Mobile Edition
Print
Get the Mag
Weekly UpdatesAbstract
Value recovery and cost of a modified tree-length (MTL) logging system that measures product dimensions using a Waratah harvester were compared to that of a tree-length (TL) system that estimates dimensions. Paired comparisons of residual timber value (delivered price--harvesting cost) per harvested acre on a series of three planted pine clearcuts were evaluated. One half of each site was harvested with a TL crew and the other half with a MTL crew. A cost model was developed to compare the cost of the TL and MTL systems. The MTL system recovered more volume per acre than the TL system. Although not statistically significant, the MTL system also recovered slightly more residual timber value per acre than the TL system on three sites despite modeled logging costs that were $0.93 per ton higher. The MTL system also showed consistent, but not statistically significant, increases over percent cruised value compared to the TL system. Even with the small sample size, a sensitivity analysis with delivered prices suggests that MTL systems can recover more value than TL systems with current markets or a 5-percent increase in sawtimber and chip-n-saw precut product prices considering only value recovery potential, not harvesting cost differences.
**********
Numerous studies have found that cut-to-length (CTL) harvesting systems recover more value at a lower additional sorting cost than tree-length (TL) systems (Gingras 1996, Gingras and Soucy 1999). Other studies have shown that CTL systems in the Southeast can recover up to 90 to 94 percent of the optimum value of a harvested stand (Boston and Murphy 2003, Conradie et al. 2004). Regrettably, most markets in the southeastern United States demand TL products. Some TL harvesting operations in the southeastern United States are beginning to use harvesters to aid in product sorting and bucking at roadside. These modified tree-length (MTL) systems produce bucked logs similar to a CTL operation.
Advantages of using a harvester to measure and buck logs may be higher value recovery for the landowner, measurements and information about products from the harvesting site, and increased loader production by allocating the product sorting function to the harvester. Disadvantages may be higher logging costs from adding a high-cost piece of equipment to a system and lack of markets in the Southeast for cut log products. No studies have examined the cost or value recovery of MTL systems in comparison to TL systems in the Southeast. This study evaluated these harvest systems using a tract level paired harvest approach.
Methods
Study sites
To conduct the paired comparisons on harvest tracts, four tracts were identified that our industry cooperator had scheduled for clearcut in 2007. All of the study sites were loblolly pine (Pinus taeda L.) plantations between 24 and 33 years of age located in central Georgia. Each tract was divided roughly in half to form two blocks of approximately equal acreage. A TL or MTL crew was randomly assigned to harvest each block. For this test, the same TL crew and MTL crew was used for all four sites. Block size ranged from 43 to 67 acres.
In May 2007, a fire burned portions of Site D. The fire burned approximately 60 percent of the MTL block compared to only 25 percent of the TL block. Because the fire appeared to affect the study areas unevenly, and thus affect the value of each block unevenly, the decision was made to remove this tract from the analysis and only report results for the three remaining study sites.
The smaller sample size (n = 3) made statistical comparisons difficult. Nonparametric tests are usually tests of choice for small sample sizes, but in this case the sample size was so small that the smallest p-value attainable with a Wilcoxon signed rank test was 0.25 (Hollander and Wolfe 1999). Paired t-test results were relied on for significant differences. The small sample size was difficult to test for normality to assure that the assumptions of a parametric paired t-test were met.
Logging contractors
The TL crew's equipment consisted of two grapple skidders (a John Deere 643G2 and a John Deere 643G3), a John Deere 843H feller-buncher, a Husky 235 knuckleboom loader, two chain saws, and a delimbing gate. At times the TL crew used a Franklin KBL-28 loader instead of the Husky loader. The feller-buncher felled and bunched stems that the skidders backed through a delimbing gate before skidding to the landing. The loader operator sorted the stems on the landing, separating stems that included potential sawtimber precut products, one of the higher valued products. Chain saw operators measured the length and bucked the sawtimber precuts. Chain saw operators also topped trees and cleaned up loads on the trucks.
The MTL crew's equipment consisted of a John Deere 2054 shovel with a 622 Waratah harvester head operated as a processor on the landing, a John Deere 648G3 grapple skidder, a Tigercat 230B knuckleboom loader, and a Tigercat 724D feller-buncher with a 5500 felling head. This contractor shared the feller-buncher between two crews, although only one crew participated in this study. The feller-buncher felled trees before the rest of the crew arrived at a tract and would attempt to stay 1 to 2 days ahead of the rest of the crew. The skidder moved stems to the landing where the processor delimbed, sorted, and bucked logs. The processor operator utilized diameter and length measuring technology in the harvester head to aid in decision-making but did not use a log optimization program to make bucking decisions. The processor operator had 4 years of experience operating a processor. The loader loaded trucks and moved piles of product to keep the landing neat for the processor to work.
Paired harvests
Pre-harvest timber inventory data were calculated separately for each harvest block by the industry cooperator. Delivered prices were determined by applying observed market price differentials to Timber Mart-South delivered prices for 2007 Georgia averages (Harris et al. 2008). The residual timber value for each product class was calculated by subtracting harvesting cost from the delivered price (Table 1). Residual timber values were combined with pre-harvest volume per acre estimates to obtain value per acre estimates. Pre-harvest product values ($/acre) on each study site were compared between harvest blocks and were found to be equivalent for all of the study sites (p < 0.05).
Where multiple markets existed for the same product, the median residual timber value for that product class was used to value cruised volumes. Per acre harvested values were compared to pre-harvest estimates to determine the percentage of pre-harvest value actually harvested by each logger on each block. The percentage of cruised value harvested by each logger was compared with a t-test in SAS (SAS Institute Inc. 2002 2004). After the harvests were complete, a paired t-test and Wilcoxon signed rank test in SAS were used to compare tons per acre harvested and average residual timber value per acre harvested by each system (SAS Institute Inc. 2002-2004).
Product sorts
A total of 14 product assortments were analyzed (Table 1). All of the products were pine except for one hardwood pulpwood assortment. Not all of the products were harvested by each contractor on each site. Proximity to market and timber stand characteristics varied somewhat across the harvest blocks. The maximum number of assortments recorded on one harvest block was nine cut by the MTL crew on Site C.
Cost model
The Auburn Harvesting Analyzer was used to compare the harvesting cost per ton of MTL and TL systems and was adapted to model the logging cost of the MTL system (Tufts et al. 1985). Hourly equipment costs were estimated with the machine rate method (Miyata 1980). Both systems were assumed to have the same basic equipment configuration including a feller-buncher, grapple skidder, and knuckleboom loader. Two key differences separated the systems: the MTL crew had a harvester to serve as a processor on the landing and the TL crew had two saw-hands to be consistent with the crew in the study. Assumptions included a 5-percent return on assets as profit and a $6,000 monthly owner salary. The cost for one entrance, one push-out road, and a half-mile of road for each system were incorporated into the model. Additional machine rate assumptions are shown in Table 2.
Sensitivity analysis of paired harvests
While current markets in the U.S. South demand primarily TL products, mills in other parts of the world including Scandinavia, Europe, Canada, Australia, and New Zealand use CTL products from the woods. Mills may reduce processing costs by purchasing precut products from logging contractors; products are delivered cut to end-product specifications and mills no longer pay to handle waste material that comes from processing a TL log. The argument for purchasing TL logs rests on additional income streams from selling the waste products and the mill's ability to cut products to length more accurately than logging contractors. If mills in the U.S. South were to shift to precut products, would modified systems become more cost-effective? The sensitivity of returns to the landowner from the MTL system were examined with changing prices for precut log products. A sensitivity analysis was performed to compare the mean difference in both delivered and residual value per acre between MTL and TL systems using different price scenarios. Delivered value per acre provides a sense of value uplift from using the processor as delivered value does not reflect the impact of different harvesting costs between TL and MTL systems. Residual timber value (delivered price--harvesting cost) reflects both value recovery and harvesting cost impacts. For the sensitivity analysis, multiple assortments of the same product type were valued at the highest delivered price for that product. For example, all of the TL sawtimber product volumes (ST1 and ST2) were valued at the ST1 delivered price (Table 1). Price ratios were calculated by dividing each delivered product price by the delivered pine pulpwood (PPW) price. Pine super-pulpwood (PSP) was excluded from the analysis. The base case was the mean difference in value per acre calculated by multiplying actual harvested product volumes from the paired harvest study by their respective prices previously described.
FEED