Somidh Saha

Introduction

Trees of wild-cherry(Prunus avium L.),an important native broadleaf species in Germany,are a source of valuable timber(Hein 2009).They also provide food for wild animals,and their attractive flowers in the spring and fall leaves in the autumn increase the aesthetic value of forests.It is a fast growing and light-demanding species.The top height of cherry trees may reach 25 m on sites with deep,light,silty,and fertile soils with a good water supply and pH from 5.5 to 8.5(Savill 2013).Cherry trees are less tolerant to drought than ash(Fraxinus excelsior L.)and sessile oaks[Quercus petraea Matt.(Liebl.)];however,they are more drought tolerant than other co-occurring trees species,such as European beech(Fagus sylvatica L.),sycamore(Acer pseudoplatanus L.)and lime(Tilia cordata Mill.),in broadleaf mixed forests(Scherrer et al.2011).Its high ecological,social and economic value have made wild cherry a species of choice for establishing new broadleaf mixed forests in Germany since the 1990s(Schraml and Volz 2009;Thies et al.2009).

Cherry treesare commonly used in reforestationsatclearcuts,large gaps or in wind-thrown sites because they require open canopy conditions to grow(Savill 2013).The financial goal of a cherry forest is to produce nearly 50 crop trees per hectare ofa preferable DBHof40 cm with a straightbranchfree bole length(BFBL)up to 5–6 m within 50 years of planting(Thies et al.2009).To achieve this goal,cherry stands are often established by conventional row planting with at a 3×1 m spacing with other shade-tolerant tree species,such as lime,which can constitute up to 25%of the initial planting density.Since cherry trees are susceptible to deer browse,seedlings have to be protected during planting.A high initial planting density in row planting is supposed to ensure rapid crown closure and thus reduce the impact from competing vegetation(Drouineau et al.2000).

Conventional row planting of cherries,however,is associated with high costs.For example,planting 3300 seedlings(i.e.,3×1 m spacing)with browse protection and other site preparation costs would reach nearly Euro 16,500 per hectare in the southwestern German state of Baden-Wurttemberg.Moreover,pruning the lower green branches to achieve the desired branch-free bole length in cherry trees increases the cost.Although this pruning is a common practice because cherry trees have a weak selfpruning capacity and occlusion rate(Hein 2009).The influence of the initial growing space or planting density on the self-pruning capacity of wild cherry trees has not been properly investigated.The high initial cost of stand establishment and artificial pruning in young cherry stands established by conventional row planting has motivated foresters to search for more efficient and profitable ways of planting trees to produce high-quality timber.

Planting seedlings with close initial spacing in groups and maintaining wide spacing between groups have emerged as alternatives to row planting for regenerating high-quality oaks in mixed forests in Germany and other central European countries(Saha et al.2012,2017;Demolis et al.1997;Brang and Burgi 2004;Ruhm 1997;Gockel 1995).In oak group planting,the planting cost can be reduced by 50%due to the low number of seedlings needed for planting(Saha 2012;Ehring and Keller 2006;Petersen 2007).However,studies on oak group planting showed that reducing the number of seedlings did not hamper the attainment of the desired silvicultural and ecological attributes.For example,group planting of oaks yielded the same or more potential future crop trees(PFCTs),better tree growth and stability against wind damage and higher biomass and tree species diversity than did conventional row planting(Saha et al.2012,2013).However,whether wild cherry trees would similarly benefit by group has not yet been studied.

Competition between trees plays a crucial role in the diameter,height and timber quality of valuable broadleaf species in young broadleaf forests.Trees usually compete for light,resulting in size asymmetry among standing trees in fertile forest stands that have a sufficient water supply(Pretzsch and Biber 2010).However,the impact of neighbourhood competition on tree growth can differ between conspecific trees(i.e.,intraspecific competition)and trees from other species(i.e.,interspecific competition)due to the change in the wavelength of light with the canopy height(e.g.,red vs.far-red light)as well as the diversity of the light-use efficiency among species(Binkley et al.2013;Ballare et al.1990;Forrester 2014).For example,intra-and interspecific competition between trees influence growth,stability and the length of the BFBL in young oak trees in different ways(Saha et al.2014).The total amount of solar irradiation above the canopy of a forest during the growing period(i.e.,from April to September in Germany)is a major factor that influences growth.In addition,the availability of photosynthetically active photons under the tree canopy controls the self-pruning of branches(e.g.,BFBL).High shedding on the lower part of the tree trunk(i.e.,low density of photosynthetically active photons)accelerates the selfpruning of lower branches.Therefore,studying the influence of neighbourhood competition and light availability on tree growth,stability and quality are essential to assess the performance of trees growing in young forests.Assessing the impacts of competition on the growth and quality of PFCTs is necessary for young cherry forests.This assessment will help foresters to make appropriate decisions to mark and then fell or girdle competing trees at the time of tending and thinning.However,the influence of light availability as well as intra-and interspecific competition among young cherry trees planted in groups and rows has not yet been studied.

On December 26,1999,the severe winter hurricane Lothar damaged an estimated 185×106m3of timber in forests in Germany and other countries(Schmidt et al.2010;Schutz et al.2006).The sudden destruction of large chunk of forests and the drop in timber price after the storm created multiple challenges to finding a cost efficient method for restocking the wind-thrown forest sites(United Nations Economic Commission for Europe 2000;Brang and Combe 2001).In a forest district in southwestern Germany,cherry group and row planting were established by the local forestry administration at a site where Lothar completely uprooted a stand of Norway spruce[Picea abies(L.)K.Karst.].This planting was the first attempt to plant wild cherry trees using the group technique.In the present study,we compared the mortality,tree growth(diameter and height),tree quality(crown shape,stem form,branch-free bole length,and the number of future crop trees per hectare),and social class of cherry trees planted in groups with those in planted in traditional rows.Additionally,the influence of intra-and interspecific competition and light on tree growth and quality was assessed.

Materials and methods

Description of the study site and planting design

The study site contains bottomland forest(150 m a.s.l.)adjacent to the Rhine River.It is in the forest district of Neuried and close to the village Ichenheim in the Ortenau region of Baden-Wurttemberg,Germany.The mean annual temperature of the study site is 10.2°C,and the average annual precipitation is 832 mm.The soil is mainly gleyic cambisol,originating from alluvial deposits covered by a layer of loess,and stagnogleyic cambisol,originating from basalt loam,siltstone or sandstone(Gauer and Aldinger 2005).

There were two treatments in the reforestation study,which was established on wind-thrown sites.The first was a stand established by group planting,and the second site was a stand created by conventional row planting.Both stands were 0.5 ha in size and 14 years old.The stands were adjacent to one another,and row planting was used as a control treatment for group planting.The site characteristics did not differ between the two stands as the mean annual increment was 8.5 m3ha-1a-1in both stands as per the forest management plan.The seedlings for planting in both stands were sourced from the same nursery.Site preparation was limited to patches selected for groups since broken stumps and slash from the previous forest were kept between the groups.By contrast,site preparation was extensive in row planting,where the fallen logs,broken stumps and slash in the whole stand area had to be removed to plant trees in uniform 3×1 m rows.In each group,five cherry trees were planted with a 1×1 m spacing.Seven lime trees encircled the group,planted 1 m away from the cherry trees.Therefore,12 trees were planted in total in one cherry group.The groups were placed with a 13×13 m spacing(centre to centre),and 60 cherry groups(i.e.,720 seedlings)were planted per hectare.The distance between the trees in row planting was 3×1 m(i.e.,3300 seedlings per hectare).In the row planting stand,75%of the planted seedlings were cherry,and the rest were limes.Limes were planted as ‘trainer trees’,a practice that is common when planting valuable broadleaf trees.Trainer trees usually have lower growth rates than target trees(e.g.,cherry in this case)and provide shelter to the lower trunk of the target trees to accelerate the self-pruning of branches to promote tree quality development(von Lupke 1998).Cherry and lime seedlings were individually protected from deer browse by shelters(i.e.,growing tubes)in both forest stands.Tending operation was carried out 5 years after planting in the group and the row planting stands when a few fast-growing naturally regenerated hornbeam trees were felled and competing ground vegetation was cut.The cherry and lime trees were not thinned or pruned before data collection.

Sampling design and data collection

Ten cherry groups were randomly selected within the group planting stand as the inventory.Similarly,10 plots of 10×10 m were installed in the row planting stand.A minimum distance of 20 m was maintained between the plots or groups to cover the whole stand.Therefore,the sampling units for this study are ‘groups’and ‘plots’in group and row plantings,respectively.The fieldwork was carried out in the summer of 2015.

Trees taller than 1.3 m in height were inventoried inside each group and plot.The status of survival,diameter at breast height(DBH),branch free bole length(BFBL),stem form,crown shape,and social class were noted for each cherry tree(＞1.3 m height).The BFBL is the length of the stem from stem-base to the first green branch,with the exception of epicormic shoots(Nicolini et al.2001;United States Forest Service 2011).Stem form,crown shape,and social class were classified as an ordinal response(Pretzsch 2009;Kuehne et al.2013;Saha et al.2012;Kraft 1884).PFCTs were selected during the field inventory from the group-and row-planted cherry trees.The criteria for the selection of the PFCTs were the following:(1)the tree must belong to the dominant layer of the canopy(social class 1);(2)the tree must have a monopodial crown without steep forking(crown shape class 1);(3)the tree must have a straight stem without any bend(stem form class 1);and(4)the tree must not show any symptoms of pathogen infestations,epicormic shoots,crown die-back,or frost damage.For lime trees located in groups and plots,the status of survival and DBH were recorded.Tree quality was not assessed on lime trees because those trees were not planted for timber production,but rather served as trainer trees to achieve desirable tree quality among the high-value cherry trees.The heights of the cherry trees were estimated by the height curve created for both stands.The height to diameter ratio[HD ratio is(height/DBH)×100]was calculated for each cherry tree(von Lupke 1991).

Twenty PFCTs were selected(i.e.,10 from each treatment and one from every plot or group)as target trees for competition analysis.Only one tree per plot or group was selected as a target tree to avoid overlapping the competitors and spatial autocorrelation.The competition analysis was confined to only PFCTs because they are most important for foresters to monitor over the long term and for developing a tending and thinning strategy.Trees taller than or equal to 2 m that fell within 3 m of the selected PFCT were considered to be competitors.The DBH and distance to the target tree for each competitor were measured.Based on those measurements,Heygi’s competition index,a well-established competition index for assessing neighbourhood competition in forest trees that is highly correlated with other competition indices,was calculated(Hegyi 1974;Ammer et al.2005;Saha et al.2014).The aggregate competition is the entire neighbourhood competition for a target PFCT from all competitors irrespective of species.The aggregate competition index of a target tree was further divided into intraspecific(i.e.,cherry trees)and interspecific(i.e.,planted limes and other natural regenerating trees)competition.Hemispherical photographs were used to measure the light supply in the groups or plots.Photographs were taken in the centre of the groups and in the centre of the 10×10 m plots in row planting.The total site factor and PPFD under the canopy were calculated from the photographs using the software program Win-SCANOPY (Regent Instruments Inc.,Quebec City,Canada).The total site factor is the ratio of the average daily direct and diffuse radiation received under the canopy to that received over the canopy during the growing season(April–September).The values of the total site factor were used to observe the influence of light availability on the DBH,height and HD ratio of cherry trees.The PPFD under the canopy is the number of photons in the 400–700 nm range of the sunlight spectrum received by a unit surface area per second during the growing season under the canopy of a tree.The values of PPFD measured under the canopy were used to test the influence of light availability on the BFBL of the cherry trees.

Statistical analyses

The Shapiro–Wilk test was used to determine the distribution of continuous variables by assessing the residuals between actual values in collected data and predicted values from a theoretical Gaussian distribution.Student’s t test was applied to compare variables that followed the Gaussian distribution.For the data from a non-Gaussian distribution,the Mann–Whitney U-test was used to compare the distribution of two variables(Sokal and Rolhf 1995).Mortality,DBH,height,HD ratio,BFBL,and number of PFCT per hectare were each compared between the group and the row planting either by Student’s t-test or by the Mann–Whitney U-test.The crown shape,social class,and stem form are ordinal response variables;therefore,responses were cross-tabulated using 2×4 contingency tables for group and row plantings.Kendall’s Tau-b correlation coefficient was chosen because it is a nonparametric measure of the strength and direction of association that exists between two variables(e.g.,stem form classes vs.stand type)measured on an ordinal scale(from 1 to 4).

The competition analysis was done in following steps.First,the number of competitors and Heygi’s competition index values per target PFCT were compared between group and row plantings.Then,generalized linear model(GLM)analysis was carried out to determine the influences of intra-,interspecific competition,aggregate competition(intra-and interspecific combined),and light(total site factor or PPFD under canopy)on the DBH,height,HD ratio,and BFBL of the PFCTs.GLM analysis was performed separately for group and row plantings to separate the effect of planting type from other independent variables,such as light availability and competition.The influence of aggregate competition and light availability on the DBH,height,HD ratio,and BFBL was studied first.Next,aggregate competition was divided into intraspecific and interspecific competition to test whether a change in plant composition alters neighbourhood interactions.The GLM analysis permits the fitting of linear models to any distribution function included in the exponential family using a maximum likelihood analysis.The GLM directly fits the expected mean from the theoretical Gaussian distribution of the dependent variable,hence avoiding biases for transformed (e.g.,log-transformed)linear models(McCullagh and Nelder 1989).The independent variables were not autocorrelated because the collinearity between independent variables was checked by an auto-correlation analysis,as suggested by Dormann et al.(2013).The sample size was relatively small in this study.Therefore,bootstrapping and Monte Carlo simulations were used while performing the statistical analysis.

Results

Mortality,height,DBH and stability(HD ratio)

The mortality of planted cherry trees in row planting was significantly higher(66%)than in group planting(26%).The mortality of trainer trees in row planting was 80%,compared with only 29%in group planting.The height and DBH growth were significantly higher among cherry trees grown in groups than in row planting.The HD ratio of cherry trees was significantly lower in group planting.The mean DBH of planted lime trees were 4.3 and 5.3 cm in group and row planting,respectively.However,this difference in the DBH of lime trees between group and row planting was not significant(Table 1).

Tree quality development

The proportion of cherry trees with monopodial and steeply angled crowns(crown shape class 1 and 2)was signif icantly higher(65%)in group than row planting(44%).Similarly,the number of trees with straight or slightly crooked stems(stem form class 1 and 2)was higher in group planting(95%)than in row planting(62%).By contrast,the proportion of trees with heavily forked brushy crowns as well as deformed and crooked stems was much higher in row planting than in group planting(Table 2).The mean length of the BFBL was 2.4 m for cherry trees grown in groups and significantly higher than for trees from row planting.Interestingly,the BFBL had already reached 29%of the tree height(i.e.,8 m)in group planting,whereas in row planting,it was only 21%of the tree height(Table 1).The number of PFCTs per hectare was 190 in row planting,significantly higher than in group planting(120 PFCTs per hectare)(Table 1).Nearly 2475 seedlings had to be planted per hectare in row planting to achieve 190 PFCTs after 14 years.This result means that the success rate in row planting to produce one PFCT was 1:13.In contrast,to obtain 120 PFCTs in a group planting,only 300 seedlings(i.e.,60 groups with five seedlings in each group)were planted per hectare,for a high success rate of 1:2.5 to produce one PFCT.

Influence of plant competition and light on tree growth,stability and branch-free bole length

The PFCTs of row planting had a high number of competitors(i.e.,14 per target trees),which differed signif icantly from that of the group planting(i.e.,8 per target trees)(Table 1).The number of competitors of conspecific cherry trees was 4 in row planting,in contrast to 2 in group planting(Table 1).At the same time,the number of interspecific competitors was also significantly higher in row planting(i.e.,10 trees)than group planting(i.e.,6 trees)(Table 1).The average Heygi’s index per target tree was significantly higher in row planting than in group planting due to the high number of competitors in the row planting (Table 1).The average Heygi’s index for intraspecific competition was significantly greater in row planting,but the interspecific competition did not differ between row and group planting.The availability of light in the centre of the groups varied significantly from the plots in the row planting.The total site factor and PPFD under the canopy was considerably higher in row planting than in the centre of the groups(Table 1).

Overall,aggregate competition and light availability did not influence the DBH,height,physical stability of trees(HD ratio),or BFBL in group planting(Table 3).In the second GLM analysis,aggregate competition was then split into intra- and interspecific competition. Neither intraspecific competition nor interspecific competition was related significantly to those four response variables.The light availability inside the groups remained an insignif icant factor even when the competition was split into intraand interspecific groups(Table 3).

The PFCTs in the row planting showed different trends compared with those in the group planting(Table 3).The aggregate competition significantly reduced the DBH,height and HD ratio of the target PFCTs in the group planting.However,aggregate competition had no effect on the BFBL.Both intra-and interspecific competition significantly impacted the DBH,height,and HD ratio.Intraspecific competition also reduced the BFBL of the PFCTs,whereas interspecific competition had no effect.As in the group planting,the light availability(TSF and PPFD)was insignificant for all response variables in the row planting(Table 3).

Discussion

Difference of mortality,growth,and quality of cherry trees between group and row planting

The high mortality of cherry trees in row planting was likely a consequence of heavy competition from ground vegetation and naturally regenerating trees throughout the first decade after planting.The shading-induced asymmetric competition between higher branches of overtopping trees as well as from the same tree and lower branches of cherry trees could result in the higher mortality of lower branches and eventually whole-tree mortality(Pretzsch and Biber 2010).Similar results have been observed in other young,dense broadleaf forest stands(Henriksson 2001;Kint et al.2010).Cherry is a strongly light-demanding species that is known to be very susceptible to competition from ground vegetation during the establishment of the stand when trees try to maintain their apical dominance(Hein 2009).The forest site was fertile and located close to the Rhine River with alluvial deposits.In such sites,blackberries(Rubus spp.),stinging nettles(Urtica dioica L.),grasses(Calamagrostis spp.),and bracken ferns(Pteridium spp.)form dense bushes when an ample amountof light is available on the forest floor until 4–5 years from planting.In the study region,sycamore(Acer pseudoplatanus L.),white willow(Salix alba L.),aspen(Populus tremula L.)and hornbeam(Carpinus betulus L.)gradually replace ground vegetation and form dense thickets of tall and lanky stems(Saha et al.2013).If the site is not regularly tended,competition from ground vegetation and naturally regenerated woody plants could result in a high mortality of planted trees(Gussone and Richter 1994;Petersen 2007;Saha 2012).Perhaps this competition was the primary cause of the high rate of cherry tree death in row planting.In contrast to the row planting,in the group planting,the development of ground vegetation and natural regeneration mainly occurred in the space between groups,and this spatial segregation of competitors and planted cherry trees probably reduced cherry mortality(Rock et al.2003;Gockel et al.2001).The trainer trees planted at the periphery of the groups had increased the survival rate of cherry trees,as observed in group plantings of oaks in the past(Saha et al.2012).Such positive effect of trainer trees on the survival of cherry trees may have resulted due to the protection of cherry trees by trainer trees from the advancing blackberries and other ground vegetation,which would not be the case in rows(Rock et al.2004).In addition,the study region(upper Rhine River valley)was stricken by severe summer droughts in 2003,2010 and 2011(German Weather Service 2011;Kohler et al.2010;Rebetez et al.2006;Chakraborty et al.2013).In 2003,the stands were only 2 years old.Impacts from such drought events on mortality would certainly be higher on trees in row planting,which were already under stress due to a predisposition to intense competition in rows(Panayotov et al.2016).This claim warrants further research that could help diagnose the causes of mortality.

Table 1 Statistical comparisons ofmeanmortality,DBH,height,branch freebolelength(BFBL),potentialfuturecroptrees(PFCTs)per hectare,competition,and light condition between group and row planting

Table 2 Comparisons on crown shape,stem form,and social class of cherry trees between group and row planting

Table 3 Influences of competition and light availability on the DBH,height,tree stability(HD ratio)and BFBL of PFCTs in the group and row planting(summary of GLM analyses)

A high initial growing space(i.e.,3 m2per sapling)did not appear to be beneficial for the height,DBH,and physical stability(height to DBH ratio or HD ratio)of cherry trees in conventional row planting compared with trees that had a 1 m2initial growing space per sapling in group planting.The required initial growing space to attain desirable growth and tree quality in the stand initiation and qualification stages is a matter of debate and will vary between tree species.To date,little information is available in the published literature on the optimum spacing requirement for high-value stand establishment using wild cherry trees.Previous spacing experiments on oaks demonstrated that height and DBH growth did not differ significantly in trees with a 1 m2initial growing space compared with trees planted with an initial 2 m2growing space(Spellmann and Baderschneider 1988;Leder 1996;Kuehne et al.2013;Gaul and Stuber 1996;Knowe and Hibbs 1996).The broken stumps and coarse woody debris were mostly maintained between the groups after storm damage,which was not the case in row planting,where intense site preparation had to be taken to plant saplings with a homogenous 3×1 m spacing.Therefore,the forest floor in row planting received direct sunlight due to the absence of broken stumps and coarse woody debris,resulting in the vigorous regeneration of ground vegetation.High competition in row planting perhaps reduced the water and nutrient supply to cherry trees.The continuous shortage of resources caused poor diameter and height growth in cherry trees in row planting.The high mortality of cherry and planted lime trees did not help the surviving cherry trees.It is possible that fast-growing species quickly occupied the vacated space created by the die-back of cherry trees.The poor stability(high HD ratio)of the surviving cherry trees in row planting was due to concomitantly reduced diameter growth(Saha et al.2014).Close initial spacing and poor light availability hindered natural regeneration within the cherry groups(Saha et al.2013).Therefore,competition pressure from those trees on the height and diameter growth of cherry trees was lower in groups.Deer prefer open forest gaps to browse(Kuijper et al.2009),and they probably visit the space between the groups more frequently in the first few years after planting,hence minimizing the growth of naturally regenerated trees.As a result of these factors,cherry trees in groups dominated the forest canopy in groups,whereas most cherry trees were mainly codominant or suppressed trees in row planting.

The proportion of cherry trees growing with monopodial crowns and straight stems was found to be higher in groups.This finding supports the outcome from a past study in which oak trees with monopodial crowns and straight stems were found to be higher in group planting than in row planting,which was attributed to difference in initial growing space of trees in two contrasting planting design(Saha et al.2012,2017).Cherry trees planted in two stands were sourced from the same nursery and provenance;therefore,the difference in tree quality was not attributable to genetics.High competition from ground vegetation and other fast-growing species,such as sycamore and hornbeam,also hampered the crown shape and stem form of cherry trees in row planting,resulting in suppressed trees.When light-demanding trees,such as cherries,faced prolonged shade,they developed epicormic shoots,lost their apical dominance and were deformed(Savill 2013),which was the case for many of the cherry trees in the row planting.By contrast,the majority of the cherry trees in the group planting were either dominant or codominant due to better growth in DBH and growth.Cherry trees in groups benefitted from full sunlight,grew straight and maintained their apical dominance.Therefore,the BFBL was higher in group planting,that had already crossed the desirable length of 25%of tree height for young high-value broadleaf forests(Spiecker 1991).The high survival rate of lime trees enabled a longer BFBL for cherries in the group.It has already been proven for oaks that trainer trees such as limes can increase the BFBL in a group planting(Saha et al.2012;Ehring and Keller 2006;Petersen 2007).In absolute terms,the density of PFCTs per hectare was higher in the row planting(i.e.,190 trees per ha)than in the group(i.e.,120 trees per ha).However,the density of planted cherries was 2475 per hectare in row planting compared with only 300 per hectare in the group.Therefore,in relative terms,the success rate of group planting to obtain a PFCT is much higher than that of row planting.This result should be treated cautiously as it could vary with the density of competing vegetation,intensity of early stand tending and site characteristics.Nevertheless,this assessment confirms the result from oak group planting,where the density of PFCTs in groups did not differ from rows,although the initial planting density per hectare was almost 50%lower in group planting than in row planting(Saha et al.2012).On an average,every cherry group contained at least two PFCTs,although only one PFCT per group will be released from competition in the future,resulting in 60 PFCTs per hectare,distributed uniformly over the stand(Skiadaresis et al.2016).We could not address in the present study whether these 60 PFCTs will be sufficient to reach the commercial goal of 50 harvestable crop trees per hectare at the end of the production period.There is a lack of information on the devaluation of timber quality in future crop trees over time(Abetz and Kladtke 2002;Abetz and Ohnemus 1999),and risk assessment studies on the development of PFCTs over time are rare for the fast-growing,high-value broadleaf trees in Germany.

Influence of neighbourhood competition and light availability

In this study,neighbourhood competition was quantified by Heygi’s index for the conspecific cherries and interspecific trees from other species.The assumption was that similar competition indices from different population groups(i.e.,intra-vs.interspecific)would result in different growth and quality responses of the target cherry PFCTs.In addition to tree size and distance,competition index,species-specific traits,such as light absorption and light-use efficiency,also influence competition among trees(Binkley et al.2013).Therefore,in group and row planting stands,which present variable mixtures of planted cherries and naturally regenerated fast-growing shade-tolerant and light-demanding broadleaf trees,the same competition index may result in different responses of the target cherry trees.However,the results from the competition analysis in this study should be cautiously interpreted due to the low number of target PFCTs and because the data were gathered for only one vegetation period.

Intraspecific competition

The intraspecific competition between cherry trees had a higher negative impact on the DBH,height and tree stability of the PFCTs in widely spaced rows and had no impact on the more closely planted groups.Early in stand establishment,future crop trees in groups probably reached the top layer of the canopy and benefitted from full sunlight.Later,most cherry trees other than PFCTs became codominant;hence,the impact of intraspecific competition among groups was not significant.By contrast,the PFCTs in row the planting were not as vigorous as those in the group planting.The surviving cherry trees in the target trees’neighbourhood still exerted strong competition on the diameter and height of the target PFCTs.A recent study showed that the height and diameter growth was hampered by intraspecific competition when cherry trees were located in a conspecific neighbourhood(Loewe et al.2013).In other broadleaf species,such as in European beech trees and oak trees,intraspecific competition reduces the height and diameter in young forest stands established by artificial regeneration(Ammer et al.2005;Saha et al.2014).The negative impact of intraspecific competition on the stability in row planting shows that competition for light became intense and encouraged trees to allocate their resources to growth in height rather than to growth in diameter to maintain their position in the canopy(Waring and Schlesinger 1985).

Intraspecific competition did not influence the BFBL of future crop trees in the group planting.This result agrees with findings on an oak group planting stand of the same age from the same locality,where intraspecific competition had no effect on the BFBL of future crop trees(Saha et al.2014).Intraspecific competition in young forest stands probably allowed more far-red light to reach shaded leaves growing on lower branches and reduced the rate of selfpruning(Ballare et al.1990;Rock et al.2004).This situation was more adverse in row planting.The PFCTs grown in rows retained their lower branches on the trunk to continue photosynthesis because intraspecific competition impacted growth by reducing the crown size in the upper part of the canopy,which could be reason for the adverse relationship between intraspecific competition and the BFBL.In addition,the higher PPFD under canopy in row planting than in groups may be the reason for the lower BFBL on cherry trees in the row planting.

Interspecific competition

Interspecific competition impacted the DBH,height and tree stability in the row planting,but not in the group.The high number of interspecific competitors per PFCT in row planting exerted greater competitive pressure on tree growth that confirms result from a recent study on sessile oaks(Andrzejczyk et al.2015).In the row planting,the interspecific competitors mostly consisted of fast-growing semishade-tolerant to shade-tolerant tree species,such as sycamore and hornbeam,due to vigorous natural regeneration and the high mortality of planted lime trees.Those trees often overtopped the target PFCTs.Olano et al.(2009)reported a similar result that showed that shadetolerant,fast-growing broadleaf trees(e.g.,European beech)had a higher impact on the growth of target trees than did light-demanding,fast-growing species(e.g.,silver birch).Trees from those species overtopped the target PFCTs and created intense crown competition that resulted in poorer growth attributable to greater light interception and the aggressive acquisition of growing space(Pretzsch and Biber 2010).The results from previous studies in central Europe showed that fast-growing,early-and midsuccessional trees could exert strong competition to highvalue broadleaf trees in young stands(Wagner and Roeker 2000;Petersen et al.2009;von Lupke 1991;Ammer and Dingel 1997;Saha et al.2014;Liziniewicz et al.2016).The results from row planting also support previous studies on tree interactions in dense and young mixed-species stands,which showed strong competitive effects on target trees by fast-growing broadleaf trees(Leder 1996;Ammer et al.2005).In contrast to the case for row-planted trees,group-planted cherry PFCTs did not respond to interspecific competition.The main interspecific competitors for the cherry PFCTs were the lime trees planted as trainers.As a shade-tolerant and comparatively slowgrowing species,lime trees did not exert high competitive pressure on the dominant cherry PFCTs in comparison to naturally regenerated sycamore and hornbeam in row plantings.Therefore,lime trees are often regarded as better trainer tree species than beech and hornbeam(von Lupke 1998).The absence of any effect on interspecific competition on the BFBL is counterintuitive in this case,but supports the findings of Saha et al.(2014).The elongation of the branch-free bole is a dynamic process,and neighbourhood competition measurements taken in only one census may not be able to properly address this issue.

Overall,cherry PFCTs in the group planting were apparently not impacted by neighbourhood competition,whereas both intra-and interspecific competition slowed the growth of the PFCTs in row planting.This finding supports the crowding vs.shedding hypothesis of Canham et al.(2004).As a fast-growing,light-demanding species,cherry PFCTs were mainly subjected to intraspecific crowding inside groups,whereas in rows,trees were heavily impacted both by intraspecific crowding and interspecific shedding(Canham et al.2004).

Light availability

I was not surprised to find that light availability had no direct effect on the growth or quality of cherry PFCTs.For group plantings,light was measured inside the groups under the already closed canopy of the cherry and lime trees.By contrast,the forest canopy was also closed in the row plantings,but the heterogeneity of tree heights was higher in rows due to the high density of fast-growing,naturally regenerated pioneer and mid-successional species,which led to the higher availability of diffuse sunlight in the rows because it was reflected by the greater site factor and PPFD in the row planting than in the group.The higher light availability under the canopy in rows perhaps influenced the BFBL when all cherry trees were considered as discussed above but not in the case of PFCTs because the PFCTs were already in the canopy layer and belonged to the most vital trees.As a result,light availability under the canopy did not influence the growth of those trees.Past studies showed that light availability under a canopy in closed young forests mostly influences the abundance and growth rate of seedlings and saplings rather than trees that had already reached the canopy layer(Van Hees 1997;Finzi and Canham 2000;Canham et al.1990).

The treatment effect(i.e.,group vs.row planting)was not randomized when the plantation was established.Moreover,treatment was not replicated in multiple sites,and the stand conditions after the reforestation of storm damage differed between the group and the row plantings.Therefore,results from this study should be treated as a case study,and more studies should be done in the future to better understand the general performance trend of cherry trees planted in group and row plantings.

Conclusions

This study is the first on group planting of a high-value broadleaf species other than oaks and showed significant gains in regard to tree quality and growth when cherry trees areplanted in group plantingsratherthan row plantingsin the study area.Planting and pruning expenses can also be reduced due to the high success rate of future crop tree production and greaterself-pruning among treesin the group plantings.The future crop trees were not impacted by neighbourhood competition in group plantings,but were heavily affected by intra-and interspecific competition in row plantings.Early removal of competitors is recommended for row plantings,whereas this process can be delayed in groups and further reduces costs.The temporal homogeneity in radial growth is an important aspect for quality timber(e.g.,for veneer)production.Therefore,the influence of planting density and subsequent tending and thinning interventions on radial growth of the cherry PFCTs should be studied in future.The availability oflightunderthe canopy did not influence the growth of future crop trees.

The results from this study agree with those for oaks in group planting(see the recent review by Saha et al.2017).Cherries and oaks have some traits in common:both species tend to retain green branches and form brushy crowns when not pruned and grown in open fields.One could argue for a wider spacing during planting of species with these traits in forest regeneration programs and then vigorous tending and pruning.However,such a program should be juxtaposed with the intense monitoring of stand development,which is not always feasible,particularly in large forest districts.Intense and frequent management interventions would also increase long-term costs.By contrast,the benefits from natural pruning could be optimized in group planting.Nevertheless,a thorough financial assessment should be carried out to establish the costs and benefits in group and row plantings.

The group planting technique combines both active(e.g.,planting high-value tree species)and passive(e.g.,allowing natural regeneration between groups)restoration principles of forest regeneration.Therefore,in addition to creation of high-value mixed forests in wind-thrown sites,group planting can also be used for forest restoration and afforestation(Saha et al.2017).Group planting with other high-value broadleaf tree species should be established in different forest ecosystems around the world.

AcknowledgementsI thank Gabriel Burns for help in data collection and Jurgen Bauhus for his cooperation.I acknowledge the support of Gunter Hepfer(Forester,Neuried/Missenheim)for support in this research.

Abetz P,Kladtke J(2002)The target tree management system.Forstwiss Centbl 121(2):73–82

Abetz P,Ohnemus K(1999)Varification of the future-crop-treenorms(ZB-norm)for beech in a thinning experiment.Allg Forst Jagdztg 170(9):157–165

Ammer C,Dingel C(1997)Investigating the effects of strong competition by inferior tree species on growth and quality of young European oaks.Forstwiss Centbl 116(6):346–358

Ammer C,Ziegler C,Knoke T (2005)Assessing intra-and interspecfic competition in thickets of broadleaved tree species.Allg Forst Jagdztg 176(5):85–94

Andrzejczyk T,Liziniewicz M,Drozdowski S(2015)Effect of spacing on growth and quality parameters in sessile oak(Quercus petraea)stands in central Poland:results 7 years after planting.Scand J Forest Res 30(8):710–718

Ballare CL,Scopel AL,Sanchez RA(1990)Far-red radiation reflected from adjacent leaves—an early signal of competition in plant canopies.Science 247(4940):329–332

Binkley D,Campoe OC,Gspaltl M,Forrester DI(2013)Light absorption and use efficiency in forests:why patterns differ for trees and stands.For Ecol Manage 288:5–13

Brang P,Burgi A(2004)Trupppflanzung im Test.Zurcher Wald 36(5):13–16

Brang P,Combe J(2001)Extensive Verjungungsverfahren nach Lothar,1st edn.Eidgenossische Forschungsanstalt fur Wald,Schnee und Landschaft,Birmensdorf-Zurich,pp 1–17

Canham CD,Denslow JS,Platt WJ,Runkle JR,Spies TA,White PS(1990)Light regimes beneath closed canopies and tree-fall gaps in temperate and tropical forests.Can J For Res 20(5):620–631

Canham CD,LePage PT,Coates KD(2004)A neighborhood analysis of canopy tree competition:effects of shading versus crowding.Can J For Res 34(4):778–787

Chakraborty T,Saha S,Reif A(2013)Decrease in available soil water storage capacity reduces vitality of young understorey European Beeches(Fagus sylvatica L.):a case study from the Black Forest,Germany.Plants 2(4):676–698

Demolis C,Franc?ois D,Delannoy L(1997)Que sont devenues les plantations de feuillus par points d’appui?Office National des Fore?ts.Bull Tech 32:27–37

Dormann CF,Elith J,Bacher S,Buchmann C,Carl G,Carre G,Marquez JRG,Gruber B,Lafourcade B,Leitao PJ(2013)Collinearity:a review of methods to deal with it and a simulation study evaluating their performance.Ecography 36(1):27–46

Drouineau S,Laroussinie O,Birot Yves,Terrasson D,Formery T,Roman-Amat B (2000)Joint evaluation of storms,forest vulnerability and their restoration.European Forest Institute,Joensuu,p 48

Ehring A,Keller O (2006)Eichen-Trupp-Pflanzung in Baden-Wurttemberg.AFZ/Der Wald 9:491–494

Finzi AC,Canham CD(2000)Sapling growth in response to light and nitrogen availability in a southern New England forest.For Ecol Manage 131(1):153–165

Forrester DI(2014)A stand-level light interception model for horizontally and vertically heterogeneous canopies.Ecol Model 276:14–22

Gauer J,Aldinger E(2005)Waldokologische Naturraume Deutschlands—Forstliche Wuchsgebiete und Wuchsbezirke,mit Karte 1:100.000.Freiburg i.Br.:Mitteilungen des Vereins fur Forstliche Standortskunde und Forstpflanzenzuchtung,p 1–324

Gaul T,Stuber V (1996)Der Eichen-Nelder-Verbandsversuch Gohrde.Forst und Holz 51:70–75

German Weather Service(2011)Drought conditions in Europe 2011.Offenbach-Germany,German Weather Service,p 1–3.(http://rcccm.dwd.de/EN/ourservices/specialevents/drought/20110510_dryness2011_statement_for_WMO_en.html),Accessed 2 Nov 2016

Gockel H(1995)Die Trupp-Pflanzung,Ein neues Pflanzschema zur Begrundung von Eichenbestanden.Forst und Holz 50:570–575

Gockel H,Rock J,Schulte A(2001)Aufforsten mit Eichen-Trupppflanzungen.AFZ/Der Wald 5:223–226

Gussone H,Richter A(1994)Eichen-Nester—Zweiter Bericht der Versuche mit Nesterpflanzungen in Norddeutschland.Forst-und Holzwirt 49(11):300–304

Hegyi F(1974)A simulation model for managing Jack-pine stands.Stockholm,Sweden.Royal Coll For 30:74–91

Hein S(2009)Modelling natural pruning of common ash,sycamore and wild cherry.In:Spiecker H,Hein S,Makkonen-Spiecker K,Thies M(eds),Valuable broadleaved forests in Europe.European Forest Institute Research Report 22.Brill,Leiden,pp 103–122

Henriksson J(2001)Differential shading of branches or whole trees:survival,growth,and reproduction.Oecologia 126(4):482–486

Kint V,Hein S,Campioli M,Muys B(2010)Modelling self-pruning and branch attributes for young Quercus robur L.and Fagus sylvatica L.trees.For Ecol Manage 260(11):2023–2034

Knowe SA,Hibbs DE(1996)Stand structure and dynamics of young red alder as affected by planting density.For Ecol Manage 82(1–3):69–85

Kohler M,Sohn J,Nagele G,Bauhus J(2010)Can drought tolerance of Norway spruce(Picea abies(L.)Karst.)be increased through thinning?Eur J For Res 129(6):1109–1118

Kraft G (1884)Beitrage zur lehre von den Durchforstungen,Schlagstellungen und Lichtungshieben.Klindworth’s Verlag,Hannover,pp 1–156

Kuehne C,Kublin E,Pyttel P,Bauhus J(2013)Growth and form of Quercus robur and Fraxinus excelsior respond distinctly different to initial growing space:results from 24-year-old Nelder experiments.J For Res 24(1):1–14

Kuijper DPJ,Cromsigt J,Churski M,Adam B,Jedrzejewska B,Jedrzejewski W(2009)Do ungulates preferentially feed in forest gaps in European temperate forest? For Ecol Manage 258(7):1528–1535

Leder B(1996)Weichlaubholzer im Eichen-und Buchen-jungbestanden.Forst und Holz 51(10):340–344

Liziniewicz M,Andrzejczyk T,Drozdowski S(2016)The effect of birch removal on growth and quality of pedunculate oak in a 21-year-old mixed stand established by row planting.For Ecol Manage 364:165–172

McCullagh P,Nelder JA(1989)Generalized linear models,2nd edn.Chapman and Hall,London,pp 1–532

Nicolini E,Chanson B,Bonne F(2001)Stem growth and epicormic branch formation in understorey beech trees(Fagus sylvatica L.).Ann Bot 87(6):737–750

Olano JM,Laskurain NA,Escudero A,De La Cruz M(2009)Why and where do adult trees die in a young secondary temperate forest?The role of neighbourhood.Ann For Sci 66(1):105

Panayotov M,Kulakowski D,Tsvetanov N,Krumm F,Barbeito I,Bebi P (2016)Climate extremes during high competition contribute to mortality in unmanaged self-thinning Norway spruce stands in Bulgaria.For Ecol Manage 369:74–88

Petersen R(2007)Eichen-Trupp-Pflanzung—erste Ergebnisse einer Versuchsflache im NFA Neuhaus.Forst und Holz 62(3):19–25

Petersen R,Schuller S,Ammer C(2009)Early growth of planted pedunculate oak(Quercus petraea)in response to varying competition by birch(Betula pendula)over 8 years.Forstarchiv 80:208–214

Pretzsch H(2009)Forest dynamics,growth,and yield.Springer,Berlin,pp 1–664

Pretzsch H,Biber P(2010)Size-symmetric versus size-asymmetric competition and growth partitioning among trees in forest stands along an ecological gradient in central Europe.Can J For Res 40(2):370–384

Rebetez M,Mayer H,Dupont O,Schindler D,Gartner K,JrP Kropp,Menzel A(2006)Heat and drought 2003 in Europe:a climate synthesis.Ann For Sci 63(6):569–577

Rock J,Gockel H,Schulte A(2003)Vegetationsdiversitat in Eichen-Jungwuchsen aus unterschiedlichen Pflanzschemata.Beitr Forstwirtsch u Landsch okol 37:11–17

Rock J,Puettmann KJ,Gockel HA,Schulte A(2004)Spatial aspects of the influence of silver birch(Betula pendula L.)on growth and quality of young oaks(Quercus spp.)in central Germany.Forestry 77(3):235–247

Ruhm W(1997)Alternative—Kulturbegrundung von Eichenmischwald.osterreichische Forstzeitung 108(7):29

Saha S (2012)Development of tree quality,productivity,and diversity in oak(Quercus robur and Q.petraea)stands established by cluster planting.Ph.D.Dissertation,Freiburg:University of Freiburg,p 1–130,(https://freidok.uni-freiburg.de/dnb/download/9960),Accessed 3 July 2017

Saha S,Kuehne C,Kohnle U,Brang P,Ehring A,Geisel J,Leder B,Muth M,Petersen R,Peter J,Ruhm W,Bauhus J(2012)Growth and quality of young oaks(Quercus robur and Q.petraea)grown in cluster plantings in central Europe:a weighted meta-analysis.For Ecol Manage 283:106–118

Saha S,Kuehne C,Bauhus J(2013)Tree species richness and stand productivity in low-density cluster plantings with oaks(Quercus robur L.and Q.petraea (Mattuschka)Liebl.).Forests 4(3):650–665

Saha S,Kuehne C,Bauhus J(2014)Intra-and interspecific competition differently influence growth and stem quality of young oaks(Quercus robur L.and Quercus petraea(Mattuschka)Liebl.).Ann For Sci 71(3):381–393

Saha S,Kuehne C,Bauhus J(2017)Lessons learned from oak cluster planting trials in central Europe.Can J For Res 47:139–148

Savill PS(2013)The silviculture of trees used in British forestry,2nd edn.CAB International,Oxford,pp 1–280

Scherrer D,Bader MKF,Korner C(2011)Drought-sensitivity ranking of deciduous tree species based on thermal imaging of forest canopies.Agric For Meteorol 151(12):1632–1640

Schmidt M,Hanewinkel M,Kandler G,Kublin E,Kohnle U(2010)An inventory-based approach for modeling single-tree storm damage—experiences with the winter storm of 1999 in southwestern Germany.Can J For Res 40(8):1636–1652

Schraml U,Volz KR(2009)Do species matter?Valuable broadleaves as an object of public perception and policy—European Forest Institute Report 22.Leiden:S.Brill.Publishers,p 213–236

Schutz JP,Gotz M,Schmid W,Mandallaz D(2006)Vulnerability of spruce(Picea abies)and beech(Fagus sylvatica)forest stands to storms and consequences for silviculture.Eur J For Res 125(3):291–302

Skiadaresis G,Saha S,Bauhus J(2016)Oak group planting produces a higher number of future crop trees,with better spatial distribution,than row planting.Forests 7:289–304

Sokal RR,Rolhf FJ(1995)Biometry,3rd edn.W.H.Freeman,New York,pp 1–880

Spellmann H,Baderschneider A(1988)Erste Auswertung eines Traubeneichen-Pflanzverbands-und Sortimentsversuches im Forstamt Hardegsen/Solling.Forst und Holz 19:447–450

Spiecker H(1991)Controlling the diameter growth and the natural pruning of Sessile and Pedunculate oaks(Quercus petraea(Matt.)Liebl.and Quercus robur L.).Schriftenreihe der Landesforstverwaltung Baden-Wurttemberg 72:1–135

Thies M,Hein S,Spiecker H(2009)Results of a questionnaire on management of valuable broadleaved forests in Europe—European Forest Institute Report 22.Leiden:S.Brill.Publishers,p 27–42

United Nations Economic Commission for Europe(2000)Effects of the december 1999 storms on European timber markets.Geneva:Food and Agricultural Association,p 1–17,(https://www.unece.org/fileadmin/DAM/timber/docs/rev-00/02.pdf),Accessed 3 July 2017

United States Forest Service(2011)Field guides,methods and procedures.Phase 2 field guide—crowns:measurements and sampling-version 5.1,Washington DC:The United States Forest Service,p 1–310.(https://www.fia.fs.fed.us/library/field-guidesmethods-proc/docs/Complete%20FG%20Document/core_ver_5-1_10_2011.pdf),Accessed 3 July 2017

van Hees AFM(1997)Growth and morphology of pedunculate oak(Quercus robur L)and beech(Fagus sylvatica L)seedlings in relation to shading and drought.Ann For Sci 54(1):9–18

Loewe V,Gonzalez M,Balzarini M(2013)Wild cherry tree(Prunus avium L.)growth in pure and mixed plantations in South America.For Ecol Manage 306:31–41

von Lupke B(1991)Einfluss der Konkurrenz von Weichlaubholz auf das Wachstum junger Traubeneichen.Forst und Holz 46:166–171

von Lupke B(1998)Silvicultural methods of oak regeneration with special respect to shade tolerant mixed species.For Ecol Manage 106(1):19–26

Wagner S,Roeker B(2000)Birkenanflug in Stieleichenkulturen.Untersuchungen zur Dynamik der Konkurrenz uber 5 Vegetationsperioden.Forst und Holz 55:18–22

Waring RH,Schlesinger WH(1985)Forest ecosystems:concepts and management.Academic Press Inc,Orlando-Florida,pp 1–340