Th·Is About . file w This F as c r e ile: ated b . Missc Y sca ans l d nning en n ' ,·l ed the P h o we by th ver s rinte e soft d PUb orne w li cat io a rnistak re ha v n es rn e b e e n co . ay re rrec t rnam. ed · I · 3 I BRINGING TIMBER QUALITY CONSIDERATIONS INTO FOREST MANAGEMENT DECISIONS: A CONCEPTUAL APPROACH Roger D. Fight, Thomas A. Snellgrove, Robert 0. Curtis, and Dean S. DeBell Until recently, development of the timber industry in the Douglas-fir region was based on the availability of large vol­ umes of "old growth." Large volumes per acre and large trees meant low logging and milling costs per unit of volume. Large, old trees grown slowly in dense stands produced a high proportion of wood with narrow and uniform growth rings and relatively few knots. The highly desirable appearance, strength, stability, and machining properties of such wood have established and maintained this region in a strong position in domestic and international markets for logs, lumber, and plywood. The timber resource is now far along in the transition from unmanaged forests to extensively managed "young grqwth." It will soon be composed of more intensively managed planta­ tions that have been under some degree of control from the time of establishment. Wood from these new stands will have characteristics very different from wood from the old-growth forest, and the way these stands are managed will affect these characteristics. Any of a wide range of possible management regimes can be imposed on these new stands; and these re­ gimes will affect the characteristics of the wood produced, its value, and its suitability for different products and markets. The relative importance of these changes can most appropri­ ately be judged by putting them in a monetary context. Studies on wood quality for Douglas-fir (Pseudotsuga men­ ziesii [Mirb.] Franco var. menziesii) have been conducted for years, but the researc has several shortcomings. The first and most fundamental is that there are no examples of the wood grown over a rotation in stands grown under a wide range of silvicultural regimes. Practices such as fertilization and wide initial spacing have been introduced too recently to provide good information on their ultimate effects, and recent and cur­ rent practices by no means represent the full range of regimes that could be applied. Second, little has been done to assess effects in monetary terms. Third, the work has not been incor­ porated into a comprehensive analysis. It is known that silvi­ cultural decisions affect timber yields and costs of manage­ ment, logging, and manufacturing, as well as wood quality and 20 product yield. It is important, therefore, to have a "holistic" analysis that looks at effects on costs and returns in timber pro­ duction, logging, and manufacturing. This paper illustrates the possible magnitude of the changes in product quality and value, describes what is needed in a hol­ istic analysis, and suggests how to deal with an uncertain fu­ ture. It concludes by citing some of the overly simplistic solu­ tions that can be avoided by using a holistic analysis. AN ILLUSTRATIVE STUDY Information now available on relationships between product value. and wood quality or tree size comes mainly from experi­ ·ence with natural stands. These relationships may be quite dif­ ferent for future managed stands, which have been under stocking control from the time of establishment. Relationships will also be different for stands grown under different manage­ ment regimes. The potential effect of different regimes on product quality and value is illustrated using results of a slash pine (Pinus elliottii Engelm.) study. Comparable studies for Douglas-fir are not readily available. This study is used to point out the potential losses in yields and quality of products that may be associated with highly accelerated growth. These results do not, however, address the overall economics of the issue, which must consider the effects of revenues, costs, and interest rates. This study compared the quality and yields of mill run lum­ ber from a conservatively managed slash pine stand (hereafter referred to as the "control") with that of lumber from a rapidly grown ("test") stand. The study was conducted at the School of Forestry, Stephen F. Austin State University (Burkhart et al. 1984). Results indicate that there may be marked and undesir­ able changes in the properties of lumber from rapidly grown slash pine compared with that from denser, conservatively thinned stands. Much of the lumber manufactured from the rapidly grown trees did not meet current standards for struc­ tural grade material. The study used samples of 25 trees from each of two planta­ tions on similar sites in eastern Texas. Pertinent information on the sample is given in Table l. The test trees had been planted at 12 by 12 foot (3. 7 by 3. 7 m) spacing (300 trees/acre or 74 1/ ha) and were subsequently thinned at age 18 to 250 trees per acre (6 18/ha). The 25 test sample trees were cut at age 20. The control trees were planted at 6 by 6 foot ( l . 8 by l .8 m) spacing ( 1,200 trees/acre or 2,965/ha). Subsequent thinnings at ages 12, 20, 25, and 35 years reduced the stand to 245 trees per acre (605/ha) at age 35. The stand was thinned again to a basal area of 85 to 90 square feet per acre ( 19.5 to 20.7 m2/ha) at age 48. The 25 control sample trees had similar dbh to the test trees from the first plantation, and were cut at age 50. The effects of these two contrasting regimes on average tree volume, percentage of volume suitable for sawlogs, and per­ centage of volume in juvenile wood are shown in Table 2. The test trees had more crooked tops and lower specific gravity than the control trees. The test trees produced logs with high taper, one or two rings per inch (0.4 to 0.8 ring/em), and a substantial amount of juvenile wood. All logs were processed into 1 and 2 inch lumber through a profiling, chipping headrig. The lumber was then kiln dried and planed following conventional practice. All lumber was graded visually according to the current standards of the South­ em Pine Inspection Bureau. Subsequently, a subsample of the boards were tested for stiffness (modulus of elasticity) and bending strength (modulus of rupture) to provide some mea­ sure of how this lumber meets performance standards. . Because the control trees were taller, they produced more sawlogs per tree. To facilitate comparisons of recovery be­ tween the two samples, only the top log and butt log from each tree were used. Recovery was compared on three bases: ( l ) proportion o f the log recovered as lumber, (2) the average quality and value of the lumber, and (3) the average value of the log (which combines the effects of items 1 and 2). Lumber recovery factor (the nominal board feet of lumber per cubic foot of log) was used to compare the proportion of the log recovered as lumber (Table 3). Less lumber per cubic foot of log was recovered fro the test sample because the trees had higher taper and more crooked logs. TABLE 4. TABLE 1. Tree measurements of slash pine in eastern Texas for test and control regimes (twenty-five trees each regime). 50=year Average Average Dbh Site Index Age Height (yrs) Regime (in.) (em) (ft) (m) Test 14.5 36.8 72 21.9 20 105 Control 15.1 38.4 105 32.0 50 105 TABLE 2. Comparison of test and control slash pine trees in eastern Texas. Juvenile Sawlog Average Volume as a Tree Volume Percentage of Wood as a Percentage of Sawlog Volume** (ft3) (m3) Tree Volume* Test 26.5 0.75 77 55 Control 47.0 1.33 87 16 Regime *Based on a 4 inch (10.2 em) top. **Calculated on the assumption that the first ten rings from pith are juvenile wood, TABLE 3, Lumber recovery of the slash pine logs. Control Test (board feet/ft3) % Reduction Butt.logs 6,3 5,0 21 Top logs 6.7 5.8 13 The quality of timber could be expressed as percentage of lumber recovery by grade, but a more useful index of quality for this study is the average value of the lumber produced, in U.S. dollars per thousand board feet. The average value is cal­ culated by multiplying market prices for lumber by grade by the amount produced in each grade. Table 4 shows the percent­ age of No. 2 and Better lumber recovered and the average value of all the lumber produced. Percentage of No. 2 and Better lumber recovered and average value of lumber produced from slash pine trees in eastern Texas. Average Value No. 2 of Lumber* and Better Logs Control Test Reduction ' (percent) Butt logs Top logs 70 64 47 28 Control Test % Reduction ($U.S, /MBF) 33 230 196 15 56 200 175 12 *Based on lumber prices from Temple-EasTex for September 24, 1982, Timber Quality Considerations 21 No. 2 and Better grade is the most common grade used in residential construction, and the proportion of No. 2 and Better recovered is therefore a good index of how well the resource quality meets the current market demands for housing. The re­ duction in the percentage recovery of No. 2 and Better between the control and test (Table 4) was partly offset by gains in lower grades, evident by the smaller reduction in the average value of all lumber produced. The product value per cubic foot of log combines the effect of product volume and quality. In Table 5, product value is expressed in dollars (U.S.) per cunit (100 cubic feet). Based on current markets and visual grading of lumber, a mill sawing 200 cunits (566 m3) of logs per day would recover about $10,000 (U.S.) per day less in lumber from the test trees than from the control. Again, this does not consider the economics of rotation length. These comparisons of product volumes and values do not fully express the differences in lumber quality. The lumber grades commonly used are based on visual inspection of the piece and were developed for timber grown in natural stands; they may not be indicative of lumber quality of wood grown in the future under radically different conditions. Two common measures of performance of wood are extreme fiber strength (modulus of rupture, MOR) and stiffness (modulus of elas­ ticity, MOE). Most structural uses of lumber depend, in part, on these two measures. Grades of lumber assigned after visual inspection are supposed to meet certain published requirements for strength and stiffness (design criteria). Table 6 compares lumber from the control and test samples in terms of these strength and stiffness requirements. Although much of the lumber from rapidly grown trees sampled in this study may meet the published requirements for strength, only a small per­ centage will meet the requirements for stiffness. Stiffness is a critical factor, however, because in most wooden structures se­ rious problems arise from bending long before breakage oc­ curs. The consequences of such problems could be disastrous in terms of product liability if these differences were to go un­ recognized. When recognized, their importance will be re­ flected in the value of products. The previous tabulation showed that log values were one-fourth to one-third less for the fast grown material; these values would be reduced even more if the lumber was graded to meet the design criteria. How well this example will fit Douglas-fir is not known. The study was based on slash pine, and it was a relatively small sample from only one geographic area. Nevertheless, the known differences in wood properties between juvenile wood and mature wood in Douglas-fir, and the fact that juvenile , wood is produced over a longer time span in Douglas-fir, strongly suggest that similar effects will occur. Whether the magnitudes will be more or less cannot be determined without similar recovery studies for Douglas-fir. Such studies should be conducted to assess effects on the full array of potential . 22 Fight, Snellgrove, Curtis, DeBell TABLE 5. Product value of the slash pine logs, Test Control $U. s . I cunit $U. s. I $U. s. I m3 cunit $U. s . I % m3 Reduction Butt logs 146 52 98 35 33 Top logs 134 47 102 36 24 Cunit • 100 cubic feet, TABLE 6. Percentage of No, 2 and No, 3 grade lumber from the control and test slash pine trees from eastern Texas meeting strength and stiffness requirements, No. 3 No. 2 Logs Stiffness Strength Stiffness 100 100 100 100 100 100 100 100 100 19 100 19 77 8 96 9 Strength Control: Butt logs Top logs Test: Butt logs Top logs products for trees grown under a range of management re­ gimes. The'.above example illustrate negative effects of one set of management practices. Some practices, such as pruning, will have positive effects and will improve wood quality. A recov­ ery study of Douglas-fir trees that had been pruned about thirty-five years before (Cahill et al., this volume) showed about twice as much select grade lumber and over three times as much high grade veneer as was recovered from an unpruned "control" sample. REQUIREMENTS OF A HOLISTIC TIMBER ANALYSIS In broad terms, a holistic analysis should include everything that affects either costs or returns in timber management, log­ ging, transportation, manufacturing, and marketing. Note the limitation that this is a holistic "timber" analysis. Many other considerations besides the financial return to timber production go into silvicultural decisions. it is important, however, to un­ derstand the timber values and implications before trade-offs with other values can be reasonably considered. The SILMOD model (Sutton 1984) developed for radiata pine (Pinus radiata) in New Zealand is an example of a holistic timber analysis. Although this approach may have been done by some corpora­ tions, it has not as yet been used to guide management deci­ sions on public forests or on the majority of private forests in North America. Figure 1 shows the more important compo­ nents of a holistic analysis of silvicultural regimes. Necessary components for a holistic analysis are discussed below. GROWTH AND YIELD ' EVALUATION MANAGEMENT I.. COST OF SILVICULTURAL - LOGGING COST REGIMES NET VALUE OF PRODUCTS Figure 1. Important components of a holistic analysis of silvicultural regimes. Growth and Yield Yield information for coast Douglas-fir is available from published sources and stand simulators such as DFSIM (Curtis et al. 1981, 1982), DFIT (Bruce et al. 1977), TASS (Mitchell and Cameron 1985), and SPS (Arney 1985), as well as from various proprietary models. From the standpoint of a holistic analysis, the concern with these models is that they provide little beyond volume and diameter that can be used to deter­ mine product potential. Furthermore, the relationships of tree size to product yields and values that would be used in con­ junction with these models are based primarily on natural stands, and many of them may not be applicable to the man­ aged stands of the near future. Other tree characteristics that may affect product quality and value and that can be manipulated by silvicultural practices in­ clude the proportion of juvenile wood, specific gravity, ring width, ring width uniformity, knot size and kind (live versus loose), bole taper, and live crown length. All these attributes are potentially predictable using existing modeling methodol­ ogy. Management Costs Management costs are probably the least troublesome data needs. The biggest management costs are those that occur with stand establishment and early spacing control. These costs are being incurred today, and records from current management provide a good basis for the data needed. Records are also available on other commonly used silvicultural practices such as fertilization. Logging Costs Current logging costs for coast Douglas-fir are available on a piecemeal basis from a variety of sources. Logging cost rela­ tionships for young-growth coast Douglas-fir that have been developed specifically to meet silvicultural planning needs are available in Fight et al. (1984). Harvesting of stands now being established or treated, however, is far enough in the future that changes will probably occur in cost relationships associated with the developments of technology for logging small trees. The critical need in logging cost information is that it properly account for the effect of diameter and volume harvested per entry, because these variables significantly influence logging costs and are significantly affected by silvicultural decisions. Product Value The area most deficient of necessary information for a holis­ tic analysis is in the determination of the net value of products. Figure 2 outlines information needed. An estimate is needed for the value of products and the cost of manufacturing trees from managed stands when those trees are allocated to prod­ ucts, bucked, and processed. The difference between value of products and cost of manufacturing is the net value. If net product value can be determined from trees with different char­ acteristics that will be produced in managed stands, the results can be used to develop.relationships between net value and tree characteristics; these relationships can then be used in a holis­ tic analysis of silvicultural regimes. Several problems will occur in developing the net value rela­ . tionships. As with logging, the manufacture of products will TREES FROM MANAGED STANDS I ALLOCATE LOGS TO PRODUCTS I I I VALUE OF COST OF PRODUCTS MANUFACTURING I I I NET VALUE OF PRODUCTS Figure 2. Infonnation needed to detennine net value of products. Timber Quality Considerations 23 occur far enough in the future that changes in technology will change manufacturing costs and even product lines. Further­ more, current information on manufacturing cost is generally regarded as proprietary and is not readily available. The criti­ cal need for information on manufacturing costs is that it prop­ erly reflect the effect of log size. Appropriate grade and yield recovery also poses a serious problem. Although numerous mill recovery studies have been conducted for Douglas-fir (Ernst and Fahey, this volume), vir­ tually all of the studies are for extensively managed stands grown under "natural" conditions. Results of these studies will not be representative of the trees grown under intensive management. The authors know of no studies to determine re­ covery from the range of tree descriptions that will be repre­ sented in future managed stands; yet the slash pine example shows that management can have dramatic effects. Further­ more, the Douglas-fir studies deal with only a few of the tree characteristics that are influenced by silvicultural regimes. In those studies, size of trees has been used as a surrogate vari­ able for a number of quality-related characteristics. In man­ aged stands those characteristics will become more important and will not necessarily be correlated with size. A final problem will occur in developing product prices that are relevant to future markets. Various econometric models might be used to get projections of timber prices. These mod­ els only project prices for some unspecified mix of volume of products by grade. Relative prices for different grades may change dramatically as a result of changes in the abundance of and demand for lower grade and higher grade material. Fur­ thermore, these econometric models cannot be expected to be very successful in anticipating the effect of technological change on relative prices. UNCERTAINTY IN SILVICULTURAL PLANNING Because of the need to relate silvicultural decisions to prices, costs, and technological change, any analysis necessar­ ily involves numerous and uncertain assumptions. Many in­ vestment analyses ha:ve dealt with uncertainty by ignoring it. The danger in ignoring it is that the analyst may develop a sil­ vicultural strategy that is rigidly tied to a particular view of the future; this strategy may prove very costly under other reason­ able scenarios. A safer way to deal with uncertainty is to do extensive sensi­ tivity analysis. The approach to sensitivity analysis that seems most relevant to silvicultural planning proceeds as follows.. First, develop alternative futures that are internally consistent. Internal consistency requires, for example, that the cost and price assumptions be based on the same set of broad economic assumptions about wages, economic gr<?wth, and price infla­ tion. Broad professional participation is important to ensure 24 Fight, Snellgrove, Curtis, DeBell that relevant trends are recognized and that the alternative fu­ tures are internally consistent. The next step is to translate these alternative futures into sets of assumptions needed for the analysis. The analysis is repeated for each alternative future. The "best" strategy is developed by the same subjective deci­ sion process that has always been used, but with a better under­ standing of the range of outcomes that can be anticipated from alternative strategies. One consideration should be how much flexibility is provided by a particular strategy to shift to another strategy if in midrotation there are major unanticipated changes in technology or markets. An illustration of the approach pro­ posed here is discussed in another paper in this volume (Fight and Briggs). CONCLUSIONS There are several ways one could err badly in developing management strategies for coast Douglas-fir. Failure to recog­ nize how product yield and value may be affected by silvicul­ tural decisions may result in an overly simplistic solution such as ''maximize volume production and let technology take care of quality problems.'' It may be true that if enough Douglas-fir of similar characteristics is produced, technology will find a use for it. It is also very likely that the market will impose a penalty by paying a premium to those who produce wood that is better suited to meeting the needs of timber products manu­ facturers. Failure to evaluate the monetary consequences of silvicul­ tural decisions on wood quality may result in an overly sim­ plistic solution such as "minimize the amount of juvenile wood that is produced.'' Juvenile wood does have undesirable properties for many uses, but it is unlikely that minimizing ju­ venile wood is the most cost-effective means of managing the ''juvenile wood problem.'' Failure to consider alternative futures and how technology might change cost relationships in manufacturing and logging may result in an overly simplistic solution such as ''grow Douglas-fir on all sites to a specified target'size." It is unlikely that requirements in manufacturing are so rigid that the advan­ tages of a common tree size would outweigh all other consider­ ations in all circumstances. The examples above may seem overly simplified, but most of these "solutions" have been discussed before. The organi­ zational structure of corporations and public agencies and the difficulties of interdisciplinary communication often lead to tunnel vision and to solutions like these that are much less than optimum. Forest managers need to recognize that there are no right or wrong answers; there are simply various management strate­ gies with different probabilities of success. All the issues being addressed in this volume have a role in helping make a better estimate of the most desirable management strategy. A holistic analysis offers strong hope of identifying the key issues and focusing that information in a way that will give the forest manager the best possible information of what strategy to pur­ sue. doches. Texas. 16 p. Curtis, R. 0., G. W. Clendenen, and D. J. DeMars. 1981. A new stand simu­ lator for coast Douglas-fir: DFSIM user's guide. USDA For. Serv. Gen. Tech. Rep. PNW-128. Pac. Northwest For. and Range Exp. Stn., Port­ land, Oregon. 79 p. Curtis, R. 0., G. W. Clendenen, D. L. Reukema, and D. L. DeMars. 1982. Yield tables for managed stands of coast Douglas-fir. USDA For. Serv. Gen. Tech. Rep. PNW-135. Pac. Northwest For. and Range Exp. Stn., Portland, Oregon. 182 p. REFERENCES Fight, R. D., C. B. LeDoux, and T. L. Ortman. 1984. Logging costs for man­ agement planning for young-growth coast Douglas-fir. USDA For. Serv. Gen. Tech. Rep. PNW-176. Pac. Northwest For. and Range Exp. Stn., Arney, J. D. 1985. A modeling strategy for the growth projection of managed Portland, Oregon. 10 p. stands. Can. J. For. Res. 15:511-518. Bruce, D., D. J. DeMars, and D. L. Reukema. 1977. Douglas-fir managed Mitchell, K. J., and I. R. Cameron. 1985. Managed stand yield tables for yield simulator: DFIT user's guide. USDA For. Serv. Gen. Tech. Rep. coastal Douglas-fir: Initial density and precommercial thinning. Land Manage. Rep. 31. Ministry of Forests, British Columbia. 69 p. PNW-57. Pac. Northwest For. and Range Exp. Stn., Portland, Oregon. 26 Sutton, W. R. J. 1984. New Zealand experience with radiata pine. The H. R. p. Burkhart, L. F., M. D. MacPeak, and D. Weldon. 1984. Quality and yield of lumber produced from fast growth, short rotation slash pine: A mill study. MacMillan Lectureship in Forestry. University of British Columbia, Van­ couver. 21 p. Unpublished report. On file at Stephen F. Austin State University, Nacog- 1986. In: Oliver, Chadwick Dearing; Hanley, Donald P.; Johnson, Jay A., eds. Douglas-fir: stand management for the future: Proceedings of a symposium; 1985 June 18-20; Seattle, WA. Contribution no. 55. Seattle: College. of Forest Resources, University of Washington. Reproduced by USDA Forest Service, for official use. Timber Quality Considerations 25