The productivity and cost of harvesting whole trees from early thinnings with a felling head designed for continuous cutting and accumulation

ABSTRACT

The main problem with the utilization of untended stands for energy use is that small-diameter trees are expensive to harvest, and the value of the harvest fails to exceed the harvesting cost. Until now, the equation has appeared difficult to solve, but a wood harvesting innovation – the Risupeto II accumulating felling head – which works on a continuous basis, could provide a solution to the harvesting cost problem of young stands with a dense undergrowth. The novel accumulating felling head is attached to the boom of the crawler excavator, and trees are harvested at full-tree length. The crosscutting is done later during loading by a forwarder equipped with a grapple saw. The objectives of the present study were: 1) to define the productivity; 2) produce time consumption models for the Risupeto II harvesting unit in selective thinnings and 3) to determine the harvesting cost of whole trees from the early thinnings based on the above described two-machine configuration. Time study data from the 16 study plots consisted of 3,099 whole trees where the average tree volume varied between 14.2 and 52.0 dm³ resulting in 73 m³ of whole trees. The productivity varied between 11.2 and 26.6 m³/PMh on the time study plots indicating that felling-bunching productivity was relatively high compared to the latest studies with the Nordic harvesting technology in early thinning. According to the profitability analysis, the harvesting costs could be covered by the revenue from energy wood sales.

KEYWORDS:

• Energy wood
• excavator
• harvesting
• cutting device
• young stands
• selective thinning

Introduction

Selective thinning of young stands improves the value of the growing stock and its durability against abiotic damage, pests, and diseases (Varmola and Salminen 2004; Huuskonen and Hynynen 2006; Uotila and Saksa 2014; Haikarainen et al. 2021). Removing less valuable tree species that hamper the growth of the desired tree species also increases the growth of the remaining trees (Heikkilä et al. 2007, 2009; Uotila et al. 2020; Haikarainen et al. 2021). The main problems with the utilization and forest management of untended stands in Finland are that small-diameter trees are too expensive to harvest for energy use, and on the other hand, not all forest owners can afford to tend their sprawling stands to production condition (Vestlund et al. 2006; Ahtikoski et al. 2008; Oikari et al. 2010; Belbo 2011; Bergström et al. 2012; Fernandez-Lacruz et al. 2015; Karttunen and Laitila 2015, Erber et al. 2016; Nuutinen et al. 2021). Tree volume governs the productivity in small tree harvesting and it is important to identify the minimum tree volume that makes harvesting economic for each situation: Below such a size, productivity does not reach the required level, and the value of the harvest fails to exceed the harvesting costs (i.e., cutting and forwarding) (Oikari et al. 2010; Petty 2014; Ahtikoski et al. 2021; Austin et al. 2021; Laitila and Väätäinen 2021).

Until now, the equation has appeared difficult to solve, but a wood harvesting innovation – the Risupeto II accumulating felling head – which works on a continuous basis could provide a solution for the economic problem of untended young stands with a dense undergrowth, as at least part of the costs can be covered by revenue from energy wood sales. Working on a continuous basis means that the accumulating felling head can cut and accumulate whole trees in an upright position during a continuing movement and the work is only interrupted when the accumulating felling head is full and ready to bunch the whole trees on the ground (Bergström et al. 2007, 2012; Bergström 2009, 2017; Belbo 2011; Grönlund et al. 2015). It may be that the threshold of economic productivity for felling bunching is more likely to be achieved when whole trees can be harvested at their full length without having to cut them into shorter pieces during cutting. It is more feasible and economic for a forwarder equipped with a grapple saw to carry out a crosscutting procedure for the bunches of whole trees later during loading (Bergström and Di Fulvio 2014; Jylhä and Bergström 2016). With a robust felling head capable of continuous cutting and accumulating while continuously moving the boom/head, it is possible to improve the felling-bunching productivity compared to multi-tree handling with conventional accumulating felling heads equipped with a saw bar or shear blade cutting elements (Laitila and Väätäinen 2021). In addition, the semi-sharp disk sawblades shatter the cutting surface of the stump, which may decrease regrowth (Laitila and Väätäinen 2021).

The novel accumulating felling head is attached to the boom tip of the medium-sized crawler excavator (Laitila and Väätäinen 2021). The advantages of excavators produced in high volumes include a substantially lower purchase price than for conventional forest machines, such as harvesters and forwarders, and outside of the harvesting season, the option of removing the harvesting equipment and using the base machine in the work for which it has been originally designed (Väätäinen et al. 2004; Laitila and Väätäinen 2013). The versatility of machinery represents one way of achieving year-round employment and ensuring the availability and stability of a professional workforce. It also improves the profitability of the machine-contracting business (Laitila and Väätäinen 2013). In Finland, the use of excavators in logging is relatively rare whereas the majority (over 70,000 hectares annually) of soil preparation in forest regeneration areas is done with excavator-based machinery (Laine et al. 2019). According to a survey (Palander et al. 2012a), the mobility of excavator-based harvesters was considered to be poorer than that of wheeled harvesters on slopes and in rocky terrain. On peatlands and other flat land, excavators’ mobility was rated good (Palander et al. 2012a).

In Sweden, there has been an extensive effort to develop methodology and technology for the devices that can continuously cut and accumulate trees during linear crane movement in early thinnings (Bergström 2009, 2017; Bergström et al. 2012, 2022; Bergström and Di Fulvio 2014; Grönlund et al. 2015). As a result of the development work, harvester mounted accumulating felling-head, Bracke C16 (www.brackeforest.com), exist within the market (Bergström and Di Fulvio 2014; Bergstöm et al. 2022). In addition, Bracke MAMA and Flowcut accumulating felling heads are at the prototype phase (Bergström and Di Fulvio 2014; Grönlund et al. 2021a, 2021b). Bracke accumulating felling heads cut trees using a 1.9 cm (¾ inch) saw chain, which is fitted to a circular disc (Bergström and Di Fulvio 2014). The Flowcut head cuts trees with a circular saw blade (Grönlund et al. 2021a) or using a 1.9 cm (¾ inch) saw chain fitted to a saw bar that is fixed at both ends (2021b).

The prototype of the Risupeto accumulating felling head was studied in the integrated clearing and harvesting of roadside and field-edge brushwood for fuel (Laitila and Väätäinen 2021). The prototype was operated without any unnecessary breakdown delays, and the quality of the clearing work was good. The felling-bunching productivity figures were relatively high compared to the productivity of conventional Nordic harvesting operations in short-rotation forestry (Jylhä and Bergström 2016) and harvesting power-line corridors for energy (Fernandez-Lacruz et al. 2013). These results inspired tan evaluation of the harvesting productivity and quality of the Risupeto accumulating felling head in early thinnings, as well. A new and improved commercial version, Risupeto II, was selected for this study. Already a few Risupeto II accumulating felling heads have been manufactured and sold, and a serial production is about to begin at Reformet Oy (www.reformet.fi) in Finland.

The objectives of this study were 1) to define the productivity of felling bunching; 2) produce time consumption models or parameters for a machine unit capable of continuous felling and bunching of whole trees in early thinnings and 3) to assess the cost and profitability of energy wood harvesting in varying stand conditions in early thinnings. The harvesting cost and profitability analysis was performed by a spreadsheet-based system analysis, using productivity figures for felling bunching from this study, an existing whole-tree forwarding model, and official price statistics for delivery sales and the cost parameters for the whole-tree harvesting.

Materials and methods

Field study

Machinery studied

Risupeto II is an accumulating felling head that is capable of continuous boom movement while cutting and accumulating multiple trees in selective thinnings (Figure 1). The felling head cuts trees with two parallel disk sawblades and accumulates cut trees in the collecting chamber using continuously rotating collecting arms made of a thick rubber mat (Figure 1). When the collecting chamber of the felling head is full, the chamber gate is closed with two sliding latches, and the accumulated whole-tree bunch is moved to the pile and dropped down. The unloading of the whole-tree bunch is done by tilting the felling head downward and rotating the disk saws and collecting arms in the opposite direction from that during cutting. The opening and closing of the chamber gate latches are integrated with the rotation and stopping movements of the collecting arms and disk sawblades. The width of the hydraulically powered accumulating felling head is 95 cm, and the maximum cutting diameter with one cut is 30 cm. The height of the accumulating felling head is 165 cm, and its mass with the extender (optional extra) is 1,250 kg. The device is primarily intended for 16–20 tonnes crawler excavators to avoid stability problems during whole tree harvesting. The stability of the excavator can be still increased with additional counterweights and forestry-equipped heavy-duty tracks.

Figure 1. The felling bunching of whole trees in the early thinning was done with the Risupeto II felling head, capable of continuous cutting and accumulation. In time studies, the Risupeto II accumulating felling head was attached to the boom tip of the Kobelco SK140SRL-7 crawler excavator (see lower right-hand corner). Photos: J. Laitila and Reformet Oy.

The studied Risupeto II accumulating felling head was attached with a tilt rotator to the boom tip of the new forestry equipped Kobelco SK140SRL-7 (www.kobelcocenter.fi) crawler excavator (18.5 tonnes, 86 kW), providing a total reach of 10 meters with the extender (Figure 1). The rotation speed of the disk sawblades, which were made of high-strength wear-resistant steel, was 990 rotations per minute. The base machine’s hydraulic oil flow and pressure were 284 l/min and 340 bar. The excavator was equipped for logging with a short tail-swing extension (180 mm equipped with an additional counterweight). The excavator’s width, with 700-millimeter-wide sheltered heavy-duty tracks, was 2,790 mm, and the ground clearance was 410 mm. The operator was skilled, with a year’s experience of working with the Risupeto II accumulating felling head and whole-tree harvesting in early thinnings.

Time-study and stand measurements

Time studies were carried out in a dense early thinning stand in Iitti, Southeast Finland (61°02ʹN 26°12ʹE). The stand was 15–25 years old. The area had been cultivated and planted with spruce, after which all silvicultural work had been completely neglected. In establishing the time-study plots, the aim was that the number and volume of trees should be varied between the time-study plots (Table 1). The length of the rectangular time-study plots was 25 m, and the widths equaled the measured work path, which was an average of 20.04 m wide (measured after thinning). The boundaries of the time-study plots were marked with ribbons and poles at the stand. Strip roads were not marked in advance, but they were planned by the operator during thinning. The operator also chose the trees to be removed (thinning from below), in accordance with silvicultural recommendations (Äijälä et al. 2019). The total number of time-study plots in the birch spruce mixture forest stand was 16, each with a surface area of 500.9 m².

Table 1. Harvesting properties of the time-study plots 1–16. The estimated share of tree species of the cutting removal as well as height and DBH were estimated based on stump diameter measurements on circular sample plots and existing regression models (Repola et al. 2007; Laitila and Väätäinen 2020). Standard deviation in parentheses.

Time study plot number	Initial stand, number of trees per hectare	Removal, number of trees per hectare	Average volume of harvested trees, dm^3	Cutting removal, m^3/ha	Norway spruce, %	Silver birch, %	Other broadleaved tree species, %	Scots pine, %	Mean stump diameter of cutting removal, mm	Estimated mean height of cutting removal, m	Estimated mean DBH of cutting removal, mm
1	5,671	4,871	19.5	95.1	32	68	-	-	69 (31)	6.2 (2.0)	57 (26)
2	5,811	5,011	18.1	90.8	47	53	-	-	64 (37)	5.7 (2.1)	53 (31)
3	5,791	4,991	18.8	93.6	35	65	-	-	68 (30)	6.3 (1.9)	57 (25)
4	4,873	3,973	25.7	102.2	60	40	-	-	77 (44)	5.9 (2.8)	64 (37)
5	4,095	2,995	21.6	64.7	51	48	1	-	79 (51)	6.7 (2.7)	66 (43)
6	4,833	4,033	25.1	101.0	59	41	-	-	103 (51)	7.9 (2.4)	86 (43)
7	8,267	7,367	14.2	104.3	31	67	2	-	65 (32)	5.9 (2.1)	54 (26)
8	4,613	3,913	25.8	101.0	52	43	5	-	68 (34)	5.8 (2.2)	57 (29)
9	3,116	2,516	34.0	85.6	61	22	17	-	88 (41)	6.7 (2.4)	76 (36)
10	2,897	1,897	31.0	58.8	41	55	4	-	80 (29)	6.9 (1.7)	67 (24)
11	4,673	3,773	22.5	84.8	63	20	17	-	77 (39)	6.2 (2.0)	65 (34)
12	2,117	1,517	52.0	78.9	52	44	4	-	116 (44)	8.7 (2.0)	97 (37)
13	6,790	5,890	20.0	117.6	30	61	9	-	61 (30)	5.8 (1.9)	51 (25)
14	4,554	3,554	24.4	86.6	34	62	3	-	73 (39)	6.3 (2.3)	60 (33)
15	4,075	3,075	30.5	93.9	32	65	3	-	92 (47)	7.4 (2.8)	77 (39)
16	3,396	2,496	39.5	98.6	45	33	2	21	99 (53)	7.3 (2.7)	83 (45)

Terrain conditions on the time study plots were estimated by the Finnish classification system (Anon 1990). The factors assessed were the load-bearing capacity, the roughness of the terrain surface, and the steepness of the terrain. Based on the measurements, the study site was classified as terrain class 1 (easy/normal in Finnish conditions).

The time study of the felling bunching of whole trees in early thinnings with the Risupeto II accumulating felling head was conducted during the 1^st and 2^nd of November 2021 during daytime in natural light conditions. The weather was cloudy but not rainy, and the temperature was 5 to 10 °C during time studies. The cutting work was recorded using a video camera (GoPro 7), and the time study was carried out by analyzing the collected video material with a continuous time-study method. The camera was placed in the cabin of the excavator in a position that enabled it to fully observe the working elements and the number of removed trees during the cutting operation as well as the markings in the forest (i.e. the boundaries of the time-study plots), with minimal obstruction for the machine operator. All time-study plots were cut continuously without any additional pauses between the plots. In time studies, cutting was carried out according to the normal pace to which the operator was accustomed.

The harvested whole trees were weighed during unloading by the Valmet 860.1 forwarder loader, equipped with an Epec Load Optimizer crane scale with an accuracy of 2 kg. Forwarding of whole trees was completed immediately after the cutting trial. The weighed biomass was converted to m³ (solid) based on the relative proportions of tree species on the time-study plot (Table 1). The number of forwarder loads weighed was the same as the number of time-study plots. The weighed cutting removal from the time-study plots was converted to m³ with tree species distribution-specific green density (kg m⁻³) values, produced by the Finnish Forest Research Institute (Lindblad et al. 2010; Ministry of Agriculture and Forestry 2010) which are presented in official tables for each region, month and freshness class. The average green density was 900 kg m⁻³ in the time-study plots.

The respective values for the total number of whole trees harvested and volume were 3,099 and 73.0 m³ during the time study. The average volume of the harvested trees varied between 14.2 and 52.0 dm³, the harvesting intensity ranged from 1,517 to 7,367 harvested trees per hectare, and the harvesting removal was 58.8–117.6 m³ per hectare on the time-study plots (Table 1). The average volume of whole trees harvested (dm³) in a time-study plot was calculated by dividing the cutting removal (m³) by the number of harvested trees (Table 1). The number of trees harvested was obtained from the time study, in which the number of trees in each felling head bunch during cutting was observed and recorded. The total number of felling head bunches were 415 in the time study.

The width and the spacing of strip roads were measured at 25-meter intervals along the strip road after thinning. The strip roads’ width was measured via the “SLU method,” in which the distances to the nearest trees were measured at right angles from the middle of the strip road, along a distance of 10 m on each side (Björheden and Fröding 1986). The measurement point on the trees was the stump height level, and the width of the strip road was the sum of the two distances (Björheden and Fröding 1986). These measurements were accurate to within 1 cm. The distance between two parallel strip roads was measured perpendicularly from the middle of the left-hand strip road to the middle of the right-hand strip road. The accuracy of these measurements was 10 cm.

The data of the remaining trees were collected after thinning from two circular 50 m² sample plots, located as Figure 2 shows on the time-study plot. The number of remaining trees, the mean height (m), the mean overbark diameter (mm) at a height of 1.3 m (DBH), and the basal area per hectare (m²) were recorded in the tree sample plots. The initial number of trees on a time-study plot was obtained by summing the number of harvested and remaining trees (Table 1).

Figure 2. Location of the sample plots for stand measurements of remaining trees and stumps on the time-study plot.

The relative proportion of tree species on the time-study plot (Table 1) for the crane-scale measurement described above was estimated using regression models for tree height and DBH (Laitila and Väätäinen 2020), in which the stump diameter at 10 cm height was the independent variable, and the tree biomass models of Repola et al. (2007). The stump diameter was measured at a height of 10 cm from the ground to an accuracy of 1 mm, while the tree species was determined visually on the circular sample plots (Figure 2). The average stump diameter was 75 mm (SD 40) at the stand level, and ranged between 61 and 116 mm on the time-study plots (Table 1). Based on the stump diameter measurements, the estimated share of tree species of the total cutting removal was 2% for Scots pine (Pinus sylvestris), 44% for Norway spruce (Picea abies), 50% for silver birch (Betula pendula), and 4% for other broadleaved tree species, such as aspen, alder, rowan, and willow. The estimated mean height of trees ranged between 5.7 and 8.7 m within the time-study plots (Table 1).

Analysis of the time-study data

The video material was analyzed by continuous timing from a TV screen using a Rufco-900 handheld field computer. An experienced researcher was responsible for collecting all the time-study material. The recording accuracy of the field computer was 0.6 seconds. The working time was recorded through the application of a continuous timing method in which a clock ran continuously, and the exact time of the end of the work element was recorded with distinct numerical codes (Harstela 1991; Magagnotti et al. 2013). The accumulating felling had the highest priority in the time study, followed by the piling and moving elements. The time for the highest prioritized work element was recorded if multiple work elements were performed simultaneously. The auxiliary time use (e.g., planning of work and preparations) was included in the work phases in which it was observed. Productive machine (PM) time, the working time excluding all delays (IUFRO 1995), was divided into the following work phases, as listed in Table 2.

Table 2. Work elements with detailed definitions of the whole-tree harvesting with the Risupeto II accumulating felling head in the time study.

Work element	Definition of the work element
Moving between working locations	Begins when the excavator starts to move (heavy-duty tracks turning) and ends when the excavator stops moving to perform another activity at the working location, i.e., machine positioning to perform accumulating felling.
Felling accumulation	Begins when the boom starts to swing toward the first removable tree, and continuous cutting and accumulation starts. The work element ends when the accumulating felling head is full, and the head with the tree bunch begins to move to the pile. The number of whole trees in each work element is observed and recorded.
Arranging the brushwood tree bunch in a pile	Begins when the last tree has been cut and accumulated to the head and ends when the head with the tree bunch is tilted and laid (i.e., bunched) on the ground. The tree bunch is unloaded from the accumulating chamber by rotating the disk saws and collecting arms in the opposite direction to that during cutting.
Delay	Time not related to productive felling-bunching work, e.g., personal breaks, repairing or maintenance of the base machine or accumulating felling head. The reason for the interruption was recorded.

For the productivity modeling of the whole-tree cutting with the Risupeto II accumulating felling head, the recorded plot-wise time-study data and the measured harvested whole-tree volumes (m³ measured plot-wise by crane scale) were combined as a data matrix. The time consumption of working elements of accumulating felling and arranging the whole-tree bunch into a pile were combined for the felling-bunching time. In modelling the productive machine, time consumption was expressed in seconds per harvested whole tree, and the productivity in solid cubic meters per productive machine hour time (m³ per PMh).

The time consumption of felling bunching was formulated by a regression analysis, in which the volume of harvested whole trees (dm³) and the number of whole trees per felling-bunching cycle were used as independent variables. Various transformations and curve types were tested to achieve symmetrical residuals and to ensure the statistical significance of the coefficients. Tables 4 and 5 show the results of the regression model analyses. According to the regression analyses, the intensity of harvested whole trees per hectare was not a statistically significant independent variable for the time consumption of moving, and the moving time per tree was therefore calculated from the average time consumption values of all the time-study plots, regardless of the harvesting intensity (Figure 8). The regression analysis was carried out with IBM SPSS Statistics 21.0 statistical software.

Table 3. Basic stand data for time-study plots 1–16 after harvesting.

Time-study plot number	Number of remaining trees per hectare	Mean DBH of remaining trees, mm	Mean height of remaining trees, m	Basal area of remaining trees, m^2/ha	Initial stand, number of trees per hectare	Estimated mean height of cutting removal, m
1	800	92	12.3	8.0	5,671	6.2
2	800	110	13.1	9.0	5,811	5.7
3	800	101	13.2	8.0	5,791	6.3
4	900	103	10.8	9.5	4,873	5.9
5	1,100	105	10.5	8.5	4,095	6.7
6	800	104	10.5	9.0	4,833	7.9
7	900	96	12.2	10.0	8,267	5.9
8	700	119	12.3	9.5	4,613	5.8
9	600	107	10.6	9.0	3,116	6.7
10	1,000	116	11.2	8.5	2,897	6.9
11	900	135	12.0	11.0	4,673	6.2
12	600	115	11.4	9.0	2,117	8.7
13	900	122	13.8	9.5	6,790	5.8
14	1,000	133	14.8	9.0	4,554	6.3
15	1,000	118	12.7	12.5	4,075	7.4
16	900	116	10.0	13.5	3,396	7.3

Table 4. Regression model statistics for accumulating felling bunching of whole trees with the Risupeto II accumulating felling head (N_Number = number of trees in the accumulating felling head).

Work phase model	Dependent variable	R^2	F-test		N	Term	Constant/Coefficient		t-test
			F-value	p			Estimate	Std. error	t-value	p
Felling bunching	TFelling bunching	0.59	19.845	<0.001	16	Constant exp(NNumber)	11.025 –0.099	1.824 0.022	6.044 –4.455	<0.001 0.001

Table 5. Regression model statistics for the number of trees in the Risupeto II accumulating felling head (x = the average volume of the harvested whole trees on the study plot).

Work phase model	Dependent variable	R^2	F-test		N	Term	Constant/Coefficient		t-test
			F-value	p			Estimate	Std. error	t-value	p
Number of trees in the felling head	NNumber	0.57	18.783	<0.001	16	Constant x^−1	3.768 84.279	0.848 19.446	4.445 4.334	0.001 0.001

Cost and profitability analysis

In the system analysis, the productivity of felling bunching (m³ per PMh) was based on this study’s results (observed plot-wise productive machine time consumption and plot-wise cutting removal (m³) measured by crane scale), and the forwarding productivity of whole trees was calculated using the model by Laitila et al. (2007), assuming that using a grapple saw for crosscutting increased loading time by 50% (Jylhä et al. 2015). In addition, the grapple load size during unloading was set at half (0.32 m³) the normal level (Laitila et al. 2007) due to a smaller grapple size and the extra weight of saw cassette accessories. The payload of the forwarder was set to 7.1 m³ in line with the work of Laitila and Väätäinen (2021).

The productive machine hour (PMh) productivities of felling bunching by Risupeto II and forwarding were converted to operating hour productivities––also known as scheduled machine hour (SMh) productivity––using coefficients of 1.3 and 1.2, respectively (Laitila 2008). The operating hour costs, €92.90 for the felling bunching and €84.00 for the forwarding, were acquired from the study of Laitila and Väätäinen (2021) and updated to the current cost level with the cost indexes of forest machinery and vehicles produced by Statistics Finland (2022).

The profitability analysis considered the roadside price of the energy wood and the harvesting (i.e., felling bunching and forwarding) cost, both in € per m³. The roadside prices of €23.00, and €25.10 m⁻³ for the harvested whole trees as energy wood was based on the official price statistics of delivery sales between 2020 and 2021 in Finland (Volumes and … 2022). To assess the profitability, the energy wood harvesting method in varying stand conditions, the roadside prices of whole trees as incomes, and the harvesting cost of whole trees as expenses were summed in 16 scenarios. Each scenario represented the stand conditions of a time-study plot in the field study (study plots 1–16). The forwarding distance was assumed to be 300 m, and the total length of the strip road network at the thinning stand was assumed to be 600 m per ha, based on an average strip road spacing of 20 m (Niemistö 1992).

Results

Harvesting quality

The strip-road spacing observed within the time-study stand varied between 17.0 to 23.2 m, and the average spacing was 20.04 m (SD 2.1 m) (Figure 3), which did not differ from the recommended minimum spacing of 20 m (Anon 2021). The average width of the strip road was 452 cm (SD 39 cm) on the time-study stand (Figure 3), which did not exceed the recommended maximum width of 460 cm (Anon 2021). The harvesting quality (Table 3 and Figure 4) achieved the recommended silvicultural standards as regards for remaining trees (Äijälä et al. 2019; Anon 2021).

Figure 3. The average spacing and width of strip roads within the time-study stand.

Figure 4. A view of the time-study stand after selective thinning.

The structure of work elements

Moving between working locations represented 10% of the productive machine time (Figure 5). Felling accumulation took most of the time of work as a whole, representing a 67% share, whereas the share of arranging the tree bunch in a pile was 23%. The absolute values were 1,800 s, 11,920 s, and 4,033 s, respectively. During the time studies, the accumulating felling head and base machine operated without any unnecessary breakdown delays.

Figure 5. Average proportion of working elements in the time study.

Modeling time consumption

At first, the best modeling approach was sought to predict the time consumption of the felling-bunching operation (Table 4). A good explanation (R² = 0.59) was obtained by using the average number of trees in the accumulating felling head as an independent variable (Table 4). However, the number of trees in the head cannot be directly monitored/obtained from the stand data, and another regression analysis for estimating the number of trees in the accumulating head (Table 5) with the stand parameters was therefore executed. In this, the average tree size of the removal resulted in good prediction power (R² = 0.57) (Table 5). The statistical criteria for accepting the regression models and their variables were that residuals were linear and systematically distributed against predicted values, and t-test p-values showed significance (p < 0.005) for each approved variable of the regression model. The productivity factors in the accumulating felling bunching of whole trees were therefore the average volume of harvested trees and the number of trees in the felling-bunching cycle (Figure 6 and Table 5). The time consumed (Table 4) in accumulating felling bunching per harvested whole tree (T_{Felling-bunching}) depended on the number of trees in the felling head (Table 5) in a felling-bunching cycle (Figure 7 and Table 4). Because the whole-tree volume affected the number of trees that fit the accumulating felling head, the number of trees cut per productive machine hour decreased as the whole trees’ average volume grew (Figures 6 and 7). There were between 5 and 10 trees in the accumulating felling head in a felling-bunching cycle, depending on the average volume of harvested whole trees (Figure 6).

Figure 6. The average number of trees (N_Number) harvested per felling-bunching cycle as a function of the average whole-tree volume (dm³) for each study plot (n = 16).

Figure 7. The time consumption of felling bunching (T_{Felling-bunching}) per whole tree harvested as a function of the number of trees per felling-bunching cycle.

According to the regression analysis, the removal intensity of harvested trees (1,517–7,367) per hectare was not a statistically significant independent variable for the time consumption of moving between working locations, although some reduction of time as a function of removal intensity was visually detectable (Figure 8). The moving time per tree (T_Moving) was thus defined to be the constant value (0.62 s per tree) generated from the whole time-study data of all study plots.

Figure 8. Time consumption of “moving between working locations” work element as a function of tree removal per hectare in felling bunching of whole trees. The average moving time per tree was 0.62 s.

Total time consumption and productivity

The total time consumption per harvested whole tree (T_Total) with the Risupeto II accumulating felling head in the time-study site was the sum of the two working elements (Eq.1)(1) $T_{Total}=T_{Moving}+T_{Felling-bunching}$(1) :

(1) $T_{Total}=T_{Moving}+T_{Felling-bunching}$(1)

The PMh productivity, expressed as the number of whole trees harvested per PMh, was converted to m³ by multiplying the number of harvested whole trees by the average volume of harvested whole trees (x).

A comparison of observed productivities in time-study plots on the regression model curve resulted in a good fit as a function of average tree volume (Figure 9). Accumulating felling-bunching productivity was sensitive to the average volume of harvested whole trees (Figure 9). While the average volume of whole trees doubled, from 14 to 28 dm³, the productivity increased by 54%.

Figure 9. The accumulating felling-bunching productivity of whole trees with the Risupeto II felling head as a function of the average volume of harvested whole trees.

Cost and profitability of harvesting whole trees

The harvesting costs for whole trees were below the average roadside prices for whole trees (Figure 10) when the cutting removal was 59–118 m³ per hectare, the average volume of harvested trees was 14–52 dm³, and the forwarding distance was 300 m. In the system analysis, the productivity of felling bunching was based on the observed time consumption on the time-study plots (the observed PMh was converted to SMh with a coefficient of 1.3).

Figure 10. The harvesting cost of whole trees at the roadside landing with the two-machine configuration by work phase when the cutting removal is 59–118 m³ per hectare, average volume of harvested trees is 14–52 dm³, and the forwarding distance is 300 m.

The profitability (i.e., the cost margin between harvesting costs and roadside prices) was at its lowest at €2. 30–€4. 40 per m³ and at its highest at €8. 40–€10.50 per m³. The total harvesting costs at the roadside landing were in a range of €14. 60–€20.70 per m³, and of this, the cost of felling bunching constituted €4.50–€10.80 per m³, and forwarding €9.00–€10.90 per m³ (Figure 10). Extending or shortening the forwarding distance by 150 m increased or decreased the harvesting costs by €1.60 per m³. Compared to the conventional forwarding cost, the use of a grapple saw increased forwarding costs by €1.90–€2.40 per m³ in the same harvesting conditions as in time-study plots 1–16.

Discussion

Due to the relatively small number of observations (16 time-study plots) from which the models were extrapolated, the results should only be applied to similar stand and harvesting conditions (similar tree volumes, density of trees, biomass removal, and non-frozen harvesting period in easy terrain conditions). The modeled productivity function of accumulating felling bunching of small-diameter whole trees was based on the output of one machine operator, one site, and one machine, which may result in biased productivities when comparing results from a larger productivity study of several operators, machines, sites, and conditions in the future. However, this was the first productivity study of a freshly developed harvesting solution to assess the magnitude and potential of such cutting-felling-bunching technology and harvesting methods. With the produced regression models and productivity figures, this study provides the first productivity estimates for an innovative multi-tree harvesting system in the presented site and harvesting conditions for different types of cost calculations and sensitivity analyses.

Several studies have shown that the operator has the most important impact on harvesting productivity (e.g. Sirén 1998; Ovaskainen 2009; Purfurst 2010; Palander et al. 2012b). There are differences in human factors, such as the operator’s motor skills, work planning and the decision-making process at the stand, which have an effect on productivity. Previous studies have also shown that productivity curves based on follow-up studies are lower than the productivity curves calculated based on time studies (Ryynänen and Rönkkö 2001; Sirén and Aaltio 2003). The reasons for this include the fact that time studies based on brief work on sample plots do not fully correspond to real-world work throughout the year. Accordingly, a long-term follow-up study would give a more reliable picture of productivity in practice, as well as of the functional and technical utilization rate of the base machine and accumulating felling head in variable stand conditions at different times of the year (Sirén 1998; Sirén and Aaltio 2003; Spinelli and Visser 2008; Purfurst 2010; Eriksson and Lindroos 2014).

To have a perspective of the potential of the studied harvesting concept, study results need to be compared with some previous studies of energy wood harvesting by accumulating felling heads designed for continuous cutting and the accumulation of whole trees (Figure 11). The potential of the Risupeto II accumulating felling head for harvesting whole trees in selective thinning was better than expected compared to the results of the study of the previous Risupeto prototype, which operated in clearing all the trees next to roadsides and field edges (Laitila and Väätäinen 2021). The comparison showed that the productivity of thinning was significantly higher than that of a harvester fitted with Bracke MAMA and C16 accumulating felling heads in selective thinning (Bergström and Di Fulvio 2014), and a little lower than in brushwood clearing (Laitila and Väätäinen 2021). The production potential of the Risupeto II is evident when considering the difference in difficulty between operating in the selective thinning and clear-cutting of whole trees (Jylhä and Bergström 2016).

Figure 11. Felling-bunching productivity as a function of the average volume of harvested whole trees compared to previous studies of the continuous cutting and accumulation of whole trees. *.Laitila and Väätäinen (2021), **Jylhä and Bergström (2016), *** Bergström and Di Fulvio (2014)

A video camera proved an appropriate data collecting tool and enabled the analysis of short working elements. The placing of the camera enabled the working elements in the continuous accumulating felling-bunching process to be fully recorded and the number of harvested trees accounted, with no interference in the operator’s work view. Compared to whole tree harvesting with conventional accumulating harvester heads or felling heads (Kärhä 2006; Laitila et al. 2016; Laitila and Väätäinen 2020; Bergström et al. 2022) the number of working elements was smaller because the boom movements were integrated with continuous cutting and accumulation of whole trees. In addition, trees were harvested as full-length bunches without having to cut them into shorter pieces during cutting.

In this study, a static Excel spreadsheet-based system analysis was applied without taking any stochasticity or interactions between felling bunching and forwarding operations into account in the system. There may be an imbalance between the productivities of felling bunching and forwarding in whole tree harvesting, and this may result in increased harvesting costs due to shift arrangements. In this study (Figure 10), the operating hour productivities of felling bunching and forwarding were 8.6–20.5 and 7.7–9.3 m³ per SMh. The imbalance was greatest for the largest whole-tree volumes (>30 dm³).

Overall, forwarding is significantly improved by accumulating felling bunching, because the harvested volume is concentrated in fewer locations, thus increasing grapple loads and thereby speeding up forwarding (Laitila et al. 2007; Erber et al. 2016; Läspä and Nurmi 2018). However, the use of a grapple saw for the crosscutting of whole trees during loading may increase the productivity imbalance of machines on the site if there is a need to crosscut bunches before loading. In the cost analyses, it was expected that the use of a grapple saw would decrease both the loading and unloading productivity. The first work phase was due to extra work for the crosscutting and re-bunching before loading, and the latter was due to a smaller grapple load.

From the forwarding perspective, tree sections should be as long as possible to maximize the forwarder’s load (Björheden 1997). The payload is one of the most important productivity and cost factors in forwarding and lengthening the forwarded wood (i.e., timber logs or whole trees) is the easiest way to increase its size. The maximization of the payload is facilitated if the felled trees are crosscut to 6–7 meters or even to 8 meters (Kärhä 2006). In this case, the strip road network must be carefully planned, and sharp turns avoided (Kärhä 2006). The disadvantages are that the excessive lengthening of forwarded trees complicates the mobility and loading work. Furthermore, the forwarders tend to become tail-heavy (Laitila et al. 2007). In selective thinning, the smallest trees are removed in the first instance, which reduces the need for crosscutting whole trees during loading. In this study, the estimated cutting removal mean height varied between 5.7 and 8.7 m in the time-study plots (Table 1). However, in some study plots, the length of some trees was as much as 12–13 m (for trees harvested from the strip road).

Time studies of the forwarding work with a grapple saw cannot be found in the literature, and for more reliable forwarder productivity results, it is important that practical work studies are established. However, one might reasonably expect the use of a grapple saw to reduce bucking time through its greater handling capacity with full tree-length bunches compared to crosscutting of single trees into forwarding lengths during the accumulating felling process. In practice, the Risupeto II could be used for the cutting of such a tall tree with the prerequisite to cut the standing tree in two phases: The top is cut first, and after felling, the head is moved vertically downward, and the tree is cut again at stump height.

In geometrical area-based cutting, all trees from specific areas are harvested in the one crane movement cycle and adjacent areas will stay untreated (Läspä and Nurmi 2018), which produces more heterogeneous stand structures than selective thinning (Nuutinen et al. 2021; Bergström et al. 2022). In the so-called boom-corridor thinning trees are harvested from narrow corridors, which favors accumulating felling heads that can continuously cut and accumulate trees during linear crane movement (Bergström 2009; Bergström et al. 2012; Bergström and Di Fulvio 2014). The boom-corridors are about 1–2 m wide and 10 m deep and the boom-corridors are systematically placed on both sides of the strip road (Läspä and Nurmi 2018). The boom-corridors have a fan-shaped orientation or perpendicular orientation to the strip roads (Läspä and Nurmi 2018).

In the conventional thinning, where the aim is to optimize the future production of high-value roundwood, the operator selects the trees to be thinned and the permanence of residual stand hinders crane movements, limiting the theoretical accumulation capacity of the harvester head (Belbo 2011; Petty 2014). Whereas in geometrical cutting as well as clear cutting all the biomass from a specific area is harvested systematically, there is no remaining tree stand which will hamper harvesting, crane movements are effective, and the time spent in tree selection can be avoided. Simulations of hypothetical harvester technology combined with boom-corridor thinning suggest productivity can be boosted by 40–200%, with the greatest effect being seen with continuous cutting and accumulation (Bergström et al. 2007; Sängstuvall 2018). In the most recent field study (Bergström et al. 2022), the boom-corridor thinning yielded on average 16% higher harvester productivity compared to selective thinning using the Bracke C16 accumulating felling head.

The management practice in the study was the selective thinning of the young stand from below to regulate the spacing of vital trees, as well as to improve the quality properties of the remaining stand and physical logging conditions of future cuttings (Äijälä et al. 2019). Harvesting conditions were challenging due to the small tree size, high tree density, and dense undergrowth. Geometrical cutting (Bergström et al. 2010, 2022; Sängstuvall et al. 2012; Ahnlund Ulvcrona et al. 2017; Läspä and Nurmi 2018) or clear-cutting of young dense stands for fuel (Jylhä and Bergström 2016) are not traditional Finnish forest management practices, and the above treatments were therefore excluded from the study.

Conclusions

In conclusion, the study’s results proved that the Risupeto II accumulating felling head provided sufficient profitability for economical thinning operations in small and over-dense young stands. It was noteworthy that profitability calculations did not include any subsidies used to boost forest management operations in Finland for energy wood and timber harvesting in young stands (Petty and Kärhä 2014). Risupeto II therefore has even greater potential to offer an economical solution for the treatment of untended young stands. Risupeto II provided to be a cost-effective harvesting solution to the problem of untended young stands with a dense undergrowth, because the harvesting costs can be covered by the revenue from energy wood sales in typical early thinning harvesting conditions. For example, Belbo (2011) lists strategies to reduce harvesting costs in small tree harvesting:

• Improving the accumulation and packaging capacity of the accumulating harvesting head.

• Employing a machine system with lower capital costs than dedicated forest machine systems.

• Continuous felling and bunching of trees during crane movement.

The Risupeto II accumulating felling head, which is mounted on the boom of the forestry equipped crawler excavator, meets all of the following criteria.

The main factors for the high felling-bunching productivity for the small-sized energy wood cutting were that the 1) accumulating felling head was capable of continuous felling-bunching whole trees without any unnecessary breakdown delays or stops while cutting trees, 2) trees were harvested as non-delimbed, which increased harvesting removal, 3) trees were harvested as full-length bunches without having to cut them into shorter pieces during cutting (crosscutting during forwarding), 4) the accumulating felling head was attached firmly to the boom tip of the crawler excavator (cf. mounting felling head on a swing damper) enabling precise control and continuous work, 5) the disk sawblades made of wear-resistant steel was an appropriate choice for work in tough conditions within the forest with lots of small stones in the soil and vegetation and 6) stand density was favorable for a continuous harvesting process with a relatively fast fill of the accumulating felling head while controlling the boom movement.

An efficient accumulation process is crucial for the productive harvesting of small trees, and continuous harvesting enabled both continuous and simultaneous cutting and accumulation without fixing each tree inside the felling head before cutting from stump. In addition, it was crucial that none of the recently felled trees dropped from the accumulation chamber of the felling head. The observed accumulation and packaging capacity of the accumulating felling head was relatively high by varying between 5 and 10 whole trees and 127–282 dm³.

Disclosure statement

No potential conflict of interest was reported by the author(s).