Canada bluejoint foliar δ¹⁵N and δ¹³C indicate changed soil N availability by litter removal and N fertilization in a 13-year-old boreal plantation

Abstract

Canada bluejoint grass [Calamagrostis canadensis (Michx.) P. Beauv., hereafter referred to as bluejoint] outcompetes overstory tree species such as white spruce [Picea glauca (Moench) Voss] by creating a thick litter layer and competing for the available nitrogen (N). This study was conducted to investigate the effects of bluejoint litter layer (with or without litter removal) and N fertilization on soil water and N availabilities using principal component analysis (PCA) and foliar δ¹⁵N and δ¹³C of bluejoint in a plantation in north-central Alberta, Canada. PCA using soil properties and understory growth data demonstrated that N fertilization was more effective in changing the soil environment and resource availabilities for bluejoint growth than litter layer removal. The increase in soil N availability by N fertilization was linked with increased bluejoint foliar δ¹⁵N (by around 3‰) in fertilized plots, as a result of greater N isotopic fractionation in the fertilized plots. The more negative δ¹³C (by around 1‰) of bluejoint in litter layer-removed plots suggested that litter layer removal increased soil water availability, indicating that the litter layer reduced soil water availability on the site. Therefore, results from this and previous studies showed that the litter layer decreased both soil water and N availabilities. Although the exact mechanisms of the benefit of the litter layer for bluejoint remains unknown, bluejoint likely adversely impacted tree growth by competing for N due to its strong N acquisition ability under soil resource-limiting conditions.

Key words:

• Canada bluejoint grass (Calamagrostis canadensis (Michx.) P. Beauv.)
• δ¹³C
• δ¹⁵N
• understory competition
• white spruce (Picea glauca (Moench) Voss)

INTRODUCTION

Canada bluejoint grass [Calamagrostis canadensis (Michx.) P. Beauv., hereafter referred to as bluejoint] is a fast-growing understory species that can outcompete overstory tree species such as white spruce [Picea glauca (Moench) Voss] (Staples et al. 1999; Hangs et al. 2002), aspen (Populus tremuloides Michix.) (Landhäusser and Lieffers 1998), jack pine (Pinus banksiana Lamb.) (Bell et al. 2000; Hangs et al. 2002, 2003b), and black spruce [Picea mariana (Mill.) BSP] (Bell et al. 2000) in boreal forest ecosystems. The mechanisms for the strong competitiveness of bluejoint have been extensively studied, especially after Lieffers et al. (1993) published a review paper describing the ecological characteristics of bluejoint. The previous studies addressed the effects of soil nutrient availability and temperature on the competitiveness of bluejoint (Landhäusser and Lieffers 1994) and reported that bluejoint has a high nitrogen (N) fertilizer use efficiency (Staples et al. 1999) due to a strong N uptake ability as compared with other boreal plant species (Hangs et al. 2003a).

Bluejoint proliferates after a site experiences a disturbance, such as clear-cutting, which induces changes in N cycling in the site, much like other competitive understory species (Chang et al. 1996; Prescott et al. 1996; Björk et al. 2007; Laungani and Knops 2009; Rossiter-Rachor et al. 2009). One of the characteristics of the bluejoint community is the creation of a thick (up to 50 cm) litter layer on the ground above the organic soil layer (specifically layers of litter, gragmented litter, and humus, the LFH layer ) due to the mass production of bluejoint biomass and thus the accumulation of bluejoint litter, which has a slow decomposition rate (Lieffers et al. 1993). This litter layer is a unique characteristic of bluejoint and it lowers soil temperature because of its insulating effect (Hogg and Lieffers 1991; Matsushima and Chang 2007a) and thus further decreases the litter decomposition rate (Matsushima and Chang 2007a). Detrimental effects of bluejoint litter on aspen growth (Landhäusser and Lieffers 1998) and root sucker regeneration (Landhäusser et al. 2007) have been also reported. In a white spruce plantation in north-central Alberta, we investigated the effects of bluejoint removal in combination with litter removal and N fertilization on white spruce growth and soil N dynamics in a series of experiments for 2 years (2003 and 2004) (Matsushima and Chang 2006; 2007a; Matsushima et al. 2012). These earlier studies indicated that N availability for overstory species (white spruce in this case) increased as a result of bluejoint removal, and this translated to increases in tree height and diameter growth. The increased N availability was evidenced by the increased N isotope ratio of tree foliage due to N isotopic fractionation associated with N loss which is likely to be proportional to N availability (Matsushima et al. 2012). Despite the availability of a large dataset on the relationship between bluejoint grass and white spruce growth, we do not have a full understanding of the integrative impact of litter removal and N fertilization on soil resource (water and N) availability and the consequent effects on the understory vegetation.

This study was conducted to evaluate the impacts of bluejoint litter layer treatments (removal vs. intact) and N fertilization on soil resource availability using principal component analysis (PCA) on the multivariate dataset including soil and bluejoint data obtained from the 2-year field experiment (Matsushima and Chang 2006, 2007a, 2007b; Matsushima et al. 2012). In addition, to assess the beneficial effects of litter layer accumulation on bluejoint, the changes in natural N (δ¹⁵N) and carbon (C; δ¹³C) isotope abundances in bluejoint in response to litter layer removal and N fertilizer application were investigated. As plant δ¹⁵N and δ¹³C are affected largely by soil N and moisture availability (Choi et al. 2005a; Matsushima et al. 2012), these are expected to serve as indicators of changed soil dynamics with litter layer removal and N fertilization.

2. MATERIALS AND METHODS

2.1. Study site

The details of the field experiment were described in our previous publications (Matsushima and Chang 2006; Matsushima and Chang 2007a; Matsushima et al. 2012). Briefly, the study site (54°7′N, 115°50′W) was located approximately 20 km southwest of the city center of Whitecourt in Alberta, Canada, with a mean annual precipitation of 578 mm and a mean annual temperature of 2.6°C (Environment Canada 2004). In this region, the mean annual potential evapotranspiration was 508 mm (The National Atlas of Canada 1974). The soil in the study site was an Eluviated Sombric Brunisol according to the Canadian System of Soil Classification (Soil Classification Working Group 1998). The site was clear-cut in November 1991, and white spruce was planted in 1992. In 2003, when we started the experiment, the understory of the site was dominated by bluejoint (50–100% cover; 82% on average). A large amount of understory dead shoot (litter) of bluejoint accumulated on the ground (579 g m⁻², 43% C and 0.64% N on average) above the LFH layer (5–10 cm thickness) above the mineral soil surface. Where the site was less infested by bluejoint, there was relatively more fireweed (Epilobium angustifolium L.).

2.2. Experimental design

Three treatments were applied to 24 experimental plots in a 2 × 2 × 2 split-split plot design with three replications. Each experimental plot was 8 × 8 m and included 5–14 trees. At the initiation of the experiment, in June 2003, tree height and diameter at the ground level were 183 and 3.7 cm on average, respectively, and cumulative growth of height and diameter in two growing seasons in the control plot were 69 and 1.7 cm, respectively (Matsushima and Chang 2006). The three treatments applied were as follows: (1) understory intact vs. understory removed (U) at the whole-plot level, (2) no N fertilization vs. N fertilization (N) at the split plot level, and (3) litter layer intact vs. litter layer removed (L) at the split-split plot level. As the objective of the present study is to investigate the effects of litter layer and N fertilization on soil resource availability using understory (bluejoint) data, the U treatment was not included in this paper. The effect of the U treatments on the overstory tree growth and soil resource availability have been extensively reported in other publications (Matsushima and Chang 2006; 2007a, 2007b; Matsushima et al. 2012). This experiment therefore had the following four treatment combinations: control, L, N, and LN. Urea was broadcast applied at 200 kg N ha⁻¹ to the N fertilization (N) plots in June 2003, and the understory litter was manually cut at ground level above the organic layer and removed from the litter layer removal (L) plots in 2003 and 2004. Dead shoots of bluejoint accumulated as almost-intact leaves up to 50 cm above the organic layer. We treated this litter as the litter layer and manually removed it, and the organic layer was not disturbed by this treatment.

2.3. Data acquisition

The data (except for bluejoint δ¹⁵N and δ¹³C) used for the PCA analysis have been reported in detail in our earlier publications (Matsushima and Chang 2006, 2007a, Matsushima et al. 2012), and this paper used the average values over the two growing seasons. Here, we briefly provide a description of the sampling and measurement protocols. Current-year needles of white spruce were collected from upper-mid crown locations in September 2003 and September 2004. Needle samples were dried in an oven at 70°C for 24 h and unit needle weight (weight of 100 needles) was determined. Subsamples of needles were ground to fine powder and analyzed for N concentration and δ¹⁵N and δ¹³C (see definitions for their calculation below) of white spruce. The aboveground bluejoint samples were also collected from each experimental plot on the same dates in 2003 and analyzed for δ¹³C and δ¹⁵N.

The organic and mineral soils (0–10 cm) were sampled in September 2004, air-dried, and then used to determine δ¹⁵N values for the bulk soil N. The gravimetric water contents of organic and mineral soils (0–10 cm) were monitored monthly from June to September of 2003 and 2004. Similarly, the soil temperature at 10 cm during the experimental period (from June to September in 2003 and 2004) was recorded by three HOBO^® H8 Temp Loggers (Onset Computer Corporation, Bourne, MA, USA) at each experimental plot. The soil microbial biomass C (MBC) and N (MBN) of organic and mineral soil were measured by the fumigation-extraction method (Vance et al. 1987) from June to August 2004. The nitrification and N mineralization rates were measured using the in situ incubation method (intact core incubation in a buried bag) (Eno 1960) in June to August 2004. Plant Root Simulator (PRS)^™ probes (Western Ag Innovations Inc., Saskatoon, Canada), which were equipped with ion-exchange membranes, were used to estimate the potential soil N supply rates (soil available N) from June to August 2004.

Because bluejoint δ¹³C and δ¹⁵N were not reported in the earlier publications (Matsushima and Chang 2006, 2007a, 2007b, Matsushima et al. 2012), we describe them in detail. The bluejoint samples were ground using a ball mill (MM-200, Retsch GmbH 88 Co., KG, Germany), weighed adequately for isotopic analysis into tin capsules (0.07 mL), and analyzed for N concentration, δ¹⁵N and δ¹³C using a mass spectrometer connected to an elemental analyzer (Integra-CN, PDZ Europa, UK). The δ¹⁵N and δ¹³C were calculated as:

(1)

where R_sample and R_standard are the ratio of N (¹⁵N/¹⁴N for δ¹⁵N) and C (¹³C/¹²C for δ¹³C) isotopes for sample and standard, respectively. The standards for N and C are atmospheric N₂ and the Pee Dee Belemnite standard (PDB), respectively. The accuracy and reproducibility of the measurements of δ¹⁵N and δ¹³C were checked using an internal reference and found to be better than 0.3 and 0.2‰ for δ¹⁵N, and 0.2 and 0.1‰ for δ¹³C, respectively. The internal reference material, glycine (δ¹⁵N: +2.0 ± 0.1‰; δ¹³C: –31.86‰ ± 0.1‰), was calibrated against IAEA-N2 (ammonium sulfate, δ¹⁵N = +20.3‰) for δ¹⁵N and against IAEA-C6 (sucrose, δ¹³C = –10.8‰) for δ¹³C.

2.4. Statistical analysis and PCA

Analysis of variance (ANOVA) and multiple comparison (Tukey’s test) were conducted to investigate the statistical significance of all data using the mixed model procedure of PASW Statistic 17.0 (SPSS Inc., Chicago, IL, USA); a P value of < 0.05 was taken to indicate statistical significance.

The PCA was performed using the same software to examine the impact of L and N treatments on soil resource availability using the data set of soil properties (δ¹⁵N of organic and mineral soils including average water content in organic and mineral soils, microbial biomass C and N in organic and mineral soils, average soil temperature, N availability measured using ion-exchange membranes, nitrification and N mineralization rates in organic and mineral soils, and bluejoint growth parameters including N concentration in foliage, biomass accumulation and N uptake. These PCAs enable us to visually depict similarities and differences of the site ecological characteristics along a small number of extracted new axes to withdraw integrative understanding of the whole data set as affected by the treatments. Only two principal components for each diagram were selected because their scree plots showed decreased eigenvalues after the third component was added.

3. RESULTS AND DISCUSSION

3.1. PCA

PCA using soil nutrient availability data has widely been used to describe site and treatment traits such as forest burning history (Durán et al. 2012), edaphic and climatic characteristics (Zeglin et al. 2007), and urbanization effects on forest soils (Pouyat et al. 1995). Similarly, we analyzed our soil and bluejoint growth data using PCA to examine the impacts caused by N fertilization and litter layer removal on site similarities and differences. In the PCA biplot, that is based soils data, PC1 and PC2 explained 37 and 18% of variations of the analyzed soil variables, respectively (Fig. 1A, Table 1). PC1 was positively correlated with water availability in organic and mineral soils, MBC and MBN in organic soil (Table 1). PC2 did not have strong correlation with specific soil properties but was correlated positively with total N and pH of mineral soil and negatively with available N, nitrification and mineralization rates of mineral soil (Table 1). In this biplot, the control and L plots are located together on the lower side of PC2, whereas the N and LN plots are located on the upper side of PC2 with one exception (i.e. one plot of control on the positive side of PC2) (Fig. 1A). In the other PCA biplot, that is depicted using understory growth data, PC1 and PC2 explained 66 and 27% of variations, respectively (Fig. 1B, Table 2). Growth and nutrient data of understory vegetation (N concentration, aboveground biomass and N uptake of bluejoint) may reflect soil resource availability such as water and N for its growth (Casper and Jackson 1997; Matsushima and Chang 2006). This PCA biplot delineated similar grouping to the other PCA; control and L plots are grouped together on the left side, and N and LN are on the other side of PC1 (Fig. 1B). These results may indicate that N fertilization had relatively greater impacts on the soil environment and resource availabilities than litter layer removal.

Table 1 Principal component (PC) loadings (correlation coefficients between principal component scores and original data) of principal component analysis (PCA) using soil property data.

Soil properties	PC 1 (37%)*	PC 2 (18%)
Water availability in organic soil	0.903	−0.124
Water availability of mineral soil	0.909	−0.012
Soil temperature	−0.637	0.397
δ^15N of organic soil	0.620	−0.063
δ^15N of mineral soil	0.822	0.297
Available N	−0.414	−0.401
MBC of organic soil	0.879	−0.207
MBC of mineral soil	−0.386	−0.344
MBN of organic soil	0.968	−0.144
MBN of mineral soil	−0.281	−0.130
Nitrification rate of organic soil	0.050	−0.423
Nitrification rate of mineral soil	0.351	−0.538
Mineralization rate of organic soil	−0.426	0.411
Mineralization rate of mineral soil	−0.177	−0.535
Total C of organic soil	0.634	0.549
Total N of organic soil	0.864	0.354
Total C of mineral soil	−0.103	0.545
Total N of mineral soil	−0.089	0.800
pH of organic soil	0.767	0.000
pH of mineral soil	−0.104	0.778

*: The percent of variation explained by the PCA axis. δ¹⁵N, natural abundance of nitrogen-15; MBC, microbial biomass carbon; MBN, microbial biomass nitrogen.

Table 2 Principal component (PC) loadings (correlation coefficients between principal component scores and original data) of principal component analysis (PCA) using bluejoint [Calamagrostis canadensis (Michx.) P. Beauv.] growth data.

Bluejoint growth parameters	PC 1 (66%)*	PC 2 (27%)
Foliar N concentration in 2003	0.507	0.834
Foliar N concentration in 2004	−0.765	0.529
Cumulative aboveground N uptake of 2003 and 2004	0.960	0.204
Cumulative above ground biomass of 2003 and 2004	0.946	−0.227

*: The percent of variation explained by the PCA axis.

Figure 1 Results of principal component analysis: (A) using soil environmental properties and (B) using the growth data of bluejoint [Calamagrostis canadensis (Michx.) P. Beauv.] (the lists of used properties and their principal component loadings are reported in Tables 1 and 2). Treatment code: L stands for litter layer removal; N stands for nitrogen fertilization.

3.2. δ¹⁵N

Plant δ¹⁵N is regarded as a surrogate of soil N dynamics, particularly for the N loss process that results in ¹⁵N enrichment of the remaining soil N due to the N isotopic fractionation (Robinson 2001; Stevenson et al. 2010). In the present study, application of N fertilizer increased bluejoint foliar δ¹⁵N by 2.4 and 3.2‰ in 2003 (after the first growing season after treatment) and 2004 (after the second growing season), respectively, but litter removal and their interaction did not affect δ¹⁵N in both years (Table 3). Because the applied urea (δ¹⁵N: –0.87 ± 0.03‰) was ¹⁵N-depleted compared to the soil N (1.1‰ for organic soil and 4.0‰ for mineral soil) (Matsushima et al. 2012), the higher foliar δ¹⁵N of bluejoint in the N treatments (1.4 and 2.3‰ in 2003 and 2004, respectively) and in the LN treatment (2.7 and 2.8‰ in 2003 and 2004, respectively) than the control (–0.1 and –0.7‰ in 2003 and 2004, respectively) suggests that application of ¹⁵N-depleted urea did not lower foliar δ¹⁵N of bluejoint. Although plants applied with ¹⁵N-depleted fertilizer have lower δ¹⁵N than unfertilized plant in the short period after N fertilization, δ¹⁵N of fertilized plants is likely to increase with time due to increases in the δ¹⁵N of fertilizer-derived N due to N isotopic fractionation associated with N loss (Choi et al. 2002; Yun et al. 2011). Choi et al. (2005a) also reported that application of urea (δ¹⁵N = −2.7‰) increased foliar δ¹⁵N of loblolly pine (Pinus taeda L.) from –3.8 ± 0.4‰ to –0.2 ± 0.3‰. Therefore, the increased foliar δ¹⁵N by fertilization is plausible to be attributed to N isotopic fractionation associated with N loss via ammonia (NH₃) volatilization and leaching of ¹⁵N-depleted nitrate (NO₃^–) produced from incomplete nitrification since N loss causing N isotope fractionation is likely to be parallel with soil N availability (Högberg 1997). A significant positive linear relationship between foliar N concentration and δ¹⁵N of bluejoint in 2003 (Fig. 2A) supports this hypothesis.

Table 3 Effects of litter layer removal and nitrogen (N) fertilization on the natural abundance of ¹⁵N and carbon-13 (¹³C) in bluejoint [Calamagrostis canadensis (Michx.) P. Beauv.] foliage determined in September 2003 and 2004. Bold numbers indicate statistical significance (P < 0.050) among the treatments at each sampling time. Treatment codes: L stands for litter layer removal; N stands for N fertilization.

	δ^15N		δ^13C
Treatment	2003	2004	2003	2004
Control	−0.1 ± 0.2	−0.7 ± 0.4	−26.7 ± 0.1	−27.4 ± 0.5
L	−0.7 ± 0.6	−0.5 ± 0.3	−28.2 ± 0.1	−28.2 ± 0.1
N	1.4 ± 0.6	2.3 ± 1.2	−26.7 ± 0.5	−27.8 ± 0.3
LN	2.7 ± 0.8	2.8 ± 0.9	−27.1 ± 0.2	−28.6 ± 0.3
Effects	Probability > F
Nitrogen (N)	0.003	0.002	0.119	0.328
Litter (L)	0.531	0.623	0.017	0.045
N*L	0.107	0.816	0.106	0.968

Figure 2 Correlation relationships between nitrogen (N) concentration and the natural abundance of (A) ¹⁵N and (B) carbon-13 (¹³C) in bluejoint [Calamagrostis canadensis (Michx.) P. Beauv.] foliage determined in September 2003 and 2004.

Although both ammonium (NH₄⁺) and NO₃^– are available for plant uptake, plant species show different preferences for either NH₄⁺ or NO₃^– depending on soil conditions such as relative abundance of soil NH₄⁺ and NO₃^–, pH and temperature (Choi et al. 2005b). Since bluejoint roots are known to prefer NH₄⁺ to NO₃^– (Hangs et al. 2003a), uptake of NH₄⁺ that experienced nitrification and thus enriched with ¹⁵N under NH₄⁺ abundant conditions by urea application seemed to induce increases in foliar δ¹⁵N. However, in a previous study (Matsushima et al. 2012), we found that δ¹⁵N of soil mineral N (NH₄⁺ and NO₃^–) measured at the end of a 2-year experiment was not affected by N fertilization, probably due to dilution of the δ¹⁵N signal of mineral N by complicated soil N process including N mineralization that produces ¹⁵N-depleted N relative to the existing mineral N pool (Choi and Ro 2003). This suggests that measurement of δ¹⁵N of mineral N in the active growing season rather than at the end of the growing season such as in this study may better reflect changed soil N dynamics by N fertilization.

While foliar δ¹⁵N of bluejoint was affected by fertilization in both years (P = 0.003 for 2003, P = 0.002 for 2004, Table 3), the response of foliar δ¹⁵N of the overstory tree (white spruce) to fertilization was slightly found only in 2004 (P = 0.832 for 2003, P = 0.105 for 2004, Table 4). This indicates that bluejoint is affected by changes in soil N availability and dynamics more responsively than white spruce, probably due to the well developed root system and high N uptake ability of bluejoint (Hogg and Lieffers 1991; Hangs et al. 2004).

Table 4 Effects of litter layer removal and nitrogen (N) fertilization on the natural abundance of ¹⁵N and carbon-13 (¹³C) in white spruce [Picea glauca (Moench) Voss] foliage determined in September 2003 and 2004. Bold numbers indicate statistical significance (P < 0.050) among the treatments at each sampling time. Treatment codes: L stands for litter layer removal; N stands for N fertilization.

	δ^15N		δ^13C
Treatment	2003	2004	2003	2004
Control*	−0.1 ± 1.5	−2.5 ± 1.2	−27.5 ± 0.2	−27.8 ± 0.1
L	−0.5 ± 1.5	−3.1 ± 0.8	−27.8 ± 0.2	−28.0 ± 0.6
N*	−0.3 ± 0.7	−0.9 ± 0.6	−27.6 ± 0.1	−27.3 ± 0.5
LN	−0.1 ± 0.2	−0.7 ± 0.8	−27.5 ± 0.1	−27.4 ± 0.2
Effects	Probability > F
Nitrogen (N)	0.832	0.105	0.202	0.019
Litter (L)	0.885	0.740	0.211	0.237
N*L	0.178	0.506	0.108	0.888

*: Data for control and N was reported in Matsushima et al. (2012).

3.3. δ¹³C

Bluejoint foliar δ¹³C was significantly decreased by litter removal throughout the two growing seasons by –0.95 in 2003 and –0.82‰ in 2004, but not by fertilization (Table 3); meanwhile foliar δ¹³C of white spruce did not respond to litter removal (Table 4). According to the C isotope discrimination model for C3 plants (Farquhar et al. 1989), plant δ¹³C is determined by the ratio (C_i/C_a) of intercellular (C_i) to atmospheric (C_a) carbon dioxide (CO₂) concentration that is controlled by the balance of CO₂ supply via stomata and consumption by carboxylation. Therefore, many environmental conditions including soil water and nutrient availability affect plant δ¹³C by influencing either stomatal conductance, carboxylation rate or both (Farquhar et al. 1989). For example, an increase in water availability is likely to decrease plant δ¹³C by facilitating CO₂ supply via stomata; whereas a better nutrient condition tends to lead to an increase in plant δ¹³C by enhancing carboxylation (Choi et al. 2005a). Therefore, the decreased foliar δ¹³C of bluejoint suggests that soil water availability was improved by litter removal. This collaborates with the finding that litter layer removal significantly increased water content of organic soil layer at the study site (Matsushima and Chang 2006). As bluejoint rhizomes prefer low bulk density conditions such as the organic layer and the interface between organic and mineral soils (Lieffers et al. 1993; Landhäusser et al. 1996), bluejoint may be able to benefit from the increased soil water availability in the organic layer.

Litter layer removal could result in an increase in δ¹³C by enhancing carboxylation rate via increased N availability as indicated by increased mineralization and nitrification rates (Matsushima and Chang 2007a), but our results showed the reverse trend. Therefore, the decreased δ¹³C by litter removal suggested that changed soil water availability affected bluejoint photosynthesis more greatly than nutrient availability. The non-significant relationship between foliar N concentration and the corresponding δ¹³C of bluejoint (Fig. 2B) also highlights less significant contribution of N availability than water availability to gas exchanges, probably due to the inherently high N uptake ability of bluejoint (Hangs et al. 2003a). The role of the thick litter layer of bluejoint, especially in its effects on the competitive ability of the different plant species in a community, has not been addressed adequately in previous studies (Hogg and Lieffers 1991; Lieffers et al. 1993; Landhäusser and Lieffers 1998; Landhäusser et al. 2007). The δ¹³C response of bluejoint to litter layer removal reported here and in previous studies (Hogg and Lieffers 1991; Matsushima et al. 2007a) is linked to decreased N mineralization due to lowered temperature in plots with a litter layer. Although the presence of a litter layer can reduce both soil water and N availabilities, bluejoint can survive better than overstory trees due to its strong N uptake ability and high N use efficiency (Staples et al. 1999; Hangs et al. 2003a).

In conclusion, the PCA based on soils and bluejoint growth data in combination with foliar δ¹⁵N suggested that N fertilization significantly changed resource availability and growing environment much more than litter layer removal. Although the exact mechanism for the benefit of the litter layer for bluejoint is not clear yet, the foliar δ¹³C data in this and our previous studies (Matsushima and Chang 2006, 2007a, 2007b; Matsushima et al. 2012) indicate that although creation of the litter layer may reduce soil water availability for bluejoint as well as overstory trees, bluejoint may further adversely impact tree growth by competing for N due to its high N acquisition ability under N-limiting conditions.

ACKNOWLEDGMENTS

Initial funding for this study was provided by the Natural Science and Engineering Research Council (NSERC) of Canada, the Canadian Foundation for Innovation, the Weyerhaeuser Company Ltd. and Weldwood of Canada Ltd. We thank Dave Swindlehurst, Weyerhaeuser Company Ltd., and Paul Godin, Millar Western Forest Products Ltd., for assistance in site selection and in providing information on site history.