July, 1999, vol.10 no.2

Single- and Double-Grip Harvesters – Productive Measurements in Final Cutting of Shelterwood

Dan Glöde
The Forestry Research Institute of Sweden
Uppsala, Sweden

ABSTRACT

To compare the performance and cost of two machine types, a time study of single-grip harvesters (SGH) and double-grip harvesters (DGH) was conducted in the final cutting of three shelterwood stands in central parts of Sweden. A randomized block design was used with one block in each stand and the treatments SGH and DGH, respectively. The stands were characterized by dense to relatively dense advanced regeneration under approximately 200 trees/ha. No significant differences were found in mean harvesting time between SGH and DGH or between stands. Fewer trees, but approximately the same volume per hour, were harvested when shelterwood stands were cut as compared with clear–cutting of ordinary stands. Despite few stems per hectare and the dense regeneration hindering the operator’s field of view, final cutting was done with fairly high productivity (15.9–34.0 m3/E15-h) and a low harvesting cost (2.7 – 6.0 USD/m3). It was concluded that both machine types gave acceptable results regarding cost and productivity. The SGH could be recommended as a good choice in general due to the low cost per machine hour as compared with the DGH, while the properties of the DGH would be beneficial in shelterwood stands with a large proportion of trees with large diameter (e.g. >7 cm) branches.

Keywords: Time study, harvesting time, normative time, productivity, cost.

INTRODUCTION

Conventional forestry practice in Sweden has been to clear-cut, scarify and plant. However, the high cost of silviculture, together with demands from the general public and market concerning environmental considerations in forestry practices has led to increasing interest in natural regeneration [6]. One example of this is natural regeneration of Norway spruce (Picea abies, L. Karst) under shelterwood, i.e. under 160 – 400 stems per hectare [17].

The shelterwood system is an old silvicultural method that has been modified and revived in Sweden [20]. One modification is the use of harvesters instead of manual or motor-manual felling. Leikola [12] argued that the development of careful logging methods is one of the most important research tasks in the area of natural regeneration. Westerberg [20] presented a four-stage shelterwood system for regeneration of Norway spruce; (1) Preparatory cutting, (2) Seed cutting, (3) Removal cutting and (4) Final cutting. The harvesting technique used in the three first stages is similar to late thinning in conventional stands and is commonly done with single-grip harvesters (SGH). However, little research has been published on final cutting of shelterwood with SGHs or double-grip harvesters (DGH) since the method still is in the introductory stage. Hence, comparative studies of these machines are needed [7], [19].

Compared to conventional stands, shelterwood stands typically contain trees with a larger mean stem volume. This together with the advanced regeneration will affect harvester performance differently than in conventional stands. Studies of different harvesters made during clear–cutting of ordinary stands show the DGH to have a higher productivity than the SGH when trees exceed 0.5–0.8 m3/tree [8], [9]. This would suggest that the DGH should be more productive in final cutting of shelterwood as the mean stem volume often is high in these stands. However, Westerberg & Berg [21] found that productivity was higher for a SGH than for a DGH in final cutting of a shelterwood with 200 Scots pine trees (Pinus sylvestris) per hectare. Mäkelä [14] summarized the results from two studies of final cutting of shelterwood with SGHs and DGHs [13], [18] and concluded that SGHs were best suited for removal of seed- and shelter trees. Mäkelä [15] also concluded that the DGH was the most expensive method when removing seed- and shelter trees in both studies.

As part of the development towards more diversified silvicultural methods in Sweden [20] the use of seed–tree and shelterwood systems increased markedly to 40% of the regenerated area in 1993, and thereafter declined somewhat [4]. Some of the shelterwoods created in the early 1990s will soon mature for final cutting. Hence, it was important to compare SGHs and DGHs in final cutting of shelterwood to identify potential difficulties, needs for development and differences in harvesting time and work pattern.

The aim of this study was to compare the productivity and cost of SGHs and DGHs in final cutting of shelterwood. The assumption was that the DGH, although more productive, would turn out to be a more expensive alternative due to its higher operational cost. Consequently, the hypothesis to be tested was that there is no difference between the SGH and the DGH in harvesting time at final cutting of shelterwood (Table 1).

MATERIAL AND METHODS

Experimental sites and design

A randomized block design with three blocks and two treatments was used. Three experimental sites, each used for one block, were selected for the study. The sites were located in Sweden at Avesta (lat. 60°09’N) at Graninge (lat. 63°06’N) and at Lövsjön (lat. 63°01’N). The treatments included final cutting of shelterwood with a SGH and a DGH in each block. The size of the block was 2.0 hectares at Avesta, 3.8 hectares at Graninge and 2.1 hectares at Lövsjön. The ground had high bearing capacity at all sites. The terrain was flat and the surface was even at Avesta and Graninge while there was a moderate slope (11-18°) and a somewhat uneven ground surface at Lövsjön.

Within each block, the centres for circular plots were marked with aluminum sticks in a grid system of 10×10 m at Avesta and 12×12 m at Graninge and Lövsjön. Each centre was used for circular plots with a radius of 1 m in which the number of plants and plant heights were registered (Table 1). Some of the centres were also used for plots with a radius of 10–20 m, where various properties of shelterwood trees were measured (Table 1).

Time studies

The block at each test site was split into two treatment units and in such a way that the stand properties of the two was as similar as possible (Table 1). Each treatment unit was formed to allow the harvester to work normally without interfering with the adjacent unit. The length or width of the treatment units was not shorter than twice the length of the average tree. Treatment was randomized between the two units

Table 1. Shelterwood stand data and regeneration stand data before final felling.

The weather conditions were fairly similar during the studies, except for the deep snow cover at Lövsjön (Table 2). The diameter at east height (DBH) was measured and marked on all trees in the treatment units. Continuous time studies (cmin) using SIWORK3 [1] software and a Husky Hunter computer were made for defined work cycle elements on each harvester (Table 3). All element times were measured as effective times (E0) [2]. For each processed tree, the DBH and tree species were registered. The volume of the measured trees was calculated according to Näslund [16] using approximately 30 height measured trees per block. All volumes were calculated as solid volume under bark (m³). The travelling distance was registered for each driving occasion with the help of a measuring line or measuring stick. The working width was measured at every 20 m along the striproads as the distance between marked stumps on one side of the striproad perpendicular to the other side. The area within the working width was calculated as working width multiplied with driven distance divided by the treatment unit area.

Machines and operators

The harvesting at Avesta was done with a Valmet 892/960 SGH and a FMG/ÖSA 706/260 DGH at a cost of 90.4 and 92.6 USD/E15-h, respectively. The harvesting at Graninge and Lövsjön was done with the same operators using a Valmet 892/955 SGH and a Timberjack 1880 Master DGH to a cost of 90.4 and 102.3 USD/E15-h, respectively. Machine costs were chosen as prevailing (Mellanskog Forest Owner Ass.) reimbursement to contractors and harvester types. All operators were experienced and highly skilled. The operators were instructed to work at their normal sustainable pace and to minimize damage in the advanced regeneration. To promote future timber production, the felling instructions were to drive and fell in blank spots of the regeneration. However, if the operator observed single plants in the otherwise blank spots, then he should drive and fell in denser parts of the regeneration.

Table 2. Weather conditions during the time studies.

Table 3. Definitions of the work cycle elements for the single-grip harvester (*) and for the double-grip harvester (#).

Calculations and statistical analysis

For the SGH the felling + processing time (y=a+b*x) was calculated by linear regression analysis as the function of stem volume (x) and tree species. For the DGH the felling + boom–in + processing 1+2 time was calculated by linear regression analysis as the function of stem volume and tree species.

Differences in mean stem volume and in mixture of tree species between treatments were equalized by merging stand data from the two treatments in each block to common (normative) mean values per block.

The normative felling and processing time were calculated with consideration to the effect of tree species and mean stem volume on the elements felling and processing for the SGH, and felling, boom–in and processing 1+2 for the DGH. The time functions per tree species were weighted to one time function per treatment (y=a+ b*x). The weighting was done on the basis of each tree species’ proportion of the harvested volume per block. These functions were used to calculate harvesting time at an equal mean stem volume per block.

Normative travel time was calculated based on the number of stems per hectare. For each treatment the travel distance per hectare was multiplied by the area of the block and divided by the speed, resulting in travel time per hectare. Driven time per hectare was then divided by the mean number of stems per hectare and block, resulting in normative driven time in cmin/tree.

The elements boom-out, start/wait and halt are most likely affected by the number of stems per hectare in the shelterwood and the size, sort and amount of advanced regeneration, but no reliable normative factors or functions were found. The elements disturbance and miscellaneous were included in productive time because they might reveal differences in the harvester’s capacity to work in extensive regeneration.

The sum of all elements except felling and processing for the SGH and felling, boom–in and processing 1+2 for the DGH were added as a constant value to the intercept (a) in the normative time functions for the SGHs and the DGHs resulting in a curve for total time per tree over varying mean stem volume. The factor of 0.71 was used as the relation between productivity per effective time (m³/E0-h) compared with per gross effective time (m³/E15-h). The latter includes effective time with the addition of delay times shorter than 15 minutes [2].

An analysis of variance using Microsoft Excel [3] was carried out. If p<0.05, the result of the statistical analysis was called significant. The following model was used to test differences in total normative harvesting time per tree between SGH and DGH:

Y_ij = µ

The SGH at Lövsjön had the shortest driven distance, the least working width, i.e. area within reach of boom, and the largest number of harvested trees per set–up place (Table 6). The deep snow at Lövsjön made travel speed low for the DGH as compared with the SGH (Table 6).

The difference in productivity between the two harvesters was largest at Avesta where the SGH had approximately 20% lower mean productivity than the DGH (Figure 1, Table 6). However, the SGHs mean productivity at Graninge and Lövsjön was approximately 9 and 12% higher, respectively, as compared with the DGH (Figure 1, Table 6).

At Avesta the SGH spent 30% of the total time on the elements disturbance, start, and driving (Table 6). The average stem took approximately 22% longer time for the SGH to process than for the DGH at Avesta. The difference in harvesting time between the two harvesters at Graninge and Lövsjön was smaller than at Avesta (Table 6). The SGH had a lower harvesting time, and higher productivity, than the DGH at Graninge and Lövsjön (Table 6). The DGH at Graninge and Lövsjön used 4–6 times as much time on the element miscellaneous as compared with the mean time for miscellaneous in the other treatments.

yij = m + ui + tj + eij where,

yij = total normative time consumption per tree in treatment ij

m = m total mean

ui = fixed effect of block i, where i = 1–3

tj = fixed effect of treatment j, where j = 1–2

eij = residual effect in block i, treatment j

(0, se2)

RESULTS

There was no significant difference between SGH and DGH or between blocks in mean harvesting time according to the analysis of variance (Table 4).

Most stems per hectare were harvested in the spruce dominated shelterwood at Lövsjön and the largest trees were harvested in the pine dominated shelterwood at Graninge (Table 5).

Table 4. Analysis of variance for normative harvesting time for the single- and the double-grip harvesters in final cutting of the shelterwood.

Table 5. Stand data as means of harvest per treatment unit and normative stand data as means per block

The sum of all elements except felling and processing for the SGH and felling, boom–in and processing 1+2 for the DGH were added as a constant value to the intercept (a) in the normative time functions for the SGHs and the DGHs resulting in a curve for total time per tree over varying mean stem volume. The factor of 0.71 was used as the relation between productivity per effective time (m³/E0-h) compared with per gross effective time (m³/E15-h). The latter includes effective time with the addition of delay times shorter than 15 minutes [2].

An analysis of variance using Microsoft Excel [3] was carried out. If p<0.05, the result of the statistical analysis was called significant. The following model was used to test differences in total normative harvesting time per tree between SGH and DGH:

The SGH at Lövsjön had the shortest driven distance, the least working width, i.e. area within reach of boom, and the largest number of harvested trees per set–up place (Table 6). The deep snow at Lövsjön made travel speed low for the DGH as compared with the SGH (Table 6).

The difference in productivity between the two harvesters was largest at Avesta where the SGH had approximately 20% lower mean productivity than the DGH (Figure 1, Table 6). However, the SGHs mean productivity at Graninge and Lövsjön was approximately 9 and 12% higher, respectively, as compared with the DGH (Figure 1, Table 6).

At Avesta the SGH spent 30% of the total time on the elements disturbance, start, and driving (Table 6). The average stem took approximately 22% longer time for the SGH to process than for the DGH at Avesta. The difference in harvesting time between the two harvesters at Graninge and Lövsjön was smaller than at Avesta (Table 6). The SGH had a lower harvesting time, and higher productivity, than the DGH at Graninge and Lövsjön (Table 6). The DGH at Graninge and Lövsjön used 4–6 times as much time on the element miscellaneous as compared with the mean time for miscellaneous in the other treatments.

Table 6. Working pattern, harvesting time and productivity.

The regression coefficient (b) was generally lower for the DGH than for the SGH (Table 7). The harvesting time curves at Graninge had a flatter slope than at Avesta and Lövsjön (Table 7, Figure 1).

The harvesting costs for the SGH and the DGH, respectively were 5.7 and 4.7 USD/ m³ at Avesta, 2.7 and 3.3 USD/m³ at Graninge and 4.8 and 6.0 USD/m³ at Lövsjön.

DISCUSSION

Experimental design

The aim of this study was to compare productivity between two types of harvesters in a new silvicultural method under equal and normal working conditions. This demanded large treatment units which made it difficult to create equal stand properties. The problem was solved by calculating normative stand values. However, there is also the risk for differences in operator performance. Bergstrand [7] discussed the accuracy of time studies on forest technology and the problem with difference in operator performance, which can result in a 20–50% variation in machine productivity. He pointed out that it would be necessary to include approximately 400 operators in order to reach a confidence level of 95%, which he concluded to be economically unfeasible. Appelroth [5] argued that analyses of differences between work methods using the comparative time study are incorrect unless work rating is done, because the rate of working can vary unpredictably between zero and maximum for a given operator.

Furthermore, if the rate of working is much below the maximum, the output can not be used for comparison of methods or machines, only for comparing the operators. Appelroth [5] suggested that the researcher doing time studies also should do the rating. However, subjective performance rating may be used with an acceptable result under easy conditions, like manual planting, but is doubtful in other operations with more complex sequences of work elements [19]. Bergstrand [7] considered performance rating as very difficult and concluded that the only strategy that could give statistically reliable results from forest technology studies is replication, which we usually can not afford. All the above arguments are relevant and hence, the result in this study must be considered with the human factor, the operator, in mind. However, Bergstrand [7] concludes that this does not exclude comparative studies from resulting in indications and guidance for development of new products and methods, which corresponds with the aim of this study.

Table 7.Time functions per treatment and tree species.

^a Treatment SGH in Lövsjön consisted of 4% of the total volume of pine but the pine trees were too few to base any functions upon.

Analysis of variance

The large mean stem volume in shelterwood stands makes it reasonable to assume that the DGH should have a higher productivity than the SGH but also higher harvesting cost due to higher machine cost. However, the productivity of the DGH was not found to be higher than of the SGH. Hence, the null hypothesis, i.e. no difference in time consumption per processed tree between blocks and treatments, could not be rejected. Differences between the treatments and blocks when dividing the time consumption in volume dependent elements (e.g. processing and felling) and not volume dependent elements (e.g. boom–out, driving, miscellaneous, etc) was also examined using the same model as in Table 4. No significant differences were found between SGH and DGH (p=0.870) and blocks (p=0.750) in time consumption for non-volume dependent elements. Hence, there do not seem to be any large differences between SGHs and DGHs concerning boom work and driving in the shelterwood stands. In time consumption for volume dependent elements, statistical differences were found between blocks (p=0.034) but not between SGH and DGH (p=0.166). The significant differences between blocks are not surprising since there were differences in mean stem volume.

Harvesting cost

With no statistically proven difference in harvesting time and productivity between the two harvesters, the SGH would be the first choice, from an economical point of view, due to a lower machine cost per hour. This was also the case at Graninge and Lövsjön, with Avesta as the exception. However, it is important to remember that stand properties will affect maintenance costs. If a SGH regularly works in stands with a high mean stem volume and large diameter branches, as at Avesta, it is likely that the maintenance cost would rise, resulting in a machine cost closer to the DGH.

Perhaps the only situation where the harvesting cost of the DGH is lower than that of the SGH is where the shelterwood stand consists of trees with large diameter branches, as at Avesta. In stands with few, easily delimbed, trees per hectare the stronger processing unit of the DGH is not an advantage. Where there are few trees per hectare there are also fewer opportunities for the DGH to simultaneously process one tree while felling another. Hence, the DGH, as compared with the SGH, probably reduces its productivity more when converting from final felling of ordinary stands to shelterwood stands, provided that mean stem volumes are equal.

Time functions and productivity

The coefficient of the slope (b) was generally lower for the DGH which indicates that it is less sensitive to variation in tree volume than the SGH. This is in line with experiences from earlier studies [8], [9]. The size of b is also generally lower than given by Brunberg [8] where b=0.107 for the SGH and 0.078 for the DGH. The lower regression coefficient in this study may be explained partly by technical development and partly by less variation in tree volume in a shelterwood than in an ordinary stand.

The purpose for transformation from effective (E0) to gross effective (E15) time is to give an idea of an approximate level of productivity in practice. The relation 0.71 between m³/E0-h compared with m³/E15-h, is chosen in this study because it is an approximation based on experience and was used in earlier machine studies in shelterwood [21]. A comparison between E0- and E15-time was made by Brunberg [10] who found the relation to be 0.64 – 0.68 depending on whether the element miscellaneous was included or not. Hence, the productivity presented in m³/E15-h might be over-estimated.

Between the three blocks the average productivity was highest at Graninge for both harvesters. This is probably explained by easily delimbed pine trees, the mean stem volume being 60 to 80% larger and the harvested volume 35 to 45% higher at Graninge than at Avesta and Lövsjön.

Harvesting time

Differences in harvesting time between the two harvesters at Avesta occurred mainly in the elements start/wait, halt, driving and disturbance. A magnetic valve for the tilt–up function of the harvester head on the SGH malfunctioned and was replaced. This problem may have contributed to the time start/wait and halt. The SGH also had problems with delimbing and the main part of the disturbance time occurred during processing. Mean diameter of the branches causing problems while delimbing was measured in a parallel pilot study. The result showed that the SGH had problems when delimbing branches with a mean diameter of 7 cm for pine and 8 cm for birch. Corresponding branches causing problems for the DGH had 2 cm larger diameter. Brunberg [11] found that the SGHs Timberjack 1270/762B and Valmet 911/960 had delimbing problems when branches had a mean diameter of 7.2 and 7.0 cm, respectively, as compared with 9.5 cm thick branches for the DGH Timberjack 1880. Hence, the delimbing problems for the SGH at Avesta might be due large diameter branches.

The fact that the DGH in Graninge had a higher total harvesting time than the SGH, is surprising considering the large mean stem volume. However, the pines at Graninge had small green crowns and thin branches which might have favoured the SGH since the DGH was unable to use the advantage of its powerful processing unit and, hence, could not gain time on the SGH when processing trees with thick branches. The thin branches might have contributed to the time consumption curves having a flatter slope than in other studies, e.g. Brunberg [8], [9]. A lot of the time which the DGH in Graninge spent on the element miscellaneous consisted of clearing the field of view with the boom before felling, which probably contributed to a higher time consumption than for the SGH.

The shelterwood at Graninge had a larger mean stem volume and standing volume per hectare as compared with the shelterwood at Avesta. However, the regeneration was by no means as dense at Graninge as at Avesta. Consequently, the time spent on start, halt and driving could be expected to be lower at Graninge than at Avesta, but they were of equal size. This could be due to the fact that there were fewer trees per hectare at Graninge and that the number of stems per hectare had a larger effect on travelling time and its share of the total time than volume per hectare.

The SGH at Lövsjön encountered more problems in the deep snow than the DGH when felling, sometimes having to “shovel“ away snow from around the trees with the harvester head before they could be felled. The trees were also full of cold thin snow which frequently fell down during positioning, obstructing the operator’s field of view for a moment. This probably contributed to the increased time for felling as compared with the DGH.

The fact that the curves of harvesting time for the SGH and the DGH in Lövsjön do not cross each other until very large mean stem volumes is probably due to difficulties with driving in deep snow. When comparing time for travelling and travel speed between the two machines it is obvious that the DGH had difficulties. Both machines had chains on the front wheels and belts on the bogies, but the SGH had flat (5×5 cm) steel pegs welded on the belts, which increased its mobility. Another explanation of the shorter time spent on driving and more trees per processing site for the SGH could be that the trees were standing in groups while they were more scattered on the DGHs part of the block.

Normative harvesting time

Normative time was calculated for factors concerning both harvester types. In addition, normative time of the elements disturbance at Avesta and felling, miscellaneous and driving at Lövsjön was calculated. This was done by reducing the element disturbance for the SGH at Avesta to the mean of the same element of the other treatments. Furthermore by reducing felling, for the SGH at Lövsjön, to the same relative level as was the case between the harvester types at Graninge. It was also done by reducing the miscellaneous and driving for the DGH at Lövjön to the mean of the other treatments. This changed the mean time consumption for the SGH and the DGH to approximately 130 and 127 cmin/tree, respectively. However, we could still not reject the null hypothesis since the p–values were 0.206 and 0.700 for cause of variance in block and treatment, respectively.

CONCLUSIONS

There were differences in productivity between the SGH and the DGH, mainly attributed to the structure of the shelterwood stands. Both harvesters could be a good choice depending on stand data, alternative logging and logistics. The SGH could be recommended as a good choice in general, due to the low cost per machine hour as compared with the DGH, while the properties of the DGH would be beneficial in shelterwood stands with a large proportion of trees with large diameter (e.g >7 cm) branches.

Despite the few stems per hectare and dense regeneration obstructing the view of the operator, the costs are considered low for final cutting of the shelterwood compared to clearcutting of ordinary stands. Fewer trees but approximately the same volume per hour was processed as in an ordinary stand, which resulted in similar costs.

ACKNOWLEDGEMENTS

I would like to thank Prof. I. Wästerlund, Dept. of Operational Efficiency, Prof. L.Sennerby-Forsse, and a number of colleagues at the Forestry Research Institute of Sweden: S.Berg, S.Frohm, U.Hallonborg, H-Ö.Nohrstedt, J.Sonesson, M.Thor and Å.Thorsén, for valuable comments on the manuscript. I also wish to acknowledge the operators and the personnel at Graninge Skog & Trä and at the Forest Owners Associations Mellanskog and Norrskog for their cooperation and assistance during the studies. I also thank former Project leader Svante Scherman for encouraging an ambitious experimental design.

REFERENCES

[1] Anon. 1988. SIWORK3, version 1.1. Work study and field data collection system based on Husky Hunter handheld computer. Danish Forest and Landscape Research Institute. Skovbrynet 16, Dk–2800, Lyngby.

[2] Anon. 1978. Forest work study nomenclature. Norwegian Forest Research Institute, Ås. pp. 83-99. ISBN 82-7169-210-0.

[3] Anon. 1996. Microsoft Excel for Windows 95. Version 7.0a. Analysis ToolPak. Copyright 1985 – 1996. Microsoft Corporation.

[4] Anon. 1997. Forestry in Sweden. The Forestry Research Institute of Sweden. Uppsala. ISBN 91 7614 088 1.

[5] Appelroth, S-E. 1989. The Analysis and Interpretation of Forest Work Study Results. 8 p. In: Sutton, R.F. and Riley, L.F. (eds.). Proceedings of a Symposium on the Equipment/- Silviculture Interface in Stand Establishment Research and Operations. Forestry Canada, Ontario Region. Minister of Supply and Services Canada 1989. Catalogue No. Fo. 46–14/401E. ISBN 0–662–17513–1.

[6] Berg, S. 1995. Felling of Standards in Shelterwood Systems. In Heding, N. (ed.) Forestry operations in Multiple – Use Forestry – A NSR Project. pp. 47–60. Danish Forest and Landscape Research Institute. Lyngby. ISBN 87–89822–54–4.

[7] Bergstrand, K-G. 1987. Planning and Analysis of Time Studies on Forest Technology. The Forest Operations Institute of Sweden. Meddelande no 17. 58 p. Uppsala. (In Swedish with English Summary).

[8] brunberg, T. 1988. Instruments for Setting Productivity Norms for Harvesters in Final Fellings. The Forest Operations Institute of Sweden. Redogörelse no. 4. 1988. 20 p. Uppsala. (In Swedish with English Summary).

[9] brunberg, T. 1992. Produktionsnorms-resultat från studier av skördare i slutavverkning med automatisk datainsamling. The Forestry Research Institute of Sweden. Stencil no 1992–11–13. 16 p. Uppsala. (In Swedish).

[10] brunberg, T. 1995. Basic data for productivity norms for heavy–duty single–grip harvesters in final fellings. The Forestry Research Institute of Sweden. Redogörelse no. 7. Uppsala. (In Swedish with English Summary).

[11] brunberg, T. and Westerlund, M. 1994. Studie av Timberjack 1270, Timberjack 1880, och Valmet 911 vid Djupadal, SCA Skog AB. The Forestry Research Institute of Sweden. Stencil no. 1994–12–07. Uppsala. (In Swedish).

[12] Leikola, M. 1982. Natural regeneration of coniferous forest. Tidsskrift for Skogbruk. Vol. 90. Häfte 1. pp. 114–121. Norway. (In Swedish with English summary).

[13] Mäkelä, M. 1990. The removal of seed trees and shelterwood using Pika 75 and FMG 707/12 s harvesters. Metsätehon Katsaus; no. 19. 4 p. Helsinki, Finland. (In Finnish with English summary).

[14] Mäkelä, M. 1992. Harvesting technique for removal of seed and shelter trees. Metsätehon Katsaus; no. 6. 6 p. Helsinki, Finland. (In Finnish with English summary).

[15] Mäkelä, M. 1995. Timber Harvesting in Conjunction with Natural Regeneration. In Heding, N. (ed.) Forestry operations in Multiple–Use Forestry – A NSR Project. pp. 63–69. Danish Forestry and Landscape Institute. Lyngby. ISBN 87–89822–54–4.

[16] Näslund, M. 1940. Funktioner och tabeller för kubering av stående träd – tall, gran och björk i norra Sverige. Meddel. från Statens Skogsför- söksanstalt, Häfte no. 22, pp. 87–132. Stockholm. (In Swedish).

[17] Opsahl, W. 1950. Barskogen. Pleie og naturlig foryngring. 237 p. 1950. Oslo, Norway. (In Norwegian).

[18] Peltoniemi, T. 1991. Removal of seed trees and shelter trees mechanically, manually and by using a combination of both. Metsätehon Katsaus; no. 18. 4 p. Helsinki, Finland. (In Finnish with English summary).

[19] Samset, I. 1990. Some observations on time and performance studies in forestry. Communications of the Norwegian Forest Research Institute. no. 43.5. 80 p. 1990. Ås.

[20] Westerberg, D. 1995. Profitable Forestry Methods – Maintaining Biodiversity as an Integral Part of Swedish forestry. 5 p. In: Bamsey, C.R. (ed). Innovative Silviculture Systems in Boreal Forests. Clear Lake Ltd., Edmonton, Alberta, Canada. ISBN 0–9695385–3–7.

[21] Westerberg, D. and Berg, S. 1994. Felling of standards: trial method for determining productivity, costs and damage to advance growth. The Forestry Research Institute of Sweden. Redogörelse no. 10. Uppsala. (In Swedish with English summary).