- Email: [email protected]
An alternative skidding technology to the current use of crawler tractors in Alpine logging operations
Journal of Cleaner Production 31 (2012) 73e79
Contents lists available at SciVerse ScienceDirect
Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro
An alternative skidding technology to the current use of crawler tractors in Alpine logging operations Raffaele Spinelli a, *, Natascia Magagnotti b, Rubén Laina Relaño c a
CNR e IVALSA, Via Madonna del Piano, Pal. F, I-50019 Sesto Fiorentino (FI), Italy CNR e IVALSA, Via Biasi 75, I-38010 San Michele all’Adige (TN), Italy c Universidad Politecnica de Madrid, Ciudad Universitaria s/n., E-28040 Madrid, Spain b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 26 November 2011 Received in revised form 26 February 2012 Accepted 27 February 2012 Available online 9 March 2012
The study investigated the possibility of replacing aggressive crawler tractors used in Alpine logging operations with a new rubber-tired mini-skidder, in the same size and cost class. Replacement proved desirable in terms of environmental protection and labor safety, and offered substantial economic benefits. While it required a moderate additional initial investment, extraction cost was reduced between 30 and 50%. On the average skidding distance of 130 m, productivity was 3.2 and 4.1 m3 per scheduled machine hour, respectively for the crawler and the skidder. Depending on extraction distance, skidding cost ranged from 19 to 47 V m3 for the crawler, and from 13 to 23 V m3 for the skidder. A model was developed to estimate machine productivity, extraction cost, and energy use for both machines (conventional and innovative) under a variety of work conditions typical of non-industrial private forestry. This information can be used for operational costing, planning and optimization. It can also serve to measure energy savings obtained through machine replacement. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Skidding Productivity Energy Cost Comparison
1. Introduction Non-industrial private forestry (NIPF) accounts for two-thirds of the Italian forest area (INFC, 2008). NIPF holdings are often too small for cost-effective industrial harvesting (Kittredge et al., 1996), whose overall cost-efficiency is heavily dependent on intense annual use of machinery and adequate tract size (Moldenhauer and Bolding, 2009). That explains the relatively small size of logging firms and the large popularity of intermediate harvesting technology (Rickenbach and Steele, 2006). This is based on versatile, low-investment machinery, which is more appropriate for smallscale users due to its lower capital cost (Dubey, 2008). Despite the recent advances of dedicated harvesting technology, forestry-fitted farm tractors are still the backbone of the Italian logging fleet (Spinelli and Magagnotti, 2011a). Farm tractors are used for a variety of forest harvesting tasks, especially extraction snjar et al., 2008). Conversion (Picchio et al., 2009) and transport (Su to extraction work requires appropriate implements, such as winches, trailers and loaders (Spinelli and Magagnotti, 2011b). In this respect, rubber-tired farm tractors are limited by poor off-road
* Corresponding author. Tel.: þ39 (0)55 5225641; fax: þ39 (0)55 5225643. E-mail addresses: [email protected] (R. Spinelli), [email protected] (N. Magagnotti), [email protected] (R. Laina Relaño). 0959-6526/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.jclepro.2012.02.033
mobility and are only used on very favorable terrain (Magagnotti and Spinelli, 2011a). Hence many loggers use one or more crawler tractors to address less favorable terrain conditions. An unpublished survey conducted in the Italian Alps found that 10% of the farm tractors used in forestry are tracked (92 tractors out of 888). Compared to conventional farm tractors, crawlers offer better maneuverability and superior traction capacity (Vaughan, 1988). For these reasons they are often used for bunching (De Lasaux et al., 2009) and skidding (Hill, 1991). Unfortunately, crawlers generate a greater site impact than other extraction systems commonly used in Italy. A recent study shows that damage to residual trees is twice as high and severe soil disturbance four times as high as the average levels, when crawler tractors are used (Spinelli et al., 2010a). Hence, there is keen interest in finding a rubber-tired replacement that can offer the same operational advantages without the heavy environmental impact. Such a replacement should have similar size, maneuverability, purchase price and operating cost as the small crawlers tractors currently in use. A low-investment cost is crucial to selection, given the smallscale of most logging firms in the region. Crawlers could be replaced with conventional rubber-tired skidders, but the acquisition of these machines would require a much larger initial investment and a proportionally higher utilization rate. On the other hand the availability of the mini-skidder may fulfill all these needs and qualify as a suitable alternative (McElroy, 2006).
74
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
The goal of this study was to compare the technical, economic and energy performance of a conventional crawler and an innovative rubber-tired mini-skidder, used under the same conditions, assumed as typical of small-scale Alpine logging. The hypothesis was that the mini-skidder offered equal or superior productivity, and equal or lower extraction cost, compared to the crawler. If not, very few loggers would consider substituting the mini-skidder for their crawlers. 2. Materials The crawler tractor selected for the test was a FIAT 55-85, powered by a 40 kW diesel engine and weighing 3200 kg (Fig. 1). This model is built by the Italian manufacturer specifically for hillside agriculture, but is often converted into a crawler skidder by small-scale loggers. The machine was fitted with a roll-bar, a nose guard and a mechanical forestry winch bolted to the rear end. The winch had a 60 kN maximum pulling force and its single drum contained 50 m of steel cable with a diameter of 10 mm. The machine was rented from a local logger. The purchase price for the complete new machine was estimated to 38,000 V by combining the listed prices of the tractor and the winch (the latter including installation). The mini-skidder was a Hittner Ecotrac V55, powered by a 40 kW engine and weighing 3600 kg (Fig. 2). This machine is built in Croatia specifically for small-scale logging and its geometry closely resembles that of a conventional skidder (Horvat et al., 2007). The machine on test was equipped with a front blade, a rear anchor plate and a hydraulic double-drum winch installed over the rear axle. Each drum had 35 kN maximum pulling force and contained 40 m of steel cable with a diameter of 10 mm. The operator sat inside a well-protected cab, with a heating and ventilation system. The machine was rented directly from the manufacturer, which quoted a purchase price for the new machine of 54,000 V. Two teams were chosen for the test. Each team consisted of two operators e driver and assistant (chokerman). These tasks were fixed, such that drivers and assistants could not switch roles. The two teams rotated between the two machines, alternating every ca. 3 h. This way each team would work as much with the crawler as with the skidder. The teams were selected according to the following specifications: 1) all operators were experienced loggers, proficient with operating tractors and winches; 2) no one had ever operated a crawler or a skidder before. This was done in order to
Fig. 2. The mini-skidder with its built-in double-drum winch.
prevent different operator experience from skewing the comparison (Gullberg, 1995). The tests lasted 6 days, but the first day was used for instruction and adaptation. Data collected during this day were excluded from the analysis and only used to check if the adaptation period was long enough to reach a work routine that was steady state. Both machines were tested on the same site, in the Regional forest of Pramosio (Table 1). The stand was a typical sprucedominated Alpine forest under continuous-cover forestry (CCF) management (Larsen, 1995). The treatment was individual-tree selection thinning, designed to remove about one-third of the total standing mass. Harvest trees were marked by the forester beforehand, who selected mature and younger low-quality trees. Specific attention was given to maintaining biodiversity during tree selection, which is a common practice in the whole region (Primmer, 2011). About 150 trees were felled and processed motormanually, with chainsaws. Stems were delimbed and cross-cut in 2.3, 4.2 and 5.2 m lengths, depending on small end diameter and stem quality. There was one main prepared skid trail within the stand which led to a large landing. The average and maximum slope gradients of the prepared skid trail were 10% and 29%, respectively. The average and maximum slope gradients of the paths on the
Table 1 Description of the test site. Municipality Province Longitude Latitude Altitude Slope gradient Skid trail gradient Road density Species Management Current treatment Age Removal Removal Residual density Avg. tree DBH Avg. stem height Avg. stem volume Fig. 1. The forestry-fitted crawler tractor with its single-drum winch.
m a.s.l. % % m ha1
years m3 ha1 trees ha1 trees ha1 m m m3
Cleulis e Paluzza Udine 46 570 22.2600 N 13 010 91.2100 E 1390 42 10 100 Picea abies (90%) e Abies alba (10%) High forest Thinning 100 111 100 203 37 25 1.115
Notes: a.s.l. ¼ above sea level; m3 are measured outside bark; Avg. ¼ Average.
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
observed values was tested against normality in order to select the appropriate analytical procedures. The significance of differences between treatments was determined with the t-test or the ManneWhitney test, depending on compliance with the normality assumptions. Tractor rates were calculated with the method described by Miyata (1980). It was assumed that each machine was available for 500 scheduled hours and was depreciated over a 15-year period. These values were meant to reflect the usage pattern of small-scale forest operations (Spinelli et al., 2010b). Fuel consumption was measured by starting with a full tank and refilling it at the end of the work day. The costs of insurance, repair and service were obtained directly from the owners of the tractors. Labor cost was set to 16 V SMH1 inclusive of indirect salary costs. The calculated operational cost of all teams was increased by 20% to account for overhead costs (Hartsough, 2003). Both direct and indirect fossil energy uses were estimated, reflecting the same principles followed by Pellizzi (1992) in his energy analysis of Italian agriculture. The direct energy used by the tractor was estimated by multiplying the measured diesel consumption by an energy content of 37 MJ L1 (Bailey et al., 2003), and then inflating this value by 1.2 in order to account for the additional fossil energy used in the production and transportation of diesel fuel (Pellizzi, 1992). The indirect energy consumption represented by tractor manufacturing, repair and maintenance was estimated as 30% of the total energy use of the tractor (Mikkola and Ahokas, 2010). Further detail on financial and energy cost calculations is shown in Table 2. Finally, the impact of annual use on skidding cost was gauged by changing the figures for annual use from 100 to 1600 h per year, and assuming the average productivity figures obtained from the study. Maximum service life was assumed to be 12,000 h or 15 years, whichever was reached first. At the annual use of 100 h per year, the machine was depreciated on 15 years and 1500 h. If the annual use was assumed to be 1600 h, then the machine was depreciated on 12,000 h and 7.5 years. The break-even annual use for replacing the crawler tractor and the minimum annual use required to make it economically efficient to invest in the miniskidder were determined. The time study consisted of 131 tractor turns (75 for the skidder and 56 for the crawler), extracting 145 m3. Overall, 45 h of productive time-study data was collected.
Table 2 Costing and energy use: assumptions, cost items and totals. Factor
Unit
Crawler
Skidder
Power Initial investment Resale value Service life Annual use Interest rate Fuel consumption Depreciation Interest Insurance Diesel Lube R&M Labor Subtotal Overhead Total (subtotal þ overhead) Direct energy Indirect energy Total energy
kW Euro Euro years h year1 % dm3 h1 Euro year1 Euro year1 Euro year1 Euro h1 Euro h1 Euro h1 Euro h1 Euro h1 Euro h1 Euro h1 MJ h1 MJ h1 MJ h1
40 38,000 7600 15 500 4 2.1 2027 953 953 2.9 0.9 3.2 32.0 46.9 9.4 56.2 91.9 40.4 132.3
40 54,000 10,800 15 500 4 1.3 2880 1354 1354 1.8 0.5 4.6 32.0 50.1 10.0 60.2 57.7 25.4 83.1
75
Cost in Euro (V) as of November 18, 2011. 1 V ¼ 1.37 US$.
forest floor were 23 and 33%, respectively. The loaded tractors traveled downhill on the prepared skid trail and uphill on the forest floor. The tractors rotated between loading sites, so that both machines would be tested over the same range of extraction distances and with the same distribution of short, medium and long trips. 3. Methods A time and motion study was carried out to evaluate machine productivity and to identify those variables that were most likely to affect it (Bergstrand, 1991). Each skidding cycle was timed individually using a stopwatch, separating productive time from delay time (Björheden et al., 1995). Delays caused by the study were removed from the analyzed data. Extraction distances were determined with a hip chain. No correction was made for slope gradient, so that distances represented the actual paths covered by the tractors. The volume of all logs in each load was determined by measuring their length and their diameter at mid-length, outside bark. Regression analysis of the time-study data was conducted to develop a model to predict skidding cycle time as a function of statistically significant independent variables (SAS Institute Inc., 1999). Indicator variables were used to represent different machines and teams (Olsen et al., 1998). The distribution of
4. Results The skidder was significantly faster than the crawler; its speed was between 50% and 85% faster (Table 3). The loads dragged by the
Table 3 Skidding conditions and performance with statistical evaluation. Factor
Distance on trail Distance in forest Winching distance Travel speed unloaded on trail Travel speed unloaded in forest Travel speed loaded on trail Travel speed loaded in forest Load size Load size Net cycle time Total cycle time Net productivity Gross productivity
Unit
m m m km h1 km h1 km h1 km h1 n pieces m3 min min m3 h1 m3 h1
Crawler
Skidder
Test of difference
Mean
SD
Mean
SD
p
Test
91.1 29.2 10.9 2.8 1.3 2.0 0.9 4.2 0.928 16.3 19.3 3.6 3.2
124.5 21.6 5.2 0.9 0.5 1.2 0.4 0.9 0.451 5.6 9.7 1.6 1.6
105.7 32.7 16.0 4.2 2.2 3.7 1.5 5.6 1.243 15.9 19.3 4.7 4.1
130.6 24.5 6.9 1.6 1.5 1.7 0.8 1.8 0.512 4.7 9.2 1.7 1.6
0.5063 0.4769 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0004 0.7569 0.9366 0.0009 0.0027
t-test MW t-test MW MW MW MW t-test t-test t-test MW t-test t-test
Notes: Net cycle and productivity exclude preparation and delay time, whereas Gross cycle and productivity include both preparation and delay time; SD ¼ Standard Deviation; Test ¼ test type, where t-test is the standard two-tailed unpaired t-test, and MW is the non-parametric ManneWhitney test.
76
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
Table 4 Comparison of the performance of the two test teams. Factor
Unit
Team A
Team B
Difference
Mean
SD
Mean
SD
p
Test
Crawler
Travel speed unloaded on trail Travel speed unloaded in forest Travel speed loaded on trail Travel speed loaded in forest Load size Net productivity Gross productivity
km h1 km h1 km h1 km h1 m3 m3 h1 m3 h1
2.7 1.4 2.2 1.0 0.780 3.498 3.050
1.1 0.5 1.5 0.3 0.291 1.515 1.315
2.8 1.0 1.9 0.6 1.217 3.770 3.475
0.4 0.4 0.7 0.3 0.566 1.977 2.054
0.5388 0.0075 0.9514 0.0039 0.0021 0.6098 0.3390
MW MW MW MW MW t-test t-test
Skidder
Travel speed unloaded on trail Travel speed unloaded in forest Travel speed loaded on trail Travel speed loaded in forest Load size Net productivity Gross productivity
km h1 km h1 km h1 km h1 m3 m3 h1 m3 h1
4.1 2.0 3.9 1.5 1.024 4.314 3.672
1.7 1.2 1.7 0.5 0.334 1.801 1.607
4.3 2.5 3.6 1.6 1.492 5.169 4.732
1.6 1.9 1.7 1.2 0.568 1.445 1.525
0.5512 0.6589 0.6713 0.5761 <0.0001 0.0484 0.0108
t-test MW MW MW t-test t-test t-test
Notes: SD ¼ Standard Deviation; Test ¼ test type, where t-test is the standard two-tailed unpaired t-test, and MW is the non-parametric ManneWhitney test.
skidder were one-third larger than those dragged by the crawler, and this difference also proved significant to the statistical tests. As a result of its higher speed and larger payload, the skidder offered a significantly higher productivity compared to the crawler. The two test teams had different work styles. In particular, Team B assembled significantly larger loads, regardless of machine type (Table 4). The difference was substantial, and in the range of 50%. Team B also drove the crawler at lower speeds on the forest floor. This difference was not found once the crawler was on the prepared skid trail, or with the skidder. Hence, Team B may have felt less confident about the stability of the crawler, or it may have assembled excessive loads, which limited the speed of the crawler. Team B achieved a higher productivity than Team A, both with the crawler and with the skidder. This difference was higher (20e28%) and statistically significant for the skidder. It was lower (8e14%) and without statistical significance for the crawler, probably due to the compensating effect of the higher travel speed achieved by Team A. Table 5 reports the regression equations used for predicting the time taken by each individual work task. These equations may help in understanding process dynamics, as well as the potential impact of work conditions on task duration. All regressions are highly significant and can explain a large proportion of the overall
variability, except for the equation predicting loading time, which only accounts for about one-third of the variation in the original data. However, all other equations explain between 60% and 95% of the variability. All equations contain the same independent variables: skidding distance, winching distance, load size, machine type (crawler or skidder) and team. The equations in Table 5 were used to create the graphs representing gross productivity as a function of skidding distance for the two machines and the two teams (Fig. 3). The skidder is substantially more productive than the crawler, and the difference increases with distance, due to its higher speed and larger payload. The productivity for Team B is between 7 and 25% higher than for Team A. The higher travel speeds achieved by Team A with the crawler cannot compensate for the larger loads assembled by Team B, but they reduce the difference between the two teams as distance increases. The graphs in Fig. 3 were used to calculate skidding cost, both in financial and energy terms (Table 6). Replacing the crawler with the mini-skidder reduced skidding cost between 30 and 50%, depending on extraction distance. Due to the larger payload and higher speed, savings increase with distance. Energy savings are even larger because of the compounded effect of higher
Table 5 Regressions models for estimating time consumption and productivity. Gross skidding productivity in m3 h1 ¼ Load size * 6000/Total time, Where: Total time (min * 102) ¼ Travel empty on prepared skid trail þ Travel empty in forest þ Loading þ Travel loaded in forest þ Travel loaded on prepared skid trail þ Unloading þ Delays Travel empty on trail (min * 102)
r2 0.950 F ¼ 788 p < 0.0001 19.917 þ 1.158 Distance T þ 0.813 Crawler * Distance T 0.540 Team A * Distance T
N ¼ 126
Travel empty in forest (min * 102)
r2 0.597 F ¼ 64 p < 0.0001 38.431 þ 1.994 Distance F þ 2.982 Crawler * Distance F 1.889 Team A * Crawler * Distance F
N ¼ 129
Loading (min * 102)
r2 0.322 F ¼ 21 p < 0.0001 306.764 þ 17.708 Distance W þ 79.869 m3 þ 54.316 m3 * Crawler
N ¼ 130
Travel loaded in forest (min * 102)
r2 0.719 F ¼ 106 p < 0.0001 N ¼ 125 29.517 þ 7.337 Crawler * Distance F 3.868 Team A * Crawler * Distance F þ 2.974 m3 * Distance F
Travel loaded on trail (min * 102)
r2 0.919 F ¼ 359 p < 0.0001 26.497 0.895 Distance T þ 2.422 Crawler * Distance T 0.633 Team A * Crawler * Distance T þ 1.339 m3 * Distance T
Unloading (min * 102)
p ¼ 0.0761
N ¼ 127
N ¼ 129
349.304 if Crawler, 396.720 if Skidder Accessory time and delays (min * 102)
0.256 * (Travelling þ Loading þ Unloading)
Distance T ¼ one-way travel distance on the prepared skid trail in m; Distance F ¼ one-way travel distance on the forest floor in m; Distance W ¼ winching distance in m; Crawler ¼ indicator variable: 1 if crawler, 0 if skidder; Team A ¼ indicator variable: 1 if Team A, 0 if Team B; Load ¼ Load size in m3.
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
Fig. 3. Gross skidding productivity as a function of distance on the prepared skid trail, machine type and team. Note: the curves in the graph were calculated for a skidding distance on the forest floor ¼ 35 m, a winching distance ¼ 15 m, an average load size ¼ 1.024 and 1.490 m3 for the skidder manned by Teams A and B respectively, and 0.780 and1.217 m3 for the crawler manned by Teams A and B (mean values from the study). Productivity figures are calculated on total time consumption, and include preparation and delay time. Volume figures are outside bark.
productivity and lower fuel consumption. The higher productivity of Team B is also reflected in a lower financial and energy cost. Cost reduction averages 15%, but ranges from 7% to 24% and decreases with distance due to the compensating effect of the higher travel speed achieved by Team A with the crawler. Fig. 4 shows the relationship between skidding cost and annual use. The skidder always offers a lower extraction cost than the crawler, but the margin increases with annual use. However, the crawler is an agricultural tractor and can be used for more tasks than just skidding. This may allow for more use year-round, especially when logging opportunities are limited. Hence a comparison was done by assuming different utilization levels for the two machines. Table 7 shows the annual use of the skidder and the corresponding additional number of hours per year the crawler should work on other jobs, if it was to achieve the same unit skidding cost. If the skidder works at least 500 h per year, it will
Table 6 Skidding cost as a function of distance on trail, machine type and team. Dist. (m) Crawler Team A Crawler Team B Skidder Team A Skidder Team B Skidding cost, Euro m3 50 24.5 100 27.0 150 29.5 200 32.1 250 34.6 300 37.2 350 39.7 400 42.3 450 44.8 500 47.3
19.5 22.0 24.4 26.9 29.4 31.9 34.4 36.8 39.3 41.8
17.2 17.9 18.5 19.2 19.9 20.5 21.2 21.9 22.6 23.2
13.0 14.0 14.9 15.9 16.8 17.8 18.7 19.7 20.6 21.6
Skidding cost, MJ m3 50 57.5 100 63.5 150 69.5 200 75.5 250 81.5 300 87.5 350 93.4 400 99.4 450 105.4 500 111.4
45.8 51.7 57.5 63.3 69.2 75.0 80.9 86.7 92.5 98.4
23.7 24.7 25.6 26.5 27.4 28.4 29.3 30.2 31.2 32.1
18.0 19.3 20.6 21.9 23.2 24.6 25.9 27.2 28.5 29.8
Dist. ¼ skidding distance on trail, in m; the assumed skidding distance on the forest floor is 35 m, and winching distance 15 m, Volume figures are outside bark.
77
Fig. 4. Skidding cost as a function of annual use.
offer a cheaper service than the crawler, regardless of how much additional work the crawler obtains.
5. Discussion The hypothesis of the study is confirmed; the new mini-skidder offers superior productivity and lower extraction cost compared to the crawler. That is the necessary economic condition for crawler replacement, which is already desirable in terms of site impact and labor safety. Most of the forestry-fitted crawler tractors inspected during the above-mentioned survey do not comply with current safety regulations (i.e. EU Directive 2008/50/CE). Upgrading to current safety standards may incur substantial cost, thus making replacement a more desirable option. In fact the Croatian Forest Administration decommissioned its substantial crawler fleet in the mid 1990s and replaced it with new skidders (Beuk et al., 2007). The superior payload capacity of the skidder depends on its unique structural design, which allows for a better lifting of the log ends and an easy transfer of the vertical load component onto the rear axle (Stoilov and Kostadinov, 2009). That increases the tractive capacity of the skidder, especially when negotiating uphill grades (Tomasi c et al., 2009). The structure of the crawler does not allow for the same lift, so only a small part of the total load weight contributes to increasing the crawler’s tractive capacity. The study also shows the importance of the human factor, and sheds light on the most efficient skidding techniques. The productivity difference between teams averaged 15%, for the same machine and work conditions. The difference was smaller than reported for complex tree felling and processing machines (Ovaskainen and Heikkilä, 2007), but the impact of human adaptation is likely to increase with task complexity (Kärhä et al., 2005). The study indicates that load maximization increases productivity and decreases cost, and that a double-drum winch certainly helps assembling larger loads.
Table 7 Annual use of the skidder and additional annual hours the crawler should work in order to achieve the same skidding cost. Hours skidder
Additional hours crawler
Skidding cost (V m3)
100 200 300 400 500
5 50 150 300 N
33 22 18 16 15
Note: volume figures are outside bark.
78
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
In fact, readers should note that the study was conducted with as much rigor as possible, and yet it occasionally violated the statistical assumption of a strict experimental design. Therefore, the regressions should not be used to determine the exact weight of each independent variable, but only to establish its general significance. The productivity levels achieved in this study equal or exceed the productivity levels reported in previous studies for similar machines. Magagnotti and Spinelli (2011b) report a productivity of 2.3 m3 h1 and an average load of 0.8 m3 for a forestry-fitted crawler tractor as used in this study. Ze ci c and Maren ce (2005) indicate a productivity of 2.4 m3 h1 and an average load of 1.1 m3 for the same type of mini-skidder tested in this study, operating over the same distance. However, stand type and the silvicultural prescriptions were different, which may explain the different productivity and load size. In any case, results are comparable, which may support the use of our productivity figures as a general reference. The energy cost ranged from 18 to 32 MJ m3 for the miniskidder and from 45 to 111 MJ m3 for the crawler. These figures are comparable with those reported by Picchio et al. (2009) (36 MJ m3) and by Magagnotti and Spinelli (2011a) (60e120 MJ m3). They are relatively low, which may be expected for intermediate mechanization (Berg, 1997), especially under Alpine conditions (Valente et al., 2011). The mini-skidder is equipped with a modern Tier III engine, which is likely to dramatically reduce emissions (Berg and Lindholm, 2005). 6. Conclusions Rubber-tired mini-skidders can effectively replace forestryfitted crawler tractors under certain conditions as outlined in this study. Replacement is desirable in terms of environmental protection and labor safety, and offers substantial economic benefit; it requires a moderate additional initial investment but allows reducing extraction cost between 30 and 50%. As compared to the crawler tractor, the mini-skidder is more environmentally friendly in terms of its fossil fuel use. Most crawlers used in forestry are now quite old. With the intent of promoting rural development, the European Union offers grants for the acquisition of new machinery, and low-impact technologies are generally favored over conventional, less-efficient solutions. Regional managers should be apprised of these new environmentally compatible products when developing grant programmes. The manufacturer may also consider adapting and registering his mini-skidder for road travel according to the Italian Law. That would enable independent, low-cost translocation between work sites. Road trafficability would give the new machine further advantage over the crawler, which cannot travel on public roads and is moved between work sites on a trailer. The study demonstrates that one can effectively control most sources of unwanted variability even when conducting the comparative test in a real work environment, such as a forest. The same methodology could also be applied to other machines or operations. Acknowledgments This study is a part of the project “Organization and rationalization of logging operations” funded by Regione Autonoma Friuli-Venezia Giulia, Servizio Gestione Forestale e Produzione Legnosa. This study was also made possible thanks to funding received from the STSM programme of Action COST FP902. Thanks are also due to Dr. Mario Di Gallo and the CESFAM team for their technical support.
References Bailey, A., Basford, W., Penlington, N., Park, J., Keatinge, J., Rehman, T., Tranter, R., Yates, C., 2003. A comparison of energy use in conventional and integrated arable farming in the UK. Agricultural Ecosystems and Environments 97, 241e253. Berg, S., 1997. Some aspects of LCA in the analysis of forestry operations. Journal of Cleaner Production 5, 211e217. Berg, S., Lindholm, E., 2005. Energy use and environmental impacts of forest operations in Sweden. Journal of Cleaner Production 13, 33e42. Bergstrand, K.G., 1991. Planning and analysis of forestry operation studies. Skogsarbeten Bulletin 17, 63 pp. Beuk, D., Tomasi c, Z., Horvat, D., 2007. Status and development of forest harvesting mechanization in Croatian State forestry. Croatian Journal of Forest Engineering 28, 63e82. Björheden, R., Apel, K., Shiba, M., Thompson, M.A., 1995. IUFRO Forest Work Study Nomenclature. Swedish University of Agricultural Science, Dept. of Operational Efficiency, Garpenberg, 16 pp. De Lasaux, M., Hartsough, B., Spinelli, R., Magagnotti, N., 2009. Small parcel fuel reduction with a low-investment, high-mobility operation. Western Journal of Applied Forestry 24, 205e213. Dubey, P., 2008. Investment in small-scale forestry enterprises: a strategic perspective for India. Small-scale Forestry 7, 117e138. Gullberg, T., 1995. Evaluating operatoremachine interactions in comparative time studies. Journal of Forest Engineering 7, 51e61. Hartsough, B., 2003. Economics of harvesting to maintain high structural diversity and resulting damage to residual trees. Western Journal of Applied Forestry 18, 133e142. Hill, S., 1991. D4H tractor and towed arch in radiata clearfell. LIRA Report 16, 8. 4 pp. snjar, M., 2007. Morphological characteristics and producHorvat, D., Ze ci c, Z., Su tivity of skidder Ecotrac 120V. Croatian Journal of Forest Engineering 28, 11e25. INFC, 2008. National Forest Inventory e Tab. 3.1 Surface by Ownership Type. www. infc.it (accessed 04.11.11). Kärhä, K., Jouhiaho, A., Mutikainen, A., Mattila, S., 2005. Mechanized energy wood harvesting from early thinnings. International Journal of Forest Engineering 16, 23e36. Kittredge, D., Mauri, M., McGuire, E., 1996. Decreasing woodlot size and the future of timber sales in Massachusetts: when is an operation too small? Northern Journal of Applied Forestry 13, 96e101. Larsen, J.B., 1995. Ecological stability of forests and sustainable silviculture. Forest Ecology and Management 73, 85e96. Magagnotti, N., Spinelli, R., 2011a. Financial and energy cost of low-impact wood extraction in environmentally sensitive areas. Ecological Engineering 37, 601e606. Magagnotti, N., Spinelli, R., 2011b. Integrating animal and mechanical operations in protected areas. Croatian Journal of Forest Engineering 32, 489e499. McElroy, N., 2006. Parcelization of the Chesapeake Bay watershed and implications for sustainable forestry. Master thesis, Virginia Polytechnic Institute and State University, Falls Church, VA. 57 pp. Mikkola, H., Ahokas, J., 2010. Indirect energy input of agricultural machinery in bioenergy production. Renewable Energy 35, 23e28. Miyata, E., 1980. Determining Fixed and Operating Costs of Logging Equipment. General Technical Report NC-55. Forest Service North Central Forest Experiment Station, St. Paul, MN, 14 pp. Moldenhauer, M., Bolding, M., 2009. Parcelization of South Carolina’s private forestland: loggers’ reactions to a growing threat. Forest Products Journal 59, 37e43. Olsen, E., Hossain, M., Miller, M., 1998. Statistical Comparison of Methods Used in Harvesting Work Studies. Oregon State University, Forest Research Laboratory, Corvallis, OR, Research Contribution n 23, 31 pp. Ovaskainen, H., Heikkilä, M., 2007. Visuospatial cognitive abilities in cut-to-length single-grip harvester work. International Journal of Industrial Ergonomics 37, 771e780. Pellizzi, G., 1992. Use of energy and labour in Italian agriculture. Journal of Agricultural Engineering Research 52, 111e119. Picchio, R., Maesano, M., Savelli, S., Marchi, E., 2009. Productivity and energy balance in conversion of a Quercus cerris L. coppice stand into high forest in Central Italy. Croatian Journal of Forest Engineering 30, 15e26. Primmer, E., 2011. Analysis of institutional adaptation: integration of biodiversity conservation into forestry. Journal of Cleaner Production 19, 1822e1832. Rickenbach, M., Steele, T., 2006. Logging firms, nonindustrial private forests, and forest parcelization: evidence of firm specialization and its impact on sustainable timber supply. Canadian Journal of Forest Research 36, 186e194. SAS Institute Inc., 1999. StatView Reference. SAS Publishing, Cary, NC, ISBN 158025-162-5, pp. 84e93. Spinelli, R., Magagnotti, N., Nati, C., 2010a. Benchmarking the impact of traditional small-scale logging systems used in Mediterranean forestry. Forest Ecology and Management 260, 1997e2001. Spinelli, R., Magagnotti, N., Picchi, G., 2010b. Deploying mechanized cut-to-length technology in Italy: fleet size, annual usage, and costs. International Journal of Forest Engineering 21, 23e31. Spinelli, R., Magagnotti, N., 2011a. The effects of introducing modern technology on the financial, labour and energy performance of forest operations in the Italian Alps. Forest Policy and Economics 13, 520e534.
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79 Spinelli, R., Magagnotti, N., 2011b. Wood extraction with farm tractor and sulky: estimating productivity, cost and energy consumption. Small-scale Forestry. doi:10.1007/s11842-011-9169-8. Stoilov, S., Kostadinov, G., 2009. Effect of weight distribution on slip efficiency of a four-wheel-drive skidder. Biosystems Engineering 104, 486e492. snjar, M., Horvat, D., Kristi Su c, A., Pandur, Z., 2008. Morphological analysis of forest tractor assemblies. Croatian Journal of Forest Engineering 29, 41e51.
79
snjar, M., Horvat, D., Pandur, Z., 2009. Factors affecting timber Tomasi c, Z., Su skidding. Croatian Journal of Forest Engineering 30, 127e139. Valente, C., Spinelli, R., Hillring, B., 2011. LCA of environmental and socio-economic impacts related to wood energy production in alpine conditions: Valle di Fiemme (Italy). Journal of Cleaner Production 19, 1931e1938. Vaughan, L., 1988. Thinning with small crawler tractors. LIRA Report 13. 6 pp. Ze ci c, Z., Maren ce, J., 2005. Mathematical models for optimization of group work in harvesting operation. Croatian Journal of Forest Engineering 26, 29e37.
Contents lists available at SciVerse ScienceDirect
Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro
An alternative skidding technology to the current use of crawler tractors in Alpine logging operations Raffaele Spinelli a, *, Natascia Magagnotti b, Rubén Laina Relaño c a
CNR e IVALSA, Via Madonna del Piano, Pal. F, I-50019 Sesto Fiorentino (FI), Italy CNR e IVALSA, Via Biasi 75, I-38010 San Michele all’Adige (TN), Italy c Universidad Politecnica de Madrid, Ciudad Universitaria s/n., E-28040 Madrid, Spain b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 26 November 2011 Received in revised form 26 February 2012 Accepted 27 February 2012 Available online 9 March 2012
The study investigated the possibility of replacing aggressive crawler tractors used in Alpine logging operations with a new rubber-tired mini-skidder, in the same size and cost class. Replacement proved desirable in terms of environmental protection and labor safety, and offered substantial economic benefits. While it required a moderate additional initial investment, extraction cost was reduced between 30 and 50%. On the average skidding distance of 130 m, productivity was 3.2 and 4.1 m3 per scheduled machine hour, respectively for the crawler and the skidder. Depending on extraction distance, skidding cost ranged from 19 to 47 V m3 for the crawler, and from 13 to 23 V m3 for the skidder. A model was developed to estimate machine productivity, extraction cost, and energy use for both machines (conventional and innovative) under a variety of work conditions typical of non-industrial private forestry. This information can be used for operational costing, planning and optimization. It can also serve to measure energy savings obtained through machine replacement. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Skidding Productivity Energy Cost Comparison
1. Introduction Non-industrial private forestry (NIPF) accounts for two-thirds of the Italian forest area (INFC, 2008). NIPF holdings are often too small for cost-effective industrial harvesting (Kittredge et al., 1996), whose overall cost-efficiency is heavily dependent on intense annual use of machinery and adequate tract size (Moldenhauer and Bolding, 2009). That explains the relatively small size of logging firms and the large popularity of intermediate harvesting technology (Rickenbach and Steele, 2006). This is based on versatile, low-investment machinery, which is more appropriate for smallscale users due to its lower capital cost (Dubey, 2008). Despite the recent advances of dedicated harvesting technology, forestry-fitted farm tractors are still the backbone of the Italian logging fleet (Spinelli and Magagnotti, 2011a). Farm tractors are used for a variety of forest harvesting tasks, especially extraction snjar et al., 2008). Conversion (Picchio et al., 2009) and transport (Su to extraction work requires appropriate implements, such as winches, trailers and loaders (Spinelli and Magagnotti, 2011b). In this respect, rubber-tired farm tractors are limited by poor off-road
* Corresponding author. Tel.: þ39 (0)55 5225641; fax: þ39 (0)55 5225643. E-mail addresses: [email protected] (R. Spinelli), [email protected] (N. Magagnotti), [email protected] (R. Laina Relaño). 0959-6526/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.jclepro.2012.02.033
mobility and are only used on very favorable terrain (Magagnotti and Spinelli, 2011a). Hence many loggers use one or more crawler tractors to address less favorable terrain conditions. An unpublished survey conducted in the Italian Alps found that 10% of the farm tractors used in forestry are tracked (92 tractors out of 888). Compared to conventional farm tractors, crawlers offer better maneuverability and superior traction capacity (Vaughan, 1988). For these reasons they are often used for bunching (De Lasaux et al., 2009) and skidding (Hill, 1991). Unfortunately, crawlers generate a greater site impact than other extraction systems commonly used in Italy. A recent study shows that damage to residual trees is twice as high and severe soil disturbance four times as high as the average levels, when crawler tractors are used (Spinelli et al., 2010a). Hence, there is keen interest in finding a rubber-tired replacement that can offer the same operational advantages without the heavy environmental impact. Such a replacement should have similar size, maneuverability, purchase price and operating cost as the small crawlers tractors currently in use. A low-investment cost is crucial to selection, given the smallscale of most logging firms in the region. Crawlers could be replaced with conventional rubber-tired skidders, but the acquisition of these machines would require a much larger initial investment and a proportionally higher utilization rate. On the other hand the availability of the mini-skidder may fulfill all these needs and qualify as a suitable alternative (McElroy, 2006).
74
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
The goal of this study was to compare the technical, economic and energy performance of a conventional crawler and an innovative rubber-tired mini-skidder, used under the same conditions, assumed as typical of small-scale Alpine logging. The hypothesis was that the mini-skidder offered equal or superior productivity, and equal or lower extraction cost, compared to the crawler. If not, very few loggers would consider substituting the mini-skidder for their crawlers. 2. Materials The crawler tractor selected for the test was a FIAT 55-85, powered by a 40 kW diesel engine and weighing 3200 kg (Fig. 1). This model is built by the Italian manufacturer specifically for hillside agriculture, but is often converted into a crawler skidder by small-scale loggers. The machine was fitted with a roll-bar, a nose guard and a mechanical forestry winch bolted to the rear end. The winch had a 60 kN maximum pulling force and its single drum contained 50 m of steel cable with a diameter of 10 mm. The machine was rented from a local logger. The purchase price for the complete new machine was estimated to 38,000 V by combining the listed prices of the tractor and the winch (the latter including installation). The mini-skidder was a Hittner Ecotrac V55, powered by a 40 kW engine and weighing 3600 kg (Fig. 2). This machine is built in Croatia specifically for small-scale logging and its geometry closely resembles that of a conventional skidder (Horvat et al., 2007). The machine on test was equipped with a front blade, a rear anchor plate and a hydraulic double-drum winch installed over the rear axle. Each drum had 35 kN maximum pulling force and contained 40 m of steel cable with a diameter of 10 mm. The operator sat inside a well-protected cab, with a heating and ventilation system. The machine was rented directly from the manufacturer, which quoted a purchase price for the new machine of 54,000 V. Two teams were chosen for the test. Each team consisted of two operators e driver and assistant (chokerman). These tasks were fixed, such that drivers and assistants could not switch roles. The two teams rotated between the two machines, alternating every ca. 3 h. This way each team would work as much with the crawler as with the skidder. The teams were selected according to the following specifications: 1) all operators were experienced loggers, proficient with operating tractors and winches; 2) no one had ever operated a crawler or a skidder before. This was done in order to
Fig. 2. The mini-skidder with its built-in double-drum winch.
prevent different operator experience from skewing the comparison (Gullberg, 1995). The tests lasted 6 days, but the first day was used for instruction and adaptation. Data collected during this day were excluded from the analysis and only used to check if the adaptation period was long enough to reach a work routine that was steady state. Both machines were tested on the same site, in the Regional forest of Pramosio (Table 1). The stand was a typical sprucedominated Alpine forest under continuous-cover forestry (CCF) management (Larsen, 1995). The treatment was individual-tree selection thinning, designed to remove about one-third of the total standing mass. Harvest trees were marked by the forester beforehand, who selected mature and younger low-quality trees. Specific attention was given to maintaining biodiversity during tree selection, which is a common practice in the whole region (Primmer, 2011). About 150 trees were felled and processed motormanually, with chainsaws. Stems were delimbed and cross-cut in 2.3, 4.2 and 5.2 m lengths, depending on small end diameter and stem quality. There was one main prepared skid trail within the stand which led to a large landing. The average and maximum slope gradients of the prepared skid trail were 10% and 29%, respectively. The average and maximum slope gradients of the paths on the
Table 1 Description of the test site. Municipality Province Longitude Latitude Altitude Slope gradient Skid trail gradient Road density Species Management Current treatment Age Removal Removal Residual density Avg. tree DBH Avg. stem height Avg. stem volume Fig. 1. The forestry-fitted crawler tractor with its single-drum winch.
m a.s.l. % % m ha1
years m3 ha1 trees ha1 trees ha1 m m m3
Cleulis e Paluzza Udine 46 570 22.2600 N 13 010 91.2100 E 1390 42 10 100 Picea abies (90%) e Abies alba (10%) High forest Thinning 100 111 100 203 37 25 1.115
Notes: a.s.l. ¼ above sea level; m3 are measured outside bark; Avg. ¼ Average.
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
observed values was tested against normality in order to select the appropriate analytical procedures. The significance of differences between treatments was determined with the t-test or the ManneWhitney test, depending on compliance with the normality assumptions. Tractor rates were calculated with the method described by Miyata (1980). It was assumed that each machine was available for 500 scheduled hours and was depreciated over a 15-year period. These values were meant to reflect the usage pattern of small-scale forest operations (Spinelli et al., 2010b). Fuel consumption was measured by starting with a full tank and refilling it at the end of the work day. The costs of insurance, repair and service were obtained directly from the owners of the tractors. Labor cost was set to 16 V SMH1 inclusive of indirect salary costs. The calculated operational cost of all teams was increased by 20% to account for overhead costs (Hartsough, 2003). Both direct and indirect fossil energy uses were estimated, reflecting the same principles followed by Pellizzi (1992) in his energy analysis of Italian agriculture. The direct energy used by the tractor was estimated by multiplying the measured diesel consumption by an energy content of 37 MJ L1 (Bailey et al., 2003), and then inflating this value by 1.2 in order to account for the additional fossil energy used in the production and transportation of diesel fuel (Pellizzi, 1992). The indirect energy consumption represented by tractor manufacturing, repair and maintenance was estimated as 30% of the total energy use of the tractor (Mikkola and Ahokas, 2010). Further detail on financial and energy cost calculations is shown in Table 2. Finally, the impact of annual use on skidding cost was gauged by changing the figures for annual use from 100 to 1600 h per year, and assuming the average productivity figures obtained from the study. Maximum service life was assumed to be 12,000 h or 15 years, whichever was reached first. At the annual use of 100 h per year, the machine was depreciated on 15 years and 1500 h. If the annual use was assumed to be 1600 h, then the machine was depreciated on 12,000 h and 7.5 years. The break-even annual use for replacing the crawler tractor and the minimum annual use required to make it economically efficient to invest in the miniskidder were determined. The time study consisted of 131 tractor turns (75 for the skidder and 56 for the crawler), extracting 145 m3. Overall, 45 h of productive time-study data was collected.
Table 2 Costing and energy use: assumptions, cost items and totals. Factor
Unit
Crawler
Skidder
Power Initial investment Resale value Service life Annual use Interest rate Fuel consumption Depreciation Interest Insurance Diesel Lube R&M Labor Subtotal Overhead Total (subtotal þ overhead) Direct energy Indirect energy Total energy
kW Euro Euro years h year1 % dm3 h1 Euro year1 Euro year1 Euro year1 Euro h1 Euro h1 Euro h1 Euro h1 Euro h1 Euro h1 Euro h1 MJ h1 MJ h1 MJ h1
40 38,000 7600 15 500 4 2.1 2027 953 953 2.9 0.9 3.2 32.0 46.9 9.4 56.2 91.9 40.4 132.3
40 54,000 10,800 15 500 4 1.3 2880 1354 1354 1.8 0.5 4.6 32.0 50.1 10.0 60.2 57.7 25.4 83.1
75
Cost in Euro (V) as of November 18, 2011. 1 V ¼ 1.37 US$.
forest floor were 23 and 33%, respectively. The loaded tractors traveled downhill on the prepared skid trail and uphill on the forest floor. The tractors rotated between loading sites, so that both machines would be tested over the same range of extraction distances and with the same distribution of short, medium and long trips. 3. Methods A time and motion study was carried out to evaluate machine productivity and to identify those variables that were most likely to affect it (Bergstrand, 1991). Each skidding cycle was timed individually using a stopwatch, separating productive time from delay time (Björheden et al., 1995). Delays caused by the study were removed from the analyzed data. Extraction distances were determined with a hip chain. No correction was made for slope gradient, so that distances represented the actual paths covered by the tractors. The volume of all logs in each load was determined by measuring their length and their diameter at mid-length, outside bark. Regression analysis of the time-study data was conducted to develop a model to predict skidding cycle time as a function of statistically significant independent variables (SAS Institute Inc., 1999). Indicator variables were used to represent different machines and teams (Olsen et al., 1998). The distribution of
4. Results The skidder was significantly faster than the crawler; its speed was between 50% and 85% faster (Table 3). The loads dragged by the
Table 3 Skidding conditions and performance with statistical evaluation. Factor
Distance on trail Distance in forest Winching distance Travel speed unloaded on trail Travel speed unloaded in forest Travel speed loaded on trail Travel speed loaded in forest Load size Load size Net cycle time Total cycle time Net productivity Gross productivity
Unit
m m m km h1 km h1 km h1 km h1 n pieces m3 min min m3 h1 m3 h1
Crawler
Skidder
Test of difference
Mean
SD
Mean
SD
p
Test
91.1 29.2 10.9 2.8 1.3 2.0 0.9 4.2 0.928 16.3 19.3 3.6 3.2
124.5 21.6 5.2 0.9 0.5 1.2 0.4 0.9 0.451 5.6 9.7 1.6 1.6
105.7 32.7 16.0 4.2 2.2 3.7 1.5 5.6 1.243 15.9 19.3 4.7 4.1
130.6 24.5 6.9 1.6 1.5 1.7 0.8 1.8 0.512 4.7 9.2 1.7 1.6
0.5063 0.4769 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0004 0.7569 0.9366 0.0009 0.0027
t-test MW t-test MW MW MW MW t-test t-test t-test MW t-test t-test
Notes: Net cycle and productivity exclude preparation and delay time, whereas Gross cycle and productivity include both preparation and delay time; SD ¼ Standard Deviation; Test ¼ test type, where t-test is the standard two-tailed unpaired t-test, and MW is the non-parametric ManneWhitney test.
76
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
Table 4 Comparison of the performance of the two test teams. Factor
Unit
Team A
Team B
Difference
Mean
SD
Mean
SD
p
Test
Crawler
Travel speed unloaded on trail Travel speed unloaded in forest Travel speed loaded on trail Travel speed loaded in forest Load size Net productivity Gross productivity
km h1 km h1 km h1 km h1 m3 m3 h1 m3 h1
2.7 1.4 2.2 1.0 0.780 3.498 3.050
1.1 0.5 1.5 0.3 0.291 1.515 1.315
2.8 1.0 1.9 0.6 1.217 3.770 3.475
0.4 0.4 0.7 0.3 0.566 1.977 2.054
0.5388 0.0075 0.9514 0.0039 0.0021 0.6098 0.3390
MW MW MW MW MW t-test t-test
Skidder
Travel speed unloaded on trail Travel speed unloaded in forest Travel speed loaded on trail Travel speed loaded in forest Load size Net productivity Gross productivity
km h1 km h1 km h1 km h1 m3 m3 h1 m3 h1
4.1 2.0 3.9 1.5 1.024 4.314 3.672
1.7 1.2 1.7 0.5 0.334 1.801 1.607
4.3 2.5 3.6 1.6 1.492 5.169 4.732
1.6 1.9 1.7 1.2 0.568 1.445 1.525
0.5512 0.6589 0.6713 0.5761 <0.0001 0.0484 0.0108
t-test MW MW MW t-test t-test t-test
Notes: SD ¼ Standard Deviation; Test ¼ test type, where t-test is the standard two-tailed unpaired t-test, and MW is the non-parametric ManneWhitney test.
skidder were one-third larger than those dragged by the crawler, and this difference also proved significant to the statistical tests. As a result of its higher speed and larger payload, the skidder offered a significantly higher productivity compared to the crawler. The two test teams had different work styles. In particular, Team B assembled significantly larger loads, regardless of machine type (Table 4). The difference was substantial, and in the range of 50%. Team B also drove the crawler at lower speeds on the forest floor. This difference was not found once the crawler was on the prepared skid trail, or with the skidder. Hence, Team B may have felt less confident about the stability of the crawler, or it may have assembled excessive loads, which limited the speed of the crawler. Team B achieved a higher productivity than Team A, both with the crawler and with the skidder. This difference was higher (20e28%) and statistically significant for the skidder. It was lower (8e14%) and without statistical significance for the crawler, probably due to the compensating effect of the higher travel speed achieved by Team A. Table 5 reports the regression equations used for predicting the time taken by each individual work task. These equations may help in understanding process dynamics, as well as the potential impact of work conditions on task duration. All regressions are highly significant and can explain a large proportion of the overall
variability, except for the equation predicting loading time, which only accounts for about one-third of the variation in the original data. However, all other equations explain between 60% and 95% of the variability. All equations contain the same independent variables: skidding distance, winching distance, load size, machine type (crawler or skidder) and team. The equations in Table 5 were used to create the graphs representing gross productivity as a function of skidding distance for the two machines and the two teams (Fig. 3). The skidder is substantially more productive than the crawler, and the difference increases with distance, due to its higher speed and larger payload. The productivity for Team B is between 7 and 25% higher than for Team A. The higher travel speeds achieved by Team A with the crawler cannot compensate for the larger loads assembled by Team B, but they reduce the difference between the two teams as distance increases. The graphs in Fig. 3 were used to calculate skidding cost, both in financial and energy terms (Table 6). Replacing the crawler with the mini-skidder reduced skidding cost between 30 and 50%, depending on extraction distance. Due to the larger payload and higher speed, savings increase with distance. Energy savings are even larger because of the compounded effect of higher
Table 5 Regressions models for estimating time consumption and productivity. Gross skidding productivity in m3 h1 ¼ Load size * 6000/Total time, Where: Total time (min * 102) ¼ Travel empty on prepared skid trail þ Travel empty in forest þ Loading þ Travel loaded in forest þ Travel loaded on prepared skid trail þ Unloading þ Delays Travel empty on trail (min * 102)
r2 0.950 F ¼ 788 p < 0.0001 19.917 þ 1.158 Distance T þ 0.813 Crawler * Distance T 0.540 Team A * Distance T
N ¼ 126
Travel empty in forest (min * 102)
r2 0.597 F ¼ 64 p < 0.0001 38.431 þ 1.994 Distance F þ 2.982 Crawler * Distance F 1.889 Team A * Crawler * Distance F
N ¼ 129
Loading (min * 102)
r2 0.322 F ¼ 21 p < 0.0001 306.764 þ 17.708 Distance W þ 79.869 m3 þ 54.316 m3 * Crawler
N ¼ 130
Travel loaded in forest (min * 102)
r2 0.719 F ¼ 106 p < 0.0001 N ¼ 125 29.517 þ 7.337 Crawler * Distance F 3.868 Team A * Crawler * Distance F þ 2.974 m3 * Distance F
Travel loaded on trail (min * 102)
r2 0.919 F ¼ 359 p < 0.0001 26.497 0.895 Distance T þ 2.422 Crawler * Distance T 0.633 Team A * Crawler * Distance T þ 1.339 m3 * Distance T
Unloading (min * 102)
p ¼ 0.0761
N ¼ 127
N ¼ 129
349.304 if Crawler, 396.720 if Skidder Accessory time and delays (min * 102)
0.256 * (Travelling þ Loading þ Unloading)
Distance T ¼ one-way travel distance on the prepared skid trail in m; Distance F ¼ one-way travel distance on the forest floor in m; Distance W ¼ winching distance in m; Crawler ¼ indicator variable: 1 if crawler, 0 if skidder; Team A ¼ indicator variable: 1 if Team A, 0 if Team B; Load ¼ Load size in m3.
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
Fig. 3. Gross skidding productivity as a function of distance on the prepared skid trail, machine type and team. Note: the curves in the graph were calculated for a skidding distance on the forest floor ¼ 35 m, a winching distance ¼ 15 m, an average load size ¼ 1.024 and 1.490 m3 for the skidder manned by Teams A and B respectively, and 0.780 and1.217 m3 for the crawler manned by Teams A and B (mean values from the study). Productivity figures are calculated on total time consumption, and include preparation and delay time. Volume figures are outside bark.
productivity and lower fuel consumption. The higher productivity of Team B is also reflected in a lower financial and energy cost. Cost reduction averages 15%, but ranges from 7% to 24% and decreases with distance due to the compensating effect of the higher travel speed achieved by Team A with the crawler. Fig. 4 shows the relationship between skidding cost and annual use. The skidder always offers a lower extraction cost than the crawler, but the margin increases with annual use. However, the crawler is an agricultural tractor and can be used for more tasks than just skidding. This may allow for more use year-round, especially when logging opportunities are limited. Hence a comparison was done by assuming different utilization levels for the two machines. Table 7 shows the annual use of the skidder and the corresponding additional number of hours per year the crawler should work on other jobs, if it was to achieve the same unit skidding cost. If the skidder works at least 500 h per year, it will
Table 6 Skidding cost as a function of distance on trail, machine type and team. Dist. (m) Crawler Team A Crawler Team B Skidder Team A Skidder Team B Skidding cost, Euro m3 50 24.5 100 27.0 150 29.5 200 32.1 250 34.6 300 37.2 350 39.7 400 42.3 450 44.8 500 47.3
19.5 22.0 24.4 26.9 29.4 31.9 34.4 36.8 39.3 41.8
17.2 17.9 18.5 19.2 19.9 20.5 21.2 21.9 22.6 23.2
13.0 14.0 14.9 15.9 16.8 17.8 18.7 19.7 20.6 21.6
Skidding cost, MJ m3 50 57.5 100 63.5 150 69.5 200 75.5 250 81.5 300 87.5 350 93.4 400 99.4 450 105.4 500 111.4
45.8 51.7 57.5 63.3 69.2 75.0 80.9 86.7 92.5 98.4
23.7 24.7 25.6 26.5 27.4 28.4 29.3 30.2 31.2 32.1
18.0 19.3 20.6 21.9 23.2 24.6 25.9 27.2 28.5 29.8
Dist. ¼ skidding distance on trail, in m; the assumed skidding distance on the forest floor is 35 m, and winching distance 15 m, Volume figures are outside bark.
77
Fig. 4. Skidding cost as a function of annual use.
offer a cheaper service than the crawler, regardless of how much additional work the crawler obtains.
5. Discussion The hypothesis of the study is confirmed; the new mini-skidder offers superior productivity and lower extraction cost compared to the crawler. That is the necessary economic condition for crawler replacement, which is already desirable in terms of site impact and labor safety. Most of the forestry-fitted crawler tractors inspected during the above-mentioned survey do not comply with current safety regulations (i.e. EU Directive 2008/50/CE). Upgrading to current safety standards may incur substantial cost, thus making replacement a more desirable option. In fact the Croatian Forest Administration decommissioned its substantial crawler fleet in the mid 1990s and replaced it with new skidders (Beuk et al., 2007). The superior payload capacity of the skidder depends on its unique structural design, which allows for a better lifting of the log ends and an easy transfer of the vertical load component onto the rear axle (Stoilov and Kostadinov, 2009). That increases the tractive capacity of the skidder, especially when negotiating uphill grades (Tomasi c et al., 2009). The structure of the crawler does not allow for the same lift, so only a small part of the total load weight contributes to increasing the crawler’s tractive capacity. The study also shows the importance of the human factor, and sheds light on the most efficient skidding techniques. The productivity difference between teams averaged 15%, for the same machine and work conditions. The difference was smaller than reported for complex tree felling and processing machines (Ovaskainen and Heikkilä, 2007), but the impact of human adaptation is likely to increase with task complexity (Kärhä et al., 2005). The study indicates that load maximization increases productivity and decreases cost, and that a double-drum winch certainly helps assembling larger loads.
Table 7 Annual use of the skidder and additional annual hours the crawler should work in order to achieve the same skidding cost. Hours skidder
Additional hours crawler
Skidding cost (V m3)
100 200 300 400 500
5 50 150 300 N
33 22 18 16 15
Note: volume figures are outside bark.
78
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79
In fact, readers should note that the study was conducted with as much rigor as possible, and yet it occasionally violated the statistical assumption of a strict experimental design. Therefore, the regressions should not be used to determine the exact weight of each independent variable, but only to establish its general significance. The productivity levels achieved in this study equal or exceed the productivity levels reported in previous studies for similar machines. Magagnotti and Spinelli (2011b) report a productivity of 2.3 m3 h1 and an average load of 0.8 m3 for a forestry-fitted crawler tractor as used in this study. Ze ci c and Maren ce (2005) indicate a productivity of 2.4 m3 h1 and an average load of 1.1 m3 for the same type of mini-skidder tested in this study, operating over the same distance. However, stand type and the silvicultural prescriptions were different, which may explain the different productivity and load size. In any case, results are comparable, which may support the use of our productivity figures as a general reference. The energy cost ranged from 18 to 32 MJ m3 for the miniskidder and from 45 to 111 MJ m3 for the crawler. These figures are comparable with those reported by Picchio et al. (2009) (36 MJ m3) and by Magagnotti and Spinelli (2011a) (60e120 MJ m3). They are relatively low, which may be expected for intermediate mechanization (Berg, 1997), especially under Alpine conditions (Valente et al., 2011). The mini-skidder is equipped with a modern Tier III engine, which is likely to dramatically reduce emissions (Berg and Lindholm, 2005). 6. Conclusions Rubber-tired mini-skidders can effectively replace forestryfitted crawler tractors under certain conditions as outlined in this study. Replacement is desirable in terms of environmental protection and labor safety, and offers substantial economic benefit; it requires a moderate additional initial investment but allows reducing extraction cost between 30 and 50%. As compared to the crawler tractor, the mini-skidder is more environmentally friendly in terms of its fossil fuel use. Most crawlers used in forestry are now quite old. With the intent of promoting rural development, the European Union offers grants for the acquisition of new machinery, and low-impact technologies are generally favored over conventional, less-efficient solutions. Regional managers should be apprised of these new environmentally compatible products when developing grant programmes. The manufacturer may also consider adapting and registering his mini-skidder for road travel according to the Italian Law. That would enable independent, low-cost translocation between work sites. Road trafficability would give the new machine further advantage over the crawler, which cannot travel on public roads and is moved between work sites on a trailer. The study demonstrates that one can effectively control most sources of unwanted variability even when conducting the comparative test in a real work environment, such as a forest. The same methodology could also be applied to other machines or operations. Acknowledgments This study is a part of the project “Organization and rationalization of logging operations” funded by Regione Autonoma Friuli-Venezia Giulia, Servizio Gestione Forestale e Produzione Legnosa. This study was also made possible thanks to funding received from the STSM programme of Action COST FP902. Thanks are also due to Dr. Mario Di Gallo and the CESFAM team for their technical support.
References Bailey, A., Basford, W., Penlington, N., Park, J., Keatinge, J., Rehman, T., Tranter, R., Yates, C., 2003. A comparison of energy use in conventional and integrated arable farming in the UK. Agricultural Ecosystems and Environments 97, 241e253. Berg, S., 1997. Some aspects of LCA in the analysis of forestry operations. Journal of Cleaner Production 5, 211e217. Berg, S., Lindholm, E., 2005. Energy use and environmental impacts of forest operations in Sweden. Journal of Cleaner Production 13, 33e42. Bergstrand, K.G., 1991. Planning and analysis of forestry operation studies. Skogsarbeten Bulletin 17, 63 pp. Beuk, D., Tomasi c, Z., Horvat, D., 2007. Status and development of forest harvesting mechanization in Croatian State forestry. Croatian Journal of Forest Engineering 28, 63e82. Björheden, R., Apel, K., Shiba, M., Thompson, M.A., 1995. IUFRO Forest Work Study Nomenclature. Swedish University of Agricultural Science, Dept. of Operational Efficiency, Garpenberg, 16 pp. De Lasaux, M., Hartsough, B., Spinelli, R., Magagnotti, N., 2009. Small parcel fuel reduction with a low-investment, high-mobility operation. Western Journal of Applied Forestry 24, 205e213. Dubey, P., 2008. Investment in small-scale forestry enterprises: a strategic perspective for India. Small-scale Forestry 7, 117e138. Gullberg, T., 1995. Evaluating operatoremachine interactions in comparative time studies. Journal of Forest Engineering 7, 51e61. Hartsough, B., 2003. Economics of harvesting to maintain high structural diversity and resulting damage to residual trees. Western Journal of Applied Forestry 18, 133e142. Hill, S., 1991. D4H tractor and towed arch in radiata clearfell. LIRA Report 16, 8. 4 pp. snjar, M., 2007. Morphological characteristics and producHorvat, D., Ze ci c, Z., Su tivity of skidder Ecotrac 120V. Croatian Journal of Forest Engineering 28, 11e25. INFC, 2008. National Forest Inventory e Tab. 3.1 Surface by Ownership Type. www. infc.it (accessed 04.11.11). Kärhä, K., Jouhiaho, A., Mutikainen, A., Mattila, S., 2005. Mechanized energy wood harvesting from early thinnings. International Journal of Forest Engineering 16, 23e36. Kittredge, D., Mauri, M., McGuire, E., 1996. Decreasing woodlot size and the future of timber sales in Massachusetts: when is an operation too small? Northern Journal of Applied Forestry 13, 96e101. Larsen, J.B., 1995. Ecological stability of forests and sustainable silviculture. Forest Ecology and Management 73, 85e96. Magagnotti, N., Spinelli, R., 2011a. Financial and energy cost of low-impact wood extraction in environmentally sensitive areas. Ecological Engineering 37, 601e606. Magagnotti, N., Spinelli, R., 2011b. Integrating animal and mechanical operations in protected areas. Croatian Journal of Forest Engineering 32, 489e499. McElroy, N., 2006. Parcelization of the Chesapeake Bay watershed and implications for sustainable forestry. Master thesis, Virginia Polytechnic Institute and State University, Falls Church, VA. 57 pp. Mikkola, H., Ahokas, J., 2010. Indirect energy input of agricultural machinery in bioenergy production. Renewable Energy 35, 23e28. Miyata, E., 1980. Determining Fixed and Operating Costs of Logging Equipment. General Technical Report NC-55. Forest Service North Central Forest Experiment Station, St. Paul, MN, 14 pp. Moldenhauer, M., Bolding, M., 2009. Parcelization of South Carolina’s private forestland: loggers’ reactions to a growing threat. Forest Products Journal 59, 37e43. Olsen, E., Hossain, M., Miller, M., 1998. Statistical Comparison of Methods Used in Harvesting Work Studies. Oregon State University, Forest Research Laboratory, Corvallis, OR, Research Contribution n 23, 31 pp. Ovaskainen, H., Heikkilä, M., 2007. Visuospatial cognitive abilities in cut-to-length single-grip harvester work. International Journal of Industrial Ergonomics 37, 771e780. Pellizzi, G., 1992. Use of energy and labour in Italian agriculture. Journal of Agricultural Engineering Research 52, 111e119. Picchio, R., Maesano, M., Savelli, S., Marchi, E., 2009. Productivity and energy balance in conversion of a Quercus cerris L. coppice stand into high forest in Central Italy. Croatian Journal of Forest Engineering 30, 15e26. Primmer, E., 2011. Analysis of institutional adaptation: integration of biodiversity conservation into forestry. Journal of Cleaner Production 19, 1822e1832. Rickenbach, M., Steele, T., 2006. Logging firms, nonindustrial private forests, and forest parcelization: evidence of firm specialization and its impact on sustainable timber supply. Canadian Journal of Forest Research 36, 186e194. SAS Institute Inc., 1999. StatView Reference. SAS Publishing, Cary, NC, ISBN 158025-162-5, pp. 84e93. Spinelli, R., Magagnotti, N., Nati, C., 2010a. Benchmarking the impact of traditional small-scale logging systems used in Mediterranean forestry. Forest Ecology and Management 260, 1997e2001. Spinelli, R., Magagnotti, N., Picchi, G., 2010b. Deploying mechanized cut-to-length technology in Italy: fleet size, annual usage, and costs. International Journal of Forest Engineering 21, 23e31. Spinelli, R., Magagnotti, N., 2011a. The effects of introducing modern technology on the financial, labour and energy performance of forest operations in the Italian Alps. Forest Policy and Economics 13, 520e534.
R. Spinelli et al. / Journal of Cleaner Production 31 (2012) 73e79 Spinelli, R., Magagnotti, N., 2011b. Wood extraction with farm tractor and sulky: estimating productivity, cost and energy consumption. Small-scale Forestry. doi:10.1007/s11842-011-9169-8. Stoilov, S., Kostadinov, G., 2009. Effect of weight distribution on slip efficiency of a four-wheel-drive skidder. Biosystems Engineering 104, 486e492. snjar, M., Horvat, D., Kristi Su c, A., Pandur, Z., 2008. Morphological analysis of forest tractor assemblies. Croatian Journal of Forest Engineering 29, 41e51.
79
snjar, M., Horvat, D., Pandur, Z., 2009. Factors affecting timber Tomasi c, Z., Su skidding. Croatian Journal of Forest Engineering 30, 127e139. Valente, C., Spinelli, R., Hillring, B., 2011. LCA of environmental and socio-economic impacts related to wood energy production in alpine conditions: Valle di Fiemme (Italy). Journal of Cleaner Production 19, 1931e1938. Vaughan, L., 1988. Thinning with small crawler tractors. LIRA Report 13. 6 pp. Ze ci c, Z., Maren ce, J., 2005. Mathematical models for optimization of group work in harvesting operation. Croatian Journal of Forest Engineering 26, 29e37.