Establishment of direct seeded seedlings of Norway spruce and Scots pine: Effects of stand conditions, orientation and distance with respect to shelter tree, and fertilisation

Available online at www.sciencedirect.com

Forest Ecology and Management 255 (2008) 1186–1195 www.elsevier.com/locate/foreco

Establishment of direct seeded seedlings of Norway spruce and Scots pine: Effects of stand conditions, orientation and distance with respect to shelter tree, and fertilisation Charlotta Erefur a,*, Urban Bergsten a, Michelle de Chantal b b

a SLU Silviculture & Vindeln Experimental Forests, S-92291 Vindeln, Sweden Department of Forest Ecology, P.O. Box 27, FIN-00014, University of Helsinki, Finland

Received 17 January 2007; received in revised form 13 September 2007; accepted 13 October 2007

Abstract The objectives of this work were to quantify the effects of stand stem density (SSD), orientation and distance with respect to shelter tree, and fertilisation on the establishment of Pinus sylvestris L. and Picea abies (L.) Karst. regenerated by direct seeding on different soil preparations. The field experiment was performed on South (648140 N, 198460 E, 225 m a.s.l.) and North (648090 N, 198360 E, 274 m a.s.l.) slopes in boreal Sweden. Regeneration and height growth in three SSDs with different light regimes, i.e. uncut forest (500 stems/ha), shelterwood (150 stems/ha), and clear-cut, were compared. Sowing was done in 2001, after using two soil preparations (mineral soil and a mixture of mineral soil and humus layer ground to a fine texture), at six distances to shelter trees (0.5, 1, 1.5, 2, 4 and 6 m). Half of the seedlings were also irrigated with fertiliser (10 mM N) from the second to the fourth growing season after seeding, i.e. until final inventories were made. The light environment did not differ significantly between different orientations and distances with respect to trees, but it was clearly different between SSDs. The establishment and growth of direct seeded seedlings depended on species and SSD as well as on soil preparation. On the North slope, the emergence was highest (50 seedlings in percent of germinable seeds for P. sylvestris and 44% for P. abies) in SSD 150. On the other hand, on the South slope the conditions in SSD 0 favoured the high emergence of P. sylvestris (41%), whereas for P. abies there was no difference between SSD 0 and SSD 150 (28% versus 30%, respectively). The soil preparation that created a mixture of mineral soil and humus layer generally favoured seedling emergence. Only for P. abies on the North slope, fertilised seedlings were taller (ca. 20%) than non-fertilised seedlings. After four years, P. abies on the North slope was most successful with nine seedlings in percent of germinable seeds remaining out of 24% emerged. The main conclusion is that for plant establishment from seed, the general conditions of the stand matter more than the orientation and distance with respect to the nearest tree and the light environment is more important than the nutritional status, i.e. light requirements cannot be moderated by nutrient supply. # 2007 Elsevier B.V. All rights reserved. Keywords: Frost heaving; Height growth; Light; Predation; Shelterwood; Soil preparation; Clear-cut; Dense forest

1. Introduction During the latter part of the 20th century, Fennoscandian forests have been harvested mainly by creating clear-cuts. This method has the advantage of being effective, but it is questioned. There is a growing interest in forestry-related amenity and recreation together with a debate on forest decline and certification, which both favour more environmentally friendly approaches to forest management (Pommerening, 2006). Furthermore, regeneration following clear-cutting is not

* Corresponding author. Tel.: +46 933 61576; fax: +46 933 10326. E-mail address: [email protected] (C. Erefur). 0378-1127/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2007.10.024

always successful as the local climate changes along with soil flora and soil nutrient and water contents (Ottosson Lo¨fvenius, ¨ rlander, 2001). At least in 1993; Marshall, 2000; Langvall and O small-scale forestry, other methods than clear-cutting, e.g. different types of continuous cover forestry systems (CCFS), are of interest. In areas that are difficult to regenerate, for example in wetlands, in areas with high risk of frost, or in stands where competition from the surrounding vegetation is severe, the use of shelterwoods, i.e. a form of CCFS, when regenerating forests can be advantageous. Shelter trees not only provide seeds for natural regeneration, but they also create a favourable environment for seedlings. The ground temperature in a day is more stable in a shelterwood than on a clear-cut since the

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195

overstorey trees to some extent prevent the long-wave radiation from leaving to the atmosphere. This will give less frequent summer frosts and less frost heaving but also less incoming radiation (Hannerz and Gemmel, 1994; Orlander and Langvall, 1997; Orlander and Karlsson, 2000). Other effects of shelter trees are that the seedlings will benefit from less wind exposure, more stable soil water content (on a clear-cut there can easily be an excess of water when the mature stand is removed, but also droughts as a result of more extreme temperatures), and less vegetation competition for seedlings (Holgen and Ha˚nell, 2000). However, leaving shelter trees when harvested may also bring disadvantages for the following regeneration. A high shelter tree density together with the location of seedlings compared to the surrounding trees may create a microclimate with competition for light, water, and nutrition (Malcolm and Ibrahim, 1993; Ottosson Lo¨fvenius, 1993; de Chantal et al., 2003b). Depending on growing site, mechanical disturbances due to snow and rain (Goulet, 1995; Winsa, 1995a) and predation (Nystrand and Granstro¨m, 1997b,a, 2000) will also affect seedlings. It is thus necessary to fully understand the complex picture of competition both between seedlings and between seedlings and mature trees. For optimum regeneration with a sufficient number of seedlings and good growth in an uncut forest the tree canopy has to be layered in a way that favours development of a forest stand with a gap structure close to that of a natural forest (Linder et al., 1997; Okland et al., 2003). In this work focus was on regeneration from seed. In a separate paper focus is on regeneration from plants (Erefur et al., in preparation). The overall aim of this work and the parallel work on plants is to understand how the environment could be moderated in combination with the selection of plant material to get a cost-effective regeneration in CCFSs. This research aims at optimizing silvicultural methods and at answering the following questions. Is it possible to say how sparse a forest stand has to be in order for individual distance to shelter tree to matter for seedling growth? At what distance will competition between shelter trees and seedlings then occur, and what type of competition will take place? Can fertilisation outweigh inadequate light conditions? The specific objectives of this study were to quantify the effects of stand stem density (SSD), orientation and distance with respect to shelter tree on seedling emergence, mortality, and height growth at two sites on slopes of opposite aspects. On height growth, the effect of

1187

fertilisation was also evaluated. Shade-tolerant P. abies (Picea abies L. Karst.) and shade-intolerant P. sylvestris (Pinus sylvestris L.) growing on two different soil preparations were compared, based on their light demands. Hypothesising that light was the main competitive factor, one further objective was to characterise the light environment in the different SSDs. 2. Materials and methods 2.1. Study site and experimental design The experiment was performed on a South slope (648140 N, 198460 E, 225 m a.s.l., 7.5% declination) and a North slope (648090 N, 198360 E, 274 m a.s.l., 12% declination) in Vindeln Experimental Forests, 60 km NW of Umea˚ in northern Sweden. Both slopes are spruce-dominated with a vegetation cover dominated by Vaccinium myrtillus L. on a moist podzolic soil with a texture of loamy sandy till (Ha¨gglund and Lundmark, 1982). Further information on the site characteristics is presented in Table 1. Three different types of stands were selected or created in May 2001 at each site, that is SSD 500 (500 stems/ha, i.e. dense forest), SSD 150 (150 stems/ha, i.e. thinned forest or shelterwood) and SSD 0 (clear-cut; created about 10 years before the thinned forest/shelterwood). All three experimental areas were ca. 50 m2, but whereas SSD 150 was thinned as a circle in the forest of SSD 500, both SSD 500 and SSD 0 were delimited within a larger area of SSD 500 and SSD 0, respectively. The climate in northern Sweden is temperate with a rather limited growing season mainly in June, July, and August. The weather conditions during the years of the experiment are listed in Table 2, showing that the summer of 2001 was wet whereas 2002 was dry and warm. The summers of 2003 and 2004 were closer to last decade normals in temperature and precipitation. Table 1 Stand characteristics

South slope North slope

Site index a

Age

Volume (m3 sk/ha)

Basal area (m2/ha)

G22 G20

144 111

340 182

37 27

The inventory was done in 1989 except for basal area that was measured in 2002. a Top height (m) in even-aged stands at 100 years of total age for Norway spruce.

Table 2 Weather conditions on the South slope (S) and North slope (N) Condition

Growing seasona 2001

Number of days Temperature sumb (day 8C) Precipitationc (mm) a b c

2002

2003

2004

S

N

S

N

S

N

S

N

169 995 561

169 1038 585

150 1228 216

150 1298 251

139 1062 392

139 1081 340

160 914 394

162 954 421

Period of the year when the daily mean temperature is above +5 8C. The sum of the mean temperature for all days with daily mean temperature above +5 8C. Measured 1.6 m above ground.

1188

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195

Fig. 1. Drawing of the experimental design in one stand stem density (SSD) with five selected trees.

The same experimental design was used on both slopes (see Fig. 1). Five shelter trees were randomly selected in the central parts of SSD 500 and SSD 150. In SSD 0, the randomly chosen locations of two hypothetical trees were marked with sticks. Two opposite blocks were delimited in the north-south direction from each selected (or hypothetical) tree. In each block, two sets of six parcels (30 cm  60 cm each) separated by 20 cm to 2 m (due to obstacles like rocks and roots) were marked at distances of 0.5, 1, 1.5, 2, 4 and 6 m from the shelter tree. No other tree was closer to the parcels than the selected shelter tree except at distances of 4 m (in SSD 500) and 6 m (in SSD 150 and 500). Each parcel was divided in two subparcels. Seedbeds were prepared in June 2001 using two soil preparations, HuMinMix and bare mineral soil, that were randomly allocated to each subparcel. HuMinMix, i.e. a mixture of humus layer and mineral soil ground to a fine texture (47% average loss on ignition on the South slope and 66% on the North slope), was done mechanically using a rototiller mounted on a clearing saw (Winsa, 1995b). To prepare the mineral soil seedbed, the humus layer was manually removed. To improve seedling emergence, micropreparation was done manually before sowing with a tool (8 cm  22 cm) consisting of 10 adjacent inverted square pyramids (5  2 pyramids, each 4 cm  4 cm and 2 cm deep) that make indentations in the seedbed (Winsa and Bergsten, 1994). The tool was pressed twice on each subparcel to create two series of 10 indentations. Ten seeds each of P. sylvestris and P. abies were sown in each subparcel (one species per series of indentations, randomly allocated) in June 2001. The seeds used in this study originated ¨ stteg 65810 N) and Norway spruce (Hissjo¨ from Scots pine (O 0 6481 N) seed orchards, with germination capacities of 96.2% and 97.3% (ISTA, 1985), respectively. From each pair of parcels at each specific distance and in each direction from the selected shelter (or hypothetical) trees, one was randomly chosen to be fertilised. Between 2002 and 2004 fertilisation was performed on each slope according to standard practices in commercial nurseries, i.e. one day every

other week starting from the first week of June until the end of August (six occasions; Gulin, personal communication). A water solution of N, P and K (10 mM N; Superba S, Hydro Supra AB, Landskrona, Sweden) was showered onto the chosen parcels based on the required amount for 40 seedlings (the highest amount of 1.2 l per parcel was during the second, third and fourth occasions), giving a total amount of 0.09 g of nitrogen/season and seedling. 2.2. Inventories Inventories were done in July 2001–2003 and at the end of the growing season from 2001 to 2004. Exceptionally, an inventory was done also in the spring of the second growing season (2002). During these inventories, all emerged (seedlings that appeared above the soil surface in 2001) or surviving (2002–2004) seedlings were recorded, as well as injury and mortality due to predation (cotyledons totally or partly missing due to browsing, mainly by rodents, birds and slugs), frost heaving (total or partial expulsion of the seedling’s roots from the soil), physiological (change in needle colour) and mechanical damage (broken), and mortality due to unknown causes (missing seedling). Mean values were calculated per 10 seed-plot. Maximum seedling emergence (hereafter referred to as emergence) was calculated as the highest number of living seedlings out of germinable seeds (laboratory germination test according to ISTA (1985) at 20 8C) per parcel during the first growing season. Height was measured from the root scar/soil surface to the tip of the terminal bud at the end of each growing season. A mean value was calculated per 10 seed-plot. Predation (most likely by slugs, rodents, and birds) mainly occurred during the first summer, so the data from the inventory done at the end of 2001 was used for analysing predation. For analysing frost heaving, the spring inventory in 2002 was used, since frost heaving happened in late 2001/early 2002. The values referred to for height growth are the ones measured at the end of the fourth growing season (2004).

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195

2.3. Determination of light environment The light environment in the different SSDs was characterised with the total site factor (TSF, the relative amount of incidental, direct + diffuse, PPFD, photosynthetic photon flux density, that penetrates below canopy for a specified period of time) by taking fish-eye photos during totally or partly overcast days in July 2006. Photos were taken at a height of approximately 1 m, by placing the camera with a fish-eye lens in a self-levelling mount in the north-south direction, at a distance of 1.25 and 5 m north and south of the three randomly selected shelter trees in each SSD at each slope (i.e. 24 photos from each slope). In SSD 0 a photo was taken at the point of each hypothetical shelter tree (i.e. two photos from each slope). The photos were analyzed using Regent Instruments computer software WinScanopy (2005) with the pixel classification done in grey scale and the time for growing season set to June 1– September 30.

1189

otherwise all three levels of SSD were included in the model. When significant effects (0.05 level) were found, Tukey’s multiple comparison test was used to test differences between treatments. Since there were large differences between treatment variances when analysing seedling emergence and mortality, the analyses were also performed on log-transformed values [log 10], according to Log [( p + 0.5)/(1.5  p)] (Sabin and Stafford, 1990) where p is the proportion of emerged seedlings out of germinable seeds. However, since the results of the analysis of variance were almost identical for untransformed and transformed values, the data were not transformed. The effects of SSD, orientation and distance in relation to tree, as well as their interactions, on TSF were tested using MiniTab’s general linear model (Anon, 2000) according to the same statistical model as explained above. 3. Results

2.4. Statistical analyses

3.1. Photosynthetic photon flux density

The effects of SSD, orientation and distance with respect to tree, soil preparation, as well as their interactions, on seedling emergence, height, predation, and damage due to frost heaving were tested using analysis of variance with fixed, crossed factors (MiniTab GLM; Anon, 2000). An example of the model is presented in Table 3. The effect of fertilisation on seedling height was also tested, along with its interactions with the other factors. Since there were significant differences in emergence and height between species and slopes, the data were analysed separately for the North and South slopes, as well as for species. Predation was tested separately for each slope, regardless of species since its effect was not significant. Frost heaving was tested separately for each species and only on the North slope, because frost heaving mainly occurred at that site. The effects of orientation and distance to tree were tested using SSD 150 and 500 only, since orientation and distance with respect to tree were only hypothetical factors of the design with SSD 0;

An average for both slopes showed a homogenous light environment within each SSD, but clearly different environment between SSDs, with less than 20% of PPFD above the canopy reaching the forest floor, i.e. being transmitted, in SSD 500; 40% in SSD 150; and 80% in SSD 0 (see Fig. 2). No significant differences were found between the different orientations or distances in relation to tree. 3.2. Emergence 3.2.1. P. abies On the North slope, the emergence of P. abies in SSD 150 was approximately twice as high as the emergence in SSD 500 and SSD 0 (Table 3, example of one ANOVA out of 28 in total; Table 4). In addition, on the south side of trees, regardless of

Table 3 General linear model for the effects of stand stem density (SSD), orientation, distance to tree, and soil preparation on emergence of P. abies on the North slope d.f.

MS

p

SSD Orientation (O) Distance to tree (DT) Soil preparation (SP) SSD*O SSD*DT SSD*SP O*DT O*SP DT*SP Error

1 1 5 1 1 5 1 5 1 5 453

47885.9 1287.7 1035.5 8670.6 178.8 881.0 204.7 505.5 4405.9 212.0 592.2

0.000 0.141 0.122 0.000 0.583 0.192 0.557 0.512 0.007 0.877

Total

479

Note: For this dependent variable and this combination of species and slope, 3way and 4-way interactions were not significant and were therefore excluded from the model.

Fig. 2. TSF values in SSD 0, SSD 150, and SSD 500 averaged between North and South slopes during growing season 2006 (filled square) and one day in July 2006 (open circle). TSF values differed between SSDs at the 0.05 level of significance according to Tukey’s multiple comparison test. Error bars denote the standard error of TSF.

1190

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195

Table 4 Maximum emergence  standard error (percent of germinable seeds) of P. abies and P. sylvestris in stands of different densities on opposite slopes

Table 5 Maximum emergence  standard error (percent of germinable seeds) of P. abies and P. sylvestris on mineral soil and HuMinMix

Stand stem density (stems/ha)

Soil preparation

0

150

500

P. abies North slope South slope

20.8  2.6 b 27.8  2.2 AB

44.2  1.7 a 30.1  1.1 A

24.2  1.5 b 25.5  1.0 B

P. sylvestris North slope South slope

30.9  2.7 b 41.3  2.1 A

49.5  1.7 a 34.1  1.4 B

26.0  1.6 b 25.1  1.0 C

Comparisons are done by species and slope. Note: Means with different letters are different at the 0.05 level of significance according to Tukey’s multiple comparison test. Capital letters are for the South slope.

SSD, the emergence of P. abies was higher on HuMinMix than on mineral soil (Table 5). On the South slope, the emergence only differed between SSD 150 and SSD 500, being higher in the former (Table 4). In SSD 150, the emergence was about 40% higher on HuMinMix (35.1%) than on mineral soil (25.1%), while in SSD 500 and SSD 0 there was no difference between soil preparations. In addition, on HuMinMix the emergence was higher in SSD 150 than in both SSD 500 and SSD 0 (Table 5). On mineral soil, there was no difference between SSDs. There was a trend of high values of emergence at least 1.5 m away from trees on the north side, whereas the lowest values were between 1 and 2 m on the south side of trees (Table 6). 3.2.2. P. sylvestris On the North slope, the emergence of P. sylvestris in SSD 150 was higher than in SSD 0 and SSD 500, both on mineral soil (not significantly though) and HuMinMix. In SSD 500, maximum emergence was much lower on mineral soil than on HuMinMix (Table 5). Furthermore, on mineral soil, maximum emergence was ca. 2 and 2.5 times as high in SSD 0 and SSD 150, respectively, as in SSD 500 (Table 5). On the South slope, in comparison, the emergence was highest in SSD 0 and lowest in SSD 500 (Table 4). The emergence was also higher on HuMinMix (34.8  1.2%) than on mineral soil (28.5  1.2%) and on the north side of trees (31.5  1.3%) than on the south side (27.7  1.2%). Further-

P. abies North slope Orientation North side South side

Mineral soil

HuMinMix

31.3  2.3 b 28.5  2.3 b

33.7  2.4 b 43.1  2.6 a

South slope Stand stem density (stems/ha) 0 30.5  3.4 ab 150 25.1  1.5 b 500 23.3  1.4 b P. sylvestris North slope Stand stem density (stems/ha) 0 36.9  4.3 bc 150 48.1  2.4 ab 500 19.4  1.7 d

25.1  2.8 b 35.1  1.6 a 27.7  1.4 b

25.1  3.0 cd 50.8  2.6 a 32.5  2.6 c

Comparisons are done by species and slope. Note: Means with different letters are different at the 0.05 level of significance according to Tukey’s multiple comparison test.

more, the emergence 1 m away from trees was lower than the emergence 4 and 6 m away (Table 6). 3.3. Mortality For P. abies seedlings, damage by frost heaving was highest on mineral soil in SSD 0, and even though the value was double, it did not differ from that in SSD 150 on mineral soil (Table 7). On the other hand, for P. sylvestris frost heaving was highest in SSD 0, irrespectively of soil preparation (Table 7). Predation (for all seedlings irrespectively of species) was highest in SSD 0 on the North slope, but in SSD 150 on the South slope (Table 8). Predation was lowest in SSD 150 on the North slope and in SSD 0 on the South slope. On the North slope, there were no differences between distances in SSD 500, whereas in SSD 150 predation was higher at 6 m away from trees than at 0.5 and 1 m (Table 8). On the South slope there were no differences between distances in SSD 150, but in SSD 500 predation was lowest at 1 m away from trees (not

Table 6 Maximum emergence  standard error (percent of germinable seeds) of P. abies and P. sylvestris at different distances from trees on the South slope Distance from tree (m)

P. abies

P. sylvestris

North side 0.5 1 1.5 2 4 6

26.3  2.3 23.9  2.4 31.6  2.7 36.4  2.6 32.8  3.1 30.8  2.7

South side ad bcd ac a ab ad

29.8  2.6 22.2  2.1 20.7  2.0 20.2  1.9 31.1  2.8 27.5  2.9

North and South sides ad bcde cde de ad ad

26.4  2.2 24.6  1.8 28.4  2.3 30.9  2.0 33.9  2.0 33.4  2.3

AB B AB AB A A

Comparisons are done by species. Note: Means with different letters are different at the 0.05 level of significance according to Tukey’s multiple comparison test. Capital letters are for average values of both orientations.

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195

1191

Table 7 P. abies and P. sylvestris seedlings damaged by frost heaving  standard error (%) on the North slope P. abies

P. sylvestris

Stand stem density (stems/ha)

Mineral soil

HuMinMix

Mineral soil and HuMinMix

0 150 500

14.3  5.2 a 6.7  2.0 ab 1.9  0.9 b

1.4  1.0 b 1.7  0.5 b 3.9  1.7 b

15.6  5.7 A 4.6  1.0 B 1.5  0.6 B

Comparisons are done by species. Note: Means with different letters are different at the 0.05 level of significance according to Tukey’s multiple comparison test. Capital letters are for average values of both soil preparations. Table 8 Predation  standard error (%) of P. abies and P. sylvestris seedlings in stands of different densities on opposite slopes

experiment, from emerged to established, is presented in Fig. 4 which shows the rather rapid stabilisation in number of seedlings after two years.

Stand stem density (stems/ha)

3.4. Height growth

0

150

500

North slope 0.5 1 1.5 2 4 6

53.9  3.3 A

19.7  1.3 13.1  2.6 16.0  3.0 17.0  2.5 18.1  3.2 21.8  2.9 32.5  3.8

C c bc ac ac ac a

27.4  1.7 23.4  5.0 24.2  3.6 29.4  3.7 30.3  3.9 34.0  4.8 23.1  3.7

B ac ac ab ab a ac

South slope 0.5 1 1.5 2 4 6

26.9  2.5 C

51.9  1.7 52.6  4.3 50.4  4.2 51.9  4.0 61.6  3.9 47.1  3.9 48.9  4.3

A ab ab ab a ab ab

38.9  1.6 39.6  4.3 26.8  3.9 37.8  4.2 36.1  3.7 46.3  3.9 46.2  3.8

B bc c bc bc ab ab

Comparisons are done by slope. Note: Means with different letters are different at the 0.05 level of significance according to Tukey’s multiple comparison test. Capital letters are for average values of SSD.

significantly lower than at 0.5, 1.5, and 2 m though). Predation was a great cause of mortality and a high number of seedlings were also killed by something other than predation or frost heaving (Fig. 3). The decline in number of seedlings during the

On both the North and South slopes, seedlings of both species were significantly taller in SSD 0 than in SSD 150 in autumn 2004 (Fig. 5). There were very few seedlings left in SSD 500 in 2004 and their heights were ca. 43–65% of the heights in SSD 150, depending on slope and species. Only for P. abies on the North slope, fertilised seedlings were significantly taller than non-fertilised ones (9.4  0.6 cm versus 7.9  0.4 cm, respectively). 4. Discussion 4.1. Light environment The analysis of the fish-eye photos showed expected differences between SSDs. The fact that there were no significant differences between orientation or distance with respect to tree, could still mean that there is variation during a day between microsites within the stand (Lieffers et al., 1999). We compared the radiation values obtained from the analysis of our fish-eye photos with ongoing reference measurements at a

Fig. 3. Mortality causes and establishment (seedlings in percentage of germinable seeds, with maximum emergence as the highest value).

1192

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195

overestimation of PPFD. An overestimation could also be due to an incorrect value for turbidity in the software. We were anyway interested in the TSF, which tells us the relative amount of incidental (direct + diffuse) radiation that penetrates below canopy for a specified period of time. 4.2. Emergence

Fig. 4. Seedlings in percentage of germinable seeds (starting with maximum emergence) for P. abies, North slope (filled square) and South slope (open square) and P. sylvestris, North slope (filled circle) and South slope (open circle) during the time of the experiment (starting in the beginning of growing season 2001 and ending at the end of growing season 2004). Error bars denote the standard error of seedlings in percentage of germinable seeds.

clear-cut at Svartberget field station (648140 N, 198460 E). During the days in July when the fish-eye photos were taken, the reference values of PPFD were ca. 20% lower than the values obtained with WinSCANOPY. The main reason for this could be that the photos were taken on totally or partly overcast days to get good contrasts for the process of analysing, but the program cannot calculate with clouds, which results in an

The emergence results are likely related to the higher radiation (and thus increased temperature) that reached the ground in SSD 0 and SSD 150 than in SSD 500 (Fig. 2). Also greater water availability in SSD 0 due to the lack of competition from shelter trees gave higher soil moisture (Taskinen et al., 2003) which can be advantageous to seedling emergence but for Pinus light has been shown to be more crucial (Beon and Bartsch, 2003). Concurrent with results obtained by Brang (1998) for the emergence of P. abies on slopes of opposite aspects, conditions in SSD 150 on the North slope were clearly more advantageous to seedling emergence of both species compared to conditions on the South slope. In contrast, for both species conditions in SSD 0 were more favourable on the South slope, but for SSD 500 slope aspect had no effect. A pure organic matter substrate will desiccate under high direct radiation, thereby reducing seedling emergence (Brang, 1998). A way to overcome this problem could be by incorporating mineral soil into the organic matter. Considering that emergence on HuMinMix was either higher or not significantly different from emergence on mineral soil, our results show that HuMinMix soil preparation offers the benefit

Fig. 5. Seedling height of P. abies (A) and P. sylvestris (B) on the North slope and P. abies (C) and P. sylvestris (D) on the South slope at the end of growing seasons 2002–2004. Open and filled squares are for stand stem density (SSD) 0 and 150, respectively. Error bars denote the standard error of seedling height.

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195

of a mineral soil substrate (high water retention capacity) without the adverse effect of substrate drying. HuMinMix should provide nutrients for future seedling growth through the incorporated humus. In accordance with the description of the light environment, orientation and distance from tree did not have significant main effects on emergence, but for the emergence of P. abies on the South slope they had a significant interaction. In this case, light may not be the only explanatory factor, though. For example, there could be less precipitation in combination with high radiation peaks on the south side of trees close to the stem. The comparison of seedling emergence between the north and south sides of trees may be irrelevant in the forest, as the north side of one tree becomes the south side of the next. However, in a sparse shelterwood stand, the orientation may play a more important role for seedling emergence. 4.3. Mortality The two most common causes of seedling damage and mortality were predation during the first growing season, and frost heaving over the first winter. In our study, seedlings growing in either SSD 0 or SSD 150 were more susceptible to predation than seedlings in SSD 500. The stand density that sustained the highest predation differed according to slope orientation though, but we could not find a good explanation for that result. Our results on seedling mortality by predation disagreed with those of Nystrand and Granstro¨m (2000) who reported that seedling predation by slugs decreased from uncut forest to shelterwood stands to clear-cut. This may be due to the fact that seedlings in our stands were predated more by birds (germinants), rodents, or other animals than by slugs. In SSD 150 on the North slope, seedling predation increased with distance from the shelter trees. This may be because seedlings are less visible in the shade created by the crown of trees. Frost heaving is a common cause of seedling injury and mortality during the first autumn and spring after seedling establishment in the boreal forest (Winsa and Bergsten, 1994; Goulet, 1995; de Chantal et al., 2003a, 2007). In this process, seedlings are either killed or injured when they are completely or partly pushed out of the soil by needle ice (Goulet, 1995). This action will break their roots or leave them partly exposed above the soil surface, making them more susceptible to desiccation (de Chantal et al., 2003a) and reducing their growth (Goulet, 1995; de Chantal et al., 2004). In this study, P. abies seedlings growing on mineral soil in SSD 0 incurred the most damage (injury and mortality combined) due to frost heaving. This result is not surprising since the risk of frost heaving is increased where temperature fluctuations are large, i.e. in clearcuts and canopy gaps (Goulet, 1995; Orlander and Langvall, 1997; de Chantal et al., 2007). Also, frost heaving is more common on exposed or disturbed soil than on soils with an intact structure (de Chantal et al., 2003a, 2006, 2007). In contrast, with HuMinMix, there was an insulating layer of humus on the soil surface, which reduces frost heaving (Winsa, 1995b; Sahle´n and Goulet, 2002).

1193

The first few years of seedling establishment are critical as seedlings are especially vulnerable to extremes of climate, predation, and pathogens. As our results show, many seedlings died due to other causes than the ones we could confirm (predation and frost heaving). After four growing seasons, there were 8.8 seedlings in percent of germinable seeds out of 23.8% emerged for P. abies on the North slope, 7.3% out of 30.2% for P. sylvestris on the North slope, 4.6% out of 22.7% for P. abies on the South slope, and 4.5% out of 28.7% for P. sylvestris on the South slope. Consequently, site conditions are important for seedling establishment and play a major role in determining the success of regeneration. Winsa (1995a) concludes that in addition to good quality of seeds, soil and microsite preparation are crucial for seedling emergence. de Chantal et al. (2004) show that exposed E/B horizon is a good alternative when regenerating P. sylvestris forests through direct seeding or natural regeneration, while the findings of Oleskog and Sahle´n (2000) indicate that Scots pine seeds germinate if proper seedbeds, i.e. the humus and intact substrates, are prepared prior to clear-cutting to maintain seed moisture contents high enough to allow germination. P. sylvestris and P. abies seedlings are less susceptible to frost heaving when planted on top of mounds or in humus according to the studies of Sahle´n and Goulet (2002). Jonsson (1999) reported that P. sylvestris was more sensitive to site conditions in a radius of 3 m than P. abies, in the establishment phase, but that in the early growth phase, site conditions affected P. abies more strongly than P. sylvestris. As such, the survival of P. sylvestris is influenced by grass coverage (i.e. shade) and allelopathic effects. 4.4. Height growth Scots pine and Norway spruce have fundamentally different requirements for growth, the former being a shade-intolerant and the latter a more shade-tolerant species. Despite this basic difference, both species will benefit in growth from increased radiation obtained by canopy gaps (de Chantal et al., 2003b) or clear-cuts, as observed in our results. For several species it has been shown that the maximum survival and growth does not occur in full light, though the optimum level of light tends to increase with the shade intolerance of the species (Lieffers and Stadt, 1994; Dai and Dai, 1996). Carter and Klinka (1992) have also shown that for some North American conifer species increasing soil moisture needs to be met with increasing PAR light in order to maintain the same growth rate. The absolute amount of light reaching the ground first of all depends on geographic location, and then it is the stand density, and size and orientation of a possible gap, which determine the amount of light (Poulson and Platt, 1989). The geographical variation is thought to be of greatest importance at higher latitudes with low sun angles so that east-west oriented gaps result in better illumination of the northern edge trees than shaded southern portions of the gap. Also larger gaps are required on north facing slopes in order to produce a desired light regime (Malcolm et al., 2001). The height and structure of the stand surrounding the gap greatly influence the proportion of light that is received as direct beam radiation compared to

1194

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195

diffuse. Careful matching of tree species to gap size and gap position can minimize early mortality and maximize growth rates. Opening sizes need not to be very large (0.1–0.2 ha minimum) in order for species to achieve growth rates similar to those found in the open conditions of a clear-cut (Coates, 2000). Only for P. abies on the North slope, fertilised seedlings were taller than non-fertilised ones. For all other slope and species combinations, after four growing seasons, the heights of fertilised and non-fertilised seedlings did not differ in SSD 0 and SSD 150. Considering this result, one may question the usefulness of fertilisation as a means to moderate the environment for coniferous seedling establishment. Both P. abies and P. sylvestris will thrive better in a competition-free environment. However, shade-intolerant species are more susceptible to belowground competition than shade-tolerant species (Tryon and Chapin, 1983). Calamagrostis, a common competitor, has efficient N uptake and can easily overtake seedlings (Hangs et al., 2004). In our fertilised parcels, there was also a lot of Luzula pilosa which became more abundant and grew taller with higher amounts of nitrogen available and therefore possibly competed with tree seedlings. This was especially the case in SSD 0. In SSD 500 light clearly seems to be limiting growth, while in SSD 150 on the North slope P. abies had enough light to make use of the fertilisation. The presence of a shelter tree can benefit seedlings, as there may be less competition from vegetation near the stem of adult trees, which promotes abundant regeneration (Kuuluvainen and Pukkala, 1989). However, adult trees may compete with seedlings for water and nutrients. Shelter trees also create shade, thus reducing the amount of light available for photosynthesis, which may eventually limit seedling growth (Mitchell, 2001). In a Loblolly pine stand a trenching experiment showed that the largest effect of overstory trees on understory development was due to light attenuation rather than competition for soil resources (Messina and Jenkins, 2000). Nevertheless, in this experiment, the presence of a shelter tree did not affect the height growth of seedlings during the first four years. The effect of competition from shelter trees may appear later when seedlings are larger and require more resources. The fact that no effect of the shelter trees could be found may be due to the restricted selection of those trees. There could have been an effect from other shelter trees than the chosen ones. However, this is unlikely except in certain places at distances of 4 or 6 m. The relationship between regeneration, gap structure, and silvicultural system has only recently been considered (Coates and Burton, 1997). Guidance on which pattern and intensity of canopy manipulation that results in satisfactory regeneration of a desired species, is needed in order to achieve successful CCFS (Malcolm et al., 2001). A good help when choosing the most suitable plant material (species and regeneration method) could be a light environment characterisation (such as in Fig. 2). The main conclusion of this study is that for plant establishment of direct seeded P. abies and P. sylvestris the general conditions of the stand, such as. . ., matter more than the individual orientation and distance with respect to the nearest shelter tree. The environment was homogenous for seedling

development, even in SSD 150, where one may expect more variation in shade and light conditions due to the sparse distribution of trees. The lack of effect of fertilisation together with light measurements indicates that light was the limiting factor for growth. For other types of plant material the result could be different though. Therefore, in Erefur et al. (in preparation) the response of planted seedlings, which have already finished the germination and early establishment phase, will be evaluated in relation to different distances from trees in different SSD. Acknowledgements This study was financed by the research program ‘‘Utilization of the boreal forest’’ and the European Union structural funds. We wish to thank the staff at Vindeln Experimental Forests for help with fieldwork and So¨ren Holm for help with the statistical analysis. References Anon, 2000. MINITAB statistical Software. Beon, M.-S., Bartsch, N., 2003. Early seedling growth of pine (Pinus densiflora) and oaks (Quercus serrata, Q. mongolica, Q. variabilis) in response to light intensity and soil moisture. Plant Ecol. 167, 97–105. Brang, P., 1998. Early seedling establishment of Picea abies in small forest gaps in the Swiss Alps. Can. J. Forest Res. 28, 4–639. Carter, R.E., Klinka, K., 1992. Variation in shade tolerance of Douglas fir, western hemlock and western red cedar in coastal British Columbia. Forest Ecol. Manage. 55, 87–105. Coates, K.D., 2000. Conifer seedling response to northern temperate forest gaps. Forest Ecol. Manage. 127, 249–269. Coates, K.D., Burton, P.J., 1997. A gap-based approach for development of silvicultural systems to address ecosystem management objectives. Forest Ecol. Manage. 99, 337–354. Dai, X., Dai, X.B., 1996. Influence of light conditions in canopy gaps on forest regeneration: a new gap light index and its application in a boreal forest in east-central Sweden. Forest Ecol. Manage. 84, 187–197. de Chantal, M., Leinonen, K., Ilvesniemi, H., Westman, C.J., 2003a. Combined effects of site preparation, soil properties, and sowing date on the establishment of Pinus sylvestris and Picea abies from seeds. Can. J. Forest Res. 33, 931. de Chantal, M., Leinonen, K., Kuuluvainen, T., Cescatti, A., 2003b. Early response of Pinus sylvestris and Picea abies seedlings to an experimental canopy gap in a boreal spruce forest. Forest Ecol. Manage. 176, 321. de Chantal, M., Leinonen, K., Ilvesniemi, H., Westman, C.J., 2004. Effects of site preparation on soil properties and on morphology of Pinus sylvestris and Picea abies seedlings sown at different dates. New Forests 27, 159. de Chantal, M., Rita, H., Bergsten, U., Ottosson Lo¨fvenius, M., Grip, H., 2006. Effects of soil properties and soil disturbance on frost heaving of mineral soil: a laboratory experiment. Can. J. Forest Res. 36, 2885–2893. de Chantal, M., Holt Hanssen, K., Granhus, A., Bergsten, U., Ottosson Lo¨fvenius, M., Grip, H., 2007. Frost heaving damage to one-year-old Picea abies seedlings increases with soil horizon depth and canopy gap size. Can. J. Forest Res. 37, 1236–1243. Erefur, C., Bergsten, U., Lundmark, T. de Chantal, M., Establishment of planted seedlings of Norway spruce and Scots pine: effects of stand conditions, shelter tree distance, and fertilisation, in preparation. Goulet, F., 1995. Frost heaving of forest tree seedlings: a review. New Forests 9, 67–94. Gulin, L., 2000. Personal communication. SCA Forest Products. Bogrundet Nursery, 860 30 So¨rberge.

C. Erefur et al. / Forest Ecology and Management 255 (2008) 1186–1195 Ha¨gglund, B., Lundmark, J.E., 1982. Handledning i bonitering med Skogsho¨gskolans boniteringssystem. Del 1–3. Forestry, S.N.B.o. Jo¨nko¨ping, Sweden. Hangs, R.D., Greer, K.J., Sulewski, C.A., 2004. The effect of interspecific competition on conifer seedling growth and nitrogen availability measured using ion-exchange membranes. Can. J. Forest Res. 34, 754–761. Hannerz, M., Gemmel, P., 1994. Granfo¨ryngring under ska¨rm – sammanfattning. Skogforsk, Redogo¨relse nr 4, 8–10. Holgen, P., Ha˚nell, B., 2000. Performance of planted and naturally regenerated seedlings in Picea abies-dominated shelterwood stands and clearcuts in Sweden. Forest Ecol. Manage. 127, 129–138. ISTA, 1985. International rules for seed testing 1985. Seed Sci. Technol. 13, 300–520. Jonsson, B., 1999. Stand establishment and early growth of planted Pinus sylvestris and Picea abies related to microsite conditions. Scand. J. Forest Res. 14, 425–440. Kuuluvainen, T., Pukkala, T., 1989. Effect of Scots pine seed trees on the density of ground vegetation and tree seedlings. Silva Fennica 23, 159–167. ¨ rlander, G., 2001. Effects of pine shelterwoods on microclimate Langvall, O., O and frost damage to Norway spruce seedlings. Can. J. Forest Res. 31, 155–164. Lieffers, V.J., Stadt, K.J., 1994. Growth of understory Picea glauca, Calamagrostis canadensis, and Epilobium angustifolium in relation to overstory light transmission. Can. J. Forest Res. 24, 1193–1198. Lieffers, V.J., Messier, C., Stadt, K.J., Gendron, F., Comeau, P.G., 1999. Predicting and managing light in the understory of boreal forests. Can. J. Forest Res. 29, 796–811. Linder, P., Elfving, B., Zackrisson, O., 1997. Stand structure and successional trends in virgin boreal forest reserves in Sweden. Forest Ecol. Manage. 98, 17–33. Malcolm, D.C., Ibrahim, K.G., 1993. Nutrient: Productivity relations in plantation-grown Sitka spruce in Scotland. Studia Forestalia Suecica 0, 87–94. Malcolm, D.C., Mason, W.L., Clarke, G.C., 2001. The transformation of conifer forests in Britain – regeneration, gap size and silvicultural systems. Forest Ecol. Manage. 151, 7–23. Marshall, V.G., 2000. Impacts of forest harvesting on biological processes in northern forest soils. Forest Ecol. Manage. 133, 43–60. Messina, M., Jenkins, J., 2000. Loblolly pine stand early development under reserve-tree silvicultural systems in east Texas. South J. Appl. Forestry 24, 11–16. Mitchell, A.K., 2001. Growth limitations for conifer regeneration under alternative silvicultural systems in a coastal montane forest in British Columbia, Canada. Forest Ecol. Manage. 145, 129–136. Nystrand, O., Granstro¨m, A., 1997a. Post-dispersal predation on Pinus sylvestris seeds by Fringilla spp.: ground substrate affects selection for seed color. Oecologia 110, 353–359.

1195

Nystrand, O., Granstro¨m, A., 1997b. Forest floor moisture controls predator activity on juvenile seedlings of Pinus sylvestris. Can. J. Forest Res. 27, 1746–1752. Nystrand, O., Granstro¨m, A., 2000. Predation on Pinus sylvestris seeds and juvenile seedlings in Swedish boreal forest in relation to stand disturbance by logging. J. Appl. Ecol. 37, 449–463. Okland, T., Rydgren, K., Okland, R.H., Storaunet, K.O., Rolstad, J., 2003. Variation in environmental conditions, understorey species number, abundance and composition among natural and managed Picea abies forest stands. Forest Ecol. Manage. 177, 17–37. Oleskog, G., Sahle´n, K., 2000. Effects of seedbed substrate on moisture conditions and germination of Scots pine (Pinus sylvestris) seeds in a mixed conifer stand. New Forests 20, 119–133. Orlander, G., Langvall, O., 1997. Markberedning och plantering. Meddelande – Skogsstyrelsen 2, 31–36. Orlander, G., Karlsson, C., 2000. Influence of shelterwood density on survival and height increment of Picea abies advance growth. Scand. J. Forest Res. 15, 20–29. Ottosson Lo¨fvenius, M., 1993. Temperature and radiation regimes in pine shelterwood and clear-cut area. Dissertation. Department of Forest Ecology, Swedish University of Agricultural Sciences. Umea˚, Sweden. Pommerening, A., 2006. Transformation to continuous cover forestry in a changing environment. Forest Ecol. Manage. 224, 227–228. Poulson, T.L., Platt, W.J., 1989. Gap light regimes influence canopy tree diversity. Ecology 70, 553–555. Sabin, T.E. Stafford, S.G., 1990. Assessing the need for transformation of response variables. Oregon State University, Special Publication 1–31. Sahle´n, K., Goulet, F., 2002. Reduction of frost heaving of Norway spruce and Scots pine seedlings by planting in mounds or in humus. New Forests 24, 175–182. Taskinen, O., Ilvesniemi, H., Kuuluvainen, T., Leinonen, K., 2003. Response of fine roots to an experimental gap in a boreal Picea abies forest. Plant and Soil 255, 503–512. Tryon, P.R., Chapin, F.S.I., 1983. Temperature control over root growth and root biomass in taiga forest trees. Can. J. Forest Res. 13, 827–833. Winsa, H., 1995a. Effects of seed properties and environment on seedling emergence and early establishment of Pinus sylvestris L. after direct seeding. Dissertation. Department of Silviculture, Swedish University of Agricultural Sciences. Umea˚, Sweden. Winsa, H., 1995b. Influence of rain shelter and site preparation on seedling emergence of Pinus sylvestris L. after direct seedling. Scand. J. Forest Res. 10, 167–175. Winsa, H., Bergsten, U., 1994. Direct seeding of Pinus sylvestris using microsite preparation and invigorated seed lots of different quality: 2-year results. Canad. J. Forest Res. 24, 77–86. WinScanopy, 2005. Re´gent Instruments Inc., Que´bec, Canada.