVEG). Values were normalised (state-wide, 0 to 1 for the two layers) then

averaged, to allow comparison between forest types.

Wet eucalypt

Area

logged (ha)

Area

remaining (ha)

Number of TASVEG

polygons (thousands)

C in biomass (Mg ha71)

FullCAM B&R Normalised & Average

E. delegatensis 38,224 276,378 4135 66(37) 101(15) 0.59

E. obliqua 60,384 464,082 7425 84(53) 106(24) 0.73

E. regnans 20,485 76,891 1230 82(53) 117(20) 0.76

12 C. Dean et al.

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

analysis of old-growth carbon stocks can provide

useful calibration points for its adjustment.

The 466(132) Mg ha71 of aboveground biomass

carbon found here, is in between that for old-growth

forests measured elsewhere, and just above the

median value for temperate forests in the review of

Keith et al. 2009. For example, it is above that found

in the Carpathian mountains (Szwagrzyk & Gazda

2007 – 268 Mg ha71, Keeton et al. 2010 – 160 Mg

ha71) but lower than other sites in Tasmania (Dean

& Roxburgh 2006 – 723 and 642 Mg ha71, for E.

regnans and E. obliqua respectively), in Victoria

(Australia) (Van Pelt et al. 2004 – *843 Mg ha71;

Keith et al. 2009 – 1819 Mg ha71 for E. regnans) and

in the USA [e.g. Van Pelt et al. 2004 – *647 Mg

ha71 for Pseudotsuga menziesii Mirb. (Franco)]. A

value of 723 rather than 850 Mg ha71 of Wood et al.

(2010) is used here for the Tasmanian E. regnans, as

the latter included root carbon. Nevertheless, the

aboveground biomass carbon measured in this work,

when compared with two other old-growth Tasmania

examples (Dean & Roxburgh 2006), aligns well

with the sequence of normalised Tasmanian state-

wide potential biomass layers (Table V): i.e.

4665 6425 723 Mg ha71 corresponding to

0.595 0.735 0.76 for E. delegatensis, E. obliqua

and E. regnans respectively.

Aboveground carbon, one of the major carbon

pools, is the easiest pool to measure in native forests.

These carbon stocks in old-growth wet-eucalypt

forests, under certain combinations of age classes,

can be much higher than those used in routine

accounting (e.g. Keith et al. 2009). With the land-

scape-level, age-averaged stock of carbon being the

pre-logging benchmark, it is necessary to account for

such potentials. Upper stock limits are likely when

the forest has significant old-growth characteristics,

with other components that may be either younger

age classes of the canopy species (recruits resulting

from non-stand-replacing fires or non-fire-induced

recruits) or rainforest understorey species capable of

continuous regeneration in the absence of fire. The

data collection and processing presented here pro-

vide a clear example of how the pre-logging stocks

can be tallied for use in regional comparisons.

Acknowledgements

The authors are indebted to numerous volunteers

who performed fieldwork and support. The authors

also thank, for helpful comments and/or data:

Stephen Roxburgh (CSIRO), Sandy Berry (ANU);

Bureau of Meteorology, Adrian Goodwin (Bushlo-

gic), Department of Climate Change and Energy

Efficiency, Forestry Tasmania, Daniel Livingston

(FPA), Private Forests Tasmania, and two anon-

ymous reviewers.

References

Ashton DH. 2000. The Big Ash forest, Wallaby Creek, Victoria –

Changes during one lifetime. Aus J Bot 48: 1–26.

Ashton DH, Attiwill PM. 1994. Tall Open-forests. In: Groves,

RH, editor. Australian vegetation, second edn. Cambridge:

Cambridge University Press. pp. 157–196.

Berry SL, Roderick ML. 2006. Changing Australian vegetation

from 1788 to 1988: Effects of CO2 and land-use change. Aust J

Bot 54: 325–338.

Dean C. 2011. Comment on ‘Carbon in trees in Tasmanian State

forest’. Int J For Res. Available: http://www.hindawi.com/

journals/ijfr/aip/212361. Accessed May 2011 17.

Dean C, Roxburgh SH. 2006. Improving visualisation of mature,

high-carbon-sequestering forests. For Biom Mod Inf Sci [serial

online]: 48–69. Available: http://www.fbmis.info. Accessed Oct

2010 16.

Dean C, Roxburgh SH, Mackey BG. 2003. Growth modelling of

eucalyptus regnans for carbon accounting at the Landscape

Scale. In: Amaro. A, Reed, D, Soares P, editors. Modelling

forest systems. Wallingford, UK: CAB International. pp. 27–

39 þ plates.

Dean C, Roxburgh SH, Mackey BG. 2004. Forecasting land-

scape-level carbon sequestration using gridded, spatially

adjusted tree growth. For Ecol Manage 194: 109–129.

Dean C, Wardell-Johnson GW. 2010. Old-growth forests, carbon

and climate change: Functions and management for tall open-

forests in two hotspots of temperate Australia. Plant Biosys

144: 180–193.

DPIW. 2010. TASVEG, Base data from theLIST, @ State of

Tasmania. Department of Primary Industries and Water,

Hobart. Available from: http://listdata.thelist.tas.gov.au. Ac-

cessed Aug 2010 10.

Elliot HJ, Felton KC, Jarman SJ, Stone MG. 2008. A history of

innovation: Eighty-five years of research and development at

forestry Tasmania. Hobart: Forestry Tasmania. p. 402.

Ellis RC, Lowry RK, Davies SK. 1982. The effect of regeneration

burning upon the nutrient status of soil in two forest types in

southern Tasmania. Plant Soil 65: 171–186.

Feller MC. 1980. Biomass and nutrient distribution in two

eucalypt forest ecosystems. Austral J Ecol 5: 309–333.

Felmingham B, Farley M, Lancaster G, Farley C. 2004. Impact of

the policy to cease clearfelling of old growth forests in 2010: An

overview of productivity, financial & employment aspects. A

report prepared for the Tasmanian Forest & Timber Indus-

tries. Symetrics Business Intelligence. Available: http://

www.nafi.com.au/userfiles/publication/Felmingham%20Report.

pdf. Accessed Oct 2010 16.

Ferre C, Leip A, Matteucci G, Previtali F, Seufert G. 2005.

Impact of 40 years poplar cultivation on soil carbon stocks and

greenhouse gas fluxes. Biogeoscience Disc 2(4): 897–931.

Florence RG. 1993. Forestry in transition in Australia: From the

primacy of wood production to ecologically sustainable

development. Commonwealth For Rev 72: 321–337.

Forestry Tasmania. 2009. A New silviculture for Tasmania’s

public forests: A review of the variable retention program.

Hobart: Forestry Tasmania. p. 107. Available: http://www.stors.

tas.gov.au/au-7-0095-01383. Accessed Sep 2010 22.

Gilbert JM. 1959. Forest succession in the Florentine Valley,

Tasmania. Papers Proc Roy Soc Tasmania 93: 129–151.

Gloor M, Sarmiento JL, Gruber N. 2010. What can be learned about

carbon cycle feedbacks from CO2 airborne feedback. Atmos

Chem Phys 10: 7739–7751. doi:10.5194/acp-10-7739-2010.

Goodwin AN. 1992. A stem taper formula for E. obliqua.

Proceedings, IUFRO Conference (S 4.01, Mensuration,

Growth and Yield) on integrating forest information over

space and time, 1992 Jan 13–17. Canberra, Australia. pp. 454–

467.

Old-growth pre-logging C, via allometry 13

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11

Helms AD. 1945. A giant eucalypt. Aust For 9: 24–28.

Hickey JE. 1994. A floristic comparison of vascular species in

Tasmanian oldgrowth mixed forest with regeneration resulting

from logging and wildfire. Aust J Bot 42: 383–404.

Houghton J. 1997. Global warming: The complete briefing.

Cambridge: Cambridge University Press.

Ilic J. 1997. Woods of eucalyptus – part1. Distinguishing three

species from the ash group (E. regnans, E. delegatensis and E.

obliqua). IAWA J 18: 27–36.

Johnson K, Scatena FN, Pan Y. 2010. Short- and long-term

responses of total soil organic carbon to harvesting in a

northern hardwood forest. For Ecol Manage 259: 1262–1267.

Kashian DM, Romme W, Tinker DB, Turner MG, Ryan MG.

2006. Carbon storage on landscapes with stand-replacing fires.

BioScience 56: 598–606.

Keeton SW, Chernyavskyy M, Gratzer G, Main-Knorn M,

Shpylchak M, Bihun Y. 2010. Structural characteristics and

aboveground biomass of old-growth spruce-fir stands in the

eastern Carpathian Mountains, Ukraine. Plant Biosys 144: 1–

12.

Keith H, Barrett D, Keenan R. 2000. Review of allometric

relationships for estimating woody biomass for New South

Wales, the Australian Capital Territory, Victoria, Tasmania

and South Australia. National Carbon Accounting System

Technical Report 5b. Canberra: Australian Greenhouse Office.

p. 113.

Keith H, Mackey BG, Berry SL, Lindenmayer DB, Gibbons P.

2010. Estimating carbon carrying capacity in natural forest

ecosystems across heterogeneous landscapes: Addressing

sources of error. Glob Chang Biol 16: 2971–2989.

Keith H, Mackey BG, Lindenmayer DB. 2009. Re-evaluation of

forest biomass carbon stocks and lessons from the world’s most

carbon-dense forests. Proc Nat Acad Sci USA. doi 10.1073/

pnas.0901970106

Lindenmayer DB, Hunter ML, Burton J, Gibbons P. 2009. Effects

of logging on fire regimes in moist forests. Conserv Lett 2:

271–277.

Meir P, Cox P, Grace J. 2006. The influence of terrestrial

ecosystems on climate. Trends Ecol Evol 21: 254–260.

Mokany K, Raison JR, Prokushkin AS. 2006. Critical analysis of

root:shoot ratios in terrestrial biomes. Glob Chang Biol 12: 84–

96. doi: 10.1111/j.1365-2486.2005.001043.x

Moroni MT, Kelley TH, McLarin ML. 2010. Carbon in Trees in

Tasmanian State Forest. Int J For Res. doi:10.1155/2010/

690462

Read J, Hill RS. 1988. The dynamics of some rainforest

associations in Tasmania. J Ecol 76: 558–584.

Richards GP, Brack C. 2004. A continental biomass stock and

stock change estimation approach for Australia. Aust For 67:

284–288.

Schmidt M, Lipson H. 2009. Distilling free-form natural laws

from experimental data, science 324: 81–85.

Szwagrzyk J, Gazda A. 2007. Aboveground standing biomass and

tree species diversity in natural stands of Central Europe. J

Vegetat Sci 18: 563–570.

Tasmanian Government. 2006. Woodchip production data native

forest. Hobart: State of the Environment Tasmania, Tasma-

nian Government. Available: http://soer.justice.tas.gov.au/

2003/table/332/index.php. Accessed Nov 2010 29.

Van Pelt R, Sillet S, Nadkarni NM. 2004. Quantifying and

visualizing canopy structure in tall forests; methods and a case

study. In: Lowman MD, Rinker HB, editors. Forest Canopies.

San Diego, CA: Elsevier Academic Press. pp. 49–72.

Walker B, Felton K. 2007. Tasmania. In: Raison, RJ, Squire, RO,

editors. Forest management in Australia: Implications for

carbon budgets. National Carbon Accounting System Tech-

nical Report 32. Part 1. Canberra: Australian Greenhouse

Office. pp. 147–192.

Wang Y, Hamilton F. 2003. Projecting stand attributes of

regrowth ash eucalypts sampled in forest inventory. Aus For

66, 75–86.

Warner A. 2007. Farm forestry toolbox version 5.0: Helping

Australian growers to manage their trees. RIRDC Publication

No 07/135. Canberra: Rural Industries research and Develop-

ment Corporation.

Woldendorp G, Keenan RJ. 2005. Coarse woody debris in

Australian forest ecosystems: A review. Austral Ecol 30: 834–

843.

Wood SW, Allen KJ, Hua Q, Bowman DMJS. 2010. Age and

growth of a fire prone Tasmanian temperate old-growth forest

stand dominated by Eucalyptus regnans, the world’s tallest

angiosperm. For Ecol Manage 260: 438–447.

Ximenes FA, Gardner WD, Kathuria A. 2008. Proportion of

above-ground biomass in commercial logs and residues

following the harvest of five commercial forest species in

Australia. For Ecol Manage 256: 335–346.

14 C. Dean et al.

Dow

nloa

ded

at 2

2:51

17

Dec

embe

r 20

11