Master by Research in Biology
Master’s thesis
Sarah Gore
Regeneration in the Rocky Pine Forest in the High Coast area of
North East Sweden.
1
MID SWEDEN UNIVERSITY
Department of Natural Science
Supervisor: Bengt-Gunnar Jonsson, bengt-gunnar.jonsson@miun.se Author: Sarah Louise Gore, sago1800@student.miun.se
Degree programme: Master by Research, 120 credits Main field of study: Biology
Semester, year: Spring, 2018
2
Regeneration in the Rocky Pine Forest in the High Coast area of North East Sweden .
Master’s Thesis Sarah Gore
April 2020
3
Tree regeneration in the rocky Pine forest of the High Coast Area in North East Sweden
Sarah Gore
Mid Sweden University, Department of natural Science
Abstract
Rocky Pine Forest (RPF) regeneration in the High Coast area of Sweden is an area of low productive and high naturalness forest. RPF’s main tree species is
Scots Pine trees (Pinus Sylvestris) and Norway Spruce (Picea abies). Natural forests in Sweden and around the world are becoming a rare phenomenon and
it is important to understand natural forest dynamics in order to maintain high production value and biodiversity value in forest industry. This study used 15 transects over 5 sites to look at forest dynamics in this area focusing on trees
≤1.3m or ≤ 10cm Diameter at Base Height (DBH) following a previous study on the same transects looking at trees ≥1.3m and or >10cm DBH. All trees along the 100m transect and 5m either side were mapped and samples were taken
from a selection of Pine and Spruce trees to assess growth rate. Moose browsing severity at each Pine tree was taken. Dendrochronological techniques were implemented to assess tree age and used to calculate growth rate. Out of Pine and Spruce, Pine tree numbers were much higher than Spruce
with 72% Pine and 28% Spruce and no Spruce trees under the age of 11 were recorded. However in under 50 years age classes Pine seedlings accounted for
55% of the seedling and Spruce 7%. In over 50 years of age however Spruce had a slightly larger share of 21% and Pine had 17%. Likelihood of Sprucification is small but Spruce trees do well in older ages classes so it cannot be ruled out. Browsed trees had a higher growth rate than trees with
no browsing. Most likely due to the desirability of more productive and vigorous trees to Moose in the first place. Natural unproductive forests such as
RPF can tell us a lot about forest regeneration and how to manage and maintain forests in Sweden and how forest dynamics play out in a changing
world.
4
Table of Contents
Abstract ... 3
Introduction ... 5
Aims ... 8
Methods ... 8
Results ... 13
All trees ... 18
Discussion... 23
Conclusion ... 26
References ... 28
5
Introduction
Rocky Pine forests (RPFs) make up 2.8% of forested land in Sweden (National Forest inventory). In höga kusten (The High Coast) in north east coast of Sweden the rocky Pine forests have old growth characteristics, low productivity, undisturbed and unmanaged environments. Characterised by steep terrain, rocky and sun exposed outcrops. Tree species here are largely Pinus Sylvestris or Scots Pine and to a lesser extent Picea abies or Norway Spruce (from now on referred to as Pine and Spruce respectively). The area also has small scale fire history and of low intensity, with average frequency being between 20.5 years and 65 years (Sandström et al., 2020). This whole area is a UNESCO world heritage site as it is the site of the most rapid isostatic change worldwide (Berglund, 2012).
Since the rocky Pine forests of the high coast area are slow growth and low production forests, they are not used in industry. Low production forests are defined as those that produce less than 1m³/ha annually. Currently, they comprise 18% of Swedish forested area. These forests are therefore unmanageable by Swedish law (Swedish forest agency 2014). Yet these low productive forests have received little research attention. Studying rocky Pine forests could help us assess their value in conservation management, especially by comparing them with high productivity forests (Hämäläinen et al., 2018, Hämäläinen et al., 2020)
In Sweden 72% of forests protected from management are low production forests (Swedish forest agency 2014). Low productivity forests are in general thought to have little value for endangered or declining species (Cederberg 1997). Conversely since low productive forests are generally spared from management these forests are home to long lived trees (Storaunet et al., 2005). Certain species adapted to these long lived, sun exposed forest environments will flourish in low production forests compared to high productive forest stands (Cederberg 1997). The Rocky Pine forests of the high coast area is one of two sites in Sweden with the red listed beetle Chalcophora mariana. The species prefers dead trees with sun exposed wood, allowing them to thrive in this sparse and exposed area (Artadabanken 2015). In 2019 Hämäläinen et al found that the lichen species richness was highest in low productivity forest, on thin rocky soil such as in the rocky Pine forest, compared to other low production stands in mires and production forests sampled.
In the RPF of the high coast, some trees have reached 300 years and over. According to forests statistics in conducted by SLU 12.2% of the whole of Sweden’s forested area contains forests over the age of 141 years old. Old growth characteristics or naturalness are characteristics defining forests not disturbed or managed by humans and are generally deemed to be forests that have a wide variety of tree age and size and contain examples of “old” trees. In addition natural
disturbances must be in action (Kuuluvainen, 2002b). These forests are important carbon sinks, accumulating and storing carbon over time (Luyssaert et al., 2008). Having dead or dying trees of a varied age, an old growth forest can contribute to a more heterogeneous environment with many different habitat types, leading to forests with increased biodiversity for a wide range of species (Andersson and Östlund, 2004, Bond and Franklin, 2002, Linder and Östlund, 1998).
In Sweden 80% of forests are classified as productive forests and therefore it is unusual for these forests to reach old growth status due to their economic value. Normally managed forests are even aged homogeneous stands with little species richness and biodiversity (Kuuluvainen, 2002a, Linder and Östlund, 1998, Kuuluvainen, 2009, Axelsson and Östlund, 2001).
Old growth characteristic, undisturbed, low productivity forests such as the RPF studied can give us an idea of how forest dynamics function without human interference. A so called “natural” forest could aid forest management especially in the light of land use change, biodiversity loss, and climate change. (Kellomaki et al., 2001, Hämäläinen et al., 2018).
6 It was first hypothesised that a natural forest would reach a stable “climax community” when no disturbance or human interaction had interrupted the community development. However, modern understanding of forests dynamics is that disturbance and tree mortality are a natural part of forest and ecosystem dynamics allowing for increased regeneration and successional variation (Bengtsson et al., 2000, Kuuluvainen, 2009, Angelstam, 2004). Most current forest species would therefore have adapted to disturbances and tree mortality. As previously stated, most of Sweden’s forests are managed in some way and so the natural disturbances of a forest have been changed to meet the needs of the mangers of the forest. A good example of this is fire suppression in forest industry (Kuuluvainen et al., 2017).
Fires can be caused naturally by lightning events and through the Anthropocene by humans (Granström, 1993). There is evidence of fire prevention from humans as early as 1800 but after World War 2, in the 1950s the increasing value of timber stands led to the implementation of increased fire prevention methods and a reduction in natural and anthropogenic fire disturbances (Ostlund and Roturier, 2011, Niklasson and Granstrom, 2000, Östlund et al., 1997). In addition, more efficient and large-scale forestry operations in Sweden since the 1970s have caused more
homogeneous and fragmented forest populations with less dead wood. Together, these factors have created fewer fire opportunities across all forest types, including RPF (Sandström, 2018), and changed species composition as a result (Halme et al., 2013) (Nowacki and Abrams, 2008). For example, some species such as aspen can only reproduce sexually when facilitated by fire and the beetle Melanophila acuminate relies upon fire as the larvae can only feed on tree bark that has been burnt (Schmitz and Bleckmann, 1998, Evans, 1966).
There is also evidence that Sprucification has occurred in normally Pine dominated forests.
Sprucification being the increase dominance of Spruce trees in formally Pine dominated forests. As fire frequency declines, there are fewer gaps in the forest canopy preventing seedling survival due to lack of light. Spruce seedlings are generally believed to be shade tolerant giving them an advantage over Pine and seedlings in stands without fire (Linder, 1998, Engelmark et al., 1998, Linder and Östlund, 1998, Esseen, Esseen, 1997)
Increasing numbers of Spruce trees in stands from fire suppression and forest industry aggravate wind disturbance. Replacing stands with fast growing even aged trees of the same species leaves them vulnerable to storms and heavy winds. In a natural forest, there would be trees of all age types and varied species. Deciduous trees and old growth trees are more resilient to wind disturbance and so strengthen the whole stand (Schelhaas et al., 2003). In a natural forest, wind disturbance can create gaps in the forest canopy which encourages seedling growth and vegetation variation through successional change (Thonicke et al., 2008).
Storm damage can make a forest susceptible to increased tree mortality in the form of insect outbreak as insects are attracted to the burnt or dying wood and more dead wood can increase ignition material. (Thonicke et al., 2008, Schmitz and Bleckmann, 1998).
Other animal disturbances come in the form of ungulate species. Preferred plant species of these herbivores can be negatively affected by the presence of herbivore species. In a study conducted by (Ellis and Leroux, 2017), it was found that preferred plant species in areas with no moose were 1.4 times taller than the same species in areas where moose were present. Pine is generally a preferred species by moose, so Pine seedlings and saplings regeneration could be limited compared to Spruce for example (Milligan and Koricheva, 2013, Zamora et al., 2001). This suggests that herbivores can shape plant communities spatially and temporally, thereby affecting regeneration of palatable species (Pastor et al., 1993) (Hester et al., 2006, Edenius et al., 2002, Felton et al., 2018)
It has even been shown that top predators indirectly affect regeneration of trees by reducing the number of herbivores such as moose, suggesting that tree regeneration in some cases could be increased when top predator numbers are high. This effect is diminished when combined with other disturbances such as fire and insect infestations (McLaren and Peterson, 1994).
7 Normal levels of browsing however have been shown to benefit the ecosystem by stabilising
succession and creating and maintaining clearings, creating new niches and allowing more species to arise within the whole ecosystem and so increasing biodiversity. Nutrient cycling within the
ecosystem has also been observed to benefit from increase in biodiversity (Stewart, 2001).
Disturbances, both large and small, create gaps in the forest providing regeneration opportunities for seedlings. In Pine forests, large trees appear to strongly influence seedling regeneration. These forests normally contain a large seedbank in the understory. If these seedlings are in the vicinity of a large tree this can negatively affect regeneration. When these trees fall, the suppressed seedlings have an almost instant increase in resource availability. Microsites produced from these fallen trees, soil troughs and peeks, aid regeneration and establishment of seedlings. This is both from the above (more light and water availability) and from the ground(root nutrients and water uptake, less competition from ground vegetation removed during the fall) (Kuuluvainen, 1994).
Microsite conditions such as temperature, sun exposure, areas covered with snow, weed cover, smooth and slightly elevated slope area have all been shown to impact on seedling growth (Colak, 2003, Tegelmark, 1998)
Ground vegetation can positively influence establishment of seedlings by favouring Pine germination and establishment over Spruce, influencing nutrient availability in soils, reducing soil temperature, increasing humidity and shaping fire severity or even fire occurrence (Zackrisson et al., 1995, Nilsson and Wardle, 2005).
The spatial position of a tree could also influence its regeneration patterns. Aggregation is a normal feature in forest ecosystems and beyond an in turn the interaction between any neighbouring tree can be good and bad. Canopy cover is light limiting and neighbour trees potentially compete for nutrients and reduce growth rate. On the other hand transfer of nutrients via the mycelial network of fungal symbiosis could increase uptake of nutrients and so increase regeneration (Dickie et al., 2002, Van Der Heijden et al., 2015).
Understanding forest dynamics and interactions could provide valuable management techniques to improve biodiversity and forest production for the future, especially in the light of anthropogenic climate change and land use change threatening forest ecosystems (Angelstam et al., 2017, Angelstam, 2004).
Regeneration is used widely in forest and tree research however there are many different types such as seedling regeneration, stem regeneration or vegetative regeneration and some which do not state a specific regeneration type. There are also growth rates calculated using change in height, biomass and branching. This can make different research hard to compare and so in this study regeneration describes the radial growth rate of the tree calculated by working out the tree age and dividing this by the diameter of the tree to give growth rate in mm per year (mm/yr)(Speer, 2010a) This work is an extension of the work by Jennie Sandström (Sandström et al., 2020) who first worked on these sites while investigating forest dynamics in the rocky Pine forest in north east Sweden. In her study she investigated the history and land use of the RPF and looked at fire frequency across all the sites. However, Jennie’s study only looked at trees above 1.3 meters tall and 10cm diameter at breast height and so excluding those trees smaller than this. In this investigation, we looked at all trees that were not sampled in the previous study. All trees below 1.3 meters and less than 10cm in diameter at breast height. In doing so the aim is to acquire a full knowledge of the forest including all the trees.
8
Aims
Study aims are to investigate the regeneration dynamics in the rocky Pine forest of the high coast area of northern Sweden. More specifically looking at:
• Has tree species composition changed over time?
• Does the browsing of Moose effect tree growth rate?
• Is growth rate different between age classes
• Does aggregation affect growth rate.
Methods
Study Area
The study was conducted in the Västernorrland County high coast area in north eastern Sweden. So called due to the coastline having the most rapid isostatic uplift in the world, currently at an 8mm rise per year. A result of the ice melt after the last ice age since this time the coastline in this area has risen by 286m creating sheer cliffs and rocky terrain. At the southern boreal zone, the forest here is primarily made up of Scots Pine (Pinus Sylvestris) and Norway Spruce (Picea abies) with a lesser proportion of deciduous trees. The soil is generally thin and nutrient poor with many exposed rocky areas.
In 2011 the county administration in Västernorrland conducted a general inventory of 26 rocky Pine forest sites to assess the conservation value of the areas. From these sites 8 of highest conservation value were selected for Jennie Sandströms study of the Rocky Pine forest looking at the structure, dynamics and history of this area. (Sandström et al., 2020)
Study Species Pine
Pine is a pioneer species and so has rapid early growth but then high seedling mortality (Engelmark et al., 1998) (Johansson et al., 2014). Pine competes best at coarse sites with low moisture and nutrients (such as RPF where growth is slow, but trees can get to advanced ages) and in gaps and after disturbances such as fire or wind blows. Pine can out compete Spruce in these areas. Generally forming multi aged stands mostly due to fire effects (Engelmark et al., 1994). Favoured for browsing by moose in winter months giving more opportunity to Spruce to increase growth (Engelmark, 1999).
Spruce
Spruce are late successional species and therefore grow more slowly as seedlings than Pine and deciduous trees. Spruce then can survive well in shade of these trees and lay relatively dormant until conditions change to favour Spruce. Such as high moisture and nutrient conditions. Spruce can survive for very long periods of time in this state until conditions become favourable(Engelmark et al., 1998) (Engelmark, 1999). Spruce trees can be extremely long lived as they can clone themselves and regenerate by layering. These clones can then produce upward growing trunks. A famous example being a Spruce tree named “old Tjikko“ in Dalana Sweden. That has been dated as 9550 cal yr Bp clone with an upward stem which is thought to have started to emerge in the 1930s when the climate began to warm (Oberg and Kullman, 2011)
Pine and Spruce can form mixed stands and stands on their own.
9 Field Sampling
In Jennie Sandströms study in 2014 all trees above 1.3m and 10cm diameter at breast height (DBH) were studied. Factors such as the sites fire history and dead wood content were investigated. Each of the 8 sites were divided into 3 randomly selected and placed transects using a numbers grid over the site map. The starting point and orientation (N, NE, E, SE, S, SW, W, NW) were then randomly selected. Each transect was at least 100m from the other. From these points a 100m by 10m transect was created and the end of the transect coordinates were recorded (Sandström et al., 2020). When using jennies data 4 years were added to look at the sites in real time.
From these 8 sites, 5 were used for this study due to availability and time constraints. Permission from landowners was obtained before any fieldwork was carried out. In September 2018 the transects were located using start and end coordinates, the direction information, a GPS, a compass and a tape measure to lay along the whole transect.
Once the exact transects were located this study focused on recording the location of all trees in the transect area and recording the species, size and browsing information then collecting trees core or disk samples from selected trees. We split up this method to every 10m along the transect. So as to select the correct number of samples. We travelled up
the plus side of the transect and then the negative side recoding tree information as we went. Then went back and collected the sample information (samples were selected in the walk stage and a paper bag or core straw was labelled and attached to the tree to locate it more easily afterwards.
It was made sure to only sample trees not sampled in previous study. That is trees under 1.3m or 10cm DBH, these trees were sampled. Tree cores were the favoured method of sampling however some trees were just too small to sample in this way. Personal judgement was used to assess if tree could be cored or not. Tree core/tree disc samples were taken where possible in 3 categories, each for Spruce and Pine trees: A) Trees below 1.3m, B) Trees above 1.3m with DBH 0-5 cm, C) Trees above 1.3m with DBH 6-10 cm. No more than 3 Pine and 3 Spruce samples of each A, B or C were taken if possible, at every 10m of the transect. Potentially 180 samples per transect. Although Spruce numbers likely not to be this many. This was to spread the sampling along the transect line to achieve a more accurate portrayal of the whole transect. If there were more than 3 potential
samples found every 10m along the transect then the first 3 were selected to avoid any bias.
Pine samples were taken at a rate of 3 cores/samples in each category every 10m along the 100m transect, to try to spread the sampling along the whole transect. As Spruce tree numbers were limited compared to Pine, as many samples as could be found in each category every 10m along the 100m transect were taken with no limit on number.
Samples were taken using a 6mm increment borer at the base of each tree, around 30cm in height. If the tree diameter at this height was too small to be cored, then the tree would be sawn down using
Figure 1. Diagram of transect layout used in the High Coast RPF in all sites. 100m transect with 5m each side.
10 and a disc sample would be taken at lease a centimetre in height so that sanding could be
completed.
Within the transect, every living tree was recorded, its location x) along the transect and y) distance at a right angle from the x coordinate. Up to 5m from the x transect line either side was recorded, equalling a 100m by 10m rectangle. Each tree was also recorded for its species, height below 1.3, base diameter, DBH in taller trees and browsing on a scale from 0 (no browsing), 1 (Little browsing), 2 (browsing), 3 (much browsing). Any other notes were also recorded. Cores were labelled and placed in paper straws and discs were put in paper bags immediately and stored in a cool dry place until processing.
Dendrochronological analysis
The tree ring analysis followed standard procedures as outlined in (Speer, 2010b). All samples were glued, mounted and labelled on wooden core mounts for cores and plywood for discs. While drying the samples were held with tape. When the glue was dry, after at least 24 hours, the samples were sanded down using first large then small grain size sandpaper, to give clarity to the tree rings. When ready each sample was scanned with a high-resolution scanner and each scan picture was saved and recorded. The high-resolution pictures were then used to count the rings of each tree by eye.
Samples too small for mounting, around 10mm or less, had their age counted with a hand lens. All samples had the diameter measured again in millimetres.
Statistical Analysis
Growth rate was calculated by dividing age by diameter giving growth rate in yrs/mm. Growth rate of each sample was separated into browsing categories 0 to 3 and this data was then imported to R where an ANOVA test was run and then a Tukey’s test to pinpoint exact difference in mean between the categories.
The data was also separated into browsed and un-browsed and a welch two sampled T test was run to test the relationship between means. However, the number of trees given the highest level of browsing (3) was only n=12 individuals and (2) n=43 (1) n=142 and no browsing (N) n=625.
Therefore, the total browsing average was compared to the total average non browsed individuals.
A two way ANOVA was run in R to compare browsing and non-browsing in each transect to look more closely at the data
Cluster Analysis
In this part of the analysis I used both my data and Jennies data.
The distance of each tree to its nearest neighbour along the x axis was calculated and then a
variance to mean ratio (VMR) was calculated (Horne & Schneider 1995) to assess aggregation. Using the binomial and Poisson distributions to then assess aggregation. Values equal to 0 have no
aggregation, binomial distribution between 0 < VMR > 1 is under dispersed, VMR equal to 1 show Poisson distributions and VMR values above one have a negative binomial distribution and are over dispersed which is associated with data sets that show clustering or aggregation in the data.
The distance between each point on the x axis was calculated and then a variance of mean (VMR) was calculate for each site showing clustering. This method does not include aggregation in a 2 or 3- dimensional sense and therefore is missing a higher level of description. But still give us an idea of aggregation.
11 Average nearest neighbour distance was calculated for each transect and then compared to average growth rate at each site. These values were then compared in a Spearman’s rank correlation in R to try to ascertain is these values correlate.
Regression analysis of Pine and Spruce trees from both studies
Tree age and growth rate in mm/yr and tree age and sample Diameter from Jennie Sandström’s data (Sandström et al., 2020). The data from Jennies study was modified to bring the data into the same time as the current sstudy. This was done by adding 4 years to the age data (Jennies data was collected in 2014 and this study data in 2018) and growth rate was used to increase diameter measurements in Jennies data by 4 years. This data was then compared in a Linear regression analysis in R. This studies data was also looked at on its own. Any data analysis completed will state if the data is solely from this studies data set or combine with data from Jennie Sandströms study (Sandström et al., 2020)
Assumptions
When comparing the species before gathering data I have assumptions that the trees should be different: 1)Expect that the species shall have differnet regeneration in realtion to their successional stage. 2)There will be in general low moisture and nutrients in soil and corse rocky substrait,
however the will be varieability and so exceptions to this. 3)I expect Pine to have higher numbers than Spruce.
12
Figure 2- Picture a) showing rocky Pine forest at Gårdberget taken looking towards Fanön. Sparse, rocky and steep terrain of the RPF. b) Map of Sweden and surrounding countries with a square around the location of the High Coast area. c) High Coast area map showing locations of each of the sites GR=Gropberget, PR=Porsmyrberget,
VR=Vårdkallberget, FA=Fanön, GÅ=Gårdberget.
a)
b) c)
13
Results
Stand structure
Table 1: Stand structure data for Scots Pine, Norway Spruce and deciduous trees ≤1.3m or ≤ 10cm DBH in 5 sites from the High Coast area of Sweden. GR=Gropberget, PR=Porsmyrberget, VR=Vårdkallberget, FA=Fanön, GÅ=Gårdberget. VMR- refers to Variation (standard deviation) to Mean Ratio of nearest neighbour distance values compared over the 5 sites.
Values that are “clustered” are *. Standard error in brackets.
GR VR PR GÅ FA Total
Pine Share (%) 53 71.5 77 58 66 66
Spruce Share (%) 25 21 8 20.5 22 16
Deciduous Share (%) 22 7.5 15 21.5 12 18
Trees ≤1.3 (%) 59.8 64.8 78.5 74.3 77.9 71.8
Trees >1.3 (%) 40.2 35.2 21.5 25.7 22.1 28.2
Average Base Diameter (mm) All Species
36 (2.0)
36 (2.1)
36 (2.2)
33 (2.1)
28 (2.5)
35 (1.0)
Max BD (mm) All species 160 104 129 140 105 160
Average Pine Age (yrs)
35 (2.6)
42 (3.2)
34 (2.4)
31 (2.9)
27 (3.0)
34 (1.3)
Max Pine Age (yrs) 143 145 164 199 140 199
Average Spruce Age (yrs)
71 (5.3)
78 (4.7)
90 (8.3)
65 (4.3)
64 (5.0)
76 (2.5)
Max Spruce Age (yrs) 158 163 224 177 105 224
Pine growth rate (mm/yr)
0.944 (0.05)
0.876 (0.05)
0.913 (0.04)
0.875 (0.05)
0.930 (0.06)
0.907 (0.02)
Spruce growth rate (mm/yr)
0.572 (0.04)
0.484 (0.03)
0.544 (0.03)
0.643 (0.04)
0.621 (0.09)
0.573 (0.02)
Average Growth rate (mm/yr)
0.83 (0.04)
0.75 (0.04)
0.84 (0.04)
0.79 (0.04)
0.88 (0.06)
0.81 (0.02)
VMR 1.95* 2.03* 0.97 1.26* 3.45* 1.97*
Over the rocky Pine forest studied Pine trees are most numerous with average share over all sites at 66%
and deciduous trees 18% and Spruce trees at 16% share. Spruce trees were on average 42 years older than Pine trees and had an older maximum tree age of 224 compared with Pine maximum age of 199.
Tree base diameter varied with average BD of 28mm compared to a maximum of 140mm. Variance to mean ratio showed that all sites but PR were aggregated.
14 Young Pine tree frequency is very high and decreases with increasing age. Young Spruce trees have a lower frequency in general and bell shaped age class pattern. Peaking in the 56-60 years age class and decreasing steadily from there. There are no Spruce trees under 11 years old. There are more Spruce in older age categories compared with Pine.
Frequency of smaller base diameters in Pine trees is very high in low diameter classes and then seems to level off in the higher diameters classes whereas Spruce frequency starts off low then rises and stays fairly steady until dropping off at 100mm. Spruce size classes appear to stay steadier in number then Pine. Pine have many smaller base diameter trees than Spruce trees.
0 10 20 30 40 50 60 70 80 90
Freuency
Age (yrs)
Pine Spruce
0 20 40 60 80 100 120 140
Frequency
Base diameter classes (mm)
Pine Spruce
Figure 3- Frequency of age classes (yrs) of Scots Pine (dark grey) and Norway Spruce (light grey) trees ≤1.3m or ≤ 10cm DBH at 5 RPF sites in the High Coast area of Sweden.
Figure 4- Frequency of base diameter (BD) classes of Scots Pine (dark grey) and Norway Spruce (light grey) trees ≤1.3m or ≤ 10cm DBH at 5 RPF sites in the High Coast area of Sweden.
15 Growth rate
Pine growth rate is higher than that of Spruce but the growth rate as shown in a paired T-test with a P-value of 3.726e-16. The difference between trees that are ≤1m and >1m from their nearest neighbour is not significantly different in a paired T-test with a p-value = 0.3421. However as nearest neighbour is calculated from a one-dimensional horizontal axis (on the x axis) it is possible that we do not have enough data on which points are aggregated along both axes. From table 2 we can see clearly the number of trees in each category and the variation from in the data. This patten follows that of the age frequency graphs.
Figure 5. a) Growth rate (yr/mm) of both Spruce and Pine trees ≤1.3m or ≤ 10cm DBH across all 5 sites compared.b) Growth rate between trees ≤1.3m or ≤ 10cm DBH who have neighbours <1 meter from them compare to trees with neighbours >1 meter away.
0 0.2 0.4 0.6 0.8 1 1.2
<50 51-100 101-150 150<
Growth rate (mm/yr)
Age (yr)
Spruce Pine
Figure 5.a) Scots Pine (dark grey) and Norway Spruce (light grey) average growth rates (mm/yr) through age classes (yrs).
b) Average growth rate difference between tree heights <1.3 meters and >1.4meters. On trees ≤1.3m or ≤ 10cm DBH from 5 RPF sites in the High Coast area of Sweden. Error bars show Standard error.
a) b)
a) b)
Figure 6. a) Average growth rate (mm/yr) of Scots Pine (dark grey) and Norway Spruce (light grey) trees ≤1.3m or ≤ 10cm DBH across all sites compared.b) Average growth rate between trees ≤1.3m or ≤ 10cm DBH who have neighbours <1 meter from them compare to trees with neighbours >1 meter away. All trees are from 5 sites in the High Coast area of sweden.
a) b)
16
Table 2. Average growth rate (mm/yr) of Scots Pine and Norway Spruce in different ages classes with n= number of trees and SE
= Standard error SD = Standard deviation. Scots Pine and Norway Spruce trees used are ≤1.3m or ≤ 10cm DBH and from high coast area of Sweden.
<50 51-100 101-150 150<
Spruce 0.52 0.62 0.56 0.43
N 57 116 45 9
SD 0.33 0.30 0.18 0.08
SE 0.04 0.03 0.03 0.03
Pine 0.93 0.88 0.64 0.54
n 446 115 22 2
SD 0.59 0.32 0.02 0.02
SE 0.03 0.03 0.02 0.02
17 Browsing
There is a significant increase in radial growth rate in browsed trees compared to non-browsed trees. ANOVA results showed a significant difference in the means of all 4 browsed categories. An F value of 18.42 and a P value of 1.43e-11. Showing a significant result. To describe the data further the Tukey’s test was run and showed this significance in mean was between the non-browsed category and the rest of the browsing category. Browsed trees have a significantly higher growth rate to non-browsed trees and this difference cannot be due to chance.
As there is quite a difference in the sample numbers between non- browsed trees and the categories of browsed trees (see Figure 7a) and b)) which could affect the validity of results. In response the browsed and non-browsed have been compared in a box plot and in a Welch Two Sample t-test t = 5.941, df = 269.76, p-value = 0.000000008714. With the P value well below 0.05 the difference between browsed and non-browsed is significantly different from each other.
Browsing and none browsing had a P value lower than 0.05 so were significantly different from each other. The sites compared means were also significantly different but not as significant as Browsing conditions. The comparison of growth rates between Browsing conditions and sites gave a Pvalue of 0.002 showing a significant difference in these means.
Looking at the frequency of each browsing category it is clear that there is not so much browsing in total on Pine trees in this area.
Figure 7. a) showing only N= no browsing trees and B = browsed trees. b) Browsing levels and associated growth rates (mm/yr). O = no browsing (n=626), 1 = Little browsing (n=142), 2 = Browsing (n=43) and 3 = Much browsing (n=12). All trees are from Scots Pine trees ≤1.3m or ≤ 10cm DBH from hight coast area of Sweden.
Sum Sq DF Fvalue Pvalue
Browsing 10.027 1 44.1248 5.72e-11
Sites 6.697 14 2.1052 0.009954
Browsing: Site 7.740 14 2.4328 0.002360
Residuals 180.204 793
Table 3 ANOVA table showing comparison between browsing and growth rate over all 15 sites of from Scots Pine trees
≤1.3m or ≤ 10cm DBH from hight coast area of Sweden.
a) a) b)
Table 3 ANOVA table showing comparison between browsing and growth rate over all 15 sites of from Scots Pine trees
≤1.3m or ≤ 10cm DBH from hight coast area of Sweden.
18
All trees
Pine
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
0 50 100 150 200 250 300 350 400 450
Growth rate (mm/yr)
Age (yrs)
Pine L Pine S
0 100 200 300 400 500 600
0 50 100 150 200 250 300 350 400 450
Diameter (mm)
Age (yrs)
Pine L Pine S
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
0 100 200 300 400 500 600
Growth rate (mm/yr)
Diameter (mm)
Pine L Pine S
Figure 8. Graphs comparing Scots Pine trees ≤1.3m or ≤ 10cm DBH (grey) and trees ≥1.3m (black) from Jennies data (Sandström et al., 2020) collected in 2016 from hight coast area of Sweden. All Diameters are BD for trees ≤1.3m or ≤ 10cm DBH and DBH for trees ≥1.3m. a) Age (yrs) compare with growth rate (yrs/mm). b) Age (yrs) compared with diameter (mm). c) Growth rate (mm/yr) compared with diameter (mm).
a)
b)
c)
19 Spruce
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
0 50 100 150 200 250 300
Growth rate (mm/yr)
Age (yrs)
Spruce L Spruce S
0 50 100 150 200 250 300
0 50 100 150 200 250 300
Diameter (mm)
Age (yrs) Spruce L Spruce S
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
0 50 100 150 200 250 300
Growth rate (mm/yr)
Diameter (mm)
Spruce L Spruce S
a)
b)
c)
Figure 9. Graphs comparing Norway Spruce trees ≤1.3m or ≤ 10cm DBH (grey) and trees ≥1.3m (black) from Jennie(Sandström et al., 2020) data collected in 2016 from hight coast area of Sweden. All Diameters are BD for trees ≤1.3m or ≤ 10cm DBH and DBH for trees ≥1.3m. a) Age (yrs) compare with growth rate (yrs/mm). b) Age (yrs) compared with diameter (mm). c) Growth rate (mm/yr) compared with diameter (mm).
20 250 years. PineS trees range from 0- 150 years and PineL trees range from 76-375 years. Pine trees in both groups make a j shape curve and Spruce make a more bell shape age frequency.
In Jennies data (Sandström et al., 2020) trees ≥1.3m and 10cm DBH Pine also have a higher frequency than Spruce trees. Frequency of age classes is much like trees below 1.3m and 10cm diameter. High frequency of young Pine trees while the highest frequency in Spruce trees was between 51 and 73 years. Both frequencies taper off with increasing age.
There is a huge decrease in numbers of seedling to adult tree individuals most prominently seen in Pine trees but also in Spruce trees. Pine trees numbers fall by over 50 % from <25years compared with 26-50. While Spruce tree numbers increase in these categories by nearly 300%.
Table 4 and figures 7 and 8 show the relationship between Pine and Spruce with Age, Diameters and Radial growth rate between the two studies. There is significant relationship between all regressions apart from age compared to growth rate in Pine trees.
Figure 10 shows proportion as a percentage of Pine and Spruce in both from Norway Spruce and Scots Pine trees ≤1.3m or ≤ 10cm DBH (PineS and SpruceS) and trees ≥1.3m (PineL and SpruceL) from Jennies data. PineS trees have the highest percentage share in each site and SpruceL trees have the lowest share in all sites.
Figure 11 shows the frequency of PineS and PineL over age classes and SpruceS and SpruceL over age classes. SpruceS trees occur in a range from 0 -200 years and SpruceL trees occur in age classes 76-
Table 4.Regression analysis in R of all graphs in Figures 7 and 8. Pine and Spruce trees ≤1.3m or ≤ 10cm DBH (grey) and trees
≥1.3m (black) from Jennie(Sandström et al., 2020) data collected in 2014 from hight coast area of Sweden. All Diameters are BD for trees ≤1.3m or ≤ 10cm DBH and DBH for trees ≥1.3m.
Age vs Growth rate
Diameter vs
Growth rate Age vs Diameter Pine
Residual standard error 0.5833 44 40.9
Adjusted R squared 0.000736 0.7467 0.7467
F-statistic 0.2988 2819 2819
DF 953 955 955
P value 0.5848. < 2.2e-16 < 2.2e-16
Spruce
Residual standard error 0.3103 27.95 26.62
Adjusted R squared 0.02202 0.5923 0.4999
F-statistic 7.012 388.9 227.9
DF 266 266 226
P value 0.008582 < 2.2e-16 < 2.2e-16
21
12
21 24
8
18 63
58
53
75
59
1 2 6 1 1
24
19 18 16
22
FA GÅ GR PR VR
Percentage spceies composition
Spruce Pine SpruceL PineL
Figure 11. -Frequency of Scots Pine (a))and Norway Spruce (b)) in different age classes (yrs). Data is from trees ≤1.3m and ≤10cm BD (Pine S and Spruce S) H and trees ≥1.3m and ≥10cm DBH (PineL and SpruceL). From the RPF in the High Coast area of Sweden
Figure 10. Bar chart showing composition as a percentage at each site comparing Norway Spruce and Scots Pine trees
≤1.3m or ≤ 10cm DBH (Pine and Spruce) and trees ≥1.3m (PineL and SpruceL) from Jennies (Sandström et al., 2020) data collected in 2016 from hight coast area of Sweden.
0 50 100 150 200 250 300 350
Frequency
Age class (yrs)
PineS PineL
0 50 100 150 200 250 300 350
Frequency
Age class (yrs)
SpruceS SpruceL
b) a)
22 Sprucification
In general Pine tree frequency is higher than Spruce with 66% Pine share and 16% Spruce share.
Spruce share was also different between the two studies with trees above 1.3m and 10cm DBH having a much lower share than the trees below 1.3m and 10cm DBH as can be seen in figure 11 Pine recruitment is much larger than that of Spruce seedling and older Pine trees reach an increased maturity compare to Spruce trees. However, the oldest tree in the below 1.3m and or 10cm DBH study was a Spruce tree at 224 years old while the oldest Pine in that category was 199. In Jennies data the oldest tree was a Pine tree at 418 years old while the oldest Spruce was 259 years old.
Dead trees
Due to an extremely warm summer in 2018 some of the trees sampled were dead as a result of these high temperature, dying only that summer. I noted down these trees and they accounted for 352 individual dead trees out of all trees recorded which was 1834. Average Height of dead trees was 33cm and average base diameter was 1.48cm.
Species Frequency Percentage
Spruce 7 2.0
Rowan 1 0.3
Pine 322 91.2
Aspen 23 6.5
Total 353 100.0
Spatial analysis
Average distance between nearest neighbour along a one-dimensional horizontal line was calculated for each transect and then compared with growth rate at each transect in a spearman’s rank
correlation test. The P value was 0.5667 and so growth rate is not affected by distance from nearest neighbour.
Table 5. Frequency and Percentage of dead trees separated into species found at all sites in the RPF of the High Coast area of Sweden, trees ≤1.3m and ≤10cm DBH.
23
Discussion
Stand structure
Trees below 1.3m or 10cm DBH in the High coast area sites are generally slow growing with the oldest tree being a 224 year old Spruce which was 109mm in diameter at the base. Giving a rough 0.5 mm growth per year (see Table 1). This can largely be explained by the sparsity of trees at these sites which has a direct correlation to forest production (Liira et al., 2007). The lack of nutrients and water retention in poor growth conditions such as the RPF is also a limiting factor of growth and regeneration as the soils in most of this area are thin and exposed with rocky surfaces (Oberhuber et al., 1998)(see Figure 2a)).
In terms of old growth characteristics these sites have a wide variation in tree age and diameters at all sites. Even when we just look at trees below 1.3m or 10cm DBH all sites contain trees over 100 years. All sites apart from one (PR) had fire signs on site, showing evidence of disturbances (Sandström, 2018) and so of old growth characteristics(Kuuluvainen, 2002b). Another old growth characteristic is the existence of rare species and as previously mentioned, the RPF in the high coast is also home to the red listed beetle Chalcophora mariana.
Growth rate
Radial growth rate was highest in Pine trees compared to Spruce (see Figure 5a)) and young Pine trees had the fastest growth rate among Pine age classes (see Figure 5a)). Growth rate among Spruce trees was highest in the age category 50-100 years. Pine and Spruce growth rates are
significantly different from each other. This could be an explanation for continued Pine dominance in spite of the relative absence of fire at these sites.
Growth rate over time or rather through age classes seems to vary on a spatial scale and between Pine and Spruce species with the youngest Pine trees having the highest growth rates while the Spruce trees have their highest growth rate between years 50 to 100 (see Figure 5a)). This does again follow the evidence that Pine trees are pioneer species and Spruce trees are late successional shade tolerant species(Engelmark, 1999). Shade tolerant Spruce and fir trees appear to have a lower needle area, photosynthesis rate in high light levels, less roots compared to the shoots so more nutrient take up than photosynthesis potential) compare to Pine trees. Giving Pine trees a regeneration advantage to Spruce trees in the exposed RPF environment. Which could be an explanation for why Pine grow faster over all compare to Spruce (Messier et al., 1999).
Browsing
Trees that have been browsed by moose have higher growth rates than trees which have not been browsed. This is an unexpected result as browsing can have a detrimental effect on the growth of trees (Pastor et al., 1993) (Hester et al., 2006, Edenius et al., 2002, Felton et al., 2018).
In a study conducted by (Danell et al., 1991) found that Pines in unproductive stands were more likely to suffer mortality due to moose browsing compared with more productive stands. The moose were found to eat more of the unproductive trees needle bearing twigs as they were thinner and slow growing. This study clearly found that slower growth and in some cases mortality is a result of browsing by moose and that trees in low production forests, such as the rocky Pine forest, are effected to an even greater degree.
Moose have been observed browsing most heavily in stands with increased food amount and quality and RPF sites are not very high in nutrients and are low production it could follow that the amount of moose browsing in this area is low. Browsed trees accounted for 3 times fewer individual trees
24 than browsed individuals. Suggesting that the browsing in this area is at a fairly low frequency and is therefore not causing a decrease in growth as seen in other studies (Edenius et al., 1993) (Ball et al., 2000). Reduced plant nutrient availability effects the plants biomass and number of shoots, this in turn makes the plant less desirable to moose browsing. It could be that the browsed trees were the most desirable trees for the moose because of their biomass, therefore having a more rapid growth rate than those trees not browsed and not selected by the moose (Edenius, 1993). In another study Edenius (1993) showed that intense simulated browsing decreased Pine growth rate in dense forest.
Conversely in the least simulated browsing condition in open forest, growth rate was like that of the control condition where no simulated browsing had taken place. suggesting that open forest environments effect growth rate and browsing interactions and could be a reason that browsing in this area and a high growth rate are correlated.
In a review by Belsky in 1986 he concluded that there was no good evidence that plant
overcompensated their growth after browsing and any studies that claimed to show this had little experimental evidence, were badly designed, conducted in lab conditions or about crop plants.
However, this review looked at plants in general and not at tree growth.
On the other hand some studies did find that mammalian and simulated mammalian browsing does positively influence tree growth(Kupferschmid and Bugmann, 2013) (Stewart, 2001). In
Kupferschmid and Bugmann’s study in 2013 they found that the timing of browsing events, tree health, strength and amount of light availability affected sapling overcompensation (increase in growth after browsing). Healthy saplings with access to light that were browsed after bud
production overcompensated while other saplings did not. Another point to consider is the time of year that the trees are browsed. Trees that are mostly browsed in winter according to Kupferschmid and Bugmann would have an increased probability of overcompensation and increased growth as browsing in the summertime directly effects the growth of the tree in its growing season.
Spatial analysis
The one-dimensional horizontal aggregation of trees does not influence growth rate. Growth rate was not significantly different between trees that had neigbours within and including 1 meter away and trees which jad neigbours over 1 meter away (See Figure 6b)On the other hand as this method is not the most descriptive as it does not include the 2 or 3-dimensional aggregation. Pukkala et al in 1994 modelled competition in mixed Spruce and Pine forests, and they found that intraspecific competition had a stronger negative affect on growth in Pine and Spruce than interspecific competition with each other. Showing the possible interactions that our 1-dimensional analysis would have missed. Conversely the VMR values do show that 4 out of 5 of the sites do show aggregation.
Aggregated trees had more competition and this negatively affected growth rate. Competition between trees influences resource uptake and so growth and even mortality, for instance, trees with the largest canopy size had the largest negative effect of surrounding trees. This could be in the form of large trees taking up the most resources, nutrients and water from smaller surrounding trees. So out competing their neighbours causing loss in growth rate, vigour or even death (Fraver et al., 2014)
Although big canopy trees do effect seedling growth, seedlings are affected more by microsite conditions and micro habitat vegetation compared with older trees. Older trees require added resources due to a larger size and so competition for these resources has a higher intensity
compared to the reduced needs of seedlings (Szwagrzyk, 1992, Collins and Good, 1987). Competition with micro habitat vegetation is also more pronounced in seedlings (Zackrisson et al., 1995, Jonsson
25 et al., 1999). The canopy cover effects the regeneration of seedlings and young trees also limiting light availability and so that plants ability to photosynthesis, grow and regenerate. Any aggregation could affect this light availability and affect growth. Fraver et al in 2009 found that regeneration is normally led by the canopy or overstory structure and density shaping the understory beneath. This study, however, was conducted in a dense forest where light is a limiting factor as opposed to the RPF where light is abundant. Yet, calculated VMR values show aggregation of trees at the study sites and this aggregation could influence regeneration as shown above.
In opposition no aggregation could also influence growth and regeneration. Areas with sparse tree cover have increased light levels, wind levels and exposed soils. This could reduce water retention and so effect regeneration potential in these trees (Oberhuber et al., 1998).
There could be an advantage to some aggregation or close proximity to a well-established tree in the form of resource exchange via the mycelial network. This network allows for exchange of and
increased efficiency of nutrient uptake by the seedling roots. In a study conducted by Dickie et al in 2002 established trees increased seedling uptake of nitrogen, phosphorus, other nutrients via the mycelial network and perhaps increase the seedling growth rate. This result could either be due to the helping hand of the established tree to the seedlings via the mycelial network or an effect of competition from a different species. Either way the growth rate of seedlings was affected by neighbouring trees (Dickie et al., 2002) more reference. Mycorrhizal fungi symbiosis and the mycelial network are especially bifacial to nutrient deficient trees in areas such as the rocky Pine forest as the fugal symbiosis helps the plant to get nutrients from their surrounding that they would not be able to do alone (Van Der Heijden et al., 2015) and any interaction through this network from established trees to seedlings and sapling could give a growth advantage (Dickie et al., 2002).
Sprucification
By looking at species composition the question of Sprucification in the area can be answered. There is a clear dominance of Pine trees in all the sites studied in both studies and there are in fact no Spruce trees recorded below the age of 11 years (see table 1 and figure 10) . If we only look at these result the Spruce population at these sites could be in danger of extinction in the area. However, if we look at figure 4 diameter frequency of both Pine and Spruce trees ≤1.3m and ≤10cm DBH the Spruce diameter is steadier compared with age through all diameter classes. As pioneer species, Pine seedlings grow rapidly and many and then have high mortality (see Table 5). Spruce seedlings have slow growth in the early stages when they are out competed by Pine but can stay growing slowly in the shade of other seedlings until the increase water and nutrients can be found (Engelmark, 1999).
This could be why Spruce trees below 1.3m and 10cm DBH survive to older age classes and have low seedling frequency compared with Pine in early age classes. Unfortunately for Spruce in general nutrient poor soils, limited water and some fire events (even if decreased) (Esseen, 1997) mean that the environment in the RPF in High Coast area has not changed enough to favour Spruce, at the moment.
In a study conducted by Hedwall and Mikusinski in 2016 based on 60 years of national forest inventory data from protected forest, they assessed changes that had occurred in species
composition. They found no change in tree species composition in protected forests. Concluding that these forests contain very few tree species and so there are limited successional pathways for tree species to follow and so they tend to stay at equilibrium. Another stated explanation could lie in Spruce trees increase likelihood to die as an older tree. Being more susceptible to climatic disturbances (Zackrisson, 1977) wind damage (Peltola et al., 2000), whereas Pine trees are more resilient to wind, climate change and fire (Hedwall and Mikusiński, 2016, Zackrisson, 1977, Engelmark, 1984, Engelmark et al., 1994). From the data in figures 10 and 11 we can see that the
26 proportion of Spruce in older age categories is very low. If there were signs of Sprucification in the future it would be likely that Spruce trees would outlive Pine and so increasing in numbers. The evidence here is to the contrary as Pine trees out live Spruce. In 1994 Engelmark also conducted a study that shows no Sprucification affect in boreal Pine and Spruce forests with a known reduction in fire frequency (Engelmark et al., 1994). Pine in general is known to grow well in low nutrient and low productivity areas such as the high coast(Liira et al., 2007, Esseen, 1997)and so it follows that it will continue to do well in this environment compares to Spruce. These findings are contradictory to some predicted outcomes of reduce fire events on the composition of Pine dominant boreal forests in this area(Linder and Östlund, 1998, Linder, 1998)
Spruce is known to be late successional species and as seen in figure 11, Spruce numbers in the earlier age classes are lower that Spruce frequency in middle age categories (51-75yrs). Numbers of Spruce were lower in every age category compared with Pine numbers. Spruce therefore are disadvantaged in these sites due to numbers. Late successional species such as Spruce are adapted to soils of later successional stage, Conversely, here in the RPF the soil is thin and shallow. Perhaps never reaching composition of those soils seen in late successional stages. This could have a major effect on the capacity of Spruce to compete with Pine and so reducing Spruce regeneration ability and could explain why Spruce has the highest growth in later stages when soil conditions are more suitable for Spruce optimal growth (Jobidon et al., 1989) (Kronzucker et al., 1997).
It is also worth considering the role of climate change and increase temperatures and so drought in these areas in the future (Schlyter et al., 2006). It is quite possible that even with a reduction in fire frequency the conditions in the RPF will not be optimal for Spruce trees. Further reducing the probability of Sprucification in this area.
Conclusion
Rocky Pine forests in northern Sweden’s high Coast area show many old growth characteristics such as varied, uneven age distribution and presence of red listed species. They do not seem to be experiencing signs of Sprucification due to the anthropogenic reduction in fire frequency. This is probably due to the adaptation of Pine to grow and thrive in this nutrient poor, low water and exposed sites compared with shade tolerant, nutrient dependant late successional Spruce trees.
Even without fire it seems that Pine are very well adapted to these sites.
Spatial analysis shows that there is aggregation in all but one of the sites (Porsmyrberget), a factor that can both aide and deter seedling and sapling growth. Mycelial networks allow exchange of nutrients between nearby established trees and young trees that can facilitate an increase chance of growth compared to trees further from the established tree. On the other hand, aggregation can be detrimental to growth, when competition from surrounding trees reduces the chance of a young tree from getting resources required for successful growth and in some cases survival.
Trees that have been browsed by moose in the RPF have a faster growth rate on average than trees with no evidence of browsing. This is perhaps evidence of overcompensation by the trees or is a result of preference by moose to pick the trees with most vigour.
Radial growth rate in RPF trees is fastest in the lowest age class <50 years as would be expected. Yet if we separate this into Pine and Spruce we see that young Pine trees are the fastest growing. In Spruce it is the slightly older trees of 50-100 years that grow at the fastest rate. Perhaps this could be attributed to the corresponding successional stage that these trees occupy and the properties of each species. Pine as a pioneer species is fast growing in high light levels as canopy cover increases
27 shade tolerant Spruce trees can grow in conditions best suited to the species. In the RPF and other areas with a sparse canopy cover and so little light limitations Pine species can dominate(Linder, 1998, Engelmark et al., 1998, Linder and Östlund, 1998).
There is much more that can be done to look even closer at regeneration in rocky Pine forests.
Three-dimensional aggregation analysis would help us to understand in more detail the growth responses in the RPF to aggregation. Investigating of both intra and inter specific competition to fully understand how aggregation affects growth in the RPF and to compare these results with dense forests to observe the relationship between aggregation and resource availability. It would also be beneficial to know the numbers and even movements of the moose population in the RPF so that browsing severity can be assessed between sites based on the population of moose in the area.
To compare these forests to RPF in other areas and their growth rates would aid us in pinpointing the relationship between tree regeneration and the forest as a whole. It would also be interesting to investigate the RPF using dendrochronological techniques to see what has affected growth in the past and use modelling techniques to look at what could affect growth in the future.
As natural or old growth forests give way to even aged, low biodiversity forests it is important that we try to understand forest dynamics and regeneration especially in the light of climate change. This knowledge could allow us to manage our forest industry in uncertain times and ensuring the
organisms thrive within them.
Acknowledgements
Thank you to Bengt-Gunnar Jonsson and Jennie Sandström and all in Mittunivesitetet Natural Science department for help and guidance. Thank you to all landowners for the permission to use their land and trees for this project. Thank you to Marijn Post for as well as Laurens Post and Emiel Driessen invaluble fieldwork assistance!
28
References
ANDERSSON, R. & ÖSTLUND, L. 2004. Spatial patterns, density changes and implications on biodiversity for old trees in the boreal landscape of northern Sweden. Biological Conservation, 118, 443-453.
ANGELSTAM, P., PEDERSEN, S., MANTON, M., GARRIDO, P., NAUMOV, V. & ELBAKIDZE, M. 2017.
Green infrastructure maintenance is more than land cover: Large herbivores limit
recruitment of key-stone tree species in Sweden. Landscape and Urban Planning, 167, 368- 377.
ANGELSTAM, P. K., T. 2004. Boreal forest disturbance regimes, successional dynamics and landscape structures. A European perspective. In P. Angelstam, M.Do ̈nz-Breuss & J.-M. Roberge (Eds.), Targets and tools forthe maintenance of forest biodiversity. Ecological Bulletins,, 51,117-136.
AXELSSON, A.-L. & ÖSTLUND, L. 2001. Retrospective gap analysis in a Swedish boreal forest landscape using historical data. 147, 109-122.
BALL, J. P., DANELL, K. & SUNESSON, P. 2000. Response of a herbivore community to increased food quality and quantity: an experiment with nitrogen fertilizer in a boreal forest. Journal of Applied Ecology, 37, 247-255.
BENGTSSON, J., NILSSON, S. G., FRANC, A. & MENOZZI, P. 2000. Biodiversity, disturbances,
ecosystem function and management of European forests. Forest Ecology and Management, 132, 39-50.
BERGLUND, M. 2012. THE HIGHEST POSTGLACIAL SHORE LEVELS AND GLACIO-ISOSTATIC UPLIFT PATTERN IN NORTHERN SWEDEN. 94, 321-337.
BOND, B. J. & FRANKLIN, J. F. 2002. Aging in Pacific Northwest forests: a selection of recent research.
Tree Physiology, 22, 73-76.
COLAK, A. H. 2003. Effects of microsite conditions on Scots pine (Pinus sylvestris L.) seedlings in high- elevation plantings. Forstwissenschaftliches Centralblatt, 122, 36-46.
COLLINS, S. L. & GOOD, R. E. 1987. The Seedling Regeneration Niche: Habitat Structure of Tree Seedlings in an Oak-Pine Forest. 48, 89.
DANELL, K., NIEMELA, P., VARVIKKO, T. & VUORISALO, T. 1991. MOOSE BROWSING ON SCOTS PINE ALONG A GRADIENT OF PLANT PRODUCTIVITY. Ecology, 72, 1624-1633.
DICKIE, I. A., KOIDE, R. T. & STEINER, K. C. 2002. INFLUENCES OF ESTABLISHED TREES ON
MYCORRHIZAS, NUTRITION, AND GROWTH OF QUERCUS RUBRA SEEDLINGS. 72, 505-521.
EDENIUS, L. 1993. Browsing by Moose on Scots Pine in Relation to Plant Resource Availability.
Ecology, 74, 2261-2269.
EDENIUS, L., BERGMAN, M., ERICSSON, G. & DANELL, K. 2002. The role of moose as a disturbance factor in managed boreal forests. Silva Fennica, 36.
EDENIUS, L., DANELL, K., BERGSTRÖM, R. & BERGSTROM, R. 1993. Impact of Herbivory and Competition on Compensatory Growth in Woody Plants: Winter Browsing by Moose on Scots Pine. Oikos, 66, 286.
ELLIS, N. M. & LEROUX, S. J. 2017. Moose directly slow plant regeneration but have limited indirect effects on soil stoichiometry and litter decomposition rates in disturbed maritime boreal forests. Functional Ecology, 31, 790-801.
ENGELMARK, O. 1984. FOREST FIRES IN THE MUDDUS-NATIONAL-PARK (NORTHERN SWEDEN) DURING THE PAST 600 YEARS. Canadian Journal of Botany-Revue Canadienne De Botanique, 62, 893-898.
ENGELMARK, O., HOFGAARD, A. & ARNBORG, T. 1998. Successional Trends 219 Years after Fire in an Old Pinus sylvestris Stand in Northern Sweden. 9, 583.
ENGELMARK, O., KULLMAN, L. & BERGERON, Y. 1994. Fire and age structure of Scots pine and Norway spruce in northern Sweden during the past 700 years. New Phytologist, 126, 163- 168.
ENGELMARK, O. H., H 1999. Coniferous forests. Acta Phytogeographica Suecica, 84: 55–74.
ESSEEN.
