Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Model for predicting net revenue of harvesting operations in coastal second-growth stands Pavel, Mihai 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1997-0187.pdf [ 5.06MB ]
Metadata
JSON: 831-1.0075261.json
JSON-LD: 831-1.0075261-ld.json
RDF/XML (Pretty): 831-1.0075261-rdf.xml
RDF/JSON: 831-1.0075261-rdf.json
Turtle: 831-1.0075261-turtle.txt
N-Triples: 831-1.0075261-rdf-ntriples.txt
Original Record: 831-1.0075261-source.json
Full Text
831-1.0075261-fulltext.txt
Citation
831-1.0075261.ris

Full Text

M O D E L FOR PREDICTING NET R E V E N U E O F H A R V E S T I N G O P E R A T I O N S IN C O A S T A L S E C O N D - G R O W T H S T A N D S by MIHAI P A V E L Dipl. Eng., University Transilvania of Brasov - Romania, 1987  A T H E S I S SUBMITTED IN PARTIAL F U L F I L L M E N T O F T H E R E Q U I R E M E N T S FOR T H E D E G R E E O F MASTER OF FORESTRY in THE FACULTY OF GRADUATE STUDIES Department of Forest Resources Management  We accept this thesis as conforming to the required standard  T H E UNIVERSITY O F BRITISH C O L U M B I A April 1997 © Mihai Pavel, 1997  in  presenting this  degree  at the  thesis  in  University of  partial  fulfilment  of  of  department  this thesis for or  by  his  or  scholarly purposes may be her  representatives.  It  is  permission.  of  H.^esJ-  &s<zt4rcc<^  The University of British Columbia Vancouver, Canada  DE-6 (2/88)  for  an advanced  Library shall make it  agree that permission for extensive  publication of this thesis for financial gain shall not  Department  requirements  British Columbia, I agree that the  freely available for reference and study. I further copying  the  ^<^^CU^(A^  granted  by the  understood  that  head of copying  my or  be allowed without my written  ABSTRACT As forest harvesting shifts from old-growth to second-growth stands, profitability is becoming an important issue. In a cooperative project involving Forest Engineering Research Institute of Canada (FERIC), University of British Columbia (UBC) and Canadian Forest Service - Pacific Forestry Center, a model was developed to predict economics of second-growth harvests.  The final result that can be obtained with the model is the net revenue produced after logging a second-growth stand. This is computed as a function of stand characteristics, company product requirements and harvest equipment used. Additional results computed by the model are total volume, distribution of volume by species and by sort, and estimate of time to harvest a block.  The model is a Windows based program, written in Visual Basic 3.0 using some third party 1  Visual Basic Extensions. The final product is a program that makes data input very easy. It ships on two diskettes with a set up kit, making installation simple.  The model was tested on two second-growth settings close to Powell River, B.C. In both cases the results were very good, value predicted by the model being within 3% of the actual value obtained. More testing is underway and considering input received from industry, some improvements are being considered.  The objective of this project and of the model, which is the final product of the project, is to demonstrate the potential benefits to users of this type of management tool and to serve as a medium term decision support tool that will predict economics of second-growth harvests.  1  Windows and Visual Basic 3.0 are trademarks of Microsoft Corporation. ii  T A B L E OF CONTENTS  Abstract  ii  Table of Contents  iii  List of Tables  vi  List of Figures  vii  Acknowledgments  viii  1. INTRODUCTION  1  1.1 Background  1  1.2 Rationale  2  1.3 Study focus  3  2. OBJECTIVES  4  3. STUDY AREAS  5  4. LITERATURE REVIEW  7  4.1 Review of Methods to Optimize Value for Individual Trees and For an Entire Stand  7  4.1.1 Review of Optimum Bucking Systems for Individual Trees  8  4.1.2 Review of Methods to Predict Stand Value Based on Cruise Data  12  4.1.3 Literature Review for Fiber Recovery and Value Loss in Harvesting Operations  19  4.2 Review of Models Used to Predict Machine Productivity and Cost of Harvesting  22  5. STUDY DESIGN  32  5.1 Model Development  32  5.1.1 Basic Features of The Model  33  5.1.2 Factors Considered During Model Elaboration  34  5.1.3 Presentation of Software Used  35  5.1.4 Data Requirements and Basic Assumptions of The Model  36  iii  5.1.4.1 Data Requirements to Compute Stand Volume and Value  37  5.1.4.2 Data Requirements to Compute Machine Productivity and Cost of Harvesting 5.1.4.3 Basic Assumptions and Limitations of The Model 5.2 Method of Calculating Block Volume and Value 5.2.1 Optimal Bucking for Cruise Data  39 40 41 48  5.2.2 Factors that Can Affect Prediction of Stand Volume and Value and Methods to Account for Them  56  5.3 Method to Calculate the Cost of Harvesting  62  5.4 Model Validation  65  6. FIELD DATA COLLECTION  67  6.1 Data Collection for Prediction of Stand Volume and Value  67  6.1.1 Field Procedures For Checking the Taper Equation, Evaluation of Breakage and Comparison Between Theoretical and Real Bucking 6.1.2 Field Procedure for Cruising 6.2 Data Collection for Harvest System Design 7. RESULTS AND DISCUSSION  67 68 69 70  7.1 Results for Evaluation of Taper Equation, Evaluation of Breakage and Comparison Between Theoretical and Real Bucking  71  7.1.1 Evaluation of Taper Equation  71  7.1.2 Evaluation of Loss due to Breakage  72  7.1.3 Comparison Between Theoretical and Actual Bucking  75  7.2 Comparison of Block Volume and Value - Predicted vs. Actual  76  7.2.1 Results for the Stillwater Block  76  7.2.2 Results for the Okeover Block  86  7.3 Results for Harvesting System Design 8. MODEL SUMMARY AND DISCUSSION  94 95  iv  9. FUTURE DEVELOPMENT OF THE MODEL 10. R E F E R E N C E S  97 100  APPENDICES 1. Presentation of the program using pseudocode  106  1 .A Computation of stand volume and value  106  1 .B Harvest System Design and computation of net revenue  113  2. Productivity equations and average productivities used in Harvest System Design  122  2.A Clear-cut  122  2.B Partial cut  126  v  LIST OF TABLES  T a b l e 1. A v e r a g e s t a n d c h a r a c t e r i s t i c s of t h e s t u d y s i t e s  6  T a b l e 2 . B i a s e s a n d S t a n d a r d E r r o r of E s t i m a t e of d i a m e t e r i n s i d e b a r k f r o m g r o u n d to t o p  72  T a b l e 3 . A n a l y s i s of b r e a k a g e for m a n u a l felling  73  T a b l e 4 . A n a l y s i s of b r e a k a g e for m e c h a n i c a l felling  74  T a b l e 5. M a t c h of l o g s f o u n d at t h e log s o r t y a r d  75  T a b l e 6. F o u r t e e n p r e d i c t i o n s of total v o l u m e a n d v a l u e for t h e S t i l l w a t e r b l o c k  78  T a b l e 7 . R e s u l t s of p r e d i c t i o n s 1-3 v s . a c t u a l , for t h e S t i l l w a t e r b l o c k  79  T a b l e 8 . R e s u l t s of p r e d i c t i o n s 4 - 8 v s . a c t u a l , for t h e S t i l l w a t e r b l o c k  81  T a b l e 9 . R e s u l t s of p r e d i c t i o n s 9 - 1 0 v s . a c t u a l , for t h e S t i l l w a t e r b l o c k  84  T a b l e 1 0 . R e s u l t s of p r e d i c t i o n s 1 1 - 1 4 v s . a c t u a l , for t h e S t i l l w a t e r b l o c k  85  T a b l e 1 1 . F o u r t e e n p r e d i c t i o n s of t o t a l v o l u m e a n d v a l u e for t h e O k e o v e r b l o c k  86  T a b l e 1 2 . R e s u l t s of p r e d i c t i o n s 1-3 v s . a c t u a l , for t h e O k e o v e r b l o c k  87  T a b l e 1 3 . R e s u l t s of p r e d i c t i o n s 4 - 8 v s . a c t u a l , for t h e O k e o v e r b l o c k  89  T a b l e 1 4 . R e s u l t s of p r e d i c t i o n s 9 - 1 0 v s . a c t u a l , for t h e O k e o v e r b l o c k  92  T a b l e 1 5 . R e s u l t s of p r e d i c t i o n s 1 1 - 1 4 v s . a c t u a l , for t h e O k e o v e r b l o c k  93  vi  LIST OF FIGURES  Fig. 1 Block diagram of the model  36  Fig. 2 Block diagram for computation of stand volume and value  45  Fig. 3 Dynamic programming representation of stem  55  Fig. 4 Block diagram representation of dynamic programming - forward pass  55  Fig. 5 Flow chart for computation of harvesting cost and net revenue  64  Fig. 6 Results per species produced with the Model (Predictions 4-8) for the Stillwater block  82  Fig. 7 Results per species produced with the Model (Predictions 4-8) for the Okeover block  90  vii  ACKNOWLEDGMENTS  I would like to acknowledge with gratitude the assistance given to me by the members of the Faculty of Forestry - Forest Operations Group and in particular to my supervisor Prof. G.G. Young. His advice and suggestions in how to prepare for the thesis project and compile the data have been of great help. I would also like to thank Dr. J.D. Nelson for his assistance during the entire period of my graduate studies.  The support of the Forest Engineering Research Institute of Canada (FERIC) for this project is also acknowledged. Special thanks goes to Bjdm Andersson for his assistance throughout this study.  My gratitude is also directed to my fellow graduate student Francois Warren for introducing me into this project and for his friendship, to Peter Dyson from FERIC for his assistance in the field work and to Dave Gizowski from Faculty of Forestry - Forest Resources Management Department, for his help in designing the user interface of the model and guidance throughout the programming process.  I would like also to give sincere thanks to people from MacMillan Bloedel Ltd. - Stillwater Division and Granet Lake Logging Ltd. from Powell River, for providing two study blocks for this project and much information during and after the harvest.  This project was funded by the Canada - British Columbia Partnership Agreement on Forest Resource Development: FRDA II.  viii  1. INTRODUCTION 1.1 Background There has always been a lot of interest associated with predicting the value of a stand before harvesting. This prediction is correlated with the inventory system used to evaluate the volume of timber and with sorts that are to be produced from the stand.  Initially, researchers focused on old growth. Later on, second-growth stands came into attention, because old-growth forests are constantly diminishing and public pressure to stop harvesting them is increasing.  In British Columbia, value prediction is performed according to Ministry of Forests scaling and grading system. The main purpose of currently used procedures, is to predict the volume according to statutory grades. This is done to estimate the value of the stand as a base for bidding and for stumpage appraisal. Later, after scaling of harvested wood, the result is used to calculate stumpage and royalties. The method of collecting stand data is regulated in the Forest Service Cruising Manual. The description of grades and legal procedures associated with scaling, are given in Forest Service Scaling Manual. Cruise data collected in the province are processed, using a cruise compilation program. The program considers the inventory zones and uses a cruise compilation loss factor table to estimate loss associated with various defects.  This paper presents the results of a research project involving the Forest Engineering Research Institute of Canada (FERIC), the University of British Columbia (UBC) and Canadian Forest Service - Pacific Forestry Center. In this cooperative project, a computer model was developed to predict the net value (revenue) of British Columbia coastal second-growth stands as a function of stand characteristics, harvest equipment used and company product requirements. In contrast with the existing method, which is designed for general use in the province, the model is  1  a tool intended to analyze economics of harvesting operations, based strictly on specific information collected from the stand under consideration.  The objective of this project and of the model, is to demonstrate the potential benefits to users of this type of management tool and to serve as a medium term (1-5 years) decision support tool that will predict economics of second-growth harvests.  1.2 Rationale The utility and applicability of the model, is related to the fact that forest managers responsible for marketing, work study and logging planning require detailed information on the potential yield and log size distribution from a stand. Information is required on the potential of the stand to yield different types and qualities of logs as a result of different bucking patterns, average piece size, and the number, size and volume of the individual log types. Detailed data by product type, specific to beatable areas are necessary to maximize the contribution from the timber resource. After completion of a literature review, it was found that there were no such models that can reliably predict the net revenue generated by a stand.  The attempt to predict the value of second-growths was inspired by the practical observation that there are few defects associated with these stands. This is in contrast with old growth, where the presence of many defects and the imprecision of correlating them with loss in wood volume, can make the prediction less reliable (accurate).  In terms of predicting the cost of harvesting second-growth stands, it is estimated (Andersson and Jukes, 1995) that the change from harvesting old-growth to second-growth stands is presenting new operational challenges for the forest industry. The consensus is that if old-growth harvesting methods are used in second-growth forests, harvesting costs will be higher than when  2  the same methods are used in old growth. However, the smaller tree size of second-growth timber allows harvest planners to choose from a variety of mechanized harvesting systems, which could lead to reduced costs. Choosing appropriate systems for harvesting second-growth stands is presently difficult because little information exists concerning related costs and productivity.  1.3 Study Focus The fact that this study focuses on second-growth is driven by the increasing importance of these stands in timber supply analysis. In 1985 approximately 12% of the annual allowable cut (AAC) of B.C.'s Southern Coastal Region, was harvested from these stands; for the same region, in 1995, second-growth timber was about 33% of the volume harvested and forecasts (Sauder, 1988) suggested that by the year 2005 second-growth volumes may be as high as 47% of total harvest. The same author has predicted that under current conditions there is only 20 to 30 years of harvestable old-growth remaining for most coastal operations.  For the purpose of this study, a second-growth forest is defined (Andersson et al., 1995) as being less than 150 years old, and having originated following harvesting or natural disturbance such as fire, wind storm, or insect infestation. The characteristics of the second-growth stands projected for harvest are diverse. They range in age from 50 to 150 years, have volumes from 400 to 1100 m /ha, and stand densities from 200 to 900 trees/ha. Many stands, especially those 3  of natural origin, contain some large trees ("veterans") from the previous stand. In those cases, tree diameter at breast height (dbh) can range from 15 to 150 cm, and tree heights from 20 to 60 m. Slopes are generally gentle, but some sites may have slopes in excess of 80%. The presence of old growth stumps and large windfalls are also characteristic of second-growth stands.  3  2. OBJECTIVES The main objective of the study is to develop a computer model that will be used to accurately predict the revenue generated by second-growth harvests. The revenue generated by a stand will be computed as the difference between the value of timber and the cost of harvesting it; other costs like stumpage and road construction are not considered. To complete the main objective, the following problems have to be addressed:  1. Establish the value of timber based on cruise data and on company sort descriptions for two silvicultural systems: clear-cut and partial cut.  To accomplish the first objective, the following secondary objectives were addressed: 1.1 Propose and test a modified method of cruising. 1.2 Develop an algorithm that will theoretically buck each tree to its maximum value. 1.3 Rank defects identified in second-growth stands, in terms of their impact on value prediction; study their effect by incrementally including them in the analysis. 1.4 Determine breakage and other loss factors for the systems studied. Quantify the amount of wood lost during harvesting, and the impact on stand value.  2. Develop a quick analysis tool for estimating the cost of harvesting a cut block. This part of the project will offer the user data about productivity and cost of different machines. The majority of data used to derive these productivity equations were developed as part of FERIC studies; additional equations were obtained from published studies carried out in the Pacific North-West.  4  3. STUDY AREAS As a part of this study, two operations were monitored; the first one was used for developing the model and the second one was used for testing it. Both blocks were situated on Crown land, near Powell River, in the biogeoclimatic region CWHvml - Submontane Very Wet Maritime Coastal Western Hemlock Variant.  The first block called # 921 Weldwood Main, was offered by MacMillan Bloedel Ltd., Stillwater Division. Species encountered in this study site were Douglas fir (Pseudotsuga menziesii (Mirb.) Franco var. menziesii), western hemlock (Tsuga heterophylla (Raf.) Sarg.) and western red cedar (Thuja plicata Donn). The total area of the block was 40 ha. The block was cruised by FERIC in November 1995. Logging started in January 1996 and lasted until April. Trees were manually felled and bucked into prime lengths (or combinations of prime lengths) then extracted to the roadside using a Madill 123 grapple yarder and two John Deere 690 excavator forwarders. Logs closer to the road (20% of total area) were loaded directly onto trucks. Results of the scale were obtained from the sortyard of the company.  The second block was offered by Granet Lake Logging Ltd. from Powell River. This block consisted of two settings, totaling 14.6 ha. It was a small business sale with the license number A 39750 and was located by Okeover Inlet. The species in the block were the same as in the first block. This stand originated after a fire that left many old growth cedars. These trees displayed poor form having the bottom almost destroyed by fire, and an average coefficient of utilization around 40%. Overall, in terms of density and tree diameters, this was a non-uniform stand as its characteristics varied greatly throughout the area. The full tree system was adopted in this block. Trees were manually felled, but not delimbed, except for large trees and the ones by the boundary. Extraction used a Madill 122 skyline yarder, with processing done mechanically in the first sub-block by a Keto 1000 processor, and manually by two buckers in the second subblock. Some additional bucking and upgrading was done at the sortyard.  5  For both study sites, average stand characteristics are given in Table 1. In the rest of this paper the two blocks will be called Stillwater and Okeover, respectively.  Table 1. Average Stand Characteristics of Study Sites Tree dbh Total height Trees/ha Avg. tree vol. Gross vol ./ha Net vol ./ha Stillwater Okeover  cm. 33 29  m 27 33  m 1.22 0.7 3  565 1021  The data in Table 1 are based on operational cruises.  6  m 690 715 3  m 627 680 3  4. LITERATURE REVIEW Maximizing tree value through optimal bucking and prediction of revenue has been of interest to foresters for a long time. However, as part of this study, only new models and methods will be presented, because older ones are obsolete either because of the programming language, the operating system, or from criteria related to scaling methods and units of measurement.  4.1 Review of Methods to Optimize Value for Individual Trees and For an Entire Stand The definition of what maximizes value depends on the viewpoint of the decision-maker. Various viewpoints will be presented throughout this chapter.  It has been long recognized that undesirable reductions in managerial flexibility and substantial opportunity losses are incurred by bucking prematurely and in retrospect, inappropriately. The move, particularly in North America, to full tree logging and trucking and the development of log merchandisers to centralize the bucking decision and maximize realizable yields, provides ample testimony to the importance of bucking.  Murphy et al. (1991), outlined that the key finding of all research carried out with a view to maximize value recovery is that it is much easier to add $1 to unit product value by improving bucking than it is to reduce $1 from unit costs.  Pickens et. al (1992) identified the main causes of poor bucking decisions as follows: •  the complexity of industry specifications.  •  unquantified or imprecise specifications.  •  the number of log grades.  •  production pressure.  •  lack of incentives.  7  •  lack of training, and  •  lack of decision aids.  Murphy et al. (1996), present the findings of one of New Zealand's larger forest companies which stated that up to US $ 1.5 million improvement in annual revenue could be gained if they were able to capture an additional one per cent in value recovery. For all New Zealand, it was estimated that value recovered could be improved by over US $ 50 million per year.  4.1.1  Review of Optimum Bucking Systems for Individual Trees  Pnevmaticos and Mann (1972), outlined that the most important penalties of poor log production are: 1) the loss of volume from inefficient log bucking practices, and 2) the loss of value due to poor choice of end products. The authors outlined the general principles and structure of a dynamic programming problem and designed a method for log optimum bucking. Because there is no standard algorithm, they identified some simplifying assumptions: •  the problem is characterized by stages. In the example analyzed a stage is defined as the smallest possible log length. This assumption drastically reduces the computations required in the optimization process.  •  sorts have the same length.  •  the taper was considered to be zero, and  •  no defects or quality remarks were considered.  With all the above simplifications, the problem looked trivial. However, this paper outlined the principles of dynamic programming in log bucking and generated further research.  Young and Waddell (1986), consider the following factors in the optimum bucking process: 1) species, 2) quality and size characteristics of the tree, 3) the various conversion plants and markets to which the logs can be allocated, 4) current product values, 5) log extraction costs, 6) transportation costs, and 7) worker safety. The combined effect of these factors makes tree  8  bucking for optimal value an important and complex decision process. They produced a hand held computer designed for fallers. The authors suggest that benefits of a bucking system, would be felt by logging divisions, the forest industry and the Crown. The system developed included information related to the tree (dimensions and other characteristics) and data about the market and extraction costs. The optimization system was comprised of two major components: 1) an office-based microcomputer and, 2) a sophisticated handheld field computer. The office-based microcomputer was used to store information about sorts that were downloaded to the handheld. For the handheld, prior to bucking, the faller was to input the relevant characteristics of the tree such as butt diameter, distance from the butt to each change in wood quality, and finally top diameter. Then, a dynamic programming algorithm used these data to determine the optimal cuts. The field computer could also be used to calculate the value of a proposed bucking policy based on a pattern chosen by the faller. This option allowed for unusual bucking situations where the faller could not buck at the optimal location due to safety considerations or if the tree has been damaged when felled. This option could also serve as a faller training guide. Once the tree characteristics were input, the faller could play "what if" analysis to determine the value of alternative bucking patterns. Linear taper was used, based on top diameter, butt diameter and the distance between these two measurements. Field tests of the system showed a net increase in value of 4%, in a study area that was bucked to higher standards than normal due to easy terrain and the additional training of the falling crew. Larger savings were expected to result in normal circumstances.  Considering significant value loss that occurred in New Zealand forests operations, Murphy and Twaddle (1986), conducted a value recovery control program. A system called AVIS (Assessment of Value by Individual Stems, Geerts and Twaddle 1984) that used dynamic programming for optimal bucking was used in conjunction with various statistical quality control techniques. The advantages and disadvantages of the sampling techniques used are underlined, but no final results (in terms of dollars recovered) are presented.  9  Sessions (1988), outlined the principles used in designing a hand-held computer used at the felling site to assist in optimal tree-bucking. The computer uses a network optimization method. It can simultaneously consider tree surface quality, taper, log lengths, diameters and grades, mill delivered prices, logging costs and transport costs. Background information must be entered into the computer before it is taken to the woods. Such information includes the delivered prices for logs of various grades, lengths, and diameters, transport costs to various destinations, and logging information by log size, scaling rules and stumpage payment rules. At the felling site the bucker enters information on tree characteristics into the handheld computer. Tree taper is entered by measuring length and diameter at various intervals along the tree. The bucker assigns a quality code to each zone of the tree after observing branch size, branch frequency and surface defects. Locations where bucking cuts must be made because of breaks or sweep, and intervals along the tree where it is not safe to buck, are also entered. Profit calculation can include costs beyond the tree itself, e.g. if slash cleanup costs can be related to the size of logs not removed from the forest, a penalty or negative profit can be used to assess the value of the tree.  Based on Sessions's (1988) work, Olsen et. al. (1990), developed BUCK, a computer optimizer that can be used for harvest planning. Apart from finding the best combination of logs to be cut from a given tree (no information is given about the method used), BUCK also evaluates the following problems: •  whether to remove or leave slash and marginal value logs. The cost of leaving slash at the stump can be weighted against yarding it to the landing or delivering it to a mill.  •  weight limitations, which may be imposed on a logging system by one of the machines used during log handling.  •  log-mix restrictions, i.e. a given percentage of a grade is required to be in logs longer than a given limit.  10  •  payload efficiency: BUCK analyzes the trade-off between size of payload and number of logs that have to be handled.  Selected sample-trees measurements can be processed on a desktop computer, and this model can be used for evaluating management choices under various logging scenarios.  Raghavendra (1994), overviews Operations Research methods in forestry. He focuses on dynamic programming and gives good theoretical explanations on how this method is applied in optimal bucking, paying special attention to logs with defects. His method consists in grouping logs into categories, thus reducing the variability existing in log descriptions.  Some optimum bucking systems have been included in mechanical harvesters. Sondell (1987), evaluates the systems available in Sweden. Although some sophisticated systems were available, like bucking to taper or bucking to value, many modern features were unavailable at that time.  The optimum bucking problem was also approached in a different stage of log processing - at the sawmill. The conditions are much better compared to woods bucking and consequently more sophisticated methods can be used.  Maness and Adams (1991), developed a method for combined optimization of log bucking and sawing strategies at the mill. The method consists of three models to address each aspect of the problem. •  A stem bucking model that solves the optimal bucking policy using dynamic programming.  •  A cutting pattern optimizer that simulates the cant sawing of a log with a given diameter, taper and length, and determines the optimal sawing pattern given lumber values by dimension and length.  ll  •  A log allocation model (based on linear programming - LP) that introduces production constraints and optimizes the output of a sawmill based on first two models.  The three models are linked and solved simultaneously, the first two programs as sub-problems to the LP. Suggested cant breakdown patterns and log bucking solutions are generated for the LP as a function of the current marginal value of both primary and secondary products. The LP then chooses between the set of "suggested" patterns rather than the full set of feasible patterns. The set of "suggested" patterns continues to grow until no new pattern increases the objective function of the LP. At this point in time the problem is solved. Obviously, it is very likely that logs at the sawmill had already been bucked or had received some sort of manufacturing. Stems traveling into the sawmill are optically and the dynamic programming model determines the best bucking policy.  4.1.2 Review of Methods to Predict Stand Value Based on Cruise Data The process of predicting stand value based on cruise data has to consider more information (constraints) than in the case of individual trees. This mainly consists of market requirements, mill demands, additional costs like trucking and sorting, etc. Some of the best models for this purpose are presented in the rest of this section.  Deadman and Goulding (1979), developed A Method for Assessment of Recoverable Volume by Log Types (MARVL), which accounts for stem quality and malformation plus log specifications and preferences. The method used to optimize revenue from each tree is dynamic programming. This method accounts for various defects, the prevalent being forking. Some features of this model follow: •  a complicated cruising method is used and the stem is partitioned into variable length intervals, each of which can be considered to be of uniform quality. One-letter  12  alphabetic codes (that incorporates various characteristics of the sort) are assigned to each interval. Forks, large branches and diameter reductions require estimation of diameters at 1.4 m above the top of the defect. Skilled staff are necessary for estimation of stem quality and dimensions. Lengths and diameters are usually estimated, although frequent checks are being made using a hypsometer and a relaskop. •  observations for each tree are recorded directly on a hand-held computer.  •  a local taper equation (fifth degree parabola) is used to compute diameter inside bark at different heights. Volume is computed using also a locally developed function.  •  a breakage simulator randomly selects breakage in such a way that all the trees have the same probability of breaking. For the selected trees, an expected relative break point height is derived from a general function based on average ground slope and total height. All material above the break point is treated as waste.  A test of the method at ten clearcutting sites (for each site a sample of twenty-five Pinus radiata tress was taken) indicated that good results were attainable. As a result of compensation between sorts, the overall error in recovered volume was 3.7 %. Although log value is the driving factor for the dynamic programming procedure, no comment is made about it. An operational system based on the method is now being used by the New Zealand Forest Service. The authors acknowledge that use of the system outside their country would not be an easy process.  Eng et al. (1986), prescribe appropriate bucking patterns to be applied to several stands over a specified period so as to meet market requirements and other side constraints. This model takes into account simultaneously the limitations of the forest resource in terms of quality and quantities and the market requirements for products. Other characteristics of this approach are:  13  •  the bucking is done at the landing and dynamic programming is used to solve this problem. The algorithm is explicitly interfaced with a linear programming module through price. The dynamic programming subproblem is used to generate activities for the linear programming, so that bucking strategies reflect the opportunity costs resulting from critical constraints on demands and resources.  •  the main assumption of the model is that trees can be grouped into classes, each defined by the size and quality of the stems found in one or more stands. In the example presented, the authors used 100 classes, which seems to make the field procedures cumbersome.  •  no useable model is supplied. The authors recommend a mixture of three components (procedural language, linear programming solver and spreadsheet).  The model was applied in Fiji and the authors state that further work on extending this methodology to medium term (3 -10 years) forecasts is merited.  Mendoza and Bare (1986), approached optimal bucking and log allocation as a two-stage decision problem. This approach allows for the formulation of a linear programming problem (usually with a large number of decision variables), as a two-stage decision problem where two interrelated and functionally dependent problems are solved at each stage. The first stage determines an optimal allocation policy for the logs; the second stage consists of generating alternative log bucking policies. The method used to solve the second stage problem is a modified dynamic programming algorithm. The following points are noted: •  trees and logs are grouped into stem classes based on length.  •  the model does not include other quantitative and qualitative bases for categorizing or grading logs.  •  log markets were not included in this formulation but they can be easily incorporated in the objective function and constraints.  14  TREEVAL (Briggs 1989) is a microcomputer model primarily developed for predicting the product value of stands grown under various silvicutural regimes. The stands of interest are often hypothetical or modeled rather than actual stands. The objective of this model is primarily for long-term strategic planning of silvicultural decisions rather than short-term or instantaneous decisions of how to buck a tree. A dynamic programming algorithm (written in FORTRAN) is used to find the combination of logs that maximizes the value of each tree. Some characteristics of this model are as follows: •  value of markets (lumber, veneer, pulp and chips) is obtained using empirical product recovery relations, prices and costs to estimate the conversion return.  •  shape abnormalities such as sweep and crook are not considered. Because the intent of the model is to forecast product value from stands grown under intensive management, it is assumed that trees developing these defects will be a minor stand component.  •  no breakage was considered. The dynamic programming formulation is applied to a situation like mill-length log bucking, i.e. the tree quality is considered known.  •  only small end diameter and length are considered as log dimensions.  •  taper equations are used to compute diameter inside bark at different heights. The user selects from three different taper equations. The first one has the coefficients already built-in; the other two require user input. One of the last two equations was developed by Kozak et. al. (1969).  Sessions et al. (1989), found that optimal tree bucking at the stump, that maximizes values of individual trees, does not reflect the price structure that log sellers face; i.e. many contracts specify that a given percentage of log volume must be contained in logs of a specific length. A method designed was to maximize individual-tree value under contracts with such volumelength restrictions. The following is a good example: if one mill offers $ 220/m and requires that 3  70% of the log volume be in logs exceeding 10m, and another mill offers $ 205 /m and requires 3  15  that 50 % of the log volume be in logs exceeding 8.5 m, the seller can determine which contract delivers the highest value considering the length-volume restrictions and the price schedules simultaneously. Input consists of description of a sample of the trees to be harvested. An optimal bucking pattern is found for each tree (using a Network Optimization method), based on the unadjusted set of log prices. The results are summed over all trees in the sample. If the percentage of long logs satisfies the requirements, log prices are not adjusted. If the long-log percentage must be increased, the prices for all long logs are increased by a multiplication factor determined through binary search. The solution proposed identifies a set of adjusted log prices that maximized the net value of each tree while meeting the length-volume requirement for the stand as a whole.  Based on the program BUCK (Olsen et al. 1990) an improved model called CRUISE/BUCK was developed by Olsen et al. (1991). BUCK was a bucking aid for felled trees ready to be cut into logs. CRUISE/BUCK applies to standing trees; measurements are done on sample plots and data are processed in the office. The CRUISE/BUCK method can estimate the type of logs which should be cut from a stand and evaluates the potential revenue if different sets of mills are chosen. The main objective of the program was to achieve more value from a timber sale by picking a set of bidders that would pay the highest net return. Input data includes market specifications (mill prices, transportation costs and harvesting costs) plus cruise data. Apart from usual measurements that are collected when cruising, three different methods were used to estimate diameters inside bark at different heights (dendrometer, relaskop and taper equations). Results were tested against diameter measurements made with calipers, after the trees were felled. The overall standard deviations were respectively 3.0 cm., 4.8 cm. and 3.6 cm. Surface quality assessments from sample trees were also estimated, using a clinometer. Each tree had an average of five different quality codes assigned. For the case analyzed, the log sortyard solution had 2.8% more volume but 6% less total value. The authors recommend taper equations, and acknowledge that the development of a surface-quality scheme is a difficult process and the person collecting field data must be well prepared in using it. Because  16  implementation of this method requires some training, it was recommended that a timber owner who cruises only occasionally would be better off hiring a consultant trained in the use of CRUISE/BUCK.  Warren (1995), conducted a study in Coastal B.C., where stands were cruised before logging, using fixed area plots, and the author developed local height as a function of DBH equations. Kozak's (1988) taper equation was used in conjunction with Mathcad , to calculate diameters on 1  the stem, given a set of sort descriptions. Apart from typical measurements taken for each tree, as a part of this project, the term "crown ratio" was introduced, which was defined as the fraction of the total height of the tree that is covered with branches. Each tree was analyzed independently and the optimization started from the butt. Step by step the author tried to fit for each tree the combination of logs that maximizes its value. Final conclusions have not been published yet, however, it seems that a drawback of this procedure is that it is a slow process, since most decisions are manual (Warren, F. - personal comm., 1996).  Nieuwenhuis and Malone (1996), describe the early stages of a method aimed at the development of cost-effective methods and data analysis software, with a view to maximize the value obtained from forest stands through optimal inventory and cross-cutting methodologies in Ireland. The decision support system which is being developed, is designed for the sawmilling industry, linking standing timber to mill specifications and demands. The project is based on the fact that relative uniformity of Irish plantations and cross-cutting practices provides an ideal context for the development of systems for maximizing the value of standing timber. As a part of this project, an attempt was made to develop a taper equation, relating lower stem taper to that of the entire stem, however, this approach did not seem to produce very good results and affects precision of the volume prediction. The authors consider that further studies will be required  1  Mathcad is a registered trademark of MathSoft Inc. 17  b e f o r e t h e m o s t a p p r o p r i a t e m e a n s of r e p r e s e n t i n g s t e m profile c a n b e f o u n d , but o v e r a l l t h e prototype software yields promising results.  M u r p h y et a l . ( 1 9 9 6 ) , p r e s e n t t h e p r i n c i p l e s a n d first r e s u l t s of a n i n t e g r a t e d v a l u e m a n a g e m e n t s y s t e m , that is u n d e r w a y in N e w Z e a l a n d . D u e to its c o m p l e x i t y , t h e d e v e l o p m e n t of this t o o l m a y e x t e n d into t h e e a r l y 2 1 s t c e n t u r y . S i n g l e s t e m o p t i m i z a t i o n of all s t e m s , c o m p u t e r v i s i o n a n d m a r k e t f e e d - b a c k m e c h a n i s m s a r e t h e k e y c o m p o n e n t s of t h e Integrated V a l u e M a n a g e m e n t S y s t e m . T h e A V I S (Geerts a n d T w a d d l e 1984) d y n a m i c programming algorithm w a s c h o s e n a s t h e k e y part of a l a r g e r s y s t e m for o p t i m i z i n g log p r o d u c t i o n in r e a l t i m e . T h e e l e m e n t that d i f f e r e n t i a t e s t h i s m o d e l f r o m t h e p r e v i o u s o n e is a c o m p u t e r v i s i o n s y s t e m . T h e g e o m e t r i c i n f o r m a t i o n ( e . g . d i a m e t e r , s w e e p , ovality a n d length) will b e c a p t u r e d t h r o u g h t h e u s e of s t e r e o i m a g e s , o b t a i n e d with digital p h o t o g r a p h y , a s t h e l o g - m a k e r w a l k s t h e l e n g t h of t h e s t e m . T h e s a m e i m a g e s will b e u s e d to a s s e s s e x t e r n a l l y v i s i b l e q u a l i t i e s a l o n g t h e l e n g t h of t h e s t e m . T h e d a t a c a p t u r e d b y t h e c o m p u t e r v i s i o n s y s t e m will b e u s e d b y t h e A V I S a l g o r i t h m to o p t i m i z e s i n g l e s t e m s . T o e n s u r e that s i n g l e s t e m o p t i m i z a t i o n a l s o m e e t s m a r k e t a n d o p e r a t i o n a l c o n s t r a i n t s , n e w p r o c e d u r e s a r e b e i n g d e v e l o p e d to a l l o c a t e t h e b e s t cutting p a t t e r n s a n d s t a n d s to e a c h l o g g i n g c r e w . T a b u s e a r c h h e u r i s t i c s a r e u s e d to s o l v e this last p r o b l e m . I n s t e a d of p r o v i d i n g t h e o p t i m a l s o l u t i o n to t h e p r o b l e m , a h e u r i s t i c m a y o n l y p r o v i d e a g o o d s o l u t i o n , but it is u s e d b e c a u s e of its ability to s o l v e difficult p r o b l e m s in a r e a s o n a b l e a m o u n t of t i m e . A l t h o u g h not m e n t i o n e d b y t h e a u t h o r s , it s e e m s that t h e m a j o r p r o b l e m will b e t h e t i m e r e q u i r e d for c o l l e c t i n g t h e i m a g e of e a c h t r e e .  A p a r t f r o m m e t h o d s d e s c r i b e d a b o v e , t h e r e a r e o t h e r m e t h o d d e v e l o p e d for s a w m i l l s o r o t h e r p l a c e s (log s o r t y a r d s ) w h e r e a h i g h e r d e g r e e of m e c h a n i z a t i o n c a n b e o b t a i n e d : •  G r e e n (1994), reviews s c a n n i n g s y s t e m s a n d a p p r o a c h e s . Optimization m e t h o d s are a l s o a n a l y z e d , a l o n g with t h e i r i m p l i c a t i o n s o n o v e r a l l activity of t h e s a w m i l l .  18  •  Hardison (1995), describes the advanced methodology employed at bucking stations in sawmills. Two trends are obvious; one consists of scanners and computer-based optimizing programs, another one uses cameras at the log deck and video monitors to give the operator a vantage point as if he were personally examining the logs on the deck.  4.1.3  Literature Review for Fiber Recovery and Value Loss in Harvesting Operations  Related to predicting volume and value of a stand, is the evaluation of losses that occur during harvesting. Some of the methods developed to estimate and to account for this follow.  Murphy (1983), identified that forest industry was in many cases production not value oriented, i.e. not much attention is paid to recovering the entire value of trees. He identified the following losses for various logging phases: •  thinning 1 - 2 %.  •  felling damage exclusively due to breakage 4 -7%.  •  felling damage due to other causes (high stumps, butt damage in the form of slabbing, side-splitting etc.) 4 - 5%.  •  extraction losses 1 - 2%.  •  processing losses 20 - 25%.  Gingras (1992), defined fiber recovery efficiency as how completely a given harvesting system recovers the merchantable fiber. In other words, this is the volume recovered divided by the sum of volume recovered plus merchantable slash. He evaluated six different systems, producing recovery indices between 91 - 99%.  19  Copithorne and Young (1994), monitored two sites in Coastal B.C. and found that total losses in log recovery value was 5.7%. Broken down by harvesting stage it was: 1) falling 3.6%, 2) bucking at stump 1.5%, 3) yarding 0.2%, and 4) bucking at the log sortyard 0.4%. The authors conducted a detailed analysis of causes that produced these losses. Typical second-growth stands, consisting of Douglas fir, western hemlock and red cedar were studied. No information is given about crew skill, but if is assumed that an average crew was monitored, then the above results can be considered average losses that occur when harvesting in coastal British Columbia stands. Apart from factors that are inherent to harvesting process, decisions taken by log makers can be of great importance in value recovery.  The operator's ability to produce optimum value under production conditions was tested by Geerts and Twaddle (1984). A program called AVIS (Assessment of Value by Individual Stems) was developed to evaluate what skidworkers can achieve in optimal crosscutting of the stems. In this logging system workers were in a confined area shared with large machines and given a set of log-making instructions (which are ambiguous and/or imprecise), then were expected to do a rapid assessment of each stem, recognize the various log qualities, and optimize the stem's value. As a part of this study each tree was scaled after being felled, but before extraction. Rather than using a general volume or taper equation, sectional measurements are made on each stem. If the tree broke, AVIS enables pieces to be grouped to calculate the optimal cutting pattern of each piece and the maximum value of each tree. The optimal crosscutting algorithm used is the same as the one used by MARVL (Deadman and Goulding 1979). There were no restrictions on the maximum or minimum quantities of any of the log types produced throughout the study period. Two studies had been conducted as part of this project. The first result showed that 26% of potential stem value was being lost from sub-optimal log-making by the skidworkers. This was reduced to 11% after the introduction of practices designed to aid the workers in their choice of individual stem cutting patterns.  20  C o s s e n s a n d M u r p h y ( 1 9 8 8 ) , s t u d i e d h u m a n v a r i a t i o n in o p t i m a l l o g m a k i n g . T w o g r o u p s of p e o p l e with a b o u t t h e s a m e a g e a n d y e a r s of e x p e r i e n c e , but differing v e r y m u c h in e d u c a t i o n a n d t r a i n i n g , m a r k e d 31 t r e e s f o r cutting into l o g s . T h e c o n c l u s i o n w a s that t h e first g r o u p that i n c l u d e d s u p e r v i s o r s (better e d u c a t e d ) w e r e significantly better t h a n t h e s e c o n d g r o u p , that i n c l u d e d s k i d w o r k e r s . T h e v a l u e lost a v e r a g e d (for b o t h g r o u p s ) 6 . 6 % , r a n g i n g f r o m 4 t o 1 0 % . T h e d a t a s h o w e d that 5 6 % of e r r o r s w e r e d u e t o d i m e n s i o n a l e r r o r s ( s u c h a s d i a m e t e r a n d length), 3 5 % w e r e d u e t o m i s s e d o r m i s i n t e r p r e t e d quality f e a t u r e s , a n d 9 % w e r e d u e to o t h e r r e a s o n s . H o w e v e r , e v e n t h o u g h all p e o p l e w e r e i n v o l v e d in forest r e l a t e d activities, a n d w e r e f a m i l i a r w i t h l o g g r a d e s p e c i f i c a t i o n s , n o n e of t h e m h a d a n y f o r m a l t r a i n i n g in t h e t a s k of l o g making.  P a r k e r et a l . ( 1 9 9 5 ) , f o u n d that a h i g h n u m b e r of l o g g r a d e s ( 1 4 o r 1 9 in t h e i r s t u d y ) , i n c r e a s e d t h e n u m b e r of o u t - o f - s p e c i f i c a t i o n l o g s . O n t h e o t h e r h a n d , cutting a g r e a t e r n u m b e r of l o g g r a d e s h a s m a n y a d v a n t a g e s , i n c r e a s i n g t h e flexibility of t h e o p e r a t i o n , a n d t h u s p r o v i d i n g c u s t o m e r s with r e q u e s t e d l o g g r a d e s p l u s a c h i e v i n g g r e a t e r v a l u e r e c o v e r y . T h e a u t h o r s c o n s i d e r that t h e o p t i m u m n u m b e r of g r a d e s t o b e cut b y a l o g g i n g c o n t r a c t o r is a b o u t t e n .  Conclusions at Section 4.1 : A f t e r r e v i e w i n g all s t u d i e s p r e s e n t e d s o f a r , n o n e of t h e m s e e m to r e s p o n d t o s p e c i f i c n e e d s of t h e forest industry in B . C . V a r i o u s a p p r o a c h e s a r e c o n s i d e r e d for b u c k i n g individual t r e e s to their optimal v a l u e . S o m e m o d e l s extend t h e s e algorithms to t h e entire s t a n d , t a k i n g into a c c o u n t b r e a k a g e a n d v a r i o u s t y p e s of l o s s , p l u s a d d i t i o n a l f a c t o r s that o c c u r in t h e l o g g i n g p r o c e s s . H o w e v e r , n o n e of t h e s e m e t h o d s c a n b e u s e d in c o n j u n c t i o n with t h e i n v e n t o r y (cruising) m e t h o d u s e d in B . C . A l s o , n o n e of t h e m w a s d e s i g n t o u s e s p e c i f i c c o m p a n y sort d e s c r i p t i o n s a s u s e d b y t h e l o c a l forest industry.  S o m e of t h e a l g o r i t h m s d e s c r i b e d , a r e in fact g o o d starting p o i n t s f o r o p t i m a l b u c k i n g of i n d i v i d u a l t r e e s . C o n c e p t s a n d i d e a s f r o m o t h e r s m o d e l s c a n b e c h a n g e d a n d u s e d to s a t i s f y s p e c i f i c r e q u i r e m e n t s of B . C . forestry. T h e r e a r e d e s c r i b e d in t h e r e m a i n d e r of t h i s t h e s i s .  21  4.2 Review of Models Used to Predict Machine Productivity and Cost of Harvesting Developing harvesting models was a major preoccupation in North American forestry. As presented in the rest of this chapter, two trends can be distinguished in this activity. The first one consisted of producing computer models that were able to chose machinery to harvest a given block with greatest economical efficiency. These models are commonly called expert systems. The second trend consisted of predicting results of the harvesting operation (mainly in terms of productivity and cost) based on stand data and existing machinery. Conclusions about these models and how the model developed in this study relates to other models, will be presented at the end of the chapter.  Gibson (1979), presents a computerized planning and evaluation system for cable logging operations. The planning model is segmented into three main levels: sale area, sub-block area and profile (deflection line). Input consists of a digitized map of the area and general data about the stand. Regression equations that predicted turn time based upon characteristics of the stand were developed and programmed into the model. Using interactive graphics, the engineer can view the area to be logged, select locations of corridors and query the system for various output, such as system mechanics and time to log any particular logging unit. The system included various cable systems (running and live skyline and hi-lead) and different silvicultural prescriptions (shelterwood, clear-cut and group selection).  Goulet et al. (1980) present the criteria for a harvesting computer model that was to be design in the southern US. The purpose of the paper was to ask the forest harvesting community for an evaluation and an idea exchange. The model objective is to provide a means for addressing questions that relate to wood flow and costs, so that intelligent management decisions about the  22  harvesting operation can be made. A tree-to-mill flow diagram of southern harvesting operations, with each node representing a harvesting phase, is presented. The main feature of the model is modularization, which is achieved by: •  executing only those nodes that are on the user's path.  •  allowing for different implementations of each node, and hence explicitly admitting into the design the great variability present within any one harvesting function.  Comments are made about the level of detail and variables that are to be implemented into the model. The variables will be production, time and costs - by machine, by function and by total system. Other information presented to potential users includes technical tests for the model, how flexible and user friendly the model should be, computer language and machine requirements for this model.  Meng et. al (1981) presented two approaches to predicting performance of logging machines. The study was primarily oriented towards evaluating new machines; however, the same methodology can be used for existing ones. The first approach was to develop simple analytical formulae, in which equations express average harvesting time per tree as a function of stand, tree and machine variables (average values are substituted for each variable). The second approach involved tree-by-tree simulation of harvest. The authors consider that the first type of analysis can be done reasonably well by hand, whereas the second requires a computer and detailed programs that incorporate assumptions about the behavior of the machine and its operator during logging. The best (to that date) Canadian simulation model for harvesting machines, CANLOG (Newnham, 1971) was selected for comparison. The Larson Shortwood Harvester was selected as the subject machine. Results from the two methods did not differ appreciably, despite the greater complexity and cost of the simulation approach. However, when compared with available field data, neither method was particularly reliable. The authors conclude that simulation may be an effective alternative where considerable interaction exists between machines or machine functions, where variable distributions are given substance through empirical studies, and where adequate program documentation exists.  23  A very comprehensive package (Harvesting Analysis Technique), developed over a period of fourteen years is presented by Stuart (1981). This is a three-part system for modeling the environment, individual machine activities and harvesting system interactions. It consists of three major components: 1) procedures for stand definition, 2) a generalized machine simulator and 3) a harvesting system simulator. These components may be linked in different ways. Stands may be selected from the library of mapped stands or with the stand generator included in the model. Each tree is identified using sixteen characteristics, e.g.: coordinates, species, dbh, diameter at ground line, double bark thickness at ground line, crown width etc.. Based on these, four more characteristics are computed for each tree. These variables are passed to the machine simulator and used in modeling machine movements and performance time. This module can be linked with a growth simulator that can accept details for various combinations of thinnings and fertilization strategies. The user can vary stand age at time of thinning and stand development after thinning. The machine simulator is capable of handling the most common harvesting machine activities. The area to be treated at one time by the machine being simulated is defined by a swath width parameter, which determines the width of the strip harvested by the machine each time it passes across the face of the stand. Trees to be harvested and the order of removal is given by individual tree data previously entered. Different functions may be simulated for each machine: travel empty, travel loaded, move, process, turn, cut, etc. The performance time for each of the functions is generated either by sampling from a user supplied distribution or through equations provided, the user must select the equation type and supply the coefficients required. The last part of the model, the Harvesting System Simulator (HSS), can be used to simulate systems up to a maximum of fourteen machines operating on areas containing up to fourteen different harvesting units. The harvesting strategy is controlled by assigning priorities to each area for each machine in the system. The model can include up to twenty different types of down or delay times. Reports include a set of production, income and cost data for the entire operation and for each machine in the system, plus other more specialized report types. The authors state that the program was in use during its entire  24  evolution, being tested in various circumstances, but they do not provide any result of these tests. Although it seems that the model was getting to a great level of detail, the authors state that the model is in a ongoing state of evolution and refinement.  Webster et al. (1983) present the major features of a feller-buncher module which is one part of a new timber harvesting simulation model. In this article, authors present some of the basic principles of the model. As in Goulet et al. (1980), the most important design criteria for the model is modularity, which allows portrayal of a variety of different harvesting operations. The user can combine different modules to depict a total harvesting system or select one module and simulate only a part of harvesting, such as felling. The article describes the features of a feller-buncher module which simulates the actions of a drive-to-tree feller-buncher with an accumulating shear head. To simulate the harvesting movements, the module requires both machine functions and the consideration of site factors which affect harvest operations. The stand to be harvested is modeled by overlaying a map of the stand with a grid. Cruise data are used to describe the grid cells, giving the average percent slope of the land in each grid cell, the maximum bunch cross-sectional area, the maximum bunch volume etc. Tree characteristics are: coordinates, dbh, volume, species and status (harvestable tree, leave tree or obstacle). In addition to the physical characteristics of each grid cell, the model considers the harvest pattern desired. The feller-buncher is assumed to produce a uniformly arranged bunch of trees having a volume and total bunch diameter matching the requirements of a companion skidder. It can move through the plot avoiding leave trees and obstacles, selecting trees, harvesting them and depositing bunches along a predetermined pattern. When a tree is selected as a candidate for shearing, its dbh and volume are examined to see if the shear head and bunch will hold the tree. If not, the next closest tree is examined. Several influencing factors are chosen, like machine interaction and stand characteristics (density, growth pattern, distribution of trees by species and dbh, and obstacles and leave trees, stand brushiness. Delay elements like scheduled delays, equipment failures, and machine interferences are also considered. The output produced by the model includes: productive machine hour, total length of delays, total volume of  25  wood bunched, average length of a cycle, average volume per schedule machine hour, average volume per productive machine hour etc. End of simulation reports include summaries of utilization, production and costs of the feller-buncher. The authors mention that much effort has been expanded to develop this module which faithfully mimics the operation of a feller-buncher and allows considerable flexibility. No comment is made about the time required to run the model or about any trials of the model. No more detail about the other phases of the harvest were found.  A computer model (HARVEST) that has the purpose of simulating harvesting operations with cable skidders, is described by Howard (1987). The model was written in Fortran IV and could be executed on both CP/M and MS-DOS systems. Previously developed models for predicting productive and non-productive time for felling, skidding and bucking were programmed in HARVEST. A data processing program (FOREST), which is capable of generating stand and stocking tables (from either marking tallies or inventory data) is linked to HARVEST. The model incorporates a detailed methodology for logging costs analysis. Author states that these programs in combination with user-interactive programs represent a decision support system which is common in southern New England. In the case presented, the system was used to investigate the relationships that influence logging profitability. HARVEST can also generate information helpful in the development of silvicultural and operational management strategies.  Reisinger et al. (1988), analyzed three of the most recent and comprehensive programs developed for analyzing harvesting system costs and production problems. These are: Auburn Harvesting Analyzer (a spreadsheet program developed by Auburn University - School of Forestry), Harvesting System Analyzer (a menu-driven program written in Pascal, developed by the Tennessee Valley Authority) and Harvesting System Simulator (component of the Harvesting Analysis Technique - Stuart, 1981). The authors present a description of each program, along with the results of a test using data for a mechanized tree-length system. Results of the analysis of the three programs compare closely with each other, and computed cost per unit was only  26  slightly higher than an industry average for a similar system. However, data input for all programs was tedious and time-consuming, particularly for analysis of large systems. In several cases, individual programs were not flexible enough to accept input data in different formats. Programs produced comparable results using realistic system data, even though there were some inconsistencies in the way harvesting costs and productivities were computed. The authors do not recommend a specific program as the best, because the choice ultimately rests with the user and depends on that person's specific needs.  Mellgren (1990) acknowledges that many of the logging machine types used in Canada were studied and evaluated. However, it was always recognized that the results obtained were site and operator specific. A best guess performance correction factor had to be added to forecast the performance of similar machines in different logging conditions. To assist logging managers in the selection of the optimal logging system, and for budgeting and control, the author developed a series of baseline performance curves related to average tree size. These baseline curves are for commonly used logging machines or logging systems working in what was defined as ideal conditions where machine performance would be optimized. The shortwood at roadside harvesting system was chosen as a common point for comparison. The same comparison logic can be used for tree-length or full-tree systems. Correction factors were provided so that machine performance and logging costs can be predicted for a variety of stand conditions. Results are presented in a graphical form to illustrate the analysis and comparison logic. The main purpose of the study was to present a methodology along with the appropriate equations. Potential users should substitute their own data for variables according to their experience and practice. The performance graphs and system comparisons presented in the study are meant to apply to large-scale logging operations, and may not be valid in very different conditions such as smaller-scale operations where performance is affected by various limitations.  27  Randhawa et. al (1992) describe Timber Harvester, a microcomputer system designed to automatically generate feasible harvesting systems. Harvesting system selection is based on the interaction of two major components: 1) site and stand data plus processing variables and 2) the operating attributes of the available equipment. The alternatives selected depend on the information provided in the equipment and support databases, that includes production equations and costs. The best-first search algorithm is used to find a solution. It uses production cost to guide the search process in the direction which minimizes the total production cost. The authors stress that selecting an appropriate degree of mechanization to avoid underutilization of expensive resources is a critical decision, and requires that the product mix, environmental and user constraints be matched against the available technology and required performance criteria. This tool is especially intended to aid long-term, strategic planning by both government agency engineers and private industry managers and owners. It could also be used for short term tactical planning by contractors.  Koger (1992) developed a Skidding, Trucking and Landing Simulation (STALS). This is a timber harvesting program that utilizes production by function, queuing theory and simulation technique to analyze skidding, loading and trucking interactions. Input consists of general data about the block (volume and value) and the number of machines that are to be used. Help screens with harvesting rates, cycle volumes, and equipment hourly costs are provided to assist the user in estimating values for skidding volumes and cycle times, loading cycle times, trucking volumes and cycle times and hourly equipment costs. Queuing theory is used to analyze the system. For both skidder and truck arrival times at the landing it is assumed that the probability of more than one event occurring in a small time interval is negligible, and the probability of an event occurring in this time interval is independent of other intervals. Finally, STALS performs a simulation analysis of the system. Several dynamic interactions such as equipment delays, full landing conditions, and wood shortages can be modeled using this technique. As mentioned before, to simplify the complex interactions that occur in a normal timber harvesting operation, only the activities of skidding, loading and trucking were modeled. The author acknowledges that  28  the model lacks the detail that is required in real situations, and the fact that it does not consider many non-productive times (scheduled work breaks, end-of-shift times, random down times, etc.).The emphasis in this model has been to keep data input to a minimum while obtaining as much information as possible by using queuing theory methods. No information is given about validation of this model.  Baumgras et al. (1993), developed GB-SIM ( A Ground Based Harvesting Simulation model) to estimate stump-to-truck production rates and multiproduct yields for conventional ground-based timber harvesting systems in Appalachian hardwood stands (clear-cuts and partial cuts). The model tries to link the real attributes of the operating environment (e.g. skidding distance, slope, topography) to the harvest stand attributes (e.g. stand density, product volume) and harvest system characteristics (e.g. numbers and types of equipment, utilization rates, volumes skidded). GB-SIM is a stochastic, discrete-event model that applies the next-event approach to update the system state and advance simulation time. One of the main features of the system is the random assignment of tree attributes, including X and Y coordinates that locate each tree within the harvest unit. The model simulates production cycle attributes and calculates cycle times for each type of equipment. While the authors do not expect stochastic simulation models such as GB-SIM to be applied directly by commercial loggers, these research applications can continue to provide valuable information regarding harvesting production and cash flows for the economic analyses of timber harvesting, wood utilization, and forest management alternatives.  An expert system called X-Harvester was produced by Linehan and Corcoran (1993). The system was created for commercial forests under the legal, physical and economic conditions existing in Maine, but it can be adopted for other regions. Expert Systems (ES) have the ability to evaluate both quantitative and qualitative information at the same time providing an important advantage in complex problems such as forest management. The authors acknowledge that although some earlier claims of success of ES (that they would entirely replace human decision makers in many areas) has not proven realistic, they have become  29  valuable tools to assist decision making in many areas. The model includes separate modules (which may pass data from one to another) for the following functions : •  calculate machine rate costs for individual logging machines.  •  evaluate market prices for logs and pulpwood delivered to the mill.  •  build a harvesting system from individual machines and determines how costs should be calculated.  •  evaluate stumpages for a harvest operations.  •  determines financial viability of the harvest by computing the estimated costs and revenues for the operation.  •  evaluates Maine cutting rules and performs an advisory function to evaluate physical operability, regulatory status and soils suitability for forest growth.  The model was tested by back analyzing an easy to evaluate situation (small block, a stand with only pulpwood-grade trees) and results produced agreed with the real ones. At the time it was elaborated, the system was not very flexible (data were stored in so called frames), but authors claim that future development of the model (storing data into databases) would increase its applicability. The framework of the system provides a useful foundation to assist land managers in making timber harvesting decisions.  Final Comment on Chapter 4: Similarly to the conclusion drawn at Section 4.1, at the end of this section I can state that none of existing models fit the specific needs of B.C. forestry. In terms of general approach, most of these programs are restrictive, i.e. lack the degree of flexibility that is required in this field of activity. Also, productivity equations used are inappropriate for conditions existing in Coastal B.C. However, some of these models use concepts that are generally valid, like modularity (Goulet et al. -1980, Webster et al. - 1983), that was used in my model as well. Apart from design criteria presented at the end of the previous section, the model developed in this study eliminated shortcomings exiting in older models, and has the following features:  30  •  various harvesting systems and various scenarios can be easily analyzed for a block.  •  for all phases of the harvesting system, the user is offered a large variety of methods.  •  for each one of these methods, existing or new machines can be used.  •  default productivities and costs are offered for each phase, but these can be always overwritten by the user, based on experience and knowledge.  Each one of the features and a detailed description of the entire model are given in the remaining parts of the thesis.  31  5. STUDY DESIGN The model developed as part of this project is an office tool, intended to be used by planners and foresters working in second-growth forests that are to be clear-cut or partial cut. According to the main objective, the final result of the model will be the net revenue generated by the block; however, during the process of obtaining this result, other important results like distribution of volume by species and sorts, need to be computed and reported. The process of developing the model is presented in the rest of this chapter.  The phases identified in the development of the model are: 1. define the problem and objectives. 2.  study the logging activity as a physical system and determine its discrete components (various phases of the process).  3.  identify input and output pathways; construct flow charts for the system and for each phase.  4.  convert the flow chart to computer code.  5. collect a first set of input data and use it to build the model. 6.  refine the model using results produced by the first set of data.  7. collect a second set of data and use it to verify the model. The very last step consists of offering the software to forest industry for production situations.  5.1 Model Development A step by step description of the methods used for model development and the factors considered throughout this process are described next.  32  5.1.1  Basic Features of The Model  The design of the model was mainly determined by the consideration of steps taken during planning of the harvest. These steps must provide a good estimate of the volume to be logged and find the best methods to do it (i.e. accomplish the work in an economically efficient way without producing damage to the site or generating undesirable side effects).  In British Columbia, the physical process of planning a harvest consists of: •  selection of the block.  •  survey boundaries and calculate the total area.  •  cruise the block according to B.C. Ministry of Forests regulations. The cruise is analyzed with a standardized cruise compilation program, that will estimate volume by species according to Ministry grades. The grading system assigns a grade to each log produced and these are used to calculate the value of stumpage.  •  engineer road layout and, if cable systems are to be used analysis of deflection lines.  •  select equipment to ensure economical feasibility of the operation and compliance with environmental and legal requirements.  From the above description, one can see that there is little information about the value of the wood produced from a block and consequently of the revenue that will be generated. Since this value is dictated by the sorts manufactured, my model was designed to incorporate company sort descriptions. The model was also designed to accept standard cruise data as the basic stand description, which ensures easier implementation of the model. However, some modifications to incorporate additional tree description data to Ministry cruising methods were considered. More details about the data will be given in the rest of this chapter. Full description of office procedures are given in the user manual (Pavel et al., 1997) which is available upon request from FERIC.  33  Considering its main features, my model will: •  calculate the volume by sort and value of timber, using cruise data and company log sort descriptions.  •  allow various harvesting systems to be quickly specified for a block (environmental and legal constraints are considered to be satisfied).  •  estimate the cost of harvesting a block, using productivity functions and cost data provided by the model or supplied by the user.  •  estimate the net revenue from harvesting a block.  In this model, net revenue is obtained by subtracting the cost of harvesting from the value of the timber. As mentioned in the introduction, other data necessary to compute the net revenue are stumpage and the cost of roads, however, these are not included in this model. The user can supply them and manually make adjustments.  5.1.2  Factors Considered During Model Elaboration  The specific needs that are satisfied by the model are described below: •  need to consider alternatives to clear-cutting.  •  need to determine value in fluctuating log markets.  •  need to evaluate many, complicated log sort descriptions.  •  the need to stratify a block according to specific conditions related to site and legal requirements. For each stratum, the user has the option to select from a wide variety of harvesting equipment and systems for each phase of the harvesting system.  •  need to consider interactions between harvesting phases, and  •  need to estimate net revenue.  34  With the above features, the model should be a useful tool for planning second-growth harvests because it responds to the above needs.  5.1.3  Presentation of Software Used  A preliminary analysis of the problem showed that the operating system used as a platform for this project has to be common in the B.C. forest industry. In terms of programming language used, it was desirable that it has the capacity to address all needs of the model and create an easy and attractive interface. Windows was found to be the most suitable operating system, and 1  Visual Basic (current version when this program started was 3.0) was considered to be the most suitable programming language. The interface of the model was designed so that it will be both user friendly and error resistant.  Data are stored in an Microsoft Access 2.0 database. To improve the user interface and to make it more portable, the following third party Visual Basic Extensions (VBX) were used: •  TrueGrid.VBX 2.1 - this improves the capabilities of Visual Basic to display data stored 2  in a database. •  VS.VBX 5.0 - this is a tab control, which is very useful when displaying multiple choices 3  and an elastic which ensures that various forms always fill the screen, regardless of screen size.  Error proofing consists of directing the user input in various instances and making available in some cases only the choices that are consistent with previous decisions.  1 2 3  Windows, Visual Basic 3.0 and Access 2.0 are trademarks of Microsoft Corporation. TrueGrid.VBX is a trademark of Apex Software Corporation. VS.VBX is a trademark of VideoSoft Company. 35  Finally, the model has to provide an easy installation method. This was also solved using the capabilities of Windows and Visual Basic 3.0. A more detailed description of all these features is given in the user manual. The program has more than 4500 lines of code (including comments).  The end product is an independent program, i.e. the user will not need to have other software installed, other than Windows 3.0 or a later version, plus DOS. The program ships on two diskettes with a set up kit making installation simple.  5.1.4  Data Requirements and Basic Assumptions of the Model  Input of the model consists of cruise data and sort descriptions. Based on these data the model computes the volume by sort and total value of the wood in the cut-block. Next, the cost of harvesting processes for that block can be computed using either default productivities or input provided by user. The block diagram in Figure 1 illustrates the principles of this model:  Sort descriptions Optimization algorithm Cruise data  Stand value Stand volume, volume by species and sort, avg. tree vol., avg. dbh, etc.  Productivity by phase for each machine in all strata in the block; cost to operate the machinery.  Net revenue Cost of harvesting (sum of costs by phase for all strata) and estimate of time to log the block.  Figure 1. Block diagram of the model.  36  Data requirements will be further detailed according to computation of stand value and computation of harvesting costs.  5.1.4.1 Data Requirements to Compute Stand Volume and Value  To compute volume and value of the wood harvested from a block, the model needs cruise data and sort descriptions.  The main objective of the cruise is to obtain an estimate, to a specified degree of accuracy and with a specified level of confidence, of the volume of timber on the block. Apart from that, the cruise provides the essential data for determining stumpage rates, for establishing conditions of sale and for planning of the logging operations by the licensee. All details about this activity in B.C. are given in the Ministry of Forests Cruising Manual (1983). The manual sets out the minimum cruising standards including specifications for the statistical design of the cruise, the accuracy of the field measurements and a standard compilation procedure. The reliability of any cruise is based on statistical concepts and this is a function of the intensity of sampling, the uniformity of the timber on the area cruised and the degree of fit of the volume equation to the particular stand. A more detailed description of the method will be given in Chapter 5.2.  The type of data collected for each tree in the cruise are: •  general data about the tree: current number, total height, species, dbh, tree class, crown class.  •  pathological remarks: conk, blind conk, scar, fork or crook, frost crack, mistletoe, rotten branch, dead or broken top.  •  quality remarks: spiral grain, sweep, lean, location of the first live limb and the first stub, data about knots in the first two 5-m. logs.  •  other type of data collected only for damaged stands.  37  T h e t r e e is d i v i d e d into t h i r d s for r e c o r d i n g p a t h o l o g i c a l a n d quality r e m a r k s . C o d e s a r e a s s i g n e d to c o r r e s p o n d i n g thirds ( o n e o r m o r e  T h e r e is n o s t a n d a r d for log sort d e s c r i p t i o n . E a c h c o m p a n y d e s c r i b e s t h e m in a different m a n n e r a n d i n c l u d e s d e t a i l s that m a k e a c e r t a i n l o g a c c e p t a b l e in a s p e c i f i c sort. T h e s e d a t a a r e g e n e r a l l y not m a d e p u b l i c b y forest c o m p a n i e s . A n a n a l y s i s of c o m p a n y s o r t s s h o w e d t h e following c o m m o n features: •  d a t a r e l a t e d to d i m e n s i o n s : 1.  l e n g t h - u s u a l l y g i v e n a s a p r i m e l e n g t h , i.e. a l e n g t h p r e f e r r e d b y t h e mill a n d for w h i c h a p r e m i u m p r i c e is p a i d .  2.  d i a m e t e r s - m i n i m u m t o p d i a m e t e r , m a x i m u m t o p d i a m e t e r , m i n i m u m butt d i a m e t e r ( l e s s c o m m o n ) , m a x i m u m butt d i a m e t e r o r a c o m b i n a t i o n of t h e s e .  •  i n f o r m a t i o n a b o u t k n o t s : t h e m a x i m u m d i a m e t e r of k n o t s a c c e p t e d a n d if t h e y a r e r e q u i r e d to b e w e l l s p a c e d . S o m e t i m e s this is d o n e b y s p e c i f y i n g t h e p o r t i o n of t h e s u r f a c e of t h e log that m u s t b e c l e a r .  •  o t h e r d e f e c t s : s w e e p , c r o o k , twist, v a r i o u s t y p e s of rot, heart c h e c k o r splits, c o n k , rotten k n o t s , p r e s e n c e of A m b r o s i a , shatter, c a t f a c e , w e a t h e r c h e c k , butt s h a k e , goiter, b u r l s , p o w d e r w o r m , b a r k s e a m s , flare butt.  •  ring c o u n t : for v a l u a b l e t i m b e r s u c h a s s a w l o g s a n d g a n g l o g s , s e c o n d - g r o w t h b l o c k s u s u a l l y h a v e a thrifty a n d a h i g h g r a d e s o r t — b a s e d o n t h e n u m b e r of r i n g s p e r unit length.  •  v a l u e : a s e x p e c t e d , s o r t s with l a r g e r d i a m e t e r s , c l e a r of d e f e c t s a n d g r e a t e r ring c o u n t h a v e a g r e a t e r v a l u e s t h a n s o r t s that d o not satisfy t h e s e r e q u i r e m e n t s .  A s o n e c a n s e e from the a b o v e descriptions, the data collected about trees w h e n cruising a n d t h e s p e c i f i c r e q u i r e m e n t s for a log a r e not d e s c r i b e d in a s i m i l a r m a n n e r . S o m e of t h e s e s p e c i f i c a t i o n s a r e difficult to quantify, o r their q u a n t i f i c a t i o n m a y not b e n e c e s s a r y b e c a u s e t h e y  38  are not found in second-growth stands. To match tree descriptions in these two data sets (cruise and sort descriptions) some simplifications had to be made as described in Section 5.1.4.3.  5.1.4.2 Data Requirements to Compute Machine Productivity and Cost of Harvesting  The data requirements of this part of the model are mainly related to the level of detail desired and the objectives set forth. In the logging activity, when predicting machine productivity (m /hour) three levels of detail can be used: 3  •  Mean times for harvesting phase activities; this approach is efficient in terms of programming but offers the least amount of detail.  •  Regression models; this method also has advantages and disadvantages. The advantage is that, for most logging activities researchers selected the same independent variables for their regression models, which seems to make them portable. The main problem is that these equations were developed in specific conditions (different machines were monitored and different range for variables resulted) making their predictions less reliable in other conditions.  •  Simulation; represents the most detailed level, and it seems to provide good results for scheduling systems and interactions between phases within systems, but this is the least desirable method in terms of programming effort.  From the characteristics presented above, and all models that were presented in the literature review, it was concluded that simulation is not an appropriate method to be used in this project. This decision was because there are an infinite variety of situations existing in practice, and the accuracy offered by these models does not justify the effort to develop them. Therefore, in this study, productivity functions based on regression were chosen. When these were not available, averages were used. The majority of equations programmed in the model were developed by the Forest Engineering Research Institute of Canada (FERIC) in coastal second-growth stands.  39  When not available, other data developed in similar stands in Pacific North-West were used. All productivity functions and averages programmed in the model are presented in Appendix 2.  5.1.4.3 Basic Assumptions and Limitations of the Model  The main assumption made for the calculation of volume and value was that coastal secondgrowth trees are generally clear of surface and hidden defects. From experience (confirmed also during the field work that was undertaken during this project), it was known that only a few defects can be found in second-growth, i.e. shallow scars that do not affect the sort, fork or crook, dead or broken tops and sweep. To improve accuracy of predictions made by the model, cruises recorded more detailed data about these defects, as described in the field procedures (Chapter 6).  As expected, log dimensions have a major impact on grading. In this study, logs that can be bucked from a tree are based on total height of that tree (which is considered to be measured or estimated accurately), and diameters on the stem at different heights which are computed from a taper equation. Of great importance in log grading, was knot dimensions and spacing. More detailed information about this was collected in this project, as described in the field procedures.  The approach taken in this project consisted of considering the above mentioned factors to compute stand volume and value; all other data collected when cruising will be used as criteria for partial cut.  To compute the cost of harvesting, the main assumption was that stands are uniform, i.e. volume is evenly distributed over the entire area of the block. Another assumption, based on experience, was that blocks will not be stratified into more than 5 strata. In this part of the model, for each phase, the user can select from a wide variety of harvest system configurations. Several configurations can be simulated and the best one selected for the cutblock conditions.  40  D u e to t h e l a r g e v a r i e t y of s y s t e m s in t h e industry, a l m o s t n o r e s t r i c t i o n s a r e i m p o s e d w h e n a s y s t e m is configured.  T h e m o s t c o m m o n s p e c i e s in C o a s t a l B . C . w e r e i n c l u d e d in t h e m o d e l , n a m e l y D o u g l a s - f i r , w e s t e r n h e m l o c k a n d w e s t e r n r e d c e d a r . C o m m e n t s a b o u t i n c l u d i n g a d d i t i o n a l s p e c i e s in t h e m o d e l , a r e m a d e in t h e last c h a p t e r .  T h e f o l l o w i n g s e c t i o n s of this c h a p t e r , i n c l u d e a d e s c r i p t i o n of m e t h o d s u s e d t o c o m p u t e t h e total b l o c k v a l u e a n d c o s t of l o g g i n g .  5.2 Method of Calculating Block Volume and Value T h e f o l l o w i n g is a d e s c r i p t i o n of t h e B . C . i n v e n t o r y m e t h o d that w a s u s e d a s a f o u n d a t i o n to c o m p u t e v o l u m e a n d v a l u e of t h e b l o c k .  T h e cruise plan a n d the a c c u r a c y requirements are affected by the timber selling p r o c e d u r e s a n d t h e t i m b e r a p p r a i s a l . T h e m o s t c o m m o n a r e s c a l e - b a s e d s a l e s , for w h i c h t h e c r u i s e p r o v i d e s t h e b a s i s for e s t i m a t i n g t h e s t u m p a g e rate w h i l e t h e a m o u n t billed is b a s e d o n t h e s c a l e .  T h e t e c h n i q u e of c o l l e c t i n g c r u i s e d a t a c o n s i s t s of s y s t e m a t i c s a m p l i n g u s e d in c o n j u n c t i o n with s a m p l i n g with p r o b a b i l i t y p r o p o r t i o n a l to s i z e ( P P S ) ; i.e. p l o t s a r e p l a c e d s y s t e m a t i c a l l y in t h e b l o c k ( u n l e s s s p e c i a l c i r c u m s t a n c e s i n d i c a t e o t h e r w i s e ) a n d in e a c h plot, t h e p r o b a b i l i t y for o n e t r e e to b e m e a s u r e d (cruised) is p r o p o r t i o n a l to its s i z e (dbh). In B . C . ( a n d b a s i c a l l y in entire N o r t h A m e r i c a ) , t r e e s a r e s e l e c t e d u s i n g a p r i s m . T h i s t y p e of plot is u s u a l l y c a l l e d a p r i s m plot, o r a v a r i a b l e r a d i u s plot o r a point s a m p l e plot.  41  Brief description of variable radius plots method (adopted from LeMay and Marshall, 1991). The method establishes several fixed area plots of different sizes at each point. As a practical method of doing this, instead of measuring different radius for each tree size, an angle (called the critical angle) is projected from plot centre, using a prism. The plot size varies directly with the tree size and so the probability of a particular tree being "in" increases with the size of the tree. Consequently, the number of stems per hectare represented by each tree varies inversely with the tree dbh. In point sampling, the probability of tree selection is proportional to basal area and the large diameter trees therefore, are sampled with the same intensity as the small diameter trees. However, because the tree dbh and the plot size are inversely proportional, the. basal area per hectare for each tree tallied remains constant for a given critical angle and is called the basal area factor (BAF) of that angle. Therefore, for each critical angle there is a corresponding BAF. The BAF represents the basal area per hectare for a tallied tree and is defined as: BAF = ba * Ft t  where ba - basal area for tree "t". t  Ft - tree factor, which is the number of stems per hectare represented by each tree.  BAF is a function of the critical angle and is known for each prism. More exactly, each prism is calibrated as a certain BAF. To calculate the volume per hectare obtained with this method, the volume per tree must be calculated first. Next, the tree factor is calculated for each tree. The volume per hectare represented by each tree is then the volume per tree times the tree factor. The volume per hectare represented by each plot is the sum of volume per hectare represented by each tree for all tallied trees. The volume per hectare can be broken down by species. To obtain the stems per hectare represented by a plot, the tree factors have to be summed over all tallied trees. As with volume, the number of stems can be broken down by species, or by species and dbh classes. Average volume per tree is obtained by dividing the volume per hectare by the  42  stems per hectare represented by the plot. The average diameter can be calculated in a number of ways, the most common method is to weight each diameter by the tree factor: \(Ft*dbh ) Ave. dbh = ^= t  [1 ]  2>  The combination of several variable radius plots to obtain estimates of the stand is performed using per hectare estimates for the plots. The way in which per hectare attributes are combined will depend on the sampling design chosen. With all the above comments, for each species, the volume per hectare for sort "i" can be computed with formula [2]: yi=— n  [2]  where: y - volume (m/ha) for sort i. 3  i  y , - total volume from sort "i", (m) produced by all plots. 3  t  i  - current number of the sort, i = 1 .. m m - maximum number of sorts in the set.  n - total number of plots in the cruise. Total volume from sort "i" produced by all plots (y .) is computed with formula [3]: ;  n  ?.•.=__>*•  [3]  ;'=i  j - current number of plot (cluster), j - 1 .. n yij - volume (m/ha) of sort "i" produced in plot "j". 3  The total volume per hectare for a given species £y) can be computed by summing up volumes for each sort:  43  ? = __>,«=i  [4]  A complete description of theoretical basis of the method can be found in LeMay and Marshall (1991) and Avery and Burkhart (1994). Details are provided in the Cruising Manual, developed by the B.C. Ministry of Forests (1983).  The sample size (which is the number of plots to be placed in a block) is a function of the t-statistic, the coefficient of variation and the allowable sampling error. The coefficient of variation (which is the standard deviation expressed as a percentage of the mean), can be obtained by running a small initial sample (pilot sample), or may be estimated from experience gained from sampling similar stands. Since n is dependent upon t, and t is indirectly dependent upon n through its degrees of freedom, the sample size needs to be computed by iteration. In B.C., for scale-based sales, the allowable sampling error is 15% at a significance level of 95%. A block diagram for the process of computing stand volume and value is given in Figure 2.  44  Cruise data  Sort descriptions  Silvicultural system general data about operation (felling method, baf, min.diam., etc.) Yes  >  Is this a clear-cut ? No >  t  Display pre-harvest stand table  >  t  Select criteria for partial cut i  Display post harvest stand table  >  t  Accept stand structure ?  No  Yes Compute stand volume and value, volume by species and by sort, plus average stand characteristics.  Figure 2.  Block diagram for computation of stand volume and value.  The steps taken in this part of the model are: •  input a set of sort definitions that will be used to optimally "pencil buck" the trees.  •  input a set of variable plot cruise data for the area.  45  input general data about operation (BAF, method of felling, minimum diameter recovered, etc.) and select a silvicultural system. for each tree, a dynamic programming algorithm is used to "buck" the combination of logs that will maximize the value. Merchantable height and diameter at different heights on the stem are computed using the following taper equation (Kozak, 1988).  ^  _  _  Q  a  x  _ D j£  bZ t  2  +b  2  In (Z + 0 . 0 0 1) + b  3  -J~Z + b e 4  z  +b  s  ( D IH  )  [5]  where: dj - diameter inside bark at height hi (cm.) ao,a,a,bi,...b - regression coefficients 1  2  5  D - diameter outside bark at breast height i.e. dbh (cm.)  X=(l- ^717)(l-Vp)  [6]  V  h| - height from the ground (m) H - total height (m) p - inflection point (percentage of total height where function which represents the bole shape changes form) Z = hi / H  [7]  Values for regression coefficients (a,ai aa.b^.^bs) and percent inflection points (p) 0  were produced by Kozak by species, by mature and immature trees and by inventory zones in British Columbia (Kozak, A., 1996 - Pers. comm.). With this equation, the merchantable height for a given top diameter can not be calculated directly and must be obtained by iteration. Volume of individual logs is computed with Smalian's formula. volume and value of each log is accumulated into corresponding plot/species/sort. by averaging values for all plots, volumes and values per hectare, by species and by sort are obtained.  46  •  total volume and value per hectare are obtained by summing the above results.  •  given the total area, per hectare results are extended to the entire block.  •  other parameters of interest computed are: average dbh, average tree volume, average piece volume.  The items listed above make a complete description for analysis of a clear-cut. For a partial cut, the user will be asked to input the area of corridors, then the analysis will consist of two parts: •  corridors will be analyzed as clear-cuts.  •  for the block itself, the model presents the stand table. Then, an input form is displayed, where criteria can be input for the partial cut. The input from this form is used by the program to extract from the cruise data only the trees that correspond to the partial cut criteria. The following partial cut parameters are considered, separately for each species: 1. crown class. 2. diameter class. 3. tree class. 4. presence of pathological indicators (conk, scar, mistletoe etc.) or existence of trees displaying poor form (spiral grain, sweep).  According to criteria set in the previous form, the stand table after the partial cut is displayed. At this point, the user can switch back and forth between the form that displays the stand table after logging and the form that accepts criteria for the partial cut, until an acceptable stand structure is obtained. The analysis will be performed for trees that were selected, and results will be extended from a per hectare basis to the area of block minus the area of corridors.  With the above features, the model has the desired power and flexibility to address specific needs of both clear-cuts and partial cuts.  47  5.2.1  Optimal Bucking of Cruise Data  Cruised trees are analyzed one by one by a dynamic programming algorithm, which maximizes the value from each tree: 1. the tree is divided in "stages". The total number of stages for a given tree is based on the total merchantable height of the tree, which is determined by stump height and minimum diameter that has to be recovered (usually 10 cm.). In the program, each stage can take information about the following items: •  diameter of the stem.  •  diameter of knots.  •  knot spacing.  •  presence of sweep.  •  presence of fork or crook.  2. for each stage, diameter (or more exactly the large end diameter) is computed using the taper equation. 3.  information from the cruising sheet is attached to corresponding stages and used in the decision process (e.g. if sweep was recorded from 5 to 15 m., this will be attached to corresponding stages and used in the optimization algorithm).  4. the algorithm (forward pass) runs from the top to the bottom of the tree and finds for each stage the log that will maximize the value of the tree, from top up to that stage. 5. once the entire tree has been analyzed, a subroutine (backward pass) picks up the combination of logs that maximizes the value for the entire tree. This starts at the butt of the tree and proceed towards top.  Volume and value is accumulated per sort within each species and each plot. Results per hectare are computed based on individual plots. By multiplying these numbers by block area, block results are obtained.  48  A detailed description of the dynamic programming method, including how was it applied in this project is given in the remainder of this section.  Description of Dynamic Programming Algorithm Hillier and Lieberman (1974), state that there is not a standard mathematical formulation of the dynamic programming problem. Rather, dynamic programming is a general type of approach to problem solving, and the particular equations used must be developed to fit each individual situation. The same authors present the basic features which characterize dynamic programming problems. Based on these features and other studies (Pnevmaticos and Mann 1974, Raghavendra 1994), my final dynamic programming formulation, adjusted for the log bucking problem is presented as follows.  1. The problem can be divided into stages, with a policy decision required at each stage. The total number of stages for a tree will be found by dividing the merchantable height by the unit length of a stage, which is established in such a way to provide good results in a reasonable solution time. The policy decision consists in maximizing tree value up to each stage.  2. Each stage can have a number of states associated with it. In general, the states are the various possible conditions in which the system might be at that stage of the problem. In our case, for each stage there is only one state. However, there are a number of different decisions that can be taken at any stage, represented by the various logs (different sorts and different preferred lengths for each sort) that can be bucked.  3. The effect of the policy decision at each stage is to transform the current state into a state associated with stage (n-k), where k is the length of the log to be tried, expressed in stage intervals.  49  4. Given the current state, an optimal policy for the remaining stages is independent of the policy adopted in previous stages.  5. The solution procedure consists of two passes: the forward and the backward pass. It begins by finding (as the first step in the forward pass) the optimal policy for each state of the last stage (top of the tree). The solution of this one-stage problem is usually trivial.  6. The forward pass moves from the top to the butt of the tree, stage by stage, each time finding the optimal policy for each stage, until it finds the optimal policy when starting at the initial stage (butt).  7. The backward pass moves from the butt to the top of the tree in steps equal to log lengths. This very last step consists in picking-up the logs that produce the highest value of related stages, and thus the combination of logs that maximizes value of the entire tree.  In this approach, a problem may be produced by defects that extended over a small number of stage intervals. A good example for this is fork or crook. If this defect is present on a tree, the value will increase to the stage just before the defect, and then, because none of the sorts considered accepts this defect, the value will decrease for a number of stages. Specifically, the value of the tree will be less than the value before the defect, at least for a number of stages corresponding to the minimum preferred length. A solution to that, would be to advance stage by stage in the backward pass. However, this will dramatically increase the computing time for this algorithm (which is already very time consuming). To account for this problem, bucking will be enforced outside the dynamic programming procedure, at stages where a fork or crook is present. Consequently, optimization will be carried out for two stem sections, instead of one entire tree.  50  Recursive relationship formulations for dynamic programming in log bucking, were done by Pnevmaticos and Mann (1973), MacPhalen (1978), Faaland and Briggs (1984), Dykstra (1985), Young and Waddell (1986), Eng et al. (1986), Briggs (1989), Sessions et al. (1989) and Raghavendra (1994).  The dynamic programming recursive equation used in this project was developed starting from those used by Young and Waddell (1984) and Eng et al. (1986). Due to the different approach taken in this model, some modifications have been made.  The general form of the dynamic programming recursive relationship is given as follows: Let: n - current stage d - stage resolution (m) k - decision in stage intervals (no. of stage lengths to make log) I - log length (m) I = kd D - tree diameter at stage n (cm) n  D . - tree diameter at stage (n-k) (cm) n  k  C = 0.0000392699 - constant used for computing log volume with Smalian's formula U , (qi, q2, q3, q4, qs) - unit price ($/m ) for species S and sort g, subject to all the 3  s  g  following conditions (that can take true/false values): •  qi - condition related to small end diameter  •  q - condition related to large end diameter  •  q - condition related to knot diameters  •  q - condition related to knot spacing  •  q - condition related to sweep  2  3  4  5  51  V(D , D _K, I) - volume of log with diameters D respectively D .k and length I (m) 3  n  n  n  V(D„, D . I) = l ( D n  kl  2 n  n  + D . )C  [8]  2  n  k  R[V(D , D -k, I), U ,g (qi, q2, q3, q4, qs)] - log value based on volume and unit price ($) n  n  s  R[V(D , Dn. , I), U ,g (qi, q n  k  s  2 >  q , q , q )] = V(D„, D . I) * U , (qi, q , q , q , q ) 3  4  6  n  kl  s  9  2  3  4  6  [9]  The total length of the tree is divided into N stages as defined by the stage resolution (d) and the recursive relationship becomes as follows:  f„ = max { R[V(D , D . , I), U , (qi, q , q , q , q )] +  k  n  n  k  s  g  2  3  4  B  where f - maximum value of the tree from the top to stage n. n  f_k - maximum value of the tree from the top to stage (n-k). n  k - set of feasible sort lengths at stage n.  Figure 3 illustrates the computational scheme:  52  ta}  [10]  n+1 n n-1  n-k  R[V(D„, D ^ , I), U .g(qi, q , qs, q4, qs)] !< =»i n  2  s  32 1  Tn-k  f„ = max { R[V(D„, D -k, I), U ,g (qi, qa, qs, q4, qs)] + fn-k} s  n  Figure 3. Dynamic programming representation of stem.  Essentially, the principle is that given the optimal result for each of the preceding (n-1) stages, the sum of the value at the n-th stage after cutting a log of length k stages plus the value of the optimal result for the first (n-k) stages, can be calculated for all feasible log lengths. The log type and value of k that give the maximum value are retained as the optimal combination (decision) for the n-th stage. A block diagram for the dynamic programming algorithm is presented in Figure 4. The following notations have been used, assuming a given stage length: Let p - number of stages corresponding to minimum pulpwood length q - number of stages corresponding to the minimum sort length for a given species. It is well known that p < q N - total number of stages for a given tree n - current stage  53  Each stage will store the following information: •  value to stage n  •  log length - length of the log to be bucked at stage n to maximize tree value up to that stage  •  log volume  •  log value  •  sort name - name of the sort to which the above log was assign  Before the dynamic programming algorithm is started, data on current diameter, knot diameter and knot spacing are stored at each stage of the tree. Sorts are arranged in increasing order of length.  For an individual tree, at the end of algorithm described in Figure 4, each stage stores information that will maximize its value; i.e. maximum value up to that stage, log length, log volume, log value and the corresponding sort for that stage.  54  Tree Description: each stage stores information about tree diameter, knot diameter, knot spacing and sweep  I  For n = 1 to p assign Value of stage (n) = 0 Log volume = 0 Log value = 0 For n = p+1 toq assign Value of stage (n) = value pulpwood Log volume = volume computed with Smalian's formula Log value = value pulpwood n=q ^1 n = n + 1 Value of stage (n) = 0 Current sort = 1 Bucking length = k = length of current sort converted to stage intervals Check if all conditions for current sort are satisfied; i.e. Small end diam., Large end diam., Knots diam., Knots spacing, Sweep Are all conditions met ?  No  Yes  Log value = Value pulp * Log vol.  Log value = Value of current sort * Log volume  Temporary value = Log value + Value of stage (n - k) k Is Temporary value > Value of stage (n) ?  j-  No  Yes Value of stage (n) = Temporary value Save the best decision and value so far (all data necessary for this stage) Current sort = Current sort + 1  Bucking length = k = length of current sort converted to stage intervals No No  . |  (Is k > n) OR (at the end of sort set) ?  ]  Yes Is this the last stage of this tree (n = N) ?  j  Yes All the data necessary for this tree was computed  Figure 4. Block diagram representation of dynamic programming - forward pass.  55  Once the optimal decision and optimal value have been determined for each stage in the forward pass, the optimal bucking policy for the tree can be determined through a backward pass. The process starts at stage N (the tree butt) and obtains the optimal length of the first log to be bucked. Since the optimal solution has been determined and saved for all stages in the tree, the optimal length of the second log is known at the stage where the tree is bucked for the first log. This process continues until the complete bucking policy for the tree is determined.  The part of program related to computation of stand volume and value is described using pseudocode in Appendix 1 A. The forward pass of the dynamic programming algorithm is presented in the subroutine DynamicProgramming; the backward pass is described in the subroutine BackPass.  5.2.2  Factors That Can Affect Prediction of Stand Volume and Value and Methods to Account for Them  In the process of analyzing cruise data and optimizing value that is obtained from each tree, we acknowledge the existence of some factors that may affect prediction. In the first part of this section there is a presentation of these factors. Methods used to account for them are introduced later. 1. Factors due to sampling method.  2. Factors due to inventory procedure (independent of people doing the work and that can not be avoided). •  Errors due to defects that could not be detected during cruising (different types of rot and other internal defects).  •  Errors in estimating wood quality. This refers mainly to estimating the numbers of rings per inch. This is an important criterion in grading and may produce bias when evaluating the total value of the block. It is not possible to get these data during cruising, without a major impact on productivity and cost.  56  3. Factors that produce a loss in the total volume and value recovered due to harvesting procedures: •  Breakage during falling. Breakage will produce some volume loss because some broken pieces will become non-merchantabe and, value loss because it may produce downgrading of logs.  •  Breakage during primary transportation. Depending on the extraction method, additional breakage may occur while logs or trees are being moved from felling site to roadside.  •  Non-optimum bucking. Bucking takes place at the stump or at the landing. In the case of a full tree system, it may also take place at the sort yard. Non-optimum bucking occurs more frequently in the bush, and produces a loss in value, because the taller may be under production pressure or concerned about safety. Errors in bucking may also occur in other locations, however, with less severe impacts.  4. Factors that affect computations in the program (equation error). This consists primarily of errors in computing the taper of trees cruised. The following methods will be used to account for factors described above: 1. To analyze the sampling error, cruises will be more intensive than is legally required. Results produced by the model will be compared to results obtained after harvesting the block. Comparisons will be done by species and by sort using a t test. The following hypothesis will be tested: H : Value produced by model = Actual value obtained after scaling 0  Hi : Value produced by model * Actual value obtained after scaling The significance level (a) chosen is 95%; the t test will be:  Hr  1111  m  x  57  where: n  - total number of plots.  x - mean volume per species and sort computed from all n plots, v  - real volume after scaling.  S- - standard error of the mean. X  This is a two-tailed test, and the critical region consists of intervals (- °°, - t .i,a/2) and n  (tn-i,o/2,  °°). H is rejected if t test is in the critical region. If H is rejected the two values are 0  0  statistically different. If t test is outside critical region, we fail to reject H , and the two values 0  are statistically the same.  Because the distribution of volume by species and sort determines the value of the wood, this test will be the main validation of the model.  2. For inventory procedure factors, the following methods are considered: •  It is usually expected that defects that could not be seen during cruising will produce downgrading for some sorts and may cause an overall loss of wood. As presented earlier, it was known from experience that in coastal second-growth stands, little defect of this kind is encountered. The assumption was confirmed during field work carried out during this study, i.e. no defect of this nature was identified. Consequently, the model was not design to account for this type of problem when computing volume and value of individual trees.  •  Ring count is difficult to collect during cruising but should be considered because in some second-growth blocks there is a significant volume that satisfies the requirements of more valuable sorts (high grade instead of thrifty). In the model, a combined sort should be included, and its value should be weighted by the value of the two sorts in question. The numbers used in this process will come from the user's experience. For  58  example, if it is known from experience that about 20% of the saw log volume in a region is high grade and the rest is thrifty, these percentages should be used to compute a weighted average price for the combined thrifty/high grade saw log sort introduced in the set.  To analyze breakage, non-optimum bucking and equation error, fixed area plots were placed in the Stillwater block. To account for the three different harvesting methods used in the block, two plots were located in the area of each system, each one in terrain with different slope conditions. Within each plot, trees were cruised the same way as for the rest of the block, and numbered in such a way to ensure easy and accurate retrieval of each log. Trees were located after felling and bucking and marked accordingly. The length and diameters of logs were recorded. Logs were also tracked to the log sortyard, and their final dimensions (as taken by scalers), volume and company sort were recorded. The detailed field procedures for this phase will be fully described in Chapter 6.  3. A description of methods chosen to account for each factor follow: •  Breakage during falling: the approach taken to solve this problem is similar to the one taken by Deadman and Goulding (1979). The main purpose was to develop a frequency function that will predict the height of break point relative to total height. The differences relative to the previously mentioned function are: I. The breakage frequency will be a function of species and dbh class. Trees will be grouped in 20 cm. classes for which it is assumed that break point is the same . II. As other studies proved (Warren 1994), due to good falling techniques (the tree is felled parallel to the contour lines so as to minimize breakage), ground slope is not a significant factor in the breakage function, except for very steep slopes (greater than 60%). Therefore it will not be considered in this model.  59  In this model, the portion of stem above the break point is considered recoverable. Obviously, this will overestimate the volume harvested and the user will be warned about it in the manual that will accompany the software. Analysis was carried out on what happens if this assumption is not true, and the entire portion of the stem above break point is unrecoverable.  Based on practical observation that the volume of breakage differs greatly between manual and mechanical felling, breakage analysis was conducted separately for these two methods. Although not collected as part of this project, data from another FERIC project (Young, 1996) were used to analyze mechanical falling.  Breakage during primary transportation: previous studies (Copithorne and Young, 1994) showed that in Coastal B.C. this is about 0.2 % of volume. In this project, it will be assumed that loss in volume and value that occurs during this phase is negligible.  Non-optimal bucking: theoretical (pencil bucking) of standing trees (as cruised before felling) in the fixed area plots, was compared with what was actually achieved at the log sortyard.  Sampled logs from the fixed area plots (measured and marked in bush) were retrieved at the roadside and repainted to ensure easy retrieval. Next, logs were tracked to the log sortyard and scaling and grading were recorded. Company sort and B.C. Ministry of Forests grades were assigned to the log. At this stage, scalers from MacMillan Bloedel Ltd., made a substantial contribution, recording on their hand-held computers, the identification number of logs.  60  On the other hand, the cruise for fixed area plots were analyzed with the algorithm developed as a part of this project. For each tree, the combination of logs that maximized its value was determined.  The purpose of the analysis was to compare the value produced by each method, for as many trees as possible. Obviously, that necessitated that all logs (or at least the most representative ones) be identified at the sortyard. As presented in next chapter, due to practical considerations, this was only partially achieved. However, valuable conclusions could still be drawn, by comparing for each tree and for each individual log results from the model and the scalers at the sortyard.  4. To evaluate the effect of taper equation estimates, differences between predicted and measured diameters were evaluated. Diameters measured at stump height and at each one-tenth of the height, were compared to diameters produced by the taper equation at the same points. Predicted vs. real results were checked using the same statistics used by the author of the taper equation (Kozak, 1988), i.e. average bias (B) and the standard error of the estimate (SEE); formulae are given below:  n  B =  i=l  [12]  n where  Yi- measured value  Yi estimated value n number of observations  [13]  61  where m - number of parameters used in estimation, all the other parameters are the same as above. For these statistics, results were compared to those in Kozak (1988).  Considering all factors outlined in this chapter and methods to account for them, the t-test will be used to compare prediction vs. actual result by species and by sort. This will be presented in Chapter 7 - Results. The remaining factors are the taper equation estimates, evaluation of breakage and comparison between theoretical and real bucking. To evaluate the effect of these three factors, specific field work was done as presented in next chapter.  Method to Calculate the Cost of Harvesting  5.3  The cost of harvesting is determined by analyzing and comparing the results of several harvesting options. When this result is computed, the revenue generated by the block can be found, by subtracting the cost of logging from total value of the stand. In the second part of the model, the user defines the harvesting system by selecting the desired harvesting technique for each phase. The options offered for each phase are the following: •  Falling Mechanical: Feller buncher, Feller processor. Manual: Falling only, Falling and bucking.  •  At-stump processing: Delimber, Processor.  •  Log alignment: Excavator log alignment.  •  Primary extraction: Cable extraction: Grapple yarder, High lead yarder, Super snorkel Ground based extraction: Choker skidder, Grapple skidder, Short wood forwarder, Hoe forwarder  •  Roadside processing: Manual, Delimber, Processor.  62  •  Loading: Front end loader, Hydraulic loader.  The model allows stratification of the stand. Up to five different harvesting techniques can be used in different portions (strata) of the same cutblock.  The cost by phase and the total cost of logging is based on productivity equations and cost per hour of machines used in each phase. The bulk of these equations were developed by FERIC. In case the appropriate function was not found in a study carried out in the coastal region of B.C., equations developed in similar conditions were considered. The main parameters of the block (total volume, total value, average piece size etc.) will be available for this submodule, to compute the cost of the harvesting system.  The user is given full control over this part of the model. There is no restriction in configuring various harvesting systems and, values offered by productivity equations programmed in the model, can be overwritten anytime.  This portion of the model is represented in Figure 5.  63  Stratify t h e st,and a n d s e l e c t h a r v e s t i n g s y s t e m p h a s e s for e a c h st r a t u m .  >  t  D i s p l a y default m a c h i n e t y p e , productivity a n d c o s t b y phase.  Next phase  No  Next stratum  S e l e c t o n e of e x i s t i n g m a c h i n e s o r input a n e w o n e ; v a l i d a t e productivity a n d cost. f  Is this t h e last p h a s e for this s t r a t u m ? Yes  No  >  t  Is t h i s t h e last s t r a t u m for the block analyzed ? Yes i  Display harvesting costs and  NET REVENUE  F i g u r e 5.  F l o w c h a r t for c o m p u t a t i o n of h a r v e s t i n g c o s t a n d net r e v e n u e .  T h e s t e p s t a k e n in this part of t h e m o d e l a r e d e s c r i b e d a s f o l l o w s : •  e a c h b l o c k c a n b e split into five strata b a s e d o n l o g g i n g s y s t e m .  •  for e a c h s t r a t u m , a h a r v e s t i n g s y s t e m c a n b e c o n f i g u r e d u s i n g t h e o p t i o n s a v a i l a b l e for e a c h p h a s e .  •  b a s e d o n t h e s y s t e m , t h e u s e r will c o m p u t e t h e c o s t of e a c h p h a s e a n d e a c h stratum. Productivity, cost, yarding distance a n d other d a t a c a n b e c h a n g e d . At e a c h p h a s e , t h e u s e r c a n u s e t h e default m a c h i n e o f f e r e d b y t h e m o d e l o r input a n e w one.  •  b a s e d o n a b o v e n u m b e r s , total c o s t for h a r v e s t i n g t h e entire b l o c k will b e c o m p u t e d .  64  •  the revenue generated by the block will be computed by subtracting from the value of the stand the cost of the harvesting process.  The user interface of this section of the model consists of three forms. The code attached to each one of these forms is presented using pseudocode, in Appendix 1B.  Productivity equations (or averages) used in the model and machinery costs are presented in Appendix 2. As mentioned before, these data will be considered only as recommendations that can be anytime overwritten by the user, based on his experience and on better knowledge of available machinery and field conditions.  5.4 Model Validation As presented in the objectives of this study, after its completion, the model was validated. The Stillwater block was used to develop the model and, as part of model development, all factors that affect prediction were analyzed, as described in Chapter 5.2.2: •  taper equation was assessed.  •  a breakage model was elaborated.  •  an analysis of bucking was performed.  Predicted results by species and sort were compared with the actual ones. With all the above factors included, the model was validated by analyzing the Okeover block. This block was cruised using the method proposed in the project and the final result was compared with what the model predicted. For both blocks, different runs were produced, by incrementally including specific cruise data that were collected.  For the second part of the model (prediction of cost of harvesting) no validation was planned. This is due to the nature of this module, more exactly, due to the fact that it offers only some  65  approximate values for a specific harvesting phase, that will be validated by the user based on experience, machine availability and considering the specific conditions of the block.  Results obtained in the process of model validation are presented in Chapter 7.  66  6. FIELD DATA COLLECTION The first part of this chapter will describe the field procedures used to compute stand volume and value. Next, specific data collection methodology related to the design of the harvesting system, will be presented.  6.1 Data Collection for Prediction of Stand Volume and Value First, field procedures that were used to account for factors that can affect prediction of stand volume and value are introduced. As described in Chapter 5.2.2, these factors are: the taper equation, breakage and differences between optimum (theoretical) and actual bucking. Finally, the specific cruising procedures that were used in this project will be described.  6.1.1 Field Procedures For Checking the Taper Equation, Evaluation of Breakage and Comparison Between Theoretical and Real Bucking For fluency of presentation, a partial description of the method used to account for these factors was given in the previous chapter. This method consisted of laying out six fixed-area plots in the Stillwater block.  For statistical reasons, it was desirable that at least 30 trees (all logs of the tree) be retrieved after falling and tracked to the sortyard. From experience it was decided that at least twice the number of standing trees should be marked. Given the stand density of 550 stems/ha, a plot area of 0.125 ha was chosen. To account for specific terrain conditions, the shape of the plot was circular for grapple yarding and hoe-forwarding (radius of 20 m) and rectangular for direct loading (15 m x 80 m). In these six plots a total of 393 trees were recorded, with an average of 65 trees / plot.  67  Each tree in the plot was cruised the same way as for the rest of the block), and numbered to ensure easy and accurate retrieval of each log. Trees were located after felling and in-bush bucking and marked accordingly. The following data were recorded for each tree: •  number of the tree.  •  species  •  diameter at breast height (dbh)  •  stump height  •  diameter at stump height inside bark (as an average of two measurements) or, if this was not accessible, butt diameter of first log was measured.  •  information collected for each log is the following: 1.  log number  2. diameter outside bark (dob) at large end - only if previous log was not found or the stump was not accessible. 3. total length / length where breakage occurred - if this is the case. 4. diameter outside bark (dob) at the small end. 5. bark thickness (measured with a bark gauge).  Specific methods and formulae used to solve each of the three problems addressed in this section were given in Chapter 5.2.2.  6.1.2 Field Procedure for Cruising In general, cruises performed as part of this project were done in accordance with B.C. Ministry of Forests regulations. However, more details were collected for individual trees, to test if this would improve accuracy of prediction, namely: •  data collection about knots was completely changed; diameter and spacing of knots were estimated, by dividing the bole into uniform portions with respect to these  68  c  parameters. While cruising, it was noted that downed trees (naturally or for road right-of-way) could be used by cruiser to check his calls about diameter and knot spacing. •  for all defects recorded in the cruise tally sheet, height of occurrence was measured if possible, or at least estimated.  •  for defects that affect a large portion of the stem (sweep was the only defect of this nature that was encountered), apart from height of occurrence, total length was measured or estimated.  Higher intensity cruising was performed during field trials, with a view to analyze how sampling intensity affects prediction. For the Stillwater Block, 39 plots including 191 trees, were laid out, and for Okeover block 26 plots including 162 trees were used. The cooperating companies placed 21 operational cruise plots (92 trees) in Stillwater block, and 15 plots (96 trees) in the Okeover block.  6.2 Data Collection for Harvest System Design No specific field procedures were carried out as part of this submodule. As mentioned in the study design, existing equations developed as part of other studies were implemented. The user has the possibility either to use the results provided by these equations, or to overwrite them, based on experience, knowledge of specific site conditions and machinery used.  This submodule is mainly theoretical, the main concern being how to offer a user-friendly and comprehensive interface, rather than how to collect field data.  69  7. RESULTS AND DISCUSSION The model was thoroughly tested both at the individual tree level (the combination of logs that maximizes value for each tree) and at the stand level. It was noticed that due to the large number of computations and their complexity, this part of the program was very time consuming. To account for this problem, the model was designed to accept two stage lengths for the dynamic programming algorithm, i.e. 0.1 m. and 0.2 m.. The first choice is designed to handle detailed analysis, whereas the second one can be used for quick analysis. Accuracy of results produced by these two options depends mainly on the lengths of sorts that are to be bucked. If these lengths are multiples of 0.2 m., then the second option can produce results very close to the first one. If the above condition is not met, then results may be quite different.  In terms of time performance, when running the model on a 486 machine, with 16 MB RAM, it required about 3 seconds per tree for the version quick analysis (0.2 m. stage length) and almost double for detailed analysis (0.1 m. stage length). For clear-cuts, to find the total time to analyze a cruise, these numbers should be multiplied by the number of trees. For partial cuts, time will increase proportionally to the number of trees that will be processed (removed) outside corridors (in the residual stand).  Although the model was designed primarily to handle data in the format that was recorded in this project, it can be very easily used to analyze standard cruises. For cruises performed according to B.C. MoF regulations, the model offers default values for fields that are not included in the tally sheet. Details on how to address this process are given in the user manual.  While the model was developed to assess both clear-cuts and partial cuts, only clear-cut results were validated. The rest of this chapter will present results produced by the model. The first subchapter will introduce evaluation of factors that may affect prediction of stand volume and value.  70  7.1 Results for Evaluation of Taper Equation. Evaluation of Breakage and Comparison Between Theoretical and Real Bucking Results presented in this chapter were obtained using principles and formulae described in Section 5.2.2. Specific comments and additional information is given for each item addressed.  7.1.1  Evaluation of Taper Equation  Overall, 283 measurements were used to check the taper equation and to develop the breakage function. Results are presented in Table 2. The format chosen to present these data is the same that was used in Kozak (1988). Diameters were measured at the ends of logs and then, grouped together by percentages of total height of the tree. Both bias and standard error of the estimate are presented in cm.. No attempt was made to represent them in percentages, since this way of presentation is affected by the base to which the comparison is made, and many times misrepresents results.  In the data analysis process, some measurements were eliminated. The most frequent case was when details pertaining to a tree were not recorded; e.g. it was noted that a log was missing, therefore the height where a diameter was measured could not be computed. Eight parameters are used by the taper equation to estimate a diameter. Therefore, the standard error of the estimate (SEE) was computed only when the number of measurements (n) available was greater than or equal to 9.  71  Table 2. Biases and Standard Errors of Estimate - S E E of diameter inside bark from ground to top. Height from ground <0.5 20% 30% 40% 50% 60% 70% 80% 90% All  n 24 3 9 6 9 7 16 5 10 89  Douglas-I ir Bias (cm) 1.03 1.10 1.09 1.05 0.97 1.08 1.00 0.84 1.01 1.02  SEE (cm) 1.78  21  Cedar Bias (cm) 0.88  4 5 7 8 10 4 59  1.01 1.16 1.03 1.05 1.12 1.09 1.05  n  SEE (cm) 1.84  2.04  2.09 1.14 1.76  1.54 1.69  54  Hemlock Bias (cm) 0.97  7 19 8 5 22 16 4 135  0.97 0.96 0.99 0.95 0.90 1.08 1.19 1.00  n  SEE (cm) 1.98  1.71  2.02 1.21 1.73  Results in the table compare well with results presented in Kozak (1988). According to Kozak (1997 - Personal comm.), an acceptable limit would be 2.0 cm. for bias and 4.0 cm. for S E E . Results obtained in this project are well under these limits. As one can see in Table 2, values for bias are positive for all species and all classes. This is despite the fact that some real measurements were greater than theoretical ones. However, this doesn't affect accuracy of computations because of the small difference between values, and because of the scaling rule used in B.C.: all diameters are rounded-off to the nearest even number.  In conclusion, the taper equation predicted diameters very well in this study.  7.1.2  Evaluation of Loss due to Breakage  This analysis was performed to estimate a ratio of breakage, in terms of how many trees will break (by species and by dbh classes), and percentage of total height where breakage will occur. Also, the portion of the stem that has to be trimmed off after breakage was estimated according to the same criteria.  72  To extend applicability of the model, analysis was carried out both for manual and mechanical felling. Results obtained for manual felling are presented in Table 3:  Table 3. Species CE CE CE CE DF DF DF DF HE HE HE HE  DBH limit (cm) 20 40 60 >60 20 40 60 >60 20 40 60 >60  Analysis of breakage for manual felling. Breakage frequency (%) 50 100 100 100 50 100 100 100 50 100 100 100  Height of break point (% of Total Height) 60 74 77 72 60 60 69 59 60 66 67 59  Length of trim (m) 0.4 0.8 0.8 0.8 0.4 0.8 0.8 0.8 0.4 0.8 0.8 0.8  The frequency of breakage and length of trim are presented after rounding numbers. The actual numbers were within ± 10% of numbers displayed. This solution was chosen to avoid complications related to programming tasks.  Average heights of breakage (as a percentage of total height) per species are: Cedar 71%, Douglas-fir 62%, Hemlock 63 %. As one can see in the Table 3, the length of stem that had to be trimmed off due to breakage was the same for each species. Also, this length was the same for all trees with dbh greater than 40 cm.  For mechanical falling, the only species available was Douglas-fir. Results obtained from the analysis are presented in Table 4.  73  Table 4 . Analysis of breakage for mechanical felling of Douglas-fir. Species  DBH limit (cm)  Breakage frequency (%)  Height of break point (% of Total Height)  Length of trim (m)  DF DF DF DF  20 40 60 >60  50 100 100 100  81 75 85 78  0.4 0.8 0.8 0.8  Results obtained for Douglas-fir were extended to other species as well. For the same reasons presented above, breakage frequency and length of trim were rounded.  As one can see comparing Tables 3 and 4, the only difference between the two falling methods consists of the height (as percentage of total height) where breakage occurs. For mechanical felling, this averaged 80% for Douglas-fir and was used for all species in the model.  The model also compares the height of break point and the merchantable height for each tree. Depending on the result produced by this comparison, a tree can be analyzed as one or two separate portions. The model assumes only one break point for a tree, i.e. if the top log is long enough, it is considered fully recoverable and analyzed by the optimization algorithm. Obviously, this will produce an overestimation of volume of the stand. Comments about this situation are made in the remainder of this chapter. Detailed description of these procedures are provided in Appendix 1.  Although none of the blocks analyzed in this study were mechanically felled, test runs of this method evaluated differences between the two falling methods. Results are presented in Section 7.2.  74  7.1.3 Comparison Between Theoretical and Actual Bucking The comparison of bucking efficiency was seriously affected by the fact that the cooperating company changed the method of scaling during the project and manually scaled only one in seven loads. This resulted in 241 logs being lost. Eventually, an agreement was reached and special attention was given to loads that included logs marked in this project. However, due to this perturbation, only logs from three plots (out of six) were used for comparison between theoretical and real bucking. Although each plot was selected to represent one of the harvesting systems used, this reduced the size of the sample to half.  A number of 83 logs were found at the sortyard, of these, 8 were eliminated because the values measured did not agree with the initial cruise data. It is believed that this was produced by misreading of identification numbers that were covered with mud during logging. The final situation is presented in the Table 5. In Table 5 and in the rest of this section, the word match is used when actual and predicted logs were the same. Since each tree is analyzed in the same way, the analysis was not carried out by species.  Table 5 . Match of logs found at the log sort yard. No. Logs Found From a Tree One log Two logs Three logs TOTAL  No. Trees  No. Logs  No. Logs That Matchec  30 21 1 52  30 42 3 75  19 29 0 48  As shown in the table, there was only one tree for which three logs were found. Unfortunately the bottom log was missing and the ones found did not match, thus making any analysis impossible.  A note should be made that in the first class (one log found) ten of the logs recorded were log number two. It is obvious that, when butt diameter of a real second log and of the second log produced by the model were equal, the previous log (if found) would have matched as well. This happened in six cases. When two non-consecutive logs were found (logs number one and three)  75  and they matched the model logs, the log in between would also match. This happen three times in this analysis. These missing logs were assumed to validate model performance.  Overall, results for sampled trees that were tracked from the woods to the sortyard, can be considered good, but not enough logs were recovered to validate the model. Because all logs were not retrieved for any plot, it is difficult to evaluate the role played by the breakage model. The final scale for the entire block will be used as the main criteria to test model performance.  7.2 Comparison of Block Volume and Value - Predicted vs. Actual Analysis of block value and the influence of different factors was conducted for the two blocks. Different runs were produced to account for all factors that could influence the final result. To ensure maximum accuracy, all these runs used a dynamic programming stage interval of 0.1 m.  Each set of values obtained with the procedure described above, was compared with the scale obtained after the entire block was harvested. This comparison helped draw conclusions about the effect of more detailed tree description on accuracy of predictions.  Results are presented separately for each block.  7.2.1 Results for the Stillwater Block Fourteen runs were considered in the analysis of the Stillwater block. Factors included in each one of these runs and results produced are presented in Table 6.  These predictions are split into four groups as presented below: •  First group (Predictions 1 -3): excludes the effect of breakage. This is an unrealistic assumption but, it was considered to calculate the maximum volume and value per  76  hectare for the stand. Downgrading factors were included in groups of two, i.e. fork and sweep first and then knot diameter and spacing. •  Second group (Predictions 4-8): includes the effect of breakage. Since this is what happens in real life, the analysis was conducted in more detail, and factors were included one by one. The most complete is Prediction No.8. which represents the actual function programmed in the model.  •  Third group (Predictions 9-10) was done to test the capabilities of the model. As presented before, the model assumes that from each tree that breaks (based on frequency presented in previous chapter), the top section is recoverable. Of course the shattered portion between the two sections is lost. Prediction 9 was done assuming that breakage is so severe, that the top log is entirely lost. Since a breakage frequency function was developed for mechanical falling with no chance to test it, Prediction 10 was done to estimate volume recovery had the first block (Stillwater) been mechanically felled.  •  Fourth group (Predictions 11-14) is described below. 1.  Prediction 11: presents results obtained by the cooperating company, based on their operational cruise, and using Ministry of Forests compilation program. Because ministry grades are used, for this prediction, only volumes can be compared with the model.  2.  Prediction 12: includes results obtained using company cruise in the model.  3.  Prediction 13: was carried out to analyze the influence of the sample size and to make it comparable with the sample size used by the company. Even number plots were used in this run.  4.  Prediction 14: same as above but using odd number plots.  All predictions are described in Table 6. The actual total volume recovered from the block was 630 m /ha. Due to an agreement with the cooperating company, value is not disclosed. Instead, 3  the percentage between value at different runs versus actual value is presented.  77  Table 6 .  Prediction No. 1 2 3 4 5 6 7  8 9 10 11 12 13 14  Fourteen predictions of total volume and value for the Stillwater block . 1  Factors Breakage Fork Sweep KnotDiam. Knot Sp. X  X  X  X  X  Volume  X  Volume m /Ha 694 694 693 665 665 665 665  110 110 110 106 106 106 106  112 111 108 107 106 106 104  X  664  105  103  608 681 561 550 761 570  97 108 89 87 121 90  99 107  X X  X  X  X  X  X  X  X  X  X  X  X  X  Top Section Not Recoverable Mechanical Falling Company Cruise & MoF Compilation Company Cruise Analyzed with the Model Feric Cruise - even no. plots Feric Cruise - odd no. plots  3  (%)  Value  (%)  -  86 120 86  The four groups of predictions will be presented one by one in the rest of this chapter. Results will be presented by species and by sort. For reasons presented in the previous chapter (i.e. percentages may be misleading), percentages are given only for totals by species and the grand totals . For comparison, each table will also contain the actual scale result obtained at the log sortyard.  First Group - Predictions 1-3: Results are presented in Table 7. As expected, the volume per hectare produced by these runs was above the actual volume. Total volume was 110% of actual and volumes per species ranged from 101% to 140%. The overestimation of volume for hemlock may look severe, but this species represented only 21% of total volume in the actual scale. Volumes of the three runs are not exactly the same because of rounding and because including a defect may cause some portions of a tree to become unmerchantable. Some variations may be due to Smalian's formula.  1n this table Knot Diam. stands for Knot Diameter and Knot Sp. is the abbreviated form for Knot Spacing. 1  78  Table 7 .  R e s u l t s of p r e d i c t i o n s 1-3 v s . a c t u a l , b y s p e c i e s a n d sort. Actual  Sp.  Sort  Vol/Ha  m CE  DF  Saw Log  3  m  Prediction 2  %  Vol/Ha  m  3  %  Prediction 3 Vol/Ha  m  3  8  6  6  6  22  38  38  33  Chip & S a w  11  13  13  17  Utility  2  0  0  0  Pulp Total C E  14 57  8 65  8 65  9 65  S a w Log  281  265  264  265  Gang  106  167  167  124  Chip & S a w  20  0  0  17  Utility  9 27  2  2  24  114%  Total D F  443  13 447  Saw Log  15  28  28  23  Gang  77  121  121  119  Chip & S a w  10  0  0  2  Utility Pulp Total H E  VOLUME VALUE  101%  14 447  114%  101%  %  3  Gang  Pulp HE  Prediction 1 Vol/Ha  16 446  114%  101%  1  10  10  15  27 130  23 182  140%  23 182  140%  23 182  140%  630  694  110%  694  110%  693  110%  112%  111%  108%  B y c o n s i d e r i n g fork a n d s w e e p (there a r e 11 t r e e s with fork a n d 1 t r e e with s w e e p ) , v a l u e p e r h e c t a r e d r o p p e d b y 1%. K n o t c h a r a c t e r i s t i c s p r o d u c e d a n a d d i t i o n a l d r o p in v a l u e of 3 % .  Second Group - Predictions 4-8:  v a l u e s a r e p r e s e n t e d in T a b l e 8 . S i m i l a r to t h e first g r o u p ,  v o l u m e p e r h e c t a r e i s a b o u t t h e s a m e for t h e s e five p r e d i c t i o n s but b e c a u s e b r e a k a g e w a s c o n s i d e r e d , t h e v o l u m e p r e d i c t e d w a s r e d u c e d b y 5 % for p r e d i c t i o n N o . 8 a n d 4 % for t h e o t h e r s .  B y i n c l u d i n g d e f e c t s o n e b y o n e , v a l u e p e r h e c t a r e d r o p p e d f r o m 1 0 7 % to 1 0 3 % relative to a c t u a l v a l u e . T h e s e v a l u e s a r e r e a s o n a b l e b e c a u s e , t h e m o d e l a s s u m e s that for b r o k e n t r e e s , t h e t o p s e c t i o n is entirely r e c o v e r a b l e . N o t e that t h e o n l y t r e e that d i s p l a y e d s w e e p d i d not p r o d u c e a n y d e c r e a s e in v o l u m e o r v a l u e . A s for p r e d i c t i o n of v o l u m e p e r s p e c i e s , a g a i n t h e w o r s t p r e d i c t i o n is for h e m l o c k b e c a u s e of t h e s m a l l s a m p l e .  79  Prediction 8 represents the final version of the model. Overall, overestimates of 5% in volume and 3% in value were achieved.  The detailed distribution of volume per sort for each species for Predictions 4 - 8 is given in Figure No. 6.  80  : 0  SI  s  :CM •CO  00  c o  SI -D CD  Itt  CC  I  co in cn  o o  I CO  co  CM  O CM CM I -  •CM CO CM  s  co CD  CO CM  |rc  I5>  IS  |CO  0  TJ  co  X  1>o  CO  in  co  I  co m  in  co  ol  :CO •CO  s  s  CO  iri:  CD CO  :1 -  I T3 X CD  in  00  0 -  I co c o o  CM 00 CM  m  CM  CM :CM  •CO in •I - CD CD  CD  s  CO  ro 3  o  (0  w >  op  4  co c  g Tj ~o CD Q. O  in 32 3 (0 CD  o  0.  m  CO  CD  CM  Jo l>  CM  I COCM  CD O  CM  iCO  oa  _ J  ° CO  CO  9- • co 'sz  I©  o :  i  o _i  CO oei CO co CO  .9-1  sz  O  s  5  CO  111 O  CO  o  CD  I- iO CM :CO  s  5  co  ol  iJ2' in co  CO  I-  15  >  O)  cnl  •00 CM •CM  CD CO CO CO  co  '1 o  in co co  •CO •CO  CO  CDIX  co  o>  :05 CM ;CM  CO  I CDm CO  I"* c o  CO 0>  is  t_  rr n cc  co  TJ X CD  o  ol  :CO •CO  I ID c o  o i CO 31  0_i  co CO  ob  9co co 'sz CD  CO  IS  o :  o co CD  LU  • LU  2  ,_l  o< > ; Q4 > IX  Figure 6. Results per Species Produced with the Model (Predictions 4 - 8)  SawLog Figure 6a.  Gang  Chip&Saw  Utility  Pulp  Distribution of volume per sort for Western Red Cedar.  Figure 6b.  Distribution of volume per sorts for Douglas Fir.  • Actual SPred.4 DPred.5 UPred.6 HPred.7 BPred.8 Figure 6c.  Distribution of volume per sorts for Western Hemlock.  82  As seen in Table 8 and Figure 6, by introducing factors one by one, the volume distribution gets closer to the actual. The exceptions are the sorts Chip & Saw for Cedar and Utility for Hemlock. A possible explanation for this, may be that a detailed description for these sorts was never obtained and the description used in the analysis had to be mainly inferred from the B.C. MoF Scaling Manual.  The Student t-test described in the study design, was used to compare for each species and sort, if the values produced by Prediction No.8 and the actual values, are statistically the same. This test proved that for the entire distribution only three sorts are statistically different: cedar pulp wood sort and hemlock chip & saw and utility sorts. Various factors could produce this effect, but none of them can be clearly identified. One of them could be the incomplete description of the two sorts mention above. Another factor could be that these two species are less represented in the stand (both comprise 35% of total volume), therefore the sampling intensity is lower for these species. For the most representative species in the stand (Douglas fir), real and predicted values are statistically the same for all sorts.  Third Group - Predictions 9-10: Prediction 9 shows results of the model, assuming the top log of broken trees is not recoverable. This produced an underestimation of volume and value, as displayed in Table 9. Prediction 10 was done as if the block was mechanically felled, the result being an overestimation of volume and value. In fact, these values are very close to the ones produced by Prediction 3, when all defects were considered but, it was assumed that there is no breakage. This shows that height of breakage for mechanical felling is close to merchantable height and therefore, does not produce great losses. Results of these predictions are presented in Table 9.  83  Table 9 . Results of predictions 9-10 vs. actual. Sp.  Actual Vol/Ha  Sort  DF  HE  Saw Log Gang Chip & Saw Utility Pulp Total CE Saw Log Gang Chip & Saw Utility Pulp Total DF Saw Log Gang Chip & Saw Utility Pulp Total HE VOLUME VALUE  Prediction 10 Vol/Ha %  m  m  m  8 22 11 2 14 57 281 106 20 9 27 443 15 77 10 1 27 130 630  6 28 20 0 7 61 254 113 0 6 8 381 23 119 1 11 12 166 608  6 30 19 0 8 63 264 123 11 24 16 438 23 119 2 16 20 180 681  3  CE  Prediction 9 Vol/Ha %  3  3  107%  86%  128% 97% 99%  111%  99%  138% 108% 107%  As expected, the real scale value is between Predictions 8 and 9. Obviously, Prediction 10 yields a value greater than all preceding three values.  Fourth Group - Predictions 11-14: Results are presented in Table 10. Brief comments about each one of these predictions follow: •  Prediction 11: represents the actual values calculated by the MoF cruise compilation program from the operational cruise. It underestimates the total volume by 11%; an underestimation of 22% occurs in total volume of Douglas Fir, and an 21% overestimation of Hemlock. Since this prediction does not use company sorts, there are no data either for distribution of volume per sort or about values obtained.  84  •  Prediction 12: is obtained by compiling the operational cruise with the model. Volumes per specie are less accurate than the previous prediction, but still comparable.  •  Prediction 13: uses the FERIC cruise but takes only even number plots. Due to the small sample taken, the result is less accurate and overestimates volume and value for all species and consequently on a per hectare basis.  •  Prediction 14: similar to the previous one, but using odd number plots. As expected, it results in an underestimation of volumes and values.  Table 10 . Results of predictions 11-14 vs. actual. Sp. Sort  Actual Prediction 11 Prediction 12 Prediction 13 Prediction 14 Vol/Ha Vol/Ha % Vol/Ha % Vol/Ha % Vol/Ha % m  CE Saw Log Gang Chip & Saw Utility PyjP.  Total CE DF Saw Log Gang Chip & Saw Utility Pulp.  HE  Total DF Saw Log Gang Chip & Saw Utility Pulp.  Total HE VOLUME VALUE  3  8 22 11 2 14 57 281 106 20 9 27 443 15 77 10 1 27 130 630  m  m  3  57  100%  347  78%  157 561  121% 89%  2 27 7 0 6 42 217 104 0 3 12 336 23 109 0 15 25 172 550  m  m  4 38 26 0 5 73 322 116 27 11 27 503 33 120 0 11 21 185 761  8 21 13 0 10 52 191 122 18 5 18 354 13 119 2 10 20 164 570  3  3  74%  76%  132% 87% 86%  128%  114%  142% 121% 120%  3  91%  80%  126% 90% 86%  As one can see, these last predictions are quite different from the actual value, but this is due mostly to the fact that they are based on a much smaller sample size, about half of what was used in previous predictions.  85  7.2.2 Results for Okeover Block The analysis was conducted in a manner similar to Stillwater; the fourteen different runs arranged in four groups are presented in the Table 11. In this study, cedar was eliminated and analysis was conducted only for Douglas-fir and Hemlock. This decision was taken because it was impossible to separate the logs produced from old-growth from those produced from second-growth trees. Another specific point for this block was that the main product that was supposed to be obtained from Douglas-fir, was peeler logs. This sort was not used in the first block, and its presence affected the analysis.  Unfortunately, actual values obtained in this second block were not disclosed (neither values per sort nor the total value). To run the analysis, some estimates have been made, based on values used for the other block. Since the real value was not available, the percentage between value of different runs is presented versus value produced by the model (Prediction No. 8). This is based on the observation that both volume per hectare and distribution of volume per species and sort produced by the model are close to actual numbers, therefore, predicted value should be close to the actual one, as well. The actual total volume obtained was 679 m/ha. 3  Table 11. Fourteen predictions of total volume and value for the Okeover block. Prediction  Factors  Volume  Volume  No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14  Breakage Fork Sweep Knot Diam. Knot Sp.  m 684 684 681 658 659 659 658 658 615 676 565 552 543 770  (%) 101 101 100 97 97 97 97 97 91 100 83 81 80 113  X  X  X  X  X  X  X X  X  X  X  X  X  X  X  X  X  X  X  X  X  Top Section Not Recoverable Mechanical Falling Company Cruise & MoF Compilation Company Cruise Analyzed with the Model Feric Cruise - even no. plots Feric Cruise - odd no. plots  86  3  Value (%)  120 120 102 116 116 116 100 100 98 102 -  92 82 118  The four groups of predictions will be presented one by one (by species and by sort) in the rest of this chapter. Each table will contain also the actual scale result obtained at the log sortyard, for comparison.  First Group - Predictions 1-3: as one can see in the Table 12, the predicted total volume is almost identical to actual; this should not be the case for these predictions since all of them are based on assumption that when breakage occurs, the top log is recoverable. Analyzing results per species, it is obvious that numbers obtained for Douglas-fir meet expectations, whereas values for hemlock, do not. As presented later, this was the case throughout the entire analysis, which suggests that hemlock was not uniformly spread in the block, and the cruise did not sample it properly.  Fork and sweep (there are 7 trees with fork and 3 trees with sweep) has almost no effect on value, but knots produced a great drop in value of 18%.  Table 12. Results of predictions 1-3 vs. actual, by species and sort. Sp.  Sort  Actual Vol/Ha m 44 259 90 64 35 492 5 79 50 40 13 187 679 3  DF  HE  Saw Log Peeler Gang Chip & Saw Pulp Total DF Saw Log Peeler Gang Chip & Saw Pulp Total HE VOLUME VALUE  Prediction 1 Vol/Ha % m 0 430 8 81 6 525 0 126 0 29 4 159 684  Prediction 2 Vol/Ha % m  3  107%  85% 101% 120%  87  3  0 430 8 80 7 525 0 126 0 29 4 159 684  Prediction 3 Vol/Ha % m 39 302 66 96 18 521 9 82 28 34 7 160 681 3  107%  85% 101% 120%  106%  86% 100% 102%  Also, to evaluate loss in value due to knots, from Tables 10 and 11 it is obvious that the value predicted when knots were considered but without considering breakage, was less than the value obtained when breakage, fork and sweep were examined.  Second Group - Predictions 4-8: values are presented in Table 13. Volume per hectare is about the same for these five predictions but, due to the lack of accuracy when computing the volume for hemlock, it is less than actual.  Similarly with the previous group, fork and sweep did not produce any significant changes, but value per hectare dropped 16% when knots were considered. In terms of distribution of volumes per sort within each species, as one can see in the Table 13 and Figure 7, for predictions No.4 8, by including more factors in the model, this distribution gets closer to the actual.  88  O)  CD  CD  cn  CD  CD  CD  O) 00  CD CO  3  O CO  w >  CD  CD  00 I  •tf w c 0  CD  1 a CD  o w 3 w  CD DC  LO  CO  CO  _0 co LL  LU X  Qj"c0  aiaj 3  a  Figure 7. Results per Species Produced with the Model (Predictions No. 4 - 8)  Figure 7a.  Distribution of volume per sort for Douglas Fir.  ca  x  u  E, <u  E _j o >  SawLog  Peeler  Gang  Chip&Saw  Pulp  • Actual HPred.4 DPred.5 ^Pred.6 HPred.7 HPred-S  Figure 7b. Distribution of volume per sort for Western Hemlock.  90  Hemlock gang was the only sort for which predicted value was different from actual, as determined by the t-test (Prediction 8 vs. real values). This is due to the strict requirement on knots size, which in reality caused some downgrading from peeler to gang. The overall lack of accuracy in prediction of hemlock volume for this block, must have played a role. This problem was solved as part of another FERIC project, although results were not available at the time cruising was performed in this second block. The solution consisted of requiring the cruiser to estimate if the tree analyzed was appropriate for peeling (by assessing the uniformity of its cross section and lack of other surface defects), and record this in an additional field. The optimization algorithm has been modified to check if dimensional requirements were met, and in case of a peeler, the cruiser's call was used to override conditions related to knots. If available in time, this method would have assured a better distribution of volume between peeler and gang, and thus improve the overall accuracy of this prediction. Also, missing the real value of sorts, must have played a role in this analysis, by changing the log that maximized tree value at a given stage in the dynamic programming algorithm. It should be noted that the only significant difference occurred for the least representative species in the stand. For Douglas-fir, in all sorts, predicted and actual volumes, were statistically the same.  Prediction 8 represents the final version of the model. Overall, an underestimation of 3% in volume was achieved. No conclusion can be made about value.  Third Group - Predictions 9-10: results are presented in Table 14. Prediction 9 shows results of the model, assuming the top log of broken trees is not recoverable; Prediction 10 simulated a mechanically felled block. Results are good for Douglas fir, but overall results per hectare are affected by the imprecision related to hemlock.  91  Table 14 . Results of predictions 9-10 vs. actual. Sp.  Actual Vol/Ha  Sort  m  3  DF  HE  Saw Log Peeler Gang Chip & Saw Pulp Total DF Saw Log Peeler Gang Chip & Saw Pulp Total HE VOLUME VALUE  44 259 90 64 35 492 5 79 50 40 13 187 679  Prediction 9 Vol/Ha %  Prediction 10 Vol/Ha %  m  m  38 299 61 47 20 465 9 82 27 24 8 150 615  39 302 66 86 25 518 9 82 27 33 7 158 676  3  3  95%  80% 91% 98%  105%  84% 100% 102%  As expected, Prediction 9 yields a value less than No. 8, whereas value given by Prediction No. 10 is greater than the one produced by No.8. However, the differences between these three predictions are smaller than in the first block, due especially to the larger trees found in the first block.  Fourth Group - Predictions 11-14: results are presented in Table 15. Brief comments about each one of these predictions are given as follows: •  Prediction 11: represents the actual values calculated by the ministry cruise compilation program from the operational cruise. It underestimates the total volume by 17%; an underestimation of 25% occurs in total volume of Douglas Fir, and an 6% overestimation of Hemlock; since this prediction does not use company sorts, there is no data about distribution of volume by sort.  •  Prediction 12: is obtained by compiling the ministry cruise with the model. Volumes per species are less accurate than the previous prediction produced, but still comparable.  92  •  Prediction 13: uses the FERIC cruise but takes only even number plots. The result is less accurate. It underestimates volumes for most sorts and consequently on a per hectare basis, however, it seems to yield the best result for Hemlock.  •  Prediction 14: similar to the previous, but using odd number plots. As expected, it results in an overestimation of total volume. In terms of distribution of volume by sort for hemlock, produced by the last two predictions, the only explanation found was that this species is spread non uniformly (in patches) in the stand and the cruise could not sample it properly.  Table 15 . Results of predictions 11 -14 vs. actual. Sp. Sort  Actual Vol/Ha m 44 259 90 64 35 492 5 79 50 40 13 187 679 3  DF  HE  Saw Log Peeler Gang Chip & Saw Pulp. Total DF Saw Log Peeler Gang Chip & Saw Pulp. Total HE VOLUME VALUE  Prediction 11 [ Prediction 12 Prediction 13 | Prediction 14 Vol/Ha % j Vol/Ha % Vol/Ha % | Vol/Ha % 3 m m m 0 10 66 297 242 358 81 15 45 88 59 46 47 30 23 367 74% 75% 401 82% 366 640 130% 0 9 8 126 87 76 0 31 23 17 15 38 10 13 5 198 106% 151 177 81% 95% 130 70% 552 565 83% 81% 543 80% 770 113% 92% 82% 118%  I  3  3  As one can see, these last predictions are different from the actual values, but this is mainly due to the fact that all of them are based on a much smaller sample size. For both blocks, another factor that may have affected predictions is the inaccuracy in measuring the total area of the block.  93  7.3  Results for Harvesting System Design  N o f o r m a l v a l i d a t i o n w a s p e r f o r m e d for t h e h a r v e s t i n g d e s i g n s e c t i o n ; h o w e v e r , this part of t h e m o d e l u n d e r g o n e d e t a i l e d v e r i f i c a t i o n a n d t e s t i n g . A d e t a i l e d d e s c r i p t i o n of t h i s m o d u l e is g i v e n in t h e u s e r m a n u a l . S i n c e this part d e p e n d s a l m o s t entirely o n u s e r ' s e x p e r i e n c e , a n a n a l y s i s will not b e o b j e c t i v e . T h e p r o g r a m will b e g i v e n to p e o p l e in i n d u s t r y to p l a n their c o s t s w h e n h a r v e s t i n g a b l o c k . T h e r e a l c o s t of h a r v e s t i n g t h e b l o c k will b e c o m p a r e d w i t h w h a t t h e y predicted using the model.  P e o p l e with m o r e e x p e r i e n c e in this field, will b e a s k e d to t a k e part in t h e a n a l y s i s , a n d m o r e information m a y b e i n c o r p o r a t e d into t h e m o d e l , d e p e n d i n g o n f i n d i n g s .  94  8. MODEL SUMMARY AND DISCUSSION The objective of this thesis was to develop a decision support tool that will predict economics of second-growth harvests. Primary input was chosen to be company sort descriptions and cruise data. Based on these, the model first computes volume by sort for each species and total volume and value for the entire block. Second, it allows forest planners to configure various harvesting systems and select the best one. Built-in productivities and costs, or input values can be used to compute the total cost for logging the block, and obtain an estimate of time necessary for each one of phases in the harvesting process. Total value of the block and the cost of logging can be then used to compute the net revenue.  An alternative (more detailed) cruise method was tested and it provided better results than the standard one. Various factors that may affect the accuracy of prediction were identified, analyzed and methods to account for them were developed and incorporated into the model. The tree characteristics included in the cruise were sweep, fork or crook and knot size. For the first two items their exact location was recorded. For knots, the distribution on the stem was estimated. All these factors proved to play a role in prediction of volume by sort and total value produced by a stand. The most important factor was knot distribution.  As part of this project, a dynamic programming algorithm that will theoretically buck each tree to its maximum value, was developed. Also, a breakage model was incorporated into the program for manual and mechanical felling. The amount of wood that is lost during harvesting and its impact on stand value was evaluated.  The model was tested at two sites. Overall, it can be considered that it performed well: an accurate distribution of volume by species and by sort was obtained. The predicted volume was within 5% and predicted value was within 3% compared to the actual numbers. The best prediction of volume by sort was achieved for those species that are the most dominant in the  95  stand. Considering the large variability existing in the B.C. forest industry, in terms of sort descriptions for each species, future validation will be very beneficial.  It is concluded that from the technical point of view, all the objectives set forth at the beginning of the study were fulfilled.  The model has the standard appearance of Microsoft products. For common tasks (new file, open file, print etc.) standard buttons, designed and used by Microsoft, were used. For jobs specific to this project, buttons have attached images suggesting the task they initiate. All buttons that appear on the start-up form, display tag tips (labels that explain what the button does) and in other instances the model displays different help messages.  Being a joint UBC - FERIC project, the model will be presented in both a Master of Forestry thesis and a FERIC report. The report plus the user manual will be submitted to Canadian Forest Service - Pacific Forestry Center, that provided partial funding for this project. It will summarize the findings of the model validation process.  During the elaboration of this project, people from industry had been contacted and asked their opinion about the model. This had a major impact on the model, especially on the part that addressed partial cuts. As a consequence, at the completion of the program, a detailed presentation will be made to the same people and the model will be discussed.  Depending on the input from the industry and other factors (especially funding) the items described in the next section of the paper will be considered for future development.  96  9. FUTURE DEVELOPMENT OF THE MODEL The features that can be added, in further development of the model are: •  related to forestry specific problems.  •  related to programming in general and to Visual Basic in particular.  The first class of improvements addresses primarily mensuration problems that aim at improving the methods used to predict the volume of second-growth stands using the B.C. Ministry of Forests cruise method: •  Include more details when predicting the best combination of logs for each tree. The current version of the model uses only dimensions (diameters and length), sweep, knot diameters and spacing. Other details collected when cruising (conk, blind conk, scar, frost crack, mistletoe, rotten branch, spiral grain) are used only as criteria for partial cutting. Although these defects are not very frequently found in secondgrowth, a method to assimilate them in the optimum bucking algorithm should be developed. This will increase the applicability of the model.  •  Assign a B.C. Ministry of Forests grade to each log produced in the optimum bucking process and calculate a value for stumpage.  •  Include other species in the model, especially hardwoods. A solution would be to include Alder and assimilate all hardwoods to this grouping. This should be done to improve accuracy of volume and value and for a better analysis of the cost of harvesting process.  •  Offer a separate menu (plus icon) that will analyze the block as if no breakage occurs. This is the theoretical maximum value that should be obtained from the block and would serve as a target for the logging crew.  •  Include a T table in the model. Currently, the model calculates the volume for each plot individually. These values can be displayed so experienced users can check if  97  the sampling error for the cruise is within the legal limit, or more plots have to be cruised. •  Include a knot distribution function. As a part of this project it was found that considering knots accounted for about 4% of the entire value of the stand. It may be of interest to try to develop an equation that will describe (per species) the diameter of knots on the stem as a function of DBH, total height and current height. By including this, it is expected that the model will perform as well as it does now, using standard cruise data.  •  Make the model capable of accepting cruise data that are input for B. C. Ministry of Forests stumpage and appraisal purposes, thus avoiding inputting it twice.  •  Attach a new field to cruise data that will record whether or not a tree is appropriate for peeling, and consider it in the optimum bucking algorithm. This was done as part of another FERIC project, because peelers were in fact the most important sorts in the operation. No final result is available so far, but is estimated that with minor effort, a significant improvement of prediction can be achieved in this type of situation. The modification was not available by the time the Okeover block was cruised, so it was not used in the analysis.  Further work on a more elaborate bucking and breakage model may not be very rewarding, as important influences include logging personnel and variability of ground conditions within the stand, neither of which may be predictable.  Apart from these improvements, which are all quantitative, the following qualitative improvements should be considered. All of them are driven by the extensivity of Visual Basic, i.e. by the modules that can be attached to this programming language. Once the final program is written, the setup kit will include the components it needs from these external modules and can be distributed royalty-free. So far, three products of this type have been used and the results were very good. They worked very well in conjunction with Visual Basic and dramatically  98  improved the quality of the overall performance of the program and of its user interface. Among these add-on tools advertised in the latest "Component Objects and Companion Products for VB" catalog, the following can be used to improve the model both quantitatively and qualitatively: •  Sylvan Maps - a powerful Geographic Mapping and Analysis custom control, which can be used to add sophisticated mapping functionality to VB programs.  •  M.4 - a software package for adding intelligence to VB applications using a expert system language with English-like syntax, which makes knowledge system development easy. Previous research in this field (Linehan and Corcoran, 1994) proved the feasibility and utility of applying this type of tool in forestry.  •  NeuroWindows - a Dynamic Link Library for VB that enables a programmer to solve using neural networks, applications that have previously defied solution with conventional programming. This may be used as a substitute for the dynamic programming procedure should the introduction of more detail make processing speed prohibitive.  All the above mentioned applications can be used to attach to the model information related to social, legal or environmental problems related to harvesting activities.  Apart from these ones, there are lots of VB add-ons that can assist the programmer in writing a professional "help" for the application and improve its graphic capabilities, making it more appealing to the user.  99  REFERENCES Andersson, B.O., Young, G.G., Pavel, M. - Estimating Harvesting System Performance on Second-Growth Coastal Sites. FERIC, Vancouver. Report in progress. Andersson, B.O., Jukes, W. 1995 - Harvesting Coastal Second-Growth Forests: Two Case Studies. FERIC, Vancouver. Technical Note TN-232.14 pp. Andersson, B.O., Warren, F. 1996 - Harvesting Coastal Second-Growth Forests: Hand Falling and Skyline Yarding. FERIC, Vancouver. Technical Note TN-239.14 pp. Apex Software Corporation. 1994 - TrueGrid Pro. Pittsburg PA. 241 pp. Araki, D. 1993 - Evaluation of a Denharco DH 550 Feller-Processor. FERIC Technical Note TN203. 7 pp. Avery, T.E., Burkhart, H.E. 1994 - Forest Mensurations. Fourth edition. McGraw - Hill Inc., 408 pp. Baumgras, J.E., Hassler, C.C., LeDoux, C.B. 1993 - Estimating and Validating Harvesting System Production Through Computer Simulation. Forest Products Journal. Vol.43, No. 11/12. pp. 65-71. Bennett, D.M. 1993 - Partial Cutting in a Second-Growth Douglas Fir Stand in Coastal British Columbia: Productivity, Costs and Soil Impacts. FERIC Technical Note TN-199.12 pp. Briggs, D.G. 1989 - Tree value system; description and assumptions. Gen.Tech.Rep. PNWGTR-239. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 24 pp. British Columbia, Ministry of Forests. 1986 - Forest Service Scaling Manual. Victoria. 400+ pp. (updated periodically). British Columbia, Ministry of Forests. 1983 - Manual of Cruising Procedures and Cruise Compilation. Victoria. 400+ pp. (updated periodically). Copithorne, R.D., Young, G.G. 1994 - Log Value Losses in Second-Growth Harvesting Operations in Coastal British Columbia. U.B.C. 14 pp. Cossens, P., Murphy, G. 1988 - Human Variation in Optimal Log Making: A Pilot Study. Proceedings of International Mountain Logging and Pacific Northwest Skyline Symposium. Portland, Oregon, pp. 76-81. Daigle, P.W. 1995 - Partcuts: A Computerized Annotated Bibliography of Partial-Cutting Methods in the Pacific Northwest (1995 update). B.C. Ministry of Forests, Research Branch. 25 pp. Deadman, M.W., Goulding, C.J. 1978 - A Method for Assessment of Recoverable Volume by Log Types. New Zealand Journal of Forest Science. Vol.9, No.2. pp. 225-239. Dykstra, D.P. 1984 - Mathematical Programming for Natural Resource Management. McGraw Hill, New York. 318 pp. Eng, G., Daellenbach, H.G., Whyte, A.G.D. 1986 - Bucking Tree-Length Optimally. Can J . For. Res. Vol.16, No.5. pp. 1030-1035.  100  Faaland, B., Briggs, D. 1984 - Log Bucking and Lumber Production Using Dynamic Programming. Management Science, Vol.30, No.2. pp. 245-257. Geerts, J.M.P., Twaddle, A.A. 1984 - A Method to Assess Log Value Loss Caused by CrossCutting Practice on the Skidsite. New Zealand Journal of Forestry. Vol.22, No.2. pp.173184. Gibson, D.F. 1976 - An Interactive Planning and Evaluation System for Cable Logging. American Society of Agricultural Engineers, Paper No. 76 - 1562. 26 pp. Gingras, J.-F. 1991 - Fibre Recovery Efficiency of Wood Harvesting Systems. FERIC Technical Note TN-186. 12 pp. Goulet, D.V., Hines, G.S., Sirois, D.L. 1980 - Critique Requkested of: Timber Harvesting Computer Simulation Model Design Criteria. Forest Prod. J . 30 (11): pp. 22-23. Green, J . 1994 - Stem City. Timber Processing. Vol.19, No.3, pp. 17-23. Hardison, L.D. 1995 - Bucking Takes On High-Tech Perspective. Southern Lumberman. May 1995, No.256, Vol.3, pp. 31-37. Hedin, I.B. 1994 - Shelterwood Harvesting in Coastal Second-Growth Douglas Fir. FERIC Technical Note TN-216.10 pp. Hillier, F.D., Lieberman, G.J. 1967 - Introduction to Operations Research. Holden-Day Inc., pp. 248-280. Howard, A.F. 1987 - Modeling the Cost and Profitability of Timber Harvesting with Cable Skidders. Northern Journal of Applied Forestry, No.4 (1987). pp. 87-92. Howard, A.F., Dodic, D. 1989 - Highlead Yarding Productivity and Costs in Coastal British Columbia: Predicted vs. Actual. WJAF 4(3), pp. 98-101. Hunt, J.A. 1995 - Commercial Thinning a Coastal Second-Growth Forest With a Timberjack Cutto-Length System. FERIC Technical Note TN-235.14 pp. Jukes, W. D. 1995 - Estimating Harvesting Productivity and Cost on Second-Growth Coastal Sites in British Columbia. U.B.C. M.Sc. Thesis. 102 pp. Kellogg, L.D., Bettinger, P. 1994 - Thinning Productivity and Cost For a Mechanized Cut-toLength System in the Northwest Pacific Coast Region on the USA. J . For. Eng., Vol.5, No.22. pp. 43-54. Kelogg, L.D., Olsen, E.D., Hargrave, M.A. 1986 - Skyline Thinning a Western Hemlock - Sitka Spruce Stand: Harvesting Costs and Stand Damage. Oregon State U., For. Res. Lab., Res. Bull.53, 21 pp. Koger, J . 1992 - Analyzing Timber Harvesting Systems Using STALS-3. Forest Prod. Journal. Vol. 42 (4). pp. 25-30. Kozak, A. 1988 - A Variable-Exponent Taper Equation. Canadian Journal of Forest Research. Volume 18, pp. 1363-1368. Kozak, A., Munro, D.D., Smith, J.G.H. 1969 - Taper Functions and their Application in Forest Inventory. Forestry Chronicle. Vol. 45, No.4. pp. 278-283.  101  LeDoux, C.B., Starnes, L.W. 1986 - Cable Logging Production Rate Equations for Thinning Young-Growth Douglas Fir. For. Prod. J . Vol.3, No.5. pp. 21-24. LeMay, V.M., Marshall, P.L. 1991 - Forest Mensuration - Forestry 238 Course Manual. The University of British Columbia, Vancouver, Canada, pp. 214. Linehan, P.E. and Corcoran, T.J. 1994 - An Expert System for Timber Harvesting Decision Making on Industrial Forest Lands. FPJ, Vol.44, No.6, pp. 65-70. MacDonald, A.J. 1987 - Productivity and Profitability of the Madill 122 When Grapple Yarding B.C. Coastal Second-Growth Timber. FERIC Special Report SR-48. MacDonald, A.J. 1988 - Productivity and Profitability of Grapple Yarding Bunched B.C. Coastal Second-Growth Timber. FERIC Special Report SR-54. Maness, T.C. and Adams, D.M. 1991 - The Combined Optimization of Log Bucking and Sawing Strategies. Wood and Fiber Science. Vol. 23, No.2. pp. 296-314. Mellgren, P.G. 1990 - Predicting the Performance of Harvesting Systems in Different Operating Conditions. FERIC, Special Report SR - 67. 22 pp. Mendoza, G.A., Bare, B.B. 1986 - A Two-Stage Decision Model for Log Bucking and Allocation. Forest Products Journal. Vol. 36, No.10. pp. 70-74. Meng, C.H., Williamson, J.D.A., Cottell, P.L. 1981 - Analytical Calculations vs. Simulation for Predicting the Performance of Prototype Logging Machines. The Forestry Chronicle. Vol.(12). pp. 279-283. Microsoft Corporation. 1993 - Microsoft Visual Basic. Version 3.0. Programmer's Guide. 713 pp. Microsoft Corporation. 1993 - Microsoft Visual Basic. Version 3.0. Professional Features, Book 1 - Custom Control Reference. 339 pp. Microsoft Corporation. 1993 - Microsoft Visual Basic. Version 3.0. Professional Features, Book 2 - Data Access Guide. 170 pp. Microsoft Corporation. 1994 - Microsoft Access. Version 2.0. User's Guide, pp. 1-314. Moshenko, D.W. 1991 - Grapple Yarding in the Interior of British Columbia. FERIC Technical NoteTN-176.10 pp. Murphy, G.E. 1983 - Impact of Harvesting on Value Recovery. Logging Industry Research Association - Proceedings of a Seminar Held in Rotorua. Murphy, G.E., Twaddle, A. 1986 - Technique For The Assessment And Control of Log Value Recovery In The New Zealand Forest Harvesting Industry. Proceedings of C O F E Meeting. Mobile, Alabama. Murphy, G.E., Twaddle, A., Cossens, P. 1991 - How to Improve Value Recovery From Plantation Forests: Research and Practical Experience in New Zealand. Proceedings, C O F E Meeting, Nanaimo, British Columbia.  102  Murphy, G.E., Lane, D.C., Cossens, G.P. 1996 - Progress Report on The Development of An Integrated Value Management System. Proceedings of the joint meeting COFE and IUFRO. Marquette, Ml. pp. 124 -129. Newnham, R.M. 1971. CANLOG - The New Canadian Forestry Service Harvesting Machine Simulator. Pulp Paper Mag. of Canada. Vol.72(3): pp. 107-112. Nieuwenhuis, M., Malone, L. 1996 - Value Maximization of Forest Stands Through Optimal Inventory and Cross-Cutting Methodologies. Proceedings of the joint meeting COFE and IUFRO. Marquette, Ml. pp. 117-123. Olsen, E.D., Pilkerton, S., Garland, J.J., Sessions, J . 1990 - Extending Strategies for Optimal Bucking to Harvesting and Site Preparation. Western Journal of Applied Forestry. Vol. 5, No.l.pp. 12-15 Olsen, E.D., Pilkerton, S.J., Garland, J.J. 1991 - Evaluating Timber Sale Bids Using Optimal Bucking Technology. Applied Engineering in Agriculture. Vol.7,No.1. pp. 131-136. Pavel, M., Young, G.G., Andersson, B.0.1997 - Decision Support Model: Predicting Net Revenue of Harvesting Coastal Second-Growth Forests. User Manual. FERIC, 30 pp. Parker, R., Park, R., Clement, B., Gibbons, W. 1995 - Effect of Number of Log Grades on Log Making Errors. Logging Industry Research Organization, Vol.20, No.23. Peterson, J.T. 1986 - Comparison of Three Harvesting Systems in a Coastal British Columbia Second-Growth Stand. FERIC Technical Report TR-73. 50 pp. Peterson, J.T. 1987a - Effect of Falling Techniques on Grapple Yarding Second-Growth Timber. FERIC Technical Note TN-107. 8 pp. Peterson, J.T. 1987b - Harvesting Economics: Grapple Yarding Second-Growth Timber. FERIC Technical Report TR-75.12 pp. Peterson, J.T. 1988 - Harvesting Economics: Two Case Studies of a Cypress 7280B Swing Yarder. FERIC Technical Note TN-115. 20 pp. Pickens, J.P., Lee, A., Lyon, G.W. 1992 - Optimal Bucking of Northern Hardwood. Northern Journal of Applied Forestry. Vol.9, No.4. pp. 149-152. Pickens, J.P. 1996 - HW-BUCK A Computerized Optimal Bucking Decision Simulator for Training Hardwood Log Buckersto Improve Value Recovery. Presented at the joint meeting of C O F E and IUFRO Subject Group S3.04-00. Marquette, Michigan. Pnevmaticos, S.M. and Mann, S.H. 1973 - Dynamic Programming in Tree Bucking. Forest Products Journal, 22 (1972). pp. 26-30. Putnam, N.E., Kelogg, L.D., Olsen, E.D. 1984 - Production Rates and Costs of Whole-Tree, Tree-Length and Log-Length Skyline Thinning. For. Prod. J., Vol. 34, No.6. pp. 65-69. Raghavendra, B.G. 1994 - Application of Operations Research in the Efficient Cross-Cutting of Logs. Indian Forester. Vol.120, No.12. pp. 1047-1056. Randhawa, S.U., Scott, T.M., Olsen, E.D. 1992 - Timber Harvester: A Microcomputer - Based System for Automatic Selection of Timber Harvesting Equipment.  103  Reisinger, T.W., Greene, W.D., McNeel, J.F. 1988 - Microcomputer-Based Software for Analyzing Harvesting Systems. South. J . of Apl. For., pp. 37-40. Rogers, R.E., MacDonald, A.J. 1989 - Ground Skidding Second-Growth Timber in Coastal British Columbia: a Case Study. FERIC Special Report SR-60.19 pp. Rutherford, D.A. 1996 - Productivity, Costs and Optimal Spacing of Skyline Corridors of Two Cable Yarding Systems in Partial Cutting of Second-Growth Forests of Coastal British Columbia. M Sc. Thesis, UBC, Faculty of Forestry. 87 pp. Sauder, E.A. 1988 - Future Logging Equipment Needs in Coastal B.C. (1989-2005). FERIC, Special Report SR-49. 44 pp. Sessions, J . 1988 - Making Better Tree-Bucking Decisions in the Woods; An Introduction to Optimal Bucking. Journal of Forestry. Vol.86, No.10, pp. 43-45. Sessions, J., Olsen, E., Garland, J . 1989 - Tree Bucking for Optimal Stand Value with Log Allocation Constraints. Forest Science. Vol.35, No.1. pp. 271-276. Sondell, J . 1987 - Computerized Bucking Techniques and Systems. Resultat. No.6. (translated by Raymond W. Lipton). Stuart, W.B. 1981 - Harvesting Analysis Technique: A Computer Simulation System for Timber Harvesting. Forest Products J. 31 (11): pp. 45-53. VideoSoft. 1995 - VSVBX 5.0. Berkeley, CA. 127 pp. Warren, F. 1994 - Forestry 430 - Advanced Biometrics. Term Project. University of British Columbia, Faculty of Forestry. 5 pp. Warren, F. - Estimating Yield of Second-Growth Softwood Stands - A Case Study. U.B.C, M.F. Thesis (in progress). Walpole, R.E. 1982 - Introduction to Statistics. McMillan Publishing Co. Inc. 521 pp. Wang, S.J., Giles, D.R. 1989 - Effects of Various Factors on Computer-Optimized Bucking System Performance. Forest Products Journal. Vol. 39, No.11/12. pp. 33-36. Webster, D.B., Padgett, M.L., Sirois, D.L. 1983 - Features of a felling module using a fellerbuncher in a timber harvesting computer simulation model. Forest Prod. J . 33(6): pp. 1116. Yaussy, D.A., Brisbin, R.L. 1988 - Predicting Individual Log Dimensions and Grade from Hardwood Cruise Data. In The 1988 Symposium on System Analysis in Forest Resources. USDA Forest Service, General Technical Report RM-161. Rocky Mountain Forest and Range Experiment Station. Fort Collins, Colorado, pp. 75-80. Young, G.G., Waddell, D.A. 1986 - A Model for the Interactive Optimization of Log Bucking Utilizing Dynamic Programming. U.B.C, Department of Supplies and Services. Research Report. 84 pp. Young, G.G. - Cut-to-length versus Tree-length Harvesting of Coastal Second-growth. FERIC Report in progress.  104  Appendices:  1. Presentation of the program using pseudocode: 1 .A Computation of stand volume and value. 1 .B Harvest System Design and computation of net revenue. 2. Productivity equations and average productivities used Harvest System Design. 2.A Clear-cut. 2.B Partial cut.  105  Appendix 1 A. Presentation of the program using pseudocode: computation of stand volume and value. There is no pseudocode description given on how to input bucking specifications and cruise data, because this is a straight forward procedure; more detailes are given in the user manual. The first procedure that is used to process the data is the following: SUBROUTINE Datalnput  Input general data about the block: Area, Silvicultural System, StageLength (used in DynamicProgramming), Falling Method, BAF, Min.Diam, Pulp Value, Min.Len.Pulp and if necessary Area of Corridors I Tf Silvicultural System is clear cut Then Call CalcVolAndValue Exit Sub Else  Display graph showing the actual structure of the stand (Stand Table) Do  r~Do Display form CriteriaForPartialCut, where the user can input his criteria for each species Validate user input for partial cut (see user manual for details) Loop Until a set of valid criteria is selected. Display graph showing the structure of the stand according to criteria set by the user Loop Until an acceptable stand structure is obtained |__End If Call CalcVolAndValue END OF SUBROUTINE _TJ  106  SUBROUTINE CalcVolAndValue Call LoadBuckPrefs - this procedure will load in the memory, using dynamic arrays, all sorts grouped by species and arranged in ascending order with respect to length. Initialize data about breakage - read data from corresponding table and store it in memory Initialize the array that will store volume and value per plot/species/sort Initialize the array that will store volume and no. of stems per plot Initialize values (TotalVolume, TotalValue and other numbers used later in the model) If Silvicultural System is clear cut Then Assign Area = Total Area Block Set stringClearCut that selects all trees from Cruise Data table, ordered by plot and tree number Call CruiseData(stringClearCut) Else ' do the analysis for corridors as if it is a clear cut Assign Area = Area of Corridors Set stringClearCut that selects all trees from Cruise Data table, ordered by plot and tree number Call CruiseData(stringClearCut) ' do the analysis for the remaining area as a partial cut Assign Area = Area Block - Area Corridors Initialize arrays and values specific to Partial Cut Set stringPartialCut that selects trees from Cruise Data table, according to criteria established by the user, ordered by plot and tree number Call CruiseData(stringPartialCut) _ E n d If Compute AverageDBH, StandDensity Compute TotVolumePerHa, TotValuePerHa, VolPerSpeciesAndSortPerHa, ValuePerSpeciesAndSortPerHa, volume and value per total block, per species, per sort for the entire block Compute AverageTreeVolume and AveragePieceVolume If Silvicultural System is partial cut Then Compute all the above numbers corresponding to Partial Cut _ E n d If Save all the numbers computed above in the corresponding database Display results for the block analyzed END OF SUBROUTINE  107  SUBROUTINE CruiseData(stringSql) Set DynaSet (set of records from Cruise Data) according to string passed to this procedure [""Do While Not End Of DynaSet Read fields from current record (Plot No., Tree No., Species, DBH, TotalHeight etc.) Assign new Plot No. to account for non sequential plots (if this is the case) Compute BasalArea, TreeFactor and other number that will be used in the program Compute MerchH (MerchantableHeight) for the tree analyzed Compute NoStages (used in the Dynamic Programming procedure) NoStages = TotalHeight / StageLength; correct NoStages with stump height Call MakeStemProfile Assign innitial values for LowerLim and UpperLim, i.e. the first and last stage used in DynamicProgramming; LowerLim = 0, UpperLim = NoStages Call ChecklfBreakage - this procedure will check if the current tree will break or not. If yes, then FlagBreakage will be set true and data will be available about where breakage occurs and the length of shatter If FlagBreakage = True Then Compute BreakH (height where breakage occurs) r If MerchH > BreakH Then ' i.e. breakage occurs below merchantable height LowerLim = (MerchH - BreakH) / StageLength UpperLim = NoStages Else ' breakage is above merchantable height Set FlagBreakage = False |_ End If | End If Do ' if breakage is below merch. height analyze first the butt log Call DynamicProgramming(LowerLim, UpperLim) Call BackPass(LowerLim, UpperLim) ' check if there is a second log to be analyzed [~Tf FlagBreakage = True Then r If MerchH > BreakH + LengthOfPulp Then ' if the top log is long enough, analyze it LowerLim = 0 UpperLim = (MerchH - BreakH) / StageLength Else ' there is breakage but the top log is too short FlagBreakage = False | _ E n d If End If Assign FlagBreakage = Not(FlagBreakage) L Loop Until FlagBreakage = True Move to next record in DynaSet L- Loop END OF SUBROUTINE  _•  108  SUBROUTINE  MakeStemProfile  For i = 0 To NoStages Compute diameter of each stage using Kozak's taper equation, based on Species, DBH, Current Height, Total Height Compute average taper based on Minimum Top Diameter, Diameter at inflection point, and (Total Height - Height of inflection point). Compute diameter for the last 2 m. (close to the stump) based on average taper calculated above Compute each diamter using BCMoF scaling manual conventions, i.e. round off diamteres to the nearest even number Next  ' the following subroutines work in a similar way; each one reads information about the ' current tree in its record and assigns it to corresponding stages Call AssignSweep ' the next procedure will read if Fork or Crook is present on the stem and put it in the ' corresponding stage. Since none of the sorts will accept this deffect, the procedure will ' also set FlagBreakage=True that will impose that the tree be bucked at the stage where ' the deffect is Call AssignForkOrCrook Call AssignDiameterKnots Call AssignKnotsSpacing  END OF SUBROUTINE  109  SUBROUTINE DynamicProgramming(FirstStage,LastStage) Compute N, which is the length of shortest sort for a species, converted to stage intervals N = MinimumSortLength / StageLength For i = FirstStage + 1 To FirstStage + N Compute Volume for each stage using Smalian's formula Assign Value = Pulp Value for each stage Assign SortName = Pulp for each stage ' Note: Logs resulted after applying this algorythm and that are shorter than Minimum ' Length of Pulp, will have a value and volume equal to 0 assigned in procedure ' BackPass _Next  ' let i be the current stage For i = FirstStage + N + 1 To LastStage Assign StageValue = 0 Assign CurrentSort = 1 ' N.B. Sorts are already arranged in increasing order of length Assign BuckLength-length of CurrentSort converted to Stagelntervals Do While (BuckLength <= i - FirstStage) And (CurrentSort <= No.Sorts For This Sp.) Call VerifyConditions (i, BuckLength) ' VerifyConditions will return ValueSort - the value of the sort that can ' be obtained at this stage ' TemporaryVal will store the value of current log + the value of the rest of the tree TemporaryVal = ValueSort + Value of Stage (i - BuckLength) If TemporaryVal > StageValue Then ' Save the best decision so far ' This is the highest value of the tree up to stage i StageValue = TemporaryVal ' These are the characteristics of log that has to be bucked at this stage ' to get the above value OptimumLength = BuckLength OptimumSort- Current Sort OptimumVolumeStage=VolumeSort OptimumValueStage- ValueSort End If  Calculate BuckLength as the length in stages of next sort Loop  ' When exit above loop have optimal decision for stage i Save solution for backward pass Next  END OF SUBROUTINE  110  To make this procedure easy to follow, conditions are presented as if are tested together. In the program, conditions are tested separately and when one of them fails, the program skips a portion of code; this ensures optimum functionality for this procedure.  SUBROUTINE VerifyConditions (CurrentStage, BuckLength) Declare boolean variable Condition Initialize VolumeSort and ValueSort to zero Assign Small End Diameter Sed = Diameter At Stage (CurrentStage - BuckLength) Assign Large End Diameter Led = Diameter At CurrentStage LCheck conditions about sort dimensions If ( Led >= MinButtDiam of Current Sort Or MinButtDiam IsNull) And ( Led <= MaxButtDiam of Current Sort Or MaxButtDiam IsNull) And (Sed >= MinTopDiam of Current Sort Or MinTopDiam IsNull) And (Sed <= MaxTopDiam of Current Sort Or MaxTopDiam IsNull) Then Condition = True Else  Condition = False GoTo label _ E n d If  ' Check for allowable defects For i = CurrentStage - BuckLength To CurrentStage If (KnotsDiam at stage i <= KnotsDiam accepted for Current Sort) And ' KnotsSpacing and Sweep are boolean values (KnotsSpacing at stage i >= KnotsSpacing accepted for Current Sort) And (Sweep at stage i <= Sweep accepted for Current Sort) Then Condition = True Else  Condition = False End If  ' The Fork or Crook was already considered in AssignForkOrCrook Next  label: Calculate VolumeSort with Smalian's formula (given Led, Sed and BuckLength) If Condition = True Then  ValueSort = VolumeSort * Unit Value of Current Sort Else  ValueSort = VolumeSort * Unit Value of Pulp End If  END OF SUBROUTINE  111  (FirstStage,LastStage) ' This procedure picks-up the combination of logs that maximizes the value of the tree  SUBROUTINE BackPass  Assign CurrentStage = LastStage Do  ' take sort name and length produced by DynamicProgramming CurrentSort = Sort Obtained at CurrentStage BuckLength = BuckLength at CurrentStage If  ( CurrentSort = Pulp) And (BuckLength < sngLengthPulp) Then ' the sort produced is Pulp but doesn't satisfy the minimum length condition CurrentVolume = 0 CurrentValue = 0  Else  ' take values produced by DynamicProgramming CurrentVolume = VolumeSort CurrentValue = ValueSort  I End If Accumulate the CurrentVolume, CurrentValue into the corresponding plot/species/sort Increment the StandDensity in the array that stores it (only for the first log  produced from a tree) Assign CurrentStage = CurrentStage - BuckLength Loop While  CurrentStage > FirstStage  END O F SUBROUTINE  112  Appendix 1B. Presentation of the program using pseudocode: Harvest System Design and computation of net revenue. The following procedures are attached to the Harvest System Design form. SUBROUTINE  HarvSy st Design  Select a system for the first stratum, according to indication displayed under the toolbar, i.e.: click on the item that you want to include in the system (only one item can be included in each phase). To deselect all items in a phase, double click on any of them. By default the first stratum is available and its percentage is 100%. If the block will be stratified in more than one stratum, then the percentage for the first one must be reduced. This principle holds also when the block includes more than one stratum and the user wants to include a new one, i.e. the sum of percentages for existing strata must be less than 100%. Select Case the Button clicked by user Case NewStratum Call NewStratum Click  Case StratumNumber (one of the buttons in the Strata frame) Call StratumNumber_Click  Case DeleteStratum Call DeleteStratum_Click  Case Phases Call Phases_Click  Case Results Call Results_Click  Case Cancel/End Unload the form. End Select END OF SUBROUTINE  _•  113  FUNCTION CheckStratum  Calculate MaxPercentage = 100% - Sum of percentages for all other existing strata. If Percentage assigned to current stratum > MaxPercentage Then Display message about the range allowed for Percentage CheckStratum = False Exit Sub End If  Update record for the last stratum analyzed If system of current stratum <> system of current stratum that was assigned previously Then  Update the system for current stratum Update machine names to "No Machine" Update the cost of harvesting the stratum to 0. End If  If PercentageOfTotalVol of current stratum <> PercentageOfTotalVol of current stratum that was assigned previously Then Update the Volume of Current Stratum Update machine names to "No Machine" Update the cost of harvesting the stratum to 0. End If CheckStratum = True END OF FUNCTION  114  SUBROUTINE NewStratum_Click If CheckStratum() = False Then Exit Sub  If Not all existing strata have a system assigned Then Display message showing the user the situation that occured Exit Sub End If  ' each block can include maximum 5 strata If number of existing strata = 5 Then Display message showing the user the situation that occured Exit Sub End If  ' all conditions are met, a new stratum can be added Add a new stratum, i.e. find the first available stratum and enable it. Make all its labels grey and make its percentage = 100% - sum of percentages for all other existing strata. Refresh foregrounds for strata buttons: make the active one red, and all the others grey. END OF SUBROUTINE  SUBROUTINE StratumNumber_Click If CheckStratum() = False Then Exit Sub  Activate the stratum corresponding to the button that was clicked: change the colour of labels (in green) according to the system previously set, and display its percentage. Refresh foregrounds for strata buttons: make the active one red, and all the others grey. END OF SUBROUTINE  115  SUBROUTINE DeleteStratum_Click  Display message asking the user to confirm that he wants to delete that stratum If the answer is No Then Exit Sub  Set to 0 the value associated to each label in the harvesting system (i.e. not selected). Delete the name of all machines for that system. Set to 0 the cost for each phase and for harvesting the entire stratum. Disable the button corresponding to that stratum. Activate the first stratum available in the block; if none is available, activate stratum number 1. END OF SUBROUTINE  SUBROUTINE Phases_Click If CheckStratum() = False Then Exit Sub  If none of existing strata has a system assigned Then Display message showing the user the situation that occured Exit Sub End If Call FormPhases_Show END OF SUBROUTINE  116  SUBROUTINE Results_Click If CheckStratum() = False Then Exit Sub If n o n e of e x i s t i n g strata h a s a s y s t e m a s s i g n e d Then Display m e s s a g e s h o w i n g t h e u s e r t h e situation that o c c u r e d Exit Sub End If Call FormResults_Show END OF SUBROUTINE  SUBROUTINE FormResults_Show Show f o r m R e s u l t s . Display p e r c e n t a g e of b l o c k a n a l y z e d s u m m i n g u p p e r c e t a g e s for a l l s t r a t a . Display i n f o r m a t i o n a b o u t e a c h s t r a t u m (in t h e grid). Calculate a n d Display t h e total c o s t of h a r v e s t i n g a n d net r e v e n u e g e n e r a t e d b y t h e entire b l o c k . END OF SUBROUTINE  S U B R O U T I N E FormPhases_Show Show f o r m E q u i p m e n t P r o d u c t i v i t y a n d C o s t . Enable o n l y b u t t o n s that c o r r e s p o n d to strata that h a v e a s y s t e m a s s i g n e d . Activate t h e first s t r a t u m that h a s a s y s t e m a s s i g n e d . Display t h e v o l u m e of t h e a c t i v e s t r a t u m . Configure t a b s o f t h e a c t i v e s t r a t u m a c c o r d i n g t o t h e h a r v e s t i n g s y s t e m c h o s e n ( e a c h t a b c o r r e s p o n d s t o a p h a s e in t h e s y s t e m ) . Activate t h e first t a b ( p h a s e ) in t h e s y s t e m ; d i s p l a y in t h e grid d a t a r e l a t e d t o t h e first phase. END O F SUBROUTINE  _•  117  The following procedures are attached to the Equipment Productivity and Cost form. SUBROUTINE EquipProdAndCost If a new tab in the same system is selected Then Call SwitchTab End If  Select Case the Button clicked by user  Case StratumNumber (one of the buttons in the Strata frame) Call StratumEquipment_Click  Case Design Call Design_Click  Case Results If CheckMachines() = False Then Exit Sub Call FormResults_Show Case End If CheckMachines() = False Then Exit Sub  Unload the form. End Select END OF SUBROUTINE  _•  118  SUBROUTINE SwitchTab If CheckMachines() = True Then Update the new tab; refresh information in  (phase) that was selected.  the grid according to the new tab  End If  END OF SUBROUTINE  FUNCTION CheckMachines Select Case  NoOfMachinesSelected  Case 0  values in the previous tab; set machine name to "No Machine", for that phase, and make cost equal to 0.  Update  Case 1 Update  the name of machine used and cost for previous phase, according to user selection.  Case Else Display  message box to warn the user that he is not allowed to select more than one machine.  CheckMachines = False Exit Function End Select CheckMachines = True END FUNCTION  _•  119  SUBROUTINE StratumEquipment_Click If CheckMachines() = False Then Exit Sub  Activate New Stratum; change foreground to gray for the other strata and red for the active one. Display the volume of active stratum. Configure the tabs for the active stratum according to system selected. Select the first tab which corresponds to the first phase in the system. Update data in the grid as per new stratum. END OF SUBROUTINE  SUBROUTINE Design_Click If CheckMachines() = False Then Exit Sub  Unload form EquipmentProductivityAndCost Call FormDesign_Show END OF SUBROUTINE  SUBROUTINE FormDesign_Show  Show form HarvestSystemDesign Enable buttons for strata that have a system assigned Activate the first stratum available; change the colour of labels (in green) according to the system previously set, and display its percentage. Refresh foregrounds for strata buttons: make the active one red, and all the others grey. END OF SUBROUTINE  120  The following procedure is attached to the Results form. SUBROUTINE Results  Select Case the Button clicked by user Case Design Unload form Results Call FormDesign_Show  Case Phasses Unload form Results Call FormPhases_Show Case End  Unload form Results End Select END OF SUBROUTINE  121  3  s  1O >  3  1o o >  CD CD  >  CD  o >CD CD i_  H II  E  I  c5  s  co  oo o>  1o Q  3 . O  TJ C  n §«--  £8 § _ r_5 Q  <? ra  CD a .  co  o oo  o  n  TJ  5 Q O O  Q  o O Q LU Z c r LU  c o TJ  o  o  a>  < z  o  » 1 03  o  _l  co LU  z X  o <  |£ c  Q o  Q  E,  11  LU Z  I'D  _i < z  00 lO  o  o  0  ci +  CO  I CM  LU  I co Io  z X  o <  II  Ic"  1 !§.  LU  a  CD  E  w .92  i n cn cn  cn  II  I I  11  O O O +  >  CM  m ci  <  II  IS  "5"  II  _L at  E F  E  DC  >  O a .  O a . TJ  2 CM X TJ C  O  |  o < DC  3  '5.  I-  CD  CD ra O  X LU  co  Q. Q. <  CO  IO  E ra CO  o co  o <  co  ai  a . a . ra  a .  8a  IS  I  a  Ic  o  CO  X  o  CD  O  CD  DC \— X LU  > CC cc a .  3  42.  ra  I CM cn cn  CD  E  a .  s  CD  >  to  CM  T3 |£ CD  S  1 .5?  I X  CD  CD  TJ TJ  TJ TJ  2  CO  O O  JZ  O  ,  "O  co a> a .  cn c 3  o  & x:o 8 CD  co  co CM  8  CO D O" LU  8"  I  I  CO  3 sz  •o  3 sz Ol  if  o  X  S>  in CM  US  1  Cost [$/h]  lO  3  s  co cn c 0  Source  1 T J C  Range for variables  <  lO CJJ OJ  c 3  I  •* OJ OJ  00  8  co  co  OJ OJ  OJ OJ  c o  e o  c CD  X  U)  ffi  C  IO OJ OJ  in  & =j  <  <  C  <  in  E  CD T J  T J  CO  c o CL  •  in T J  8  UJ  E CD  ci  o"  E. "5  > «  CM  Productivity Equation  co + co co co  co LU  z 1  o 10  Q LU  2  p  5—  0  <o «  E. o >  1  CD CD  h« o co ll  J= T J  2  co  I CM  ci  0.  CM  CD  Machine  D_  E. T J  2 a.  in  1  T J CO  2 a.  z  X  o < LU  CD  CC  5  C\J  CD  Si  E F  I  co  ai  CD  Phase | No.  CD z _l  -1  if <  0  1 is c  g 0.  "55  "ai  LU  LU  o  O J  c "CD LU  CM  E, 2  -i  o <  T I  co  CO  0)  O  s  T J  "55 LL  Q LU  o  E. CL  O  CM  <  (D Z  _ J _I  LL  m  c  !  "55 LU  CD  E  s o  8 2  CL  E. 2  CD cc  E. 2  O  o o  o  LU  o o cc LL  H  Z LU c SZ  a  X  If in  Q  H CO  — 1 <  CU  CM  <  CD  O  o  2 z a  Q CM OJ OJ  a  0_  2  H  E TB  E,  ?  T J  o o  X  CO  II  o_  10  3  LU  l-i oo  CL  •i  CL  LL  o < 2  OJ CM CM II  Q_  c  0.  <  LU Z X  CD z co co  CD z CO co LU o o cc  CD  CM  CM CM II  T J  o  8  a.  m  10  D_  m  1  3  z>  T J  1—  >> c  CO  o  o  o CD . C  CO  LU Z X  co oo ll  %  co § a. B to  £ h-  5"  0 0 I--  LU  E. T J  CD  CO LU  o  o c  CD  II  II  o o SZ  CD CD  1  0  1  2  C O J  O J  LU 00 CM  co  O)  <3 co 5 cn o X  •2  g  cr  111  o  s c co c  S3  o  3 .c CO X  5 x:  2>  X  CM  co  E.  0.  s  <-  co  Prentice 625  {CC}  2 TJ  Hydraulic  CL II  Podain  GL  Peterson -1986  Cost [$/H]  OJ CO IO  8  8  n  0.  2  Hedin -1994  Peterson -1986  Peterson -1986  Peterson -1986  Source  CM  Front End  CL  g E,  LOADING  TJ  {CC}  E,  Hahn  Range for variables  to  Processor  o  LARGE MACHINES  II  PROCESSING  CM CM  II  ROADSIDE  CO  5B  2 o  {CC}  TJ  {CC}  MEDIUM MACHINES j=  Hahn II  Std. Power Saw  PROCESSING n  Processor  Manual  ROADSIDE  Productivity Equation  co  Delimber  5A  CL  1  Machine  cvi  Phase  o TJ  (00)  CO  No.  c CO CD CO  co  CM  II  £ £  TJ T J  0.  2  <M  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0075261/manifest

Comment

Related Items