TY - JOUR
AU - Lautala, Pasi
AB - This study analyzes the cost of harvesting pulpwood from natural forests intended for the expansion of forest product opportunities in Michigan. Four sources of information were used to assess costs: (1) the USDA Forest Service Forest Inventory and Analysis database; (2) a Michigan-specific version of the USDA Forest Service Fuel Reduction Cost Simulator (FRCS); (3) primary logistics data collected from a questionnaire sent to logging firms in Michigan; and (4) primary transportation data collected from truck and rail firms. Three different harvest prescriptions were modeled: 30% selective cut, 70% shelterwood cut, and clearcut. The prescriptions were applied to fully stocked or overstocked stands analyzed from the Forest Inventory and Analysis database. Harvest systems analyzed were the following: mechanized whole-tree feller buncher with skidder and processor; mechanized cut-to-length equipment and forwarder; and chainsaws and skidder systems. Transportation analyses have been conducted for truck and bimodal (truck and rail) transportation options. Procedures and results describe the wide range of data required to analyze the cost of the logging supply chain, demonstrating the variability in the determination of a fixed cost for forest biomass removal operations. cost assessment, pulpwood, supply chain, harvesting, modeling, survey, transportation In the United States, the logging sector has been heavily affected by the economic downturn due to reduced demand for wood-frame housing, pulp and paper, and furniture. A recent regional economic analysis suggested that the total number of jobs associated with the wood products industry (direct, indirect, and induced employment) fell by 20% between 2004 and 2009 (Southern Research Station Forest Inventory and Analysis 2012). Interest is expanding for developing forest products industries, such as bioenergy, as well as new forest products markets. Understanding the supply cost of pulpwood is essential for determining the feasibility of an industry dependent on forest products. Assessing the cost of forest harvesting operations, especially in natural stands, is not straightforward. With the diverse and dynamic conditions of forest stands, area harvested, density of material removed, site conditions, terrain, harvest treatments, systems, operators' skills, configuration of equipment, and many other factors, it is difficult to estimate a uniform cost of removing biomass. To that end, this study analyzes the supply chain and logistics linked to harvesting raw pulpwood from natural forest stands in Michigan using different methods. It is based on information from (1) the USDA Forest Service Forest Inventory and Analysis (FIA) database, (2) the Michigan-specific version of the USDA Forest Service Fuel Reduction Cost Simulator (FRCS), (3) primary logistics data from a questionnaire sent to logging firms in Michigan, and (4) primary transportation data from truck and rail firms. Information from using these sources was used to model supply chain scenarios for the state. The purpose for using multiple sources and methods in assessing the cost of the harvest supply chain is to highlight how cost factors are highly dependent on the method used, stand conditions, and input type received from existing databases or key informants. Accordingly, comparing cost results is not the goal of the analysis here; rather the goal is to investigate different cost-related factors and outcomes in the supply chain of forest biomass using different methods in different stand types. Results are reported for three harvest prescriptions (30% selective cut, 70% shelterwood cut, and clearcut) applied to fully stocked or overstocked stands from the FIA database. Three harvest systems were implemented: mechanized cut-to-length equipment and forwarder; mechanized whole-tree feller buncher, skidder, and processor; and conventional chainsaw felling with skidder. The assumption is that processor cost included self-loading, and therefore no loading cost was added to the harvest system. Production equations were developed to link the FIA database and the FRCS model to generate cost assessments using standard equipment cost assessment methods (Miyata 1980). Transport of forest pulpwood and biomass is a competitive, low-margin business (Mendell and Harbor 2006), and it can be a significant component of the supply chain cost for forest biomass. Transportation of forest biomass products has some inherent features that result in a low-efficiency enterprise, including a large number of points of origin for loads, limited access through logging roads, specialized trucks that lessen opportunities for product backhauls, and a significant portion of operating hours spent loading and unloading inventory (e.g., Carlsson and Ronnqvist 2007). In most cases, biomass is transported from the forest landing to the final destination (mill or plant) by a truck in a single movement (Schroeder et al. 2007). It can be performed using a variety of truck sizes and configurations, but in the state of Michigan it is most commonly accomplished with trucks specifically designed for roundwood or chip transportation. Michigan law allows a relatively higher load than other states (164,000 lb gross vehicle weight rating for properly spaced 11-axle trucks) (e.g., Michigan Department of Transportation 2013). For longer distance movements, use of bimodal (truck + rail) transportation can become an economical transportation alternative. In a recent survey of Michigan loggers (Abbas et al. 2013), respondents were asked about their equipment and operations in regard to woody biomass transport, which provided information about the current practices within the state. According to the survey responses, rail transportation was used to move biomass only in the Upper Peninsula of Michigan. Nearly 13% of survey respondents indicated that an average of 22% of their annual production moved by rail. This result suggests that bimodal transportation is a viable option for a subset of forestry operations in the state. In a bimodal scenario, woody biomass is transported by trucks from the landing to a rail siding, where it is then either directly transferred to rail cars or temporarily unloaded for storage, before loading to rail cars. Bimodal supply chains require at least one additional handling of the load, increasing the supply chain cost. However, these costs may be offset by lower transportation unit cost by using rail. In most cases, rail cars are delivered directly to the final destination, as the additional cost for transloading back to trucks for final delivery would probably make the bimodal alternative cost prohibitive (Figure 1). In certain situations, marine (water) transportation could function as an alternative to rail, but high quantitative requirements and lack of both suitable ports and vessels make this scenario unlikely in the State of Michigan. Figure 1. View largeDownload slide Schematic diagram of single mode trucking and bimodal rail transportation for forest biomass. Figure 1. View largeDownload slide Schematic diagram of single mode trucking and bimodal rail transportation for forest biomass. Management and Policy Implications This study presents statewide scale analysis and estimated costs for the supply of pulpwood. The results demonstrate the extent to which a unit price of a ton of pulpwood is inconsistent. Furthermore, the cost of the supply of pulpwood is not much of a product issue but rather more of logistics and the conditions surrounding the product issues. In other words, procurement of wood from different conditions is a dynamic key component in determining the cost of supplying wood. If these issues are not addressed, economic and logistical considerations in the supply chain of pulpwood factors will prevent removal on several sites. Stand stocking and accessibility and transportation are integrated factors that contribute to feedstock-related market decisions. The fact that emerging pulpwood markets for bioenergy, for example, are not sought only from residue or lower value material has a very significant implication on wood products, markets, and policies. The cost of wood supply from natural stands in the long run is most likely going to increase, as more material would initially be sought from more accessible and stocked sites, leaving behind feedstock in more difficult conditions that restrict pulpwood extraction in a cost-effective manner. Forest policy and management implications are factors not only for product markets but also for the workforce of machine operators and the acceptability of the current pulpwood prices. Methodology FIA Database The USDA FIA database was queried to generate raw data for forest plots in Michigan. A Michigan-based study summarized the FIA data for 2004–2008 for eight species groups: aspen (Populus), maple (Acer), oak (Quercus), upland hardwoods, lowland hardwoods, pine (Pinus), upland softwoods, and lowland softwoods (Leefers and Vasievich 2010). Plot-level data were entered into the FRCS model to generate harvesting cost values. To reduce wide variations among stands modeled; only fully and overstocked stands were modeled. Even though a filter of fully and overstocked stands was applied, the selection still indicated very low stocking among a few stands. This was as low as 6 trees/acre with an average and maximum of 158 and 692 trees/acre for aspen stands and 238 and 1,062 trees/acre for nonaspen stands, respectively. These values are understandable, given the very diverse stocking of natural stands. The analysis assumed that aspen stands would be clearcut, whereas nonaspen stands would undergo partial cut treatments at 30 and 70% removal rates. This treatment is consistent with typical forest management treatments in the State of Michigan (Michigan Department of Natural Resources—Forest, Mineral and Fire Management 2008). The total number of FIA summarized plots analyzed exceeded 10,000 plots. The data from these raw plots were used with permission to develop the plot data used in the FRCS model for this study (Tessa Systems LLC 2011). Because the FIA was not designed for the FRCS model, plots with missing information such as the percentage of hardwoods in any one plot would eliminate the entire record. This was not a problem in our analysis, because the purpose was not to test the FIA database or the purity of the stand types but rather to develop typical plot scenarios in Michigan. Equations were developed to link between the plot data and the FRCS model. After removal of data errors of plot data and filtering stands that could not be used in the FRCS model because of missing information such as the size, volume, or density of trees in any one stand, the total number of FIA plots analyzed consisted of 1,599 and 359 stands from overstocked and fully stocked nonaspen and aspen forest types, respectively. FRCS The FRCS model developed by the USDA Forest Service (Fight et al. 2006, Dykstra et al. 2009, USDA Forest Service 2010) was used to assess the cost of harvesting natural stands within the state of Michigan. FRCS is a Microsoft Excel model application designed to estimate the cost of harvesting natural stands with optional biomass recovery as an adjunct to conventional logging operations that produce sawlogs and pulpwood. The model and its documentation may be obtained without cost (USDA Forest Service 2010). The model allows entry of site-specific data such as stand area, average yarding distance, the number of removed stems per unit area by product class (sawlog, pulpwood, or biomass), the cubic foot volume, weight, residue fraction per stem, and the mix of hardwoods versus softwoods. The average slope for both aspen and nonaspen stands was 5° per each stand type. Any of a variety of logging systems can be specified, although for this project only three harvest systems were considered: cut-to-length equipment and forwarder, mechanical whole-tree using feller buncher, skidder, and processor, and manual whole-tree using chainsaws and skidders. Options within the model permit users to specify whether estimates include the cost of moving equipment to and from the harvesting site and whether the operation involves clearcutting or partial cutting. The FRCS is a forest harvesting simulation model designed to estimate the cost of harvest operations involved in cutting and delivering trees using conventional timber harvesting operations. It uses an engineering cost method to estimate costs for individual machines (Miyata 1980). It combines machines into different harvesting systems (four ground-based systems, four cable systems, and two helicopter systems) by using the approach described by Hartsough et al. (2001) and others. The approach followed in developing the model involved reviewing relevant productivity studies published within the last 30 years. The studies were divided into two components, one used to develop the model and the other to validate the results (Hartsough et al. 2001). Production equations from additional studies were incorporated, and equipment costs were updated to December 2002. The FRCS model was designed to focus on the systems and thinnings that are designed to address the buildup of forest fuels that contribute to the risk of uncontrollable wildfire (Fight et al. 2006). Operational assumptions used both in the FRCS model and in calculating costs were based on results from the literature, face-to-face interviews, and an in-depth survey. For equipment moving cost, we assumed that the move-in and move-out distance was 20 miles each way, and the cost was amortized over the reported average harvesting area of 50 acres of a logging firm, as identified from survey results (Abbas et al. 2013). Face-to-face interviews with loggers in Michigan demonstrated that logs are typically skidded for a quarter of a mile. The average move-in and move-out cost of equipment under those assumptions was estimated per short green ton (t) of pulpwood material removed from stands. The equipment moving cost was calculated based on different stand densities, harvesting systems, and prescriptions. The move-in and move-out of equipment cost was generated by running the FRCS model for each harvesting system and prescription with and without move-in and move-out of equipment option included. Harvesting costs without the equipment moving cost were subtracted from scenarios with the move-in and move-out of equipment to identify cost per ton harvested. The model-generated transportation cost per ton was applied to the survey harvesting cost results. Survey-Based Equipment Productivity Analysis A comprehensive survey was disseminated among Michigan-based logging firms in 2009 and 2010 (Abbas et al. 2013). The survey used Dillman's Total Design Method (Dillman 2000) for mail surveys. Contact with each survey participant involved up to five contact attempts, including (1) a preliminary notice sent by mail to notify respondents about the survey and its objectives, (2) a questionnaire booklet mailed with a cover letter and a postage-paid return envelope, (3) a postcard reminder/thank you note, containing the URL to the online survey site, sent to the entire mailing list 2 weeks after the initial mailing, (4) a reminder sent to nonrespondents about 2 weeks after the previous reminder, with a replacement questionnaire and cover letter including the website to the online survey, and (5) a postage-paid return envelope and mailing of incentive checks ($20.00) to all respondents. The survey inquired about the different equipment configurations used by operators and the productivity of their systems similar to the FRCS configurations of whole-tree feller buncher, skidder, and processor, cut-to-length equipment and forwarder, and handheld chainsaws and skidder systems. Results were collected from responses that reported harvesting volumes of pulpwood for 30% removals, 70% removals, and clearcuts only in hardwood conditions to mimic 30% and 70% removals of nonaspen forest types and clearcuts of aspen trees. Survey respondents were not asked to specifically report productivity under aspen forest types or site conditions inputted in the FRCS. The difference between clearcut results for hardwood species investigated in the survey and aspen species calculated in the FRCS model was not an issue to determine clearcut volumes. Aspen (hardwoods) and pine (softwoods) forest types are two of Michigan's largest timber types and are clearcut for stand regeneration (Michigan Department of Natural Resources—Forest, Mineral and Fire Management 2008). Furthermore, comparing cost results for harvesting particular species groups is not the goal of the study; rather the goal is to identify different cost-related factors in the supply chain of forest biomass using different methods in different stand types and conditions. Transportation Transportation supply chain cost analysis is a complex process with the potential to provide a competitive advantage to a firm when properly done. Similar to equipment costing, there is typically a portion of the total transportation cost that is fixed (covering material handling, equipment maintenance, and other costs) and a portion that varies with transportation distance (e.g., fuel use and operator time). Additional “variable costs” may be added to the formula, such as interchange fees by railroads when shipments move between carriers and increased mileage charges by truckers due to substandard access roads. There may also be surcharges, such as fuel surcharges, to protect carriers from fluctuations in operating expenses. To assess the current costs of truck transportation in Michigan, costs for hauling roundwood by truck were developed based on tariff sheets obtained from several independent truckers operating within the state. Trucking rate sheets were collected through cooperation with the Lake States Shippers Association (LSSA), a regional organization dedicated to improving the competitiveness of biomass transport. Roundwood and wood chip transportation rate data were obtained from a single operator in the Lower Peninsula, and several rates were obtained for Upper Peninsula roundwood transportation, all within a time period of roughly 1 year (2009–2010). The assumption is that transportation cost included self-loading trucks, and therefore no loading cost was added to the supply system. Public rail tariff rates for roundwood were obtained directly from the website of the main operating railroad in the Upper Peninsula and examples of contract rail rates were provided by companies through the LSSA network. Rail access at the final destination was assumed. The Cost of the Supply Chain FRCS-Based Results As mentioned earlier, in the FRCS model calculations, aspen forest types were modeled under a clearcut removal scenario, whereas nonaspen forest types were modeled under partial cut scenarios. Partial removals of 30% and 70% of trees per acre were modeled to mimic selective and shelterwood cuts, respectively. The FRCS-generated results used multiple stand characteristics data input criteria described in Table 1. The FRCS model also used raw plot data fields generated from the FIA database (Tessa Systems LLC 2011) and survey results. On average, nonaspen forest type plots had larger and more trees per acre than aspen forest types. The hardwood fraction contributes to higher harvesting cost, and aspen stands have a higher portion of hardwoods. Table 1. Nonaspen and aspen mean plot characteristics. View Large Table 1. Nonaspen and aspen mean plot characteristics. View Large Plot characteristics and operational data were input in the FRCS model. Summaries of FRCS cost results per ton are presented in Tables 2, 3, and 4 for three harvesting systems (cut-to-length equipment and forwarder; mechanical whole-tree feller buncher, skidder, and processor; and manual chainsaws and skidders) for nonaspen 30% and 70% removals and aspen clearcut conditions, respectively. Table 2. Cost assessment of harvesting 30% of nonaspen stand material using different equipment configurations. View Large Table 2. Cost assessment of harvesting 30% of nonaspen stand material using different equipment configurations. View Large Table 3. Cost assessment of harvesting 70% of nonaspen stand material using different equipment configurations. View Large Table 3. Cost assessment of harvesting 70% of nonaspen stand material using different equipment configurations. View Large Table 4. Cost assessment of clearcutting aspen forest types using different equipment configurations. View Large Table 4. Cost assessment of clearcutting aspen forest types using different equipment configurations. View Large Survey-Based Results Productivity, in tons harvested per hour, was based on results reported from logging firm respondents to the survey (Abbas et al. 2013). Survey-based productivity was calculated using the standard machine costing method (Miyata 1980). Respondents were asked to report productivity based on their experience, and estimate the average production rate of their most productive equipment configuration under the various harvest conditions of 30%, 70%, and clearcut treatments. Responses were analyzed to provide production estimates as presented in Tables 5, 6, and 7. Under these conditions, respondents reported the volume of material removed on an hourly basis for their harvest system configurations: feller buncher, skidder, and processor; cut-to-length equipment and forwarder; and chainsaws and skidder. Because production was reported on an hourly basis from a regular workday these values were used as scheduled machine hours to determine the harvest cost. Labor wages and benefits were calculated to be $16.06 hour−1 (Bureau of Labor Statistics 2011) with an estimated additional 20% for benefits based on conversations with local stakeholders. Thus, each piece of equipment used in the configuration assumed a labor component of $19.43 hour−1. The size of the equipment used to harvest pulpwood was not factored in the survey questions. Therefore, it was assumed that the reported productivity was produced with either small or big equipment types. For that reason, there are two generated cost estimates, which are based on productivity reported by survey respondents. One estimate cost is based on the survey productivity with an assumption that small pieces of equipment were used, and a second estimate is based on an assumption that large pieces of equipment were used. The results of these assumptions were revealed in the different harvesting costs generated per ton. The cost of operating the equipment according to the fixed and variable cost attached to the size of the equipment used is shown in Tables 5, 6, and 7. Table 5. Survey-generated equipment configuration productivity: cut-to-length (1)/forwarder (1). View Large Table 5. Survey-generated equipment configuration productivity: cut-to-length (1)/forwarder (1). View Large Table 6. Survey-generated equipment configuration productivity: feller buncher (1)/skidder (1)/processor (1). View Large Table 6. Survey-generated equipment configuration productivity: feller buncher (1)/skidder (1)/processor (1). View Large Table 7. Survey-generated equipment configuration productivity: chainsaws (2.6)/skidder (1). View Large Table 7. Survey-generated equipment configuration productivity: chainsaws (2.6)/skidder (1). View Large Move-In and Move-Out of Equipment Cost Results Harvest cost results from the FRCS model and the survey were added to nine unique equipment moving values and added to the supply chain cost. These values were generated by modeling the 30%, 70%, and clearcut scenarios twice: once with the nine modeled scenarios (i.e., aspen and nonaspen tree stands, with 30%, 70%, and clearcut removals, using cut-to-length equipment, whole-tree feller buncher, and manual chainsaws) run in the model without the move-in and move-out of equipment and a second time with the move-in and move-out of equipment option included. The first run was subtracted from the second to identify the difference in the move-in and move-out cost, because this was not a fixed value added by the model to the harvest cost. The difference was divided by the total tons removed per acre in each scenario. This calculation resulted in 9 unique move-in and move-out cost values per site entry (Table 8), which were added as the equipment hauling cost to the harvesting cost for both the survey and FRCS model-generated results. The fuel cost was assumed to be $3.83 gallon−1 based on the US Energy Information Administration (EIA) average of year-round diesel prices (EIA CSV 2011). Table 8. Move-in and move-out of equipment: $ t−1 20 miles each way. View Large Table 8. Move-in and move-out of equipment: $ t−1 20 miles each way. View Large An understanding of the stocking is critical when the moving cost of equipment is calculated. Clearcut or partial cut treatments need not be confused with higher or lower removal rates for the different stand types analyzed (Table 1). For example, aspen trees harvested per hectare were fewer and smaller than the nonaspen stands that received the partial 70% treatments. Hence, the moving cost of equipment per ton of nonaspen partial cut pulpwood (2,131 t) was lower than that for aspen stands with the clearcut treatment (1,661 t). Stumpage Price Stumpage prices used to assess supply cost were based on current market reports for aspen and nonaspen forest species (Tessa Systems LLC 2009) (Table 9). These values were averaged and added to harvesting (Tables 2, 3, 4, 5, 6, and 7), move-in and move-out of equipment cost (Table 8), and transportation supply costs (Table 10). Table 9. Average stumpage price per species group. Data from Tessa Systems LLC (2009). cd, cord: 4 ft × 4 ft × 8 ft dimensions. * The value of 2.3 t/cd was based on the average tonnage per cord for different species in the northcentral region. Estimates are based on Hackett and Smith (1990). View Large Table 9. Average stumpage price per species group. Data from Tessa Systems LLC (2009). cd, cord: 4 ft × 4 ft × 8 ft dimensions. * The value of 2.3 t/cd was based on the average tonnage per cord for different species in the northcentral region. Estimates are based on Hackett and Smith (1990). View Large Table 10. Linear transportation distance and associated costs for trucking forest biomass. View Large Table 10. Linear transportation distance and associated costs for trucking forest biomass. View Large Transportation Cost Results Truck transportation rates are commonly based on the freight tonnage and length of haul, allowing for generalized rate estimates for various distances and tonnages in a single formula. In addition to the distance-based fee, all collected rates contained a fixed-cost value ($3.85) that accounted for the loading and unloading. Sample trucking rates demonstrated a high degree of uniformity regardless of forest biomass type or location (Figure 2). An average forest biomass transportation cost for all truck transportation movements was estimated. All of the collected individual rates per ton of material moved for various distances are presented in Figure 2. From the collected data, a linear regression was performed to develop the generalized equation for forest biomass truck transportation in Michigan (Equation 1; Figure 2, solid line).   where DC is delivery cost at a given distance ($ t−1) and Dt is distance traveled (miles). Figure 2. View largeDownload slide Transportation costs for forest biomass in Michigan, based on location (Upper Peninsula [UP] versus Lower Peninsula [LP]) and product type (roundwood versus chips). No data were collected in the Upper Peninsula on wood chip transport, but based on industry interviews, the rate structure is expected to be similar to the overall trend. As explained in the text, the overall rate equation for forest biomass transport based on collected data (solid line) was altered to more accurately reflect prevailing fuel prices in 2011 (dashed line). Figure 2. View largeDownload slide Transportation costs for forest biomass in Michigan, based on location (Upper Peninsula [UP] versus Lower Peninsula [LP]) and product type (roundwood versus chips). No data were collected in the Upper Peninsula on wood chip transport, but based on industry interviews, the rate structure is expected to be similar to the overall trend. As explained in the text, the overall rate equation for forest biomass transport based on collected data (solid line) was altered to more accurately reflect prevailing fuel prices in 2011 (dashed line). It is estimated that the loading and unloading operations required for moving logs account for roughly half of the fixed-cost portion of this price structure, based on discussions with forest industry experts in the region. To estimate the transportation cost increase due to higher fuel prices since the time data were collected, we followed the approach outlined in Lautala et al. (2011) for calculating the anticipated increase in variable cost per ton-mile of transport for each $1 increase in the prevailing diesel fuel price. By using the difference in baseline diesel fuel costs between the time of data collection ($2.67 gallon−1) and an agreed-on date that represents a significant price increase of diesel fuel ($3.83 gallon−1 in July 2011) (EIA CSV 2011), our variable costs for truck transportation were likely to increase by 18% to   Transportation costs are assessed in the context of prevailing fuel price to determine whether previous assessments will accurately reflect current market conditions for the scenario in question. Development of rate equations for rail transportation is more challenging, because rates rarely follow a generalized mathematical formula but rather are developed separately for each individual origin-destination pair. Every rail service provider has a specific policy and rate to charge customers, and rail customers can obtain these rates through railroad websites (tariff rates). However, rail service providers often reward consistent shippers of sizeable volumes with individually negotiated contract rates, adding another layer of complexity to the rate structure. In addition to company policies, other important considerations in the decision to ship (or not ship) by rail include the proximity of biomass origins to the nearest rail loading site and several operational and local considerations that are summarized in Table 11. Table 11. Important considerations for bimodal rail transportation of woody biomass. Adapted from Cheaney (2009). View Large Table 11. Important considerations for bimodal rail transportation of woody biomass. Adapted from Cheaney (2009). View Large Analysis of rail rates for wood chips was not conducted, because there were no examples of chips moving by rail in the region, but tariff rail rates for roundwood were obtained for 100 random origin-destination pairs in the study region. The analysis revealed a small variable cost per ton for the initial 100 miles, but a significant increase in variable costs for trips longer than 100 miles (Figure 3-A). For the first 100 miles of any rail trip, the majority of costs are fixed, because of capital expenses and other operational costs, largely independent of the quantity being shipped. For contract prices, the rates obtained from LSSA members suggested that the variable costs were similar to tariff rates, with a flat fee discount applied to the fixed cost of each shipment. Based on industry interviews, the majority of forest product rail movements in the region occur under contract rates, so the final rate formula included a uniform contract rate discount (Figure 3A, dashed line), whereas the rate of increase in variable cost was maintained equal between tariff and contract rates. It was also noticed that rates provided by LSSA included only shipments more than 100 miles, demonstrating the trend that most rail shipments take place in longer movements (Figure 3B). Figure 3. View largeDownload slide A. Data obtained from a rail operator in the Upper Peninsula of Michigan (solid markers) were used to construct a segmented tariff rail cost structure, with a significant increase in variable costs at roughly 100 miles. B. Comparison of the tariff rate structure to the samples of contract shipping rates for logs (open markers) from different companies illustrated how rates are usually discounted from published tariff rates. Figure 3. View largeDownload slide A. Data obtained from a rail operator in the Upper Peninsula of Michigan (solid markers) were used to construct a segmented tariff rail cost structure, with a significant increase in variable costs at roughly 100 miles. B. Comparison of the tariff rate structure to the samples of contract shipping rates for logs (open markers) from different companies illustrated how rates are usually discounted from published tariff rates. The development of bimodal rates requires merging rail and truck rates. In a recent effort to model transportation of forest products in the northern Great Lakes region, it was revealed that logs move on average 30 miles from the forest landing areas before being loaded onto rail cars (Lautala et al. 2011), so trucking charges for 30 miles are added to the rail rates to form the complete bimodal cost (Figure 4). When truck rates are compared with bimodal rates, truck transportation in the Upper Peninsula is more cost efficient than the bimodal alternative for trips less than 125 miles of total distance (combined truck and rail). This “breakpoint” is higher than that in similar analyses conducted in Finland, which found that rail becomes cost competitive with truck transportation of unchipped forest biomass at less than 100 miles of total transport, assuming less than 20 miles of truck transport before rail loading (Tahvanainen and Anttila 2011). However, the breakpoint is not fixed, but depends on multiple variables, such as the distance from forest landing to rail loading location and prevailing fuel prices. Our analysis relied on transportation rates that were collected within a fairly narrow time frame in 2009–2010 and were modified to reflect higher fuel prices during 2011. Fuel surcharges for rail transport are much less significant than those for truck transport, because of the increased fuel economy of rail transport. Further investigations should be calibrated for the current or anticipated fuel price regime or examine the sensitivity of results to fuel price (e.g., Lautala et al. 2011). The rates presented in Figure 4 are also based on three additional assumptions: that all material is directly unloaded from a log truck to a waiting rail car during a bimodal trip (“hot-loading”), that the origination and destination of rail movements is within a single railroad company (no interchange required), and that the final destination can be accessed by rail. If rail cars are not present at the siding when log trucks arrive or if intermediate storage at siding is preferred, logs will need to be unloaded to the ground and later loaded to the rail cars either by log trucks or designated loaders. This extra handling step represents an additional cost, approximately $4–6 cord−1, that would have to be included according to industry professionals. If the final destination is not accessible by rail, material would have to be loaded back to trucks for final delivery, and if an interchange between railroad companies is required, the cost formulas become highly unreliable because of various rate arrangements. In general, the additional cost and related time delay or either storage or railroad interchange would make bimodal transportation economically infeasible. Figure 4. View largeDownload slide Comparison of trucking and bimodal transportation rates for logs in the Upper Peninsula of Michigan. Figure 4. View largeDownload slide Comparison of trucking and bimodal transportation rates for logs in the Upper Peninsula of Michigan. With use of Equation 2 above and the data presented in Figure 4, it is possible to generate a series of transportation costs for movement of forest biomass within Michigan. Table 10 briefly illustrates the sample costs of truck and bimodal transportation for selected distances. Calculating the Full Supply Chain Cost The FRCS and survey-based productivity results generated multiple harvesting cost values per ton harvested. The multiple values from natural stand harvest cost calculations reveal the ever dynamic nature of harvesting from natural stands. Cost factors are site specific and depend on site conditions, harvesting prescriptions, biomass stocking, forwarding distance, equipment hauling costs, stumpage price, number of times equipment is hauled to the site, the operators' skills, and terrain (Hartsough 1989, Hartsough et al. 1997, Rummer 2002, Abbas et al. 2011). Equation 3 and Table 12 present the cost analysis of the total supply chain cost value of forest harvesting logistics, stumpage, and transportation equation. The values are provided up to 150 miles for one-way transportation distance, because beyond that distance, rail transportation should be brought into consideration, making total supply chain cost modeling more challenging. One bimodal supply chain value is calculated and listed in the last column of Table 12 for comparison purposes.   where TC is total pulpwood supply cost of delivery at a given distance ($ t−1), S is stumpage cost averaged (nonaspen, $11.43 t−1; aspen, $11.36 t−1), H is cost of harvesting ($ t−1), MEq is move-in and move-out of equipment (20 miles) both ways (Table 8), and DC is cost of delivery at a given distance ($ t−1). Table 12. Pulpwood total supply chain per green ton average cost analysis at different truck hauled distances. CTL, cut-to-length equipment and forwarder system; WT Mech, whole-tree feller buncher, skidder, and processor system; Manual, manual chainsaws and skidder system. * Highest and lowest supply costs in the systems modeled. View Large Table 12. Pulpwood total supply chain per green ton average cost analysis at different truck hauled distances. CTL, cut-to-length equipment and forwarder system; WT Mech, whole-tree feller buncher, skidder, and processor system; Manual, manual chainsaws and skidder system. * Highest and lowest supply costs in the systems modeled. View Large Results and Discussion The task of determining the cost analysis of large-scale forest harvesting technology is not straightforward. The nature of forest stands in terms of stand conditions and different operating technologies and logistic results in more diverse cost values compared with those for fixed plantation harvesting, for example. To attempt to address this issue, the project used the existing FRCS model developed by the USDA Forest Service and adapted it to more Midwestern regional characteristics. In an attempt to develop another source of production-based cost assessment, survey results were used. The FRCS assesses the cost of harvesting based on literature sources from the region and factors in very particular stand characteristics that were generated from the FIA database for Michigan. The survey data source used primary productivity and transportation data collected from equipment operators and analyzed using standard equipment cost assessment methods. The reported results do not include potential profit to the operators nor overhead and management expenses. Understandably, the cost estimates generated by FRCS were in many cases substantially different from those derived from the calculated averages of productivities from the survey results. A number of factors may account for the differences and help explain cost disparities between the FRCS and survey productivity results. The stand conditions perceived by the survey respondents may have been different from the statewide FIA averages used in the FRCS calculations. For example, the survey results produced lower costs for the clearcut stands than for partial cutting. This finding would be logical if the average tree size removed is the same, but the inventory data show the average size in the aspen clearcuts to be only two-thirds that for the partial cut stands analyzed in the FRCS. Survey respondents may also be reporting average productivities for the combination of sawlogs and pulpwood they typically harvest, rather than incremental productivities for pulpwood. The latter volumes are probably lower, because of the smaller size of pulpwood trees. Each logger reported overall system productivity for their specific mix of equipment. To generate the hourly costs to apply to these survey-based productivities, we assumed a configuration ratio of 1 feller buncher:1 skidder:1 processor for the whole-tree system; 1 cut-to-length machine:1 forwarder for cut-to-length; and 2.6 for chainsaws:1 skidder for the manual system. The 2.6 chainsaws reported productivity is for a crew with handfelling constituting more than 50% of their total productivity. In contrast, FRCS calculates productivities and costs separately for each function, e.g., cut-to-length felling and forwarding. Finally, the productivity relationships included in FRCS are from other midwestern, northern, and northeastern areas because detailed studies are not yet available for Michigan, so FRCS may not accurately reflect the typical harvesting conditions in Michigan. In this study, we applied different methods, based on a cost simulator model and survey results, to assess the cost of the harvest and delivery supply chain. This approach has highlighted the extent of which cost factors are highly dependent on the method used, stand conditions, and input received from existing databases or key informants. Table 12 summarizes the total supply chain cost results from the study. The cost results of harvesting aspen stands using the FIA database differ from those for feedstock harvested from unspecified hardwood species. The FIA database allowed an understanding of the density of aspen stands, compared with that of other hardwood stands harvested by the respondents to the survey. As a result, there was a wide range of results for both supply chain calculations. The variations in harvesting system used (manual versus mechanical), stand characteristics, stocking, tonnage removed, tree size, number of trees per acre, and size of equipment used were key factors that resulted in the very different occurrences of harvest cost values. The variability in the cost per ton because of these combined factors and inputs alone is significant. For example, the highest and lowest supply costs in the systems modeled appeared within the same clearcut prescription in the harvest supply chain (Table 12). These results help to demonstrate how methods and inputs identified are critical in determining the nonfixed cost of harvesting, in contrast with the more likely fixed product prices in the market. Determining a fixed cost per ton without knowing the stand and logistics conditions could be difficult, but results show a ballpark estimate of what could be expected under similar study situations. From the results, it was found that on-site operations represent the largest component of costs associated with the delivery of forest raw materials. The cost of harvesting alone ranges from as low as 32% to as high as 57% of the total supply cost, within 150 miles from a site using truck delivery. Harvesting plus move-in and move-out of equipment ranges from as low as 45% to as high as 63% of the total supply cost; this is within 150 miles from a site using truck delivery. Material harvesting and move-in and move-out of equipment cost plus stumpage price constitutes a range from as low as 66% to as high as 77% of total supply cost, within 150 miles using truck delivery. Of the total supply and delivery cost, harvest cost averaged 44.5%, move in/out of equipment averaged 9.5%, stumpage averaged 17.5%, and transportation averaged 28.5%. Transportation costs form a significant portion of overall supply chain costs, but the heavy competition causing the rates between truck service providers to be closely aligned facilitates general rate modeling. For biomass transport of less than 125 miles it appears that truck transportation is the most cost-effective method of transportation, assuming that a truck trip of 30 miles would be required before rail transport. For longer movements or situations involving shorter truck movements to rail sidings, it would be prudent to investigate the potential use of bimodal transport systems involving rail transport. These decisions will often require individual assessment to examine particular characteristics that may favor one transport mode over the other, such as proximity to rail sidings or prevailing fuel price. Conclusion In this study, the potential increase in demand for wood products for growing industries was explored by investigating different methods that calculate the supply options of pulpwood from existing natural stands. Information presented in this article helps identify some of the key logistics and cost factors that need to be accounted for when a new wood product or bioenergy facility dependent on pulpwood is built. Based on the analysis, biomass utilization economics is sensitive to harvesting, mobilization costs and hauling distance, and method to markets. Currently, biomass markets are only paying for a fraction of the full cost of harvesting, forwarding, and delivery of the biomass material. This fact indicates the importance of the development of a technical understanding of the supply conditions in line with market values of harvested pulpwood material to better support sustainability of growing economies. Without factoring in the full cost of the supply chain, the workforce is most likely going to direct its resources to different markets, to promote a different business to pay off equipment costs, or change profession. To conclude, companies naturally favor stable prices for the harvesting, supply, and delivery of forest biomass. In most cases the same equipment used to harvest and transport biomass is used for higher-value products such as sawlogs. In fact, usually the removal of small trees and residuals in natural stands results in lower productivity of the operator. Hence, to ensure a more cost-effective supply, forest machine and transportation operators need to focus on constraining inefficient supply chains and transportation radii wherever possible. The most effective way of ensuring a sustainable supply of biomass is to promote operations that harvest both high- and low-value products at the same time. Market dynamics will inevitably play a major role in determining whether biomass can be sustainably supplied to consuming industries. Acknowledgments: We extend a special thank you for the support provided by the Department of Energy (Award No. DE-EE-0,000,280), forest machine operators and logging firms from the State of Michigan, forestry consultants, officials from the Michigan Department of Natural Resources, and many others. We acknowledge support and input from multiple experts from Michigan Centers of Excellence; from the Michigan Department of Natural Resources; from Michigan State University, especially Prof. Michael Vasievich; and from Michigan Technological University, Stephen Chartier Jr. and Hamed Pouryousef. We also thank three anonymous reviewers and the Journal of Forestry associate editor and editor-in-chief for constructive and helpful remarks. This report was prepared as an account of work sponsored by an agency of the US Government. Neither the US Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, or service by tradename, trademark, manufactured, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the US Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the US Government or any agency thereof. Literature Cited Abbas D., Current D., Ryans M., Taff S., Hoganson H., Brooks K. 2011. Harvesting forest biomass for energy: An alternative to conventional fuel treatments: Trials in the Superior National Forest, USA. Biomass Bioenergy  35( 11): 4557– 4564. Google Scholar CrossRef Search ADS   Abbas D., Handler R., Hartsough B., Dykstra D., Srivastava A., Hembroff L., Shonnard D., Lautala P. 2013. A survey of forest harvesting and transportation logistics. North. J. Appl. For . In press. Bureau of Labor Statistics. 2011. Operational data annual statistics database . Available online at www.bls.gov/oes/current/oes_mi.htm#45-0000; last accessed May 4, 2013. Carlsson D., Ronnqvist M. 2007. Backhauling in forest transportation: Models, methods, and practical usage. Can. J. For. Res . 37( 12): 2612– 2623. Google Scholar CrossRef Search ADS   Cheaney D. 2009. How to ship by rail . Simons-Boardman Books, Inc., Omaha, NE. 464 p. Dillman D. 2000. Mail and Internet Surveys: The tailored design method , 2nd ed. John Wiley and Sons, New York. 464 p. Dykstra D., Hartsough B., Stokes B. 2009. Updating FRCS, the fuel reduction cost simulator, for national biomass assessments. In Proc., Environmentally sound forest operations. 32nd annual meeting of the Council on Forest Engineering, June 15–18, 2009, Kings' Beach, CA. CD-ROM. University of California, Davis, Davis, CA. US Energy Information Administration CSV. 2011. Today in energy. 2011 brief: US average gasoline and diesel prices over $3 per gallon throughout 2011 . US Energy Information Administration Independent Statistics and Analysis. Available online at www.eia.gov/todayinenergy/detail.cfm?id=4570; last accessed Dec. 6, 2012. Fight R.D., Hartsough B., Noordijk R. 2006. User's guide for FRCS: Fuel reduction cost simulator software . USDA For. Serv., Gen Tech. Rep. PNW-GTR-668, Pacific Northwest Research Station, Portland, OR. 23 p. Hackett R.L., Smith W.B. 1990. Pulpwood production in the North Central Region. USDA For. Serv., Resource Bull. NC-123, North Central Experiment Station, St. Paul, MN. 29 p. Hartsough B.R. 1989. Harvesting small stems and forest residues in the Pacific Southwestern United States. P. 100– 107 in Proc. of the International Energy Agency, task vi, activity 3 symposium. Harvesting small trees and forest residues, June 5–7, 1989, Auburn, AL. USDA For. Serv., Southern Forest Experiment Station, Asheville, NC. Hartsough B., Drews E.S., McNell J.F., Durston T.A., Stokes B. 1997. Comparison of mechanized systems for thinning ponderosa pine stands and mixed conifer stands. For. Prod. J . 47( 11/12): 59– 68. Hartsough B.R., Zhang X., Fight R.D. 2001. Harvesting cost model for small trees in natural stands in the Interior Northwest. For. Prod. J . 51( 4): 54– 61. Lautala P.T., Handler R.M., Hicks J.W. 2011. Spatially based model to determine price-optimal log transportation by trucks and rail in the upper Mid-West—Development and initial outcomes. Transportation Research Board 90th Annual Meeting. Transportation Research Board, Rep. No. 11-3632, Washington, DC. Leefers L., Vasievich M. 2010. Timber resources and factors affecting timber availability and sustainability for Kinross, Michigan. Prepared for Feedstock supply chain, center of energy excellence project . Project 2 Rep. Version 2.0, December 2010. Frontier Renewable Resources, LLC, Michigan State University, Michigan Technological University, Kinross, MI. Available online at www.michiganforestbiofuels.org/sites/default/files/timber-resources-affecting-availability-sustainability-appendix.pdf; last accessed Apr. 15, 2013. Mendell B., Harbor A. 2006. Evaluating the potential for shared log trucks resources in Middle Georgia. South. J. Appl. For . 30( 2): 86– 91. Michigan Department of Transportation. 2013. Michigan's truck weight law and truck user fees . Michigan Department of Transportation, Intermodal Policy Division. Available online at www.michigan.gov/documents/mdot/MDOT_013–4-16TruckWeightsMichigan_418609_7.pdf; last accessed Apr. 12, 2012. Michigan Department of Natural Resources—Forest, Mineral and Fire Management. 2008. Questions and answers about clearcutting. Issued by authority of the Michigan Department of Natural Resources . Available online at www.michigan.gov/documents/IC4270_48581_7.pdf; last accessed Apr. 12 2013. Miyata E.S. 1980. Determining fixed and operating costs of logging equipment . USDA For. Serv., Gen. Tech. Rep. NC-55, North Central Forest Experiment Station, St. Paul, MN. 16 p. Rummer R. 2002. Forest operations technology. Chapter 15 in Southern forest resource assessment , Wear D.N., Greis J.G. (eds.). USDA For. Serv., Gen. Tech. Rep. SRS-53, Southern Research Station, Asheville, NC. 635 p. Schroeder R., Jackson B., Ashton S. 2007. Biomass transportation and delivery. Sustainable forestry for bioenergy and bio-based products: Trainers curriculum fact sheets 4.5 . Southern Forest Research Partnership, Inc., Athens, GA. Southern Research Station Forest Inventory and Analysis. 2012. The inventory: An update concerning the SRS FIA program . Available online at srsfia2.fs.fed.us/program_information/March%202012%20%28The%20Inventory%29.pdf; last accessed Mar. 19, 2012. Tahvanainen T., Anttila P. 2011. Supply chain cost analysis of long-distance transportation of energy wood in Finland. Biomass Bioenergy  35( 8): 3360– 3375. Google Scholar CrossRef Search ADS   Tessa Systems, LLC. 2009. Timber supply outlook for Kinross, Michigan . Final Report. Prepared for Frontier Renewable Resources, LLC, Marquette, MI. 97 p. Tessa Systems, LLC. 2011. Dr. Michael Vasievich, East Lansing, MI. USDA Forest Service. 2010. Fuel reduction cost simulator . USDA For. Serv., Pacific Northwest Research Station, Portland, OR. Last updated March 26, 2010. Available online at /www.fs.fed.us/pnw/data/frcs/frcs.shtml; last accessed May 2, 2012. Copyright © 2013 Society of American Foresters
TI - Cost Analysis of Forest Biomass Supply Chain Logistics
JF - Journal of Forestry
DO - 10.5849/jof.12-054
DA - 2013-07-01
UR - https://www.deepdyve.com/lp/springer-journals/cost-analysis-of-forest-biomass-supply-chain-logistics-xv4PXjZ0yf
SP - 271
EP - 281
VL - 111
IS - 4
DP - DeepDyve
ER -