Printer Friendly
The Free Library
22,710,190 articles and books

Estimating the accuracy of tree ageing methods in mature Eucalyptus obliqua forest, Tasmania.



Introduction

Tree hollows provide important habitat for fauna (Gibbons and Lindenmayer 2002) and older trees are more likely to contain hollows (Mackowski 1987; Whitford 2002). The time required for a hollow suitable for faunal use to form (150 y: Mackowski 1987; 165 y: Wormington and Lamb 1999; 130 y: Whitford and Williams 2002) is substantially longer than the 60-120 y generally planned between harvesting operations in production forests (Whiteley 1999; Department of Natural Resources Many sub-national governments have a Department of Natural Resources or similarly-named organization:
Australia
  • Queensland Department of Natural Resources and Mines
Canada
  • Natural Resources Canada
 and Environment 2002a,b). Management of the hollow resource in production forest areas therefore requires an understanding of the age at which trees in different areas produce hollows suitable for use by animals.

Trees can be aged according to according to
prep.
1. As stated or indicated by; on the authority of: according to historians.

2. In keeping with: according to instructions.

3.
 the disturbance history of a site (Bradshaw and Rayner 1997b), by radiocarbon dating (Turner 1984), by tree ring counting (Banks 1997) or by using tree ring counts or tree diameter increment To add a number to another number. Incrementing a counter means adding 1 to its current value.  data to produce growth models (Lloyd and Lau 1986; Wormington and Lamb 1999; Gibbons et al. 2000; Moloney et al. 2002). The accuracy of tree age estimates can potentially influence the effectiveness of management prescriptions. For example, if there are large error margins and the average values are adopted by managers, hollows may be thought to occur in trees retained in harvested areas when they do not. It is therefore important to select an appropriate method for estimating tree age and to assess the accuracy of the technique used.

Ring counting relies on the assumption that seasonal variation in growth conditions affects the density or size of the cells accumulated. Darker latewood bands can form when cold temperatures or moisture deficit cause seasonal periods of slower growth (Pilcher and Gray 1982; Leal LEAL. Loyal; that which belongs to the law.  et al. 2004), but can also result from defoliation by grazers and fire (Mazanec 1968; Mucha 1979). Consequently, their production is not always strictly annual, with some rings being locally absent, or 'false' rings being produced. The rate of false ring production can vary with the dominance status of the tree (Mucha 1979; Brookhouse 1997; Masiokas and Villaba 2004; Bar et al. 2006). Locally absent or 'missing' rings occur most frequently on trees that are old, suppressed, have poor crown development and in areas where environmental stresses are high (Brookhouse 1997; Lorimer et al. 1999; Jonsson et al. 2002; Bar et al. 2006; Waring and O'Hara 2006; Mayfield et al. 2007; but see Rozas 2003). The annualarity of tree rings is usually assessed by cross-dating patterns of tree ring width, distinct rings or markings either between trees or with known disturbance events (Yamaguchi 1991; Rayner 1992; Brookhouse 1997; Brookhouse and Brack n. 1. An opening caused by the parting of any solid body; a crack or breach; a flaw.
Stain or brack in her sweet reputation.
- J. Fletcher.

1. Salt or brackish water.
 2006).

In eucalypts, the latewood bands often lack clear definition (Brookhouse 1997) and older trees frequently have a rotten or hollow middle. However, annual patterns in cell structure and therefore growth rings have been observed in a range of areas for a number of species (Mucha 1979; Akeroyd et al. 2002; Leal et al. 2004; Brookhouse and Brack 2006). The number of rings in the hollow middle is usually estimated using the size of the rings found in the solid wood (Rose 1993; Woodgate et al. 1994; Alcorn et al. 2001). Tasmanian old-growth E. obliqua with hollow centres were aged in this way and estimated to have an error of about 15%, although how this figure was obtained is not specified (Alcorn et al. 2001). Woodgate et al. (1994) used two mature trees with a solid centre to estimate the upper and lower age limits of the hollow middle in three other trees at the same site. The error margins ranged between 13 and 17 y for trees estimated to be 137-237 y old. Ageing eucalypts can also be complicated by swelling in the base of some old trees, or 'butt swell'. This swelling may occur asymmetrically to form 'buttresses'. Expansion of the growth rings can occur at the buttresses, with compression of the rings in the areas between (Koch, pers. obs.). Trees exhibiting butt swell will have a disproportionately large diameter at breast height Diameter at breast height, or DBH, is a standard method of expressing the diameter of the trunk of a tree.

The trunk is measured at the height of an adult's breast; this is defined differently in different situations, with foresters measuring the diameter at 1.
 (dbh) than trees of similar age without this feature, potentially biasing age estimates based on dbh. Estimating the age of the tree at a point above the swelling is one solution to this issue (Walshe 2001).

Growth models are a widely used tree-ageing technique. Known tree ages, growth rates Growth Rates

The compounded annualized rate of growth of a company's revenues, earnings, dividends, or other figures.

Notes:
Remember, historically high growth rates don't always mean a high rate of growth looking into the future.
 or tree increment widths are used to make an estimate of tree age based on dbh (Bowman and Kirkpatrick 1984; Bradshaw and Rayner 1997a; Gibbons et al. 2000). Growth rates can vary with a number of factors, including tree age, dominance and species, environmental factors, site quality, tree density, management practices and insect attack (Mucha 1979; Rose 1993; Cherubini et al. 1998; Wormington and Lamb 1999; Leal et al. 2004; Bar et al. 2006; Mayfield et al. 2007). Therefore, while linear models can be used (e.g. Bradshaw and Rayner 1997a; Gibbons et al. 2000), non-linear models may be more appropriate in some areas (Abbott and Loneragan 1983, Walshe 2001). Caution is often advised when using growth models (Abbott and Loneragan 1983; Rose 1993), but few studies examine the error associated with their age estimates. Some papers using ring-width data to produce growth models indicate the error of these widths, but do not indicate the error of the resulting age estimate (Abbott and Loneragan 1983; Rose 1993). Lindenmayer et al. (1999) checked the accuracy of their age estimates against independent fire mapping, but did not report an error.

The aims of this paper are to develop an appropriate growth model for estimating the age of Eucalyptus obliqua Eucalyptus obliqua, commonly known as Australian Oak, Brown Top, Brown Top Stringbark, Messmate, Messmate Stringybark, Stringybark and Tasmanian Oak,[1] is an hardwood tree native to south-eastern Australia.  L'Herit. trees in Tasmania and to develop methods for approximating the error of age estimates when using either ring counting and extrapolation (mathematics, algorithm) extrapolation - A mathematical procedure which estimates values of a function for certain desired inputs given values for known inputs.

If the desired input is outside the range of the known values this is called extrapolation, if it is inside then
 or a growth model. Error estimates are measured using several techniques, but all are a measure or indication of the accuracy of the tree ageing process.

Methods

Site selection and data collection

Thirty-eight forestry coupes in mature wet or dry E. obliqua forest were examined between January 2004 and May 2005 (18 dry sites, 8 damp sites and 12 wet sites; Fig. 1, Table 1). Sites were selected on the basis of an impending im·pend  
intr.v. im·pend·ed, im·pend·ing, im·pends
1. To be about to occur: Her retirement is impending.

2.
 logging operation and included previously harvested and unharvested areas. Site altitude ranged from 60 to 560 m above sea level. A wide range of aspects, topographies and soil types were included. The specific location examined at each site was determined by the forest contractor according to logistical constraints. The first tree encountered of a suitable size (at least 50 cm dbh over bark) was examined. The nearest neighbour with a dbh of at least 50 cm that was at least 20 m from all other trees was then selected until 11-13 trees had been examined. Tree and site variables were obtained on site and remotely (Table 1). The trees were felled as part of normal forestry operations and a slice of wood was cut from the stump of 329 trees (between four and twelve trees per site). The average stump height was 77 cm ([+ or -] 30 s.d.). Where possible, wood samples were cut from an area of the stump with clear rings and solid wood that was as close as possible to the centre. All samples were planed and sanded as required (using up to 1000 grit sandpaper sandpaper, abrasive originally made by gluing grains of sand to heavy paper sheets. Today sandpaper is made primarily with quartz, aluminum oxide, or silicon carbide grains, and is graded according to the size of the grains. ) for clear identification of tree rings. Tree rings were identified by the denser and darker coloured 'reverse latewood' bands that have been shown to be roughly annual in E. obliqua in a similar climate in Victoria (Brookhouse and Brack 2006). The number of tree rings was counted on one radius for each sample. Additional counts were prohibited by the small size of the wood sample generally obtained. Terminology used in this paper and specific to it is defined in Table 2.

[FIGURE 1 OMITTED]

Method for estimating tree age

For each wood sample, the number of rings counted, the length of the sample and an estimate of the distance missing on that radius to the central ring of the tree was recorded. Often the middle of the tree could be seen when cutting the sample, but it was not possible to include it because of rot or because the distance was greater than the bar of the chainsaw. In such cases the missing distance of the radius could be measured accurately. When trees were hollow in the middle, the missing distance was estimated as best as possible based on the curvature of the rings that were visible. The 70 samples that contained the tree centre will be referred to as 'complete samples'. The other 259 samples with a section missing will be referred to as 'incomplete samples'. The distance between every ten rings was measured from the inside out to determine ring width or 'decadal increments' (Table 2). For each tree, the length of the radius was plotted against the number of 'decadal increments' and a regression line Noun 1. regression line - a smooth curve fitted to the set of paired data in regression analysis; for linear regression the curve is a straight line
regression curve
 was fitted that passed through the origin. The slope of this relationship, from now referred to as the 'Increment Formula', ranged between 0.09 and 1.05. For incomplete samples, the increment formula was used to estimate the number of rings likely to have occurred in the missing section of wood. Tree ages estimated using this technique are referred to as ring count and extrapolation ages (RCAE RCAE Riverina College of Advanced Education (Australia)
RCAE Reliability Computer Aided Engineering
 ages) (Table 2). No account was taken of the time taken for the tree to reach the height at which the wood slice was cut or of the difference in age between stump height and breast height. Alcorn et al. (2001) estimated that E. obliqua trees grow to 0.3-0.5 m by age 1 y in wet forest in south-western Tasmania. We deemed this error to be sufficiently small sufficiently small - suitably small  for the purpose of this study to be disregarded.

We considered several sources of error when estimating the age of trees (Fig. 2). When a wood sample is available there is error from tree ring counting and from extrapolating the number of rings on the missing centre of a tree. When no wood sample is available for a tree, a growth model may be constructed from trees at the same site or from trees at another similar site. The accuracy of the age estimate will depend on the location of the trees from which the growth model is constructed.

Error in ring counts

Although tree rings are assumed to be annual in this study, false rings, missing rings or observer error can affect the accuracy of ring counts. Greater error is expected for workers unfamiliar with eucalypt growth rings (Mucha 1979). To assess the accuracy of the ring counter in this paper, the rings of 16 E. obliqua samples were counted by Koch and two other experienced counters. There was an average difference of 7% (4.5 s.d., maximum 17.8, minimum 1) between the number of rings counted by Koch and the first experienced ring counter and an average of 16% (16.2 s.d., maximum 55.6, minimum 1.2) with the second. Ring counts made by Koch were consistently higher than that of the first experienced ring counter but were evenly distributed above and below the counts of the second experienced ring counter (Fig. 3). However, the difference between the two experienced ring counters was an average of 16% (maximum 60), suggesting the senior author of this paper was as accurate as other experienced counters. Differences in counts did not appear to be related to tree age and are more likely to have been influenced by ring clarity (Fig. 3).

Ring count error can be estimated by comparing the estimated age with known disturbance events. It was not possible to accurately cross-date the tree ages in this study (see Appendix 1). We therefore adopt the error rates estimated by Brookhouse (1997), who used the same tree species growing in similar climatic conditions on mainland Australia. Brookhouse (1997) found that ring counts over-estimated tree age by 3% in co-dominant, 7% in dominant and 8% in suppressed trees.

Error in extrapolating

For incomplete wood samples, the number of rings on the missing distance of wood was estimated by fitting a straight line to the growth increment data and extrapolating it back to the origin. Extrapolation using a quadratic function A quadratic function, in mathematics, is a polynomial function of the form , where .  was also examined but the difference in R2 value between the two fits was on average 0.0129 (s.d. 0.0189, maximum 0.124) and was not significantly related to any tree variable examined. However, the increment formulas were significantly associated with tree age and shape (P < 0.05), suggesting a deceleration deceleration /de·cel·er·a·tion/ (de-sel?er-a´shun) decrease in rate or speed.

early deceleration
 in dbh growth rates with increasing age and senescence senescence /se·nes·cence/ (se-nes´ens) the process of growing old, especially the condition resulting from the transitions and accumulations of the deleterious aging processes.

se·nes·cence
n.
. Some trees showed accelerating growth rates, some decelerating and others a combination of the two, which made the uniform application of a decelerating function inappropriate. Visual examination of the growth rates revealed no consistent trend within sites. In the vast majority of cases a straight line was an adequate (and often more appropriate) fit for the data than a quadratic quadratic, mathematical expression of the second degree in one or more unknowns (see polynomial). The general quadratic in one unknown has the form ax2+bx+c, where a, b, and c are constants and x is the variable.  curve. No significant association was found between the increment formulas and tree dominance (P > 0.05).

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

The data collected for the present study could not be used to establish the error of extrapolation because the complete samples collected were not from older trees. Consequently, data collected from another source in dry forest in eastern Tasmania was used (von Platen A long, thin cylinder in a typewriter or printer that guides the paper through it and serves as a backstop for the printing mechanism to bang into. It is typically made of a hard rubber or rubber-like material. See carriage and typewriter. , unpublished). The sites were similar to dry forest sites we examined in eastern Tasmania. The relevance this information has to wet forest is uncertain, but no similar information was available for wetter forests. Six slabs were cut from large, complete E. obliqua tree stumps (Table 3). A high-resolution photograph was taken of each sanded slab and a computer program used to identify growth rings, which were verified either by Koch or von Platen. Von Platen (University of Tasmania (body, education) University of Tasmania -

ftp://ftp.utas.edu.au/.
, pers. comm., 2008) examined these slabs in more detail and found that ring counting resulted in an error of plus or minus six rings over an entire radius which was not cumulative. We recorded the age of each tree (number of growth rings) and the distance between every ten rings. We then 'removed' ten rings at a time and the number of rings in the 'absent' area was estimated using linear extrapolation, as outlined above, to give a series of new RCAE ages. Linear regression Linear regression

A statistical technique for fitting a straight line to a set of data points.
 analysis was done using SPSS A statistical package from SPSS, Inc., Chicago (www.spss.com) that runs on PCs, most mainframes and minis and is used extensively in marketing research. It provides over 50 statistical processes, including regression analysis, correlation and analysis of variance.  (2005). The dependent variable was the absolute percentage difference between the original ring counted age and the new RCAE age. The independent variable was the percentage of sample removed. The resulting regression model will be referred to as the extrapolation error model (Table 2). This model predicts the percentage error in age estimation associated with extrapolation according to the percentage of the wood sample that was missing. We multiplied the RCAE age by the percentage error predicted to obtain an estimate of extrapolation error in years for each tree we studied. The extrapolation error was added to the ring count error (see above) to estimate total error for the RCAE Method.

Growth models and associated error

Modelling the influence of site and tree variables on growth rates

To identify the most important variables influencing the growth rate of individual trees, Bayesian linear modelling with uninformative un·in·for·ma·tive  
adj.
Providing little or no information; not informative.



unin·for
 priors was done in WinBUGS 1.4 (Spiegelhalter et al. 2003). The 25 trees with an RCAE age greater than 400 y were not included as the ages of these trees may have been erroneous (see discussion), leaving a sample size of 304 trees. The dependent variable was the RCAE age of the tree, square-root transformed. The independent variables considered are outlined in Table 1. All continuous variables were standardised to reduce autocorrelation Autocorrelation

The correlation of a variable with itself over successive time intervals. Sometimes called serial correlation.
 between successive samples (i.e. the mean was subtracted from the data, which was then divided by the standard deviation). Variables were entered into the model in a forward stepwise stepwise

incremental; additional information is added at each step.


stepwise multiple regression
used when a large number of possible explanatory variables are available and there is difficulty interpreting the partial regression
 manner. The initial 1000 samples were discarded as a 'burn in' and the following 10 000 samples were used to calculate the Deviance Information Criterion There are a number of statistics that can act as an information criterion. They include:
  • Akaike's information criterion
  • the Bayesian information criterion, also known as the Schwarz information criterion
  • Hannan-Quinn information criterion
 (DIC DIC diffuse intravascular coagulation; disseminated intravascular coagulation.

DIC
abbr.
disseminated intravascular coagulation


Disseminated intravascular coagulation (DIC) 
). A difference in DIC value of less than two indicates a lack of difference in the models, while a difference of three or more indicates that the model with the smaller DIC value is increasingly superior (McCarthy 2007). Bayesian methods were used because they allow complex correlation structures (such as hierarchical models) to be easily modelled. The variable 'site number' was added as a random factor to determine the unexplained variation between sites. For the final model (excluding site number), 100 000 samples were used to calculate the mean, standard deviation, 2.5th and 97.5th percentiles of the coefficients. The percentiles represent a 95% Bayesian credible interval In Bayesian statistics, a credible interval is a posterior probability interval, used for purposes similar to those of confidence intervals in frequentist statistics.

For example, a statement such as
. The residuals (observed value minus predicted value) were examined to assess the fit of the model. The results of this model were used to produce the 'predictive model', which predicts the age of a tree based on site and tree characteristics.

Tree growth rates vary between sites of different quality (Abbott and Loneragan 1983) as does the floristic composition and, therefore, the forest type category assigned to an area (Ashton 1981). To determine if a difference in growth rate was observed between the broad forest types considered here, ANOVA anova

see analysis of variance.

ANOVA Analysis of variance, see there
 analysis was done using R (R Development Core Team 2006) between the slope of the increment formula and forest type. The increment formula was calculated using trees only 400 y old or less. One site was removed due to poor correspondence, leaving 18 dry sites, 8 damp and 11 wet sites in the analysis.

Examining alternative growth models

The predictive model is one way of estimating tree age. Many growth models, however, use only dbh to predict tree age (e.g. Bradshaw and Rayner 1997a; Gibbons et al. 2000). Consequently, Bayesian regression analysis In statistics, a mathematical method of modeling the relationships among three or more variables. It is used to predict the value of one variable given the values of the others. For example, a model might estimate sales based on age and gender.  was done between tree age and dbh for all trees with an RCAE age of 400 y or less, to produce the 'overall formula', and for the trees within a site, to produce a series of 'site formulas' (Table 2). The three growth models (predictive, overall and site models) were used to predict the age of all trees. The accuracy of the models was examined in three ways. Firstly the percentage of trees where the RCAE age fell within one standard deviation of the predicted age was determined. Then the RCAE age was regressed against the predicted age for each model and the slope coefficient and adjusted R2 values of the relationship were examined. Finally the error of the predictions was inspected by visually examining the relationship between the standard deviation and predicted age of each tree. The same data were used to produce and test the growth models to maximise the data available for model construction. Ideally, a new data set should have been used to verify the accuracy of these models.

Results

Ring counts and linear extrapolation

When extrapolation methods were used to account for a 'missing' length of wood, the error of the age estimate increased with the percentage of wood sample removed. The age estimated from the modified wood sample was not consistently greater than or less than the age as determined by ring counting alone until about 70% of the sample was removed. When more than 70% of the sample was removed there appeared to be greater inaccuracy in the age estimate for most trees, with ages being overestimated more than underestimated. The relationship between the absolute percentage change in age estimate and the percentage of sample missing was roughly linear (Fig. 4). The model showed a reasonable fit (average difference between predicted and ring counted age was 3.15 y), but the 95% confidence intervals showed a lack of precision in the results.

The error of the RCAE method was determined for each tree and was, on average, about 10% of the estimated tree age (Fig. 5). The maximum error rate for the trees examined was [+ or -] 80 y but for most trees (88%) the error was [+ or -] 30 y or less.

Growth models and associated error

Modelling the influence of site and tree variables on growth rates

When using site and tree variables, the best model predicting tree age included tree dbh, soil phosphorus, average temperature and stand age classification (Table 4). The stand age categories 'regrowth' and 'regrowth with mature' were combined into one category which will be referred to as simply 'regrowth'. The negative relationships with the variables other than dbh mean that trees have faster growth rates and so are younger for the same size. If the random factor 'site' was included, there was a decrease in DIC value of 36 units, indicating there was unexplained variation between sites. However, as the model is to be used for predictive purposes, the coefficients are indicated for a model excluding the random factor. The WinBUGS code for this model is provided in Appendix 2 or can be obtained from the authors. Examination of the residuals indicated that trees younger than 120 y were likely to be over-estimated and trees older than 250 y to be underestimated in age when using the predictive model. The average absolute residual was 45.3 [+ or -] 35.4 s.d., indicating low levels of accuracy.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

After removal of the site with extremely poor fit, there was a significant difference in growth rate between the different forest types (F = 3.42, P = 0.045, df = 2). The difference was significant only between dry forest and the others, but there was a large amount of overlap (Fig. 6). The mean and standard deviation of the Increment Formulas was 0.453 [+ or -] 0.09 for dry forest, 0.354 [+ or -] 0.13 for damp forest and 0.383 [+ or -] 0.09 for wet forest. If these values are used to estimate the age of a tree about 100 cm in dbh (i.e. around 500 mm in radius), the ages produced are: 226.5 [+ or -] 45 y for dry forest, 177 [+ or -] 64 y for damp forest and 192 [+ or -] 43 y for wet forest.

[FIGURE 6 OMITTED]

Examining alternative growth models

A higher percentage of trees had the RCAE age within one standard deviation of the predicted age when using the predictive model and overall formula than for the site formulas (Table 5). However, the regression of predicted age against RCAE age had a greater [R.sup.2] value for the site formulas. The slopes of the site formulas ranged from 0.99 to 2.83. The relationship between the RCAE and predicted age had a slope coefficient closer to one for the site formulas than for either the predictive model or overall formula (Fig. 7), but the error of the predictions varied greatly between sites (Fig. 8b). The variability in error magnitude was related to the number of samples obtained and the strength and linearity of the relationship between tree dbh and age at a site. Therefore, the site formulas provided accurate and precise estimates of tree age, but only when sufficient samples were used to construct the formula. The number of samples required depended on the strength of the linear association between tree age and dbh at the site.

When no information on growth rates is available at a site, a choice between the predictive model and the overall formula needs to be made. The two alternatives had a similar percentage of trees whose RCAE age lay within one standard deviation of the predicted age. However, the [R.sup.2] value of the association between RCAE age and predicted age was higher for the predictive model (Table 5). In addition, the slope coefficient of the relationship between RCAE and predicted age was slightly lower for the overall model than for the predictive model (Fig. 7), indicating slightly inferior accuracy. The overall formula was calculated by regressing RCAE age against tree dbh for all trees combined, so a constant error rate of about 76 y was predicted across all tree ages (Fig. 8c). This error margin is larger than that of the predictive model until trees reach about 300 y. Larger error margins means the RCAE age has a greater likelihood of lying within one standard deviation of the predicted age, but results are less precise. It therefore seems that when no site specific information on growth rate is available, the predictive model is the superior choice. The predictions made by the predictive model were relatively well correlated with RCAE age (Table 5, Fig. 7a) and the error margins were flexible, increasing with estimated tree age (Fig. 8a). The relationship between tree age and error for the predictive model was: error = 0.153age + 30.4.

Discussion

Two main methods for ageing trees were examined in this study, ring counting with extrapolation and growth models. For the first method, error can be associated with both the ring counting and the extrapolation. Only the second source of error could be examined in this study due to a lack of precise information on disturbance history which prevented cross-dating. Ring counting error was estimated using work by Brookhouse (1997). The error resulting from extrapolating was estimated at about 15% of the proportion of the wood sample that was missing. This held true until about 70% of the wood sample was removed, when the error rates showed a less consistent pattern. The extrapolation error was added to the ring counting error and the total error was estimated to increase linearly with estimated tree age at a rate of roughly 10%.

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

Other studies have found that accurate age estimates can be obtained for most trees, but substantial error can occur. Banks (1990) cross-dated several eucalypt species and found that the greatest inaccuracy of ring counting was only a few years. Banks (1993) cross-dated radii ra·di·i  
n.
A plural of radius.


radii
Noun

a plural of radius
 and estimated the error of ring counting in Eucalyptus regnans Noun 1. Eucalyptus regnans - tree having wood similar to the alpine ash; tallest tree in Australia and tallest hardwood in the world
mountain ash

eucalypt, eucalyptus tree, eucalyptus - a tree of the genus Eucalyptus
 F.Muell. from the Central Highlands Central Highlands is the name for several mountainous regions located in the center of the nations or geographical regions.
  • Central Highlands (Central America)
  • Central Highland (France)
  • Central Highlands (Iceland)
 of Victoria to be 2%. Banks (1982) examined the rings of E. pauciflora Sieb. ex Spreng. in alpine NSW NSW New South Wales

Noun 1. NSW - the agency that provides units to conduct unconventional and counter-guerilla warfare
Naval Special Warfare
 and found most ages were accurate, although some large discrepancies occurred. For example, ring counts ranged between 24 and 50 on one tree due to some compression in the wood. Banks (1997) also found variable accuracy when using radiocarbon dating to check ring counts. Two trees estimated by ring counting to be 170 and 130 y old showed reasonable accuracy, with radiocarbon dating estimating the trees to be between 116 and 176 y old. However, for one tree the radiocarbon dated age was double the estimated ring count age due to indistinct in·dis·tinct  
adj.
1. Not clearly or sharply delineated: an indistinct pattern; indistinct shapes in the gloom.

2. Faint; dim: indistinct stars.

3.
 outer rings (ring count age: 160; radiocarbon dated age 354 [+ or -] 30). Because radiocarbon dating and cross-dating were not available in the present study, a different technique was used to assess the accuracy of tree age estimates. The error rates estimated were relatively high. The trees we used were not selected for ring clarity or soundness of wood and the frequent use of extrapolation to cater for incomplete samples may have influenced ring count accuracy. The length of the missing section and the accuracy with which this length is measured has been shown to have a large influence on tree age estimation accuracy (Rozas 2003).

The estimates of tree age and associated error in the current study are believed to be appropriate for most trees, but the error may be underestimated for the older and larger trees for several reasons. Firstly, the error rate of ring counting was obtained for trees less than 100 cm in dbh (Brookhouse 1997). Many of the trees we examined were larger than 100 cm and ring counting can become more difficult in older trees (Banks 1982). Secondly, a linear extrapolation was used to estimate the age of the missing wood sample, despite the knowledge that trees often decrease in growth rate and produce narrower rings as they age (Rayner 1992; Banks 1993; but see Bradshaw and Rayner 1997b). Thirdly, when just a small amount of wood was available for ring counting, which occurred most often in older trees, only the outer rings could be used to determine the extrapolation slope (increment formula). The accuracy of the tree age estimate was found to decrease as the proportion of wood sample available diminished, and was extremely inaccurate if more than 70% of the wood sample was missing. As these outer rings were formed during a period of reduced growth, the process of linear extrapolation is likely to overestimate o·ver·es·ti·mate  
tr.v. o·ver·es·ti·mat·ed, o·ver·es·ti·mat·ing, o·ver·es·ti·mates
1. To estimate too highly.

2. To esteem too greatly.
 the age of the older and larger trees. Fourthly, the model estimating the extrapolation error was produced from only six trees. While they were relatively large trees (maximum dbh 165 cm), they were smaller than the biggest trees used in this study. It is uncertain if this model provides accurate error estimates for larger trees. Finally, excessive error is likely to have occurred in trees exhibiting butt swell (which occurs mostly in older stands), and the magnitude of the butt swell increases with stem size (Walshe 2001). Therefore, while the ages and associated errors are assumed to be reasonable for most trees, there is reason to suspect that this study underestimates the error in age estimates for the largest and oldest trees. Twenty-five trees were estimated to be between 400 and 735 y old with an average dbh of 188 cm [+ or -] 67 s.d. Fourteen of these older trees did not include 400 y within the error margins. Butt swell and or buttressing was noted in 13 of these 25 trees which would have increased the error of the age estimates. Excessive error in these trees may also have occurred when a very small proportion of the radius was examined (four trees) or the rings were especially difficult to count (three trees). For three trees no reason for excessive error could be found and for the other two trees the size and condition of the stump suggested that the ages were appropriate (415 [+ or -] 35 y and 438 [+ or -] 36 y). These 25 trees were removed from growth model construction due to the potentially greater error in the estimation of tree age.

Tree growth rates were related to tree dbh and a number of site-level variables: soil phosphorus, average temperature of the site and stand age. Trees were found to grow faster on soils with high phosphorus concentrations, on warmer sites and in younger stands. Although the broad classification of forest type was not included in the predictive model, it was found that the trees in dry forest grew significantly more slowly than trees in either damp or wet forest. This effect was expressed in the model by the inclusion of soil phosphorus levels. Higher soil phosphorus levels corresponded to an increase in forest wetness for the sites examined (Koch, unpublished data). This effect means that trees found in dry forest will be, on average, older for the same size than trees found in either damp or wet forest (for a tree 100 cm in dbh a mean difference of 36.5 y is expected with damp forest and 28.5 y with wet forest). However, there was a large degree of overlap in tree growth rates (increment formulas) between the forest types. Therefore forest type alone is not an accurate predictor of growth rate.

The variables conspicuous by their absence from the predictive model were tree shape and crown class. Other studies have shown that more senescent se·nes·cent
adj.
Growing old; aging.
 trees have slower growth than younger, healthier trees (Banks 1993; but see Bradshaw and Rayner 1997b). In addition, subdominant sub·dom·i·nant  
n. Music
The fourth tone of a diatonic scale, next below the dominant.

adj.
1. Zoology Less than dominant; ranking below one that is dominant:
 trees are less able to compete for sunlight and so are likely to have slow growth. The lack of a relationship with tree dominance may be because the data contained few suppressed trees or because some of those classified as suppressed were merely young and had not reached the canopy. However, up to 50% of annual rings annual rings, the growth layers of wood that are produced each year in the stems and roots of trees and shrubs. In climates with well-marked alternations of seasons (either cold and warm or wet and dry), the wood cells produced when water is easily available and  can be 'missing' in suppressed trees (e.g. regrowth Re`growth´   

n. 1. The act of regrowing; a second or new growth.
The regrowth of limbs which had been cut off.
- A. B. Buckley.
 E. diversicolor F.Muell: Rayner 1992), which means that these trees may have been older than they appeared. Therefore, the age of subdominant trees may have been less accurate than estimated despite the lack of a significant influence found in the current work. In relation to tree senescence, the relationship between tree senescence and stand age was not straightforward (33.1% of trees examined had dead tops in mature forest, 20.0% in mature forest with regrowth, 10.9% in regrowth with mature forest and 36.8% in regrowth forest) meaning that stand age is not an obvious surrogate for tree senescence in the current study. Although tree shape did not contribute significantly to the growth model, it was significantly related to the slope of the increment formula. This suggests that linear extrapolation may not be appropriate for these older trees (although no single other transformation was appropriate either), and there may be an undetected decrease in growth rate with increasing senescence. While this rationale is plausible, confirmation of tree ages by cross-dating would be required to verify it.

Upon examining the predictive ability and the error rates of the three growth models, the site formulas proved to be the best for predictive purposes. The main concern with using the site formulas was that sufficient trees need to be used to create the formula, and greater error in predictions occurred when a nonlinear association between tree age and dbh was present. While the predictive model proved superior to the overall model, the predictions made by all three growth models had relatively large error bounds, indicating a lack of precision in predictions. It is therefore recommended that such growth models should only be used as a rough guide when a more accurate alternative is unavailable. It should also be noted that these models need to be tested against new data in order to make a true assessment of their utility and accuracy. Gibbons et al. (2000) tested the accuracy of their growth model by ring counting additional trees and found that in all but three of the nineteen cases there was congruence con·gru·ence  
n.
1.
a. Agreement, harmony, conformity, or correspondence.

b. An instance of this: "What an extraordinary congruence of genius and era" 
 between the age estimates.

Implications for management of tree hollows

While some studies have shown that eucalypts can form hollows at a very young age, most agree that more than 100 y is required (Ambrose 1982; Gibbons et al. 2000; Whitford 2002). Eucalypts, however, are unlikely to have hollows suitable for use by fauna if they are less than 120-180 y old, with large hollows being rare in eucalypts less than 220 y old (Gibbons and Lindenmayer 2002). Ambrose (1982) estimated that E. obliqua in Victoria commenced hollow formation at 110 y and the maximum number of cavities occurred at 430 y. In south-eastern Australia, Gibbons et al. (2000) found that less than 50% of E. obliqua and E. fastigata Deane and Maiden trees younger than 180 y old had hollows. Using the material documented in the present paper, Koch et al. (in press) found that trees needed to be at least 100 y old before forming hollows and 140 y before they were likely to contain larger hollows more suited for use by vertebrate vertebrate, any animal having a backbone or spinal column. Verbrates can be traced back to the Silurian period. In the adults of nearly all forms the backbone consists of a series of vertebrae. All vertebrates belong to the subphylum Vertebrata of the phylum Chordata.  fauna.

Although the rate of hollow occurrence increases with tree age (Mackowski 1987; Whitford 2002), we found the error of ring count and extrapolation age estimates also increased with tree age. (The error was less than [+ or -] 15 y for 73% of trees estimated to be 100-300 y of age compared to an average of [+ or -] 42 y for trees > 350 y). However, the higher probability of having a hollow suitable for fauna in the older trees means that making management decisions is possible despite imprecision in tree age estimates. Greater accuracy is achieved for the age at which eucalypts generally begin to produce hollows (100-200 y: Mackowski 1987; Whitford 2002). The imprecision in age estimates is likely to be small enough for the younger trees to mean that age estimates are useful when making management decisions.

Greater error is expected when using growth models compared with ring counting and extrapolation, but growth models will often be more useful because they are easily applied and nondestructive. Quantifying the error in these estimated tree ages can increase the certainty in achieving management objectives. For example, if the objective is to retain a certain number of trees with hollows, based on tree age, then the lower confidence limits of the age at which trees contain hollows can be used to select the trees. Alternatively, the rate of tree retention could be adjusted to cater for inaccuracies in tree age estimation. Taking into account the uncertainty in tree-age estimation will help ensure that most of the retained trees are above the hollow-bearing age, rather than only half of them as would result from using the mean age estimate.

Acknowledgements

This project was funded by the W.V. Scott Fund, the Tasmanian Forest Practices Authority, Forestry Tasmania Forestry Tasmania is a Tasmanian state government owned corporation. It runs the state owned forestry for tourism[1] and logging (for wood chips and timber). Controversy , The University of Tasmania and the Holsworth Wildlife Research Grant. The work could not have been done without the field and laboratory assistance of Chris Spencer Chris Spencer may refer to:
  • Christopher Clarks Spencer, American football player
  • Chris Spencer (actor), actor, comedian, writer and producer
  • Chris Spencer (jazz musician), jazz guitarist based in Seattle, Washington
. Large thanks are given to all the people at Forestry Tasmania who have helped to organise field sites and who were willing to have me on their active coupes. Acknowledgement is made for the data provided by Julie von Platen, and big thanks are given to both Julie and Kathy Allen for counting and talking about tree rings. Thanks are given to Sarah Munks and Kathy Allen for reading drafts of the manuscript. Thanks also to Mick McCarthy Michael Joseph "Mick" McCarthy (b. February 7 1959, Barnsley, England) is an English-born Irish former professional footballer, who is currently the manager of Wolverhampton Wanderers.  who helped when problems were encountered with the WinBUGS code. Finally, thanks to two anonymous reviewers for their very helpful comments with this manuscript.

Appendix 1. Comparing tree age to disturbance history

One method for estimating the error in ring counts is to cross-date the estimated age with known disturbance events. Trees in Australia, particularly in wet forest, often regenerate re·gen·er·ate  
v. re·gen·er·at·ed, re·gen·er·at·ing, re·gen·er·ates

v.tr.
1. To reform spiritually or morally.

2. To form, construct, or create anew, especially in an improved state.
 after fire. Consequently, in Tasmanian wet forest there are often cohorts of trees of the same age but which vary in size (Alcorn et al. 2001). The trees we sampled from wet forest sites were subjectively divided into cohorts according to their RCAE ages. The ages of the trees within each cohort were averaged and, for those 100 y old or less, compared to the known disturbance in the area. The difference between the average age and the time of disturbance was estimated and we calculated the average difference as an indication of ring count accuracy. Exact dates of disturbance were unavailable so this examination is not conclusive. Four of five wet forest sites with trees less than 100 y old showed rough association with known disturbance in the area, indicating relative accuracy in tree ring counts (Table A1). The average error of all the sites was almost 10.5 y. Excluding the first site (because samples may not have been collected from all disturbance events), it is 5.3 y.

Because accurate information on disturbance events was not available, results from a study by Brookhouse (1997) were used to estimate the error in ring counting. Brookhouse (1997) examined E. obliqua and E. cypellocarpa L.Johnson trees of known age in Central Gippsland, Australia, and used cross-dated ring widths to examine error in ring counts. The results of Brookhouse (1997) were considered to be appropriate for use in the current study because the same tree species was examined, the understorey species were similar to those found in the wet forest sites considered in the present study and samples were taken from high-altitude sites (680-700 m asl) with climatic conditions similar to those of Tasmania (average minimum temperature of the coldest period 2.2 [+ or -] 1.6[degrees]C in the current study and below 5[degrees]C in Brookhouse (1997); annual rainfall 1170 [+ or -] 278 mm in the current study and averaging 1174 mm in Brookhouse (1997)).
Table A1. Examining the correspondence between the age estimated for
cohorts in wet forest (where the age is below 100 y) and the
approximate time of disturbance

Mean and s.d. of      Number     Estimated time of       Approximate
tree ages in         of trees    corresponding            error in
cohort (y)           in cohort   disturbance              age count

86                       1       Logged in 1950s           31 (A)
64.9 [+ or -] 4.5        4       Fire 1936 (69 y ago)        4.1
                                 Logged in 1940s
64.2 [+ or -] 5.2        4         (approx. 65 y ago)        0.8
78.2 [+ or -] 7.2        5       Fire about 80 y ago         1.8
84.5 [+ or -] 6.6        4       Fire 60-80 y ago           14.5

(A) Indicates poor correspondence. This may be a true indication of
inaccuracy in the data or may be because no trees from that
regeneration event were examined.

Appendix 2. Predictive growth model code for
WinBUGS

Inits
list(sdtau=1,base=0, a=c(NA,0), b=c(NA,0.001),d=c(NA,
0.001,0.001,0.001,0.001), g=0.001, h=0.001)
model
{
for(i in 1:**) # replace "**" with the number of data points to
be examined
{
X[i,1] <- (dbh[i]-avdbh) / sddbh
X[i,2] <- (soilp[i]-avsoilp) / sdsoilp
X[i,3] <- (avtemp[i]-avavtemp) / sdavtemp
predage[i] <- pred[i] * pred[i]
pred[i] ~ dnorm(mean[i], tau)
mean[i] <- base + g*X[i,1] + a[stand[i]] + m*X[i,2] + n*X[i,3]
}
avdbh <- 101.39
avsoilp <- 255.55
avavtemp <- 10.41
sddbh <- 46.08
sdsoilp <- 134.54
sdavtemp <- 0.988

base ~ dnorm(14.47, sdbase)
sdbase <- 1/(0.204*0.204)
tau<-1/(sdtau*sdtau)
sdtau ~ dnorm(2.167,sdsdtau)
sdsdtau<- 1/(0.01772*0.01772)

a[1] <- 0
a[2] ~ dnorm(-1.069, sda2)
sda2 <- 1/(0.327*0.327)
a[3] ~ dnorm(-1.983, sda3)
sda3 <- 1/(0.294*0.294)

g ~ dnorm(2.635, sdg)
sdg <- 1/(0.170*0.170)
m ~ dnorm(-0.762, sdm)
sdm <- 1/(0.142*0.142)
n ~ dnorm(-0.602, sdn)
sdn <- 1/(0.142*0.142)
}

Data format:
Dbh [ ] soilp[ ] avtemp[ ] asp[ ] stand[ ]

Dbh, soil phosphorus and average temperature are continuous
variables and aspect and stand are binary variables (Stand: 1
= mature forest; 2 = mature with regrowth, 3 = regrowth with
mature or regrowth).


Revised manuscript received 8 May 2008

References

Abbott, I. and Loneragan, O. (1983) Growth rate of jarrah jar·rah  
n.
An Australian tree (Eucalyptus marginata) widely grown for its hard red-brown wood.



[Nyungar (Aboriginal language of southwest Australia) jarily.
 (Eucalyptus marginata
"Jarrah" redirects here; for other uses, see Jarrah (disambiguation).
Jarrah (Eucalyptus marginata) is one of the most common species of Eucalyptus tree in south western Australia.
) in relation to site quality in cut-over forest, Western Australia Western Australia, state (1991 pop. 1,409,965), 975,920 sq mi (2,527,633 sq km), Australia, comprising the entire western part of the continent. It is bounded on the N, W, and S by the Indian Ocean. Perth is the capital. . Australian Forestry 46, 91-102.

Akeroyd, M.D., Leaney, F.W., Mathieson, M. and Moloney, D. (2002) Dating spotted gum (Corymbia citriodora Corymbia citriodora is a tall tree, to 50 metres in height, from temperate and tropical eastern Australia. It is also known as lemon-scented gum, spotted gum and lemon eucalyptus. ) tree rings in south-eastern Queensland using [sup.14]C measurements of cellulose. Australian Forestry 65, 265-267.

Alcorn, P.J., Dingle, J.K. and Hickey, J.E. (2001) Age and stand structure in a multi-aged wet eucalypt forest at the Warra silvicultural systems trial. Tasforests 13, 245-259.

Ambrose, G.J. (1982) An ecological and behavioural study of vertebrates using hollows in eucalypt branches. PhD thesis, La Trobe University 1. u/r = unranked

2.AsiaWeek is now discontinued. Student life
During the 1970s and 1980s, La Trobe, along with Monash, was considered to have the most politically active student body of any university in Australia.
, Bundoora, Victoria Bundoora is a suburb of Melbourne, Victoria, Australia. The word Bundoora is Aboriginal for "the favourite haunt of the kangaroo". Its Local Government Area is the City of Banyule and the City of Whittlesea. , Australia.

Ashton, D.H. (1981) Fire in tall open-forests (wet sclerophyll Sclerophyll is a type of vegetation that has hard leaves and short internodes (the distance between leaves along the stem). The word comes from the Greek sclero (hard) and phyllon (leaf).  forests). In: Gill, A.M., Groves, R.H. and Noble, I.R. (eds) Fire and the Australian Biota. Australian Academy of Science, Canberra, pp. 339-366.

Banks, J.C.G. (1982) The use of dendrochronology dendrochronology: see dating.
dendrochronology

Method of scientific dating based on the analysis of tree rings. Because the width of annular rings varies with climatic conditions, laboratory analysis of timber core samples allows scientists to
 in the interpretation of the dynamics of the snow gum Noun 1. snow gum - small to medium-sized tree of Australia and Tasmania having smooth white to light-grey bark shedding in patches or strips
Eucalyptus coriacea, Eucalyptus pauciflora, ghost gum, white ash
 forest. PhD thesis, Australian National University, Canberra, Australia.

Banks, J.C.G. (1990) The fire history for two forest types in Glenbog State Forest, NSW. Report to the Joint Scientific Committee on the South-East Forests, Department of Forestry, Australian National University, Canberra, Australia.

Banks, J.C.G. (1993) Tree-ring analysis of two mountain ash Mountain Ash, town, Wales
Mountain Ash, Welsh Aberpennar, town (1981 pop. 26,231), Rhondda Cynon Taff, S Wales. A former mining community, it depended upon the great coal mines nearby, which were developed in the 19th cent.
 trees Eucalyptus regnans F.Muell. from the Watts and O'Shannassy Catchments, Central Highlands, Victoria. A report to the Central Highlands Old Growth Forest project, Australian National University, Canberra, Australia.

Banks, J.C.G. (1997) Tree ages and ageing in yellow box. In: Dargavel, J. (ed.) Australia's Ever-Changing Forests III. Proceedings of the third national conference on Australian forest history. Centre for Resource and Environmental Studies, Australian National University, Canberra, pp. 17-28.

Bar, A., Brauning, A. and Loffler, J. (2006) Dendroecology of dwarf shrubs in the high mountains of Norway--a methodological approach. Dendrochronologia 24, 17-27.

Bowman, D.M.J.S. and Kirkpatrick, J.B. (1984) Geographic variation in the demographic structure of stands of Eucalyptus eucalyptus (y'kəlĭp`təs): see myrtle.
eucalyptus
 delegatensis R.T.Baker on dolerite dol·er·ite  
n. Chiefly British
A dark, fine-grained igneous rock; diabase.



[French dolérite, from Greek doleros, deceitful (from its easily being mistaken for diorite)
 in Tasmania. Journal of Biogeography The Journal of Biogeography (Blackwell Publishing), first published in 1974, is the leading international scientific journal in the subject of biogeography. Papers dealing with all aspects of spatial, ecological and historical biogeography are considered for publication.  11, 427-437.

Bradshaw, F.J. and Rayner, M.E. (1997a) Age structure of the karri forest For the Karri Forest Region, see .
Karri forest is tall open forest of Eucalyptus diversicolor (Karri), one of the tallest hardwoods in the world. It occurs only in the south-west corner of the Southwest Botanical Province of Western Australia, in the Warren
: 1. Defining and mapping structural development stages. Australian Forestry 60, 178-187.

Bradshaw, F.J. and Rayner, M.E. (1997b) Age structure of the karri forest: 2. Projections of future forest structure and implications for management. Australian Forestry 60, 188-195.

Brookhouse, M.T. (1997) Identification and analysis of growth rings in Eucalyptus obliqua L'Herit. and E. cypellocarpa L.Johnson. BSc (Hons) thesis, Australian National University, Canberra, Australia.

Brookhouse, M. and Brack, C. (2006) Crossdating and analysis of eucalypt tree rings exhibiting terminal and reverse latewood. Trees--Structure and Function 20, 767-781.

Cherubini, P., Dobbertin, M. and Innes, J.L. (1998) Potential sampling bias in long-term forest growth trends reconstructed from tree rings: a case study from the Italian Alps. Forest Ecology Forest ecology is the scientific study of patterns and processes in forests. The management of forests is known as forestry. Forest Ecosystem
Scope of Forest Ecology
 and Management 109, 103-118.

Department of Natural Resources and Environment (2002a) Central Gippsland Forest Management Area--Estimate of Sawlog The term sawlog refers to that part of a tree stem that will be processed at a sawmill. This is in contrast to those other parts of the stem that are designated pulpwood. The differences are that sawlogs will be greater in diameter, straighter and have a lower knot frequency.  Resource. Department of Natural Resources and Environment, East Melbourne, Australia.

Department of Natural Resources and Environment (2002b) Benalla-Mansfield Forest Management Area--Estimate of Sawlog Resource. Department of Natural Resources and Environment, East Melbourne, Australia.

Forest Practices Authority (2005) Forest Botany botany, science devoted to the study of plants. Botany, microbiology, and zoology together compose the science of biology. Humanity's earliest concern with plants was with their practical uses, i.e., for fuel, clothing, shelter, and, particularly, food and drugs.  Manual. Forest Practices Authority, Tasmania, Australia.

Gibbons, P. and Lindenmayer, D. (2002) Tree Hollows and Wildlife Conservation in Australia Conservation in Australia is an issue of state and federal policy. Australia is one of the most biologically diverse countries in the world, with a large portion of species endemic to Australia. Preserving this wealth of biodiversity is important for future generations. . CSIRO Publishing CSIRO PUBLISHING is an Australian-based science and technology publisher. They cover a range of scientific disciplines including agriculture, chemistry, plant and animal sciences, natural history and environmental management. , Collingwood, Victoria Collingwood is an inner city suburb in Melbourne, Victoria, Australia. It is in the Local Government Area of the City of Yarra. Collingwood is one of the oldest suburbs in Melbourne and is bordered by Smith St, Alexandra Parade, Hoddle St. and Victoria Parade. , Australia.

Gibbons, P., Lindenmayer, D.B. and Tanton, M.T. (2000) Hollow formation in eucalypts from temperate forests in southeastern Australia. Pacific Conservation Biology conservation biology
n.
The branch of biology that deals with the effects of humans on the environment and with the conservation of biological diversity.
 6, 218-228.

Jackson, M.L. (1964) Soil Chemical Analysis. Prentice-Hall Inc., Englewood Cliffs, New Jersey Englewood Cliffs is a borough in Bergen County, New Jersey, United States. As of the United States 2000 Census, the borough population was 5,322. The borough houses the world headquarters of CNBC and the American headquarters of Unilever. .

Jonsson., S., Gunnarson, B. and Criado, C. (2002) Drought is the major limiting factor A factor or condition that, either temporarily or permanently, impedes mission accomplishment. Illustrative examples are transportation network deficiencies, lack of in-place facilities, malpositioned forces or materiel, extreme climatic conditions, distance, transit or overflight rights,  for tree-ring growth of high-altitude Canary Island pines on Tenerife. Geografiska Annaler Series A--Physical Geography 84A, 51-71.

Koch, A.J., Munks, S.A., Driscoll, D.A. and Kirkpatrick, J.B. (in press) Does hollow occurrence vary with forest type? A case study in wet and dry Eucalyptus obliqua forest. Forest Ecology and Management.

Leal, S., Pereira, H., Grabner, M. and Wimmer, R. (2004) Tree-ring structure and climatic effects in young Eucalyptus globulus Labill. grown at two Portuguese sites: preliminary results. Dendrochronologia 21, 139-146.

Lindenmayer, D.B., Mackey, B.G., Mullen, I.C., McCarthy, M.A., Gill, A.M., Cunningham, R.B. and Donnelly, C.F. (1999) Factors affecting stand structure in forests. Are there climatic and topographic determinants? Forest Ecology and Management 123, 55-63.

Lloyd, P.G. and Lau, J.A. (1986) Timber, Vegetation and Habitat Resources of the Chiltern Regional Park. Department of Conservation Forests and Lands, Western Australia.

Lorimer, C.G., Dahir, S.E. and Singer, M.T. (1999) Frequency of partial and missing rings in Acer saccharum in relation to canopy position and growth rate. Plant Ecology 143, 189-202.

Mackowski, C.M. (1987) Wildlife hollows and timber management in blackbutt Blackbutt can refer to:
  • Blackbutt, Queensland, Australia
  • The Blackbutt Reserve in New South Wales, Australia
  • Any of several species of Eucalyptus tree, especially Eucalyptus patens and Eucalyptus pilularis.
 forest. MSc thesis, University of New England The University of New England can refer to:
  • University of New England, Maine, in Biddeford, Maine
  • University of New England, Australia, in New South Wales
, Armidale, Australia.

Masiokas, M. and Villalba, R. (2004) Climatic significance of intra-annual bands in the wood of Nothofagus pumilio Nothofagus pumilio (Lenga Beech or Lenga) is a deciduous tree or shrub native to the central area of the Andes range, in the temperate forests of Chile and Argentina down to Tierra del Fuego.  in southern Patagonia. Trees--Structure and Function 18, 696-704.

Mayfield, A.E., Allen, D.C. and Briggs, R.D. (2007) Site and stand conditions associated with pine false webworm. Populations and damage in mature eastern white pine plantations. Northern Journal of Applied Forestry 24, 168-176.

Mazanec, Z. (1968) Influence of defoliation by the phasmatid Didymuria violescens on seasonal diameter growth and the pattern of growth rings in alpine ash. Australian Forestry 31, 3-14.

McCarthy, M.A. (2007) Bayesian Methods for Ecology. Cambridge University Press Cambridge University Press (known colloquially as CUP) is a publisher given a Royal Charter by Henry VIII in 1534, and one of the two privileged presses (the other being Oxford University Press). , Cambridge, UK.

McMahon, J.P., Hutchinson, M.F., Nix, H.A. and Ord, K.D. (1995) ANUCLIM User's Guide. Centre for Resource and Environmental Studies. The Australian National University, Canberra, Australia.

Moloney, D., Wormington, K. and DeStefano, S. (2002) Stag retention and use by arboreal arboreal

pertaining to trees, treelike, tree-dwelling.
 marsupials in eucalypt forests of southeast Queensland, Australia: implications for management. In: Laudenslayer, W.F.J., Shea, P.J., Valentine, B.E., Weatherspoon, P.C. and Lisle lisle  
n.
1. A fine, smooth, tightly twisted thread spun from long-stapled cotton.

2. Fabric knitted of this thread, used especially for hosiery and underwear.
, T. (eds) Proceedings of the Symposium on the Ecology and Management of Dead Wood in Western Forests. USDA USDA,
n.pr See United States Department of Agriculture.
 Forest Service General Technical Report PSW-GTR-181, Albany, CA. http://www.fs.fed.us/psw/publications/documents/ gtr-181/028_Moloney.pdf

Mucha, S.B. (1979) Estimation of tree ages from growth rings of eucalypts in northern Australia The term northern Australia is generally considered to include the States and territories of Australia of Queensland and the Northern Territory. The part of Western Australia (WA) north of latitude 26° south — a definition widely used in law and State government policy . Australian Forestry 42, 13-16.

Pilcher, J.R. and Gray, B. (1982) The relationships between oak tree growth and climate in Britain. Journal of Ecology The Journal of Ecology (not to be confused with another journal called Ecology) is a scientific journal concerning plant ecology. It was first published in 1913, and is the oldest peer-reviewed, international ecological journal.  70, 297-304.

R Development Core Team (2006) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.

Rayner, M.E. (1992) Application of Dendrochronology, Stem Analysis and Inventory Data in the Estimation of Tree and Stand Ages in Karri Forest. Department of Conservation and Land Management, Western Australia, 19 pp.

Rose, P.W. (1993) Production of habitat hollows by wheatbelt eucalypts. Rose and Bending Forest and Environmental Consultants Final Report, Save the Bush Research Grant 1991/92--Project R053 Department of Conservation and Land Management, Western Australia.

Rozas, V. (2003) Tree age estimates in Fagus sylvatica and Quercus robur Quercus robur,
n See oak.
: testing previous and improved methods. Plant Ecology 167, 197-212.

Smith, A.P. and Lindenmayer, D.B. (1988) Tree hollow A tree hollow is a semi-enclosed cavity which has naturally formed in the trunk or branch of a tree. These are predominantly found in old trees, whether living or not. Hollows form in many species of trees, and are a prominent feature of natural forests and woodlands, and act as a  requirements of Leadbeater's possum possum
 or phalanger

Any of several species (family Phalangeridae) of nocturnal, arboreal marsupials of Australia and New Guinea. They are 22–50 in. (55–125 cm) long, including the long prehensile tail, and have woolly fur.
 and other possums and gliders in timber production ash forests of the Victorian Central Highlands. Australian Wildlife Research 15, 347-362.

Spiegelhalter, D., Thomas, A., Best, N. and Lunn, D. (2003) WinBUGS User Manual. MRC See Maximum return criterion.  Biostatistics biostatistics /bio·sta·tis·tics/ (-stah-tis´tiks) biometry.

bi·o·sta·tis·tics
n.
The science of statistics applied to the analysis of biological or medical data.
 Unit, Cambridge, UK.

SPSS (2005) SPSS for Windows 14.0, Rel. 12.0.1. SPSS Inc., Chicago.

Turner, J. (1984) Radiocarbon dating of wood and charcoal in an Australian forest ecosystem Forest ecosystem

The entire assemblage of organisms (trees, shrubs, herbs, bacteria, fungi, and animals, including people) together with their environmental substrate (the surrounding air, soil, water, organic debris, and rocks), interacting inside a defined
. Australian Forestry 47, 79-83.

Walshe, T.V. (2001) Describing natural disturbance regimes for use as a forest management tool. PhD thesis, The University of Melbourne
  • AsiaWeek is now discontinued.
Comments:

In 2006, Times Higher Education Supplement ranked the University of Melbourne 22nd in the world. Because of the drop in ranking, University of Melbourne is currently behind four Asian universities - Beijing University,
, Melbourne, Australia.

Waring, K.M. and O'Hara, K.L. (2006) Estimating relative error in growth ring analyses of second-growth coast redwood (Sequoia sempervirens). Canadian Journal of Forest Research 36, 2216-2222.

Whiteley, S.B. (1999) Calculating the sustainable yield The sustainable yield of natural capital is the ecological yield that can be extracted without reducing the base of capital itself, i.e. the surplus required to maintain nature's services at the same or increasing level over time.  of Tasmania's State Forests. Tasforests 11, 23-34.

Whitford, K.R. (2002) Hollows in jarrah (Eucalyptus marginata) and marri Noun 1. marri - very large red gum tree
Eucalyptus calophylla, red gum

eucalypt, eucalyptus tree, eucalyptus - a tree of the genus Eucalyptus
 (Corymbia calophylla Corymbia calophylla or the Marri (also known as Eucalyptus calophylla (R. Br.)) is a bloodwood native to Western Australia.

It is distinctive among bloodwoods for its very large buds and fruit (colloquially "honky nuts" in Western Australia).
) trees I. Hollow sizes, tree attributes and ages. Forest Ecology and Management 160, 201-214.

Whitford, K.R. and Williams, M.R. (2002) Hollows in jarrah (Eucalyptus marginata) and marri (Corymbia calophylla) trees II. Selecting trees to retain for hollow dependent fauna. Forest Ecology and Management 160, 215-232.

Woodgate, P.W., Peel, W.D., Ritman, K.T., Coram, J.E., Brady, A., Rule, A.J. and Banks, J.C.G. (1994) A study of the old-growth forests of East Gippsland. Department of Conservation and Natural Resources Department of Conservation and Natural Resources can refer to agencies of various governments: United States
  • Alabama Department of Conservation and Natural Resources
  • Nevada Department of Conservation and Natural Resources
, Melbourne, Australia. Unpublished.

Wormington, K. and Lamb, D. (1999) Tree hollow development in wet and dry sclerophyll eucalypt forest in south-east Queensland, Australia. Australian Forestry 62, 336-345.

Yamaguchi, D.K. (1991) A simple method for cross-dating increment cores from living trees. Canadian Journal of Forest Research 21, 414-416.

Amelia J. Koch (1,2,3), Don A. Driscoll (4) and J.B. Kirkpatrick (1)

(1) School of Geography and Environmental Science, University of Tasmania, Private Bag 78, Hobart, Tasmania, 7001, Australia

(2) Email: ajkoch@utas.edu.au

(3) Current address: Tasmanian Forest Practices Authority, 30 Patrick St, Hobart, Tasmania, 7000, Australia

(4) Centre for Resource and Environmental Studies, Australian National University, W.K. Hancock Building (43) Biology Place, Canberra, ACT 0200, Australia
Table 1. Tree and site attributes measured

Site and tree attributes    Description

Dbh                         Tree diameter at breast height (1.3 m)
                            over bark measured using a diameter tape
                            (cm)

Tree shape                  A classification of the shape of a tree,
                            based on definitions by Smith and
                            Lindenmayer (1988): (1) tree with full
                            round crown; (2) mature tree with large
                            branching; (3) mature tree with dead
                            branches; (4) mature tree with dead top
                            but big crown; (5) mature tree with dead
                            top and small crown.

Crown class                 The relative dominance of each tree was
                            categorised as: (1) dominant; (2)
                            co-dominant; (3) sub-dominant or
                            suppressed.

Tree height                 The height of the tree was measured after
                            felling (m).

Burn damage                 The burn damage of the tree was assessed
                            as: (0) no evidence of fire; (1) bark
                            damage; (2) wood exposed and damaged by
                            fire; (3) severe fire damage where the
                            base of the tree forms a bridge.

Tree species                Tree species was determined.

Forest type                 The forest type was determined using the
                            RFA classification available from the
                            Forest Botany (Forest Practices Authority
                            2005) and then reduced to the classes:
                            Dry: (DRY hOB) understorey dominated by
                            bracken or low shrubs; Damp (DRY shOB)
                            understorey dominated by shrubs over 2 m,
                            often including broad-leaved species;
                            Wet: (WET all types) understorey
                            dominated by broad-leaved species.

Eastings                    A GPS was used to determine the eastings
                            of the sites in GDA. Values were
                            confirmed using a map of the area.

Northings                   A GPS was used to determine northings of
                            the sites in GDA. Values were confirmed
                            using a map of the area.

Aspect                      A compass was used to determine the
                            direction directly downslope. The aspects
                            were divided into four categories (N, S,
                            E, W).

Altitude                    A GPS was used to determine the altitude
                            of the sites (m). Values were confirmed
                            using a map of the area.

Topography                  The average topography of the site was
                            categorised as: (1) ridge; (2) upper
                            slope; (3) mid-slope; (4) lower slope or
                            gully.

Slope                       A clinometer was used to determine the
                            slope of the site in degrees.

Rock                        Parent rock type of the substrate; this
                            was obtained from the plans developed for
                            the harvesting of the coupe (Forest
                            Practices Plans) and was classified as:
                            (1) granite; (2) dolerite; (3) sediment.

Soil type                   Soil type as obtained from Forest
                            Practices Plans; this was classified as:
                            (1) loamy; (2) clayey; (3) sandy.

Soil pH                     Four soil samples were collected from the
                            top 10 cm. The samples were mixed and the
                            pH was measured using a probe in a 1:5
                            soil: distilled water solution.

Soil nitrogen               Soil was air-dried, ground, sieved and
                            measured for total nitrogen using the
                            Kjeldahl method (Jackson 1964).

Soil phosphorus             As for soil nitrogen, but measured for
                            extractable phosphorus using the Bray and
                            Kurtz method (Jackson 1964).

Stand age                   Categorised according to the information
                            from the photographic interpretation (PI)
                            of the vegetation age structure within
                            each plot (obtained from Forestry
                            Tasmania's concise PI type maps): (1)
                            mature; (2) mature with regrowth; (3)
                            regrowth with mature; (4) regrowth.

Stand height                Categorised according to the information
                            from the photographic interpretation (PI)
                            of the vegetation age structure within
                            each plot into: (0) unknown; (1) 20 m;
                            (2) 30 m (3) 37 m; and (4) 50 m.

Average temperature         Annual mean temperature values obtained
                            from ESOCLIM (McMahon et 1995) using data
                            on the latitude, al. longitude and
                            altitude of the site.

Rain                        Annual mean precipitation values obtained
                            from ESOCLIM using data on the latitude,
                            longitude and altitude of the site.

Radiation                   Annual mean radiation values obtained
                            from ESOCLIM using data on the latitude,
                            longitude and altitude of the site.

Table 2. Definitions of terms used throughout the text

Term                  Definition

RCAE method           The method for estimating tree age, where ring
                      counts are done and the age of the missing stem
                      centre is estimated using linear extrapolation.

RCAE age              The age of the tree as estimated using the RCAE
                      method.

Decadal increments    The distance between every ten rings on a wood
                      sample.

Increment formulas    The slope of the fitted line between the radius
                      and the number of rings for each tree, used to
                      extrapolate the age of the missing stem centre.

Extrapolation error   The model predicting the absolute percentage
model                 error in tree age based on the percentage of
                      wood sample that was missing.

Site formulas         The models predicting tree age, produced from
                      Bayesian regression analysis between the tree
                      diameter and estimated age for all trees at a
                      site.

Overall formula       The model predicting age of a tree produced from
                      Bayesian regression analysis between the tree
                      diameter and estimated age for all trees
                      combined.

Predictive model      A growth model predicting the age of a tree
                      produced by Bayesian regression model between
                      the RCAE age of a tree and site and tree
                      variables.

Term                  Sampling details

RCAE method

RCAE age              RCAE ages were estimated for 329 samples.
                      Between 4 and 12 samples were examined for each
                      of 38 sites.

Decadal increments

Increment formulas    Three hundred and twenty-nine increment formulas
                      were constructed. Between 4 and 40 decadal
                      increments were used to construct each formula.

Extrapolation error   Six wood samples from dry forest that were not
model                 used in any other part of the current study were
                      used to construct this model.

Site formulas         Thirty-eight site formulas were constructed.
                      Between 4 and 12 trees were used to construct
                      the formula for each site.

Overall formula       Three hundred and four samples were used to
                      construct this model.

Predictive model      Three hundred and four samples were used to
                      construct this model.

Table 3. Details of the trees examined for establishing error in
linear extrapolation

Tree       Ring      Sample radius     Tree dbh
number     count         (cm)            (cm)

1           220          62.6            140
2           270          57.4            125
3           230          55.9            165
4           221          64.3            130
5           155          46.0            120
6           146          32.6            100

Table 4. Coefficients (mean, standard deviation, 2.5th and 97.5th
percentile) of the explanatory variables included in the best
Bayesian linear model for predicting tree age using only trees with
an RCAE age of 400 y or less (n = 299)

Variable                 Mean    s.d.    2.50%    97.50%   DIC (A)

Intercept               14.470   0.204   14.070   14.870    1538
Dbh                      2.635   0.170    2.309    2.966    1379
Mature                    0                                 1344
Mature with regrowth    -1.069   0.327   -1.705   -0.436
Regrowth (B)            -1.983   0.294   -2.565   -1.403
Soil phosphorus         -0.762   0.142   -1.043   -0.493    1332
Average temperature     -0.602   0.142   -0.877   -0.324    1316

(A) DIC values are the Bayesian equivalent of an AIC value to
indicate progressive model improvement by addition of each new
variable

(B)(Regrowth includes the stand age categories 'regrowth with
mature trees', and 'regrowth'

Table 5. Comparing the accuracy of the growth models

                     Fraction of trees within     Adjusted [R.sup.2] of
                    one standard deviation (A)       regression with
Method                           (%)                     RCAE age (B)

Predictive model               67.8                       0.642
Site formulas                  84.2                       0.699
Overall formula                74.5                       0.426

(A) The percentage of trees whose RCAE (ring count and extrapolation)
age was found to lie within one standard deviation of the predicted
age for the three growth models considered (n = 329)

(B) The adjusted R2 value of the regression analysis between the
predicted results from the growth models and the RCAE age
COPYRIGHT 2008 Institute of Foresters of Australia
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2008 Gale, Cengage Learning. All rights reserved.

 Reader Opinion

Title:

Comment:



 

Article Details
Printer friendly Cite/link Email Feedback
Author:Koch, Amelia J.; Driscoll, Don A.; Kirkpatrick, J.B.
Publication:Australian Forestry
Article Type:Report
Geographic Code:8AUST
Date:Jun 1, 2008
Words:9766
Previous Article:The politics of community participation in natural resource management: lessons from community forestry in Nepal.
Next Article:William Edwin (Ted) Hillis: 9 February 1921-3 February 2008.
Topics:



Related Articles
Pre-commercial thinning and fertiliser enhance growth in young native Eucalyptus obliqua (L'Herit.) stands in Tasmania.
Population differentiation within Eucalyptus obliqua: implications for regeneration success and genetic conservation in production forests.
Growth stress in Eucalyptus dunnii.
Achievements in forest tree improvement in Australia and New Zealand 9. Genetic improvement of Eucalyptus nitens in Australia.
Long-term response of Eucalyptus regnans F.Muell. to commercial thinning in a regrowth eucalypt forest in Tasmania.
Victoria's tallest trees.
Vulnerability of Tasmanian giant trees.
The effect of tree spacing on the production of flowers in Eucalyptus nitens.
Tree hollow incidence in Victorian state forests.
Premature decline of eucalyptus and altered ecosystem processes in the absence of fire in some Australian forests.

Terms of use | Copyright © 2014 Farlex, Inc. | Feedback | For webmasters