Eucalyptus nitens, recovery and economics of processing 15 year old trees for solid timber
Report Date: May 2015
Author: Dean Satchell, Sustainable Forest Solutions, R.D. 1 Kerikeri, Northland 0294
+64 9 4075525
Special thanks and acknowledgement go to:
- MPI Sustainable Farming Fund
- Neil Barr Farm Forestry Foundation
- John Fairweather Specialty Timbers
- North Canterbury, South Canterbury, South Otago and Southland branches of NZFFA
- NZFFA Eucalyptus Action Group
- NZFFA Research committee
Appendix 2: Sawn timber price estimates
Appendix 3: Literature review - Value-based survey pricing methods
Appendix 4: Literature review - Estimating profitability of growing E. nitens for solid timber production
Appendix 5: Sawmilling method
Appendix 6: Flooring price survey instrument
Appendix 7: Survey results table
Appendix 8: Survey analysis
Appendix 9: Wood physical properties, test results
Appendix 10: Glossary of terms
Appendix 11: Case study stand plot
Appendix 12: Comparison between levels of internal and surface checking
Appendix 13: Air drying experiment
Appendix 14: Sensitivity analysis
Appendix 4: Literature Review
Estimating the profitability of growing E. nitens for solid timber production
This study examines the economic viability of growing Eucalyptus nitens for solid timber products using discounted cash flow analysis. Because E. nitens has not been processed or marketed in New Zealand to date the approach was to value sawlogs based on residual value to the grower, which involves subtracting processing costs from sawn timber revenues.
A small stand of pruned and thinned 15-year-old trees was available in Canterbury for harvesting, as was a small-scale processing operation that was recently set up specifically for processing eucalypt for solid timber. This provided the opportunity for assessing revenues and expenditures as a case study to estimate profit to the grower.
Eucalyptus nitens is a fast growing and cold-hardy eucalypt species grown in cooler regions of New Zealand (McKenzie, Turner, & Shelbourne, 2003, p. 63). The species is commonly found on farms and in small plantations throughout rural Canterbury, Otago and Southland, with some stands pruned in anticipation of solid timber processing. The Southland Plantation Forest Company Ltd has established approximately 10,000 hectares of E. nitens in Southland for hardwood chip/fibre. E. nitens is considered to be a well-suited plantation eucalypt species for cooler climates because of frost tolerance, and good growth and form (Beadle et al., 2008, p. 46; Washusen et al., 2008, p. 4).
E. nitens is being managed for solid timber production in other countries, with approximately 30,000 hectares of Tasmania’s eucalypt plantations (primarily E. nitens and E. globulus) being pruned and thinned (Washusen, 2011) in anticipation of processing of solid timber products. In Chile pruning and thinning of E. nitens is also a common practice (Beadle et al., 2008, p. 46), with approximately 25,000 hectares being pruned and thinned for solid timber production (Valencia, 2014).
Pinus radiata is the dominant plantation forestry species grown in New Zealand with 90% of the plantation forest area in this species, followed by Douglas fir holding 6% of the plantation forest area (NEFD 2011). There may be a case for diversifying plantation timber species to reduce risk (Nolan, Greaves, Washusen, Parsons, & Jennings, 2005, p. 83). Species diversification may also provide specialty forest products that are quality differentiated and additional to those being currently produced from radiata pine, with opportunities both for export and import substitution that could improve the total revenue produced from forest products.
E. nitens timber has good strength properties and appearance characteristics, but is considered to be very difficult to saw and season. If E. nitens were to yield adequate recoveries of good quality solid-wood products this species could earn a wider role in New Zealand plantation forestry (McKenzie et al., 2003, p. 64). Published studies to date have demonstrated low sawn timber grade-recoveries from the species.
It is well known that investors are cautious of start up forest industries (Nolan et al., 2005, p. 83). Because plantation forestry involves longer time frames than other crops, investors would require a comparatively higher level of confidence in projected returns from forest plantation ventures before considering commitment of capital, especially for ‘emerging’ species lacking a solid history of production, improvement and research.
Hardwood products are mostly imported into New Zealand with negligible domestic production. Historically, hardwood products sourced from old-growth tropical rainforests have been abundant and inexpensive. With chain of custody, legality and sustainability certification of forest products now common practice, assurance of supply and stability of prices are uncertain for hardwood products from old growth and tropical forest origin.
For an industry to develop around plantation hardwood products, log values must reflect both an adequate return to the grower, along with product returns that both cover the costs of production and generate a profit (Nolan et al., 2005, p. 83). Serious growers and investors who would develop an industry of a size that takes advantage of scale efficiencies would require some knowledge of expected returns, even if only returns from operating at a scale in proportion to an emerging industry.
Because of serious processing degrade of sawn timber products, profitability has not been adequately demonstrated to date from growing E. nitens for solid timber.
The objective was to assess economic returns to the tree grower for 15 year-old E. nitens that were pruned and thinned, using best practice contemporary processing methods.
To achieve this overall research objective it was necessary to also examine stand characteristics, wood processing costs, sawn timber grade recoveries, product pricing, novel products and wood degrade issues. Contemporary best practice processing methods using small-scale processing equipment were applied in an attempt to produce grade recoveries sufficient for a profit to the grower.
This study aimed to estimate the profitability of growing E. nitens for solid timber products as a case study using discounted cash flow analysis. Discounted cash flow analysis is based on the theory of compound interest and brings all future costs and revenues into the present by adjusting these to take into account interest over the number of years required for the cost or revenue to be realised (Brown, 2000, p. 54). Discounted cash flow analysis is generally recognised as the accepted method of forest valuation in New Zealand (Bloomberg, Bigsby, & Sedcole, 2002, p. 27).
Estimating profitability requires pricing of logs and trees. A method commonly used for appraising log values is the residual value method (Fairfax & Yale, 1987, p. 125). The residual value method requires price estimates for final products and deducts processing costs and required profits to determine the residual log price (Fairfax & Yale, 1987, p. 125; Liggett, 1995, p. 84). This price represents the maximum amount the sawmill would pay for the log (Mbugua, 2003, p. 24).
The residual value approach has been used since the turn of the 20th century in Western U.S.A. for appraisal of log values (Fairfax & Yale, 1987, p. 125). Historically U.S.A. Forest Service appraisals involved estimating sales value of timber products from a price index of past sales (Fairfax & Yale, 1987, p. 125), with gross product values determined by the grade volumes multiplied by the lumber prices (Nagubadi, Fight, & Barbour, 2003, p. 2). Transactional evidence is now favoured over residual value methods in the U.S.A. for valuing logs of species where comparable sales data for logs is available (Nagubadi et al., 2003, p. 1).
In New Zealand, the residual value method has been used for calculating radiata pine log grade prices that were used to explain regional variation in sawlog prices according to their attributes (Bloomberg et al., 2002). In Australia, Innes at al. estimated log values for plantation E. nitens using the residual value method (Innes, Greaves, Nolan, & Washusen, 2008).
Neither log price data, nor product sales price data for sawn products is available for E. nitens in New Zealand. In the absence of sales price data, alternative methods were required to price sawn timber products for determining log residual value. This literature review describes methods for estimating prices in the absence of market data.
In addition to sawn timber product values, it is necessary to consider other factors which affect E. nitens log residual value. Residual value is directly influenced by processing costs and, by its nature, the process of producing sawn timber involves many steps and outcomes can vary. It is important to consider the nature of timber processing in order to assess how best to maximise residual value. Important factors affecting log residual value are the quality of the sawn wood products as grade recoveries, the volumes of sawn timber produced and the costs involved in production of these. Grade recoveries are influenced by silviculture, log position, log diameter and sawmilling methods. Sawn volumes and costs are influenced by log diameter, log length and sawmilling methods. Each of these factors is considered. Firstly, however, attention is given to the application of residual value methods to E. nitens.
Studies of E. nitens residual value
There are clear inconsistencies between production, grading and pricing methods used in contemporary studies that sought to quantify levels of defect, degrade and product recoveries for E. nitens sawn timber in order to estimate log residual value. The need for relevant research was identified by Nolan et al. (2005, p. v) when they made recommendations for improving production efficiency for sawn E. nitens, stating:
Work in these areas should be deliberate comparative studies, operating across species to a standard methodology that integrates growing and milling results, and provides improved assessment data for plantation inventory and economic modeling.
There have been subsequent studies on this topic but they have not met the standards recommended by Nolan et al.
Innes et al. (2008) assessed product recoveries from three production methods in order to estimate log residual values for E. nitens. However, the three production methods could not be adequately compared because grading methods, green sawn thickness, product values and methods for quantifying defect and degrade all differed between them. Furthermore, pricing of timber products was based on assumptions that were arbitrary, producing an average log value delivered to the sawmill of only A $72.37 per cubic metre. The conclusion that growing and processing thinned and pruned E. nitens was uneconomic may have resulted from both poor study design and poor application of grading and pricing methods that led to underestimating log revenues, discussed in more detail below.
Although this study compared returns from three processing methods, grading was not consistent between processing methods (Innes et al., 2008, p. 8). Only one method (Gunns Lindsay Street) produced straightened and machined final products for grading while another method (ITC Newood) did not allow for edge distortion, thus artificially elevating the recovery results for this method (Innes et al., 2008, p. 19). ITC Newood boards were not edged, but board faces were dressed to final product thickness for grading. In the third method (ITC Heyfield) dry boards were edged but faces were only skim dressed. These skim dressed boards were then assessed for surface checking to determine product value, despite acknowledgement that this has lower accuracy compared with grading final products (Innes et al., 2008, p. vi). This is because skip and other sizing defects are difficult to accurately assess and grade for until boards are sized to final product dimensions. Furthermore, dressing to final product dimensions exposes measurable levels of internal checking to the board surface for grading, beyond those exposed during skim dressing. Innes et al. (2008, p. 10) then downgraded select and medium-feature boards to high-feature grade in proportion to observed levels of internal checking for the two methods that involved dressing board faces to final product thickness. However, such downgrading was not undertaken on the technique involving skim-dressing board faces because an assessment of internal checking was not made. It was not explained why levels of checking from the two methods where board faces were dressed to final product thickness were artificially elevated according to levels of internal checking. It could be argued that the apparently arbitrary allowance given for internal checking would have had an unnecessarily high negative impact on grade recoveries from these two processing methods, despite this allowance being described by Innes et al. as a significant loss in value (2008, p. 38). Moreover, because high-feature grade held only one third of the value of Medium-feature grade in this study (2008, p. 10) the impact of study design on log residual value would have potentially resulting in a serious underestimate of profitability. The level of this discount was not discussed nor justified and appears to be arbitrary.
These shortcomings illustrate the importance of methods that produce grades that infer end products that can be priced, rather than grading and pricing quality levels in an arbitrary manner. Furthermore, accurate comparison of the three sawmilling methods was not possible because of serious inconsistencies in design and implementation of the experiment.
Innes et al. acknowledged that total product values were not consistent between processing methods being compared (Innes et al., 2008, pp. 19, 28). Although the differences in sawmilling methods were highlighted in this report by comparing recoveries (2008, pp. 18, 28), this result was clearly confounded by significant differences (Innes et al., 2008, p. 8) in grading procedures between the three production techniques used. Other inconsistencies in assessing log values from the three different production methods used by Innes at al. (2008) included:
- Not specifying green sawn thickness for two of the three sawing methods reported (2008, pp. 4, 6). The third method produced 32 mm green thickness material for a dry skim-dressed board nominal thickness of 25 mm (2008, p. 3). There are clear inconsistencies in sizing that would confound comparisons between recoveries from the different sawing systems.
- A minimum allowable board length of 1.8 m was reported (2008, p. 9) for Gunns Lindsay Street while a discounted value was allowed for lengths shorter than 1.8 m for ITC Heyfield (2008, p. 28). Furthermore, there was no reject grade for ITC Newood (Innes et al., 2008, pp. 9, 28).
These differences highlight grading methods that varied significantly between treatments. A minimum allowable board length of 1.8 m for one processing method would very likely significantly underestimate comparative profitability and bias the results. Furthermore, the 1.8 m length selected from which to apply discounts was not justified by any market information, nor was the discount level justified.
Another recent Australian study illustrates inconsistencies between reports and between sawing and processing treatments that were applied. Washusen et al.(2008, p. 23) in estimating sawn product recovery from E. nitens, reported that boards were skim dressed on the face and back to the nominal size but edges were not dressed. Washusen et al. (2008, p. 24) used the same wholesale prices as Innes et al.(2008) for the three grades used, namely select, standard and high feature grades as specified in Australian standard 2796.1 (Washusen et al., 2008, p. 23). However it was not reported whether grading was undertaken on both faces, or only on the best face. In addition to the grade rules specified in the standard, it was decided arbitrarily that select and standard grades were to also be free of surface checking, sapwood and skip on the skim-dressed surfaces (Washusen et al., 2008, p. 32). Where surface checking or sapwood was present, boards were downgraded from select and standard to high-feature grade (2008, p. 34). Surface checking was recorded as present only if total length of checks on the graded face of the board exceeded 20 mm (2008, p. 44). Furthermore, “a discount of 10 per cent was applied to boards of Select and Standard grades > 1.8 m and < 3.0 m and a discount of 50 per cent was applied to select and standard grade boards < 1.8 m” (Washusen et al., 2008, p. 23). The resulting recovery of select and standard grades was low, as were product values (Washusen et al., 2008, p. 32), resulting in only A$169 average sawn timber revenue per log cubic metre.
There are clear inconsistencies between production, grading and pricing methods used in the available contemporary studies that sought to quantify levels of defect, degrade and product recoveries for E. nitens sawn timber in order to estimate log residual value. Grade recoveries combine with prices as sawn timber revenue, a key component of log residual value. Improved methods would be consistent between all variables except the one being tested and would have a logical basis.
In the absence of product sales data for E. nitens sawn products, improved estimates of log residual value would consider product prices according to product grades and sizes based on product applications, market demand and market competition. Careful product grading into profiles and pricing of these according to market requirements would provide credibility to an economic analysis of E. nitens profitability using the residual value method of pricing logs.
Standardising of grading methods would allow comparisons between studies or treatments to be undertaken by researchers. Standardised methods would also need to be objective in terms of well-defined grade criteria and product profiles that can be priced from market data.
Sawmill productivity is largely dictated by the suitability of the processing equipment for the logs being sawn. A successful processing operation will cost-effectively utilise as much of the log volume as possible (Nolan et al., 2005, p. 82). Maximising revenue from production of sawn timber requires knowledge of products and pricing of these to target products that produce the highest revenue. However, the full range of products and by-products being produced contribute to overall profitability, including high value products along with lower value products produced at low cost (Nolan et al., 2005, p. 81).
Producing higher grades of timber in lower volumes might be preferable to simply maximizing volume production (Nolan et al., 2005, p. 81). The tradeoff between higher volume production of lower value products and lower volume production of higher quality products at higher cost requires skilled judgment calls by the processor. Good decisions would be based on experience and knowledge gained over time from marketing a range of products and from a good understanding of species log characteristics such as propensity for degrade.
The influence degrade has on product quality and value is largely determined by product application. For example, high-value cabinetry and furniture applications require material free of internal checks, whereas product value for tongue and groove flooring will not be affected by internal checks unless these are exposed on the profiled surface (Blakemore & Northway, 2009, p. 4). Clearly there are many factors that influence target sawn product preference and sawmilling decisions on log utilisation.
For this case study, sawn products selected were assumed to be those most suitable for sawing from the case study logs. Assumptions were required on products that would yield greatest returns from the logs. In consultation with industry expertise these were selected as:
- Solid timber flooring from nominal 100 mm, 125 mm and 150 mm widths; and
- panel-lamination stock from 75 mm and 50 mm widths.
Although there is not an established market for E. nitens logs in New Zealand, by estimating sawn timber prices and log revenue, processing costs could be deducted for log residual value. The approach taken in this case study was to employ best practice processing methods identified from the literature, assuming that industry players could reproduce an equivalent productivity on average if they were to utilise these methods. Cost-efficiency improvements would be expected from this benchmark in future as methods improve, with this case study representing average best current practice.
Log quality is an important factor that determines sawmill profitability. Average log quality needs to be good enough to produce sufficient volumes of higher-grade material for returns to exceed the cost of production (Nolan et al., 2005, p. 83). As log quality increases, production of higher quality appearance grades increases per cubic metre of log processed (Nolan et al., 2005, p. 82). Therefore, sawmills should be prepared to pay more for logs where improved quality is definable and returns justify the higher log price. This is the reasoning behind residual log value. For example, hardness ratings appear to increase with tree age in E. nitens (Farrell & Mihalcheon, 2009, p. 11), suggesting that older trees are more suitable for applications such as flooring where surface hardness is desired. If flooring timber were to fetch a price premium in the market, so too could older logs that produce higher quality flooring timber.
Log quality is partly a function of external log quality characteristics such as diameter and form (Alzamora & Apiolaza, 2010). However, other characteristics potentially affecting E. nitens log quality may not be visible, such as diameter of knotty core inside pruned logs. Other measurable descriptors that are not visible but may identify quality include tree age, log position and hardness.
Quality characteristics that could potentially influence grade recoveries and therefore log value may not even be detectable in a freshly cross-cut sawlog, such as propensity for end splits, movement, checking and collapse. If such properties were measurable prior to purchase or processing and a relationship could be established between log revenue and these characteristics, log price could potentially be predicted empirically from such characteristics. If log value could be predicted based on quality descriptors, growers could potentially improve their profit by targeting improved log quality.
Market demand for timber products influences prices obtained for the grades supplied. Greatest market demand and value for hardwood timber is as clearwood (Nolan et al., 2005, p. iv). Select grade timber (graded to Australian standards), with its low level of feature, dominates Australia’s appearance hardwood market and fetches the highest price premium (Nolan et al., 2005, p. 7). This suggests that growers should target production of clearwood from their trees and logs. Two silvicultural practices facilitate production of clearwood from trees, pruning and thinning.
The purpose of pruning trees is to improve wood quality and thus value of the resulting crop. It is well documented that pruning and thinning of plantation eucalypts is necessary for adequate recoveries of clear appearance grade timber (Shield, 1995, p. 135).
Pruning does not necessarily produce high recoveries of clearwood. Innes et al. reported that “Grade and overall recovery from thinned and pruned butt logs of 26 year old E. nitens was no better than from unpruned top logs from the same trees” (2008, p. 19). The logs were described as pruned too late and with a larger knotty core than would be expected from early-pruned stems (Innes et al., 2008, p. 38). In contrast McKenzie et al. (2003, p. 62) reported that the knotty core in E. nitens was effectively restricted by pruning in four lifts to 2, 4, 6, and about 8 m at ages 2, 3, 4, and 6 years respectively (2003, p. 65). Recovery of knot-free timber averaged 50% of log volume (McKenzie et al., 2003, p. 72). The implication is that pruning must be practiced properly for high recoveries of clearwood.
As log diameter decreases, the percentage of the log taken up by defect core increases, leading to lower recoveries of higher grades from smaller pruned logs (Washusen et al., 2008, p. 6). Thinning is practiced to reduce crop stocking and increase remaining tree diameters. Washusen et al. (2009, p. 52) found that sawn recoveries as a percentage of volume increased with larger tree diameters and suggested that thinning regimes targeting larger diameters might improve grade recoveries (p. 50).
Silvicultural practices potentially increase sawn timber recoveries and grade recoveries and therefore could improve returns to the grower. Final crop stocking for sawn timber production should be low enough to ensure pruned logs mature into diameters large enough to maximise log residual value. However, to date no attempts have been made to optimise final crop stocking and rotation length for solid timber production from pruned E. nitens for specific processing equipment. This study aimed to produce costs and revenues along with graded sawn recoveries for the range of tree diameters present in the case study stand.
The literature points to a number of factors that could influence grade recoveries according to log position, suggesting that log position appears to be an important variable for predicting revenue from logs.
With the exception of end-splitting, which was observed to increase in the second log, Washusen et al. reported that upper logs posed fewer processing issues than did buttlogs (2009, p. 52). This implies that upper, unpruned logs might produce profitable sawn grade recoveries. Height in tree was also found to influence both surface and internal checking levels, with less checking occurring at 6 m height than 0.5 m above the ground (Blakemore et al., 2010, p. 32). Collapse has consistently been found to be worse lower in the tree (McKenzie et al., 2003, p. 72; Purnell (1988) as cited in Shelbourne et al., 2002, p. 360). Deflection (movement off the saw), however, was reported by Washusen et al. (2008) to be higher from second logs than from buttlogs (p. 41). Washusen et al. also found that radial shrinkage increased significantly from the first pruned log to the second pruned log (2009, p. 49) and suggested this could result in more skip in boards sawn from the second log.
Because log position appears to be an important variable for predicting revenue from logs, there is a need to examine the effect log position has on recoveries.
Investigating best practice processing methods that maximise sawn timber and grade recoveries while also minimising production costs would be required to demonstrate economic value from E. nitens sawn timber production.
To date, research into production of solid timber from E. nitens has primarily examined wood quality issues, processing systems and the relationship between these two. Although a range of studies have quantified sawn recoveries and grade-limiting defects from young, pruned plantation E. nitens in both New Zealand and Australia, to date only Innes at al. (2008) published estimates of product value per cubic metre of sawlog and per hectare for plantation E. nitens in the public domain.
Innes et al. (2008, p. 1) reported on the profitability of using existing traditional and contemporary Australian native forest eucalypt sawing and drying systems as a first step in exploring utilisation of sawn plantation E. nitens. However, Innes et al. identified a range of shortcomings in the production techniques used in their study that resulted in high levels of checking, knots, collapse and distortion (2008, p. 23), thus generating low overall product values (2008, p. 39).
Best practice sawmill methods are those that have been documented to minimise degrade in sawn products. However, there is little published research that reports improvements in economic value from processing E. nitens into solid timber products from best practice methods. In order to identify methods most suitable for producing greatest returns from E. nitens sawn timber, this section will consider traditional processing methods, then improvements to these before considering contemporary methods.
Traditional processing of ash eucalypt
Cold climate eucalypt species such as the ash group have a set of processing challenges that if not carefully managed result in low grade recoveries, particularly with smaller log diameters such as from the Australian second growth resource.
Conventional single sawing methods, as traditionally practiced on old growth Australian native ash species, can have high processing costs and low nominal sawn recoveries. Washusen (2011, p. 8) attributed these inefficiencies primarily to large saw kerfs, large allowances for green board oversizing and slow throughput because of reciprocating carriages and regular log turning. These inefficiencies are compounded as log diameter and length is reduced (2011, p. 8).
In order to avoid skip and undersizing of products, conventional processors of native forest ash eucalypts in Australia target a mean green board thickness as high as 31 mm to produce dried boards with a nominal thickness of 25 mm (Washusen et al. 2006, cited by Washusen, 2011, p. 8). This oversizing of green board thickness is practiced to eliminate the presence of skip on the faces of boards dressed to nominal sizes, but has the consequence of reducing nominal recoveries as a percentage of the log volume.
Issues with traditional processing are exacerbated with production of appearance timber from plantation E. nitens. Washusen et al. (2009, p. 53) practiced conventional ash sawmilling methods on E. nitens and attributed high levels of sawn board defects to shortcomings in conventional processing practices, including poor sawing accuracy, inappropriate weighting of drying stacks, lack of control of drying rate in ambient conditions and steam reconditioning applied at sub-optimum moisture contents. Washusen et al. (p. 53) concluded that both sawing and drying strategies needed to be improved for the plantation resource.
Improvements to traditional processing methods
Washusen (2011, pp. 1, 23) suggested that improvements to sawing strategies should include better sawing accuracy and sawmill efficiency, together with “correct oversizing” of green sawn boards. Suggested improvements to drying strategies included better control over moisture gradients in the boards, optimised steam reconditioning and correct weighting of drying stacks (Washusen, 2011, pp. 1, 23).
Innes et al. (2008, p. vi) concluded that for high value sawn products to be produced from E. nitens plantations, the sawing equipment would need to be optimised for the plantation resource, with specific processing techniques required to control distortion, collapse and checking. Innes et al. observed that dry recoveries varied considerably between sawmilling technologies (2008, pp. 18, 38) and concluded that for economic production of sawn appearance grade timber from plantation pruned and thinned E. nitens, both sawing and drying techniques would need to be improved to provide greater control over distortion and checking (2008, p. 41).
Contemporary sawmill technologies
Washusen suggested that sawmill efficiencies could be improved from traditional methods as practiced on old-growth ash eucalypt by utilising twin and multi saws to symmetrically release stresses on opposite sides of the log or flitch at once (Washusen, 2011, p. 9). Currently available sawmill systems utilising this technology are restricted to small maximum log diameters, including 45 cm for Whittaker’s Timber Products small log line (Washusen, 2011, p. 10), 25cm for the Hewsaw R200 and 34 cm for the Hewsaw R250 (Washusen, 2011, p. 12). However, small sawlog diameter has been identified as a factor that reduces sawn timber and grade recoveries.
In an attempt to improve processing systems for E. nitens, Blakemore et al. (2010) applied contemporary sawing, drying and reconditioning schedules to E. nitens timber. Boards were sawn on a Hewsaw, a multi-saw linear flow system designed for softwood sawmilling that produces flatsawn boards (Washusen, 2011, p. 16). The Hewsaw removes wood simultaneously from around the log by using chippers ahead of symmetrically oriented multiple circular saws, allowing production of long-length (5 m) flatsawn timber. Sawing longer lengths reduces end splitting as a proportion of green sawn timber (Washusen, 2011, p. 14) and sawing costs are low (Table 1, Washusen, 2011, p. 16), provided an operating log throughput of 120,000 cubic metres per annum is available (Washusen, 2011, p. 12). However, because the sawing pattern cannot be altered for logs of varying diameters, mean green sawn recoveries were reported to be less than 40% for the flatsawing strategy (Blakemore et al., 2010, p. 2), with percentage recoveries reducing as log diameters increased.
Washusen (2011) reported that no attempt was made to evaluate grade recoveries as a percentage of log volumes from two Hewsaw trials sawing E. nitens, despite acknowledging that a large proportion of the boards produced would contain pith, which is associated with drying degrade (2011, p. 15). Grade recoveries were not quantified, nor were product values assessed against production costs in these trials.
Issues with the Hewsaw system include limited grade recoveries because of proportionally large defect cores from the small log diameters, the wide flatsawn boards would be subject to cupping and checking defect and because centre boards contain pith, these would be severely devalued for appearance applications. Furthermore, for production efficiency to be optimised the log diameter range is very narrow.
Satchell and Turner reported low levels of defect and high grade recoveries from sawmilling and sizing 18 year old Eucalyptus regnans, using a sawmilling method developed for small diameter eucalypt (Satchell & Turner, 2010, p. Results). The method involved accurate placement of logs for initial saw cuts, skilled judgment calls by the sawyer, predefined log lengths according to diameters, sawing faces before edges and accurate edging for grade recoveries based on visual cues on board faces. The technique involved cutting narrow quartersawn boards in preference to wide flatsawn boards.
Controlling for defect and degrade appears to be essential to achieve adequate grade recoveries from E. nitens and study design requires attention to minimising avoidable defect and implementing sawmilling methods that maximise log revenue and minimise costs. By identifying then applying methods that are current best practice, this study approached profitability as a performance benchmark, with cost and grade recovery results that any processor using the case study equipment and methods could expect.
Specific methods applied to sawing eucalypt have been studied and the literature identifies a range of issues with these and how they influence production efficiency and recoveries.
To efficiently convert E. nitens logs into sawn timber product, the literature draws attention to three main issues: sawing pattern, growth stresses and log end-splitting. Each of these is considered in turn.
Sawmill pattern will determine the quantities produced of two distinct types of boards, quartersawn and flatsawn. Strategies tend to target one type of board or the other.
There is widespread disagreement among Australian researchers on the comparative merits of quartersawing versus flatsawing as processing strategies.
In Australia native forest ash eucalypt is generally quartersawn to minimize drying defects and improve the resulting products’ stability in service (Blakemore & Northway, 2009, p. 4; Washusen et al., 2008, p. 7). Conventional sawmills prefer log mid-diameters greater than 40 cm for quartersawing (Washusen et al., 2009, p. 41) and “it is well understood that quarter-sawing is a poor sawing strategy for small diameter eucalypts” (Washusen, 2011, p. 6).
McKenzie et al. (2003, p. 71) reported that for a mean log diameter of 480 mm, recovery of quartersawn E. nitens timber averaged 50% (2003, p. 72). Although these recoveries are high relative to Australian studies, the methods used for measuring recoveries were poorly reported and it is not clear whether the recoveries were green sawn, nominal or dry, or whether recoveries included end-splits. Straightening cuts were applied in Mckenzie et al. to remove crook when ripping boards from slabs (2003, p. 75). The log length was 5m which required application of regular face cuts to the residual log. Slabs were then reduced to half that length prior to ripping (2003, p. 72). This strategy yielded high sawn recoveries but costs were not reported.
Cost-efficiency is an important consideration when assessing profitability of a sawmilling method because marginal revenue needs to exceed the marginal cost of implementing improvements in recoveries in order to maximise log residual value. Strategies consider the tradeoff between increasing the cost of sawing for higher recoveries and the value of sawn timber recovered.
Flatsawing is a lower cost strategy than quartersawing but the risk is lower grade recoveries. The additional cost of quartersawing may be justified if returns are improved. Although Nolan et al. suggested E. nitens should be quartersawn even as small diameter logs (2005, p. 28), there has been a strong preference to flatsaw E. nitens among contemporary Australian researchers, primarily because flatsawing produces a consistently higher green-sawn percentage recovery from small diameter logs (Blakemore & Northway, 2009, p. 4). However, producing larger volumes at lower cost may not result in higher revenue. Washusen et al. (2008, p. 2) reported that product values were significantly higher per cubic metre of log sawn using a quartersawing strategy, despite lower volumes produced. This was because a high incidence of surface checking significantly devalued the flatsawn timber.
Another issue influencing sawn timber value is board width. For a given log diameter a flatsawing strategy produces a greater proportion of wider boards (Washusen et al., 2008, p. 7). As log diameter decreases, quartersawing becomes less cost-efficient (Washusen et al., 2008, p. 30), while quartersawn boards also become narrower. Based on the perception that value per cubic metre increases with board width (Washusen et al., 2008, p. 7), Washusen et al. chose a strategy of flatsawing the widest boards possible from pruned plantation E. nitens in an attempt to maximize both board width and board grade, therefore value (2008, pp. 2, 17). However, wide flatsawn boards are subject to cupping (Washusen et al., 2008, p. 42), which then results in skip defect. Washusen et al. (2009, p. 53) reported that flatsawn boards from buttlogs had mean cupping of over 2mm. To remove this level of cupping from both the face and back of a board, 5mm of wood would have required to be removed from the thickness (Washusen et al., 2009, p. 53). This would have reduced the product volume significantly and therefore log revenue. It is notable that cupping has less impact on dressed board thickness in both narrower flatsawn boards and quartersawn boards than in wide flatsawn boards.
Tasmanian plantation E. nitens timber has approximately twice the amount of tangential shrinkage than radial shrinkage (Innes et al., 2008, p. 41). As a result, flatsawn boards are much more likely to cup during the drying process and move in service when exposed to moisture variation than quartersawn boards (Blakemore & Northway, 2009, p. 4; Kingston and Risdon, as cited in Nolan et al., 2005, p. 28). Washusen et al. (2008, p. 43) found that width shrinkage was significantly lower in quartersawn E. nitens boards than in flatsawn boards. This anisotropic shrinkage leads to greater stresses on flatsawn board faces than on quartersawn board faces, resulting in checking and cupping in flatsawn boards (Blakemore & Northway, 2009, p. 4). Furthermore, restraining this cupping potential in the drying stack could lead to an increase in these tension stresses, causing further surface checking (Blakemore & Northway, 2009, p. 4). The implication is that narrow flatsawn boards, with greater freedom to move and lower stresses at work on the board face, might check less during drying than wider flatsawn boards.
Collapse on the face of flatsawn boards is likely to be expressed as checking (Blakemore & Northway, 2009, p. 4). Although collapse can be severe on quartersawn board faces, this is likely to be expressed as washboarding rather than checking (Blakemore & Northway, 2009, p. 4). Checking is a serious value-limiting defect, whereas collapse has no impact on value if removed when profiling the timber.
A range of issues need to be considered in the design of sawmilling methods that minimise degrade, maximise production volumes, minimise costs and produce a quality sawn product. Although quartersawn boards are likely to have lower levels of checking than flatsawn boards and are more stable in service, log diameter affects the cost efficiency and production efficiency with which quartersawn boards can be produced. Study design should use methods that take into account the tradeoff between sawn board quantity and sawn board quality from the log in order to maximise residual value from the sample sawlogs.
Satchell and Turner (2010) evaluated a hybrid pattern (See Appendix 5) and reported relatively low costs and high grade recoveries for the scale applicable to this study from small-diameter E. regnans. Average small end diameter was 31.8 cm with a diameter range of 25 – 43 cm. This method was selected as the most suitable for sawing the case study E. nitens logs because:
- E. nitens has a propensity for high levels of surface checking. Quartersawn output was desirable and innovations allowed cost-efficient production of quartersawn output; and
- average log diameters from the case study stand (32.94 cm) were smaller than required for traditional quartersawing patterns (>40 cm).
The outer section or periphery of a eucalypt log is in longitudinal tension, with a stress distribution progressing to longitudinal compression in the core (A. N. Haslett, 1988, p. 12). Sawing of eucalypt logs and release of stresses results in bent flitches and curvature in the residual log. Regular face cuts may be necessary to produce even thickness flatsawn boards, while edge cuts are necessary to produce straight quartersawn boards. These additional saw cuts result in recovery losses (McKenzie et al., 2003, p. 63) and increased costs.
McKenzie et al. (2003, p. 72) reported that recoveries of quartersawn timber were inversely related to growth stresses in the logs. Larger growth stresses resulted in more distortion of sawn surfaces. Therefore to produce straight boards, straightening cuts removed greater curve from slab edges, consequently reducing sawn timber recoveries.
The radius of curvature (also called ‘deflection’) tends to be greater when sawing smaller diameter logs because stress gradients decrease as diameter increases (Nolan et al., 2005, p. 29; Shield, 1995, p. 134; Washusen et al., 2009, p. 6).
When producing quartersawn material, as log length increases more wood is removed from both edges to straighten the resulting board or cant from ‘spring’ curvature. Sawn recoveries reduce per log cubic metre as log diameter declines, provided log length is constant (Nolan et al., 2005, p. 30). Reducing log length counteracts the effect movement has on recoveries from smaller diameters. However, the consequence is that sawing costs increase because there is more handling for a given volume of logs.
As log length increases, producing even thickness flatsawn material with a single saw requires face cuts of increased thickness on the residual log, flitch or cant to straighten the face. The tradeoff is that production efficiency improves by flatsawing longer logs.
Some contemporary sawing systems such as twin or multiple sawing lines can overcome thickness variation in flatsawn boards by sawing simultaneously on opposite sides of the log (Shield, 1995, p. 136) to symmetrically release stresses on both faces. This allows longer flatsawn lengths to be milled without requiring face cuts, thus improving sawmill efficiency. Crook distortion on edges of quartersawn material is not overcome (Innes et al., 2008, p. 24), because only stress on board faces is relieved by symmetrical cuts. Edge straightening cuts of quartersawn boards over long lengths remain at “the expense of substantial recovery loss” (Blakemore et al., 2010, p. 26).
Washusen et al. (2008, p. 41) reported that slabs quartersawn from the centre to the periphery distorted significantly more than the half log from which the slab was cut. The explanation given by Washusen et al. (2008, p. 41) was that “distortion increases as the sawing process continues”. A more plausible explanation is that half logs do not undergo full stress release between the pith and the periphery perpendicular to the saw cut halving the log. This is because half-log deflection induces counteractive stresses at the peripheries adjacent to the saw cut that were not under tension in the direction of the deflection. This resulting counteractive stress would constrain the full release of tension. Once slabs from this central area are sawn from the residual log they release their tension freely. As the width of a quartersawn slab increases, the stress gradient will be steeper, resulting in increased curvature expressed as crook (Nolan et al., 2005, p. 29). Sawmilling strategies can take advantage of counteractive stresses by releasing these into slabs before applying straightening cuts, thereby increasing production efficiency.
It is important to understand that the radius of deflection is not influenced by log length. However, the level of deflection does increase as the log gets longer. The length tradeoff is therefore between:
- value lost from short logs caused by short board lengths, higher sawing costs and end-splits; and
- value lost from long logs as sawn recovery lost from straightening cuts.
In summary, distortion resulting from stress release does have a negative impact on log residual value. Removing distortion to produce a high quality straightened sawn product reduces sawn percentage recoveries and increases production costs because of the larger number of cuts required. This is a ‘fact of life’ with eucalypt sawmilling and indirectly influences log residual value because there are costs and benefits with different sawmilling approaches.
As log diameter decreases, resulting lower sawn recoveries and higher costs imply an increasingly negative impact on log residual value, compounded by the shorter log lengths required for adequate sawn recoveries.
To ensure sawn recovery remains high as a percentage of the log, log length must be restricted in proportion to the diameter because quartersawn slabs experience edge curvature and with single saws flatsawn slabs experience face curvature. Larger diameter logs, because these move less in proportion to their diameter, can be sawn at longer lengths.
The sawmill method described in Satchell and Turner (2010) produces straightened boards by first sawing slabs and then edging these. Although log length could potentially be optimised according to diameter for greatest residual value, for this case study log length was set to 3.0 m as industry best practice for the diameter range being sawn. Movement off the saw can be easily measured with this sawmill pattern because logs are halved before other saw cuts are made. This offered the opportunity to quantify the effect movement had on sawn recoveries according to log diameter and log position at a standard length of 3.0 m and movement was measured for each sample log.
Log end splitting and log length
Increased levels of log end splitting reduce grade recoveries. Factors influencing levels of end-splits have been reported in the literature.
Sawn timber volume losses from end splitting have been found to be greater in flatsawn boards than in quartersawn boards and to increase progressively with tree height (Washusen et al., 2008, p. 40).
Longer log lengths can reduce volume losses attributable to board end splitting (Blakemore et al., 2010, p. 2). Washusen reported only 1.2 – 2.9% loss in green sawn recovery caused by board end-splits from sawing 5 m log lengths with a Hewsaw (Washusen et al. cited by Washusen, 2011, p. 14).
Satchell and Turner (Satchell and Turner 2010, p. Results) reported end splitting losses of only 1.4% of nominal sawn recovery, from machine harvested small diameter 6m E. regnans logs, cross cut in half to an average log length of 3 m immediately prior to sawing and sawmilled within 28 days of harvesting.
End-splits increase with time after cross cutting of logs. If logs are sawn immediately after cross cutting, end splits are not likely to impact on recoveries to any significant extent, regardless of log length. Therefore best practice is to saw logs as soon as practicable after harvest, to avoid defect resulting from end-splits that would otherwise impact on log residual value.
Although pruning is considered to be essential to overcome defects associated with branches and to produce high value appearance clearwood from plantation buttlogs (Washusen, 2011, p. 3), pruned logs do not necessarily produce high recoveries of clearwood.
Innes et al. (2008, p. 23) reported the presence of knots to be the primary reason for downgrading boards from pruned logs. Inclusion of knotty core in boards from pruned logs can be seen in plates 5-8 (Innes et al., 2008), indicating that edging was not practiced properly to target clearwood from pruned material.
Grade sawing involves a tradeoff between volume production and higher grade recoveries. Sawmilling best practice involves judgment calls aimed at producing grade recoveries that maximise the value of sawn timber produced from the log. Where practicable, sawmilling methods should exclude knotty core from boards. This can be achieved by visually assessing where slabs are to be edged and taking care to edge out knotty core for production of clearwood.
Edging of unpruned logs also requires similar judgment calls to target higher grades by excluding corewood and where practicable excluding knot defect.
Sometimes knot defect can be excluded by edging and other times this is best docked out from the board length. Knot defect inevitably reduces grade recoveries from headlogs, however the impact knots and pruning have on E. nitens log value is not well understood. By examining relationships between levels of knot defect in boards and log attributes such as log position and log diameter, improved forest management practices could result that lead to higher log prices for the grower.
Summary and proposed methods
A range of methods are available to the sawmiller that can potentially reduce defect in the resulting sawn timber. Sawmilling pattern can affect volume recovery, costs and grade outcomes along with levels of checking that develop during drying. Growth stresses and end splits can reduce sawn timber recoveries and increase costs for producing these. Shorter log lengths reduce recovery losses caused by growth stress, but at the expense of increased sawing costs. End splits are unavoidable but defect resulting from these can be minimised by sawmilling soon after cross cutting logs.
This study aimed to use best practice methods identified in the literature for converting case study logs into sawn timber, in order to maximise log residual value and therefore returns to the grower. Sawmilling method as practiced in Satchell and Turner (2010) was selected, with 3 m long lengths sawn soon after cross cutting to length to minimise defect from end splits.
This study also aimed to quantify knot defect present in boards from pruned buttlogs and unpruned headlogs, along with recoveries of sawn timber to examine the effect knots have on grade recoveries and log residual value.
In order to produce high-quality appearance hardwood products from plantation E. nitens logs, significant levels of defect must not develop in the drying process (Washusen, 2011, p. 18). To avoid levels of drying degrade that impact seriously on economic value of sawn products, E. nitens timber must be dried slowly and with care (Nolan et al., 2005, p. 31), particularly in the early stages (Nolan et al., 2005, p. 32).
Washusen rated checking as the drying defect having the greatest impact on product quality (2011, p. 19). Blakemore and Northway (2009, p. 45) identified checking as “the major limitation for processing pruned plantation grown E. nitens into appearance grade products”.
Drivers of checking degrade
Initial drying that takes place too rapidly has been associated with increased internal checking and collapse in E. nitens (Nolan et al., 2005, p. 31).
Normal shrinkage (i.e. shrinkage not associated with collapse) across a tangential board face can cause high levels of stress to develop, leading to surface checking (Washusen, 2011, p. 19). Jacobs (as cited in McKenzie et al., 2003, p. 63) reported that collapse in the flatsawn face of a board can “show as heavy open checks with distortion of the surface as well”.
Collapse shrinkage, unlike normal shrinkage that occurs below fibre saturation point, is also implicated in much of the surface and internal checking experienced in low to medium density eucalypt species (Blakemore & Northway, 2009, p. 9; McKenzie et al., 2003, p. 63; Washusen, 2011, p. 19).
Shelbourne et al. (2002, p. 378) reported that checking in kiln dried E. nitens disks was representative of checking in sawn boards and found that:
- Checking levels appear to be variable more at the individual tree level than provenance or site, with a large range of checking levels evident between trees (2002, p. 373);
- checking levels decrease with tree height (2002, p. 378); and
- checking was more frequent in the transition wood zone (2002, p. 371).
Internal checking in E. nitens may not be related to tree age, with high levels of checking reported by Yang and Waugh (as cited in Shelbourne et al., 2002, p. 360) in 15, 25 and 29 year old trees.
Levels of internal checking in dry E. nitens boards increase with board thickness (Blakemore et al., 2010, p. 18). Blakemore et al. (as cited in Washusen, 2011, p. 22) reported that thin section quartersawn E. nitens could be processed to be virtually free of surface and internal checking. As board thickness decreases, earlywood collapse may express as washboarding rather than internal checking (Blakemore & Northway, 2009, p. 4).
Washusen et al., using conventional native forest ash eucalypt processing strategies, found the incidence of surface checking to be high (2008, p. 45). Haslett & Young reported high levels of checking in quartersawn boards from 30 year old E. nitens timber kiln-dried after what they described as “careful air drying” (T. Haslett & Young, 1992, p. 8). However, Haslett and Young (1992, p. 9) reported that twisting of boards during drying was also a serious problem, suggesting that drying was not practiced carefully. This is because twisting is preventable and well known to be caused by either not weighting the stack properly or drying the wood too fast. Because no detailed description of the drying process was outlined in the report, neither the timber nor the level of care can be clearly implicated as causing these poor results.
Innes et al. proposed further research to address the drying of E. nitens (2008, p. vii) because using current Australian industry drying methods, 15-40% of boards from pruned and thinned E. nitens buttlogs contained significant levels of internal checking (2008, p. 40). However, review of subsequent literature has not revealed any clear understanding of drivers that cause checking, nor methods to prevent the problem. Indications from the literature are that to limit checking to levels that do not impact heavily on economic value, current best practice includes:
- Slow drying of timber, especially during the early stages of drying;
- quartersawing where practicable; and
- limiting board and product thickness.
This study did not seek to develop specific methods for reducing checking, nor to quantify impact on economic value of tree age or variation between individual trees for the species. The intention was to undertake best practice methods identified in the literature, then assess their influence on case study log and stand residual value.
This study practiced methods that were intended to minimise levels of checking given what is currently understood of the issue. Product thickness was sawn to 28 mm and timber was slow air-dried.
Log position and checking
Checking of sawn timber from E. nitens buttlogs has been reported as the most serious value-limiting defect in several studies (Innes et al., 2008; McKenzie et al., 2003, p. 76; Washusen, 2011; Washusen et al., 2008, 2009). Innes et al. (2008, p. 23) reported that levels of checking were higher in pruned buttlogs (the primary reason for downgrade in 25-30% of boards) compared with unpruned upper logs (the primary reason for downgrade in 8-10% of boards) from the same trees. Blakemore et al. (2010, pp. 3, 31) reported that levels of both surface and internal checking decreased with height in the sawlog and also found that unpruned buttlogs yielded greater levels of internal checking than pruned buttlogs, especially before reconditioning. Washusen et al. (2008, p. 2) found that upper pruned logs produced significantly higher product values per cubic metre of log input than lower pruned logs, primarily because of reduced surface checking.
The importance of log position has been clearly identified in the literature as influencing levels of checking. This study aimed to quantify levels of checking according to log position to examine the impact log position has on log residual value.
Steam reconditioning and checking
Traditional processing of Australian native forest eucalypt involves steam reconditioning once the moisture level within the boards reaches fibre saturation point (Washusen et al., 2008, p. 20). Contemporary reconditioning strategies have been developed in Australia where greater collapse recovery is achieved by reconditioning at lower moisture contents than traditionally used (Blakemore & Langrish, 2007).
Blakemore et al. (2010, pp. 31, 42) reported that the prevalence of visible internal checking was dramatically reduced from that reported previously by reconditioning only once board moisture content was below 20% (2008, p. 18) or fibre saturation point (2008, p. 20). However, closed checks (i.e. those that were not visible) were not reported (Blakemore et al., 2010).
Checks can be ‘closed’ by steam reconditioning, which makes them less visible (Blakemore et al., 2010, p. 2). However, although the visibility of checks can be significantly reduced (Blakemore et al., 2010, pp. 41, 42), “The impacts of closed checks in reconditioned sawn boards from plantation-grown E. nitens in a range of downstream manufacturing processes and in product service needs to be determined.” (Blakemore et al., 2010, p. 3). Closed checks on the surface of a board are visible in finished products and “the closed-up hairline cracks remain” (Blakemore et al., 2010, p. 44). These could open up later in service or ‘feather’ during secondary processing, resulting in defective product (Blakemore & Northway, 2009, p. i). The impact closed surface checks could have on appearance product values has not yet been studied, but closed surface checks could potentially be a worse problem than open surface checks which are clearly visible and result in downgrading of the material (Blakemore & Northway, 2009, p. i). Washusen (2011, p. 1) identified this issue as an important knowledge gap in need of further research. Blakemore and Northway (2009, p. 16) described it as follows:
Conventional industry wisdom is that such closed checks pose a serious problem, in that if exposed when machining or moulding is undertaken, they will result in a feathering effect on the surface of the product. This may not be a significant problem for products such as quartersawn flooring, but it is a problem for backsawn products and cabinetry components such as high-value kitchen cupboard doors.
It can be concluded that reconditioning does not actually remove checking degrade but does confound attempts to measure levels of checking in research experiments. A standardised approach to quantifying levels of checking would allow future research to potentially compare conflicting results and isolate the causes of checking.
Direct comparisons of checking degrade between processing studies would require either an identical reconditioning process or none at all. By describing closed checks as “not visible” (Blakemore & Northway, 2009, p. 16), this implies that closed checks may not be visible to graders or researchers. Thus true levels of checking degrade resulting from the processing treatment being tested would not be measurable. In contrast, by not steam reconditioning timber, levels of degrade caused by surface checking such as length and width of checks or their prevalence could be accurately measured and quantified. However, assumptions would need to be made on levels of skip caused by collapse.
Collapse degrade and steam reconditioning
Collapse shrinkage, unlike normal shrinkage that occurs below fibre saturation point, is considered to be severe ‘abnormal shrinkage’ because it occurs in timber above fibre saturation point (Blakemore & Northway, 2009, p. 9).
Collapse shrinkage of E. nitens can be recovered by steam reconditioning. Up to 95% of collapse shrinkage in air-dried boards can be recovered (Blakemore & Northway, 2009, p. 13), effectively reducing the levels of shrinkage in the board. This allows green sawn sizing to be reduced without an increase in skip defect, resulting in increased nominal sawn recoveries. The extent to which boards should be oversized in the sawing process would thus be dependent on whether the boards were to be steam reconditioned. The cost of steam reconditioning should be economically justifiable by the improved nominal recovery and value that results from reduced green sizing.
The case study timber was not steam reconditioned. This allowed for accurate measurement of levels of checks on profiled surfaces. Although the green thickness of 28 mm was selected as industry best practice for eucalypt species in New Zealand, the optimum thickness for highest residual value for unreconditioned E. nitens was unknown. Therefore a scenario was designed to enable an economic comparison to be made between residual value for unreconditioned and reconditioned timber, assuming that all collapse would have been recovered sufficiently if reconditioned to have been profiled without exhibiting skip.
This approach also allowed for an evaluation of skip defect on the profiled case study boards given the chosen green thickness, while also allowing for the importance of steam reconditioning to be quantified in economic terms.
Summary and approach to drying
In conclusion the literature is not currently clear about the degree to which levels of checking can be predicted or explained in sawn E. nitens timber. This study aimed to develop and document methods for measuring checks suitable for standardising and applying to future research, to facilitate improved research methods examining causes of checking. To date methods for defining or measuring checks have not be standardised.
Methods selected as best practice to minimise levels of checking were to saw the timber to 28 mm green thickness and air dry the timber as slowly as practicable.
This study aimed to:
- Apply best practice drying methods to reduce levels of checking as much as practicable;
- quantify levels of checks on unreconditioned profiled surfaces and examine their impact on sawn timber value; and
- assess the economic impact of steam reconditioning on log residual value.
Two drying methods are applied to drying appearance hardwood timber. These are:
- Air drying followed by kiln drying; or
- kiln drying from green.
Air-drying is the predominant method used in Australia for drying eucalypt species that are known to require slow drying (T. Innes, pers. comm.).
McKimm et al. (as cited in Shelbourne et al., 2002, p. 360) found air drying of 20 year old E. nitens followed by kiln drying resulted in less internal checking than kiln drying from green.
The largest sawmill in Australia processing ash eucalypt (Heyfield, Victoria) exclusively uses air-drying and wraps stacks with permeable cloth in warmer periods of the year (T. Innes, pers. comm.). Air flow, temperature and humidity all influence the rate of drying and during drier, warmer periods of the year wrapping freshly sawn timber stacks slows the rate of drying and avoids excessive degrade (T. Innes, pers. comm.). Air-drying appeals to producers because it is low cost and does not require energy nor high capital input (T. Innes, pers. comm.). Priest et al. (as cited in Bekele, 1995, p. 5) reported that 25 mm thick Eucalyptus grandis air dried with less degrade than kiln dried material and produced a more uniformly dry material when finished in a kiln than material dried from green in a kiln. Gough (as cited in Bekele, 1995, p. 5) also found that best results were obtained from finishing air-dried Eucalyptus timber in a solar kiln. In contrast Washusen et al. (2000, p. 7) reported that slow air-drying of 40 mm thick E. globulus produced high levels of degrade similar to E. globulus that was slow kiln-dried from green. McKimm et al. (as cited in Shelbourne et al., 2002, p. 360) found air drying of 20 year old E. nitens followed by kiln drying resulted in less internal checking than kiln drying from green.
The literature points to conflicting results in the few reports available that compare the two standard methods for drying Eucalyptus timber. It was decided that the risk was too high to dry the case study timber from green in a kiln and to use Australian industry standard practice methods for ash eucalypt and slow air dry the timber, followed by finishing in a kiln.
The three elements that control drying rate are air temperature, relative humidity and airflow across the timber surfaces (Langrish & Walker, 2006, p. 1). Unlike temperature and humidity, controlling air flow can be achieved at low cost, such as by wrapping with permeable cloth.
Two methods were available for air drying the case study timber, either drying in a protected environment such as a ventilated shed, or drying outdoors. No literature was available suggesting which would produce the better grade recoveries or pointing to which method would produce higher log residual value. It was assumed that wrapping of stacks would benefit the drying process and it was decided that both air-drying methods should be tested in order to compare resulting sawn timber value.
Price is one of the main determinants of profitability in an investment analysis. Because Eucalyptus nitens is not sawn commercially in New Zealand, market transaction data for sawn timber is not currently available for estimating log value.
Satchell and Turner (2010) used the residual value method for estimating log values for 18 year old Eucalyptus regnans in New Zealand from sawn timber products, because there was not an established market in New Zealand for sawlogs. Board prices were estimated from those for more commonly available eucalypt species.
In Australia, Innes et al. (2008) utilised the residual value method for estimating E. nitens plantation profitability using native forest eucalypt sawing and drying systems. More recently Forestry Tasmania produced a commercial in confidence report on profitability of plantation E. nitens using the residual value approach to value sawlogs (Pearn et al., 2013). Price data for similar available species were used for the appraisals in these studies because E. nitens timber was not yet available on the market in Australia from which to derive prices.
E. nitens is not a recognised hardwood species in New Zealand timber markets. Consequently, market prices for products are not available. Issues and opportunities for pricing E. nitens sawn products are outlined below.
Pricing E. nitens sawn timber products
Hardwood solid timber product values have traditionally been linked to grade, board thickness, width and length (Nolan et al., 2005, p. 7). Innes et al. used product prices for native regrowth ash eucalypt in assessing the economic viability of processing plantation E. nitens, with high arbitrary discounts applied to lengths shorter than 1.8m when pricing E. nitens timber (2008, pp. iv, 9, 10). Another published Australian study on economic viability of E. nitens solid timber products discounted board lengths of less than 3m by 10% while boards of less than 1.8 m were discounted 50% from current hardwood wholesale prices (Washusen, 2011, p. 14). No basis for these discounts were provided in the reports and no evidence is available to suggest that the market discounts lengths less than 1.8 m to these levels. Sawn appearance timber is usually available in New Zealand in random lengths, meaning that short lengths are mixed with longer lengths in the packet of timber. Although this suggests that average piece length may be more important to customers than discounts for shorter lengths, the level to which length influences price for appearance timber products is not evident in the literature, nor in the market. Arbitrary discounts do not adequately reflect true market value for the levels of the quality being discounted, suggesting that empirical methods in use for pricing of quality characteristics in products new to market could be adopted as an improved method for pricing E. nitens timber.
In an attempt to produce credible estimates of market prices for the range of products sawn, market survey methods were developed as part of this study to estimate market prices for E. nitens sawn products in the absence of market sales data. These methods were designed to overcome some of the inadequacies of previous work examining profitability of growing E. nitens.
Plantation E. nitens timber is not currently available, nor marketed in New Zealand. Plantation ash eucalypt is available inconsistently and only in small quantities. A recent survey of specialty timber merchants and users by Future Forests Research (unpublished, 2012) described market recognition and availability of plantation ash eucalypt in New Zealand as negligible.
A well known and established species in the marketplace tends to hold consumer preference. This preference generates price premiums and is based on the perception of suitability for purpose (Nolan et al., 2005, p. 8). Established products and species can be resilient and hold strong loyalties, while new products may be accepted only slowly (Nolan et al., 2005, p. 80). For example, hardwood has traditionally been available in long clear lengths but these are likely to become increasingly difficult to obtain into the future (Innes et al., 2008, p. iv). Substitution of long lengths for shorter lengths might meet market resistance because of the perception that installation costs increase. Market penetration and product acceptance could be slow, even if end matching of short floorboard lengths offered low installation costs. Intangible qualities such as reputation could potentially be estimated as price discounts or premiums. hereThis study surveyed flooring timber experts in New Zealand and asked them to quantify the influence on price of four intangible qualities of flooring timber.
Product substitution and price
High quality logs of extant old-growth hardwood timber species are expected to become increasingly scarce into the future (Chen & Wood, 2011, p. i). Although opportunities might arise for product substitution based on demand for old growth timber not being fully supplied, estimating such demand over time would remain speculative.
Market acceptance, demand and price of a product new to market might best be predicted by comparing quality with established products and soliciting market feedback on suitability for purpose.
Products with quality differentiation tend to hold higher value in the market along with less price volatility than commodity products (Nolan et al., 2005, p. 80). Seasoned appearance hardwood products are differentiated and include those used for decorative structural applications (Nolan et al., 2005, p. 5).
Both physical and appearance characteristics determine whether a species or product is suitable for a differentiated end-use. Matching wood quality characteristics with technical requirements for specific timber products would be necessary to offer quality differentiated E. nitens products (Nolan et al., 2005, p. 91). Physical properties such as movement in service for appearance products, surface hardness for flooring products and strength/stiffness for structural products, could each influence market price of E. nitens timber for the differentiated application. These quality characteristics could also be compared between species when pricing a product new to market.
Innes et al. used product values derived from native regrowth ash eucalypt for valuing plantation E. nitens timber and considered plantation E. nitens to have the potential to meet market requirements currently satisfied by Australian native forest ash eucalypt (Innes et al., 2008, p. iv). Washusen et al. also considered plantation E. nitens timber suitable for meeting market requirements currently satisfied by Australian native forest ash eucalypt (2008, p. 4). However, Blakemore et al. considered the appearance of plantation E. nitens timber to vary from Australian native forest ash eucalypt because of wider annual growth rings (2010, p. 44). Differences in appearance if quantified as a comparative premium or discount might provide some credibility to an estimate of price for a new species if empirical methods were employed. Assessing the effect appearance has on price could involve market feedback mechanisms: The value individuals place on appearance could be compared between two species, revealing preference and willingness to pay for the species being compared.
Beadle et al. suggested that quality and performance of a product new to market would need to meet or exceed those for the species being substituted (Beadle et al., 2008, p. 53). It might be more plausible however, to assume that levels of a quality, if lower in a substitute product, would not meet outright market rejection but instead be discounted. Surface hardness for Tasmanian plantation E. nitens, at around 4.5 to 5.3 kN (Janka hardness), was considered by Blakemore et al. to be marginal for flooring applications (2010, p. 44). Consumer preference for a harder species only means the consumer is willing to pay less for the softer species.
Price premiums or discounts could be measured for different levels of quality characteristics. Price adjustments from the product being substituted could represent the different levels in the new product being priced.
Comparisons were used by this study to produce price estimates for E. nitens as a timber species new to the market. Comparisons were made by survey respondents for different levels of quality characteristics between species. Price discounts and premiums were determined for 15 year old E. nitens compared with Victorian ash based on appearance, hardness and movement in service levels.
Both grade and size of hardwood appearance sawn products influence their value to the consumer (Nolan et al., 2005, p. 7). Nolan et al. measured unweighted average market prices of wholesale eucalypt hardwood flooring product profiles, with prices measured relative to a benchmark grade, width and thickness (2005, p. 7). Prices were found to increase with higher grades (i.e. less feature) along with thicker and wider boards (Nolan et al., 2005, p. 7). Nolan et al. also stated that longer boards attract a price premium over shorter boards (2005, p. 7) but did not quantify this.
For results of an investment analysis to be credible where the products are not yet available and sold in the market, methods should attempt to accurately estimate what buyers in the market would be willing to pay for each product based on grade, width and length levels.
Quantifying sawn timber revenue from a log requires prices for the full range of profiles produced. In response to the need for market prices for E. nitens timber to establish log residual values, this study reported price estimates as discounts and premiums for levels of width, length and grade quality characteristics for E. nitens flooring timber.
A market survey was performed as part of this study to elicit respondents’ judgements of price for the E. nitens flooring product profiles. These were represented as discounts and premiums for the levels of quality characteristics produced in this case study.
Two value-based social survey methods were employed to price the E. nitens flooring product profiles produced in this case study:
- Dollar metric pricing of utility and part worth utilities using the graded-pairs comparison approach.
- Constant-sum allocation pricing of part-worth utilities.
Both methods are self-explicated stated preference approaches that directly estimate utilities for each product profile.
Discounts and premiums were quantified for the different levels of quality characteristics that made up each product profile.
Results were a monetised numeric estimation of maximum acceptable price to the consumer relative to the given price of a reference product (Monroe, 1990, p. 122). Rather than assuming that E. nitens and Australian ash eucalypt hold equivalent economic value, by taking differences in appearance, hardness and movement in service into account, price for E. nitens substitute products were estimated without market prices being available. By directly comparing quality attributes, including appearance, the survey prompted survey respondents to use normal consumer behaviour in comparing price and levels of attributes between species and products in deciding a maximum acceptable price they would be willing to pay for each E. nitens product.
This case study utilised survey discounts and premiums for quality levels to price the E. nitens sawn flooring products in order to estimate log revenue for both survey pricing methods.
Residual value approach for pricing products
The residual value method can be applied to infer price for a product used to manufacture other products (Kengen, 1997, p. 44). Such indirect pricing offers an alternative to arbitrary discounts for quality characteristics not readily priced in the market such as board length. For example, short floorboard lengths could be priced by estimating the value of a finished floor laid from short lengths, from which installation costs would be deducted to result in the price for the raw product. By comparing with a floor laid from long lengths and with a known price, a discount for the new product could be estimated.
Glue-lamination and finger-jointing of short length appearance hardwood may offer opportunities for product innovations such as pre-finished laminated parquet flooring and laminated panels (Shield, 1995, p. 137). Where price for an innovative manufactured product is not yet available from sales data because it is has not achieved market penetration, market surveys offer an option for empirically estimating product prices. By deducting costs for a residual value, products such as timber ‘shorts’ that are traditionally perceived to be of low value, could then potentially be priced from the manufactured product.
This study estimated some product values by taking into account innovations and changes underway in the market because of the long time frames involved with growing trees and because products and prices establish log residual values.
Product residual value was selected for pricing short floorboard lengths. This study priced short lengths of solid flooring E. nitens from their appearance as a finished floor by comparing the appearance with a floor made from longer lengths and asking survey respondents to judge utility. Costs of producing and installing end-jointed and end-matched product were estimated and subtracted for residual product values for two levels of short lengths.
Product residual value was selected for pricing case study boards of widths too narrow for standard flooring but suitable for laminating into panels (75 mm and 50 mm widths). Sample panels were produced and sold and production costs were quantified.
The literature reviewed indicates that any study that seeks to estimate profitability of growing E. nitens for solid timber products needs to consider a wide number of factors. While the concept of residual value is straightforward, its application to the process of sawing and producing timber products necessitates the consideration of many factors, including the application of best practice sawmilling and drying methods for converting logs into sawn timber products.
Important components identified in the literature review relating to residual value include: Processing costs, grade recoveries and defect levels, log diameter, log position, checking, collapse, end-splits and movement. The literature also shows factors that have the potential to improve residual value and profitability such as application of best-practice silvicultural methods, best-practice sawmill and drying methods and selection of highest-value products to saw. These factors and their influence on residual value will inevitably change as improvements are made over time.
The general research objective of this study is to estimate the profitability of growing E. nitens for solid timber products as a case study. The case study was a small woodlot of well managed E. nitens, grown for 15 years, pruned to 6.5 m and thinned for solid timber production. To achieve this general objective, methods were also developed that could be applied consistently in future studies assessing log residual value for E. nitens.
Taking into account the findings of the literature review suggests the following specific research questions:
- What is the estimated residual value and resulting net present value for E. nitens in this case study?
- What is the impact of degrade and defect on sawn timber value?
- Is it likely to be profitable to grow E. nitens under the case study scenario?
- What effect does drying method have on wood product quality and value?
- Does log position in the tree affect wood product quality, and if so, case study log residual value?
- Does log diameter affect processing costs and sawn timber value, and if so, case study log residual value?
A design to address these research questions was developed and the key features of the design are:
- To apply sawmilling method as practiced in Satchell and Turner (2010), with 3 m long lengths sawn soon after cross cutting to length to minimise defect from end splits and sawn to 28 mm thickness;
- to apply best practice drying methods to reduce levels of checking as much as practicable;
- to quantify levels of checks on unreconditioned profiled surfaces and examine their impact on sawn timber value;
- to assess the economic impact of steam reconditioning on log residual value.
- to trial two methods of air drying that offer different costs to compare the outcomes in economic terms. The study compared air drying outdoors with the more expensive option of air drying indoors in a ventilated drying shed;
- selection of products to saw that represented the least risk and greatest value. Solid timber strip flooring was selected as the target product for processing from case study logs for the purpose of determining log residual value. This was assumed to be the most marketable and least risky product to produce and price. The market for solid strip flooring is negligible for flooring board widths under 100 mm, so the product selected for 75 mm and 50 mm board widths was laminated appearance panels;
- selection of methods for pricing final products that represented how the market would value these. Prices for products produced in this research were identified as an important component that determines economic viability of growing E. nitens for sawn timber. Methods for pricing product profiles as accurately as possible were identified and selected for estimating prices for sawn products. These were discounts and premiums according to product profiles for both the graded pairs and constant sum allocation methods here reported in this study, modified for short timber lengths as residual product values in the graded-pairs method. Panel laminating stock was priced according to residual product values.
Disclaimer: The opinions and information provided in this report have been provided in good faith and on the basis that every endeavour has been made to be accurate and not misleading and to exercise reasonable care, skill and judgement in providing such opinions and information. The Author and NZFFA will not be responsible if information is inaccurate or not up to date, nor will we be responsible if you use or rely on the information in any way.