Back to Hot Topics
|
TECHNICAL REPORT
GREENHOUSE ALLIES PROJECT
Measurement of carbon sequestration
in small non-industrial forest plantations.
(continued)
4 In situ sampling of plantation trees
The methods used for sampling in the field were established at the workshop in May
1999 and described in the Scope Report. In summary, a sub-sampling methodology
was used to assess the standing volume of trees on the plantation, and either the
conversion figures used in the VSW or allometric equations were used to calculate
the standing biomass and thence standing carbon on the property. As one of the key
objectives of the project was to assess the VSW and to establish the optimal accuracy
of measurement, the approach taken was to initially assume the VSW protocol, with
some modifications, as a ‘given’ and the accuracy of the estimate is a product of the
sampling methodology.
4.1 Field data records and modified VSW protocol
The plantings on each property were overviewed, the total area planted estimated,
and some plantings were separated into subsets called strata. These were reflective
of differences in site quality, in species type, or planting age across the site which
may have affected the growth of the trees and hence the consistency of the sampling
effort. The area of each stratum thus identified was estimated and within each
several circular plots were randomly located. The radius of these plots was generally
consistent across the site, and certainly within a stratum. Within each plot the
diameter at breast height (dbh) of all the trees was measured and a sub-sample of
those trees was selected for full measurement (height). This method is one
commonly used in forestry practice in monocultures, and has been found to given an
effective and accurate estimate of the standing volume of timber and its variability.
The success of this field sampling method is dependent on good replication of fully
measured individuals within the plots in order to calculate the expected heights (and
hence volumes) for those individuals not measured in full. In mixed species plantings
it became apparent that several species were undersampled in the plot sizes chosen,
and prediction was not possible from those data alone. A search was undertaken to
obtain predictive relationships from the Subtropical Tree Site Management Database
(Specht and Digby 1998), which holds numerous field measurements of cabinet
species (the collection of which was variously funded by the Rural Industries
Research and Development Corporation and Save the Bush), recently supplemented
by eucalypt species (in work funded by the National Heritage Trust), all on farms
and in plantings similar to those being measured in this project. These data were
amalgamated with those obtained in this project to get the largest representation of
individuals of all the species possible, with no separation according to site. The
relationships obtained (with replicates up to 500 individuals for some species, but
down to as little as 10 for poorly represented individuals) were used to calculate
heights from the diameters measured in the field. Equations for a total of fifty two
species were obtained: 34 rainforest species, 16 eucalypts and for brush box and
radiata pine. Some species could not be replicated and default equations were used
for those few.
Once the diameters and heights are estimated for the trees within each plot, a
volume of the stem can be calculated using a standard (or, if available, a specific)
volume equation. A general non-specific equation was adopted for this project. This
equation, based on Phillips (1994), had been used by the author for cabinet timber
species in a Rural Industries Research and Development Corporation project and
also for Eucalyptus grandis in the Soil Tree and Grass (STAG) project of C.S.I.R.O.
Division of Forest Products.
The volume thus calculated was converted to wood density by multiplying the oven
dry density figures per cubic metre obtained for each of the species at 1.3 m height in
the plantations (Figure 6). Again, defaults were used for species that were not
sampled.
This stem biomass was then converted to total above-ground biomass using the
harvest index (0.68 for Pinus radiata and 0.70 for eucalypts). This is then followed by a
root:shoot ratio of 0.2, and finally the carbon conversion of 0.5. These equations and
conversions were written into a syntax file in the Statistical Package for the Social
Sciences (SPSS©) and this was used to calculate plot and stratum figures of dbh,
height, above ground biomass, total biomass and total carbon for each property.
4.2 Confidence of the estimate
The confidence of the estimate of carbon on a property is critical for the landholder if
they are to participate in carbon trading or gain other benefits that may be
associated with sequestering carbon. At this stage in the development of the
equations and conversion figures for calculating the various items on the path from
diameter to carbon, it shall be assumed that these are error-neutral: that is, they be
assumed to be the best protocol available. The major source of error, therefore, is in
the estimation of the carbon on a property using sampling methods such as those
described above. In other words, it is assumed that the total stand will not be
measured, therefore the accuracy and precision of the sub-sampling method adopted
will need to be determined.
The first step in the assessment of error (the inverse of the confidence) in the
measurement of standing biomass and carbon at a property is the assessment of the
area planted with the trees to be measured. This may sound simple, and is not
incorporated directly into the estimation method that follows, but is critical and often
not measured by the landholder. This will exclude those areas which have failed, thus
not always the entire area initially planted.
The second step is identifying any internal variability in a planting in order to isolate
any such variability within a calculation unit. The area of this unit (a stratum) needs
to be estimated in order to correct the overall variability on the property according
to the importance of the contributing parts. In a stylised depiction of a plantation
area broken up into three separate sub-units, the variability found in each sub-unit
will be weighted according its contribution to the overall planting (Figure 11). If the
variability of the trees in stratum 3, for example, is high, it would have a lesser effect
on the overall figure for the property than if one of the larger strata were similarly
variable.
Figure 11: A stylised property showing three different strata with plots randomly located within
each. The total area of the planting estimated in the first carbon measure should exclude the area of
dead trees. Each stratum will have an error based on the variation both within and between the
plots and the effectiveness of the plots (plot size and number).
In order to correct for the contributions of the trees, the plots and the strata to the
overall property carbon content and the variability of that estimate a formula for
estimating overall variance was used from Schreuder et al. (1993), p. 48, which takes
these into account. A sample calculation of this can be seen in
Appendix A.
4.3 Results
The data collected in the field using this method allows the calculation of field values
of biomass and carbon as well as the confidence of the estimate. This is shown for
fifteen properties in Table 3. Several properties are not included here: Johnston,
Coward and SF NSW Walcha are not included because of uncertain stratum area
estimates, Machin is not included because of unknown planting age at the time of
calculation, Jones, Dickson and Thomas are not included because the trees were too
young, and Greening Australia is not included because we require more speciesspecific
information from their plots to proceed.
The amount of carbon held in the biomass in these plantations, despite their
sometimes small area, can be considerable. The confidence of the estimate is,
however, more variable. There is little confidence, for example, in the estimate for
the very small trees at the Chandler property. This might be expected to improve
with age as there is a small variability of 10% at Envirocom and 17% at the Moody
property at five years of age. The highest confidence intervals (the largest error of
the estimate) are for very young plantings (to 2 years at least) and for some of the
cabinet timber plantings. The younger the trees, the higher the variability between
the plots. The methods adopted by most of the cabinet timber planters, Harvey-
Jones, for instance, purposely include fast-growing, low wood density species such as
Elaeocarpus grandis (blue fig/silver quandong) in order to shelter the slower, denser
species such as Heritiera trifoliolatum (white booyong). These two types are often
planted in alternating rows or pairs of rows, resulting in patchiness and in this case,
inadequacy of the plot sample size to reliably incorporate the variability and get a
precise estimate.
Table 3: The standing carbon estimated using the modified VSW protocol for six properties in
northern NSW together with a 90% confidence interval for the estimate.
There is considerable argument about the role of mixed-species plantings in carbon
sequestration, in particular the eco-forestry models of plantation growth. These,
despite the difficulties of establishing appropriate and feasible initial estimation
techniques, are designed never to be completely felled, resulting in potentially good
carbon sequestration in contrast to more simply measured, but more dramatically
cycling models.
A plot of the standing carbon per hectare per site against age (Figure 12) shows a
positive relationship of carbon with age, which is confirmatory that the calculations
are realistic. The high values which occur in the Northern Rivers (Envirocom and
Fayle) are clearly outliers to the general trend. As the documentation for Envirocom
(stratum and total areas) were very reliable and precise, this high value is assumed to
be a consequence of site or management factors.
Figure 12: The relationship between standing carbon per hectare and the average age of the planting
for the Mid-north Coast and the New England sites, with an overlay of the Northern Rivers sites.
The amount of carbon fixed per year for the measured properties gives us a good
idea of the potential of the plantings in greenhouse gas reduction. An estimate of the
amount of carbon stored in the year immediately prior to the measurement of the
trees in 1999 or early 2000 gives a good idea of the possible rate of acquisition for
these young plantings both now and in the future. This last year is taken, as, when
the plants are young (for example 0 to 4 years old), there is little actual growth while
the plants become established, so an estimate incorporating the growth in those
years as an indication of future potential would be misleading. An estimate of the last
year’s growth is obtained using the exponential growth pattern of all plant (and
indeed animal) life (Waring and Phillips 1970, Charles-Edwards et al. 1986):
This function describes the growth of the plant from its early establishment in the
paddock (the lag phase) to the exponential phase where it grows most vigorously.
This relationship is not likely to be suitable for older trees undergoing linear growth,
or for those in the maturation phase, when the plant begins to reach its final size,
which might be expected in many of these plantings start at a minimum of 30 years
of age. Using this equation, the amount of carbon acquired during 1999 shows a
marked effect of age, but also an effect of, presumably, site quality (Table 4). Sites
like Envirocom, showing a carbon acquisition rate of 13.95 tonnes ha-1 of carbon
between 4 and 5 years of age, may be compared with 7.85 t ha-1 acquired at the
Moody property by trees of the same age, commensurate with the much lower
rainfall at the Moody property. Internal stratum effects are not shown here, but
considerable variation between strata on the various properties was evident,
particularly in the case of Fayle, where one stratum, in a gully, had a very high
growth rate in comparison to other strata. These factors may well prove most
valuable in future planning for plantation establishment and management, once
error affects like accurate stratum areas are corrected.
Table 4: The carbon acquisition rate for the last year of growth for each of the measured properties
using a fitted exponential growth model with time since establishment. Note that the rate per year
is highly dependent on an accurate stratum area estimate.
The average amount of carbon fixed per hectare in the last year of growth for
plantings greater than 5 years of age is 12.91 tonnes ha-1. The data for younger age
strata are not included, as their growth rate is not assumed to be representative of
the potential growth rate due to their young age. Using a conversion figure of 3.67
this is equivalent to a CO2 fixation rate of 47.37 tonnes ha-1 yr-1.
Go to the TOP of the Page
Go to the Table of Contents
Go to the NEXT Section
|
