Plant Ecology

, Volume 192, Issue 1, pp 21–33

Water and nitrogen addition differentially impact plant competition in a native rough fescue grassland

Authors

    • Department of BiologyUniversity of Alberta
  • Bryon H. Shore
    • Department of BiologyUniversity of Alberta
  • James F. Cahill
    • Department of BiologyUniversity of Alberta
Original Paper

DOI: 10.1007/s11258-006-9222-4

Cite this article as:
Lamb, E.G., Shore, B.H. & Cahill, J.F. Plant Ecol (2007) 192: 21. doi:10.1007/s11258-006-9222-4

Abstract

We examined how water and nitrogen addition and water–nitrogen interactions affect root and shoot competition intensity and competition–productivity relationships in a native rough fescue grassland in central Alberta, Canada. Water and nitrogen were added in a factorial design to plots and root exclusion tubes and netting were used to isolate root and shoot competition on two focal species (Artemisia frigida and Chenopodium leptophyllum). Both water and nitrogen were limiting to plant growth, and focal plant survival rates increased with nitrogen but not water addition. Relative allocation to root biomass increased with water addition. Competition was almost entirely belowground, with focal plants larger when released from root but not shoot competition. There were no significant relationships between productivity and root, shoot, or total competition intensity, likely because in this system shoot biomass was too low to cause strong shoot competition and root biomass was above the levels at which root competition saturates. Water addition had few effects on the intensity of root competition suggesting that root competition intensity is invariant along soil moisture gradients. Contrary to general expectation, the strength of root competition increased with nitrogen addition demonstrating that the relationship between root competition intensity and nitrogen is more complex than a simple monotonic decline as nitrogen increases. Finally, there were few interactions between nitrogen and water affecting competition. Together these results indicate that the mechanisms of competition for water and nitrogen likely differ.

Keywords

Aboveground competitionBelowground competitionNitrogen fertilizationSoil moisture

Introduction

The predicted patterns of root, shoot, and total competition intensity along gradients of resource availability and productivity (Grime 1973, 2001; Newman 1973; Tilman 1988) have prompted numerous experimental studies with conflicting results and no resolution (reviewed by Goldberg and Barton 1992; Keddy 2001; Craine 2005; Schenk 2006). Understanding why and if root and shoot competition intensity varies with productivity and resource availability remains key to linking the mechanisms of resource competition to the consequences of competition for community structure (Keddy 2001). It is generally agreed that the intensity of shoot competition increases with increasing productivity since shading is closely correlated with neighbour size (Tilman 1988; Grime 2001; Keddy 2001). In contrast, the relationships between root competition intensity and productivity vary widely between studies (e.g. Goldberg and Barton 1992; Belcher et al. 1995; Twolan-Strutt and Keddy 1996; Peltzer et al. 1998; Cahill 1999; Keddy 2001; Sammul et al. 2006). Additionally, root and shoot competition are interdependent with the relative strength of one competitive form dependent upon the level of the alternative form (Cahill 1999, 2002a).

Belowground, plants compete for multiple resources with differing physical properties (Casper and Jackson 1997). The addition of a limiting soil resource should reduce the intensity of root competition by reducing the degree of deficiency for the limiting resource relative to other resources (Taylor et al. 1990; Casper and Jackson 1997; Davis et al. 1998), but experimental studies have found that the outcome can depend on the resource involved. Root competition intensity typically declines as nitrogen levels increase (Wilson and Tilman 1991, 1993, 1995; Peltzer et al. 1998; Cahill 1999), though Cahill (2002a) observed no change in intensity and Brewer (2003) an increase following fertilization. Competition for phosphate may be similar to nitrogen since Santos et al. (2004) found that when root competition was severe phosphate addition reduced belowground competition intensity. In contrast to the studies of mineral nutrients, several studies have shown that water addition can increase productivity without significant effects on the strength of root competition (Burger and Louda 1995; Haugland and Froud-Williams 1999; Semere and Froud-Williams 2001; Weigelt et al. 2005). Minimal effects of both water and mineral resources on total competition intensity have been found (Wetzel and van der Valk 1998; Fynn et al. 2005), but no study has examined root competition for both water and mineral resources in a single experiment. Without such an experiment it is difficult to determine whether these differences in outcome result from differences in the mechanisms of competition for water and mineral resources or are due to factors such as the concentration of studies of competition for water in more arid systems.

Equally important to differences in mechanisms of competition for belowground resources may be interactions between resources such as the close links between soil moisture, the nitrogen cycle, and plant nitrogen uptake (Fitter and Hay 2002; Booth et al. 2005; James and Richards 2005). If for example, available nitrogen or nitrogen-use efficiency increase with water addition, then water addition could cause increased nitrogen supply and reduced competition. Such interactions may be important in natural systems where productivity gradients often follow multiple covarying resource gradients (Keddy 2001). Interactions between resources have been largely ignored in studies of competition except in arid and semi-arid systems where the frequency and intensity of pulses of the nitrogen and water that follow rainfall are important (Goldberg and Novoplansky 1997; Novoplansky and Goldberg 2001).

In this study, we used a field experiment in native rough fescue grassland to examine the intensity of root and shoot competition along productivity gradients created by water and nitrogen addition. The rough fescue grasslands in the aspen parkland region of western Canada represent a unique opportunity to study the interactions of water and nitrogen as both resources can be limiting to plant growth in this community (E.G. Lamb unp. data). This is the first study to compare the effects of both nitrogen and water addition on the intensity of competition and the relationships between competition and productivity. Specific questions examined in this study include: (1) whether water and nitrogen addition have similar effects on the intensity of root and shoot competition, (2) if the relationships between competition and productivity are a function of the resource used to create the productivity gradient, and (3) whether interactions between nitrogen and water have important consequences for the outcome of competition.

Methodology

Experimental design

This experiment was conducted in native rough fescue grassland on the University of Alberta Research Ranch near Kinsella, Alberta, Canada (53°05 N, 111°33 W). The study site is a savanna-type habitat in the aspen parkland ecoregion (Sims and Risser 2000), containing a mixture of aspen (Populus tremuloides Michx.) stands and rough fescue (Festuca hallii (Vasey) Piper) prairie. The soils are thin, moderately well drained, black grassland soils over glacial till (Howitt 1988). Root competition in this community is strong (Cahill 2003a, b), and both nitrogen and water availability can limit plant growth (E.G. Lamb unp. data).

One hundred and twenty 1 m by 1 m plots were established in a 20 m by 24 m grid on a south-facing slope in the spring of 2003. The two nitrogen treatments (control and 5.44 g m−2 year−1 ammonium-nitrate applied in two 2.72 g m−2 doses in mid-May and late June) and water (control and 7.5 l week−1) were applied to the plots in a factorial design. There were 30 replicates of each nitrogen by water treatment combination. Each of these replicates contained four subplots with the following competition treatments: all neighbours (AN), shoot neighbours (SN), root neighbours (RN), no neighbours (NN). Root exclusion tubes (10.2 cm diameter and 10 cm deep) made of PVC pipe were used to separate focal plants from RN in the SN and NN treatments. Plastic netting was used to hold neighbouring shoots back from focal plants in the RN and NN treatments. Root and rhizome connections in the treatments without root exclusion tubes were cut to ensure a similar soil environment to that in the root exclusion tubes. The existing plants within each tube/trenched area were sprayed with herbicide (Roundup©). The nitrogen treatment was applied in both 2003 and 2004 while the watering treatment was begun in 2004. Cattle grazed the site in September 2003, but were not present during the experiment.

The experiment was begun in 2004, providing a year delay between the setup and the beginning of the experiment to allow the neighbouring plants to re-establish around the root exclusion tubes. In May 2004 the germination of Artemisia frigida Willd. (a perennial forb) and Chenopodium leptophyllum (Nutt ex Moq.) S. Wats (an annual forb) from the soil seedbank was encouraged by watering all plots weekly for 3 weeks. If more than one seedling was in a plot, one was randomly selected for study and the rest removed. Once the initial watering ceased, plots in the watered treatments continued to receive water at a rate of 7.5 l per week, or the equivalent of an extra 7.5 mm week−1 of rain for a total of 97.5 mm, a 60% increase over natural rainfall (160.5 mm) measured at the Viking, Alberta station (53°16 N, 111°46 W) during the same period (Environment Canada National Climate Archive; http://www.climate.weatheroffice.ec.gc.ca).

Focal plants were harvested in the third week of August 2004, after 14 weeks of growth. At harvest many of the Chenopodium had finished flowering, but none of the Artemisia had flowered. Shoot biomass was harvested, dried, and weighed. Root biomass was not harvested due to the difficulty in accurately extracting the root systems of the focal plants from the NN and RN treatments, as the roots in those treatments would be intertwined with those of the neighbouring plants (Cahill 2002b).

Soil moisture content, photosynthetically active radiation (PAR) transmission through the vegetation, and root and shoot neighbour biomass were measured in each plot immediately following the harvest of the focal plants. PAR was measured above and below the canopy using a handheld light meter (AccuPAR model PAR-80; Decagon Devices, WA). Shoot biomass was measured by clipping the live vegetation from a 20 cm by 50 cm quadrat. Root biomass was measured by washing the roots from 5.3 cm diameter root cores taken to a depth of 12 cm. Gravimetric soil moisture content was measured by collecting and weighing the wet and dry masses of a small soil sample (∼35 g) sieved from the root biomass core. Soil moisture was measured 5 days after a water application. As the intervening weather had been hot and dry, the differences in moisture content between treatments likely represent minimum differences.

Statistical analysis

The effects of the nitrogen and water treatments on environmental conditions and the neighbouring plant community were analysed using general linear models with the nitrogen and water treatments as fixed factors. The response variables were arcsine-transformed % soil moisture, arcsine-transformed % light transmission, and ln-transformed root and shoot biomass. The effects of the nitrogen and water treatments on relative allocation to root and shoot biomass were analysed using a general linear model with aboveground biomass as the response variable, root biomass as a covariate, and nitrogen and water treatments as fixed factors. All interactions including the covariate were included in an initial model but non-significant interactions were removed (Engqvist 2005) leaving a significant root biomass–water interaction in the final model. Since the root biomass–water interaction indicated that the effects of water on shoot biomass were dependent on root biomass levels, these data were divided into four groups (root biomass <500, 500–800, 800–1,100, and >1,100 g m−2) and the effect size (least-squares means) of the water treatment was estimated for each group. All analyses were conducted using proc GLM in SAS 8.02.

Focal plant survival rates were analysed using a log-linear model (G-test) with species, nitrogen addition, water addition, root, and shoot competition as fixed factors. A log-linear model is a generalized linear model with the number of surviving plants per treatment combination as the response variable and a Poisson error distribution (SAS Institute 2004). The initial number of plants (between 8 and 20 depending on the treatment) was used as an offset variable to standardize for the different numbers of starting plants between treatment combinations. A saturated model including all possible main effects and interactions was fit to these data using proc GENMOD in SAS 8.02.

Focal plant biomass was analysed using a mixed model with species, nitrogen addition, water addition, root competition, and shoot competition as fixed factors and plot as a random factor. Models were fit using proc MIXED in SAS 8.02. Satterthwaite approximate degrees of freedom were used since these data were unbalanced. Changes in the intensity of root and shoot competition caused by the resource addition treatments are indicated by significant competition–resource interactions.

Competition intensity, or the relative difference in performance between plants with and without neighbours, was directly examined using log response ratios (lnRR) (Hedges et al. 1999). The lnRR was chosen because, among the competition indices in common use, it has statistical properties that are best suited for linear analysis (Hedges et al. 1999; Weigelt and Jolliffe 2003). These indices were calculated for shoot (ln[SN/NN]), root (ln[RN/NN]), and total (ln[AN/NN]) competition following Cahill (1999). Positive values of the lnRR indicate facilitation while increasingly negative values indicate increasing intensity of competition. These indices are intended for pairs of focal plants, but in this study the biomass of each focal plant from an SN, RN, or AN treatment was divided by the mean biomass of NN plants from the same species by water by nitrogen treatment combination. This procedure was used because the use of focal plants germinated from the seedbank and mortality during the experiment left very few plots with plants of the same species in competition treatments appropriate for pairing. Indices of competition are problematic because they require the assumption that competitive ability does not vary with plant size, however without pairing plants this assumption could not be tested nor could statistically more rigorous alternatives such as ANCOVA be used (Lamb et al. 2006). Since we were interested in the relationships between competition, productivity, and the resource addition treatments we standardized the response ratios to eliminate differences in competitive ability between the species using z-scores. The z-scores were analysed using general linear models with nitrogen and water addition as fixed factors and neighbouring plant biomass as a covariate using PROC GLM in SAS 8.02. Separate analyses were run for each combination of competition intensity (root, shoot, and total) and productivity (root, shoot, and total). In each GLM all possible interactions including the covariates were included in the initial models but non-significant interactions were removed from the final models (Engqvist 2005).

Results

Soil moisture content was higher with water addition (F1,109 = 12.18, P < 0.001), but was not affected by nitrogen addition (F1,109 = 0.37, P = 0.546) (Fig. 1a). Light transmission through the plant canopy was reduced by both nitrogen (F1,115 = 49.34, P =  < 0.001) and water addition (F1,115 = 12.42, P = 0.001) (Fig. 1b). Shoot biomass increased with both nitrogen (F1,115 = 26.03, P = <0.001) and water addition (F1,115 = 6.09, P = 0.015) (Fig. 1c). In contrast, root biomass increased with water addition (F1,104 = 6.40, P = 0.013) but not nitrogen addition (F1,104 = 0.17, P = 0.677) (Fig. 1d). Nitrogen addition increased relative allocation to shoot biomass, however water addition increased relative allocation to shoot biomass only when root biomass was <500 g m−2 (P = 0.09) and 500–800 g m−2 (P = 0.02), but not when root biomass was 800–1,100 g m−2 (P = 0.85), or >1,100 g m−2 (P = 0.44) (Table 1, Fig. 2). There were no significant water by nitrogen interactions, indicating that nitrogen and water did not interact to affect productivity in this community.
https://static-content.springer.com/image/art%3A10.1007%2Fs11258-006-9222-4/MediaObjects/11258_2006_9222_Fig1_HTML.gif
Fig. 1

Percent soil moisture (a), light transmission through the canopy (b), shoot (c), and root (d) standing biomass (g m−2) in the four resource addition treatments. Error bars are one standard deviation

Table 1

Results from the general linear model testing the effects nitrogen and water addition on shoot biomass with root biomass as a covariate

Source

DF

MS

F

P

Root biomass

1

0.002

0.03

0.873

Nitrogen

1

1.915

23.21

<0.001

Water

1

0.533

6.47

0.012

Nitrogen by water

1

0.002

0.03

0.863

Root biomass by water

1

0.485

5.87

0.017

The initial model included all covariate-fixed factor interactions, but the non-significant interactions were removed from the final model

https://static-content.springer.com/image/art%3A10.1007%2Fs11258-006-9222-4/MediaObjects/11258_2006_9222_Fig2_HTML.gif
Fig. 2

Relationships between root and shoot biomass allocation in the four resource addition treatments

Artemisia seedling survival (90%) was higher than Chenopodium (71%) (χ12 = 6.98, P = 0.008), and nitrogen increased Chenopodium survival (79% vs. 57%; χ12 = 5.96, P = 0.015; nitrogen by species interaction χ12 = 10.43, P = 0.001). No other main effects or interactions were significant (P »0.1), indicating that neither water nor competition altered survival.

Nitrogen increased focal plant biomass, with Chenopodium experiencing larger benefits than Artemisia (Table 2, Fig. 3). Both species were larger when released from root but not shoot competition. Root competition intensity increased with nitrogen addition (significant root competition by nitrogen interaction). Chenopodium was a poorer belowground competitor than Artemisia (significant species by root competition interaction). The only significant effect of water addition was a four way species by nitrogen by water by shoot competition interaction (P = 0.049). This complex interaction appears to indicate that when shoot competition was removed Artemisia increased slightly in biomass under all combinations of the nitrogen and water treatments, while Chenopodium only increased in biomass when one of the resources was added but not when neither or both were added. There were no root by shoot competition interactions indicating that the two modes of competition were independent.
Table 2

Results from the mixed model testing the effects of species, nitrogen and water addition, and root and shoot competition on focal plant biomass

Effect

Num DF

Den DF

F value

P

Species

1

289

5.70

0.018

Nitrogen

1

131

15.12

<0.001

Water

1

131

0.02

0.885

Shoot competition

1

250

3.22

0.074

Root competition

1

261

32.08

<0.001

Nitrogen by water

1

131

0.55

0.459

Nitrogen by shoot

1

250

1.91

0.168

Water by shoot

1

250

0.76

0.385

Nitrogen by root

1

261

12.28

<0.001

Water by root

1

261

0.04

0.844

Species by water

1

289

0.02

0.891

Species by nitrogen

1

289

4.13

0.043

Species by shoot

1

284

0.10

0.750

Species by root

1

291

3.93

0.048

Nitrogen by water by shoot

1

250

1.71

0.192

Nitrogen by water by root

1

261

0.64

0.426

Shoot by root

1

245

2.34

0.127

Water by shoot by root

1

245

0.48

0.487

Nitrogen by shoot by root

1

245

1.43

0.233

Species by nitrogen by water

1

289

0.04

0.840

Species by water by shoot

1

284

1.50

0.221

Species by nitrogen by shoot

1

284

0.50

0.479

Species by water by root

1

291

0.06

0.804

Species by nitrogen by root

1

291

3.25

0.072

Species by shoot by root

1

280

0.17

0.679

Nitrogen by water by shoot by root

1

245

0.80

0.370

Species by nitrogen by water by shoot

1

284

3.90

0.049

Species by nitrogen by water by root

1

291

0.01

0.934

Species by nitrogen by shoot by root

1

280

0.84

0.359

Species by water by shoot by root

1

280

1.67

0.197

Species by nitrogen by water by shoot by root

1

280

2.60

0.108

https://static-content.springer.com/image/art%3A10.1007%2Fs11258-006-9222-4/MediaObjects/11258_2006_9222_Fig3_HTML.gif
Fig. 3

Focal plant aboveground biomass (g) in all combinations of resource and competition treatments. Error bars are one standard deviation. Note the logarithmic scale on the y-axis

Both root and total competition intensity increased with nitrogen addition, but the only significant effect involving water addition was a decline in total competition intensity with shoot biomass as a covariate (Table 3, Fig. 4). There were no significant competition–productivity relationships or nitrogen by water interactions for any measure of competition intensity (Table 3; Fig. 5).
Table 3

Summary of the competition–productivity relationships in this study

Productivity measure

Model terms

SCI

RCI

TCI

Shoot biomass

Water

F1,71 = 0.36

F1,73 = 0.28

F1,80 = 5.71*

Nitrogen

F1,71 = 2.10

F1,73 = 6.86*

F1,80 = 5.91*

Water by nitrogen

F1,71 = 1.40

F1,73 = 1.48

F1,80 = 0.03

Shoot biomass

F1,71 = 0.02

F1,73 = 0.00

F1,80 = 1.50

Root biomass

Water

F1,63 = 0.18

F1,64 = 0.06

F1,72 = 3.09

Nitrogen

F1,63 = 2.35

F1,64 = 7.20**

F1,72 = 8.08**

Water by nitrogen

F1,63 = 1.93

F1,64 = 0.63

F1,72 = 0.00

Root biomass

F1,63 = 0.02

F1,64 = 0.29

F1,72 = 0.11

Total Biomass

Water

F1,63 = 0.13

F1,66 = 0.01

F1,72 = 3.24

Nitrogen

F1,63 = 2.41

F1,66 = 8.52**

F1,72 = 7.91**

Water by nitrogen

F1,63 = 1.98

F1,66 = 0.66

F1,72 = 0.00

Total biomass

F1,63 = 0.08

F1,66 = 1.60

F1,72 = 0.25

F-values are reported for each term in general linear models with z-scores of shoot (SCI), root (RCI), and total (TCI) competition intensity as the response variables, nitrogen and water treatment as fixed factors, and productivity (shoot, root, and total biomass) as covariates. A significant covariate would indicate a significant competition–productivity relationship. The initial models included all covariate-fixed factor interactions, but since these interactions were non-significant they were removed from the final models

*P < 0.05, **P < 0.01

https://static-content.springer.com/image/art%3A10.1007%2Fs11258-006-9222-4/MediaObjects/11258_2006_9222_Fig4_HTML.gif
Fig. 4

Competition intensity measured as z-scores of log-response ratios in the four resource addition treatments. More negative values of the z-scores indicate increasing intensity of competition. Error bars are one standard deviation

https://static-content.springer.com/image/art%3A10.1007%2Fs11258-006-9222-4/MediaObjects/11258_2006_9222_Fig5_HTML.gif
Fig. 5

Competition–productivity relationships for total (TCI), root (RCI), and shoot (SCI) competition intensity measured as z-scores of log-response ratios and root, shoot, and total neighbour biomass. More negative values of the z-scores indicate increasing intensity of competition

Discussion

Both nitrogen and water were limiting to shoot biomass, though only water addition increased root biomass. Allocation to shoot biomass increased with nitrogen addition, consistent with observations that plants allocate relatively less biomass to roots following fertilization (Reynolds and D’Antonio 1996; Poorter and Nagel 2000). Such a change in allocation could also result from resource-driven changes in community composition but this is unlikely given the short duration of this experiment and that most species are long-lived perennials. Water addition increased relative allocation to shoot biomass at low levels of standing root biomass, but did not affect allocation at high root biomass. Most studies have found that, similar to nitrogen, water addition leads to decreased allocation to roots (Poorter and Nagel 2000), though increased root allocation has been observed in sandy soils with minimal water holding capacity (e.g. Pregitzer et al. 1993; Weigelt et al. 2000, 2005). The switch in biomass allocation with water addition could be explained if the plots with high root biomass were on poorer soils with lower water holding capacity than plots with lower root biomass.

The lack of water effects on survival is surprising since many studies have found that both water and nitrogen addition can increase seedling survival (e.g. Bertiller et al. 1996; Davis et al. 1998, 1999; Liancourt et al. 2005). Mortality from water stress is most prevalent among very young seedlings (Bertiller et al. 1996), suggesting that the 3 weeks of watering used to induce germination may have supported seedlings past the stage where they were most vulnerable to water stress. The lack of significant competition or competition–resource interaction effects on survival suggests that competition-induced mortality is likely only important in this system when resources are extremely limited (Cahill 2003a, b).

There were no significant competition–productivity relationships in this study, even though both water and nitrogen directly affected both productivity and competition intensity. The minor role of shoot competition in this system likely explains the lack of shoot competition–productivity relationships, but the reasons for the lack of strong root competition–productivity relationships are less clear. The simple patterns of change in the intensity of root and total competition along resource and productivity gradients predicted by many theories (Newman 1973; Grime 1973, 2001; Tilman 1988) are inadequate to explain these data. These results also differ from recent studies that found patterns of competition intensity and productivity at variance with the prevailing theories (Twolan-Strutt and Keddy 1996; Peltzer et al. 1998; Cahill 1999; Fynn et al. 2005; Sammul et al. 2006). The patterns observed in this study could arise if the competition–productivity relationship is non-linear (Belcher et al. 1995; Arii and Turkington 2001), however given the wide range of patterns found in previous studies, broadly applicable root competition–productivity relationships may not occur. In particular, clear relationships may not occur in communities with very high levels of root biomass, such as the present study. Cahill and Casper (2000) found that root competition intensity saturated at root biomass levels of approximately 300 g m−2 in a productive old field, well below the 400–1,400 g m−2 observed in this study. Saturation of root competition at similar levels in rough fescue grassland could explain why resource addition-induced changes in root biomass have minimal effects on root competition intensity, but leaves open the question of why plants still alter root biomass allocation in response to resource addition. The relationship between competition intensity and fine root biomass may be obscured since 25% or more of the belowground biomass in grasslands can be in organs such as thick roots and rhizomes dedicated to functions other than resource capture (Pucheta et al. 2004), or because it is necessary for plants to overproduce roots to prevent neighbours from gaining advantage in a “Tragedy of the Commons” scenario (Gersani et al. 2001).

Shoot competition was much weaker than root competition, indicating that shoot competition is unlikely to be an important process structuring this plant community. The intensity of shoot competition generally increases with increasing productivity since the degree of shading a plant experiences is closely correlated with the relative size of its neighbours (Tilman 1988; Grime 2001; Keddy 2001). Similar studies in low-statured plant communities have found aboveground competition to be unimportant at shoot biomass levels much in excess of those in this study (e.g. Belcher et al. 1995; Peltzer et al. 1998). Peltzer et al. (1998) suggested that the light penetration to the soil surface in these systems may be sufficient, even at the highest standing biomass levels, to preclude significant shoot competition. In this experiment transmission of PAR was 42% ± 2.25 SD in plots receiving both nitrogen and water. At full sunlight (1,200–1,800 μmol m−1 s−1) leaves at the bottom of the canopy in these plots would receive 500–750 μmol m−1 s−1, well above the photosynthetic compensation point for most plants (Fitter and Hay 2002). The lack of shoot competition and root–shoot competition interactions also supports the prediction that such interactions should not be expected without significant shoot competition (Cahill 1999).

The increase in root competition intensity with nitrogen addition is in direct contradiction to a large body of experimental evidence (Wilson and Tilman 1991, 1993, 1995; Casper and Jackson 1997; Peltzer et al. 1998; Cahill 1999; Schenk 2006). While we do not doubt that in general root competition intensity declines following fertilization, a great deal of variation is masked within the general pattern. For example, Fig. 4 in Wilson and Tilman (1995) shows that of eight species studied, two experienced an increase in root competition intensity with fertilization. In addition, other studies have observed either no change (Cahill 2002a) or an increase in root competition intensity following fertilization (Brewer 2003). This variation could be explained if, similar to the non-linear relationships proposed between competition intensity and productivity (Belcher et al. 1995; Arii and Turkington 2001), the relationship between root competition intensity and resource availability is non-linear. This variation could also be explained through an interaction between root competition and the timing of nitrogen availability. Experimental nitrogen applications can produce pulses of available nitrogen that last for only a few days (e.g. Jackson et al. 1989; Dell and Rice 2005). Given that root competition is size-symmetric (Casper and Jackson 1997; Cahill and Casper 2000; Schenk 2006), nitrogen capture from an ephemeral pulse should be proportional to a plants root system size. Since plant size is likely a function of the degree of root competition it previously experienced (Cahill and Casper 2000), a plant with low root competition should benefit proportionally more from a nitrogen addition than a plant experiencing severe competition. This scenario could cause an increase in competition intensity following nitrogen addition by increasing the size differences between plants with and without competition.

Several studies have shown that water addition can increase productivity without significant effects on the strength of root competition (Burger and Louda 1995; Haugland and Froud-Williams 1999; Semere and Froud-Williams 2001; Weigelt et al. 2005). As with nitrogen there are exceptions, for example the significant declines in intensity with watering in two of nine species combinations found by Weigelt et al. (2005), but the general pattern of invariant root competition intensity along moisture gradients is well supported. In contrast to the clear patterns of root competition, total competition intensity can be highly variable along moisture gradients. This study found a decline in total competition intensity with increasing moisture while other studies have found either increases (e.g. Kadmon 1995; Briones et al. 1998; Corcket et al. 2003) or few effects (e.g. Wetzel and van der Valk 1998; Haugland and Froud-Williams 1999; Novoplansky and Goldberg 2001; Fynn et al. 2005; Liancourt et al. 2005; Weigelt et al 2005). If root competition is invariant, changes in total competition intensity should be correlated with changes in shoot competition intensity, however, the only study to isolate both root and shoot competition along a moisture gradient in the presence of strong shoot competition found no changes in either root or total competition even though shoot competition intensity increased with increasing moisture (Haugland and Froud-Williams 1999). Without more studies in water-limited systems with significant shoot competition, generalizations on shoot and total competition patterns along moisture gradients are difficult to make.

The very different effects of water and nitrogen on root competition intensity suggest that the mechanisms of competition differ between the two resources. The lack of water effects on root competition intensity is contrary to the theory that the addition of a limiting resource will reduce the intensity of competition for that resource by reducing the degree of deficiency (Taylor et al. 1990; Casper and Jackson 1997; Davis et al. 1998). Competition for water remains poorly understood (Casper and Jackson 1997; Schwinning and Weiner 1998), but given that the transpirational demands of competing plants can reduce water availability (e.g. Burger and Louda 1995; Davis et al. 1999) why should plants not compete more strongly when water is limiting? A potential explanation may lie in the different mechanisms of uptake between water and mineral resources. Plants can expend energy to enhance mineral resource uptake through a variety of mechanisms including interception by root growth, increasing bulk flows of water by increasing transpiration, producing more ion uptake enzymes, and by encouraging mycorrhizal associations (Casper and Jackson 1997; Fitter and Hay 2002). In contrast, plants can enhance water uptake only through root growth and by lowering leaf water potential to increase transpiration rates (Fitter and Hay 2002). Constraints on the ability of plants to actively compete for water could explain why competition intensity does not increase as moisture levels fall.

Finally, even though nitrogen and water had independent significant effects on neighbour plant biomass and the intensity of competition, there were few interactions between the two resources. The only significant interaction involving water and nitrogen was a four-way species by nitrogen by water by shoot interaction affecting focal plant biomass. This complex interaction appears to indicate that the two species differed in their responses to shoot competition depending on the resource treatment combination such that removing shoot competition increased Artemisia performance under all combinations of nitrogen and water treatment, while Chenopodium only increased when one of the resources was added but not when neither or both were added. The lack of interactions affecting root competition is surprising given the close links between soil moisture, nitrogen cycling, and plant uptake rates (Fitter and Hay 2002; Booth et al. 2005; James and Richards 2005). The lack of interactions suggests that factors other than soil moisture regulate the availability of nitrogen in this system, and that increased soil moisture promoted growth through mechanisms independent of nitrogen. A wide range of mechanisms could be responsible since water stress can impact plants independently of nutrient availability through mechanisms ranging from reduced CO2 capture to disruptions in xylem and phloem transport and reduced protein synthesis (Fitter and Hay 2002). While this study suggests that water–nitrogen interactions may only be important in systems with close links between nitrogen availability and water, interactions between other belowground resources remain to be evaluated.

Conclusion

This study demonstrates that nitrogen and water addition can have very different effects on root competition intensity, even though both resources limit productivity. The increase in root competition intensity following nitrogen addition suggests that competition for mineral resources is more complex than currently thought and that monotonic declines in competition intensity with increasing resources predicted by theory should not always be expected. The lack of change in root competition intensity following water addition suggests that the mechanisms of competition for water and mineral resources are very different, and that plants may not compete strongly for water even when it is limiting. The lack of significant relationships between competition and productivity in this study likely occurred because shoot biomass levels were too low to cause significant shoot competition and root biomass levels were far above the level at which root competition saturates. Finally, the lack of nitrogen–water interactions suggests that in rough fescue grassland the availability of the two resources are not closely linked.

Acknowledgements

We thank D. Gabruck, S. Roehr, A. Pfeiffer, and N. Fernando for assistance in the field, the Cahill lab group, K. Ketilson, and an anonymous reviewer for helpful suggestions. Financial support came from an Alberta Ingenuity studentship to E.G.L., an NSERC PGS-M scholarship to B.H.S., an Alberta Conservation Association Biodiversity grant to E.G.L. and J.F.C., and an NSERC Discovery grant to J.F.C.

Copyright information

© Springer Science+Business Media, Inc. 2006