Macronutrient Stocks in Scots Pine Stands of Different Densities

A positive nutrient balance is crucial to sustaining forest productivity. Differences in stand densities usually mean different aboveground biomass stocks and different proportions of tree compartments. These differences can be reflected in the different macronutrient stocks between stands of different densities, because various tree compartments have different element concentrations. In this study, 82-year-old stands of Scots pine were compared, and specifically, the concentrations of the elements in tree compartments and the amounts of macronutrients in aboveground biomass were compared. The nutrients considered in this study were nitrogen, phosphorus, potassium, calcium, magnesium, and sulfur. A positive correlation between stand density and the level of stored macronutrients was found for nitrogen, phosphorus, and potassium. This result means that forest managers can influence nutrient balances by regulating stand densities or by harvesting methods (SOH: stem-only harvesting or WTH: whole-tree harvesting).

Nutrient stocks in forest ecosystems are changing due to natural processes and human operations-where some forests are experiencing some increase in nutrient content or availability, other forests are experiencing a depletion of their nutrient stocks in the soil.Many nutrient balance studies have been conducted to estimate these changes [8,[11][12][13][14].A nutrient budget is defined as the difference between nutrient inputs and outputs of an ecosystem over a specified time period [15].The budget studies are based on a mixture of empirical and assumed input and output levels of nutrients [16].In managed forests and forest plantations, harvesting is the main operation that causes nutrient losses.Nutrients removed from forest ecosystems via woody biomass can be compensated for by inputs from atmospheric deposition and, in mineral soils, by weathering [13,17].Maintaining a positive nutrient balance is essential to achieving sustainable forestry [8].
As forests are harvested more intensively, nutrient export from the ecosystem also increases [8,18].However, the magnitude of nutrient removal depends on the harvesting method, because nutrient concentrations differ between tree compartments [19].Branches, needles, and tree crowns generally have higher concentrations of nutrients than the stem [13,17,20].In conventional stem-only harvesting (SOH), these parts of trees, together with small unmarketable or damaged trees, constitute logging residues and are left on the site.In contrast, whole-tree harvesting (WTH), especially for bioenergy Forests 2018, 9, 593 2 of 11 purposes, includes the removal of these residues from the site.Nutrient balance calculations indicate that in many cases, nutrient removals by WTH exceed the replenishment rate of plant-available nutrient pools in soil by mineral weathering and atmospheric deposition [21][22][23][24].
Utilization of forest biomass for energy has substantially grown in recent years, and this trend is expected to continue to rise [8,13,25].This utilization of forests for biomass causes an increased demand for forest residues as an available energy source.However, complete removal of aboveground biomass could deplete soil nutrient pools and decrease site productivity over the long term [24,26].Therefore, it is important to know how forest management practices could affect nutrient stocks in the aboveground biomass of a stand so that harvesting intensity can be properly adjusted to preserve or improve a positive nutrient balance.
A main goal of forest management is to increase stand growth and improve its quality on a sustainable basis.Stand density is one of few things that can be controlled efficiently.Controlling stand density by initial spacing and thinning is a major tool for achieving this goal [27].Thinning reduces the number of trees on a site and increases the growth of the remaining trees [28][29][30].
The relationship between stand density and the accumulation of nutrients in biomass varies depending on the considered period (Figure 1).Over the long term, the number of trees decreases, but overall biomass increases with the nutrient stock [13,31].In the short term, the situation could be the opposite.Directly after thinning, the number of trees drops considerably and overall biomass and nutrient stocks are smaller than before thinning [32,33].However, a few years after thinning, after a period of stimulated growth, stand biomass could be the same as it was before.It is not yet clear if macronutrient stocks follow the same pattern.
logging residues and are left on the site.In contrast, whole-tree harvesting (WTH), especially for bioenergy purposes, includes the removal of these residues from the site.Nutrient balance calculations indicate that in many cases, nutrient removals by WTH exceed the replenishment rate of plant-available nutrient pools in soil by mineral weathering and atmospheric deposition [21][22][23][24].
Utilization of forest biomass for energy has substantially grown in recent years, and this trend is expected to continue to rise [8,13,25].This utilization of forests for biomass causes an increased demand for forest residues as an available energy source.However, complete removal of aboveground biomass could deplete soil nutrient pools and decrease site productivity over the long term [24,26].Therefore, it is important to know how forest management practices could affect nutrient stocks in the aboveground biomass of a stand so that harvesting intensity can be properly adjusted to preserve or improve a positive nutrient balance.
A main goal of forest management is to increase stand growth and improve its quality on a sustainable basis.Stand density is one of few things that can be controlled efficiently.Controlling stand density by initial spacing and thinning is a major tool for achieving this goal [27].Thinning reduces the number of trees on a site and increases the growth of the remaining trees [28][29][30].
The relationship between stand density and the accumulation of nutrients in biomass varies depending on the considered period (Figure 1).Over the long term, the number of trees decreases, but overall biomass increases with the nutrient stock [13,31].In the short term, the situation could be the opposite.Directly after thinning, the number of trees drops considerably and overall biomass and nutrient stocks are smaller than before thinning [32,33].However, a few years after thinning, after a period of stimulated growth, stand biomass could be the same as it was before.It is not yet clear if macronutrient stocks follow the same pattern.The Scots pine has the largest geographic distribution of all pine species and is probably the most widely distributed coniferous species in the world [29].The large geographic distribution is a  The Scots pine has the largest geographic distribution of all pine species and is probably the most widely distributed coniferous species in the world [29].The large geographic distribution is a consequence of a wide ecological amplitude, regarding both climatic conditions as well as soil requirements (from soils rich in humus to very poor).It is also an economically important species [30,34,35].To estimate how macronutrient stocks change, five sample stands of Scots pine (Pinus sylvestris L.) were compared.The stands were the same age and maturity, had different densities (476-836 trees per ha), and were 5-10 years past thinning.The aim of the study was to determine the influence of stand density on the macronutrient stocks (N, P, K, Ca, and Mg) in the aboveground biomass of mature Scots pine stands.
For Scots pine stands of a similar age, we predict that there will be a correlation between stand density and the amount of accumulated nutrients.Stands with lower densities should have less nutrients stored, which could be due to removing many trees that had nutrients contained within them [36,37].Another explanation for this potential result could be that a lower number of trees means less leaf fall [38].

Materials and Methods
This study was carried out in 5 sample stands of Scots pine (Pinus sylvestris L.), located in the Drawno Forest District, northwestern Poland (15 • 50 -16 • 0 E, 53 • 10 -53 • 13 N).This area is characterized by nutrient-poor habitats on podzolic soils, where the dominant tree species is Scots pine (Figure 2).Scots pine mostly forms uniform stands with a small admixture of other tree species, usually birch.The average annual rainfall in this area is 589 mm, the average temperature is 7 • C, and the growing season lasts 200-220 days.It was assumed that the stands have been established in a similar way (the same initial spacing) and similarly managed (the same owner and manager).Since the Second World War, the stands have been managed by the State Forests National Forest Holding.The sample plots, 0.5 ha each (Table 1), were selected so that tree density would be the distinguishing variable.Plots were established in evenly-aged, 82-year-old, single-species, single-layer stands growing on the same soil type, classified as Carbic Podzol, within the same forest complex and on flat terrain.The stands had not been thinned over the previous 5 years.
Forests 2018, 9, x FOR PEER REVIEW 3 of 12 consequence of a wide ecological amplitude, regarding both climatic conditions as well as soil requirements (from soils rich in humus to very poor).It is also an economically important species [30,34,35].To estimate how macronutrient stocks change, five sample stands of Scots pine (Pinus sylvestris L.) were compared.The stands were the same age and maturity, had different densities (476-836 trees per ha), and were 5-10 years past thinning.The aim of the study was to determine the influence of stand density on the macronutrient stocks (N, P, K, Ca, and Mg) in the aboveground biomass of mature Scots pine stands.
For Scots pine stands of a similar age, we predict that there will be a correlation between stand density and the amount of accumulated nutrients.Stands with lower densities should have less nutrients stored, which could be due to removing many trees that had nutrients contained within them [36,37].Another explanation for this potential result could be that a lower number of trees means less leaf fall [38].

Materials and Methods
This study was carried out in 5 sample stands of Scots pine (Pinus sylvestris L.), located in the Drawno Forest District, northwestern Poland (15°50′-16°0′ E, 53°10′-53°13′ N).This area is characterized by nutrient-poor habitats on podzolic soils, where the dominant tree species is Scots pine (Figure 2).Scots pine mostly forms uniform stands with a small admixture of other tree species, usually birch.The average annual rainfall in this area is 589 mm, the average temperature is 7 °C, and the growing season lasts 200-220 days.It was assumed that the stands have been established in a similar way (the same initial spacing) and similarly managed (the same owner and manager).Since the Second World War, the stands have been managed by the State Forests National Forest Holding.The sample plots, 0.5 ha each (Table 1), were selected so that tree density would be the distinguishing variable.Plots were established in evenly-aged, 82-year-old, single-species, single-layer stands growing on the same soil type, classified as Carbic Podzol, within the same forest complex and on flat terrain.The stands had not been thinned over the previous 5 years.On the sample plots, diameter at breast height (DBH) was measured for each tree on each plot, and the height of 20% of the trees was measured-i.e., every fifth tree.Based on this data for each plot separately, Näslund's height curves were developed to establish the height of each tree [39].
The dry mass of all the trees on the five experimental plots was calculated using allometric equations [40].On each 0.5-ha plot, 10 model trees were selected, with DBH representing the range of the measured diameters, resulting in 50 model trees in total.Model trees were felled in 2013 and divided into defined compartments: stem, dead branches, thick branches (diameter of more than 5 mm), thin branches, and needles.From the stem, samples were taken as 10-cm cross sections that were cut every meter along the stem.The fresh mass of each compartment was measured.After weighing the fresh compartments, the samples were dried at 65 • C to a constant mass to determine the dry mass.Using the sample proportion of dry and fresh mass, the dry mass of each compartment of each model tree was calculated.
Allometric equations were developed based on the data obtained for the 50 model trees across all plots.These allometric equations made it possible to calculate the dry mass of the aboveground component of every individual tree, using DBH and height as predictive variables.Mass was calculated for the following compartments: (1) stem, (2) dead branches, (3) thick branches, (4) thin branches, and (5) needles.The results were used to calculate the mass of each individual tree and the collective mass for the experimental plots.
Chemical analyses were conducted on a select number of trees that were used for biomass estimation.For each sample area, 3 out of the 10 model trees were selected.From the list, ordered by increasing DBH, the second, fifth, and ninth trees were chosen.Across the whole study area, a total of 15 trees were selected.Materials for the chemical analysis were taken beforehand from the samples used to determine dry mass.
Stem wood was divided into a marketable component and the top component (diameter below 7 cm).The two selected sample cross sections, the cross sections closest to the middle of each component, were taken for chemical analysis.The remaining cross sections were used to determine the proportion of wood and bark dry mass, which were previously not separated.Samples from other tree components, dead branches, thick branches, thin branches, and needles, were also taken for chemical analysis.The material was ground, marked, and sent to the laboratory.Laboratory samples were mineralized in HNO 3 and by using an inductively coupled plasma optical emission spectrometer (Optima 7300 DV ICP-OES) by Perkin Elmer Inc. (Waltham, MA, USA), and the contents of the following elements were determined: P, K, Ca, and Mg.Without mineralization, the content of N and S was determined by using a LECO TruMac CNS Macro Analyzer (Leco Corporation (Saint Joseph, MI, USA)).
To create a model for each tree in the test areas, one of 3 sample trees with the closest DBH was assigned.Based on the dry mass and percentage content of the nutrients in the assigned sample tree, the nutrient masses in each tree compartment within the test area were calculated.
The next step was determining the nutrient content of each tree component for all experimental plots.For each tree, the nutrient content was estimated based on the dry mass of each given component and the content of 7 macronutrients in the sample from the specific compartment.To calculate the element content for the remaining trees (other than model trees), specific coefficients were used based on proportions found in model trees that had the closest DBH to a certain tree.In this way, the basis for Forests 2018, 9, 593 5 of 11 calculations in smaller trees was the smallest model tree, and for the bigger trees, was the biggest model tree.The mass of any element contained in a tree compartment was calculated by multiplying the dry mass of this compartment by a specific coefficient (ratio) found in a model tree with the closest DBH.
Statistical analyses were performed using the "Multivariate Platform" tool in JMP 10.0 statistical software (SAS Institute Inc., Cary, NC, USA).Obtained as results were Pearson correlation coefficients (r) and the corresponding levels of significance.On this basis, an assessment of the correlations between the stand density and (1) aboveground biomass, (2) stand volume, (3) basal area, (4) mean DBH, (5) mean height, and (6) nutrient stock was performed.Additionally, the correlations between DBH of 15 model trees and concentrations of macronutrients in different tree compartments (stem wood, stem bark, branches, and needles) were calculated.

Results
The biomass of various tree compartments (stem wood, stem bark, thick branches, thin branches, dead branches, and needles) from five sample stands of Scots pine of different densities is shown in Table 2.The concentrations of six macronutrients in the dry mass of various tree compartments (estimated basing on 15 model trees) is shown in Table 3.Four macronutrients (N, P, K, and S) had the highest concentrations in the needles, while two macronutrients (Mg and Ca) had the largest concentrations in the stem bark.Among the compartments with the lowest concentrations of elements were those found in stem wood, except for Mg and K, which had the lowest concentration in dead branches (Table 3).Four macronutrients (N, P, K, and S) had a negative correlation between DBH and concentration for branches and bark, and one (Ca) for stem wood and bark (Figure 3).
The amounts of the six analyzed macronutrients in the aboveground biomass of Scots pine stands of different densities is shown in Table 4.The most abundant macronutrient was N, with 238.08-291.77kg•ha −1 in all plots; the highest amount of N was in the stem wood, and the lowest amount was in the stem bark.The second most abundant macronutrient in terms of amount per hectare was Ca (185.69-237.16kg•ha −1 ), with the largest amount in the stem bark and stem wood and the lowest amount in the needles.The third most abundant macronutrient was K (77.33-108.50kg•ha −1 ), most of which was in the stem wood, with the smallest amount in the stem bark.Next, Mg (38.64-71.39kg•ha −1 ) was highest in the stem wood and lowest in the needles.The last two nutrient levels were P (22.33-33.04kg•ha −1 ) and S (15.80-22.21kg•ha −1 ).Nitrogen was the most abundant element in the stem wood (68.17-126.37 kg•ha −1 ), branches (54.03-75.38 kg•ha −1 ), and needles (54.78-60.84kg•ha −1 ).In the stem bark, the most abundant element was Ca (65.21-80.38 kg•ha −1 ).The amounts of the six analyzed macronutrients in the aboveground biomass of Scots pine stands of different densities is shown in Table 4.The most abundant macronutrient was N, with 238.08-291.77kg•ha −1 in all plots; the highest amount of N was in the stem wood, and the lowest amount was in the stem bark.The second most abundant macronutrient in terms of amount per hectare was Ca (185.69-237.16kg•ha −1 ), with the largest amount in the stem bark and stem wood and the lowest amount in the needles.The third most abundant macronutrient was K (77.33-108.50kg•ha −1 ), most of which was in the stem wood, with the smallest amount in the stem bark.Next, Mg (38.64-71.39kg•ha −1 ) was highest in the stem wood and lowest in the needles.The last two nutrient levels were P (22.33-33.04kg•ha −1 ) and S (15.80-22.21kg•ha −1 ).Nitrogen was the most abundant element in the stem wood (68.17-126.37 kg•ha −1 ), branches (54.03-75.38 kg•ha −1 ), and needles (54.78-60.84kg•ha −1 ).In the stem bark, the most abundant element was Ca (65.21-80.38 kg•ha −1 ).The analysis of correlation showed a statistically significant relationship between stand density and macronutrient content for five cases (Table 5).The nutrient content in the total aboveground biomass was positively correlated with stand density for N (r = 0.970, p < 0.01), P (r = 0.909, p < 0.05), and K (r = 0.887, p < 0.05).Additionally, a positive correlation between stand density and P content was found in stem bark (r = 0.913, p < 0.05) and branches (r = 0.901, p < 0.05).Statistically significant, negative correlations were found between stand density and mean DBH (r = −0.850,p < 0.001) and mean tree height (r = −0.497,p < 0.05).No correlation was found between stand density and total aboveground biomass, biomass of tree compartments (stem wood, stem bark, branches, and needles), stand volume, or basal area.

Discussion
The obtained results confirm the widely known fact that nutrient concentrations vary greatly between tree compartments and are usually lowest in stem wood and highest in stem bark and tree crowns-mostly needles [36,41,42].The practical application of this information could be important: by employing various utilization methods, forest managers can significantly increase or decrease the amount of nutrients removed from forest ecosystems.
In mature stands (for Scots pine-over 30 years old), nutrient concentrations are rather constant and do not change with age [31].For the stands of the same age, nutrient concentrations are correlated with tree size [43].In spite of this, there is no evidence for a correlation between nutrient stock and thinning intensity [32].
By analyzing the relationship between stand density and the amount of stored nutrients, two situations should be distinguished: (1) when differences in densities are a consequence of different ages and (2) when stands of different densities are of similar ages.In the first case, changes in density result from stand age, when the number of trees decreases but the overall biomass increases, along with the amount of stored nutrients (Figure 1).In this case, the relationships are rather obvious-the amounts of stored nutrients should be positively correlated with biomass growth and stand age and negatively correlated with density.In our study, mature stands of a similar age but of different densities were compared.In this case, the obtained results might be distorted by recent thinning.Directly after thinning, the density is affected by thinning intensity (the amount of removed biomass).Stands with lower densities have lower biomass and a lower amount of stored nutrients.Consequently, the correlation between nutrient stocks and density should be positive, which is the opposite of the case of stands of different ages (Figure 1).To avoid such a distortion of results, the sample stands selected for our study had not been thinned in the last 5-10 years.
In our study, a positive correlation between stand density and the amount of stored nutrients was found for three out of the six elements (Table 5).Stand density was correlated with N, P, and K content in the aboveground biomass and additionally with P content in the stem bark and branches.There could be three explanations of correlation between stand density and nutrient stock: (1) different biomass of stands with different densities, (2) different proportions of tree compartments with different nutrient concentrations, and (3) difference in tree size in stands of different densities, or a combination of the above.
In our study, the main cause of such correlation was the variability in tree size that had different nutrient concentrations.The lower the number of trees, the bigger their DBH, and at the same time, larger trees have lower concentrations of some macronutrients.Concentrations of N, P, and K in branches and needles were negatively correlated with DBH (Figure 3), so the stock of these nutrients was lower in stands of lower densities, where trees were bigger.It could be expected that the same is true for Ca and S, the concentrations of which were also negatively correlated with DBH in some compartments.Their correlation was not statistically significant, which could be caused by the low number of model trees used for estimating concentrations (only three for each stand).
A strong negative correlation between the mean DBH of the trees and the stand density is commonly known and has been reported by multiple authors [34,[44][45][46][47][48][49].
Differences in stand biomass could not be the reason for correlation between stand density and nutrient stock, because the biomass of any compartment was not correlated with stand density.The differences in proportions of tree compartments' biomass in different stands also could not explain this correlation.When using stand-average concentrations of nutrients for calculations (without the division into trees of various size), resulting stocks were not correlated with stand density.
Research on the relationship between stand density and the accumulation of nutrients is very limited [36,50].In Korea, mature stands of Japanese red pine Pinus densiflora with a high density (842 trees•ha −1 ) and a low density (450 trees•ha −1 ) were compared, and N storage was nearly the same for both stands [37].In Canada, an estimation of macronutrient (N, P, K, Ca, and Mg) removal from the biomass of five commercially utilized tree species in stands with three different densities was carried out; density was not found to influence nutrient export at various harvesting intensities [51].The results of this research differed from our research, because the N, P, and K content showed a correlation with stand density.On the other hand, a pattern similar to the one in our study was found in plantations of loblolly pine Pinus taeda and slash pine Pinus elliottii in Georgia (USA).The amount of accumulated N, P, and K was significantly higher in denser stands, although the stands were young (4 years old) and the variations in density were due to different initial spacing [36].
The amount of nutrients in the aboveground biomass of the Scots pine stands are in the following order: N > Ca > K > Mg > P > S. Similar results were also obtained by other researchers [13,16,24,52].The ambiguity in the obtained results indicates the need for further, deeper investigations on other stands that account for various geographical locations and tree species.

Conclusions
The concentration of macronutrients in tree biomass vary greatly between tree compartments and also between trees of different DBH.Tree size is connected with stand density, which could be regulated by initial spacing and by thinning intensity.
Accumulations of N, P, and K in mature, same-age stands of Scots pine were higher in stands with higher densities.This discovery could be useful in practical forest management.More intense thinning causes a higher utilization of biomass, likely a better income from the final felling of a less dense stand [53], and a lower export of nutrients, due to the lower concentration in biomass in less dense stands.

Figure 1 .
Figure 1.Long-term and short-term changes in aboveground biomass, stand density, and aboveground nutrient stocks in forests.

Figure 1 .
Figure 1.Long-term and short-term changes in aboveground biomass, stand density, and aboveground nutrient stocks in forests.

Figure 2 .
Figure 2. Distribution of sample plots (SP) in the southeastern part of the Drawno Forest District.

Figure 2 .
Figure 2. Distribution of sample plots (SP) in the southeastern part of the Drawno Forest District.

Figure 3 .
Figure 3. Average concentrations of macronutrients in tree compartments (SW-stem wood, SBstem bark, BR-branches, FL-foliage) in model trees (N = 15) divided into small, medium, and big trees (according to DBH).Oblique lines show in which compartments the concentration of macronutrients was significantly correlated (p < 0.05) with tree size.All correlations were negative.

Figure 3 .
Figure 3. Average concentrations of macronutrients in tree compartments (SW-stem wood, SB-stem bark, BR-branches, FL-foliage) in model trees (N = 15) divided into small, medium, and big trees (according to DBH).Oblique lines show in which compartments the concentration of macronutrients was significantly correlated (p < 0.05) with tree size.All correlations were negative.

Table 4 .
Macronutrient stocks (kg•ha −1 ) in stem wood, stem bark, branches, needles, and total (all aboveground biomass) for Scots pine stands of different densities.

Table 1 .
Main characteristics of five 82-year-old Scots pine sample stands.

Table 1 .
Main characteristics of five 82-year-old Scots pine sample stands.

Table 2 .
Aboveground biomass (Mg•ha −1 ) of different compartments of five sampled Scots pine stands.

Table 5 .
Pearson's correlation coefficients (r) between stand density and nutrient stock and between stand density and nutrient concentration in different tree parts of the sampled Scots pine stands.