Introduction

The conversion of Norway spruce monocultures into more structured mixed stands, especially on sites naturally dominated by broadleaved, has become the main focus of silvicultural practices in many Central European countries over the last two decades, particularly in public forests (Otto 1995). European beech (Fagus sylvatica L.), the most frequently used species in conversion practice, leads to higher resilience to natural disturbances and higher adaptability to changes in site conditions (Pretzsch 2003), and also improves soil properties and biodiversity (Ammer et al. 2008). Other native and non-native species with high adaptative and productive capability such as Douglas-fir (Pseudotsuga menziesii Mirb.) can also increase the resilience of future mixed forests, e.g., by increasing of risk distribution in regard to climate change (Lüpke 2004, 2009). A widely used method to transform pure Norway spruce stands consists of target diameter harvesting followed by planting beech and/or other species beneath the canopy of mature spruce trees (“conversion under continuous cover scheme”) (Lüpke et al. 2004; Ammer et al. 2008). But in stands with high windthrow risks (e.g., on waterlogged or shallow soils), continuous cover methods are not applicable and a faster conversion using small-scale clear cuts followed by planting of beech and/or other species can be necessary (“plain conversion”) (Lüpke et al. 2004).

In tree plantations on clear cut areas, natural vegetation (herbs, shrubs and tree species) invades faster and grows more rapidly than on areas shaded by canopy trees, thereby often threatening a successful regeneration by hampering growth and survival of planted saplings (Collet et al. 1996; Löf 2000; Nilsson and Örlander 2003; Löf and Welander 2004; Rose and Rosner 2005; Balandier et al. 2006; Parker et al. 2009). The young saplings compete with the ground vegetation mainly for water (Cole and Newton 1986; Zutter et al. 1999; Löf 2000), in particular during dry growing seasons (Zutter et al. 1986; Nilsson and Örlander 1995), and for nutrients (Zutter et al. 1999; Löf 2000; Provendier and Balandier 2008). The latter factor can be exacerbated in dry periods because a reduced water uptake limits the nutrient uptake further (Örlander et al. 1996; Coomes and Grubb 2000; Löf 2000; Fotelli et al. 2001, 2005; Löf and Welander 2004).

Several studies have shown that various ground vegetation types (herbs, shrubs or tree species) compete differently for resources with saplings (Zutter et al. 1986; Coll et al. 2003; Rose and Rosner 2005; Parker et al. 2009). On the other hand, plant species respond differently to limited resources, which influence their capacity to compete with surrounding vegetation (Kolb and Steiner 1990; Löf 2000; Löf and Welander 2004). Reactions of tree species to competition of ground vegetation can include an increased biomass allocation to roots, thereby raising the potential for water and nutrient uptake, and a smaller leaf area contributing to less transpiration (Valladares and Niinemets 2008). In numerous studies, morphological traits like specific leaf area (SLA) and specific root length (SRL) were proved to be indicators of plant strategy to limited resources (Curt and Prevosto 2003; Prevosto and Balandier 2007; Provendier and Balandier 2008). Nevertheless, little research on planted tree sapling interactions in forests has been focused “on mechanisms of competition and competitive ability among various functional groups of herbaceous plants” (Provendier and Balandier 2008).

In temperate zones, recently clear cut stands often present two types of ground vegetation that rapidly establish and compete with planted tree saplings: gramineous and small shrubs. The two groups differ considerably in morphology and growth dynamics (Balandier et al. 2006). Under full light conditions, gramineous vegetation rapidly develops a high dense, fasciculate root system in superficial soil horizons in early spring (Coll et al. 2003), while small shrubs, like Rubus fruticosus and R. idaeus, grow over a longer period (Fotelli et al. 2001, 2002). As a morphological response of Rubus fruticosus to water stress, Fotelli et al. (2001) found a tendency to increased root growth, leading to a greater ability to exploit soil water compared to beech saplings. Rose and Rosner (2005) pointed out that the impact of herbaceous competition on seedling establishment and growth is typically more severe in earlier stages of the development, and of woody competition in later stages. This is confirmed by Harrington et al. (1995) who found that Douglas-fir basal-area growth was significantly restricted by herbaceous cover in years 2 and 3, and by shrub cover in year 3 through 5. Most investigations about herbaceous competition on beech saplings were carried out in the first 2 years after planting (Löf 2000; Coll et al. 2004; Löf and Welander 2004; Provendier and Balandier 2008) when herbaceous competitors are most important. How beech saplings react to ground vegetation competition over 3 and more years after planting and whether a response is dependent on the type of vegetation is not clear. Information about the response of Douglas-fir to different ground vegetation types in Europe is actually totally lacking. Although vegetation control over several years after planting is usually not an option in present-day practical forestry (Löf and Welander 2004), for the establishment of a new forest it may be of interest to know whether certain tree species are better competitors than others.

The objective of this study was to evaluate the reaction of young beech and Douglas-fir saplings on competition of two ground vegetation types, gramineous and small shrubs, under natural conditions. We used a competition exclusion experimental design with application of a herbicide. We tested the following hypotheses:

  1. 1.

    Removal of ground vegetation has a positive effect on nutrient availability and growth of tree saplings which is more pronounced in small shrubs than in a gramineous cover, and more distinct in a drier growing season than in a season with normal rainfall.

  2. 2.

    Beech and Douglas-fir saplings react to weed control by morphological modifications (lower SLA, SRL and fine root investment).

  3. 3.

    Content of main nutritional elements in leaves/needles of the tree saplings is negatively related to total dry mass of the surrounding vegetation and its content of main nutritional elements.

Materials and methods

Study sites and treatments

The study was carried out on two sites in the Solling Mountains (Lower Saxony, Germany, 51°47′N, 9°37′E). Both sites are characterized by well-drained dystric cambisol (podzolic brown earth). The Solling climate is classified as humid and moderately sub-continental. For the first site—Neuhaus at 500 m a.s.l.—the following long-term mean values are given: 6.5°C annual temperature, 1,050 mm annual precipitation, of which 470 mm during the growing season; for the second site—Otterbach at 300 m a.s.l.—the respective values are: 7.5°C annual temperature, 900 mm annual precipitation, of which 420 mm during the growing season. During the observation period of this study (2007 and 2008), the actual measured precipitation deviated from those long-term means and reached the following values during the growing season (measured by the Northwest German Forest Research Station): at Neuhaus: 801 mm in 2007 and 216 mm in 2008; at Otterbach: 770 mm in 2007 and 375 mm in 2008. Particularly at the Neuhaus site, 2008 was exceptionally dry, while 2007 was a fairly wet year on both sites, especially during the growing season.

At both sites, two clear cutting plots (1 ha each) in ca. 95-year-old pure Norway spruce stands were established in autumn 2003 by the Northwest German Forest Research Station in Göttingen, in which 2-year-old European beech and 3-year-old Douglas-fir saplings were planted in spring 2004. The research area was fenced against game browsing.

Two vegetation types were distinguished on each clear cut: (1) gramineous, with the main species Agrostis capillaries, Calamagrostis epigejos, Deschampsia flexuosa, Epilobium angustifolium, Holcus mollis, and Juncus effusus, and (2) small shrubs, dominated by Rubus fruticosus and Rubus idaeus (Table 1). Species were regarded as dominant when they occupied at least 75% of the total vegetation coverage of each sample plot.

Table 1 Description of the two ground vegetation types based on a survey on 1-m2 circular plots around each control sapling (dominant species, cover, average shoot length and total dry mass per m2) at the start of the experiment in July 2007 (mean values ± SE)
Full size table

In Otterbach, we sampled for each vegetation type 56–60 beech and 50–58 Douglas-fir saplings. In Neuhaus, suitable plots for the small shrubs type were scarce because both Rubus sp. were much less frequent. Therefore, we had to accept only 28 Douglas-fir and 48 beech sample saplings for the small shrubs type whereas for the gramineous type we found 50 Douglas-fir and 62 beech sample saplings. Unfortunately, during the run of the experiment, 11 beech and 21 Douglas-fir saplings had to be excluded from further observations, mainly because wild boars broke into the experimental sites and grubbed them out, after storms partly destroyed the fences in January 2007 and February 2008 (for more details, see Table 2). No damage by mice was recorded during the 2 years of the study.

Table 2 Diameter and height of sampled saplings before start of the experiment in autumn 2006
Full size table

Within both sites and within both vegetation types, no significant differences in diameter and height of the sampled saplings could be detected between herbicide-treated and control plots at the start of the experiment. However, within the small shrubs type, saplings of both species had significantly greater heights (at both sites) and smaller diameters (only at Otterbach) compared with the gramineous type (Table 2).

As significant differences in diameter and height between beech and Douglas-fir occurred within each site and within each vegetation type, we did not consider species as a main factor in a multifactorial analysis of variance (ANOVA).

Half the sampled saplings were randomly selected and ground vegetation on a circular plot of about 1 m2 (radius = 0.6 m) around each sapling was treated with the herbicide Glyphosphate (GLYFOS®, 2.5% concentration in water) by direct foliar application, three times in 2007 (end of May, beginning of July and end of September) and once in 2008 (beginning of July). During herbicide spraying, the saplings were protected by a plastic bowl. The treatments completely destroyed the surrounding ground vegetation without damaging the tree saplings.

Ground vegetation assessments

The following data were collected on 1-m2 circular plots around each control sample sapling for each species of the ground vegetation in mid-July 2007: cover as % of total surface area (by visual estimation) and length of 5–10 individuals per species in cm by measuring with a ruler. After harvest of half the sample saplings in autumn 2007, we repeated this vegetation assessment in July 2008, getting roughly similar mean values of cover and shoot length. This is in line with a broad-scale vegetation assessment at the same sites and in the same years published by Heinrichs and Schmidt (2009). We estimated the above-ground dry mass and the concentration of main nutritional elements of each ground vegetation species using the model PhytoCalc (Bolte et al. 2002), amended by correction factors of Heinrichs et al. (2010). PhytoCalc allows a non-destructive estimation of dry mass and nutrient content of ground vegetation by using the relationship between species biomass, cover and mean shoot length. It was initially developed for closed forests, but the correction factor of Heinrichs et al. (2010) made it applicable also to clear cut areas. Finally, total dry mass and total content of nutrients per square meter were obtained as sums across the species.

Saplings growth and biomass measurements

All sample saplings were manually excavated and collected from end-September to mid-October, half of them in 2007 (120 beeches and 81 Douglas-firs) and half in 2008 (96 beeches and 82 Douglas-firs), and divided into leaves/needles, branches, stems, fine (diameter < 2 mm) and coarse (diameter > 2 mm) roots. For every sapling, the following data were recorded: length of the last annual terminal shoot, and diameter at 1 cm above ground, all to the nearest millimetre. A stem disc was taken from 1 cm above ground for measuring the width of the last annual ring in two perpendicular directions. For further analyses, arithmetic means of the two measurements were used.

Wood components (branches, stems and roots) were dried at 65°C to a constant weight for 5 days and non-wood components (leaves/needles) for 3 days, and afterwards weighed to the nearest 0.1 g.

Foliage area

A subsample of 50 randomly chosen leaves per beech sapling (when the number was less, we took all leaves) was used for leaf area measurements with the LI-3100 Area Meter (LI-COR). For each Douglas-fir sapling, two subsamples of 100–120 randomly chosen needles of current year and previous year needles were scanned and processed with WinFOLIA (Regents Instruments, Quebec, Canada) to estimate one-sided leaf area. For the rest of the leaves/needles, only dry weights were determined. Using the ratio of leaf weight to leaf area of the sub-samples, we calculated total leaf/needle area for each sample sapling.

Fine root measurements

Fine roots (diameter < 2 mm) of sampled saplings were separated from coarse roots, watered, and sorted according to vitality (live, dead) using the criteria of Murach (1984). Only live fine roots were included in the following analyses. After scanning, they were processed with WinRHIZO (Regents Instruments) to obtain fine root length. Finally, roots were dried and weighed. Based on these measures, specific root length (SRL, ratio of fine root length to dry weight, m g−1) was determined. Additionally, fine root nitrogen concentration (N mg g−1) was measured in 2008 in the same way as the foliar nitrogen content (see the following paragraph).

Foliar nutrient content

Between August 13 and 20 in 2007 and in 2008, we sampled 7–9 beech leaves and 100–150 Douglas-fir needles per sample sapling to determine foliar nutrient content. The leaves/needles were randomly chosen from the last terminal shoot. Whenever the amount of leaves/needles did not suffice, we completed it with leaves/needles of current year branches at the top of the crown. Foliar nitrogen concentration (N mg g−1) was measured with an elemental analyzer (Model 150; Carbo Erba, Italy), foliar K, Mg and Ca concentrations (mg g−1) photometrically with a flame atomic absorption spectrometer (SpectrAA 300; Varian, USA), and foliar P concentration colorimetrically with the continuous-flow method (Skalar, The Netherlands).

Data analysis

Differences between treated and control saplings in initial diameter and height, annual diameter and length growth, length to diameter ratio, SLA and SRL were tested using ANOVA with treatment as factor, both per year (2007 and 2008), site (Neuhaus and Otterbach), and vegetation type (gramineous and small shrubs). Multifactorial ANOVA was used for testing effects of vegetation type (gramineous, small shrubs), treatment (control, herbicide), site (Neuhaus, Otterbach) and year (2007, 2008) on foliar nutrient concentration (N, P, K, Mg, Ca, mg g−1), fine root nitrogen concentration (N, mg g−1), diameter and length increment, and relative fine root biomass (as % of total plant biomass for 2008) by the two species (beech and Douglas-fir). In case the ANOVA yielded a significant treatment effect, we examined the differences between mean values by Scheffé post hoc test. The Pearson correlation coefficient was used to characterize the relationship between leaves/needles nutrient content (N, P, K, Mg and Ca, g) of total leaves/needles mass and surrounding vegetation dry mass (g m−2) and nutrient content of surrounding vegetation (N, P, K, Mg and Ca, g m−2), respectively. All data analyses were performed using Statistica 7.1 (StatSoft 2005).

Results

Ground vegetation

On gramineous plots, the most representative species was Agrostis capillaris with an average cover of 15% in Neuhaus and 7% in Otterbach, followed by Calamagrostis epigejos and Deschampsia flexuosa. Epilobium angustifolium, Holcus mollis and Juncus effusus were more developed in Neuhaus (Table 1).

On the small shrubs plots Rubus fruticosus occupied 27% of the area in Neuhaus and 19% in Otterbach. The respective figures for Rubus idaeus were 21% in Neuhaus and 53% in Otterbach. In combination with greater heights of both species on the latter site, the small shrubs competition was more severe in Otterbach.

Mean values of above-ground vegetation dry mass estimated by Phytocalc ranged from 206 g m−2 of the gramineous type at Neuhaus to 428 g m−2 of the small shrubs type at Otterbach (Table 1). Dry mass of the small shrubs type was significantly heavier in Otterbach than in Neuhaus (different small letters in Table 1), and within each site, superior to the gramineous type (different capital letters in Table 1).

Diameter and length growth response to weed control

For both species, the multifactorial ANOVA proved the factor “Treatment” to be significant for annual diameter growth (Table 3). Saplings on herbicide treated plots achieved higher diameter increments in almost all cases (Fig. 1a, b), but significantly only in Otterbach for beech in the first year (2007) and for Douglas-fir in both years after removal of small shrubs competition, and for beech in the second year after release from gramineous competition (Fig. 1a, b). In no case was a significant influence of ground vegetation removal on length increment detectable (Table 3).

Table 3 Multifactorial ANOVA results for testing effects of vegetation type (gramineous, small shrubs), treatment (control, with herbicide), site (Neuhaus, Otterbach) and year (2007, 2008) on annual diameter increment (mm), annual length increment (mm) and length to diameter ratio by species (Beech and Douglas-fir)
Full size table
Fig. 1
figure 1

Response of annual diameter increment (a, b), annual length increment (c, d) and length to diameter ratio (e, f) to herbicide treatment. Filled symbols treated saplings, open symbols control saplings, Gr gramineous, Ss small shrubs. Circles depict data for 2007, and squares data for 2008. Box-plots represent the mean values with ±SE. **Significant differences (P < 0.05) between herbicide and control treatments, tested by ANOVA and Scheffé post hoc test

Full size image

All four factors (vegetation type, treatment, year, and site) had a significant influence on length to diameter ratio in both species (Table 3). In beech saplings, it decreased after release from small shrubs on both sites in 2008 and also in Otterbach in 2007. Douglas-fir showed a similar reduction, being significant in Otterbach in both years after release from small shrubs and in Neuhaus after release from gramineous vegetation in 2007.

Effect of weed control on leaf and fine root morphology (SLA and SRL) and on biomass allocation to fine roots

Weed control did not affect the SLA of Douglas-fir and beech saplings in the first year. In the second year (2008), there was still no response of Douglas-fir, but beech reduced its SLA in every case, though being significantly only with small shrubs in Otterbach (Fig. 2a).

Fig. 2
figure 2

Effect of weed control on specific leaf area (a) and specific fine root length (b) after 2 years of herbicide application (2008). Filled symbols treated saplings, open symbols control saplings, Gr gramineous, Ss small shrubs. Diamonds depict beech saplings, triangles Douglas-fir saplings. Box-plots represent the mean values with ±SE. **Significant differences (P < 0.05) between herbicide and control treatments, tested by ANOVA and Scheffé post hoc test

Full size image

As with SLA, in the first year, no clear trend could be found for specific fine root length (SRL). But 2 years removal of small shrubs led to smaller SRL in both species and on both sites (significant for beech at Neuhaus and Douglas-fir at Otterbach), whereas the removal of gramineous vegetation still had no consistent effect (Fig. 2b).

No significant effect of weed control on relative fine root fraction in percent of total plant biomass could be detected in either year. Only a weak, but consistent, trend appeared as a slight reduction of investment in fine root biomass after removal of gramineous vegetation in both species and on both sites.

Effect of weed control on foliar nutrient concentration and fine root nitrogen concentration

Beech foliar nitrogen concentration showed a significant response to year, to interactions of treatment and vegetation type (TR × VT; i.e., higher nitrogen concentrations only after removal of surrounding gramineous vegetation), and of treatment and site (TR × site; i.e., higher nitrogen concentrations only at Neuhaus after vegetation release) (Table 4). Douglas-fir foliar nitrogen concentration was responsive to site and year, but not to treatment. Foliar P and K concentrations in both species were not affected by removal of either vegetation type. Beech foliar P concentration was significantly related to year and site, but in Douglas-fir no significant relationship to any of the tested factors could be detected. Foliar K concentration was significantly related to year and site for both species.

Table 4 Multifactorial ANOVA results (P values) for testing effects of vegetation type (VT, gramineous, small shrubs), treatment (TR, control, with herbicide), site (site—Neuhaus, Otterbach) and year (year—2007 and 2008) on foliar nutrient concentrations (N, P, K, Mg, Ca, mg g−1) and on fine root nitrogen concentration (N, mg g−1) for beech and Douglas-fir
Full size table

Beech foliar Mg and Ca concentrations proved to be very responsive, including a significant response to herbicide application (TR), but surprisingly this response was negative at Otterbach in 2008: here, Mg concentration reached 1.28 mg g−1 on control gramineous plots compared to 0.88 mg g−1 on treated plots, and 1.06 mg g−1 on control small shrubs plots versus 0.78 mg g−1 on treated plots; foliar Ca concentration decreased from 9.56 to 6.69 mg g−1 as an effect of gramineous control and from 6.54 to 5.02 with small shrubs. Douglas-fir was less responsive with only a few elements being significantly different between sites and years (e.g., higher Mg concentrations at Neuhaus).

Beech possessed significantly higher fine root nitrogen concentrations on treated than on control plots with both vegetation types, e.g., at Neuhaus, the N concentration rose from 8.1 to 9.2 mg g−1 after removal of gramineous vegetation, and from 9.5 to 10.7 mg g−1 after removal of small shrubs vegetation. Douglas-fir N concentration increased only after removal of small shrubs at Neuhaus (from 8.1 to 10.1 mg g−1).

Ground vegetation dry mass and content of nutritional elements as explanatory variables for leaves/needles nutrient content of the saplings

So far we have only used the vegetation type (in two grades: gramineous and small shrubs) as the explanatory variable. To get a step further in understanding the competitive influence of ground vegetation on the nutrient uptake of the tree saplings, it seemed to be helpful to drop the type of vegetation, to merge the data of the vegetation types and to build two new ground vegetation variables: (1) above-ground dry mass (g m−2), and (2) content of main nutritional elements (N, P, K, Mg and Ca, g m−2). We included only control plots, determined the variables separately for each year, site, and tree species, and calculated Pearson correlation coefficients between these variables and the content of main nutritional elements of the total mass of leaves/needles of the saplings.

In Neuhaus, none of the main foliar nutrients was significantly correlated with dry mass or with ground vegetation nutrient content (data not shown). In Otterbach, all correlation coefficients were negative—many of them significantly—indicating a more intense ground vegetation competition than in Neuhaus (Table 5). This was more pronounced in 2008 which points to a higher competitive effect of the ground vegetation on nutrient supply of the tree saplings during the drier growing season in 2008 as compared to 2007 with normal rainfall. Douglas-fir foliar nutrient content was more strongly negatively correlated with the ground vegetation variables than beech foliar content indicating a higher sensitivity of Douglas-fir saplings to competition for nutrients.

Table 5 Correlation coefficients between leaves/needles main nutrient content per sapling (N, P, K, Mg and Ca, g), ground vegetation dry mass (VDM, g m−2), and main nutrient content of ground vegetation (N, P, K, Mg and Ca, g m−2), at the site Otterbach
Full size table

Discussion

Effect of weed control on growth and functional traits of beech and Douglas-fir saplings

The removal of ground vegetation competition led to augmented annual diameter growth of planted saplings (Fig. 1a, b). The enhancement was particularly noticeable at Otterbach, showing a stronger intensity of competition than at Neuhaus which is in line with about 40% greater values of competitive vegetation dry mass there (Table 2). Diameter growth of saplings not subject to weed control was strongly reduced which is a frequently reported observation (e.g., Zutter et al. 1986; Coll et al. 2004). Particularly, saplings surrounded by small shrubs significantly decreased their diameter growth in our study which accords with an investigation of Harrington et al. (1995) who found that Douglas-fir basal-area growth was significantly limited by shrub cover in years 3 through 5 after planting. But in our study, gramineous vegetation also had a considerable effect on diameter growth as beech showed a significant increase after release from competition in 2008, a year with a considerably lower precipitation during the growing season. An exacerbating competition for water between field vegetation and tree seedlings under drier conditions and thereby an increasing impact on growth have been observed in previous studies (Zutter et al. 1986; Nilsson and Örlander 1995).

Although in our study height increment response to weed control was not significant, a general tendency appeared, consisting in a higher length increment after removal of gramineous vegetation and an opposite behaviour on small shrubs plots (Fig. 1c, d). This pattern has been reported in other studies on Douglas-fir, beech or further tree species saplings, in which the removal of herbaceous vegetation increased height growth (Coll et al. 2004; Rose and Rosner 2005; Parker et al. 2009), whereas the control of woody competitors decreased it (Wagner and Radosevich 1991; Parker et al. 2009).

Since sapling diameter growth was more responsive than height growth to vegetation control, the length to diameter ratio in our study decreased after release from vegetation competition, as in Zutter et al. (1986). Particularly at Otterbach this effect proved to be significant for both species and years on small shrubs plots. This result accords with results of an investigation of Prevosto and Balandier (2007) who found with beech seedlings a weak but significant increase of height to diameter ratio with increasing competition by neighbouring Scots pine and birch seedlings. It is well known that competition reduces tree species diameter growth more than height growth, reflecting the general hierarchy of photosynthate allocation which attributes the least priority to diameter growth of stem and branches under increasing shortage of resources (Oliver and Larson 1996; Röhrig et al. 2006).

A lower aboveground sapling growth as a consequence of vegetation competition can partly be provoked by lower biomass partitioning to leaves and higher partitioning to roots, particularly on soils with limited nutrient and water supply (Lambert and Poorter 1992; Valladares and Niinemets 2008). Since fine root mass is more important for uptake of water and nutrients than total root mass, and generally partitioning to roots is greater in drier conditions (Persson 1983; Coomes and Grubb 2000), we expected biomass partitioning to fine roots especially in the fairly dry growing season of 2008 to be lower as an effect of weed control. Instead, our results did not support this expectation. We just found a slight decrease of this trait under gramineous competition at both sites and in both species. The lack of a significant decrease in fine root biomass partitioning on sites without ground vegetation might be explained by the fact that fine root biomass alone may not always accurately indicate the capacity of roots for water and nutrient uptake (Lehmann 2003). Curt et al. (2005) hypothesized that beech response to vegetation competition consists in modifying its morphology at the fine-root level in order to raise its efficiency of soil resources uptake. The main strategy of plants to increase their nutrient uptake efficiency and adapt to limited soil conditions is to develop longer and thinner roots (i.e. higher specific fine root length, SRL) (Ostonen et al. 2007). SRL is the main trait often used to characterize the soil exploitation strategy of species and to describe the fine root morphological acclimatization to variations in soil water and nutrient regimes (Fitter et al. 1991; Bauhus and Messier 1999; Curt and Prevosto 2003; Provendier and Balandier 2008). In our study, SRL decreased with improved soil conditions caused by small shrubs control in both species, yet significantly only at Neuhaus for beech and at Otterbach for Douglas-fir. Although we were also expecting a lower SRL as an effect of gramineous control, we did not observe significant changes in SRL. Instead, a trend for beech on both sites and Douglas-fir at Otterbach appeared, having an even lower SRL in gramineous vegetation. Provendier and Balandier (2008) also found similar results for beech saplings. A possible explanation could be that tree root elongation is restrained by the very dense structure of the grass root system in the upper soil of 0–20 cm depth (Coll et al. 2003). This is in accordance with the investigation of Coll et al. (2004), in which beech saplings with grass competition had poorly developed root systems, thus limiting their ability to absorb water and nutrients.

Whereas SLA is often used to characterize the morphological adaptation of trees to light changes, it could also describe modifications in water and nutrient supply (Prevosto and Balandier 2007; Provendier and Balandier 2008). Beech saplings without competing vegetation showed lower SLA as saplings with grass (Provendier and Balandier 2008) or woody competition (Prevosto and Balandier 2007). Similar results were found by Zutter et al. (1986) for loblolly pine seedlings.

Likewise, beech SLA in our study decreased significantly after removal of small shrubs in the drier year 2008 at Otterbach with its denser ground vegetation cover and thereby stronger competition than at Neuhaus. The lack of Douglas-fir response in SLA to weed control was consistent with the very small SLA response across a light gradient from target diameter cutting to clear cutting in the same experiment (Petritan et al. 2010). The significant response of beech under small shrubs competition could be explained by a generally higher SLA plasticity of this species to light variations (Petritan et al. 2010). Furthermore, our beech saplings were about 60 cm shorter than Douglas-fir saplings, implying more stress by shrub presence. However, beech saplings with weed control received only slightly more light according to our measurements at Otterbach: the removal of small shrubs led to a negligible increase in light availability on top of the beech saplings from 83 to 90% TSF (Total Site Factor as photosynthetically active radiation in percent of above canopy radiation, estimated by hemispherical photos; unpublished data). But we can assume that the weed control considerably increased the light availability of the deeper branches. Thus, it can be concluded that competition for light could be an additional cause for increased SLA values in beech under competitive stress, whereas reduced uptake of water and nutrients might play a leading role in situations with sufficient light availability.

Effect of ground vegetation on saplings nutrient concentration

Many studies confirm that the exclusion of ground vegetation competition led to higher nutrient availability, particularly in nitrogen content (Morris et al. 1993; Löf 2000; Coll et al. 2004; Löf and Welander 2004; Provendier and Balandier 2008). Largely, our results agree with these findings. Significant influences of year and site on most of the main foliar nutritional elements (Table 4) could be explained by differences between the study years in growing season precipitation, with 2008 having been drier than 2007, and regarding the sites, by a stronger vegetation competition at Otterbach than at Neuhaus. Regarding the release treatment, beech N foliar concentration responded to an interaction between weed control and vegetation type as it significantly increased after removal of gramineous competition at Neuhaus. Also, fine root nitrogen concentration rose at Neuhaus in both species after removal of small shrubs competition, and in beech after removal of gramineous vegetation. Contrary to our expectations, beech foliar Mg and Ca concentration significantly decreased at Otterbach after release from ground vegetation of both types. Zutter et al. (1999) found similar results in loblolly pine plantations where they observed at age 6 lower concentrations of Ca and Mg as a result of weed control. They explained this in part by the large increases in crown foliage mass as a response to weed control and a concomitant dilution of nutrient concentrations. This dilution effect could explain our results with beech saplings as they considerably increased their foliage mass after release from ground vegetation (at Otterbach from 36 to 50 g per plant under small shrubs and from 44 to 66 g under gramineous cover).

Conclusions

The present study provides insight into the impact of gramineous and small shrubs competition on growth of planted beech and Douglas-fir saplings on sites with moderate nutrient and good water supply. The results show that, 5 years after planting, small shrubs compete more intensively for soil resources than gramineous species. Competition release from small shrubs resulted in stronger improvement of growth and functional traits of the planted saplings than release from gramineous vegetation. Particularly at the Otterbach site with more developed ground vegetation, the removal of small shrubs increased diameter growth of both species. Beech SLA and SRL decreased significantly after release from small shrubs competition. An unusually dry growing season like the one in 2008 can severely intensify the competition for soil resources. In our study, we could show this by means of the negative correlation coefficients between a sapling’s leaves/needles nutrient content and surrounding ground vegetation’s dry mass and/or nutrient content which were higher in 2008. According to our findings, weed control as a tool to improve growth of planted beech and Douglas-fir saplings was only advantageous on sites with well-developed small shrubs competition (mainly by Rubus fruticosus and R. idaeus), or under fairly dry conditions.

The transfer of our results into practical forest management is restricted due to the short observation period of 2 years. This might explain the lack of a significant response of sapling length growth and fine root biomass partitioning in our study. We expect that a longer experimental period would provide a considerably better understanding of the mechanisms of tree species response to weed control.