Effects on stem growth of Scots pine 33 years after thinning and/or fertilization in northern Sweden

ABSTRACT

Thinning and fertilization are two common and important stand treatments in forest management. In terms of area treated, thinning is the single most common form of stand treatment. The extent of forest fertilization on the other hand, has varied widely in recent decades and is currently not very common. Thinning is done primarily to promote stand properties while fertilization is done to increase growth before future final felling. After thinning stands of Scots pine, overall growth decreases, while growth of residual trees increases. An experiment was established outside Vindeln in northern Sweden where the long-term growth effects after thinning and/or fertilization were evaluated after 33 years. Experimental set-up was a randomized block design including 12 replications of four treatments. Treatments were control, fertilization, thinning, and thinning and fertilization combined. Thinning decreased overall and annual volume growth ha⁻¹, and increased green crown size and diameter growth at breast height (1.3 m, DBH) for the individual trees. No positive growth responses to fertilization could be seen after 33 years. In summary, this study showed that thinning can have long term effects on the growth of a Scots pine stand in northern Sweden. Possible reasons for the lack of positive response following fertilization are discussed.

KEYWORDS:

• Thinning
• fertilization
• long-term growth

Introduction

To increase diameter growth of individual trees, and to enable a higher economic return when second thinning and/or clear-cut is carried out several years ahead, thinning is carried out. In Sweden in 2015/2016 a total of 291,000 ha of productive forestland were thinned. The proportion of thinning of the total amount of extracted volume of all felling (73.8 million m³) during that season was 30.8% (Skogsdata 2017). Today, most stands are thinned one or more times during a rotation period. Stands of higher site indices tend to be thinned more often. The annually thinned area is also associated with thinning grade, i.e. the proportion of basal area or volume extracted. Lower thinning grades usually leads to a need to thin more often as the stand density increases quicker afterwards. In Sweden today it is common with thinning grades between 20% and 40%, i.e. with 20–40% of the standing basal area removed at thinning. For Scots pine (Pinus sylvestris L.) a direct effect of thinning on individual stems is that it promotes a change in growth distribution from the upper part of the stem to lower parts (Valinger 1992). Theories on how thinning affects different stand variables and different species have been discussed for a long time. For example, in a study by Li (1923) thinned plots with Weymouth pine (Pinus strobus L.) was compared with un-thinned control plots. There, the author reported that growth was greater in the thinned plots than in the control. However, the most common result after thinning today is that thinning reduces stand volume growth in Scots pine both in the short and long term, as shown by e.g. Eriksson and Karlsson (1997) and Mäkinen and Isomäki (2004). For the remaining individual tree, growth increases after thinning. These studies show that the greater the thinning grade, the lower the overall total growth.

Thinning is also expected to reduce competition for light among the remaining trees. In a study by Albaugh et al. (2017), thinning to different numbers of remaining trees was tested on Loblolly pine (Pinus taeda L.). Annual growth was measured on each single tree during six years. In their study, the individual trees diameter growth increased significantly with increasing thinning grade, while height growth was not affected to the same extent. Similar result with increased diameter growth after thinning is presented in a number of other studies with pine (cf. Mäkinen and Isomäki 2004; Juodvalkis et al. 2005; Moschler et al. 2007). However, these results are not entirely consistent with other more short-term studies. For example, Crecente-Campo et al. (2009), who studied Scots pine in Spain, found no significant growth increase for individual trees four years after thinning.

Forest fertilization is another forest management measure, aimed at raising soil productivity and thus increase the diameter growth of trees. Fertilization is done with nutrients that limit growth. Under boreal conditions it is almost always nitrogen (N) which is the most important growth-limiting factor in boreal forests (Tamm et al. 1999). There are experiments in which several nutrients other than nitrogen have been added to varying degrees. However, Jacobson and Pettersson (2010) found that no significant growth effects was found for nutrients other than nitrogen under boreal conditions. Fertilization has been a common treatment in forest management in both Sweden and Finland since the 1960s (Saarsalmi and Mälkönen 2001; Lindkvist et al. 2011). In Finland fertilization almost ceased totally in the 1980s (Saarsalmi and Mälkönen 2001). In Sweden, there was a peak during 1970s when almost 200,000 hectares year⁻¹ were fertilized (Lindkvist et al. 2011). That figure has since fallen, and during 2013 only 23,900 ha were fertilized (Swedish Statistical Yearbook of Forestry 2014). When fertilizing, a common dosage of fertilizer in Sweden is 150 kg ha⁻¹, which is expected to give an approximately growth increase of 10–20 m³ ha⁻¹ during a period of 7–11 years (Pettersson 1994). The Swedish Forestry Act (Skogsstyrelsen 2017) sets some limitations in Sweden for where fertilization is recommended. In the northern part of the country fertilization with 150 kg N can be performed three times during a rotation, whereas in the southernmost part fertilization is generally not permitted. Reasons for this are e.g. that the effect of fertilization at highly productive sites in southern Sweden is small and that the deposition of nitrogen in that area exceeds 10 kg N year⁻¹ (Zetterberg et al. 2006). Further, for environmental reasons it is recommended that zones be left unfertilized adjacent to water, wetlands, protected land, and key biotopes.

To be able to increase growth of the trees even further, i.e. to obtain an additive growth increase, fertilization directly after thinning have been shown in experiments to be a strategic measure (e.g. Möller and Pettersson 1979). Directly after thinning is performed the trees have an extended room for expanding the crowns with the aid of fertilizer leading to higher singletree crown growth and thus an overall increase in volume growth. This has been reported for several different species, like e.g. Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco), Scots pine, and lodgepole pine (Pinus contorta var. latifolia Loudon) (e.g. Brix 1981; Valinger et al. 2000; Pinno et al. 2012). An identified drawback with the combined treatment is that stands become more susceptible to damage from snow and wind afterwards, and that the periods of elevated susceptibility are extended compared to when the treatments are carried out on their own (Valinger and Lundqvist 1992).

The aim of this study was to evaluate if the growth effects 12 years after thinning and fertilization, singly and combined, for Scots pine also could be detected after 33 years. Hypotheses was that the initial positive volume growth effects of fertilization after 12 years in the same field experiment (cf. Valinger et al. 2000) had disappeared and that the loss of total stand volume growth as a result of thinning was still detectable. For this, a randomized block experiment established 1983 in northern Sweden was used.

Material and method

The study was performed at Svartberget experimental forest outside Vindeln in north Sweden (64°13′N, 19°46′E, 200 m a.s.l.). The experiment was established in a stand dominated with Scots pine (P. sylvestris L.). The stand was established in 1939 through direct seeding and natural regeneration. It was pre-commercially thinned in 1972 and when the experiment was established and first measured in 1983 the mean height was 12 m and the mean diameter at breast height (1.3 m, DBH) was 13.7 cm of the 1352 stems ha⁻¹ that was measured at that point of time (Table 1). The basal area was 20 m²ha⁻¹ and the total volume, calculated according to Näslund (1940), was 116 m³ha⁻¹ (Table 1). The site index (H₁₀₀), estimated according to Hägglund and Lundmark (1981), was T24 (dominating height 24 m at 100 years of age). This indicated a site quality of 5.2 m³ha⁻¹ year⁻¹. The soil was a mesic sandy silty till, and the ground vegetation was dominated by Vaccinium vitis idaéa L. and Vaccidium myrtillus L.

Table 1. Starting values before the treatments were executed in 1983 and number and proportion of trees cut during the 1983–2017 period.

Variable	Treatments
	T0F0	T0F1	T1F0	T1F1
DBH 1983, cm	13.6a	13.8a	13.7a	13.8a
Number of stems ha^−1 1983 before treatment execution	1368.5a	1350.9a	1355.6a	1334.3a
Number of stems ha^−1 after treatment execution 1983	1368.5a	1350.9a	738.0b	738.9b
Number of trees ha^−1 cut between 1983 and 2017	210.2a	179.6ab	138.0b	121.3b
Volume cut between 1983 and 1995, m^3	24a	25a	22a	19a

Note: Values not followed by the same letter are significantly different (p < .05).

The stand was split into 48 plots with an area of 900 m² each, i.e. a total area of 4.32 ha. The plots were sorted into 12 blocks of four treatments with basal area differences of less than 1 m²ha⁻¹. The four treatments were assigned randomly to the plots within each block, giving 12 replications of each treatment. Within each block the following treatments were allotted: Control (T₀ F₀), fertilization with 150 kg N ha⁻¹ (T₀ F₁), thinning with 40% basal area removal (46% of the stems) (T₁ F₀), and thinning and fertilization (T₁ F₁).

The thinning was performed in the autumn of 1983 and the manually applied fertilization was done with urea during the spring of 1984. Thereafter no other treatments have been executed in the stand except destructive measurements of eight trees per plot at two of the blocks annually between 1984 and 1989, i.e. a total of 384 sample trees. The destructively cut sample trees were chosen among trees with a DBH as close as possible to the arithmetic mean within each block (Valinger 1992). In 1990 an additional number of trees within six of the blocks were extracted, in two of them six trees and two in four with arithmetic means as close as possible to mean DBH in each plot as possible (Mörling and Valinger 1999). Furthermore, a few trees have been extracted due to damage from wind and snow. This has led to a statistically significant higher number of extracted trees from the control plots but not as regards the proportion of trees extracted (Table 1). During the period 1983–1995 the extracted total volume for the control plots were 24 m³ ha⁻¹, for fertilization 25 m³ ha⁻¹, for thinning 22 m³ ha⁻¹, and for thinning and fertilization 19 m³ ha⁻¹. As these figures did not differed statistically significant it was judged that the performed cuttings during the 33-year period did not affect the outcome of this study.

The experimental stand was re-measured for the first time in May 2017, i.e. in spring before the annual growth starts, where all trees were numbered and measured for DBH. Thereafter every 15th tree was allotted for measurements of height, crown height and double bark thickness (in total 256 sample trees). With these data, volume for each sample tree was calculated according function number 100–04 for Scots pine in northern Sweden above latitude 60°E (Brandel 1990) as$\begin{array}{rl}{V}_{100-04}=& {10}^{-1.2715}x{D}^{2.13211}x(D+20.0{)}^{-0.13543}x{H}^{1.58121}\\ & x\left(H-1.3\right)-0.73435x{K}^{0.06595}x{B}^{0.10998}\end{array}$

where V_100–04 = volume (dm³), D = DBH (cm), H = height (m), K = crown height (m), and B = double bark thickness (mm).

The calculated volume for the sample trees were inserted into an equation with the DBH for all measured trees. The resulting equation were then applied to all trees to calculate the volume (V, m³) for all trees, and thus for all the respective treatment plots. The resulting model had the following expression (Figure 1):$V=0.0244{\mathrm{e}}^{0.0109xD}$where D = DBH (mm). R² for this equation was 0.9491.

Figure 1. Volume (m³) as a function of DBH (mm).

The DBH for each treatment plot were calculated from the DBH of each measured tree. Thereafter figures for basal area and volume were multiplied to represent one hectare.

Multiple pairwise comparisons between treatment means were made using Tukey’s studentized test on all main effect means with the use GLM within the SAS statistical package (SAS Institute 2015). If p < .05, the result of the statistical analysis was regarded as significant.

Results

After 33 years mean DBH of individual trees was statistically larger after treatments that were thinned in comparison to un-thinned treatments (p < .05, Table 2). Variation in mean DBH varied between 18.5 and 23.8 cm. Diameter growth was faster after thinning in comparison with un-thinned treatments (p < .05, Table 2). Total variation in diameter growth during the 33 years period was between 5 and 11 cm.

Table 2. Values for the calculated variables measured in May (spring) 2017, i.e. after 33 years of growth following the treatments thinning and fertilization, singly and in combination.

	Treatments
Variable	T0F0	T0F1	T1F0	T1F1
Mean DBH, cm	19.4a	19.8a	22.4b	22.7b
Diameter growth 1983–2017, cm	5.8a	6.0a	8.7b	8.8b
Basal area 2017, m^2ha^−1	35.7a	37.3a	24.1b	25.7b
Height 2017, m	20.1a	20.4a	20.1a	20.1a
Crown height, m	12.5a	12.6a	11.2b	11.6b
Standing volume, m^3	257.8a	269.0a	182.7b	195.0b
Volume growth 1983–2017, m^3	142.1a	151.7a	112.2b	124.7b
Volume per tree, m^3	0.14a	0.15a	0.21b	0.23b
Annual growth, m^3ha^−1yr^−1	4.3a	4.6a	3.4b	3.8b
Self-thinning 1995–2017, %	1.4a	1.1a	1.7a	1.5a

Note: Values not followed by the same letter are significantly different (p < .05).

Standing basal area per hectare was approximately 30% lower after treatments including thinning in comparison to un-thinned treatments (p < .05, Table 2). Height growth was unaffected by all treatments after 33 years of growth (p > .05, Table 2). The mean heights varied less than 30 mm within each treatment and all sample trees were between 17.5 and 22.8 m in height. Crown height was significantly higher in un-thinned treatments in comparison with the thinned treatments (p < .05, Table 2). Total variation in crown height varied between 9.5 and 14 m.

After 33 years standing volume was larger in un-thinned treatments than after thinning (p < .05, Table 2). The difference was approximately 30%. Standing volume varied between 127 and 300 m³.

Volume growth during the 33 years period was larger in treatments that were not thinned in comparison with the ones that were thinned (p < .05, Table 2). The difference was approximately 20%. Volume growth during the studied period varied between 90 and 170 m³. Volume growth per tree was significantly higher for trees on thinned plots in comparison to trees on un-thinned plots (p < .05, Table 2). Annual growth was significantly larger in un-thinned treatments than after thinning (p < .05, Table 2). Total variation between the included treatments was 2.9 and 5.1 m³ ha⁻¹ year⁻¹. No significant differences in the proportion of dead trees due to natural self-thinning between 1995 and 2017 were found for the studied treatments (p > .05). The self-thinning within the treatments plots varied between 1.1% and 1.7%.

For all analyses made, no significant effect of fertilization was found between fertilization in combination with the un-thinned treatment or thinning (p > .05, Table 2).

Discussion

The main impression after the 33 years was that no statistically significant effects in growth following fertilization were detected. This was not totally unexpected as e.g. Pettersson and Högbom (2004) and Saarsalmi et al. (2006) have done similar observations. Studies not in agreement with these results are e.g. Binkley and Reid (1984) who show growth differences in Douglas-fir (P. menziesii (Mirb.) Franco) 20 years after fertilization. However, their study included dosages of nitrogen higher than 150 kg N per hectare. Eriksson and Karlsson (1997) who conducted a review of long-term fertilization experiments throughout Sweden showed that growth is higher in fertilized stands than in un-fertilized controls. Similar positive results after fertilization are also found in Finland by Laakkonen et al. (1983) and in Norway by Børja and Nilsen (2009). The experiment evaluated in the present study showed a positive volume growth response for single trees following fertilization during the first five years after treatment (Valinger 1992), probably an effect of increased needle and branch growth. After 12 years, the positive volume growth effect after fertilization was undetectable for single trees but still somewhat higher than control for trees in the combined thinning/fertilization treatment (Valinger et al. 2000). A lack of positive effect 15 years after fertilization for lodgepole pine (P. contorta Dougl. var. latifolia) was found by Ghebremichael et al. (2005) and for Douglas-fir by Brix (1982).

The immediate response to fertilization could explain the results. During the first three years, there was a 29% increase in needle weight after fertilization in the un-thinned treatment and by 42% in the thinned treatment (Valinger 1993). A growth increase in the crown, represented by branch elongation and shoot axes produced, most pronounced in the upper part of the crown, was the result of fertilization initially. In a study with Douglas-fir Brix (1981) reported a similar distribution of growth during the first years after fertilization. In the present study, the growth distribution most likely led to a more rapid closure of the canopy and a shading of lower branches. After 12 years, we could also see that the weight of the crown was lower after fertilization, which indicate that the crown had become smaller and thus has fewer needles left for wood production (Valinger et al. 2000). For the trees on thinned and fertilized plots, there was still a positive response to fertilization after 12 years. Now, after 33 years, a similar result as after fertilization singly appeared. In a similarly designed experiment with thinning and fertilization in Black spruce (Picea mariana (Mill.) B.S.P.) Weetman et al. (1980) also reported that most of the growth response to urea was gone after 10 years. In their case mostly due to mortality losses. A study by Omule et al. (2011) of growth effects on Douglas-fir 32 years after treatment also reports of the absence of positive effects of fertilization, on its own or in combination with thinning, on volume growth. A possible explanation is that the site index for the experimental site is higher than normal for north Swedish pine forests. The extent to which nitrogen fertilization affects growth is highly dependent on different site and stand characteristics (Jacobson 2005). The experimental site was well in line with his recommendations to fertilize, i.e. site indices between 16 and 30 m. Therefore, it is difficult to judge this as a reason for the lack of response. In his report, he announces prerequisites for a positive outcome of the fertilization.

Effects of thinning were clearer. They were especially pronounced for growth variables like DBH, crown height, and volume. At tree level, crown and DBH had grown larger after thinning but standing volume and volume growth were lower than for treatments including thinning. The increase in diameter growth of remaining trees after thinning cannot compensate for the loss in production of the removed trees. Enhanced diameter growth following thinning comes as no surprise, it has been reported in several studies, e.g. Mäkinen and Isomäki (2004), Juodvalkis et al. (2005), and Moschler et al. (2007). Eriksson and Karlsson (1997) and Mäkinen and Isomäki (2004) have earlier reported that thinning leads to decreased total volume growth for Scots pine. Trees remaining after thinning have more growing space and can therefore accumulate wider crowns. The thinned treatment showed a statistically significantly lower crown height despite the fact that no differences in total tree height could be detected. According to Oliver and Larson (1996) this can be explained by the increased light in the stand following thinning. Branches live as long as they receive light, permitting photosynthesis in the needles can continue. No differences in crown height could be seen after fertilizing which indicates that it is light rather than nutrients that limit crowns length. Larger crowns is a prerequisite for a larger photosynthesizing apparatus at the single tree level, something normally leading to enhanced diameter growth. Furthermore, thinning leads to increased swaying of trees as the mechanical perturbation from abiotic factors such as wind and snow affect the remaining stems (Peltola 1996). These increased loads lead to increased lower bole growth (cf. Jacobs 1954; Peltola et al. 2002) which can also explain part of the results.

At establishment, there was concern about the risk of damage from snow and wind. Wind and snow damage did also occur, but not to an extent that could jeopardize the experiment. When the extent of damage was assessed 33 years after establishment, it was observed that the natural self-thinning had been low in the 1995–2017 period, and that it was almost equal between the treatments. Most damage had occurred during the first years after establishment, which is well in line with earlier findings in experiments on thinning and fertilization in e.g. Douglas-fir (Omule et al. 2011) and in Scots pine and Norway spruce (Elfving 2010). The low incidence of damage in the present study was surprising since a larger proportion of self-thinning because of crowding is likely to occur in un-thinned plots. There are, however, results presented by Rouvinen et al. (2002) and Elfving (2010) who conclude that mortality rates are low in mature un-thinned Scots pine stands. They report figures between 0.6% and 1.8% per year for western parts of Russia and Sweden, respectively. Most damage in these studies was found on small trees. A possible reason for the low rate of self-thinning in our study could be that the initial stem density was too low, or that the period studied was not long enough. Another reason could be that no heavy machinery was used when establishing the experiment and thus no strip roads were established, meaning that rooting conditions were unaffected throughout the 33-year period.

Growth during the 33-years period differed significantly between thinned and fertilized treatments. Both annual and total volume growth were higher in treatments that were un-thinned, and the same was found for the total standing volume. Results were in accordance with those previously published in similar studies (e.g. Eriksson and Karlsson 1997; Mäkinen and Isomäki 2004). No differences in volume growth between fertilized and un-fertilized treatments could be detected. After the initial effects had waned off (Valinger et al. 2000), there was uncertainty about what effects would last longer. In this study, no positive effects on volume growth after fertilization on volume growth could be detected. Trees on un-thinned and thinned plots had grown at the same rate regardless of the use of fertilization.

Urea was the fertilizer used in this study. The expected period of increased growth following urea fertilization was about six years in comparison with the, at the time of establishment, more commonly used ammonium nitrate, which has an expected time of increased growth of 7–11 years. Another aspect on the effects of the urea was used in this study is whether other effects would have been observed with other fertilizers, e.g. ammonium nitrate or Skog-CAN, or if fertilization had been done more than once? Finnish studies indicate varying results after the use of urea in comparison with other N fertilizers according to Saarsalmi and Mälkönen (2001). It is thus possible that a use of a fertilizer with a more extended effect could have enhanced volume production more. Choice of fertilizer was however limited by the risk of contaminating a river close to the experimental site, ammonium nitrate has a larger risk of polluting streams through leakage (Jacobson 2005). Today’s use of Skog-CAN which is the most common fertilizer when fertilizing and more environmental friendly (Letho 1995) would eventually have given a different result. Whether repeated fertilization would have resulted in a more positive outcome is also hard to evaluate. For Swedish conditions, Jacobson and Pettersson (2010) indicate positive results after repeated fertilization in Scots pine and no further increase by adding micronutrients. However, results from studies in British Columbia for lodgepole pine (P. contorta Dougl. var. latifolia Engelm.) indicate that repeated fertilization can lead to deficiencies for micronutrients as copper and iron which leads to premature loss of foliage (Amponsah et al. 2005) and thus potentially a decreased growth (Kishchuk et al. 2002). Also in this latter study a deficiency of micronutrients were detected that affected growth negatively and that they conclude that S is needed for a positive response to repeated N fertilization.

Our hypotheses of growth effects of thinning and fertilization, singly and combined, in northern Sweden proved correct. The initial positive volume growth effects of fertilization detectable some years after the experiment was established had waned off after 12 years (cf. Valinger et al. 2000). The decrease in total volume growth following thinning was still detectable. In this study, thinning was the treatment that affected trees and stand characteristics. It increased tree DBH and thereby volume growth at trees level. At stand level, thinning decreased total growth. Fertilization had no effects 33 years after treatment. Height growth was un-affected by any of the treatments. Thinning resulted in longer crowns and thus a possibility to retain lower branches for longer periods of time. Results indicate that fertilization in combination with first thinning cannot be recommended in northern Swedish forestry. Should fertilizer be used the recommendation from the present experiment would be in accordance with recommendations by Jacobson and Pettersson (2010) that fertilization should be carried out approximately a decade before final harvesting.

Disclosure statement

No potential conflict of interest was reported by the authors.