EVALUATING TREES AND TREE STEMS AND/OR LOGS

Field of the Invention

The invention relates to the ranking and/or selection of logs derived from a stand of trees for wood quality dependent purposes.

Background

We have determined an effective technique for ranking and/or selection of logs for stiffness and related wood fibre properties, to a degree adequate for commercial benefit, can be achieved within a stand based solely on log physical parameters and without the additional use of log-by-log time-of-flight acoustic speed measures.

This has immediate potential commercial application as these log physical measures are available from the processor head software - so a software based tool could be built into the processor head control system to enable stiffness based segregation. This solution would be easy to implement compared against getting physical probes (as in our

WO2006/049514) to be robust enough to survive and work effectively in such a harsh working environment.

The timber industry faces a need to efficiently utilise its rather variable forest resources. This is true whether in New Zealand or elsewhere. Timber classification, for example, machine stress grading, is currently done at the end of the production chain although some earlier interventions have more recently been proposed. Stress grading, an example of late stage timber classification, results in wastage from processing which ultimately proves to be inappropriate. It is therefore more efficient to measure the log properties early in the chain and process the logs accordingly.

As used herein the term "logs" refers to logs cut from a tree or tree stem but can in certain circumstance also be applicable to the tree stem itself, cants cut therefrom, beams, boards, etc. It is of course possible to assign different destinies to different logs to be cut from the same stem of a felled tree which will maximise the extracted value of harvesting of the particular tree. This has been recognised in New Zealand and elsewhere by software packages such as those of the New Zealand Forest Research Institute (i.e. the AVISTM Software) or previously used by LIRO (i.e. the New Zealand Logging Industry Research Organisation).

In our New Zealand Patent Specifications 333434 and 337015, the full content of which is hereby included by way of reference, we disclose techniques useful early in the harvesting/processing chain reliant upon a FFT (Fast Fourier Transform) analysis of the wave pattern induced by reflecting sonic or the like stress waves in the tree-stem, logs, etc as a result of usually an impact induction of the sound or stress wave. Such apparatus and its methodology allows a reasonable estimation of modulus of elasticity (MOE) for a green felled log on the basis of an estimation of density of approximately 1000kg/m3 and thus at that early stage worthwhile decision making can be made.

Other systems contemplated also rely upon resolution of reflected sound waves. See, for example, New Zealand Patent Specification 506496 of Weyerhaeuser Company where breakdown decisions are to be made reliant upon reference to a price table that apparently takes heed of end product values and log characteristics that can be imputed to each of such logs from some average estimation for the felled tree stem.

Speed of transmission type testing as opposed to resonance based testing has been disclosed in US Patent 5307679 (Ross).

New Zealand Patent Specification 533153 (discloses hand held apparatus where two spiked probes are linked via an electronic unit with a timing device whereby stress waves imparted through one probe and its spike is timed in its passage to the second spiked probe thereby to enable the travel time of the stress wave from the first probe to the second probe to be determined for tree assessment purposes. Such apparatus is dependent upon manual positioning and manual activation to input the stress wave e.g. using an impact.

Other systems, rather than relying upon resolution of the fundamental frequency by FFT, instead rely upon analysis of a time elapsed process. One such example is that of Boardman, Graham and Tsehay as disclosed in New Zealand Patent Specification 507297 the full content of which is here included by way of reference. That specification discloses the prospect of trees of lesser quality being identified early thus avoiding the milling of trees of inferior quality. NZ507297 in selecting trees for their assessed strength

characteristics therein, is reliant on a time taken for the sound or stress wave to travel along the tree from its input point to a sensing spike. Such information is stated as being associable with a GPS location and data recording. There is also reference of manual marking by paint or the like.

US patent 7043990 (Wang et al) discloses a method of evaluating a timber comprising the acts of:

determining a multi-variable regression model relating a modified modulus of elasticity (MOE) of the timber to at least two variables including a measured MOE of the timber determined by a non-destructive evaluation technique and a physical property of the timber, wherein the physical property can be determined without applying a force to the timber;

determining the measured MOE of the timber by the non-destructive evaluation technique;

determining the physical property of the timber; and calculating the modified MOE of the timber based at least in part on using the measured MOE and the determined physical property of the timber in the multi-variable regression model.

US patent 6026689 (Snyder et al) relates to a system relating a measured speed of a stress wave to maximise the value of wood products produced.

We see a clear safety prospect advantage for limited sample felling or sampling analysis (whether felled or not) and with the felling and tree stem fate and/or felling and at least some log cut decisions and/or cut execution occurring with harvester usage or processor usage.

The present invention envisages a useful advantage to be derived from an encounter between a harvester with its harvesting head and a still standing tree, the tree as it is being felled and/or a recently felled tree. In the case of a recently felled tree or tree stem, a merchandiser may alternatively be used to carry out the same or similar function. Reference herein therefore to the term "harvester" refers to a machine having a felling head and/or processing head and/or a merchandiser of any of the kinds contemplated in the aforementioned patent specification or as otherwise understood in the tree felling art as being a harvester or merchandiser. The term "processor" could in less preferred forms be a pruning head but in more preferred forms a harvester/processor combination.

The present invention also and/or alternatively recognises that there is an advantage to be derived immediately prior to felling, during felling and/or soon after felling to have characterised a particular part of the tree stem and immediately to mark the tree stem as to a characteristic, fate or the like dependent on the physical measure or to immediately process in part the tree stem using a harvester head or merchandiser. The present invention sees an advantage in value extraction from a plantation of reliance upon remote operation from a vehicle of testing apparatus in conjunction with the harvester head and thereafter immediately using the harvester head responsive to the physical measure and/or remotely activating some marking or log cutting procedure, or instructing some marking or log cutting procedure, or a combination of both.

Various forms of harvester are known.

See, for example, the pages 1 to 8 of Timber West Journal September/October 2002 that discusses harvesting and felling head heads specs of the following manufacturers: AFM-Forest Ltd, Caterpillar, Davco Manufacturing, Denharco, Fabtek, Gilbert Tech, Hahn Machinery, Hultdins, Keto/Hakmet, Loewen Forestry Equipment, LogMax, Pierce Pacific Manufacturing mc, Ponsse, Quadco, Risley Equipment, Rotobec, Tigercat Industries mc, Valmet/Timbco/Partek, Waratah, etc.

Examples of such harvesting heads are disclosed in, for example, Valmet US Patent 4537236, Waratah US Patent 4412569, and Haim US Patent 4382457, amongst many others. The present invention also as an option recognises an advantage in, at the harvester head or merchandiser (and preferably under the control of the operator of the harvester), having an appropriate inputting of data to the harvester operator or any recording or optimising apparatus or both of the outcome of any such testing and/or the harvester head marking of tree products derived from a standing tree prior to, at and/or immediately post felling and/or harvester head processing.

It is to one or more of these advantages that the present invention is directed.

US Patent Specifications 6182725 and 6341632 of Bengt Sorvik (both here included by way of reference) relates to a harvester being used in circumstances where data from pre-analysis of a forest region and location of individual trees in that region are tied to an accurate knowledge of the harvesting machine location of its harvesting head so as to appropriately harvest and/or harvesting machine process and mark tree parts immediately prior to felling, during felling or post felling. Such a system however is dependent on multiple inputs from diverse sources.

We propose, and it is an object of the invention, in preferred forms of the invention to use regression coefficients, which are some function of stand categorisation and/or calibration, in relation to easy on site determinable individual tree features.

Another or alternative object is identifying/making/segregating a log or batch of logs to meet a target mean (average) MOE of resultant product (veneer, or lumber, or even pulp wood fibre).

Summary of the Invention

In one aspect the present invention can be said to broadly consist in A method of evaluating logs for wood quality dependent use, said logs to be derived from an eligible stand of trees, said method comprising :

a. attributing a mean value to the stand of trees that is an indicator of stiffness of logs likely to be extracted, and

b. comparing the mean value to reference data in order to derive, for that

attributed mean value, a regression model relating an indicator of log stiffness to at least one log physical characteristic, and

c. using the regression model to evaluate suitability for wood quality dependent use of each log derived, or to be derived, from the stand by reference to said at least one physical characteristic of that log.

In another aspect the present invention can be said to broadly consist in A

processing head of a tree harvesting apparatus comprising at least;

• a saw • jaws to embrace a tree to be harvested

• means to obtain information about the dimensional characteristics of the tree to be harvested, and

• associated software for processing information regarding the dimensional characteristics of the tree in order to evaluate whether the tree is appropriate, or not, for wood quality dependent use or wood quality dependent use processing

and wherein the associated software for processing information regarding the dimensional characteristics of the tree employs a method as herein described to evaluate whether the tree is appropriate, or not, for wood quality dependent use or wood quality dependent use processing.

In another aspect the present invention can be said to broadly consist in A method of evaluating a log to be derived from an eligible stand of trees, said method comprising :

attributing a mean value to the stand of trees that is an indicator of stiffness of logs likely to be extracted, and

if the mean value attributed is above a threshold considered eligible for the stand to provide logs for structural purposes (eg structural timber, veneers, LVL, high strength pulp or paper, etc),

using a data set providing a coefficient for log position in the tree, a coefficient for some dimensional aspect of each such log (eg LED), and a regression constant all for that mean value attributed, either passing as structural or failing as structural each log derived from the stand by reference to its log position and its dimensional aspect.

In another aspect the present invention can be said to broadly consist in A method of passing or notpassing logs being derived from a stand of trees as appropriate, or not, for a downstream use or a particular downstream use processing;

said method comprising :

a. attributing a mean value to the stand of trees, or deriving a mean value of the stand of trees or some sample thereof representative of the stand, that is an indicator of stiffness of logs to be extracted from or within the stand, and b. provided that mean value is above a threshold value to provide sufficient logs for structural use or structural use processing from the stand or within the stand,

c. using a data set appropriate for such mean value providing a regression

coefficient for log position in the tree, a regression coefficient for each of at least one dimensional aspect of a log from such a tree and a regression constant, and

d. passing or not passing each log as derived from the stand, or from within the stand, reliant upon its log position in its tree, and said least one of its said dimensional aspects.

In another aspect the present invention can be said to broadly consist in A method of passing or notpassing logs being derived from a stand of trees as appropriate, or not, for wood quality dependent use or wood quality dependent use processing; said method comprising :

a. attributing a mean value to the stand of trees, or deriving a mean value of the stand of trees or some sample thereof representative of the stand, that is an indicator of stiffness of logs to be extracted from or within the stand, and (provided that mean value is above a threshold value to provide sufficient logs for wood quality dependent use or wood quality dependent use processing from the stand or within the stand),

b. using a data set with a regression constant appropriate for such mean value, such a data set providing a regression coefficient for stem or log physical measures or derivations thereof such as a regression coefficient for log position in the tree, a regression coefficient for each of at least one

dimensional aspect of a log from such a tree and a regression constant, and c. passing or not passing each log as derived from the stand, or from within the stand, reliant upon its log position in its tree, and at least one of its said dimensional aspects.

In another aspect the present invention can be said to broadly consist in A method of passing or notpassing logs being derived from a stand of trees as appropriate, or not, for wood quality dependent use or wood quality dependent use processing; said method comprising:

a. attributing a mean value to the stand of trees, or deriving a mean value of the stand of trees or some sample thereof representative of the stand, that is an indicator of stiffness of same or different length logs to be extracted from or within the stand, and (provided that mean value is above a threshold value to provide sufficient logs for wood quality dependent use or wood quality dependent use processing from the stand or within the stand),

b. using a data set of regression data appropriate for such mean value, such regression data providing :

a regression coefficient for log or notionalised log position in the tree, a regression coefficient for each of at least one, or the, dimensional aspect of a log or notionalised log from such a tree and

a regression constant, and

c. passing or not passing each log as derived from the stand, or from within the stand, reliant upon its log position in its tree, or a log position notionalised by reference to its extent and position in the tree stem, and at least one of its said dimensional aspects or the dimensional aspect.

In another aspect the present invention can be said to broadly consist in A method of passing or notpassing logs being derived from a stand of trees as appropriate, or not, for wood quality dependent use or wood quality dependent use processing; said method comprising :

a. attributing a mean value to the stand of trees, or deriving a mean value of the stand of trees or some sample thereof representative of the stand, that is an indicator of stiffness of logs to be extracted from or within the stand, and

(provided that mean value is above a threshold value to provide sufficient logs for wood quality dependent use or wood quality dependent use processing from the stand or within the stand),

b. using a data set with a regression constant appropriate for such mean value, such a data set providing a regression coefficient for stem or log physical measures or derivations thereof such as a regression coefficient for log position in the tree, a regression coefficient for each of at least one

dimensional aspect of a log from such a tree and a regression constant, and c. passing or not passing each log as derived from the stand, or from within the stand, reliant upon its log position in its tree, and at least one of its said dimensional aspects.

d. attributing a refreshed mean value to the stand of trees, or deriving a

refreshed mean value of the stand of trees or some sample thereof representative of the stand, that is an indicator of stiffness of logs to be extracted from or within the remainder of the stand, and (provided that refreshed mean value is above a threshold value to provide sufficient logs forwood quality dependent use or wood quality dependent use processing from the remainder of stand or within the remainder of the stand), e. using a data set with a regression constant appropriate for such refreshed mean value, such a data set providing a regression coefficient for stem or log physical measures or derivations thereof such as a regression coefficient for log position in the tree, a regression coefficient for each of at least one dimensional aspect of a log from such a tree and a regression constant, and f, passing or not passing each log as derived from the remainder of the stand, or from within the remainder of the stand, reliant upon its log position in its tree, and at least one of its said dimensional aspects.

In another aspect the present invention can be said to broadly consist in A method of streaming logs of lengths a, b, and c ( and d, or e etc if desired ) derived from trees of a stand, said method comprising or including

a. attributing a mean value to the stand of a stiffness characteristic, b. if of a mean value applicable for logs to be derived, applying a

regression data set to the stand applicable to the attributed

characteristic, and

c. cutting and streaming the logs individually by reference to the

regression data;

d. wherein the regression data set has historically been derived , whether with logs of that length series or not;

e. and wherein the streaming is reliant on both at least:

i. log position or a notionalised log position value appropriate for the length sequence a,b,c etc and

ii. some physical characteristic of each log.

In another aspect the present invention can be said to broadly consist in a method of ranking and/or selecting logs to be derived from a stand of (preferably same species and same or similar age) trees but with a mix (as arises when there is growth variation amongst the trees of a stand) of diameters or other observable physical characteristic(s), such ranking and/or selecting to be of a degree to provide commercial benefit, the method comprising or including :

from a sampling of the stand to be representative of the stand for stiffness, deriving at least one value relatable by some observable physical characteristic(s) for each non sampled tree to provide a predictor of stiffness for at least the buttlog and up to three or more logs above the butt of each nonsampled tree, and

(provided the sampling of sufficient logs in the sampled trees has been to derive a value or values indicative of stiffness for structural purposes)

relating the value(s) of the sampling to at least one observable and/or measurable physical characteristic (not of itself or themselves a measure of acoustic speed by a time- of-flight or resonance technique) of each nonsampled tree for the purpose of fate determination based on stiffness or estimate of stiffness of processed product from each log to be derived from each such tree.

In another aspect the invention is a method of ranking and/or selecting logs to be derived from a stand of (preferably same species and same or similar age) trees but with a mix of diameters or other observable physical characteristic(s), such ranking and/or selecting to be of a degree to provide commercial benefit, the method comprising or including :

from a sampling of the stand to be representative of the stand for stiffness, deriving at least one value using a regression analysis relatable as a coefficient by some observable physical characteristic(s) for each non sampled tree to provide a predictor of stiffness for at least the butt log and up to three or more logs above the butt of each nonsampled tree, and

(provided the sampling of sufficient logs in the sampled trees has been to derive a value or values indicative of stiffness for structural purposes)

relating the value(s) of the sampling to at least one observable and/or measurable physical characteristic (not of itself or themselves a measure of acoustic speed by a time- of-flight or resonance technique) of each nonsampled tree for the purpose of fate determination based on stiffness or estimate of stiffness of processed product from each log to be derived from each such tree.

In another aspect the invention is a method of ranking and/or selecting logs to be derived from a stand of (preferably same species and same or similar age) trees but with a mix of diameters or other observable physical characteristic(s), such ranking and/or selecting to be of a degree to provide commercial benefit, the method comprising or including :

from a sampling of the stand to be representative of the stand for stiffness, deriving a coefficient or coefficients each relatable as a coefficient by some observable physical characteristic(s) for each non sampled tree to provide a predictor of stiffness for at least the buttlog and up to three or more logs above the butt of each nonsampled tree, and

(provided the sampling has not excluded the stand as insufficiently stiff for structural timber purposes)

relating the coefficient(s) to at least one observable and/or measurable physical characteristic (not of itself or themselves a measure of acoustic speed by a time-of-flight or resonance technique) of each nonsampled tree for the purpose of fate determination based on stiffness or estimate of stiffness of processed product from each log to be derived from each such tree.

In another aspect the invention is a method of ranking and/or selecting logs to be derived from a stand of same species (and preferably same or similar age) trees but of a mix of diameters or other observable physical characteristic(s), such ranking and/or selecting to be of a degree to provide commercial benefit, the method comprising or including :

from a sampling of the stand to be representative of the stand for stiffness, deriving at least one value relatabie by some observable physical characteristic(s) for each non sampled tree to provide a predictor of stiffness for at least one log (preferably at least two logs) above the butt or butt log of each nonsampled tree, and

(provided the sampling has been to derive a value or values indicative of sufficient logs in the sampled trees of stiffness eligible for structural purposes)

relating the value(s) of the sampling to at least one observable and/or measurable characteristic (not of itself or themselves a measure of acoustic speed by a time-of-flight or resonance technique) of each nonsampled tree for the purpose of fate determination based on stiffness or estimate of stiffness of processed product to be derived from each such tree.

In another aspect the invention is a method of ranking and/or selecting logs to be derived from a stand of same species (and preferably same or similar aged) trees but of a mix of diameters or other observable physical characteristic(s), such ranking and/or selecting to be of a degree for commercial benefit, the method comprising or including : from a sampling of the stand to be representative of the stand for stiffness, deriving at least one value using a regression analysis relatabie as a coefficient or as coefficients, by some observable physical characteristic(s) for each non sampled tree to provide a predictor of stiffness for at least one log (preferably at least two logs) above the butt or butt log of each nonsampled tree, and

(provided the sampling has been to derive a value or values indicative of sufficient logs in the sampled trees of stiffness for structural purposes)

relating the coefficient(s) of the sampling to at least one observable and/or measurable characteristic (not of itself or themselves a measure of acoustic speed by a time-of-flight or resonance technique) of each nonsampled tree for the purpose of fate determination based on stiffness or estimate of stiffness of processed product to be derived from each such tree.

In another aspect the invention is a method of ranking and/or selecting logs to be derived from a stand of the same species (and preferably same or similar aged) trees but with a mix of diameters or other observable physical characteristic(s), such ranking and/or selecting to be of a degree for commercial benefit, the method comprising or including : from a sampling of the stand to be representative of the stand for stiffness, deriving a coefficient or coefficients each relatabie by some observable physical characteristic(s) for each non sampled tree to provide a predictor of stiffness for one log (preferably at least two logs) above the butt or butt log of each nonsampled tree, and (provided the sampling has not excluded the stand as insufficiently stiff for structural purposes)

relating the coefficient(s) of the sampling to at least one observable and/or measurable characteristic (not of itself or themselves a measure of acoustic speed by a time-of-flight or resonance technique) of each nonsampled tree for the purpose of fate determination based on stiffness or an estimate of stiffness of logs or process product to be derived from each such tree.

In another aspect the invention is a method of evaluating a log to be derived from an eligible stand of trees, said method comprising :

attributing a mean value to the stand of trees that is an indicator of stiffness of logs likely to be extracted, and

if the mean value attributed is above a threshold considered eligible for the stand to provide logs for structural purposes (eg structural timber, veneers, LVL, high strength pulp or paper, etc), using a data set providing a coefficient for log position in the tree, a coefficient for some dimensional aspect of each such log (eg LED), and a regression constant all for that mean value attributed, either passing as structural or failing as structural each log derived from the stand by reference to its log position and its dimensional aspect.

The eligibility/calibration step may involve measuring and calculating a mean, for example HM200 Hitman acoustic resonance-based velocity as an indication of stiffness from a sample of logs from the sand; for instance logs which could have been extracted during road line salvage a year or two prior to the stand harvest operation.

Alternatively a sample of the first say 30-50 logs extracted from a stand could be measured for acoustic speed with the Hitman HM200 tool, and an average velocity derived.

Still further, a stand may be prior categorised/calibrated for stiffness using other stand level measures such as age, stocking (stems per hectare), green crown ratio, site index, etc.

The associated regression coefficients could be derived from trial work (correlating stand measures to average stand MOE) or form a look-up table of mean stand MOE (or HM200 velocity equivalent) based on prior trial work.

In yet another aspect the invention is a system for, and/or a method of, evaluating trees and/or tree stems and/or logs, using any of the methods identified as requiring a mean categorisation and for each log coefficients for that mean related to log position and some dimensional (or adjusted dimensional) aspect.

In yet another aspect the invention is a method of evaluating trees and/or tree stems of a stand of trees as they are being broken down, or are to be broken down, into logs Log

1 (butt log), Log 2, Log 3 etc, ; said method being dependent upon a previous assessment of a stand mean from a sampling representative for stiffness of the stand and, from a data set, allocating a stiffness value to specific logs being cut or to be cut by means of the equation

PHi = xa + by + z

where

PHi = Processor head index (MOE; GPa)

a = LED (large end diameter; mm)

b = Log number (1 = Butt log, 2 = 2 ^nd log, 3 = 3 ^rd log)

In another aspect the invention is a harvester/processor adapted by inputs, software and an embedded algorithm and related output(s) to perform a method of the present invention or to perform the mandatory steps (and optionally the optional steps) set out herein,

The following apply to all aspects of the invention set out above in the summary of invention section :

In some embodiments;

• If the mean value attributed in step a. does not satisfy a threshold considered eligible for the stand to provide logs for a particular wood quality dependent use or range of wood quality dependent uses, or

· the regression model derived in step b. indicates that the stand will not

provide logs, or a sufficient quantity of logs, for a particular wood quality dependent use or range of wood quality dependent uses,

• or both,

the downstream processing of the logs is not targeted for those wood quality dependent uses.

In some embodiments said indicator of stiffness is one or more selected from:

• modulus of elasticity

• acoustic velocity and

• density

In some embodiments said at least one log physical characteristic is one or more selected from:

• a dimensional characteristic and

• an indicator of log position in the stem.

In some embodiments said dimensional characteristic is, or is derivative of, one or more selected from :

• diameter

• large end diameter

• small end diameter and • taper

In some embodiments said indicator of log position in the stem is, or is derivative of, one or more selected from :

• log length

· log number and

• distance to midpoint of log

In some embodiments said wood quality dependent use is one or more selected from :

• structural timber

· veneer

• high strength pulp or paper

• LVL

• low strength pulp or paper

• non-structural timber

· pulp wood fibre

In some embodiments the regression model is a linear regression model.

In some embodiments deriving the regression model involves deriving at least:

• a coefficient for said at least one physical characteristic and

• a regression constant.

In some embodiments said regression model is a multivariable model relating log stiffness to at least two log physical characteristics.

In some embodiments said regression model is a multivariable model which relates log stiffness to at least one dimensional characteristic and at least one indicator of log position in the stem .

In some embodiments deriving said regression model includes deriving at least two regression co-efficients, including a co-efficient for each of:

• a dimensional characteristic and

• an indicator of log position in the stem,

and a regression constant.

In some embodiments the regression model accords with the equation PHi = xa + yb

+ z

where PHi = Processor head index for stiffness (MOE;GPa)

a = log dimensional characteristic

b = Log number (1 = Butt log, 2 - 2 ^nd log, 3 = 3 ^rd log)

x and y = the regression co-efficients of a and b respectively, and

z = the regression constant.

In some embodiments said reference data comprises one or more reference data sets, each data set relating one or more log physical characteristics to an indicator of log stiffness for a stand of trees, said stand of trees having been characterised according to a mean value that is an indicator of stiffness for that stand.

In some embodiments said reference data comprises at least two of said reference data sets.

In some embodiments said reference data comprises a library of reference data sets for different stands of trees, each having been characterised according to a mean value that is an indicator of stiffness for that stand.

In some embodiments comparing the attributed mean value to the reference data involves selecting one or more reference data sets relating to trees with similar stand characteristics to the stand from which logs are being derived.

In some embodiments comparing the attributed mean value to the reference data results in identifying that the attributed mean value is different to a characterising mean value of a stand of trees in the reference data, and deriving a regression model by interpolating accordingly from the regression relationships of the reference data. In some embodiments comparing the attributed mean value to the reference data results in either:

• matching the attributed mean value to a characterising mean value of a stand of trees in the reference data, or

• identifying that the attributed mean value is different but sufficiently near to a characterising mean value of a stand of trees in the reference data, and using a regression model that matches the regression relationship of the reference data set.

In some embodiments the mean value attributed to the stand of trees is obtained by one or more methods selected from :

· assessing or measuring physical characteristics of one or more trees in the stand,

• conducting testing on one or more trees in the stand, and

• consulting reference materials and/or historically collected data or

information.

In some embodiments the mean value attributed to the stand of trees is, or has been, obtained by assessing or measuring physical characteristics of and/or conducting testing on a sample of trees taken to be representative of the stand. In some embodiments the mean value attributed to the stand of trees is obtained by testing the acoustic velocity of one or more trees.

In some embodiments acoustic velocity is tested using probes on the processing head of a tree harvesting apparatus.

In some embodiments evaluating suitability for wood quality dependent use involves determining whether for each log, the log stiffness predicted by the regression model satisfies a threshold considered to make the log eligible for a particular wood quality dependent use or range of wood quality dependent uses.

In some embodiments determining that a log is eligible for a particular wood quality dependent use or range of wood quality dependent uses results in one or more selected from :

• sorting of the log,

• marking of the log, and

• processing of the log according to said use or uses for which it is determined to be suitable.

In some embodiments evaluating suitability for wood quality dependent use involves for each log, deriving said at least one physical characteristic of the log by

measurement of the log.

In some embodiments said at least one log physical characteristic is a dimensional characteristic, and

wherein evaluating suitability for wood quality dependent use involves for each log, deriving said dimensional characteristic of the log by taking dimensional measurements of the log.

In some embodiments measurements are obtained from the logs at the harvesting site.

In some embodiments measurements are obtained from the logs using the processing head of a harvester.

In some embodiments wherein measurements obtained from the logs are not a measure of acoustic speed by a time-of-flight or resonance technique.

In some embodiments the method is used to evaluate a series of logs.

In some embodiments the method further comprises the steps of:

d. for a log derived, or to be derived, from the stand, obtaining an indicator of actual stiffness by measurement or testing,

e. adjusting the mean value attributed to the stand according to the

measurement of actual stiffness,

f. comparing the adjusted mean value to a reference data set in order to derive, for that attributed mean value, an adjusted regression model relating log stiffness to at least one log physical characteristic, and

g. using the adjusted regression model to evaluate suitability for wood quality dependent use of each subsequent log derived, or to be derived, from the stand by reference to said at least one physical characteristic of that log.

In some embodiments steps d. to g . are performed at intervals over the series of logs evaluated. Other aspects of the invention may become apparent from the following description which is given by way of example only and with reference to the accompanying drawings.

This invention may also be said broadly to consist in the parts, elements and features referred to or indicated in the specification of the application, individually or collectively, and any or all combinations of any two or more of said parts, elements or features.

As used herein the term "stress wave" envisages any sound, compression or other type wave that might be created by a suitable impact or sound input and which will run along and across the tree or tree stem, log or the like, and whether measured by a resonance technique for speed or by elapsed time.

As used herein "elapsed time" relates to the time of passage or time of flight of a stress wave between sensors irrespective of whether or not one of those sensors is the input for the stress wave.

MOE means modulus of elasticity and also refers to related measures of stiffness, strength, and wood fibre characteristics.

As used herein the term "and/or" means "and" or "or", or both.

As used herein the term "s" following a noun means the plural and/or singular forms of the noun.

As used herein the term "comprises" or "comprising" or any variation thereof means consists only of and/or includes.

As used herein "tree" or "trees" is not restricted to conifers. "Conifer species" includes Radiata Pine, Loblolly Pine, Slash Pine, Hoop Pine, Parana Pine, Brazilian Pine, Yellow Pine, etc and non pine conifers such as Douglas Fir, Spruce, etc. Species may include outside of the conifers, hardwoods such as Aspen, Sweetgum, Eucalyptus and others.

As herein defined "regression analysis" includes (and is not limited to) any kind of regression analysis and not just that of the equation PHi = xa + yb + 2 given later.

As used herein "resonance technique" is not restricted to those using our HITMAN type apparatus (eg HM200 apparatus). Any such technique is included.

As used herein "Structural purposes" include structural timber, veneer, high strength pulp or paper, LVL, etc. Non structural purposes can be low strength pulp or paper, non structural timber, etc. See our WO05/072314.

As used herein "LED" is large end diameter of a log and includes large end diameter adjusted relative to a standard mean diameter according to variation in mean diameter of logs from a stand or logs representative of a stand.

As used herein "mean" includes an approximate average as well as any exact mathematical mean.

As used herein "dimensional characteristic/s" and "physical characteristic/s" includes measures of stem or log dimension and measures of stem or log dimension adjusted relative to a standard mean according to variation in that mean measure of logs from a stand or logs representative of a stand.

Detailed Description of the Invention

The invention will now be described by way of example with reference to the accompanying drawings, in which :

Figure 1: shows a tree, to be cut into logs dependent upon log number from the butt and large end diameter (LED),

Figure 2: shows variation in the stiffness of boards from different longitudinal positions along a tree stem, plotted for five different radial distances from the heart of the tree stem (taken from Ping Xu, J. C. F, Walker (2004) Stiffness gradients in radiata pine trees, Wood Science and Technology 03/2004; 38(1): 1-9. )

Figure 3: shows a graph plotting stem diameter at breast height (cm) relative to the number of stems per hectare in stands of trees at two different locations (taken from Christoph Krieger (1998), The Effects of Tree Spacing on Diameter, Height and Branch Size in White Spruce, Management Notes No 13, P.E.I. Department of Agriculture and Forestry, Forestry Division, Prince Edward Island, Canada).

Figure 4; shows a regression analysis of large end diameter against end product stiffness for logs derived from two different stands of radiata pine trees, a first located in Nelson and a second located on the West Coast.

Figure 5: shows a regression analysis of log diameter against actual lumber Modulus of Elasticity for logs derived from three different stands of loblolly pine trees, each located in the Southern USA.

Figure 6: shows an exemplary harvester with a processor head, as can be employed in carrying out the method of the present invention,

Figure 7: shows, an exemplary processor head, as can be employed in carrying out the method of the present invention, with probes for measuring acoustic velocity.

Figure 8: shows, in plan view, an further exemplary processor head, as can be employed in carrying out the method of the present invention with an articulable arm for obtaining dimensional measurements.

Despite common knowledge that stiffness is inversely correlated with diameter (within stand) and depends on the position a log is cut from within a stem (butt, second, third, or subsequent log or logs), nobody has understood that these parameters can be used as effective stiffness segregators on a processor head (without the addition of individual log-by-log or stem acoustic speed). Nobody has implemented a segregation system based on these simple measures. With the hindsight from our inventions, we presume this is because stands of different average stiffness require the level of calibration we derive from (preferably regression analysis and) sampling so that the simple log physical measures could deliver logs of a specific target mean stiffness.

We have found that for a specific stand, the use of a pre-measured stand average acoustic speed, or even a predicted stand average stiffness based on other stand measures such as site index, age, green crown ratio (depth of green crown relative to total tree height), can be adequate calibration or categorisation from which to derive a regression model. The regression model can subsequently enable effective segregation of logs to be carried out within stand at a log-by-log level using physical parameters (such as diameter and position within stem) alone.

Different wood quality is required for different end-uses of a felled tree stem, for example LVL veneer, plywood, structural lumber, appearance lumber, fibre cement pulp, or low coarseness pulp wood fibre. Wood quality pertains not only to structural stiffness but also high or low measures of related fibre properties.

The speed at which sound travels through wood (acoustic velocity) can be indicative of suitability for wood quality dependent use. For example acoustic velocity may be used to identify a tree stem suitable for a particular end use, for example a "structural' (high stiffness) use, a fibre cement "high pulp strength' use, a Now pulp coarseness' use, or other wood quality dependent use of commercial significance.

Other indicators of wood quality can be the density of the wood and its Modulus of

Elasticity (MoE). While acoustic velocity, density and MoE are related to a number of commercially relevant properties of wood, for the purposes of the present invention it is of particular interest to note that they are indicators of stiffness.

The speed of sound through the wood and the density of the wood are related to the stiffness of the wood by the following equation (equation 1) :

MoE = V ² x p

Where:

MoE is the modulus of elasticity in GPa

V is the speed at which sound travels through the wood in km/s p is the density of the wood in kg/m ³

A tree has a number of other physical characteristics which may impact on, be related to, or be correlated with its wood quality. For example the species, age, location, climate, altitude, soil type, nutrition, and genetics can all affect tree growth.

Some of the physical characteristics are dimensional characteristics, for example the height of the tree, diameter, height to base of green crown, green crown ratio (depth of green crown relative to total tree height), taper and closeness of growth rings in the tree stem. As shown in Figure 1, the stem 6 of a tree 5 can be cut at cut positions, for example those indicated as 7, to provide a butt log 1, and one or more subsequent logs 2, 3, 4 etc. The physical characteristics of the tree stem vary over its height, thus the subsequent logs, 2, 3, 4 etc will have differing wood properties to those of the butt log 1, and be devoid of the convergence (taper) extremes common with butt logs 1.

The logs 2, 3, 4 etc may also have differing wood properties amongst themselves and some lesser convergence tapering to smaller diameters higher up the tree stem 6. The large end diameter (LED) of each log may be taken at its wider end 8, while the small end diameter (SED) may be taken at the end 9 which is less wide due to taper.

There may also be variation between the properties of wood at the heart of the stem and the outer wood and/or bark.

As an example of variation in wood properties within a tree stem , Figure 2 shows the stiffness of boards cut from logs of a tree stem at different log positions of the tree, plotted for five different radial distances from the cross sectional centre of the tree stem.

In methods according to one aspect of the present invention the position of a log within the tree stem may be given by assigning to a butt log the log number 1, and the subsequent logs (2, 3, 4 etc) may be assigned corresponding log numbers according to their proximity to the butt log. In alternative embodiments of the present invention, the log position in the tree stem may instead be expressed as a distance measurement.

Typically this measurement will be taken from the butt of the stem to the midpoint M of the butt log and each subsequent log, as shown as measurements 20 and 21 in Figure 1. Expressing log position in this manner better takes into account variation in the length of logs that may be cut from a single tree stem. These measurements may be considered as notional log numbers, even though not necessarily integers. Some embodiments of a method according to the present invention are performed on a stand of trees. A stand may encompass a group of trees within a defined geographic location, however typically stands are also defined to encompass trees with some of the same or similar physical characteristics. These result in stand characteristics which are indicative of the population of trees within the stand as a whole. For example, a stand may be defined to encompass a group of trees within a geographical boundary that are:

• of the same or similar age, and/or

• of the same or similar species, and/or

• grown in the same or similar conditions (e.g. high up on a ridge, or low down in a valley) and/or

· subjected to the same or similar silviculture regimes (e.g. thinning, pruning).

However within the stand there will always be at least some variation in some of the physical characteristics of the trees. For example, growth variation within a stand may arise as a result of conditions such as the adjacent density of tree stocking. Figure 3 demonstrates the effect of tree spacing on the average mean diameter of the trees within a stand of the same species, there being a greater mean diameter for trees which are less densely stocked.

The method of the present invention may be used to derive logs of a specified wood quality from a stand of (preferably same species and same or similar age) trees but with a mix (as arises when there is growth variation amongst the trees of a stand) of diameters or other observable physical or dimensional characteristic(s). A trial was run to see whether, despite the growth variation in a stand of trees of the same species, logs with wood quality suitable for LVL veneer could be selected from the stand using a method according to some aspects of the present invention.

In the trial, a sample of logs were cut, with the acoustic velocity and physical characteristics of each measured on a log-by-log basis. The sample was of 119 logs from 40 Pinus radiata trees harvested in the Nelson, NZ area from a 34 year old stand. Three logs were obtained from each tree, except for one tree which only yielded two logs.

Most logs were cut to 5.5m in length except for eight logs which were cut to 2.7m in length.

Acoustic velocity was obtained by using a HITMAN™ resonance device HM200.

In this trial, the log physical characteristic of particular interest was log diameter. The small end diameter (SED) and large end diameter (LED) of each log was measured over what bark was remaining (generally very little bark was left on the logs). The log position (e.g. log 1, log 2, log 3 etc) was also recorded.

In this trial, logs were selected based on whether or not they were of a quality suitable for veneer. For veneer, an arbitrary desired average veneer stiffness greater than 9.246 kN/mm ² (log supply target stiffness) was chosen. This corresponds to a batch of logs selected from within the trial sample and having HM200 acoustic velocity of greater than 3.1 km/s.

Measuring the acoustic velocity of each log in the sample gave a mean acoustic velocity of 3.24 km/sec. This was assumed to be representative of the stand mean acoustic velocity, and indicated that the stand would be suitable for yielding LVL logs because the average was above 3.1 km/sec.

We have used the trial LVL veneer MOE out-turn to derive the regression coefficients for the 3.24km/sec mean HM200 velocity trial stand, relating a mean value of veneer MOE to log position and large end diameter (LED) by the equation (equation 2).

PHi = xa + yb + z

where

PHi = Processor head index (MOE;GPa) a = LED (large end diameter; mm)

b = Log number (1 = Butt log, 2 = 2 ^nd log, 3 = 3 ^rd log) and

z = the intercept constant This gave the following :

LED coefficient (x) -0.0062

Log number coefficient (y) 0.4413

Intercept (z) 13.5656

Physical characteristics such as LED and log position are characteristics which can be measured on site during harvesting.

The regression coefficients could be used (in conjunction with a processor/harvester to obtain real time, onsite measurements) as an input with log large end diameter (or large end diameter adjusted according to mean LED of the stand being harvested relative to mean LED of the trial stand) and log position for the assessment of stiffness of the logs from each stem, and to determine whether or not the logs meet the log supply target stiffness.

The regression equation derived above was used to predict the stiffness of each of the logs (as a processor head index), on the basis of LED and log position. Those logs with a predicted stiffness suitable to meet the log supply target of 3.1 km/sec were selected. The number of selected logs was recorded as the yield from the sample.

To verify this method, an alternative selection process was run where the logs were selected on the basis of acoustic speeds measured to indicate actual stiffness. Those logs with an actual stiffness suitable to meet the log supply target of 3.1 km/sec were selected, and the number of logs was recorded as the yield from the sample.

It was found that selecting logs on predicted stiffness yielded almost as many suitable logs as selecting the logs on measured acoustic speed.

Hence the conclusion that choosing a stand with a mean average stiffness predicted for its logs of say, 3.24 km/sec, yet on site relying on the dimensional aspect and log position criteria alone, will lose little (eg no more than 17% in the trial) of the potential yield of all logs actually measurable as meeting the 3.1 km/sec log supply target and thus deemed suitable for LVL. The 17% loss in yield that was observed is considered to be acceptable for selecting and delivering logs according to a log supply target stiffness in a commercial application.

Such logs can be cut to any suitable length (eg 2.7 or 5.5 metres for LVL veneer as in this example).

The new learning from the Nelson trial which was an individual log trial (veneer outturn was tracked and measured for MOE log-by-log) was that simple log physical measures such as log position, plus diameter, provided almost as effective segregation as when time of flight acoustic velocity was added into the regression analysis. A commercial advantage arises where at felling, proper segregation and directing of the logs to downstream processors, and their cut length, can be safely carried out without the risk to workers on that site arising from the alternative use of the HM200 resonance based device to measure acoustic velocity, and without a significant loss of non-structural logs being cut to the wrong length (which can happen when acoustic speed is not measured until after logs have been cut).

In our trial such segregation did not drop yields by more than 17% over the more dangerous on site manual log-by-log segregation methods that might otherwise be used, while enabling log length cutting decisions, log segregation and subsequent delivery to processing facility to be made upon suitability for wood quality dependent purposes.

This new approach is equally applicable to a pulp and paper application, although the trial generating the above presented data set was targeting veneer for LVL manufacture.

Yet this Nelson trial was only using logs from a single stand. Even for trees of the same species, there is likely to be variation in wood quality between different stands. For example, Figure 3 shows growth variation between stands of the same species at two different locations: where at a first location, trees stocked at a density of 8000 stems/ha have an average diameter of 11 cm, whereas trees in a second location stocked at a density of 8000 stems/ha have an average diameter of 12.2 cm.

If we were to use the same physical log measures as a basis for segregation in another stand with the same or similar diameter range, but a lower average HM200 log velocity (ie lower average stiffness) the resulting segregation would be of a batch of logs with lower average MOE than the trial stand produced, and potentially unsuited to structural purposes.

For example, Figure 4 shows LED plotted against the end product MoE for the radiata pine trial stand in Nelson, and for another stand of radiata pine located on New Zealand's West Coast, with a regression line derived for each data set. A log supply target stiffness of 10.1 GPa is shown by the line labelled 23. It can be seen that regardless of what physical measures (diameter) are chosen as a basis for selecting or segregating trees from the Nelson stand to meet the target stiffness, selecting or segregating trees of the same diameter from the West Coast stand will not give a batch of logs that meets the target stiffness.

Therefore, even where there are similarities in the physical characteristics of trees between one stand and the next, there is a need to calibrate the regression model used according to the stand to be harvested in order to get useful results.

From Figure 4, it can be seen that the data set from the Nelson trial stand and the data set from West Coast stand could be used together as reference data. The Nelson stand has a mean acoustic velocity of 3.24 km/sec (and can be characterised by this mean value), while the West Coast stand has a mean acoustic velocity of 2.46 km/sec (and can be characterised by this mean value).

If we were to interpolate between the regression lines for the Nelson and West Coast data sets we could generate regression lines for radiata pine stands with mean acoustic velocities intermediate of 3.24km/sec and 2.46 km/sec (or outside of these ranges by assuming a continuing data trend).

Thus if we were to encounter a third stand of radiata pine trees with a mean acoustic velocity of, for example 2.8 km/sec, we could compare this to the reference data sets that we have for 2.46 km/sec and 3.24 km/sec stands. Noting that 2.8 falls somewhere between 2.46 and 3.24, we could use the reference data to generate a regression line (in this case by proportionately adjusting both the intercept constant and the regression coefficients of the reference regression lines) by which to predict the stiffness of logs from the third stand based on their LED measurement.

A person skilled in the art will appreciate that interpolations from the reference data may be made on a directly proportional basis, or according to non-linear relationships, or with reference to other known or experimentally determined relationships, and/or interpolation techniques.

In some embodiments of the method of the present invention reference data may be stored as a library or collection containing data sets for a number of different stands, each characterised by its stand characteristics (e.g. age, species, location) and by a mean value which is an indicator of stiffness.

In order derive a regression model for log evaluation, it may be desirable to select as reference data only those data sets of the library which pertain to stands with the same or similar stand characteristics as the stand from which logs are being evaluated. For example, if you are evaluating Spruce logs, then you may select only Spruce data sets as reference data.

A person skilled in the art will appreciate that various regression models could be employed. In some embodiments the regression models are linear, and may be single or multivariable. It will be appreciated that in some cases a multivariable regression may give more accurate correlations from which to predict stiffness.

A person skilled in the art will appreciate that a number of different log physical characteristics could be used as inputs into the regression model. In some embodiments the log physical characteristics used as inputs are chosen to be dimensional characteristics of the log which can be measured on site in real time (perhaps using the processing head of the harvester to take the measurements).

Below is a further example of how data can be interpolated from a reference data set to derive a regression model. Data from the Nelson trial stand (characterised by a mean average acoustic velocity of 3.24 km/sec) can be interpolated by shifting the 'z' intercept up and down, which gives the following results:

Stand Mean HM200 Ckm/sec) 3.14 3.24 3.34

LED coefficient (x) -00062 -0.0062 -0.0062

Log number coefficient (y) 0.4413 0.4413 0.4413

Intercept (z) 13.1627 13.5656 13.9686

The data set given above as an example (data for two of the three stands being simulated from the Nelson trial stand data, to be of equal, higher, and lower average

MOE), suggest a means for calibrating each of the higher and lower stiffness stands relative to the trial stand, based upon their average MOE, by adjusting the 'intercept' variable in the regression of PHi (MOE) vs log number plus diameter in the regression applied. Thus if we were to encounter a new stand of radiata pine trees with a mean acoustic velocity of 3.34 km/sec, we can interpolate from the known regression line for the trial stand of 3.24 km/sec (in this case by proportionately increasing the intercept constant) to derive a new regression line which can be used to predict PHi for logs of the new stand based solely on their physical measures of diameter and log number. Below is a further example, of how reference data can be interpolated to derive a regression model for evaluating logs in a stand.

Data was obtained from three different stands of Loblolly pine trees from the same geographic area in the Southern USA. Logs from the three different stands were cut, and for each log the diameter, density and acoustic velocity was measured and recorded. The stiffness of each log could then be calculated by the relationship set out in equation 1.

Logs from the first stand (stand A) had a mean acoustic velocity of 3.0 km/s and mean density of 0.46 kg/m ³ . Logs from the second stand (stand B) had a mean acoustic velocity of 3.21 km/s and mean density of 0.48 kg/m ³ . Logs from the third stand (stand C) had a mean acoustic velocity of 3.28 km/s and mean density of 0.49 kg/m ³ .

Figure 5 shows log diameter against stiffness (MoE) plotted for each log in each stand. For each stand, a regression equation of the form y = ax + c has been derived for this relationship, where y = MoE, a is a regression coefficient, x = diameter (mm) and c is a regression constant.

If the stands are categorised or ranked first, second and third according to either a) mean acoustic velocity or b) density, then it can be seen that this categorisation/ranking corresponds with the position of the regression line for that stand on the Fig 5 plot.

If we were now to encounter a fourth stand of Loblolly pine trees from the same geographic location in the Southern USA, we could measure mean acoustic velocity and/or mean density for the fourth stand (or a sample thereof), and rank that stand relative to the three existing stands on this basis. Based on the ranking of the fourth stand, we could then interpolate from the regression lines of the three existing stands in order to predict a regression equation for the relationship between log diameter and lumber stiffness in the fourth stand. For example, if the mean acoustic velocity of the logs in fourth stand was found to be 3.1 km/sec, then the existing data could be interpolated to provide a regression line for the fourth stand labelled as 24 in Fig 5.

Hence the learning that simple (processor head measurable) physical log parameters such as diameter and log number/position can provide effective segregation for stiffness, and could deliver a segregated batch of logs with predictable average MOE, provided the stand is categorised/calibrated for average stiffness first.

In order perform some versions of the method, there is a need to obtain an indication of the mean stiffness of logs in the stand to be felled. The mean stiffness value (for example as indicated by mean acoustic velocity, density, or some other measure) is used to categorise, characterise or rank the stand relative to the reference data sets, and an appropriate regression model is devised from the reference data taking the

characterisation or ranking of the stand into account.

In some embodiments the mean stiffness value is derived prior to harvesting from a sample of trees which are taken to be indicative of the stand. In some cases, the sampling may occur at a time prior to the harvest of the trees. For example, while stands will usually be clear felled, sampling may be of trees that have been removed from an original periphery of the stand at some stage earlier (for example, by track clearing or the like).

Obtaining a mean value for stiffness can be done in various ways including :

• assessing physical characteristics of the trees (e.g. measuring green crown ratio, outer wood density, or dry wood) which are indicative of stiffness, · consulting reference materials (e.g. relating to the age of the stand, site

location and/or previous test results) or

• conducting testing (for example acoustic velocity can be obtained by using a HITMAN™ ST300 standing tree tool, a HITMAN™ HM200 log tool, or other time-of-flight or resonance device, or a bending stress test can be carried out on a sample from the stand).

In some embodiments, an attempt to better characterise the mean average value of the stand may be ongoing as the stand is processed. For example, an indicator of stiffness may be measured as stems are cut (for example using probes on the harvester head for time-of-flight based acoustic measurement). This measurement is then fed into the calculation of an adjusted stiffness value for the stand which takes into account the acoustic velocity measurements from previously cut stems (i.e. a rolling average). The adjusted value for stand stiffness is used to refresh or recalculate the regression model accordingly, and the adjusted regression model is used for predicting the stiffness of subsequent logs. It may not be necessary to adjust the average mean value and/or the regression model for each log that is evaluated on a log-by-log basis, as this may take too long and/or result in wear on the measuring/testing equipment (e.g. probes). Therefore it may be sufficient to take stiffness measurements at intervals during the evaluation process, for example at every 5 ^th log evaluated, or at every 30 minutes.

At felling, a harvester 10 with a head 11, for example as shown in Figure 2, can be used to cut the stem 6 to derive logs.

The head 11 may have a saw 12 for cutting, optionally one or more probes 13 (hydraulically or otherwise actuated) which may be used for measuring characteristics of the log, and one or more sensors 14 for obtaining information. The sensors 14 may be part of, or separate from the probe 13.

An example of a suitable processor head 11 is shown in plan view in Figure 7, wherein the head has a an articulable arm 25 shown at two limits of its preferred movement with respect to a specific tree stem, such articulation being operable under the action of a hydraulic ram. The arm 25 can embrace the tree stem to be felled, and in doing so can obtain measurements of the dimensional characteristics of the tree (for example its diameter).

The head 11 may have associated software for processing the predicted MoE (Phi) of the stem from the information measured/obtained by the head, and which may optionally reference stored data (stored at the head 11 or elsewhere on the harvester 10 or remotely). The software should also incorporate the previously obtained indication of the mean stiffness of logs in the stand to be felled. There may be input means, for example to permit an operator to manually enter at least some of the values needed for processing.

For example, the steps in the process of cutting a log (1,2,3,4 etc) from a felled stem

6 using log number and large end diameter (LED) as indicators of wood stiffness could include the steps of:

1. Head 11 clamps tree or stem (eg at lower end of butt log)

2. Operator activates saw 12

3. Hydraulics insert probes 13 (optional)

4. Software downloads log number and diameter from sensor/s 14 and/or data on processor head

5. Saw cut finishes

6. Velocity measured (optional) by probes. The probe output may contribute to a

"rolling average" mean stiffness calculated by the software.

7. Software calculates PHi by running data through PHi algorithm, compares PHi against predetermined MOE threshold, and uploads result to processor head log-making optimiser Probes 13 retracted (if previously deployed)

Log grade (structural or non-structural) and audible output indicated in cab 15. Operator confirms log making decision (structural or non-structural) for log to be cut

. Head moves to cut position 7, cuts log and repeats process

. Data stored for later download.

Although the invention has been described by way of example and with reference to particular embodiments, it is to be understood that modifications and/or

improvements may be made without departing from the scope or spirit of the invention.