Density loss and respiration rates in coarse woody debris of< i> Pinus radiata,< i> Eucalyptus regnans and< i> Eucalyptus maculata

Density loss and respiration rates in coarse woody debris of Pinus

radiata, Eucalyptus regnans and Eucalyptus maculata

Jens Mackensena,1, Jurgen Bauhusb,*

aUnited Nations Environment Programme, Division of Policy Development and Law, P.O. Box 30552, Nairobi, KenyabAustralian National University, School of Resources, Environment and Society, Forestry Program and Cooperative

Research Centre for Greenhouse Accounting, Canberra, ACT 0200, Australia

Received 28 January 2002; received in revised form 6 September 2002; accepted 11 September 2002

Abstract

This study compared field and laboratory decomposition rates of coarse woody debris (CWD) (.10 cm diameter) from three tree species:

Pinus radiata, Eucalyptus regnans, and Eucalyptus maculata. For this purpose, the density loss of logs on the ground sampled from

chronosequences of sites following harvesting was determined using the water replacement technique. P. radiata logs were sampled 1, 2.5, 6,

and 9 years following harvesting, and logs of E. regnans and E. maculata were collected from sites that were harvested 1, 3.5, 6.5, and 12 and

1.5, 6.5, and 11.5 years ago, respectively. In addition, the C/N ratio of wood was determined and current respiration rates of logs from these

different age classes were measured through laboratory incubation. The times for loss of 95% of material (t0.95) determined from density loss

for these species were 24 years for P. radiata, 43 years for E. regnans, and 62 years for E. maculata. The decomposition rates of CWD

derived from laboratory respiration were 6.1, 5.9 and 11.9 times higher than the decay rates from density loss in P. radiata, E. regnans, and E.

maculata, respectively. This points to severe constraints of decomposition through adverse conditions in the field. The changes in respiration

rates and C/N ratio with age of decaying logs indicated that the single component, negative exponential decay model could be applied

satisfactorily only to P. radiata. In the case of the eucalypt species, substrate quality (expressed through respiration rates) declined in the

oldest samples. This may be explained by the loss of rapidly decomposing sapwood and the retention of more decay-resistant heartwood. In

these cases, a two-component model will be more suitable to describe the density loss of decaying wood.

q 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Coarse woody debris; Wood density; Wood respiration; Decomposition

1. Introduction

Recent estimates for CO2-emissions from land use

changes and forestry activities are subject to large

uncertainties (NGGIC, 1997a,b). These uncertainties have

a range of sources, such as the area being cleared, carbon

density of vegetation, fraction of cleared biomass which is

burned, and the rate at which the remaining biomass decays.

In the absence of better information, the release of carbon

from land-use change and forestry activities through the

decay of residual biomass is being calculated according to

the Intergovernmental Panel on Climate Change (IPCC,

1997) default, which is equal to a linear decay of litter over

10 years. The IPCC default for decay of above- and below-

ground litter appears to be a coarse estimate of actual carbon

losses from decay. It does not consider the climatic and

environmental variables controlling decay nor does it take

into account other variables linked to decay such as size

distribution of litter, density, chemical composition and the

position in which litter is decaying.

The problem of obtaining more reliable decay estimates

lies in the long-term nature of the decay process, in

particular of the coarse woody debris (CWD), which may

constitute the majority of the residual biomass following

land-use change and forestry activities. Few studies have

determined the decay of individual pieces of wood over a

long time (Arthur et al., 1993; Stone et al., 1998). To

overcome this problem, many studies have assessed the

decay of CWD from chronosequences, where the period for

which wood was decaying was known (Grier, 1978; Graham

0038-0717/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.

PII: S0 03 8 -0 71 7 (0 2) 00 2 55 -9

Soil Biology & Biochemistry 35 (2003) 177–186

www.elsevier.com/locate/soilbio

1 Tel.: þ254-2-62-4251; fax: þ254-2-62-4324.

* Corresponding author. Address: School of Resources, Environment and

Society, Australian National University, Forestry Program, Canberra, ACT

0200, Australia. Tel.: þ61-2-6125-2748; fax: þ61-2-6125-0746.

E-mail addresses: [email protected] (J. Bauhus), jens.

[email protected] (J. Mackensen).

and Cromack, 1982; Fahey, 1983; Busse, 1994). A major

difficulty with this approach is in the assigning of

decomposition periods to certain logs, in particular where

dendrochronological approaches cannot be employed.

Decaying logs have to be traced back to certain harvesting

or disturbance events. In a particular stand, however,

different logs may have resulted from different events,

logs on the forest floor might have started to decay while

standing and alive, and they might not have decayed in this

position, e.g. on the ground or suspended, for the entire

period. In addition, it may be difficult to assign decaying

logs to tree species in mixed-species forest stands.

In this study we wanted to investigate how reliable decay

estimates of CWD were by assessing mass loss of three

different tree species sampled from chronosequences. For

this purpose we also determined the respiration rates and the

carbon-to-nitrogen ratios of the decaying wood. While the

C-to-N ratio can be regarded as a measure of the decay

status of litter, the respiration rate can be viewed as an

indicator of substrate availability to decay organisms, and as

such as a measure of substrate quality (Swift, 1977; Swift

et al., 1979). Nitrogen limits wood-decomposing micro-

organisms (Cowling, 1972). Therefore N is immobilised by

microbes during the wood decomposition process while C is

released through respiration (Harmon et al., 1986). As a

consequence the C-to-N ratio declines, and can thus be used

as an indicator of the decomposition stage within a given

wood type. We assumed that a comparison of mass loss

(wood density decrease), respiration rate, and C-to-N ratio

would permit the assessment of the reliability of mass loss

measurements. Respiration rates should also indicate the

applicability of constant decay factors in models of CWD

decomposition for these species.

2. Material and methods

2.1. Study sites

We measured decomposition rates of CWD using two

complementary methods: wood density decrease (method

A) and CO2 emission rate (method B). Both methods were

applied to samples of three tree species: Monterey pine

(Pinus radiata D. Don), Mountain Ash (Eucalyptus regnans

F. Muell.) and Spotted Gum (E. maculata Hook.). These

species represent important forest ecosystems in south-

eastern Australia. They also represent low-, medium- and

high-density wood species, respectively. Details about

sample area, climate and management systems are provided

in Table 1.

P. radiata samples from a chronosequence following

harvesting, including 1, 2.5, 6, and 9 years, were collected

from similar sites within the Australian Capital Territory.

The external cylindrical shape of log sections sampled was

intact. This was not representative of the sites from which

they were sampled, because it is a standard practice to crush

the woody residues following harvesting to facilitate site

preparation and planting. However, this mechanical frag-

mentation was not specifically addressed in our study.

E. regnans samples were collected from sites in the

Toolangi district, Victoria, that were clear-fell harvested 1,

3.5, 6.5, and 12 years ago. The regeneration practice in these

forests is clearfelling followed by a hot burn of logging

residues. Care was taken to avoid sampling charred

material. This could not be avoided in all cases. However,

the density of charred samples was not significantly

different from unburned samples.

E. maculata samples were collected from North Broo-

man State Forest, South Coast of NSW. The different sites

were harvested 1.5, 6.5, and 11.5 years ago. Sampling was

concentrated on pure patches of E. maculata in otherwise

mixed forests. Initially it was intended to sample older age

classes as well, but it was not possible to determine the

species with certainty in the field once the bark had fallen

off. For the ages that were sampled in this study, the logs

could be linked to stumps, that were either still carrying the

bark or had coppiced which allowed species identification.

2.2. Sample processing and determination of wood density

and respiration rate

Method A included sampling of cross-sections of CWD

of known age since clearing and determination of wood

density per disc. By establishing a chronosequence of sites

since disturbance, the decrease in wood density or weight

loss compared to initial density was used to calculate

decomposition rates. To reduce the variability of decay

stages within age-cohorts, samples were only collected from

logs that were positioned on the soil surface. Two cross

sections of approx. 5 cm width were sampled within a

distance of 0.3–1 m of each other per log. Tree disks were

cut using a chainsaw. Between 16 and 22 logs ranging in

diameter from 10 to 30 cm were sampled per species and

age class (Table 1).

Samples were stored in polyethylene bags for up to 5

days during fieldwork and were later stored cool (4 8C) until

further processing. Bark, mosses, fungal fruiting bodies and

soil were removed from samples prior to measurements.

Fresh weight and diameter of each disk was measured and

its volume was determined gravimetrically by water

displacement (Krankina and Harmon, 1995). Samples

were then oven-dried at 105 8C to a constant weight.

Water content and wood density were calculated

subsequently.

P. radiata samples of the two oldest age classes were

often fragile. Therefore one block of 25 cm in length was

sampled per log. To avoid disintegration this section was

wrapped with adhesive tape prior to cutting. For these

samples the volume was determined by measuring the exact

length and diameter at both ends, from which the cylindrical

volume could be calculated. Although this method neglects

decomposition due to fragmentation and biological

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186178

transformation, it is considered a standard method for

estimation of decomposition rates (Healey and Swift, 1971;

Christensen, 1984).

To measure respiration of wood samples (Method B),

five log sections per age class and species were collected for

incubation. These samples were log sections that were cut

out between the disks collected for method A. Incubation

samples had an average length of 30 cm and diameters

between 10 and 30 cm. Storage of samples was the same as

for the wood density samples. Water content and wood

density of these samples were derived from the accompany-

ing cross-section samples (see method A).

Samples were kept moist and were incubated in airtight

bags (REXAM, one side metallised, biaxially orientated

polypropylene film) at constant temperature (25 8C) in

controlled-climate chambers. Samples were incubated for

three days and CO2 emissions were determined using soda

lime. This method is well established for soil respiration

measurements (Edwards, 1982). Dried soda lime (20 g) was

placed into the incubation bag with the sample. Before and

after incubation, the soda lime was dried at 105 8C and

weighed. The gain in weight after incubation equalled the

adsorbed amount of CO2, corrected for the chemically

released water. The factor applied to correct for the released

water was 1.69. From this the amount of C released per unit

of mass remaining was calculated.

The C and N concentrations in wood were determined

using the combustion furnace technique of an automated

C/N-analyser (LECO-CHN-2000). Prior to analysis the

samples were finely ground and oven-dried at 60 8C for

24 h.

2.3. Calculations and statistics

Various mathematical models have been used to describe

decomposition patterns. The single-exponential model (Eq.

(1)) is based on the assumption that the decomposition rate

is proportional to the amount of matter remaining. It is the

most common model used to describe decomposition

patterns (Jenny et al., 1949; Olson, 1963; Wieder and

Lang, 1982; Means et al., 1985; Harmon et al., 1986;

Lambert et al., 1980). Other models like the multiple-

exponential model, lag time model or linear model exist but

are less frequently used (for discussion see Mackensen and

Bauhus (1999)).

X ¼ X0 e2kt ð1Þ

where X is for e.g. present wood density, X0 is initial wood

density, k is decomposition constant, t is the time. This type

of model was fitted to the mass loss data.

Assuming an exponential decay model, the time to

decompose 50% (t0.5) or 95% (t0.95) of CWD can be

calculated according to Eqs. (2) and (3) (Olson, 1963):

t0:5 ¼ 2lnð0:5Þ=k ¼ 0:693=k ð2Þ

Table 1

Location, climate, stand history and number of samples (n ) for sampled stands of P. radiata, E. regnans and E. maculata

P. radiata

Sample region ACT Canberra, Pierces Creek and Tidbinbilla Forestry Districts (358250S,

1488550E)

Climate Median prec. 791 mm yr21, mean daily min-max temp. 7.3 8C/19.5 8C

Sampling date 21/6–24/6/1999

Number of samples Age-class 1 20 logs @ 2 discs per log ðn ¼ 40Þ

Age-class 2.5 20 logs @ 2 discs per log ðn ¼ 40Þ

Age-class 6 20 logs @ 1 block (25 cm) per log ðn ¼ 20Þ

Age-class 9 20 logs @ 1 block (25 cm) per log ðn ¼ 20Þ

E. regnans

Sample region Central Highlands, Victoria, Toolangi DNRE Forest District (378330S,

1458270E)

Climate Median prec. 1300–1500 mm yr21, mean daily min–max temp. 4 8C/23 8C

Sampling date 26/5–29/5/1999

Number of samples Age-class 1 20 logs @ 2 discs per log ðn ¼ 39Þ

Age-class 3.5 20 logs @ 2 discs per log ðn ¼ 33Þ

Age-class 6.5 20 logs @ 1 block per log ðn ¼ 62Þ

Age-class 12 20 logs @ 1 block per log ðn ¼ 37Þ

E. maculata

Sample region South Coast, New South Wales, North Brooman State Forest (358300S,

1508150E)

Climate Median prec. 1100 mm yr21, mean daily min–max temp. 11.2 8C/21.2 8C

Sampling date 28/6–30/6/1999

Number of samples Age-class 1.5 22 logs @ discs per log ðn ¼ 44Þ

Age-class 6.5 16 logs @ discs per log ðn ¼ 31Þ

Age-class 11.5 18 logs @ discs per log ðn ¼ 36Þ

Samples of P. radiata and E. regnans were collected on sites that had been clear-felled previously. E. maculata samples were from selectively logged sites,

with documented harvesting history.

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186 179

t0:95 ¼ 2lnð0:05Þ=k ¼ 2:996=k ð3Þ

To test for differences between respiration rates and C-to-N

ratios of wood samples of different age, analyses of variance

were performed. Welch’s ANOVA was used in case the

variances were unequal for the different age classes (SAS

Institute, 1996). Multiple linear regression analysis was

used to explain relations between different parameters

including respiration rates, age, size, density, moisture

content and C/N ratio. Only significant (P , 0.05)

regressions are being reported.

To compare the annual decay rates determined from

density loss with daily respiration rates determined in short

term incubations, the annual decay rates were divided by

365 to convert them to daily rates. No correction factor to

convert density loss to C loss was required for this

comparison, because the C concentration in CWD of all

species remained constant over the age classes. Thus C is a

constant proportion of both remaining and lost mass. Since

decomposition processes are temperature-dependent, differ-

ences in temperature between the field and the laboratory

also have to be considered. Burger and Pritchett (1984)

provided the following formula for this:

k2 ¼ k110{logQ10=½10=ðT22T1Þ�} ð4Þ

where k2 is the desired mineralisation rate constant for the

temperature T2 (in this case the average annual temperature

at field sites), and k1 is the known mineralisation rate

constant determined at the temperature T1. We used a Q10

of 2.53 for our calculations. This is based on the analysis of

the temperature effect on decay rates of CWD (Mackensen

et al., 2003). It means that the decay rate is accelerated by

this factor with an increase in temperature of 10 8C.

Unfortunately, we could not make any corrections for

differences in water content between the incubated

samples and logs undergoing drying in re-wetting cycles

in the field.

3. Results

3.1. Pinus radiata

The density decline in logs between one and nine

years following harvesting was well described by the

decay model (Fig. 1). The decay constant of k ¼ 0:127

indicates that the time for 95% decay (t0.95) for this wood

is 24 years. At this rate 29% of the original mass are left

after 10 years. No relationship between log diameter and

density within one age group was found. The variability

in density decreased with age (Table 2). Although the

respiration rates were not constant over the chronose-

quence, they were not significantly different between the

age classes (Table 2). Multiple linear regression analysis

showed that neither age, nor wood density, moisture

content, or C-to-N ratio could explain the variation in

respiration rates. The average respiration rate across all

age classes was 0.62 mg CO2-C g (C)21 d21. When

correcting for temperature differences between the field

and the laboratory, this value was 6.1 times higher than

the decay constant derived from the density loss, when

the latter is converted to a daily decay rate. The C-to-N

ratio in the oldest wood was significantly lower than in

the youngest. The moisture content increased with age

(Table 2).

Fig. 1. Development of wood density in P. radiata logs decomposing in situ for 1, 2.5, 6, and 9 years following clear cut harvesting. Bars represent standard

deviation.

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186180

3.2. Eucalyptus regnans

The single exponential model for wood density decline

fitted the E. regnans data less well than P. radiata (Fig. 2).

Relatively little density loss occurred in logs between 6.5

and 12 years. The decomposition constant derived from the

model is k ¼ 0:041: This equals a time for 95% decay (t0.95)

of 74 years.

The laboratory respiration rates in decaying wood of E.

regnans were highly variable (Table 2), and also question

the notion of a constant decomposition rate as is assumed in

the negative exponential decay model. Respiration rates

were lowest in the youngest wood, and peaked in the 6.5

year-old material. The average respiration rate was

0.298 mg CO2-C g (C)21 d21, which was 5.9 times higher

than the decay rate determined by the negative exponential

regression model, when we corrected for temperature and

converted the annual decay rate to a daily rate. The C-to-N

ratio declined steeply over the first 6.5 years, but this decline

did not proceed to 12 years (Table 2). The respiration rates

of E. regnans wood showed a strong linear, inverse

relationship to the C-to-N ratio (Fig. 3).

3.3. Eucalyptus maculata

The approach used to derive decay rates from changes

in density was not suitable for E. maculata (Fig. 4).

Following a decline in density over the first 6.5 years,

which resulted largely from the decomposition of

sapwood, wood density increased again for 11.5 year-

old samples. Whereas in 6.5 year-old samples the

original volume was still measurable, this was not the

case for 11.5 year-old samples, which had lost all

sapwood. Thus the latter samples consisted almost

entirely of heartwood, which has a higher density than

the combined sapwood and heartwood samples at 1.5

years of age.

According to Hillis (1987) the width of sapwood in coastal

eucalypts over 15 years old remains largely unchanged.

Based on this we corrected for the missing sapwood by

determining the average sapwood width of younger samples

(2.54 ^ 0.55 cm) and calculating the ‘original’ volume of

11.5 year-old E. maculata disks. The wood density corrected

by the original log volume has been included in Fig. 4, and it

was used as the basis for the negative exponential regression.

Based on the correction for sapwood loss the estimated decay

constant was k ¼ 0:049; and the t0.95 was 62 years.

The laboratory respiration rates of E. maculata followed

a completely different pattern from that in E. regnans (Table

2). Carbon loss was highest in the youngest E. maculata

samples, supporting field observation of intensive fungal

colonisation of the sapwood in 1.5 year-old logs. The

decline of respiration from youngest to oldest samples could

not be explained statistically by the variation in wood

density, moisture content, or C-to-N ratio in E. maculata.

The variability of respiration rates within age groups

declined with age. The average respiration rate was

0.358 mg CO2-C g (C)21 d21, which is higher than in E.

regnans and 11.9 times higher than the filed decay rate

determined through density loss.

The C-to-N ratio was highest after 11.5 years. It

increased from C-to-N 410 at age 6.5 to 670 at age 11.5

years (Table 2).

4. Discussion

4.1. Log respiration rates

For all species, the laboratory respiration rates were

orders of magnitude higher than the decay rates. These

enormous differences are likely to be the result of

differences in water content. The incubated samples were

collected in winter at a very high water content, whereas

Table 2

Density (g cm23), water content (% dry weight), carbon-to-nitrogen ratio, and respiration (mg CO2-C g C21 d21) in wood samples of P. radiata, E. regnans,

and E. maculata at different stages of decomposition. Standard deviation is provided in brackets

Species Age

(yr)

Density

(g cm23)

Water content

(% dry weight)

C/N Respiration

(CO2-C g C21d21) mg

P. radiata 1 0.41(0.06) 61.1(30.5) 306.1(150.1) 0.672(0.407)

2.5 0.33(0.07) 111.0(47.1) 226.6(57.5) 0.746(0.482)

6 0.20(0.04) 173.9(50.0) 234.5(90.0) 0.405(0.174)

9 0.15(0.02) 256.1(43.7) 138.4(32.5) 0.585(0.248)

E. regnans 1 0.56(0.08) 49.6(12.5) 403.0(77.9) 0.117(0.075)

3.5 0.45(0.09) 84.0(38.5) 350.8(109.3) 0.232(0.031)

6.5 0.38(0.09) 151.7(71.6) 145.2(61.6) 0.510(0.285)

12 0.35(0.11) 183.8(96.0) 236.9(122.7) 0.331(0.208)

E. maculata 1.5 0.65(0.11) 47.3(9.3) 424.2(96.6) 0.560(0.369)

6.5 0.55(0.07) 61.6(21.2) 406.3(104.7) 0.346(0.129)

11.5 0.69(0.09) 58.8(29.9) 673.7(207.2) 0.169(0.128)

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186 181

the logs in the field undergo long dry periods. In

addition, respiration rates were based on relatively short

incubations. Thus, the C released may not only stem

from the remaining wood, but also from the turnover of

part of the microbial biomass in the wood, which might

have been accelerated by the disturbance of samples.

However, the laboratory respiration rates indicate that

there must be extended periods during which the respiration

and hence decomposition of CWD in the field is negligible.

This may explain the very slow decay process observed in

some dry forest and woodland ecosystems (Thornton et al.,

1983, 1991).

Although the laboratory respiration rates cannot be

directly converted into field decay rates, they were helpful

in establishing the decay patterns and the differences

between the different types of wood. The single component,

negative exponential decay model assumes a constant

substrate quality, and thus the decay rate is a constant

fraction of the amount of mass remaining. The laboratory

respiration rates were a useful indicator of the suitability of

the single component model. The relatively constant

respiration rates across the different age classes of decaying

logs in P. radiata confirmed the suitability of the single

component model for this species. However, the variable or

decreasing respiration rates in E. regnans and E. maculata

pointed to problems with the application of the single

component decay model.

4.2. Decay rates as determined by wood density loss

The estimation of decay rates by measuring the density of

wood samples from a chronosequence of sites is widely used.

However, the density loss pattern in E. regnans and E.

maculata depicted in Figs. 2(a) and 4(a), point to potential

Fig. 2. Development of wood density in E. regnans logs decomposing in situ for 1, 3.5, 6.5, and 12 years following clear cut harvesting. Bars represent standard

deviation.

Fig. 3. Relationship between C/N ratio and respiration rates (mg CO2 g C21 d21) in decaying CWD of E. regnans sampled at different ages following clear-

felling.

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186182

problems associated with this approach. Using the water

replacement technique to determine the volume of wood

samples takes into account the loss of density associated with

a loss of substance within otherwise solid wood. However, as

decomposition proceeds some of the original volume is also

lost e.g. through fragmentation or complete decomposition

of sapwood. Using the water replacement technique does

therefore not relate the current weight to the original volume

of the sample, but to the current volume. This may lead to an

underestimation of the density loss and consequently leads to

an underestimation of the decay rate. This is a particular

problem in samples where sapwood decays faster than

heartwood, which makes it difficult to establish the original

sample diameter once the sapwood has disappeared (see E.

maculata ). In contrast to E. regnans it was at least possible

for E. maculata samples to correct the volume for the

sapwood loss. The assumption that the small density loss in

E. regnans between years 6.5 and 12 was caused by a loss of

sapwood volume is supported by the C-to-N ratios (Fig.

2(b)). The C-to-N ratio declines steeply as decomposition

and concomitant N immobilisation increases up to age 6.5.

The subsequent increase in the C/N ratio shows that the

material at age 12 is less decomposed than that at age 6.5,

which indicates a higher proportion of heartwood. If we

follow this rational and base the determination of the

decomposition constant on the first three data points (age

class 1–6.5 years) we obtain a value of k ¼ 0:07; which

provides a time for 95% decay (t0.95) of 43 years.

A similar pattern was observed in E. maculata. In

addition, the low respiration rates in the oldest samples of

the eucalypt species indicate that the substrate quality for

micro-organisms is lower than in younger samples. This

also illustrates that both respiration rates and C-to-N ratios

can be used as a diagnostic tool to assess the validity of

density loss measurements. Where fragmentation causes the

loss of the original wood volume, this has to be

reconstructed to provide a realistic assessment of decay

rates (see also Brown et al. (1998)).

The problem caused by the loss of sapwood, including

increasing variability in density over age, did not occur for

P. radiata samples, where case-hardening preserved the

outer wood surface and thus the volume of even the older

samples (6.5 and 9 years).

The different decay rates of sapwood and heartwood in

the eucalypt species indicate that the single component

model used in the regression analysis is not applicable.

Instead a two component model simulating the decay of two

different substrates would be more appropriate (Wieder and

Lang, 1982; Means et al., 1985). However, in this study the

time span covered by the chronosequence was not long

enough to follow the decay pattern of the heartwood.

4.3. Decomposition times

The time required for 95% mass loss (t0.95) (Eq. (3)) in P.

radiata was much longer when compared to data derived

from so-called accelerated laboratory tests on wood

durability by Costa (1979). His results translate into times

for 95% decay (t0.95) of 6 years for sap- and heartwood of P.

radiata (see Mackensen and Bauhus, 1999; Mackensen

et al., 2003). However, a time for 95% decay (t0.95) of 24

years is much shorter than those found by other field studies

Fig. 4. Development of wood density in logs decomposing in situ for 1.5, 6.5, and 11.5 years following selective harvesting. Bars represent standard deviation.

The empty square at year 11.5 represents the measured density value, whereas the filled square used in the regression has been corrected for the loss of

sapwood.

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186 183

for similar species, e.g. 61 years for Pinus pinaster in West

Australia (Brown et al., 1996) and North American species

ranging from 50 to 200 years (Harmon, 1982; Foster and

Lang, 1982; Fahey, 1983; Barber and VanLear, 1984; Alban

and Pastor, 1993; Busse, 1994; Krankina and Harmon,

1995).

A time for 95% decay (t0.95) of 43–74 years for E.

regnans appears to be long when compared to results of

Accelerated Laboratory Tests for durability studies by Costa

(1979) reporting a t0.95 of only 7 years. According to

Thornton et al. (1997) E. regnans and P. radiata both

belong to durability class 4 (low durability). Durability

classes were determined from long-term field exposure of

small wood samples that were half-buried in soil. Transfer-

ring weight loss as obtained in durability studies (e.g.

Thornton et al. 1997) into time for 95% decay would

translate durability class 4 into a t0.95 of 11 years

(Mackensen and Bauhus, 1999). The much longer decay

periods found in this field study compared to durability and

laboratory studies confirm that decomposition depends on a

wide variety of factors including microclimate and decom-

poser communities.

The estimated t0.95 for E. maculata (durability class 2) of

62 years fell between the times calculated for the

comparable Eucalyptus diversicolor (durability class 2–3,

Thornton et al., 1997). Brown et al. (1996) calculated a t0.95

of 17 years for samples of Eucalyptus diversicolor including

bark from decomposition rates in a 5 year long study, while

for the 2 year long study by O’Connell (1997) calculated

t0.95 values ranged from 65 to 136 years for wood samples.

4.4. General limitations of both methods

Our sampling was concentrated on logs that were

positioned on the surface. However, it cannot be ruled out

that some of these logs may have been initially suspended

and only later fell to the ground (e.g. Næsset, 1999), and

thus will not have undergone decomposition with ground

contact for the entire period following harvesting. Although

sampling was only of stem or crown sections with clear

cutting marks, it cannot be ruled out that trees might have

been dead prior to felling or that, in the selection logging

system, the logs on the ground might have originated from

previous operations such as pole thinning.

Modification of decay through charring of logs as a

result of slash burning after logging or prescribed fires

and wildfires has not yet been determined. As charcoal is

highly resistant to decay, it is usually assumed that decay

of charred logs is reduced (see Mackensen and Bauhus

(1999)). However, our observations in the field could not

confirm this. The surface of logs is never entirely

charred, and the cracks in the charcoal allow entry of

fungal spores, if not already present before the logs were

charred. Thus we found that decomposition processes had

continued beneath a charred surface. This aspect deserves

further study.

The above considerations point to the limitations of the

chronosequence approach. They also indicate the necessity

to establish long-term experiments that allow a controlled

investigation of the different factors that may influence

CWD decomposition.

We applied a single-component, negative exponential

decay model to derive the decay rates for different woods

from density loss. However, the results for E. regnans and

E. maculata indicate that a two-component decay model,

accounting for the differences in decay resistance between

the sapwood and the heartwood, may be more appropriate.

For this study, we considered the data basis as insufficient to

apply a two-component decay model. In this case the decay

was initially faster, when sapwood was decomposing, and

then slower. However, the opposite has also been observed

(Grier, 1978; Næsset, 1999) and was interpreted as a lag

period resulting from a slow colonisation of the log with

decomposer organisms. To determine decay rates with a

greater degree of confidence, long-term studies including

logs of known volume and spanning over periods allowing

substantial decay of both sap and heartwood would be

required.

4.5. Significance of decay rates for estimating CO2

release from CWD

To determine the CO2 release from the decay of CWD

as a result of land use change and forestry activities in a

reference year, it is important to know how far back in

time estimates of the input of CWD into the decaying

pool have to go. Using the IPCC default of a 10 year

decay period of litter, the inputs would have to be

calculated for the preceding 10 years. However, the field

decay rates in this study indicate that the input of wood

into a decaying pool might have to be calculated for a

significantly longer period prior to a reference year, for

which the CO2 emission is to be established. If the input

to that pool was constant over time, the period is not

important. However, if through changes in harvesting and

land clearing activities the inputs were highly variable,

an accurate determination of the period over which input

of CWD into a decaying pool is determined will be

important.

5. Conclusions

It can be concluded that an assumed decay period of 10

years (IPCC default) is a considerable overestimation of the

rate of CWD-decomposition. Despite the variety of abiotic

and biotic parameters influencing the rate of decay, we

assume that the time for most CWD to disappear (t0.95)

exceeds 25–30 years in most cases. For many durable

species a t0.95 of more then 50 years is realistic. Instead of

assuming a linear decay of all material within 10 years, an

exponential decay function with rate constants according to

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186184

the durability class of species should be used. For species

with considerable variation in the decay resistance of sap-

and heartwood, two component decay models will be more

appropriate.

Our study has demonstrated that the combination of

different methods, including current respiration rates and

chemical characterisation of the decomposition stage (C-to-

N ratio), can assist the development of a realistic picture of

wood decay.

Acknowledgements

This study was financially supported by the Australian

Greenhouse Office. We thank Mick Morley (DNRE

Victoria), Aidan Flanagan (ACT Forests) and Ian Barnes

(State Forests of NSW) for their co-operation and assistance

with the selection of suitable sampling sites. We thank John

Kane, Sanjiva de Silva, and Scobie Mackay for their

enthusiastic help with sampling, analysis and data proces-

sing as well as Nina Stahl and Emily Backen for collecting

and managing the references. We also acknowledge the

support by Gregor Christie and Paul Smart (CSIRO,

Division of material Science and Technology), who

generously donated the foil used for incubation bags. Two

anonymous reviewers provided helpful comments on a

previous version of the manuscript.

References

Alban, D.H., Pastor, J., 1993. Decomposition of aspen, spruce, and pine

boles on two sites in Minnesota. Canadian Journal of Forest Research

23, 1744–1749.

Arthur, M.A., Tritton, L.M., Fahey, T.J., 1993. Dead bole mass and nutrient

remaining 23 years after clear-felling of a northern hardwood forest.

Canadian Journal of Forest Research 23, 1298–1305.

Barber, B.L., VanLear, D.H., 1984. Weight loss and nutrient dynamics in

decomposing woody loblolly pine logging slash. Soil Science Society

of America Journal 48, 906–910.

Brown, S., Mo, J., McPherson, J.K., Bell, D.T., 1996. Decomposition of

woody debris in Western Australian forests. Canadian Journal of Forest

Research 26, 954–966.

Brown, P.M., Shepperd, W.D., Mata, S.A., McClain, D.L., 1998. Longevity

of windthrown logs in a subalpine forest of central Colorado. Canadian

Journal of Forest Research 28, 932–936.

Burger, J., Pritchett, W., 1984. Effects of clearfelling and site preparation

on nitrogen mineralisation in a Southern pine stand. Soil Science

Society of America Journal 48, 1432–1437.

Busse, M.D., 1994. Downed bole-wood decomposition in lodgepole pine

forests of Central Oregon. Soil Science Society of America Journal 58

(1), 221–227.

Christensen, O., 1984. The states of decay of woody litter by relative

density. Oikos 42, 211–219.

Costa, E.W.B.da, 1979. Comparative decay resistance of Australian

timbers in accelerated laboratory tests. Australian Forest Research 9,

119–135.

Cowling, E.B., 1972. Nitrogen in forest trees and its role in wood

deterioration. Abstracts of Uppsala Dissertations in Science.

Edwards, N.T., 1982. The use of soda-lime for measuring respiration rates

in terrestrial systems. Pedobiologia 23, 321–330.

Fahey, T.J., 1983. Nutrient dynamics of aboveground detritus in lodgepole

pine (Pinus contorta ssp. latifolia) ecosystems, south-eastern Wyoming.

Ecological Monographs 53, 51–72.

Foster, J.R., Lang, G.E., 1982. Decomposition of red spruce and balsam fir

boles in the White Mountains of New Hamshire. Canadian Journal of

Forest Research 12, 617–626.

Graham, R.L., Cromack, K.C., 1982. Mass, nutrient content, and decay rate

of dead boles in rain forests of Olympic National Park. Canadian

Journal of Forest Research 12, 511–521.

Grier, C.C., 1978. A Tsuga heterophylla–Picea sitchensis ecosystem of

coastal Oregon: decomposition and nutrient balances of fallen logs.

Canadian Journal of Forest Research 8, 198–206.

Harmon, M.E., 1982. Decomposition of standing dead trees in the southern

Appalachian Mountains. Oecologia 52, 214–215.

Harmon, M.E., Franklin, J.F., Swanson, F.J., Sollins, P., Gregory, S.V.,

Lattin, J.D., Anderson, N.H., Cline, S.P., Aumen, N.G., Sedell, J.R.,

Lienkaemper, G.W., Cromack, J.R., Cummins, K.W., 1986. Ecology of

coarse woody debris in temperate ecosystems. Advances in Ecological

Research 15, 133–302.

Healey, I.N., Swift, M.J., 1971. Aspects of the accumulation and

decomposition of wood in the litter layer of a coppiced beech-oak

woodland. Annual Zoological Ecology Anim./hors serie, 417–430.

Hillis, W.E., 1987. Heartwood and Tree Exudates, Springer Series in Wood

Science, Springer, Berlin, 287 pp.

IPCC, 1997. Revised 1996 IPCC Guidelines for National Greenhouse Gas

Inventories, Reference Manual, Intergovernmental Panel on Climate

Change, vol. 3. Meteorological Office, Bracknell, UK.

Jenny, H., Gessel, S.P., Bingham, F.T., 1949. Comparative study of

decomposition of organic matter in temperate and tropical regions. Soil

Science 68, 419–432.

Krankina, O.N., Harmon, M.E., 1995. Dynamics of the dead wood carbon

pool in north-western Russian boreal forests. Water, Air and Soil

Pollution 82, 227–238.

Lambert, R.L., Lang, G.E., Reiners, W.A., 1980. Loss of mass and chemical

change in decaying boles of a subalpine balsam fir forest. Ecology 61,

1460–1473.

Mackensen, J., Bauhus, J., 1999. The decay of coarse woody debris.

National Carbon Accounting System, Technical Report No. 6,

Australian Greenhouse Office, Commonwealth of Australia, Canberra,

40 pp.

Mackensen, J., Bauhus, J., Webber, E., 2003. Decomposition rates of coarse

woody debris—a review with particular emphasis on Australian tree

species. Australian Journal of Bot. in press.

Means, J.E., Cromack, K.J., MacMillan, P.C., 1985. Comparison of

decomposition models using wood density of Douglas-fir logs.

Canadian Journal of Forest Research 15, 1092–1098.

Næsset, E., 1999. Decomposition rate constants of Picea abies logs in

southeastern Norway. Canadian Journal Forest Research 29,

372–381.

NGGIC (National Greenhouse Gas Inventory Committee), 1997a. Land

Use Change and Forestry. Workbook for Carbon Dioxide from the

Biosphere. Workbook 4.2, Revision 2.

NGGIC (National Greenhouse Gas Inventory Committee), 1997b. Land

Use Change and Forestry Sector 1988–1995, based on Workbook 4.2.

O’Connell, A.M., 1997. Decomposition of slash residues in thinned

regrowth eucalypt forest in Western Australia. Journal of Applied

Ecology 34, 111–122.

Olson, J.S., 1963. Energy storage and the balance of producers and

decomposers in ecological systems. Ecology 44 (2), 322–331.

SAS Institute, 1996. JMPw Start Statistics, Duxbury Press, Belmont.

Stone, J.N., MacKinnon, A., Parminter, J.V., Lertzman, K.P., 1998.

Coarse woody debris decomposition documented over 65 years on

southern Vancouver Island. Canadian Journal of Forest Research 28,

788–793.

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186 185

Swift, M.J., 1977. The ecology of wood decomposition. Science

Programme 64, 175–199.

Swift, M.J., Heal, O.W., Anderson, J.M., 1979. Decomposition in

Terrestrial Ecosystems, University of California Press, Berkeley, pp.

354, 354 pp.

Thornton, J.D., Walters, N.E.M., Saunders, I.W., 1983. An in-ground

natural durability field test of Australian timbers and exotic reference

species I. Progress report after more than 10 years exposure. Material

und Organismen 18, 27–49.

Thornton, J.D., Johnson, G.C., Nguyen, N., 1991. An in-ground natural

durability field test of Autralian timbers and exotic reference species.

Material und Organismen 26, 1–17.

Thornton, J.D., Johnson, G.C., Nguyen, N., 1997. Revised CSIRO Natural

Durability Classification—In-Groud Durability Ratings for Mature

Outer heartwood, CSIRO Forestry and Forest Products, Melbourne, 2 pp.

Wieder, R.K., Lang, G.E., 1982. A critique of the analytical methods used

in examining decomposition data obtained from litter bags. Ecology 63,

1636–1642.

J. Mackensen, J. Bauhus / Soil Biology & Biochemistry 35 (2003) 177–186186