Effects of even-aged and uneven-aged management on carbon dynamics and timber yield in boreal Norway spruce stands: A forest ecosystem model approach

Rinnakkaistallenteet Luonnontieteiden ja metsätieteiden tiedekunta

2019

Effects of even-aged and uneven-aged management on carbon dynamics and timber yield in boreal Norway spruce

stands: A forest ecosystem model approach

Kellomäki, S

Oxford University Press (OUP)

Tieteelliset aikakauslehtiartikkelit

© Institute of Chartered Foresters All rights reserved

http://dx.doi.org/10.1093/forestry/cpz040

https://erepo.uef.fi/handle/123456789/7943

Downloaded from University of Eastern Finland's eRepository

Accepted to

1

Forestry

2 3 4

5

Effects of even-aged and uneven-aged management on carbon dynamics and

6

timber yield in boreal Norway spruce stands: A forest ecosystem model approach

7 8

Kellomäki, S., Strandman, H., Peltola, H.¹⁾ 9

10

University of Eastern Finland, Faculty of Science and Forestry, School of Forest Sciences, PO Box 111, FI- 11

80101 Joensuu, Finland 12

1) Corresponding author: Heli Peltola (email: heli.peltola@uef.fi, tel.: +358 40 5880005) 13

14

Abstract

15

We used a gap-type forest ecosystem model to study how even- and uneven-aged management 16

affected the carbon dynamics and timber production in boreal Norway spruce stands. In business- 17

as-usual management, the intensity of thinnings (from below) and single-tree selective cuttings 18

followed those recommended for even-aged (BT) and uneven-aged management (BSC) in practical 19

forestry in Finland. Moreover, higher or lower basal area thresholds, and shorter or longer 20

production cycles, were used in simulations. We found that, the mean annual carbon uptake, volume 21

growth, and carbon stock in trees and harvested timber, were nearly the same under even-aged (BT) 22

and uneven-aged (BSC) management, when assuming full seed crop in latter one. However, the 23

carbon stock in the soil and ecosystem and the mean annual net ecosystem exchange were slightly 24

smaller under BT. The carbon retention time was longer under BSC. The net present value (NPV with 25

interest rate of 3%) of timber production was clearly lower under BT, when the calculation was 26

initiated at planting on clear-cutting area, in opposite to when initiating calculation a few years 27

before the second thinning. Higher basal area thresholds and longer production cycles increased 28

carbon stocks, carbon retention and timber yield, regardless of management system. On the other 29

hand, the results of uneven-aged management (BSC) were very sensitive to the success of natural 30

regeneration and ingrowth of seedlings, as a reduction of the seed crop by 25–75% from the full 31

seed crop decreases the volume growth by 44–74% and timber yield up to 46%.

32 33

Keywords: Carbon exchange, carbon retention, carbon sequestration, economic profitability of 34

forestry, even-aged management, uneven-aged management, boreal forests, gap-type forest 35

ecosystem model 36

37 38

Introduction

39

In forest ecosystems, carbon cycles through uptake into storage in trees. Carbon further cycles 40

through litterfall and dead trees into soil organic matter (SOM), further being emitted into the 41

atmosphere through the decay of SOM. Carbon uptake and emissions define the net ecosystem 42

exchange and balance. They are affected by the structure and dynamics of forest ecosystem, and 43

edaphic and climatic properties of site. In forestry, intensity of management and wood harvesting 44

also impact on the balance and retention of carbon in the ecosystem, respectively (e.g., Briceno- 45

Elizondo et al., 2006a, b; Garcia-Gonzalo et al., 2007a, b, c; Jandl et al., 2007; Lorenz and Lal, 2010).

46

This further applies for the supply of timber and forest biomass, which are used to substitute fossil- 47

based materials and energy. In the perspective of climate change mitigation, such forest 48

management regimes are needed, which can simultaneously enhance the removal of carbon from 49

the atmosphere and storing and retaining carbon in forest ecosystems, and substituting fossil-based 50

materials and energy.

51 52

In Northern Europe, Norway spruce (Picea abies Karst. (L.)) is ecologically and economically among 53

the main tree species. In even-aged management, this shade-tolerant species is widely planted on 54

medium-fertile and fertile upland sites. For example, in Finland planting of Norway spruce is usually 55

followed by precommercial management, and two to three commercial thinnings during rotations 56

of 70 to 100 years, depending on site fertility and region (Äijälä et al., 2014). Optionally, uneven- 57

aged management, with selective cutting in every 15 to 20 years can be used in Norway spruce (Äijälä 58

et al., 2014). In uneven-aged management, the production cycle extends from a selective cutting to 59

another, allowing natural regeneration and ingrowth of seedlings in canopy gaps. In selective cutting, 60

mainly larger (co-dominant and dominant) trees are harvested, but also dense groups of smaller 61

trees can be thinned, if necessary.

62 63

There are only a few of comparative studies available on how uneven- and even-aged management 64

affect simultaneously the carbon dynamics and timber production under boreal conditions. Nilsen 65

and Strand (2013) found that, carbon storage in the trees and ecosystem was greater under even- 66

aged than uneven-aged Norway spruce stands based on a long-term experiment. The amount of soil 67

carbon was, however, smaller in even-aged stands, which may partly be explained by the faster 68

decay of SOM in relation to clear-cutting than to that occurring after selective cutting (e.g., Jandl et 69

al., 2007). Based on model simulations, Pukkala et al. (2011b) and Peura et al. (2018) showed that 70

the total carbon storage in the ecosystem (trees and soil) might be of similar magnitude in Norway 71

spruce forests under both uneven- and even-aged management.Paradis et al. (2019) also suggested 72

that by increasing the rotation lengths under even-aged management and by using partial cut (to 73

lesser extent), it may be achieved higher potential to mitigate climate change. This is due to 74

increased carbon balance of forestry due to increased CO2 storage in forest biomass and harvested 75

wood products, and displacement of CO2-intensive materials, respectively.

76 77

Based on extensive data from permanent long-term experiments established in 23 even-aged and 78

26 uneven-aged Norway spruce stands in southern and central Finland, Hynynen et al. (2019) 79

showed that the basal area growth was on average 20% smaller for uneven-aged stands during the 80

15 years after cutting. The difference was the largest during the first 5 years after cutting. However, 81

based on a long-term experiment, Nilsen and Strand (2013) found that the total timber yield in 82

uneven-aged Norway spruce stands was 95% of that in even-aged stands. As a comparison, several 83

previous model-based studies have shown that timber yield under even-aged management is higher 84

than that under uneven-aged management (e.g., Lundqvist et al., 2007; Tahvonen et al., 2010;

85

Pukkala et al., 2011b; Tahvonen and Rämö 2016; Peura et al., 2018). However, the differences in 86

results between previous studies may at least partially be affected by differences in volume of 87

growing stock (stocking density) and intensity of harvesting in different management systems (see 88

e.g. Lundqvist 2017; Hynynen et al. 2019).

89 90

The economic profitability of uneven-aged management has usually been found higher than that of 91

even-aged management. This is mainly because of the avoided costs of regeneration (e.g., soil 92

preparation and planting of seedlings) and precommercial management (e.g., Pukkala et al., 2010, 93

2011b; Tahvonen et al., 2010; Tahvonen and Rämö 2016; Peura et al., 2018). The profitability of 94

uneven-aged management can be further increased compared to even-aged management if higher 95

interest rates and management costs, and/or lower timber prices are used in economic calculations 96

(Andreassen and Øyen, 2002; Laiho et al., 2011; Juutinen et al., 2018). Based on these reasons, an 97

interest in close-to-nature silviculture, and thus the use of selective cutting in Norway spruce, has 98

recently increased in Finland, as elsewhere.

99 100

In uneven-aged management, the success of natural regeneration and ingrowth of seedlings have 101

uncertainties. This is, because the amount of seed crop, success of dispersion of seeds, 102

establishment of seedlings and ingrowth of seedlings in canopy gaps, affect all the success of uneven- 103

aged management. In Norway spruce, the quantity and quality of seed crop vary substantially from 104

year to year (Koski and Tallqvist, 1978; Saksa, 2004; Saksa and Valkonen, 2011). This increases 105

uncertainty in the establishment of seedlings, compared to planting in even-aged management.

106

However, several inventories show that the establishment of seedlings is likely large enough in a 107

longer term for uneven-aged management (e.g. Lähde et al. 2002; Saksa, 2004; Saksa and Valkonen, 108

2011; Eerikäinen et al. 2014). On the other hand, there has been found a negative correlation 109

between the level of ingrowth of seedlings (and other small trees) and overstory standing volume 110

during the first 5-10 years after harvesting especially in dense stands (Lundqvist and Nilsson 2007;

111

Lin et al. 2012). Canopy openess (gaps) has also been found to affect height increment of young 112

spruce seedlings (height 0.1-2.0 m) and small trees (height 2.0-5.0 m) more than the average 113

overstory basal area or standing volume (Chrimes and Nilson 2005). In overall, the establishment of 114

seedlings might also occur over longer period under uneven-aged management than in natural 115

regeneration with seed-tree and shelter-wood cuttings in even-aged management (Pukkala et al.

116

2011a; Saksa, 2004; Laiho et al., 2011; Saksa and Valkonen, 2011; Eerikäinen et al., 2014; Valkonen 117

et al. 2017).

118 119

In this work, we used a process-based (gap type) forest ecosystem model to simulate how even-aged 120

management (with thinnings from below, removing mainly suppressed and intermediate trees) and 121

uneven-aged management (with single-tree selection cuttings from above, removing mainly 122

dominant and codominant trees) affected the carbon dynamics and timber production in Norway 123

spruce stands. The simulations were performed for medium fertile upland sites under middle boreal 124

conditions in central Finland. In the simulations, we used varying basal area thresholds in thinning 125

and selective cuttings, and varying lengths of production cycle (rotation, and the interval between 126

subsequent selective cuttings). Additionally, we used varying seed crop potential in uneven-aged 127

stands to evaluate their effects on results.

128 129

We hypothesized that the carbon sequestration and timber yield were of the same magnitude under 130

even-aged management and uneven-aged management with full seed crop, when the basal area 131

thresholds for the thinning and selective cutting, respectively, were those recommended for even- 132

aged and uneven-aged management in practical forestry. Consequently, we hypothesized that the 133

economic profitability of timber production is similar for both management regimes, if the economic 134

calculations were initiated under even-aged management at the commercial thinning phase, instead 135

of when planting on clear-cutting area. Additionally, we hypothesized, that the results for uneven- 136

aged management are sensitive to the seed crop potential, and consequently to the natural 137

regeneration success and ingrowth of seedlings, respectively.

138 139

Methods and simulations

140 141

Study layout and management options 142

In the simulations, we used a gap-type forest ecosystem model (SIMA) (Kellomäki et al., 2008). Under 143

even-aged management, the production cycle extended from seedlings planted on clear-cutting 144

area, through thinnings (from below), to clear-cutting at the end of the selected rotation length.

145

Thereafter, the clear-cutting area was replanted, followed by the same management sequence as in 146

the previous rotation. In the simulation of uneven-aged management, single-tree selection cutting, 147

at a given interval, allowed the natural regeneration and ingrowth of seedlings in canopy gaps. In 148

the baseline management under even-aged (baseline thinning, BT) and under uneven-aged (baseline 149

selective cutting, BSC) management, the basal area thresholds were those recommended for use in 150

practical forestry in Finland (Äijälä et al., 2014). In addition, higher or lower basal area thresholds, 151

and shorter and longer production cycles, were used in the simulations.

152 153

All simulations were performed under the current climate in the middle boreal zone (62N), central 154

Finland, represented by a mean annual temperature sum of 1100 degree-days, precipitation of 540 155

mm, and atmospheric CO2 concentration of 350 ppm. The simulations were performed for medium 156

fertile (sub-mesic) sites (Myrtillus site type), on moraine soil with a volumetric water content of 25 157

m³ m^-3 at field capacity, and 5 m³ m^-3 at wilting point. The amount of soil organic matter (SOM), 158

including litter and humus, was 68 Mg ha^-1 when initializing the simulations.

159 160

Under even-aged management, planting of 1800 seedlings ha^-1 with butt diameters of 2.5 cm 161

(unimproved seedlings, no breeding gain assumed), was used to initialize the simulations at the 162

clear-cut area. In the baseline thinning (BT) regime, thinnings from below were done when the 163

dominant height was 12–22 m and the basal area 24–28 m² ha^-1. After thinning, the remaining basal 164

area was 15–20 m² ha^-1, depending on the dominant height. In two additional thinning regimes, the 165

basal area was kept 20% higher/lower (BT+20 and BT-20) than in the BT over a rotation. Clear- 166

cutting was done at the end of the rotation. The simulations were performed over 525 years, using 167

seven rotations (7 x 75 years). However, the three first 75-year rotations were excluded from the 168

analysis in order to stabilize the effects of the initial stand conditions on the model output. Thus, 169

only four last rotations of 75 years, between years 225 and 525 (i.e., 300 years) was included in the 170

data analysis.

171 172

Under uneven-aged management, the simulations were initiated in the same way as for even-aged 173

management. Consequently, thinnings (from below) were done up to the year 75 from the initiation, 174

but the clear-cutting was not done, allowing natural seeding to occur. Thereafter, single-tree 175

selective cutting was done, one after another, following the given interval, thus gradually switching 176

the even-aged management to an uneven-aged system. Selective cuttings were done whenever the 177

basal area of a stand was ≥ 20 m² ha^-1, by reducing the basal area to 11 m² ha^-1. In addition to the 178

baseline selective cutting (BSC), two additional selective cutting modes were used, representing  179

20% higher/lower basal area thresholds (BSC+20 and BSC-20) than those used in the BSC over a 180

production cycle. Also in this case, only the interval of simulations between years 225 and 525 (i.e., 181

300 years) was included in the data analysis, to allow for the stabilization of the variability in the 182

structure of tree stand, and the amount and properties of the SOM. In all simulations, only sawlogs 183

(stems with a diameter of ≥ 15 cm at the top), and pulpwood (diameter ≥ 6 cm at the top) were 184

harvested. In opposite, the harvest residues (and small sized trees cut) were left on site to decay, 185

regardless of management regime.

186 187

Growth of trees under even- and uneven-aged management 188

In the simulations, the carbon dynamics and timber production were affected by the regeneration 189

success (naturally born or planted seedlings), growth, and mortality of the trees. Under both 190

management systems, seedlings were established when their height was ≥ 1.3 m. The growth of 191

established seedlings and more mature trees was calculated based on the diameter growth (Δdbh, 192

cm yr.^-1):

193 194

Δdbh = Δdbho×MTS×ML×MW×MN (1)

195 196

where Δdbho is the maximum diameter growth (cm yr.^-1) under optimal conditions. The diameter 197

growth was further scaled in the range of 0 to 1 in relation to prevailing temperature sum (TS in 198

degree-days with +5°C threshold, MTS), light conditions inside the stand (ML), soil moisture (MW), and 199

nitrogen supply (MN) (see for details, Table 1).

200 201

Table 1.

202 203

Maximum diameter growth under optimal conditions is dependent on tree diameter at breast height 204

(dbh, cm, height ≥ 1.3 m) and atmospheric CO2 concentration(ppm):

205 206

∆𝑑𝑏ℎ_𝑜 = exp⁡(𝑎1 + ^𝑏1

0.01×𝐶𝑂₂) × 𝑑𝑏ℎ × 𝑒^{𝑑𝑔𝑟𝑜×𝑑𝑏ℎ} (2)

207 208

where a1, b1, and dgro are the parameters. Stem diameter was further used to calculate the height 209

(m) of the trees (Kellomäki et al. 2008). Similarly, the mass (Mass(j), kg) of different organs (i.e., 210

foliage, branches, stem, and roots) was calculated as the function of stem diameter:

211

212

𝑀𝑎𝑠𝑠(𝑗) = exp⁡[𝑎2(𝑗) + 𝑏2(𝑗) × ^𝑑𝑏ℎ

𝑐(𝑗)+𝑑𝑏ℎ] (3)

213 214

where a2(j), b2(j), and c(j) are parameters specific to the mass component (j).

215 216

The initial amount of SOM, and the nitrogen available for growth, are related to the site type and 217

regional mean temperature sum under the current climate (Kellomäki et al., 2008). Litter from any 218

organ and deadwood (stemwood, branches, needles and leaves, stumps, and coarse to fine roots) 219

transfer carbon and nitrogen into the soil, where litter and humus decay. Consequently, nitrogen is 220

released for reuse in growth, and CO2 is emitted into the atmosphere. The simulations were 221

performed using the time step of one year and carried out on an area of 100 m². The simulations 222

were based on the Monte Carlo technique. Each management scenario was repeated 100 times, but 223

only the mean annual output values were used in the data analysis.

224 225

Emergence and ingrowth of seedlings under uneven-aged management 226

When using uneven-aged management with single-tree selection, large and mature trees (co- 227

dominant and dominant ones) produced seeds for natural regeneration in canopy gaps. Following 228

the approaches of Fox et al. (1983) and Pukkala (1987a, b), each seed crop indicates the potential 229

number of emerging seedlings, but the scarcity of stockable area (open mineral soil/seedbed) and 230

herbivory may limit the density of emerging seedlings. The properties of the seed crop may further 231

limit the number of emerging seedlings, because only a fraction of the seeds is mature and capable 232

of germinating (Kellomäki and Väisänen, 1995; Kellomäki et al., 1997):

233 234

𝑁𝑆(𝑡) = 10,000 × 𝑓(𝑆𝐶(𝑡)) × 𝑓(𝑆𝑆(𝑡)) × 𝑓(𝑈𝐸(𝑇)) × 𝑓(𝐹𝑈(𝑡)) × 𝑓(𝑀𝐴(𝑡)) × 𝑓(𝐺𝐸𝑅(𝑡)) (4) 235

236

where NS(t) is the number of emerged seedlings (seedlings ha^-1 yr.^-1) per the given seed crop SC(t) 237

(seeds m^-2), SS(t) is the fraction of stockable area, UE(t) is the fraction of uneaten seeds, FU(t) is the 238

fraction of full seeds, MA(t) is the fraction of mature seeds, and GER(t) is the fraction of germinated 239

seeds.

240 241

Only a fraction of the seedlings (SURMUL(t), (0,1)) born in a given year t survives to year t+1. The 242

initial growth of seedlings from each seed crop is followed through 12 years (Kellomäki et al., 1997).

243

Over this period, the probability of survival increases, and the sensitivity of the seedlings to death 244

decreases as a function of seedling age:

245 246

𝑆𝑈𝑅𝑀𝑈𝐿(𝑡) = 𝑆𝑈𝑅𝑃𝑅𝐵(𝑡) + 𝑌𝐹𝐿 × 𝑇𝑂𝐿(𝑡) (5)

247 248

where SURPRB(t) is the probability of seedling survival, TOL(t) is the sensitivity of the seedlings to 249

death, and YFL is a random number (0,1). Only a small fraction of emerged seedlings from each seed 250

crop survive further, with the ingrowth into the existing canopy.

251 252

The initial diameter at the stem butt (DButt) of emerged seedlings was assumed to be 0.1 cm. The 253

potential growth of butt diameter was calculated using same Eq. (1), which was used for established 254

seedlings (height ≥ 1.3 m) and more mature trees. In the case of emerged seedlings, the dbh was 255

replaced by DButt in Eq. (1). The potential growth of butt diameter was used both for naturally born 256

and planted seedlings until they reached height of 1.3 m. The butt diameter growth was further used 257

to calculate the height growth of seedlings (height < 1.3 m): PHEIHGT = pdicn x DButt, where pdicn is 258

a parameter. The ingrowth of seedlings was assumed to occur when seedling height was ≥ 1.3 m at 259

the age of 12 years since emerging (if not exceeded the height limit, seedling dies). Time span of 12 260

years used in this study equals to the period, when the number of seedlings exceeded 80% of that 261

established in the 30-year period since shelterwood cutting in the middle boreal zone in the study 262

of Räsänen et al. (1986).

263 264

Over three to five years after harvesting, the decay of logging residues binds a part of nitrogen, which 265

was available prior to cutting for tree growth. This reduces the diameter growth of seedlings and 266

more mature trees over first few years after cutting. However, the diameter growth recovers in 267

response to the increase of available nitrogen due to further decay of SOM (including the logging 268

residues from the previous cutting). Light conditions, below dominating canopy, affect the growth 269

of seedlings both before ingrowth and thereafter (e.g. Cajander, 1934; Greis and Kellomäki, 1981).

270 271 272

Model performance 273

Assessing model performance involved two main questions: (i) how well does natural seeding 274

(regeneration) produce and maintain an acceptable size distribution of trees under uneven-aged 275

management; and (ii) how are trees growing under uneven- and even-aged management through 276

the production cycles? Figure 1 shows the simulated fraction of trees in different diameter classes 277

(% of cases) over the 300-year simulation period for management option BSC, with seed crop of 0.25, 278

0.50, 0.75, and 1 of the full seed crop potential. The simulated size distribution (columns, SIMA 279

model with full seed crop = 1) is well in line with the measured size distribution (solid line, Shanin et 280

al., 2016), based on measurements for the ERIKA plots produced by the Natural Resources Institute 281

Finland. The measurements represent sample plots prior to the next selection-cutting of the plots 282

(with the average basal area of 23.8 m² ha^-1). Moreover, the insert (small Figure) in Figure 1 shows a 283

close correlation between the simulated (x-axis) and the measured (y-axis) tree size distributions.

284 285

Figure 1.

286 287

Several previous model validations have demonstrated good agreement between the simulated and 288

the measured mean annual volume growth of Scots pine, Norway spruce, and birch on upland forest 289

sites (the forest inventory plots) throughout Finland (Kellomäki et al. 2005, 2008). Simulations by the 290

current model and the empirical growth and yield model MOTTI (Hynynen et al., 2002) have also 291

showed good agreement for the mean annual volume growth for managed Norway spruce stands 292

on medium fertile upland forest sites (Routa et al. 2011a, b). This also holds for single trees growing 293

under uneven-aged management because the allometric growth of stemwood in single trees of the 294

same size is similar under both management systems.

295 296

Data analysis 297

Based on the current simulations, we investigated how different management options affected: (i) 298

the dynamics and structure of tree stands; (ii) the carbon dynamics; (iii) the amount of harvested 299

timber; and (iv) the economic profitability of timber production. We also studied the effects of: (v) 300

the size of the seed crop; (vi) the basal area thresholds for thinnings and selective cuttings; and (vii) 301

the length of the production cycle on the carbon dynamics, timber yield, and economic profitability 302

of timber production. The carbon dynamics were further assessed by considering net primary 303

production, litterfall, and carbon emission from decaying SOM (see Figure 2). Based on these, the 304

net ecosystem exchange (NEE) was calculated: NEE = NPP – RH, where NPP is the net primary 305

production and RH is the heterotrophic respiration from decaying SOM. The carbon in trees and soil 306

indicates the carbon stock in the ecosystem, where carbon is retained for a while, indicated by the 307

retention time (τ, years) (Kellomäki, 2017):

308 309

𝜏 = 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦⁡𝑜𝑓⁡𝑎⁡𝑠𝑦𝑠𝑡𝑒𝑚⁡𝑡𝑜⁡ℎ𝑜𝑙𝑑⁡𝑐𝑎𝑟𝑏𝑜𝑛

𝑅𝑎𝑡𝑒⁡𝑜𝑓⁡𝑐𝑎𝑟𝑏𝑜𝑛⁡𝑓𝑙𝑜𝑤⁡𝑡ℎ𝑟𝑜𝑢𝑔ℎ⁡𝑎⁡𝑠𝑦𝑠𝑡𝑒𝑚⁡ (6)

310 311

The retention time begins when carbon enters the system and ends when carbon leaves the system.

312 313

Figure 2.

314 315

The economic profitability of timber production was calculated using the net present value (NPV), 316

with an interest rate of 3%. In uneven-aged management, the stumpage prices for sawlogs and 317

pulpwood were assumed equal to those used for thinning other than the first commercial thinning 318

under even-aged management (Table 2). These values were higher than the stumpage prices for the 319

first thinning, but lower than those for clear-cutting. In even-aged management, the regeneration 320

cost was assumed to be 1066 € ha^-1 for 1800 seedlings planted ha^-1 including soil preparation, and 321

the tending cost for seedling stands 400 € ha^-1 five years after planting, based on the online statistics 322

of the Natural Resources Institute Finland. Three different initial stand conditions were used for the 323

NPV calculation under even-aged management over four 75-year rotation cycles, representing the 324

300-year period included in the data analysis: (i) initiation on clear-cutting area; (ii) initiation 6–7 325

years before the first commercial thinning (at an age of 20 years); and (iii) initiation 6–7 years before 326

the second commercial thinning (at an age of 45 years). Under uneven-aged management, the length 327

of the total simulation period was the same, as that under even-aged management.

328 329

Table 2.

330 331

Results

332

Timing management interventions and diameter distribution of trees 333

Regarding both management systems, Figure 3 shows the timing of thinnings and selective cuttings, 334

respectively, and the variability of the stemwood carbon, over the 300 years simulation period.

335

Under even-aged management, thinning occurred at an interval of 20–30 years (the mean interval 336

of 25 years between thinnings, equals BT) from the initiation of each rotation. Compared to the 337

baseline thinning (BT), thinning was done earlier under lower basal area thresholds (BT-20) and later 338

under higher basal area thresholds (BT+20), respectively. In uneven-aged management with full seed 339

crop potential (seed crop fraction = 1), the interval between cuttings was 15 years in the baseline 340

selective cutting (BSC). A 20% reduction in basal area (BSC-20) thresholds produced an 11-year 341

selective cutting interval, while a 20% increase (BSC+20) in thresholds gave a 23-year selective 342

cutting interval, respectively.

343 344

Figure 4 shows the mean diameter distribution for both management systems over the 300-year 345

period. Under even-aged management, the diameter distribution spanned from just-planted 346

seedlings to trees of 35–40 cm just before clear-cutting. The share of small trees (dbh < 5 cm) was 347

similar, regardless of thinning regime used (BT-20, BT, BT+20). This pattern also held for large trees 348

(dbh > 25 cm). However, the distribution was under even-aged management less skewed towards a 349

dominance of small trees than that under uneven-aged management, when assuming in the latter 350

case the full seed crop potential (seed crop fraction = 1). In the latter case, the share of small trees 351

(dbh < 5 cm) was larger, and the share of larger trees (dbh > 25 cm) smaller. When using 20% higher 352

basal area thresholds in selective cutting (BSC+20), the average number of small seedlings increased, 353

in contrast to the use of 20% lower thresholds (BSC-20), due to higher seed crop and ingrowth of 354

seedlings. A reduction of the seed crop by 25–75% reduced the number of small seedlings (dbh < 10 355

cm) under the BSC by up to 20–50%, compared to seedling numbers under full seeding potential.

356 357

Figure 3.

358 359

Figure 4.

360 361

Carbon uptake, stocks, and emission 362

The mean annual carbon uptake is shown in Figure 5, including the carbon in stems, branches, 363

foliage, and roots over the 300 years simulation period. The mean annual carbon uptake was nearly 364

the same under BT and BSC (seed crop fraction = 1). The mean annual carbon uptake was 10% lower 365

for BT-20 and BSC-20, and 4–10% higher for BSC+20 and BT+20, compared to BT and BSC.

366

Consequently, the mean carbon stock in trees was virtually the same under BT and BSC (Figure 5, 367

Table 3) over the 300 years simulation period. The mean carbon stock was 12-17% lower under BSC- 368

20 and BT-20, and 21-28% higher under BT+20 and BSC+20, compared to that under the BT and BSC.

369

The carbon stock in soil was, on average, slightly greater under even-aged management.

370 371

Regardless of management regime, the carbon stock in soil was related to the litterfall, thus 372

ultimately to the carbon uptake in trees, as further holds for the carbon stock in the ecosystem (trees 373

and soil). The reduction in annual mean seed crop had a clear effect on carbon uptake and, 374

consequently, on the total mean growth of stemwood, under uneven-aged management. The 375

reduction in yearly seeding potential, by 25–75% of the full seed crop, reduced stemwood growth 376

by 44–74%. On the other hand, carbon flows and stocks in the ecosystem (trees and soil) were less 377

sensitive to the seed crop reduction. Even under 25% of the full seed crop potential, the carbon flows 378

and stocks in the ecosystem would still be > 50% of those under the full seed crop. Differences in 379

mean annual carbon uptake, emissions, litterfall, and carbon in the trees, between uneven- and 380

even-aged (with seed crop fraction = 1) management were in general quite small. However, the 381

carbon stock in the soil and ecosystem (trees and soil), were slightly lower under even-aged 382

management.

383 384

Figure 5.

385 386

Table 3.

387 388

Carbon exchange and retention time 389

Carbon exchange (i.e., the NEE) was not sensitive to management system over the 300-year period 390

(Figure 6). The carbon exchange was only 4–6% greater under BSC-20 (seed crop fraction = 1) than 391

under BT-20. This held also for the BSC and BT. The carbon exchange was also nearly the same under 392

the BSC+20 and BT+20. Under both management systems, lower stocking reduced NEE, while higher 393

stocking produced the opposite effect.

394 395

Figure 6.

396 397

Carbon was retained in the ecosystem (in trees and soil) over 33 years (Figure 6, Table 4) when BT 398

(with a 20-year thinning interval) under a 75-year rotation length was used. The prolongation of 399

rotation length from 75-years to 100 years increased carbon retention to 37 years under BT. The 400

thinning thresholds also affected the carbon retained in the ecosystem. Carbon retention increased 401

from 36 to 42 years under BT+20 (with a 25-year thinning interval), when increasing the rotation 402

length from 75-years to 100 years. The corresponding prolongation of the rotation length also 403

increased retention under BT-20 (with a 15-year thinning interval), from 30 to 33 years.

404 405

Under BSC (seed crop fraction = 1), carbon retention was 15 years, and thus substantially less than 406

under BT. An increase in basal area threshold in selective cutting (BSC+20) increased the retention 407

to 23 years. A reduction in basal area thresholds (BCS-20%) reduced the carbon retention to 11 years.

408

However, the carbon retention was affected also by the seed crop potential. A reduction in full seed 409

crop potential for BSC by 25–75% reduced carbon retention by 5–16%. In overall, carbon retention 410

was more sensitive to the basal area thresholds than seed crop potential.

411 412

Table 4.

413 414

Carbon stock, amount of harvested timber, and economic profitability 415

The carbon stock and total amount of harvested timber were nearly the same under even- and 416

uneven-aged management (with seed crop fraction = 1), regardless of the basal area thresholds used 417

in cuttings over the 300-year simulation period. However, the amount of sawlogs and their carbon 418

stock were slightly higher under uneven-aged management, in opposite to those of pulpwood (Table 419

5). The use of a lower basal area threshold in cuttings reduced them, contrary to the use of a higher 420

cutting thresholds, regardless of management system. The average sawlog share of the total harvest 421

was 74–82% under uneven-aged management and 73–76% under even-aged management. The 422

average shares of pulpwood for them were 18–26% and 24–27%, respectively. A reduction in full 423

seed potential of 25–75% reduced the total timber yield by up to 46%, and the reduction was greater 424

for pulpwood than sawlog yield.

425 426

The NPV was 78–81% smaller under even-aged management regimes compared to corresponding 427

uneven-aged management regimes with seed crop fraction = 1, when the production cycle under 428

even-aged management was initiated by planting on clear-cutting area (Figure 7). The NPV was also 429

22–28% smaller under even-aged management regimes when initiating the NPV calculation a few 430

years before the first commercial thinning. However, initiating the calculations a few years before 431

the second commercial thinning, resulted in 3-18% higher NPV under even-aged management 432

regimes. The difference was the greatest when the BT+20 regime was used under even-aged 433

management and thee BSC+20 was used under uneven-aged management. Regardless of 434

management system, the use of a lower basal area threshold reduced the NPV (by 5–15%), in 435

opposite to the use of higher basal area thresholds, due to decrease in timber yield.

436 437

Table 5.

438

Figure 7.

439

Discussion and conclusions

440 441

Evaluation of findings 442

443

We carried out a comparative analysis on how different even-aged and uneven-aged management 444

regimes affect the carbon dynamics and timber production in Norway spruce stands in the boreal 445

conditions, based on simulations with a gap-type forest ecosystem model. The model combines the 446

regeneration (planting or natural seeding), and the cohort-based growth and mortality of trees, 447

allowing the development of both even-aged and uneven-aged forest structure as controlled by 448

management. In even-aged management, the rotation was initiated by planting of seedlings on 449

clear-cutting area, followed by thinnings from below, and ended in clear-cutting for the next 450

rotation. In uneven-aged management, the emergence and ingrowth of seedlings occurred in 451

canopy gaps, when dominating and co-dominating trees were removed in repeated single-tree 452

selective cuttings. Under even-aged management, we did not assume any breeding gain in planting 453

of seedlings, which would result in at least about 10% better growth for seedlings, compared to 454

unimproved regeneration material (see e.g. Haapanen and Mikola, 2008; Haapanen et al. 2016).

455

Additionally, it may also increase economic profitability of timber production due to enhanced tree 456

growth and earlier cuttings, despite of the higher price of improved materials (see e.g. Ahtikoski 457

et al., 2012, 2013).

458 459

In model based studies, the simulation of even-aged management with thinnings is less complex 460

than that for uneven-aged management with selective cuttings. In the latter case, the ingrowth of 461

seedlings represents the accumulation of seedlings over several years from natural seeding to 462

established seedlings. However, the effects of different factors on regeneration success are often 463

aggregated in simulation models for a lump estimation of emergence and ingrowth of seedlings 464

(e.g., Pukkala et al. 2009; Tahvonen et al. 2010; Roessiger et al. 2016; Juutinen et al., 2018). We 465

partially decomposed this aggregation by using the approach of Fox et al. (1983) and Pukkala (1987a, 466

b), in which the seeding and germination of the seeds are affected in uneven-aged management by 467

the properties of seed crop and seedbed, the prevailing climatic conditions, and the effects of 468

herbivory, respectively. Only a few emerged seedlings per each annual seed crop become therefore 469

established and grow to full maturity (Kellomäki and Väisänen, 1995; Kellomäki et al., 1987, 1997).

470

471

Our simulations produced the percentage age-distribution close to the inverse J-shaped form which 472

is typical of forest growth under single-tree selective cuttings (e.g., Eerikäinen et al. 2014; Shanin et 473

al. 2016). When assuming a full seed crop (seed crop fraction = 1), the annual mean growth of 474

stemwood was in our study 4.5–6.0 m³ ha^-1 yr.^-1 for uneven-aged management, the range being 475

quite the same as for even-aged management. It was also in the same range as that found by Laiho 476

et al. (2011) for uneven-aged Norway spruce dominated forests. On the other hand, Lundqvist 477

(2017), based the meta-data analysis, and Hynynen et al. (2019), based on the field experiments, 478

found use of uneven-aged management to reduce stemwood (or basal area) growth in Norway 479

spruce stands compared to even-aged management. In the study by Hynynen et al. (2019), the basal 480

area growth was 20% smaller under uneven-aged than under even-aged management during the 481

post-treatment period of 15 years. However, the observed differences between management 482

systems are also affected by the differences in their volume of growing stock (stocking density) and 483

cutting intensity, which both affect forest growth per unit land area (see e.g. Lundqvist 2017;

484

Hynynen et al. 2019). This could be seen also based on our findings, for example, when comparing 485

the results for even-aged management with baseline regime (BT) against those of uneven-aged 486

management with higher or lower basal area thresholds for cuttings (BSC+20 or BSC-20), instead of 487

baseline regime (BSC).

488 489

Several previous model-based studies have shown that timber yield under even-aged management 490

is equal or higher than that under uneven-aged management (e.g., Lundqvist et al., 2007; Tahvonen 491

et al., 2010; Pukkala et al., 2011a, b; Lundmark et al., 2016; Tahvonen and Rämö 2016; Peura et al., 492

2018). In our study, the amount of harvested timber was quite same order under even-aged and 493

uneven-management, when assuming full seed crop in the latter one (seed crop fraction = 1). Also 494

based on a long-term experiment by Nilsen and Strand (2013), the total timber yield in uneven-aged 495

Norway spruce stands was 95% of that in even-aged stands.

496 497

The mean annual carbon uptake, emissions, litterfall, and carbon in the trees and harvested timber, 498

were quite a same order under even-aged and uneven-management, when assuming full seed crop 499

(seed crop fraction = 1). The carbon stock in the soil and ecosystem (trees and soil) and the mean 500

annual net ecosystem exchange were, in general, slightly smaller under even-aged than uneven- 501

aged management. On the other hand, the carbon retention time was shorter under uneven-aged 502

management. The preference for higher stocking further increased both net carbon exchange and 503

carbon retention, regardless of management system. Under even-aged management, the 504

prolongation of rotation length, from 75 to 100 years, increased the NEE and carbon retention time, 505

as found in previous studies (Kellomäki, 2017). Even a moderate prolongation of the rotation length 506

can increase the capacity of the forest ecosystem to absorb carbon and enhance the retention of 507

carbon in the ecosystem (Liski et al., 2001; Pukkala et al., 2011a, b; Kellomäki, 2017). Our findings 508

on carbon stock in the soil and ecosystem (trees and soil) deviate from those of Nilsen and Strand 509

(2013), who found based on an 80-year-long experiment, that the mean carbon storage in trees 510

(including roots) was clearly greater in even- than uneven-aged Norway spruce stands. On the 511

contrary, the amount of soil carbon was in some degree lower in the even- than the uneven-aged 512

stand. Consequently, the total amount of carbon in the ecosystem (trees and soil) was higher in the 513

even- than in the uneven-aged plots.

514

515

In our study, the share of saw-logs was also slightly greater, and pulpwood smaller, under uneven- 516

aged management (seed crop fraction = 1), which is in line with the findings of Pukkala et al. (2011a, 517

b), for example. As expected, uneven-aged management was substantially more profitable than 518

even-aged management, especially when initiating the economic calculations from the clear-cutting 519

area in the latter case. This is because in the former case, the costs of regeneration (e.g., soil 520

preparation and the planting of seedlings) and precommercial thinning (e.g., Pukkala et al., 2010, 521

2011b; Tahvonen et al., 2010; Pukkala, 2016; Tahvonen, 2016; Peura et al., 2018) are avoided, unlike 522

in the latter case. The situation was the opposite when initiating the NPV calculation a few years 523

before the second commercial thinning. This emphasizes the importance of a time perspective, with 524

proper initiation of calculations in comparative analyses. Thus, the initial structure substantially 525

affected the outcome of comparison between different management regimes.

526 527

Conclusions 528

529

Based on our findings, uneven-aged management is promising management system in many ways.

530

However, its success is in the long run very much depending on the success of natural regeneration 531

and ingrowth of seedlings, as we demonstrated in this study. For example, a reduction of the seed 532

crop by 25–75% from the full seed crop reduced the number of small seedlings (< 10 cm) under the 533

BSC by up to 20–50%. As a result, also the volume growth decreased by 44–74% and timber yield 534

decreased up to 46%, respectively. Carbon flows, stocks and retention were in some degree less 535

sensitive to the seed crop reduction. In the future studies, it should be considered also the climate 536

change and associated increase in various abiotic and biotic risks to forests (see e.g. Reyer et al.

537

2017). This is, because they may partly cancel the predicted increase in forest productivity under 538

changing climate, regardless of management system.

539 540

541 542

Conflict of interest statement

543

None declared.

544 545

Funding

546

This work was supported by the Strategic Research Council of the Academy of Finland for the FORBIO 547

project (decision number 314224), led by Prof. Heli Peltola at the School of Forest Sciences, 548

University of Eastern Finland.

549 550

References

551 552

Ahtikoski A., Ojansuu R., Haapanen M., Hynynen J., Kärkkäinen R., 2012 Financial performance of 553

using genetically improved regeneration material of Scots pine (Pinus sylvestris L.) in Finland. New 554

Forest 43, 335–348.

555

Ahtikoski A., Salminen S., Ojansuu R., Hynynen J., Kärkkäinen K., Haapanen M., 2013 Optimizing 556

stand management involving the effect of genetic gain: preliminary results for Scots pine in 557

Finland. Canadian Journal of Forest Research 43, 299–305.

558

Äijälä, O., Koistinen, A., Sved, J., Vanhatalo, K. and Väisänen, P. (eds) 2014 Hyvän metsänhoidon 559

suositukset – METSÄNHOITO (In Finnish). Metsätalouden kehittämiskeskus, Metsäkustannus OY, 560

Bookwell, 264 pp.

561

Andreassen, K. and Øyen, B.‐H. 2002 Economic consequences of three silvicultural methods in 562

uneven‐aged mature coastal spruce forests of central Norway. Forestry 75(4), 483–488.

563

Briceno-Elizondo, E., Garcia-Gonzalo, J., Peltola, H., Matala, J. and Kellomäki, S. 2006a Sensitivity of 564

growth of Scots pine, Norway spruce and Silver birch to climate change and forest management 565

in boreal conditions. Forest Ecology and Management 232(1–3), 152–167.

566

Briceno-Elizondo, E., Garcia-Gonzalo, J., Peltola, H. and Kellomäki, S. 2006b Carbon stocks and timber 567

yield in two boreal forest ecosystems under current and changing climatic conditions subjected 568

to varying management regimes. Environmental Science and Policy 9(3), 237–252.

569

Cajander, E.K. 1934. Über den Höhenzuwachs der Fichten-Pflanzenbeständ nach der Befreiung.

570

Communticationes Instituti Forestalis Fenniae 19(5), 1 -59.

571

Chrimes, D., Nilson, K., 2005 Overstorey density influence on the height of Picea abies regeneration 572

in northern Sweden. Forestry 78(4), 433-442.

573

Eerikäinen, K., Valkonen, S. and Saksa, T. 2014 Ingrowth, survival and height growth of small trees in 574

uneven-aged Picea abies stands in southern Finland. Forest Ecosystems 1, 5.

575

Fox, J.D., Zasada, J.C., Casbarro, A.F. and Veldhuizen, R. 1983 Monte Carlo simulation of white spruce 576

regeneration after logging in interior Alaska. Canadian Journal of Forest Research 14, 617–622.

577

Garcia-Gonzalo, J., Peltola, H., Zubizarreta Gerendiain, A. and Kellomäki, S. 2007a Impacts of forest 578

landscape structure and management on timber production and carbon stocks in the boreal 579

forest ecosystem under changing climate. Forest Ecology and Management 241(1–3), 243–257.

580

Garcia-Gonzalo, J., Peltola, H., Briceño-Elizondo, E. and Kellomäki, S. 2007b Effects of climate 581

change and management on timber yield in boreal forests, with economic implications: A case 582

study. Ecological Modelling 209(2–4), 220–234.

583

Garcia-Gonzalo, J., Peltola, H., Briceño-Elizondo, E. and Kellomäki, S. 2007c Changed thinning 584

regimes may increase carbon stock under climate change: A case study from a Finnish boreal 585

forest. Climatic Change 81(3–4), 431–454.

586

Greis, I., Kellomäki, S. 1981 Crown structure and stem growth of Norway spruce undergrowth under 587

varying shade. Silva Fennica 15(3), 306-322.

588

Haapanen, M., Mikola, J., 2008 Metsänjalostus 2050—Pitkän aikavälin metsänjalostusohjelma.

589

[Forest tree breeding 2050—a plan for long-term tree breeding]. Working papers of the Finnish 590

Forest Research Institute 71 (In Finnish).

591

Haapanen, M., Hynynen, J., Ruotsalainen, S., Siipilehto, J., Kilpeläinen, M-L., 2016 Realised and 592

projected gains in growth, quality and simulated yield of genetically improved Scots pine in 593

southern Finland. European Journal of Forest Research 135(6), 997–1009.

594

Hynynen, J., Ojansuu, R., Hökkä, H., Salminen, H., Siipilehto, J. and Haapala, P. 2002 Models for 595

predicting stand development in the MELA system. Finnish Forest Research Institute, Research 596

Paper 835, 1–116. Helsinki.

597

Hynynen J., Eerikäinen K., Mäkinen H. , Valkonen S. 2019 Growth response to cuttings in Norway 598

spruce stands under even-aged and uneven-aged management. Forest Ecology and Management 599

437, 314-323.

600

http://statdb.luke.fi/PXWeb/pxweb/en/LUKE/. Visited on November 17, 2017.

601

Jandl, R., Lindner, M., Versterdal, L., Bauwens, B., Baritz, R. and Agedorn, F. 2007 How strongly can 602

forest management influence soil carbon sequestration? Geoderma 137(3–4), 253–268.

603

Juutinen, A., Ahtikoski, A., Mäkipää, R. and Shanin, V. 2018 Effect of harvest interval and intensity on 604

the profitability of uneven-aged management of Norway spruce stands. Forestry 91(5), 589–602.

605

Kellomäki, S. 2017 Managing boreal forests in the context of climate change. Impacts, adaptation and 606

climate change mitigation. CRC Press, Boca /London/New York, 357 pp.

607

Kellomäki, S., Hänninen, H., Kolström, T., Kotisaari, A. and Pukkala, T. 1987 A tentative model for 608

describing the effects of some regenerative processes on the properties of natural seedling 609

stands. Silva Fennica 21(1), 1–10.

610

Kellomäki, S., Peltola, H., Nuutinen, T., Korhonen, K.T. and Strandman, H. 2008 Sensitivity of 611

managed boreal forests in Finland to climate change, with implications for adaptive management.

612

Philosophical Transactions of the Royal Society B363, 2341–2351.

613

Kellomäki, S.; Strandman, H.; Nuutinen, T.; Peltola, H.; Korhonen, K.T.; Väisänen, H. 2005 Adaptation 614

of Forest Ecosystems, Forests and Forestry to Climate Change; FINADAPT Working Paper 4;

615

Finnish Environmental Institute: Helsinki, Finland.

616

Kellomäki, S. and Väisänen, H. 1995 Model computations on the impact of changing climate on natural 617

regeneration of Scots pine in Finland. Canadian Journal of Forest Research 25, 929–942.

618

Kellomäki, S., Väisänen, H. and Kolström, T. 1997 Model computations on the effect of elevating 619

temperature and atmospheric CO2 on the regeneration of Scots pine at timberline in Finland.

620

Climatic Change 37, 683–708.

621

Koski, V. and Tallqvist, R. 1978 Tuloksia monivuotisista kukinnan ja siemensadon määrän 622

mittauksista metsäpuilla. Summary: Results of long-time measurements of the quality of 623

flowering and seed crops of forest trees. Folia Forestalia 364, 1–60.

624

Lähde, E., Laiho, O., Norokorpi, Y. and Saksa, T. 2002 Development of Norway spruce-dominated 625

stands after single-tree selection and low thinning. Canadian Journal of Forest Research 32(9), 626

1577–1584.

627

Laiho, O., Lähde, E. and Pukkala, T. 2011 Uneven- vs. even-aged management in Finnish boreal 628

forests. Forestry 84(5), 547–556.

629

Lin, C.J., Laiho, O., Lähde, E. 2012 Norway spruce (Picea abies L.) regeneration and growth of 630

understory trees under single-tree selection silviculture in Finland. Eur. J. For. Res. 131, 683-691.

631

Liski, J., Pussinen, A., Pingoud, K., Mäkipää, R. and Karjalainen, T. 2001 Which rotation length is 632

favourable to carbon sequestration? Canadian Journal of Forest Research 31, 2004–2013.

633

Lorenz, L. and Lal, R. 2010 Carbon sequestration in forest ecosystem. Springer Science + Business 634

Media, Dordrecht, 277 pp.

635

Lundmark, T., Bergh, J., Nordin, A., Fahlvik, N. and Poudel, B.C. 2016 Comparison of carbon balances 636

between continuous-cover and clear-cutting forestry in Sweden. Ambio 45(2), 203–213.

637

Lundqvist L. 2017 Tamm Review: Selection system reduces long-term volume growth in 638

Fennoscandic uneven-aged Norway spruce forests. Forest Ecology and Management 391: 362- 639

375.

640

Lundqvist, L. and Nilsson, K. 2007 Regeneration dynamics in an uneven-aged virgin Norway spruce 641

forest in northern Sweden. Scandinavian Journal of Forest Research 22, 304-309.

642

Lundqvist, L., Chrimes, D., Elfving, B., Mörling, T. and Valinger, E. 2007 Stand development after 643

different thinnings in two uneven-aged Picea abies forests in Sweden. Forest Ecology and 644

Management 238, 141–146.

645

Nilsen, P. and Strand, L.T. 2013 Carbon stores and fluxes in even- and uneven-aged Norway spruce 646

stands. Silva Fennica 47, 1–15.

647

Paradis, L., Thiffault, E. and Achim, A. 2019. Comparison of carbon balance and climate change 648

mitigation potential of forest management strategies in the boreal forest of Quebec (Canada).

649

Forestry cpz004, https://doi.org/10.1093/forestry/cpz004.

650

Peura, M., Burgas, D., Eyvindson, K., Repo, A. and Mönkkönen, M. 2018 Continuous cover forestry is 651

a cost-efficient tool to increase multifunctionality of boreal production forests in Fennoscandia.

652

Biological Conservation 217, 104–112.

653

Pukkala, T. 1987a Kuusen ja männyn siemensadon ennustemalli (In Finnish). Silva Fennica 21(2), 654

135–144.

655

Pukkala, T. 1987b Simulation model for natural regeneration of Pinus sylvestris, Picea abies, Betula 656

pendula and Betula pubescens. Silva Fennica 21, 37–53.

657

Pukkala, T. 2016 Plenterwald, Dauerwald, or clearcut. Forest Policy and Economics 62, 125–134.

658

Pukkala, T., Lähde, E. and Laiho, O. 2009 Growth and yiled models for uneve-sized forest stands in 659

Finland. Forest Ecology and Management 258, 207-216.

660

Pukkala, T., Lähde, E. and Laiho, O. 2010 Optimizing the structure and management of uneven-sized 661

stands in Finland. Forestry 83(3), 131–142.

662

Pukkala, T., Lähde, E. and Laiho, O. 2011a Metsän jatkuva kasvatus (In Finnish). Joen Forest Program 663

Consulting, Joensuu, Finland, 230 pp.

664

Pukkala, T., Lähde, E., Laiho, O., Salo, K. and Hotanen, J.-P. 2011b A multifunctional comparison of 665

even-aged and uneven-aged forest management in a boreal region. Canadian Journal of Forest 666

Research 41, 851–862.

667

Reyer, C.P.O., Bathgate, S., Blennow, K., Borges, J.G., Bugmann, H., Delzon, S., Faias, S.P., Garcia- 668

Gonzalo, J., Gardiner, B., Gonzalez-Olabarria, J.R., Gracia, C., Hernández, J.G., Kellomäki, S., 669

Kramer, K., Lexer, M.J., Lindner, M., van der Maaten, E., Maroschek, M., Muys, B., Nicoll, B., 670

Palahi, M., Palma, J.H.,Paulo, J.A., Peltola, H., Pukkala, T., Rammer, W., Ray, D., Sabaté, S., 671

Schelhaas, M.J., Seidl, R., Temperli, C., Tomé, M., Yousefpour, R., Zimmermann, N.E., and 672

Hanewinkel, M. 2017 Are forest disturbances amplifying or canceling out climate change-induced 673

productivity changes in European forests? Environmental Research Letters 12, 034027.

674

Räsänen P.K., Laitinen E., Peltonen A. and Rautiainen O. 1986 Forest regeneration in the six 675

southernmost forestry board districts of Finland. Results from the inventories 1978 – 1979. Folia 676

Forestalia 637, 1 – 30.

677

Roessiger, J., Ficko, A., Clasen, C., Griess, V. and Knoke, T. 2016 Variability in growth of trees in 678

uneven-aged stands displays the need for optimizing diversified harvest diameters. European 679

Journal of Forest Research 135, 283-295.

680

Routa, J., Kellomäki, S., Kilpeläinen, A., Peltola, H. and Strandman, H. 2011a Effects of forest 681

management on the carbon dioxide emissions of wood energy in integrated production of timber 682

and energy biomass. Global Change Biology, Bioenergy 3, 483–497.

683