Thinning of Small Diameter Stands in Maine

Transcription

1 Thinning of Small Diameter Stands in Maine Patrick Hiesl 1 and Jeffrey Benjamin 2 1 Graduate Research Assistant, School of Forest Resources, University of Maine 2 Associate Professor of Forest Operations, School of Forest Resources, University of Maine Maine consists of millions of acres of small diameter stands that are in need of intermediate treatments. A recent study by Hiesl et al. (2015) has shown that a whole-tree harvesting system can economically thin such stands. Data for this study was collected from one site only and part of the analysis was based on a simulation. In this paper we present a sensitivity analysis of three key input variables of that simulation (twitch size, trucking distance, and product value). We further included one observation from a second site into our analysis. Results show that product value has the greatest impact on profit, with a change of one dollar in product value being responsible for a 10% change in profit. Unit cost of production at the additional site is more than twice as high as reported by Hiesl et al. (2015), whereas profits are less than half. This clearly shows that there is lots of variation in unit cost and productivity based on feller-buncher productivity. One of the key differences between the two sites was the experience of the operators in small diameter stands. This research shows that the thinning of small diameter stands, using a small feller-buncher, can be profitable, but can also turn into a loss. Keywords: New England, Northeast, CAT 501, biomass, wood chips 1. Introduction Precommercial thinning (PCT) is a common silvicultural treatment used in the early management of conifer forests across North America and Europe (Bataineh et al. 2013; Olson et al. 2012; Zhang et al. 2006). The effects of PCT on tree growth have been investigated and documented for a wide range of forest types (Bataineh et al. 2013; Olson et al. 2012; Pitt and Lanteigne 2008; Zhang et al. 2006), however, this treatment represents a significant financial investment by the landowner which must be carried many years before a commercial harvest. In Maine, millions of acres of forestland are in need of PCT or have already passed the point of economic feasibility. The question that now arises is How can we treat these stands? Clearly these stands need to be thinned to increase growth and yield. Whether or not this thinning is commercial or precommercial depends on whether or not any profits can be 351

2 made. One factor that determines whether a profit can be made is the cost of the equipment that is used to thin these stands, while another factor may be the operator experience and proficiency. A cut-to-length (CTL) harvesting system consisting of a harvester and forwarder is commonly used to thin softwood stands (see Hiesl and Benjamin 2013). Unpublished data of studies by Hiesl (2013) and Benjamin et al. (2013), however, indicate that thinning small diameter stands with a CTL system is undesirable due to an increase in break downs, such as thrown chains and a low productivity. The use of a whole-tree (WT) harvesting system consisting of a feller-buncher, grapple skidder, and slide-boom delimber is also undesirable due to an increase in sorting time for unmerchantable stems (unpublished data of Hiesl 2013) and the high costs of feller-bunchers commonly used in Maine. Using a smaller feller-buncher and replacing the slide-boom delimber by a chipper is one option in the search of a profitable harvesting system to commercially thin small diameter stands. In the winter of 2013/2014 the Cooperative Forestry Research Unit at the University of Maine thinned a long-term herbicide and PCT research study (Bataineh et al. 2013; Newton et al. 1992a; Newton et al. 1992b), using such a system (Site A). Results of this study show similar productivity, unit cost, and profit, across three different removal intensities (Hiesl et al. 2015). This study, however, is lacking a sensitivity analysis of the input variables and a comparison of unit cost and profit to other sites. In 2013, data was collected from a similar harvest site using the same feller-buncher (make and model) but a different operator (Site B). Data from both sites can be used to evaluate the effect of operator and site conditions on unit cost of production and profit. Our objectives were to conduct a sensitivity analysis for three of the major input variables (twitch size, trucking distance, and product value) at Site A, and to compare the unit cost and profit of Site A to data collected from Site B. 2. Methods Site Selection Detailed information about Site A is described in the publications of Newton et al. (1992a), Newton et al. (1992b) and Bataineh et al. (2013). The study site is located in Somerset County, Maine (45.20 N, W). Mean annual precipitation is 40 in., with 40% of it occurring from June through September. The site was clear-cut in 1970 and a herbicide screening trial designed to release naturally regenerated conifers from competing hardwoods was installed seven years later. Sixteen years after harvest, each herbicide treatment unit (approximately 2.5 acres each) was split, with one half being pre-commercially thinned to approximately 700 trees 352

3 per acre and the other half left unthinned. During this study only the unthinned treatment units were commercially thinned. In 2012, nine fifth-acre measurement plots were installed in a subset of the unthinned treatment units. Species, dbh, total height, and height to the base of the live crown were recorded for all trees >3 inches in dbh. Quadratic mean diameter at breast height (QMD) for these plots ranged from 4.0 in. to 5.4 in. with stand densities ranging from 1,300 to 2,225 trees per acre (Table 1). Based on stem density, all stands were dominated by balsam fir (Abies balsamea (L.) Mill.), and consisted of between 4% and 28% red spruce (Picea rubens Sarg.), 1% to 30% quaking aspen (Populus tremuloides Michx.), and up to 35% of other tree species such as paper birch (Betula papyrifera Marshall), yellow birch (Betula alleghaniensis Britt.), eastern white pine (Pinus strobus L.), and northern white cedar (Thuja occidentalis L.). Individual treatment units ranged in size from 1 to 1.8 acres (Table 1). Plot 10U was used as a training plot and was removed from further analysis. A detailed description of the experimental design at Site A can be found in Hiesl et al. (2015). In short, three different thinning prescriptions, with three replications each, were implemented. The nominal thinning prescriptions were designed to remove 33%, 50%, or 66% of the standing softwood volume using a modified thinning-from-below prescription, which included the removal of large balsam fir (dbh > 8 in.) to ensure utilization of such trees before butt rot decreases their value (Tian 2002). Site B was thinned in the summer of QMD, stand density, and basal area were similar to Site A (Table 1). Species composition, however, consisted entirely of hardwoods (American beech (Fagus grandifolia Ehrh.), bigtooth aspen (Populus grandidendata Michx.), sugar maple (Acer saccharum Marshall), and red maple (Acer rubrum L.)). The removal intensity at Site A was a 67% removal of basal area, which is comparable to the 50% volume removal prescription of Site A (Table 1). Equipment Selection, Measurements, and Simulation All treatment units at Site A were thinned using a whole-tree harvesting system consisting of a CAT 501 feller-buncher and a John Deere 648 GIII grapple skidder. The CAT 501 feller-buncher was chosen for its narrow track width and small machine size. Although this machine is not widely used in Maine, productivity data of this machine in similar high density stands showed potential for economically feasible thinnings (Benjamin et al. 2013). A truck mounted Prentiss 325 loader was used to feed a Morbark Model 23 disk chipper. Detailed 353

4 information about the feller-buncher time collection, extraction time simulations, and volume estimates can be found in Hiesl et al. (2015). Plot Table 1: Individual tree and stand attributes for all treatment units and harvest sites. Treatment QMD* Prescription Removal unit (ac) (in) (%)* (tons) Stand Density (trees/ac) Basal Area (ft 2 /ac) Hardwood Component (%) BA Removed (%) Feller-Buncher Productivity (tons/pmh) Site A 2U , U , U , U , U , U , U , U , U , Site B , # Note: QMD = quadratic mean diameter, BA = basal area, PMH = productive machine hours *Removal of standing softwood volume, #Estimated hardwood volume removal Unit cost was calculated using hourly machine rates of $103 to $135 USD/PMH for the feller-buncher, $90 to $115 USD/PMH for the grapple skidder, $40 USD/PMH for the loader, and $62 to $94 USD/PMH for the disk chipper. For the grapple skidder the number of twitches per treatment unit was estimated based on the harvest volume using an average twitch size of 3 tons. We assumed that the twitches were evenly distributed along the trails within the treatment unit. Trucking costs to the mill in this region are $2.67 USD/mile (Benjamin 2014). Round-trip trucking distance for biomass chips was assumed to be between 30 and 60 miles. The average load size per truck was 26.7 tons. Mill delivered biomass chips value was supplied by an anonymous source in the industry at $35 USD/ton. A more detailed description of the simulation setup for the wood extraction can be found in Hiesl et al. (2015). Site B was also thinned by a CAT 501 feller-buncher and for the purpose of comparing the difference in unit cost and profit we assumed that the feller-buncher removed 105 tons from an area of 1.4 acres. These values represent the average condition of a treatment unit with a 50% volume removal (Table 1). The major difference at Site B was a different operator in the feller-buncher. An analysis of variance in combination with Tukey s HSD pairwise group comparison was used to compare the unit cost of production and profit between individual treatments and sites. 354

5 Sensitivity Analysis A sensitivity analysis was conducted to understand how profit reacts to small changes of an independent variable. For this we individually changed the input variables of twitch size, trucking distance, and product value at Site A. The baseline twitch size in this study was 3 tons. We changed the twitch size from 1.5 to 4.5 tons in increments of 0.1 tons to understand the sensitivity of profit to small changes in twitch size. The round-trip trucking distance for this analysis ranged from 10 to 110 miles, with an average of 60 miles as our baseline. We increased trucking distance in 10 mile increments. To understand the impact of product value on profit we used a range of $20 to $50 USD/ton, with one dollar increments. 3. Results Sensitivity Analysis A reduction in twitch size by half resulted in a profit reduction of over 40% (Figure 1 top row). In general, a reduction of twitch size had a large negative effect on profit, whereas an increase in twitch size had a considerably smaller, but positive, effect on profit. This relationship holds true across all three removal intensities. Changing the round-trip trucking distance resulted in a change in profit of up to 50% (Figure 1 middle row). This change was positive with decreasing trucking distance but became negative with increasing trucking distance. Results clearly showed that reducing the trucking distance by 10 miles can increase profit by as much as 10%. The relationship between trucking distance and change in profit is negative linear and extends to the same amount in either direction. The biggest impact on productivity was found when changing product value (Figure 1 bottom row). An increase or reduction in product value of $15 USD/ton could lead to a change in profit of up to 150%. Even a small reduction of $1 USD/ton can decrease profit by up to 10%. The relationship between product value and change in profit is positive linear and extends to the same amount in either direction. 355

6 Figure 1: Sensitivity analysis results for three thinning treatments (33%, 50%, 66% removal) at Site A. The top row shows results for changes in twitch size, the middle row shows results for changes in round-trip trucking distance, and the bottom row shows results for changes in product value. The lines in each plot represent the different treatment units and the associated feller-buncher productivity. With the exception of the 33% removal prescription (n=2) all thinning treatments have three observations. Site and Operator Comparison For Site A, biomass harvest costs ranged from $13.55 USD/ton to $30.66 USD/ton, with an average of $20.45 USD/ton (Figure 2). For site B, the harvest costs ranged from $35.54 USD/ton to $48.53 USD/ton, with an average of $42.04 USD/ton. An analysis of variance, followed by Tukey s HSD pairwise group comparison, showed that there was no difference in unit cost of production between the individual prescriptions at Site A (p > 0.929) but between 356

7 Site A and Site B (p <0.003). Profit at Site A ranged from $4.37 USD/ton to $21.45 USD/ton, with an average of $14.55 USD/ton (Figure 3). At site B, profit ranged from -$0.54 USD/ton to - $13.53 USD/ton, with an average of -$7.04 USD/ton. An analysis of variance, followed by Tukey s HSD pairwise group comparison, showed that there was no difference in profit between the individual prescriptions at Site A (p > 0.929) but between Site A and Site B (p <0.003). Figure 2: Unit cost of production for three different treatments at Site A and one thinning treatment at Site B. 357

8 4. Discussion Figure 3: Unit profit for three different treatments at Site A and one thinning treatment at Site B. Productivity of harvesting equipment is one factor that can influence the profit that can be achieved. The literature indicates that factors such as tree size, species, twitch size, and skidding distance affect productivity (see Hiesl and Benjamin 2013b). At Site A there was no control over tree size and species. Thus we did not include these two variables in our sensitivity analysis. Further, skidding distance can have a great impact on skidder productivity (Hiesl 2013; Han et al. 2004; Kluender et al. 1997), however, in our study the skidding distance was held constant. We acknowledge that for every 100 ft increase in skidding distance the productivity will decrease by more than 4% (Hiesl 2013) and thus our results would look different with varying skidding distances. Further, a recent study by Hiesl et al. (in review), showed that with an increasing skidding distance it becomes more economical to use a second grapple skidder. Such an addition of a skidder would also change the results. Twitch size depends on the number of stems, and the average tree size, but also on the loading capacity of the skidder. The individual operator has no influence on the average tree size on a given harvest site, however, he can increase the number of stems in a twitch to increase the twitch size. Our results clearly showed that increasing the twitch size by one ton increases profits by up to 10%. Reducing the twitch size by one ton, however, decreases profits by up to 30%. This exponential behavior is not surprising as the number of twitches does not 358

9 decrease by much when increasing the twitch size, but does increase to a large number when decreasing the twitch size. Owing to the calculation of the number of twitches the change in profit follows a sawtooth-pattern. In this case the total volume removed is divided by average twitch size and rounded to the next higher whole number. Trucking distance has been shown to affect driving speed (Mousavi and Naghdi 2013). Even though a longer trucking distance increase trucking speed, the total time consumption increases as well. Such an increase in time consumption subsequently increases the trucking costs and decreases the profit. It is therefore not surprising to see that the profit increases with decreasing trucking distance and vice versa. In our study, however, we assumed a constant trucking speed and used costs provided by one logging contractor in Maine. In contrast to other states, the availability of trucks in Maine is limited and has been shown to highly influence the non-productive time of chippers (Hutton 2015). These are all factors that need to be considered when applying our results to other regions. Although not easily influenced, product value is the largest driver in whether or not a profit can be achieved. Our sensitivity analysis results clearly showed that a change of one dollar per ton can cause a 10% change in profit. Such a change in profit might be enough to warrant a longer trucking distance to a mill that pays slightly more for the product. Results of Hiesl et al. (2015) indicates that the unit cost of production and unit profit of a whole-tree harvesting system operating in small diameter stands are similar across three removal intensities. Their paper further suggests that the unit profit for such a system across three removal intensities is higher than $4 USD/ton. Data for their study has been collected from one site only and therefore might be overly optimistic. As the CAT 501 feller-buncher is not a commonly used machine in the state of Maine, additional data is limited. In 2013, however, we did a productivity study of such a machine in a stand that is comparable to Site A. The observed productivity of that machine was less than half of what has been observed by Hiesl et al. (2015). This difference lead to a unit cost of production more than twice as high as observed at Site A. Subsequently the unit profit was twice as low, not even breaking even. A difference in unit cost and profit was expected, however, a difference of such a magnitude was surprising. Equipment operators in harvesters and feller-bunchers have been shown to have a large effect on productivity (Hiesl and Benjamin 2013a; Purfürst and Erler 2011; Kärhä et al. 2004). At Site A the equipment operator had over 30 years of experience working in feller-bunchers, and over two years of experience working in such stand conditions. The operator at Site B had a few years of feller-buncher experience, but only three months of experience in small diameter stands with the CAT 501 feller-buncher. There clearly is a difference in operator experience, which might explain the difference in productivity. This 359

10 comparison also shows that choosing the right operator can make a difference between making a profit and reporting a loss. Results from Site A (Hiesl et al. 2015) show that an economical thinning of small diameter stands with the proposed whole-tree system is achievable. On the other hand, however, we showed that there is the other extreme of reporting a loss when operating in such stands. To fully understand whether or not this harvesting system can economically thin small diameter stands we need to collect data from different harvest sites and different operators. 5. References Bataineh, M.M., Wagner, R.G. and Weiskittel, A.R Long-term response of spruce fir stands to herbicide and precommercial thinning: observed and projected growth, yield, and financial returns in central Maine, USA. Canadian Journal of Forest Research 43(4), pp Benjamin, J.G Operational & Economic Aspects of Biomass Harvesting: There is No Free Lunch in the Woods. Presentation at New England Society of Foresters 2014 Winter Meeting. Nashua, NH, USA: 25 March Benjamin, J.G., Seymour, R.S., Meacham, E. and Wilson, J.S Impact of whole-tree and cut-to-length harvesting on postharvest condition and logging costs for early commercial thinning in Maine. Northern Journal of Applied Forestry 30(4), pp Han, H.-S., Lee, H.W. and Johnson, L.R Economic feasibility of an integrated harvesting system for small-diameter trees in southwest Idaho. Forest Products Journal 54(2), pp Hiesl, P Productivity standards for whole-tree and cut-to-length harvesting systems in Maine. Master Thesis. Orono, ME, USA: University of Maine - School of Forest Resources. Hiesl, P. and Benjamin, J.G. 2013a. A multi-stem feller-buncher cycle-time model for partial harvest of small diameter wood stands. International Journal of Forest Engineering 24(2), pp Hiesl, P. and Benjamin, J.G. 2013b. Applicability of international harvesting equipment productivity studies in Maine, USA: A literature review. Forests 4(4), pp Hiesl, P., Benjamin, J.G. and Roth, B.E Evaluating harvest costs and profit of commercial thinnings in softwood stands in west-central Maine: A Case Study. The Forestry Chronicle. Hutton, C.J Repair and maintenance costs, chipper productivity, and chip quality of biomass supply chains in the Northeastern U.S. Master Thesis. Orono, ME, USA: University of Maine. Kärhä, K., Rökkö, E. and Gumse, S.-I Productivity and cutting costs of thinning harvesters. International Journal of Forest Engineering 15(2), pp

11 Kluender, R., Lortz, D., McCoy, W., Stokes, B.J. and Klepac, J Productivity of Rubbertired Skidders in Southern Pine Forests. Forest Products Journal 47(11/12), pp Mousavi, R. and Naghdi, R Time consumption and productivity analysis of timber trucking using two kinds of trucks in northern Iran. Journal of Forest Science 59(5), pp Newton, M., Cole, E.C., McCormack, M.L. and White, D.E. 1992a. Young Spruce-Fir Forests Released by Herbicides II. Conifer Response to Residual Hardwoods and Overstocking. Northern Journal of Applied Forestry 9, pp Newton, M., Cole, E.C., White, D.E. and McCormack, M.L. 1992b. Young Spruce-Fir Forests Released by Herbicides I. Response of Hardwoods and Shrubs. Northern Journal of Applied Forestry 9, pp Olson, M.G., Wagner, R.G. and Brissette, J.C Forty years of spruce fir stand development following herbicide application and precommercial thinning in central Maine, USA. Canadian Journal of Forest Research 42(1), pp Pitt, D.G. and Lanteigne, L Long-term outcome of precommercial thinning in northwestern New Brunswick: growth and yield of balsam fir and red spruce. Canadian Journal of Forest Research 38(3), pp Purfürst, F.T. and Erler, J The human influence on productivity in harvester operations. International Journal of Forest Engineering 22(2), pp Tian, S Effect of Precommercial Thinning on Root Develpment and Root and Butt Decay Incidence of Red Spruce and Balsam Fir. Orono, ME, USA: University of Maine. Zhang, S.Y., Chauret, G., Swift, D.E. and Duchesne, I Effects of precommercial thinning on tree growth and lumber quality in a jack pine stand in New Brunswick, Canada. Canadian Journal of Forest Research 36(4), pp