Professional Documents
Culture Documents
01-Harvester Time Consumption in Nature Conservation Management Operations
01-Harvester Time Consumption in Nature Conservation Management Operations
To cite this article: Raul Fernandez-Lacruz, Örjan Grönlund, Tomas Johannesson, Line B.
Djupström, Jon Söderberg & Lars Eliasson (2023) Harvester time consumption in nature
conservation management operations, International Journal of Forest Engineering, 34:2,
112-116, DOI: 10.1080/14942119.2023.2174353
CONTACT Raul Fernandez-Lacruz raul.fernandezlacruz@skogforsk.se The Forestry Research Institute of Sweden (Skogforsk), Uppsala, Sweden
© 2023 The Author(s). Published with license by Taylor & Francis Group, LLC.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/),
which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.
INTERNATIONAL JOURNAL OF FOREST ENGINEERING 113
The cost for an NCM operation is proportional to the time harvester production (HPR) files, the harvest plan, follow-up
consumption of a harvester and forwarder to complete the task. inventory, and a shapefile with the stand borders were pro
NCM harvesting operations are similar to conventional thin vided. Aerial images obtained from “Basic Forest Data”
ning in cut-to-length forestry, but more time is required for (Nilsson et al. 2017) and in-field visits during April–
planning and harvesting (Grönlund et al. 2020). Time con May 2022 helped to complete stand descriptions and provided
sumption of final fellings and thinnings in Sweden has been an understanding of the aim of the NCM. For the control
studied by, e.g., Eriksson and Lindroos (2014), while few stu stands only the HPR files were collected. These stands were
dies have examined time consumption, costs, and revenues in cut in the same years, using the same fleet of harvesters and the
NCM operations. Santaniello et al. (2016) found that harvester same group of contractors as the NCM stands.
time consumption increased with levels of retention. Nordén NCM stands were selected from stands treated in 2018 and
et al. (2019) concluded that NCM in oak forests (mainly 2019 in Jämtland and Västernorrland and classified as NCM in
removal of spruce) was costly, and that both revenues and SCA Skog AB’s stand database. Harvester data was available
costs varied greatly. from 30 stands, but this data was only partial for 15 stands and
The aim of this study was to compare harvester time con one stand was excluded as it did not fit the NCM classification.
sumption in NCM operations to final felling operations in the Complete harvester data was therefore obtained for 14 stands
same geographical region and using the same fleet of harvest (Table 1).
ers. The overarching objective was to contribute knowledge NCM stands were classified into four groups according to the
needed to increase NCM operations to the extent needed to stated management goals. This classification relied on quantita
avoid failing conservation values. tive and qualitative indicators describing the stand, such as
initial stand properties, intensity of removal, tree species com
position, ground moisture, logging conditions, aerial images,
Materials and methods
and instructions in the harvest plan. The four groups were:
Study sites
(1) Broad-leaved forest (BF). Removal of spruce trees to
Felling in NCM stands and in ordinary final felling stands
improve the survival, growth, and reestablishment of
(control) was studied in areas harvested by SCA Skog AB in
shade-intolerant broad-leaved pioneer species.
Jämtland and Västernorrland (northern Sweden). Harvester
(2) Mixed conifer-broad-leaved forest (CBF). Variable den
data from 14 stands harvested for NCM purposes were col
sity thinning, especially removal of spruce, to benefit
lected, and data from 399 conventional final felling stands were
pine and broad-leaved trees, and to increase structural
collected as control.
complexity by felling trees to create structural hetero
NCM stands were felled during May–June 2018 and
geneity and other habitat features.
January–August 2019. Besides machine data in the form of
Removal d e
Total Mean
harvested stem Species
volume volume DBHb mixc
Size Age
m3 Harvester Management
Stand ha years ha−1 %a m3 cm % p/s/b size goal Remarks
1 6.0 - 167 63 0.17 16 7/84/9 XL BF Part of a swamp forest area
2 14.8 111– 246 72 0.24 17 8/74/18 XL BF -
120
3 0.9 61–70 88 46 0.18 17 11/77/12 L BF -
4 3.0 121+ 98 48 0.29 19 47/49/4 XL CBF Adjacent to a woodland key habitat
5 7.1 121+ 41 36 0.27 21 1/95/4 L CBF
6 7.2 111– 30 27 0.14 16 100/0/0 XL PF Adjacent to a woodland key habitat
120
7 6.4 121+ 83 26 0.17 16 7/90/3 XL PF Ancient remains (trapping pits, charcoal pits and house
foundations)
8 8.6 121+ 97 43 0.17 17 21/74/5 XL PF -
9 3.3 121+ 32 47 0.11 16 45/47/8 L PF -
10 2.7 121+ 81 37 0.11 13 28/68/4 L PF Adjacent to an area with high nature value
11 8.0 121+ 39 24 0.27 21 92/8/0 XL PF Adjacent to a woodland key habitat
12 1.4 121+ 77 29 0.25 18 28/59/13 XL PF Ancient remains (trapping pits)
13 0.8 - 126 57 0.18 17 0/83/17 L CF -
14 4.0 111– 100 33 0.28 20 0/94/6 L CF -
120
a Percentage of initial volume harvested.
b Mean DBH = diameter at breast height, i.e. 1.3 m above ground level.
c Percentage of removed trees by species: p, Scots pine (Pinus sylvestris), s, Norway spruce (Picea abies), b, Broad-leaved species.
d Size classification of harvester (harvester head in parentheses): large (L) denotes Komatsu 901TX (C93), Komatsu 931 (C123 or 365), John Deere 1170 (H413) or John
Deere 1270 (H415); extra-large (XL) denotes Komatsu 951 (C144) or John Deere 1470 (H415).
e BF: Broad-leaved forest, CBF: Mixed conifer-broad-leaved forest, PF: Pine forest, CF: Mixed conifer forest.
114 R. FERNANDEZ-LACRUZ ET AL.
(3) Pine forest (PF). Removal of spruce and variable density The agreement with the data provider (SCA Skog AB) pre
thinning and/or canopy opening to create structural het vented the publishing of any time consumption or productivity
erogeneity. The management goal is to create a sun- figures, so the results are presented as time differences between
exposed pine forest with late-successional habitat features, harvester sizes and relative time consumption for the different
e.g., cutting high stumps and partly removing bark along NCM treatments compared to conventional final felling.
the stem (girdling), to create dying and dead trees.
(4) Mixed conifer forest (CF). Thinning of varying intensity
and creation of dead wood to increase structural com Results
plexity, e.g., greater age range, multi-layered canopies,
Harvesting intensity varied somewhat but was generally
and other habitat features.
between 25% and 50% of the standing volume in the studied
stands. Regardless of NCM management goal, the main objec
The conventional final felling stands can be placed in a fifth
tive of the harvesting operations seems to have been removal of
group of management goals, Commercial management (CM).
spruce from the stands (Table 1).
Final felling is just one of the harvesting treatments in commer
In CM final felling operations, XL harvesters were 6.8 s per
cial management stands; all stands are also usually commercially
tree faster than large harvesters. This difference was marginally
thinned one or more times. However, time consumption in final
larger, although not significantly so, at 7.2 s for NCM felling
felling operations of commercially managed stands was chosen
operations. Mean stem volume, treatment, and machine size
as control to the time consumption in NCM treatments.
were found to significantly affect mean time consumption per
For each stand, harvester data on tree species, stem volume,
tree (Table 2). Compared to CM final felling, the average time
and time consumption (seconds, s) for each tree were automa
consumption of NCM felling was 15%, 15%, 19%, and 27%
tically registered by the harvester computer and collected in
greater for the management goals PF, CF, CBF, and BF, respec
time-stamped HPR files according to the StanForD 2010 stan
tively, and differences were deemed significant for PF, CBF,
dard (Möller et al. 2013; Möller 2014). This necessitated filtering
and BF. When comparing between NCM treatments, there
the data to remove trees harvested after a longer break or when
were a significant difference in time consumption between PF
a delay had occurred during the harvest; here, this filtering
and BF but not between any other treatment combination.
involved removing all trees with a processing time equal to or
Relative differences in time consumption per tree between
longer than 600 s (cf. Eliasson et al. 2020). Mean time consump
NCM fellings and final fellings are presented in Figure 1 for the
tion (s tree−1) and mean stem volume (m3 under bark) per tree
range of harvested mean steam volume observed in NCM stands,
were calculated for all trees with a processing time shorter than
i.e., 0.1–0.4 m3 (under bark). NCM felling operations had a relative
or equal to 600 s. Due to the limited number of stands in NCM
time consumption of 1.1–1.4 times greater than final felling opera
operations, these arithmetic means were calculated per workday
tions, depending on management goal, machine size, and mean
(n = 51). In conventional felling, splitting the data into days was
stem volume. These differences decreased with mean stem volume
not considered necessary, and means were calculated per stand
and were found to be smaller for large harvesters than XL harvest
(n = 399). The dataset of NCM fellings comprised 8,134 m3 and
ers if the same stem size was considered. The differences between
39,585 trees, while the dataset of final fellings consisted of
harvester sizes are partly a mathematical effect due to the fact that
~0.5 million m3 and 2.5 million trees.
XL harvesters were faster than large harvesters, which reduces both
the numerator and denominator.
As harvesting costs are expected to be proportional to time
Analyses consumed, NCM treatments were 10%–40% more costly than
Harvester time consumption was analyzed using a general final felling for stands with an average tree volume from 0.1 to
linear model, formulated in Eq. 1: 0.4 m3. The highest cost increases occurred for the “Broad-
leaved forest” (BF) NCM treatment, even though these stands
y¼ μ þTþMþβ�vþε (1) had high harvesting intensities.
where y (mean time per tree) was the response variable, µ was the
overall mean, T was treatment, i.e., “final felling” in CM stands or
Discussion
“management goal” in NCM stands and M machine size were
factors, β was a constant, v (mean stem volume) was a covariate, Removal of spruce was the most common measure found in
and ε was the residual error. The material was unbalanced, so the the harvest plans, despite different NCM goals. This is in
interaction T × M had to be excluded from the analysis. accordance with the findings of Grönlund et al. (2020), and it
Statistical analyses were conducted using SAS Enterprise is considered uncontroversial. Other measures consisted of
Guide 8.3 (SAS Institute Inc., USA) and results were regarded creation of lying and standing deadwood (cutting of high
as significant if the p-value was <0.05. stumps, girdling of standing trees with the purpose of
Table 2. ANOVA table for the general linear model explaining harvester time consumption per tree in the NCM stands. The
model explained 42% of the variation.
Source DF Type III SS Mean Square F Value Pr > F
Mean stem volume (v) 1 8306.00 8306.00 211.32 <0.0001
Machine size (M) 1 4232.91 4232.91 107.69 <0.0001
Treatment (T) 4 2508.57 627.14 15.96 <0.0001
INTERNATIONAL JOURNAL OF FOREST ENGINEERING 115
Figure 1. Harvester time consumption in NCM fellings with different management goals relative to final fellings as a function of mean stem volume (m3 under bark), for
each machine size. BF: broad-leaved forest, CBF: mixed conifer-broad-leaved forest, PF: pine forest, CF: mixed conifer forest, CM: commercial management (final felling).
Note: management goal CF is only represented in machine size L.
veteranisation), and marking of ancient and cultural remains felling operations, so this would contribute to reduced
by cutting high stumps. productivity.
Except for NCM treatment BF, harvesting intensity was mostly In addition, fixed costs per harvested m3 can also be
in the 25%–50% range, i.e., in the same range as ordinary com expected to increase, as the harvested volume is lower than in
mercial thinning. However, the increase in time consumption, and final felling. Furthermore, harvest planning becomes more
thereby costs, was lower than the 30% increase often stated for complicated, as forest planners normally need to specify
thinning (cf. Eriksson and Lindroos 2014; Jonsson 2015). a larger number of instructions to machine operators in the
Increased time consumption in all kinds of selective harvesting is harvest plan to achieve the specific goals of the NCM treat
an effect of restrictions caused by residual trees and seedlings and ment. This will also contribute to a higher logging cost.
the amount of care the operator has to take to avoid damaging Unfortunately, the data did not enable an analysis of how
them (Eliasson 1998). This could also explain the higher relative revenues were affected by the NCM treatments, as information
increase in time consumption for BF. Birches and aspens have on the remaining stand after felling was limited.
wider crowns, and more effort is needed to avoid snagging in these The study results are based on NCM treatments of 14 stands,
when a tree is felled, which implies that residual broad-leaved trees and the treatments were specific to each stand. This means that
cause greater restrictions to harvester work than spruce trees. In although we managed to classify NCM management goals into
addition, the BF stands were situated on soils that were normally four groups, variations within these groups can be large.
wetter than the other stands; this could also be a contributing However, the study encompasses all NCM stands treated during
factor, as additional time might be required to reinforce machine a two-year period where it was possible to retrieve complete
trails with slash to avoid soil damage. harvester data. Using data from stands with only partial data
XL harvesters were faster than large harvesters, as could be was not considered an option, as it would have created difficul
expected from earlier studies (Eriksson and Lindroos 2014; Liski ties in connecting harvester data to the other data sources. As all
et al. 2020). It seems likely that there should be interactions available stands were chosen, there was no intentional bias in the
between NCM treatment and machine size. However, investi stand selection. There was another NCM-marked stand that had
gating this would require a larger dataset than that available to all data available, but because this stand consisted of a mix of
us, in order to separate this interaction from the effects of the conventional harvesting and large set-asides, the harvesting was
operators. This was not possible in our dataset, since all work considered not to fit our definition of an NCM treatment.
days (observations) during the harvest of one stand are observa The purpose of this study was not to evaluate the ecological
tions of one or two operators using the same harvester. outcome of the forest operation. However, the field visits during
In addition to the increase in harvester time consumption due spring 2022 allowed a visual inspection of the general condition
to the residual stand, time consumption in NCM treatments is also of the stands after harvest, and enabled a subjective evaluation of
increased by additional work tasks for the harvester (cf. Santaniello how the aim of NCM was met and how well the instructions in
et al. 2016). These include girdling of standing trees for veteranisa the harvest plan were followed. The plans were normally fol
tion (Andersson-Tjäder 2017) and increasing the amount of high lowed, but with some minor deviations. Instructions varied, even
stumps and lying dead wood. for similar treatments, and this reveals a need for standardization
Increased time consumption for the harvester work is not the of instructions, with clearer definitions of goals and measures in
only thing that affects logging costs in an NCM operation. the harvest plans. A long decision chain involving several
Compared to CM final felling operations, forwarder time con players, with different points of view, increases the risk for
sumption in these NCM treatments is expected to increase in misunderstandings on how NCM treatments should be per
a similar way to that in thinning. This is mainly caused by formed. Therefore, there is a need for greater consensus and
a reduced wood concentration and increased restrictions due to better communication between forest planners, nature conserva
the residual stand (Eliasson 1998). However, it may be necessary to tion experts, production leaders, and contractors to achieve the
choose a slightly smaller forwarder than the ones used in final goals of NCM to promote forest biodiversity.
116 R. FERNANDEZ-LACRUZ ET AL.