Vegetation diversity, growth, quality and decomposition in managed grasslands

Vegetation diversity, growth, quality and decomposition in managed grasslands

Agriculture, Ecosystems and Environment 101 (2004) 73–84 Vegetation diversity, growth, quality and decomposition in managed grasslands Todd A. White ...

326KB Sizes 0 Downloads 7 Views

Agriculture, Ecosystems and Environment 101 (2004) 73–84

Vegetation diversity, growth, quality and decomposition in managed grasslands Todd A. White a,∗ , David J. Barker b,1 , Kenneth J. Moore a a

b

Department of Agronomy, Iowa State University, Ames, IA 50011, USA AgResearch, Grasslands Research Center, Palmerston North, New Zealand

Received 10 October 2002; received in revised form 5 May 2003; accepted 6 May 2003

Abstract The relationship between vegetation growth rate, quality, litter decomposition and diversity was investigated in four grazed grasslands differing in topography, management and fertility in the southern North Island of New Zealand. Vegetation samples were clipped at two sampling times (summer and autumn) from 0.5 m2 areas excluded from grazing. Crude protein (CP), in vitro dry matter digestibility (IVDMD), neutral detergent fiber (NDF), average daily growth (ADG) rate, species composition and the relative extent of tissue degradation by soil microorganisms were determined for the vegetation samples. It was hypothesized that vegetation diversity would be negatively related with measures of vegetation growth, quality and decomposition. Comparisons made across all environmental and management conditions agreed with the hypotheses. Higher vegetation diversity was associated with lower ADG rate, lower CP concentrations, lower IVDMD and higher NDF concentrations. Comparisons within management regimes (MR), however, were less consistent and possible reasons for this are examined. Litter decomposition extent was closely related to the tissue quality characteristics of the vegetation, especially NDF concentration, but not species diversity. The observed local-scale relationships between the vegetation diversity, growth, quality and decomposition characteristics were contended to be largely the result of the influence of environmental resource richness on community structure of the managed grasslands studied. There is a need to promote a functionally diverse plant community over the entire managed area rather than promoting high localized species diversity within grasslands. © 2003 Elsevier B.V. All rights reserved. Keywords: Vegetation diversity; Tissue quality; Litter decomposition; Managed grasslands; New Zealand

1. Introduction In most terrestrial ecosystems, the quantity and quality of plant tissue assimilated greatly influences the density and growth rate of both aboveground con∗ Corresponding author. Tel.: +1-515-294-7952; fax: +1-515-294-5506. E-mail address: [email protected] (T.A. White). 1 Present address: Department of Horticulture and Crop Science, The Ohio State University, Columbus, OH 43210, USA.

sumers and belowground decomposers (Begon et al., 1996). Managed grazing systems are no exception, and research on the determinants of forage growth rate, quality and litter decomposition has focused on intrinsic factors, such as the structural and chemical constituents of plant tissue (Holmes et al., 1987; Nelson and Moser, 1994; Sheaffer et al., 1998), or extrinsic factors, such as temperature, water availability, soil fertility, photoperiod, pests and diseases (Swift et al., 1979; Wilson, 1982; Buxton and Fales, 1994). However, one community level characteristic that has

0167-8809/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0167-8809(03)00169-5

74

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

been largely neglected is vegetation diversity, i.e. the total number of plant species and their relative contribution to a grassland community (Magurran, 1988). Biological diversity has long been considered important to ecosystem health (Odum, 1969) and continues to attract tremendous research effort (Kinzig et al., 2002). Research has focused on understanding the relationships between diversity and ecosystem stability (Tilman, 1996; Wardle et al., 2000), resistance to invasion of exotics (Crawley et al., 1999; Naeem et al., 2000) and, in particular, productivity (Hector et al., 1999; Mittelbach et al., 2001). While many diversity/ecosystem-functioning studies have demonstrated the importance of species number per se (i.e. a positive relationship), other studies attribute greater importance to the functional traits of the dominant species (Wardle et al., 1997b; Crawley et al., 1999). A recent review emphasized the importance of the correct ecological context (environmental characteristics, temporal and spatial scales of observation and human management intensity), to the understanding of how diversity may determine ecosystem productivity, based on two scales of diversity—species pool diversity and local diversity (Fridley, 2001). Species pool diversity (the pool of potential colonizing species) is expected to cause a positive relationship with productivity because diverse species pools are more likely to contain species of higher inherent growth rate (a sampling effect of diversity). On the other hand, local diversity (e.g. at 0.5 m2 scale) is expected to be largely a function of local environmental conditions such as resources and disturbance intensity, a relationship explained by the unimodal “humped-backed” models of Grime (1979) and Tilman (1982). Except in chronically resource poor environments, increases in fertility progressively exclude less competitive species and consequently reduce local species diversity. Ecological principles can be used to generate hypotheses about the local-scale relationship between diversity and quality of grassland vegetation. Comparative experimental screening under laboratory and field conditions has revealed some robust recurring trends between functional traits of plants and their adaptive specialization to the environmental conditions (Diaz and Cabido, 1997; Grime et al., 1997). Species associated with resource rich habitats tend to be fast-growing, morphologically plastic plants with thin short-lived nutrient-rich (N, P, K, Mg, and Ca)

leaves. In contrast, plants common to unproductive environments are often characterized by inherently slow growth rate, have fibrous long-lived thick leaves and high investment in support and storage organs (Grime et al., 1997). These are two broad extremes and while functional trait differences between grassland species may be far subtler, a few nutrient-rich low-fiber digestible species are predicted to dominate patches of fertile soil. These same species would be present in less-fertile high-diversity patches but the greater abundance of nutrient-poor less-digestible species would result in an overall reduction in vegetation quality. Decomposition of plant litter and the resulting cycling of nutrients are key ecosystem processes performed by saprophytic organisms (Swift et al., 1979). Similarly with aboveground ecosystem processes/functions, research into how vegetation diversity may influence these belowground processes has increased in recent years (Wardle, 2002). Both synergistic and antagonistic results have been observed, therefore, opinions often differ on the importance of vegetation species richness to ecosystem functions performed by soil microorganisms (Wardle et al., 1997a; Madritch and Hunter, 2002). Species identity was expected to be more important than species richness to litter decomposition because the tissue quality traits of a community’s dominant species strongly influence litter decomposition (Wardle and Lavelle, 1997; Wardle et al., 2002). This study examined the relationships between vegetation diversity, growth rate, quality and decomposition in lowland and hill-country managed grasslands of New Zealand, in particular, the hypothesis that there is an overall negative relationship between growth, quality and decomposition characteristics of New Zealand agricultural grasslands and vegetation diversity at a local scale.

2. Materials and methods On 10–11 February 1998 eight experimental sites were established in the Manawatu region of New Zealand. Two replicate sites were established within each of four different management regimes (MR) operating at AgResearch Grasslands research farms, Ballantrae (hill-country) and Aorangi (lowland). MR details are given in Table 1 (Lambert et al.,

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

75

Table 1 MR description

Locationa Soil fertilityb Vegetation historyc Major herbivore Grazing typee Stocking ratef (ssu)

MR1

MR2

MR3

MR4

Ballantrae Low Resident Sheep Continuous 10

Ballantrae Medium Resident Sheep Rotational 16

Aorangi High Resident Beef bulls Rotational 16

Ballantrae Medium Improvedd Beef bulls Rotational 16

Ballantrae: 40◦ 20 S, 175◦ 50 E, 125–350 m a.s.l. Aorangi: 40◦ 29 S, 175◦ 41 E, 25 m a.s.l. Rating based on published soil analyses for Ballantrae (Lambert et al., 2000) and Aorangi (Rijkse and Daly, 1972). Annual applications of 13 kg P ha−1 to MR1, 46 kg P ha−1 to MR2, and MR4, 15 kg P ha−1 to MR3; 1250 and 2500 kg ha−1 of ground limestone to MR2 and MR4 in 1975 and 1979, respectively. c Ballantrae, without fertilizer since 1950, was under low intensity continuous sheep grazing and dominated by low-fertility tolerant grasses (A. capillaris, A. odoratum and Cynosurus cristatus). In 1974 Trifolium repens, T. pratense, T. subterraneum and Lotus pedunculatus seed was broadcast over all Ballantrae sites. Aorangi was under permanent pasture for at least 10 years prior to this experiment. d Herbicide (Round-upTM ) applied to MR4 in autumn 1993 at 2 l ha−1 to suppress the existing vegetation. Lime coated D. glomerata cv. Grasslands Wana seed was broadcast from the air at 12 kg ha−1 . An annual application of 150 kg of di-ammonium phosphate per hectare was initiated at this time. e Continuous, animals permanently on site; rotational, higher instantaneous stocking rate for short periods (1–2 days) at intervals from 20 to 80 days. f 1 ssu ha−1 per year (sheep stock unit) = 1 breeding ewe with a lamb to weaning; 1 breeding cow = 6 ssu. a

b

2000 contains full management history of the experimental sites at Ballantrae Hill-country Research Farm). Each experimental site was divided into three blocks and within each block a poor and rich species sampling location (approximately 2 m × 2 m) was identified, giving a total of 48 sampling locations. Visual decision rules were used to identify species rich and poor areas as previous research had determined that certain species were strongly correlated with the species richness of a local area (Nicholas, 1999). At Aorangi low predicted species richness (PSR) areas were identified as being dominated by Lolium perenne L. and at Ballantrae high PSR areas were associated with the indigenous Nertera setulosa Hook.f. and Centella spp. The Ballantrae sites were based on sedimentary soils derived from sandstone, siltstone and mudstones, mainly yellow–brown earths with inter-grades to yellow–grey earths (Lambert et al., 2000). The Aorangi sites were based on Kairanga silt loam—an alluvial soil (Rijkse and Daly, 1972). The average annual rainfall at Ballantrae was 1.2 m and average daily soil temperature (0.1 m depth) ranged from 7 (winter) to 16 ◦ C (summer). Average annual rainfall at Aorangi was 0.9 m and soil temperature ranged from 7 to 18 ◦ C.

2.1. Vegetation measurements A 0.5 m2 exclusion cage (representing a plot) was placed within each low and high diversity area to prevent grazing. Herbage biomass accumulation within each exclusion cage was measured in summer (23 March 1998) and autumn (25 May 1998). At each sampling date motorized hand-shears were used to clip the vegetation within the exclusion cages down to the same height as the surrounding vegetation. Two sub-samples were removed from the clipped vegetation. One sub-sample was dissected into component species to determine botanical composition (dicotyledonous species were generally rare and therefore bulked into a single component). The second sub-sample was analyzed for dry matter, crude protein (CP), neutral detergent fiber (NDF) and in vitro dry matter digestibility (IVDMD; Forage Quality Laboratory, Iowa State University). All herbage material was dried (at 80 ◦ C) and weighed. CP concentration was estimated by multiplying total N concentration by 6.25. Total N concentration was determined by the Kjeldahl procedure where forage samples were digested with sulfuric acid and various catalysts. NDF fraction represented the total cell wall fiber content. Extraction was performed by

76

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

the Detergent Analysis System, where forage samples were boiled in neutral solution, then filtered to isolate the NDF fraction (Collins and Fritz, 2003). IVDMD was a laboratory estimate of the percentage of forage that will be digested if fed to a ruminant. The digestion involved incubating forage samples inoculated with rumen fluid for 48 h (rumen digestion) then with acidified pepsin for 24 h (lower gut digestion). On 23 March 1998, an additional sub-sample of the herbage material clipped from each of the exclusion cages was taken for litter decomposition determination. Four 5 g replicates of dried herbage material were placed in separate 0.2 m × 0.3 m litter bags. The litter bags were made of nylon mesh with a 0.001 m lattice spacing which excluded soil mesofauna. The bags were closed and placed 0.01 m below the soil surface, 0.5 m apart in a four-block randomized grid pattern within a 20 m × 40 m, fenced-off area of MR1. Bags were inserted on 8 April and removed on 7 May 1998—during which time average soil temperature at 0.1 m decreased from 15 to 10 ◦ C and gravimetric moisture increased from 66 to 89%. After excavation the bag contents were washed, dried (at 80 ◦ C, 24 h) and weighed. The residual litter weights were adjusted for any remaining soil contamination—determined by the dry ashing method (Nes, 1975). This involved incinerating samples of clean plant tissue, soil and residual litter in a muffle furnace at 600 ◦ C for 5 h. The percentage of soil contamination in the residual litter was then calculated as Ar − Ap × 100 (1) Soil contamination (%) = As − A p

individuals among species (Shannon diversity index H and evenness of diversity J ):

where Ar is the percentage of ash in residual litter, As the percentage of ash in soil and Ap the percentage of ash in clean plant tissue. Cellulose strips were used to gauge microbial activity in MR1 soil (Kurka and Starr, 1997). Twelve mesh bags, each containing 5 g of cellulose strips, were placed 0.01 m below the soil surface. Four of these bags were removed after 14 days and visually assessed for decomposition. After 28 days, 6.8 ± 1.6% of the cellulose had been degraded.

Legume and other dicotyledonous species generally contributed less to vegetation biomass than grass species (Table 2). Agrostis capillaris L. and L. perenne were the dominant grass species but their content differed significantly between MRs (P = 0.014 and 0.004 for site main effect, respectively). MR1 vegetation was dominated by A. capillaris and Anthoxanthum odoratum L., L. perenne and legumes being lesser components. While the legume content of MR2 was similar to that of MR1, L. perenne abundance was significantly higher, mirrored by a decrease in A. capillaris. Other species making notable contributions to MR1 and MR2 vegetation were Holcus lanatus L. and non-legume dicotyledonous species.

2.2. Data treatment Diversity indices were derived from the botanical composition data to quantify the number of contributing species (species richness) and the distribution of

H =

k 

pi (ln pi )

(2)

i=1

J =

H ln k

(3)

where k is the number of species and pi the proportion of individuals (i.e. tillers or growing points) belonging to species i. pi is estimated using ni /N, where ni is the number or biomass of individuals of species i and N is the total number or biomass of individuals. Vegetation diversity, quality, composition and decomposition data were analyzed as a factorial design by analysis of variance using SAS® statistical package (release 8.01). Replicates were treated as nested within MR. Significant differences between main effect means were determined by calculating Tukey test statistics for pair-wise comparisons (Zar, 1996). Significant differences between means within MR were determined using the SLICE option of the LSMEANS statement within SAS® . Relationships between species richness, vegetation quality, litter decomposition and growth rate were examined by regression analysis. Comparisons were made across all MRs and within MRs, separated by season.

3. Results

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

The vegetation composition of MR3 was quite different to that of the hill-country with L. perenne highly dominant, and legume content significantly greater than that of MR1 (P = 0.018 for legume site main effect). Similarly, in contrast to MR1 and MR2, MR3 vegetation contained no A. odoratum and very few non-legume dicotyledonous species. MR4 species composition reflected a previous D. glomerata L. establishment experiment. Other major contributors were legumes (mostly Trifolium repens L.) and L. perenne. Non-legume dicotyledonous species made only a minor contribution to MR4 vegetation. There were significant differences in vegetation composition between low and high PSR areas within MR. These differences were not always consistent between MR, for example, the content of A. capillaris was significantly greater in low than high PSR areas for MR1 but greater in high than low PSR areas for MR2 and MR4 (Table 2; P < 0.01 for MR by PSR effect sliced by MR). Where A. odoratum and non-legume dicotyledonous species were present,

77

their abundance was consistently higher in high rather than low PSR areas (P < 0.05 for MR by PSR effect sliced by MR). By contrast, L. perenne abundance did not differ between low and high PSR areas within MR. The exception was significantly higher L. perenne abundance in low PSR areas for MR2 (P < 0.001). Vegetation composition also varied with season, legume and other dicotyledonous species decreasing significantly between summer and autumn (P = 0.016 and 0.048 for season main effect, respectively). There was no significant interaction effect between MR, PSR and season for vegetation composition data. 3.1. Vegetation diversity The number of species within the sampling areas ranged from 3 to 11. On average, the lowland rotationally grazed vegetation (MR3) contained significantly less species than all the other MRs (P = 0.001, MR main effect) (Fig. 1a). Hill-country MR did not significantly differ in species richness when tested at the

Table 2 Composition of vegetation under four management regimes (MR) during summer (S) and autumn (A), expressed as a percentage of live biomass (±S.E.M)a MR

PSR

S/A

Aca

Aod

Ccr

Dgl

Hla

Lpe

Poa

OG

1

L

S A

58 ± 13 58 ± 5

11 ± 2 14 ± 1

3 5

2 0.7

12 ± 1 8±3

5±1 10 ± 2

0.1 0.4

3±2 0.9 ± 1

2±1 2±1

H

S A

33 ± 5 36 ± 3

16 ± 3 20 ± 2

6 9

0 0

5±1 5±2

8±2 14 ± 4

0 0.1

2±1 3±2

9±4 5±2

L

S A

26 ± 9 22 ± 9

4±2 8±3

1 3

0 0

12 ± 5 6±3

42 ± 3 50 ± 8

2 5

0 0

6±1 5±1

H

S A

34 ± 5 29 ± 4

12 ± 3 20 ± 4

6 5

0 0

5±1 4±1

23 ± 7 31 ± 7

0.5 1

0.4 ± 0.4 0

7±1 5±1

L

S A

10 ± 3 12 ± 5

0 0

0.1 0

0 0

0 0.1

63 ± 6 74 ± 7

0.4 1

0

H

S A

10 ± 3 11 ± 3

0 0

0 0

0 0

2±2 1±1

59 ± 6 64 ± 9

2 0.9

L

S A

6±3 3±3

0 2

0 0

62 ± 6 67 ± 9

6±3 5±3

16 ± 4 11 ± 4

H

S A

14 ± 4 12 ± 3

2 4

1 2

28 ± 7 33 ± 7

8±2 6±2

19 ± 6 18 ± 5

2

3

4

a

DdM

ADG

16 ± 2 15 ± 3

30 ± 4 21 ± 1

15 ± 2 18 ± 5

22 ± 2 15 ± 2

21 ± 3 8±2

56 ± 8 30 ± 3

13 ± 4 4±1

15 ± 2 10 ± 3

30 ± 4 23 ± 2

27 ± 3 12 ± 3

0.4 ± 0.2 0.1 ± 0.1

15 ± 3 9±1

69 ± 1 38 ± 1

15 ± 5 14 ± 4

13 ± 4 10 ± 4

0.3 ± 0.2 0.3 ± 0.3

17 ± 2 12 ± 3

51 ± 4 37 ± 2

0.5 1

0.01 0.03

9±4 11 ± 4

0 0.1 ± 0.1

20 ± 2 8±3

53 ± 6 33 ± 3

0.9 1

0.1 ± 0.1 0.3 ± 0.2

24 ± 3 20 ± 3

3±1 3±2

26 ± 3 11 ± 3

34 ± 5 29 ± 2

1 ± 0.6

Leg

DW 3±1 1 ± 0.5 21 ± 7 8±3 8±3 1 ± 0.3

PSR, predicted species richness (L, low; H, high); Aca, Agrostis capillaris; Aod, Anthoxanthum odoratum; Ccr, Cynosurus cristatus; Dgl, D. glomerata; Hla, Holcus lanatus; Leg, legumes (predominately T. repens, but also T. dubium and T. subterraneum and L. pedunculatus); Lpe, Lolium perenne; Poa, Poa spp (P. annua, P. pratensis); OG, other grasses; DW, dicotyledonous species other than legumes; DdM, dead matter, as percentage of total biomass; ADG, average daily growth rate, as kg dry matter per hectare per day.

78

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

(Fig. 1; P < 0.01 for MR by PSR effect sliced by MR for species richness, H and J ). There was no significant difference in species richness, diversity and evenness between sampling times nor any significant interaction effects. 3.2. Vegetation quality

Fig. 1. Measures of vegetation diversity for summer (䊏) and autumn (䊐) sampling under four management regimes (MR) and low (L) and high (H) PSR (±S.E.M.).

5% probability level. Mean diversity measured by the Shannon index was not significantly lower for MR3 than for the hill-country MR (Fig. 1b; P = 0.123). There was no significant difference in vegetation evenness between MRs (Fig. 1c). Within each MRs, sampling locations predicted to have a higher number of species, actually contained more species, were more diverse as measured by the Shannon diversity index and had more even distribution of species abundance

Vegetation CP content did not differ between MRs but was significantly higher in autumn than summer (Fig. 2a, P = 0.004 for season main effect). For within MR comparisons, CP content was generally higher in low PSR than high PSR vegetation (P < 0.05 for MR by PSR effect sliced by MR). However, while both summer and autumn vegetation of MR3 and MR4 had significantly more CP in low compared to high PSR areas, this was only true for the autumn vegetation of MR1 and MR2 (P = 0.002 for MR by PSR by season interaction effect). IVDMD was on average significantly higher for lowland vegetation (MR3) than for hill-country vegetation (MR1, MR2, and MR4 were not significantly different) (Fig. 2b, P = 0.042 for MR main effect). As with CP, there was a significant MR by PSR by season interaction effect for IVDMD (P = 0.006). MR3 and MR4 summer and autumn vegetation was significantly more digestible in low compared to high PSR areas. For MR1 and MR2, only high PSR vegetation from the autumn sampling was significantly less digestible than low PSR vegetation. NDF content (Fig. 2c) was on average lower for MR3 vegetation than for MR1, MR2 and MR4 (P = 0.0118 for MR main effect). However, like the other measures of vegetation quality there was a significant MR by PSR by season interaction effect for NDF (P = 0.012). Differences in NDF content between PSR areas and cutting dates differed between MR. The NDF content of low PSR vegetation was generally either the same or greater than high PSR vegetation. The exceptions were MR1 autumn and MR3 summer vegetation from high PSR areas which had greater NDF concentration. Consistent trends were apparent when the vegetation quality data were plotted against species richness (Fig. 3a–f; similar relationships were observed when Shannon H was used instead of species richness). When comparisons were made across all MR, species richness was significantly and negatively related to

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

79

Fig. 2. Measures of summer (䊏) and autumn (䊐) vegetation quality and litter dry matter degradability under four management regimes (MR) and low (L) and high (H) PSR (±S.E.M.).

MR content and in vitro dry matter digestibility and positively related to autumn NDF content. In all comparisons the proportion of the total variation in the data explained by the regression sums of squares was higher in autumn than summer. In the case of NDF, this meant a change from a non-significant to a significant relationship. The only within MR relationship between summer vegetation quality and species richness was a significant negative relationship between NDF and species richness within MR4 (Fig. 3e). At the autumn sampling, MR1 consistently agreed with the overall relationships observed between quality and species richness. Regardless of season there was no significant relationship between any of the quality measures and species richness for vegetation sampled from MR2 and MR3.

3.3. Vegetation growth rate and litter decomposition Average daily growth (ADG) rate only differed significantly between MR1 and MR3 (Table 2, P = 0.030 for MR main effect). For each MR, the ADG of low PSR was significantly higher than that of high PSR vegetation (P < 0.05 for MR by PSR effect sliced by MR). Summer growth rates were significantly higher than autumn growth rates (P = 0.003 for season main effects). Growth rate was negatively related to species richness in both summer and autumn (Fig. 3g and h; comparison across all MR). Within MR this relationship between growth rate and species richness was only observed for MR1 (autumn sampling) and MR4 (both summer and autumn sampling).

80

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

Fig. 3. Relationship of vegetation quality and growth rate with species richness (SR) during summer and autumn (CP, crude protein; IVDMD, in vitro dry matter digestibility; NDF, neutral detergent fiber; ADG, average daily growth rate) under four MRs: MR1 (䉬); MR2 (䊊); MR3 ( ); MR4 (×). Regression lines: thin, within MR; thick, over all MR. Overall R2 from regression analysis given in top left hand corner and R2 for significant within MR relationships given in top right hand corner (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

There was no overall significant difference in the extent of litter decomposition between MRs (Fig. 2d; P > 0.05 for MR main effect). There were, however, significant differences between PSR levels within MR. Litter from MR1’s high PSR plots was significantly more degradable than low PSR vegetation, the reverse being true for MR3 (Fig. 2d; P < 0.05 for MR by PSR effect sliced by MR). Litter decomposition for MR2 and MR4 did not significantly differ between PSR levels.

81

There was good agreement between the extent of litter decomposition and the CP content, in vitro dry matter digestibility and in particular, the NDF content of the vegetation from which the litter was derived (Fig. 4a–c). There was, however, no consistent relationship between diversity and litter decomposition. This was true whether comparisons were made between measured diversity and decomposition of over all MR (Fig. 4d–f) or between PSR and decomposition within MR (Fig. 2d).

Fig. 4. Relationship between vegetation quality, vegetation diversity and litter dry matter degradability (CP, crude protein; IVDMD, in vitro dry matter digestibility; NDF, neutral detergent fiber). R2 from regression analysis given in top left hand corner (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

82

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

4. Discussion The present results support the hypothesized negative relationship between diversity and both vegetation growth and quality. Across all MR, higher diversity was associated with lower ADG rates, lower MR concentrations, lower in vitro dry matter digestibility gand higher NDF concentrations. In the grassland communities studied diversity did not give rise to productivity (cf. Kareiva, 1994), rather diversity and productivity were both born from the interplay between abiotic and biotic factors. The perceived relationship between diversity, growth and quality was due to the dominant species and their functional characteristics. Lower diversity communities were consistently associated with higher fertility environments (Moore and Barker, 1999) and consistently dominated by the same species, in particular, L. perenne and T. repens (Table 2). Functional attributes, such as high morphological plasticity, low structural fiber, high nutrient content and high inherent growth rate (Grime et al., 1988), enabled these competitive species to dominate higher fertility sites and imparted high quality characteristics on the vegetation as a whole. As species diversity increased with decreasing fertility, the productive and nutritive qualities of species such as L. perenne and T. repens were diluted by increased abundance of more inherently slow-growing stress tolerant species, such as A. capillaris and A. odoratum (Table 2), resulting in a reduction of vegetation quality. Comparisons within MRs did not always agree with overall trends. Hill-country, medium fertility, rotationally grazed by sheep (MR2) and low flat country, high fertility, rotational grazed by cattle (MR3) vegetation failed to display any significant relationship between diversity and growth rate or diversity and quality (Fig. 3). Conversely, hill-country, low fertility, continuously grazed by sheep (MR1) and to a lesser extent hill-country, medium fertility, rotationally grazed by cattle (MR4) vegetation agreed with overall trends. Significant relationships were observed across different MRs that varied broadly in vegetation, soils and climate. It is therefore conceivable that the extent to which environmental conditions and vegetation vary within a MR will also be important to observing significant relationships between the diversity, quality and growth characteristics of its vegetation.

In an Australian sheep grazing study the vegetation within unfertilized plots was more heterogeneous than that of phosphate fertilized plots (Taylor et al., 1985). MR1 was of considerably lower fertility than the other MRs but there was no conclusive evidence to suggest that MR1 vegetation exhibited greater heterogeneity. The present results indicate a close association between diversity, quality and growth rate under low fertility environments. Vegetation growth and quality appeared to be more sensitive to changes in diversity under low fertility than high fertility. The leaf litter decomposition component of this study was a test of the influence of differences in community structure characteristics between MRs on litter decomposition. The divergent relationship between PSR and litter decomposition observed for MR1 and MR3 most probably resulted from species composition and not from species diversity differences between low and high PSR vegetation. For MR1, low PSR litter was less degradable than high PSR litter because of 25% more A. capillaris and 18% less non-legume dicotyledonous species (Table 2). Non-legume dicotyledonous species are often of similar or higher leaf tissue quality than many common forage grass species, including A. capillaris (Ulyatt, 1981; Haugland, 1995; Wardle et al., 1997a). Furthermore, it is likely that high PSR litter was less degradable than low PSR litter from MR3, because of less high-quality legume T. repens and more low-quality grass Elytrigia repens (L.) Beauv. (Ulyatt, 1981; Martineau et al., 1994).

5. Implications and conclusions This study highlighted the local-scale relationships between vegetation diversity, growth, quality and decomposition in New Zealand grasslands and these relationships were associated with differences in environmental resource richness. According to Hobbie (1992), resource poor environments tend to be species rich, with more slow-growing, low-quality species, which contribute low-quality litter to the soil ecosystem that retards microbial decomposition and, therefore, nutrient mineralization. Conversely, resource rich environments tend to be dominated by one or two fast-growing high-quality species that

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

contribute high quality litter to the soil, stimulating decomposition and mineralization. Accordingly, there appears to be few reasons to encourage diverse plant communities in production-orientated forage systems. As already stated by Mitchley (2001), “there can be no easy compromise between productivity and diversity—grasslands can be productive in agricultural terms, or diverse in wildlife terms, but not both”. The ecosystem relationships observed in this study were related to a 0.5 m2 scale. In grasslands managed for grazing animal production, management decisions are applied at increasing spatial and temporal scales. Grassland managers not only aim to maximize growth and quality of vegetation over their entire grazing area but also through time (Stuth and Maraschin, 2000). As spatial and temporal scales increase so does the need for greater vegetation diversity, particularly where there is considerable variability in soil type, slope, elevation, aspect and climatic conditions. Specifically, plant functional trait diversity, rather than species diversity per se, needs to be greater at larger scales. Greater and more appropriate use could be made of functional trait differences between agronomic species of the Fabaceae family. Legumes are keystone species in grassland communities because of their symbiotic associations with nitrogen-fixing Rhizobiuim bacteria (Spehn et al., 2002). This makes them highly desirable forage species because of their ability to enhance forage growth and quality. Many legumes have differing phenologies promoting temporal differences in growth patterns or in drought, cold or heat tolerance. For example, disease and low oxygen tolerance enables persistence under water-logged conditions in Lotus corniculatus L. and a tap-root promotes drought tolerance in Medicago sativa L. (Barnes and Sheaffer, 1995; Beuselinck and Grant, 1995). The key message is to promote a functionally diverse plant community over the entire managed area rather than a high localized species diversity within grasslands. Species introductions need to be “site-specific” where plant functional traits are matched with the environmental conditions that vary within the entire grazing area. Improving the accuracy and ease of identifying particular “iso-management zones” and applying the appropriate site-specific management should be a research priority.

83

Acknowledgements The authors would like to thank Nicholas Dymock for technical assistance, Yvonne Gray and the staff of the AgResearch Grassland Herbage Laboratory for processing the plant material, AgResearch Grassland Research Farm managers John Napier (Ballantrae) and Derek Saggar (Aorangi) and their staff for providing assistance with the field sites. We also thank Andrew Carran for his advice and help with the litter decomposition aspect of this experiment and two anonymous reviewers, whose comments improved the paper greatly. An AgResearch Senior Research Fellowship funded KJM. Additional funding was provided by the Foundation for Research, Science and Technology and Iowa State University.

References Barnes, D.K., Sheaffer, C.C., 1995. Alfalfa. In: Barnes, R.F., Miller, D.A., Nelson, C.J. (Eds.), Forages: An Introduction to Grassland Agriculture. Iowa State University Press, Ames, pp. 205–216. Begon, M., Harper, J.L., Townsend, C.R., 1996. Ecology: Individuals, Populations and Communities. Blackwell Scientific Publications, Oxford. Beuselinck, P.R., Grant, W.F., 1995. Birdsfoot trefoil. In: Barnes, R.F., Miller, D.A., Nelson, C.J. (Eds.), Forages: An Introduction to Grassland Agriculture. Iowa State University Press, Ames, pp. 237–248. Buxton, D.R., Fales, S.L., 1994. Plant environment and quality. In: Fahey Jr., G.C. (Ed.), Forage Quality, Evaluation, and Utilization. American Society of Agronomy, Madison, pp. 155–199. Collins, M., Fritz, J.O., 2003. Forage quality. In: Barnes, R.F., Nelson, C.J., Collins, M., Moore, K.J. (Eds.), Forages: An Introduction to Grassland Agriculture. Iowa State Press, Ames, pp. 363–390. Crawley, M.J., Brown, S.L., Heard, M.S., Edwards, G.R., 1999. Invasion-resistance in experimental grassland communities: species richness or species identity? Ecol. Lett. 2, 140–148. Diaz, S., Cabido, M., 1997. Plant functional types and ecosystem function in relation to global change. J. Veg. Sci. 8, 463–474. Fridley, J.D., 2001. The influence of species diversity on ecosystem productivity: how, where, and why? Oikos 93, 514–526. Grime, J.P., 1979. Plant Strategies and Vegetation Processes. Wiley, Chichester. Grime, J.P., Hodgson, J.G., Hunt, R., 1988. Comparative Plant Ecology. A Functional Approach to Common British Species. Unwin Hyman, London. Grime, J.P., Thompson, K., Hunt, R., 1997. Integrated screening validates primary axes of specialisation in plants. Oikos 79, 259–281.

84

T.A. White et al. / Agriculture, Ecosystems and Environment 101 (2004) 73–84

Haugland, E., 1995. Rumex longifolius DC., Ranunculus repens L. and Taraxacum officinale (Web.) Marss. in grassland. 2. Crop nutritive value in relation to proportion of dicots. Norw. J. Agr. Sci 9, 85–93. Hector, A., Schmid, B., Beierkuhnlein, C., et al., 1999. Plant diversity and productivity experiments in European grasslands. Science 286, 1123–1127. Hobbie, S.E., 1992. Effects of plant species on nutrient cycling. Trends Ecol. Evol. 7, 336–339. Holmes, C.W., Wilson, G.F., Mackenzie, D.D.S., Flux, D.S., Brookes, I.M., Davey, A.W.F., 1987. Milk Production From Pasture. Butterworths, Wellington, New Zealand. Kareiva, P., 1994. Diversity begets productivity. Nature 368, 686– 687. Kinzig, A.P., Pacala, S., Tilman, D., 2002. The Functional Consequences of Biodiversity. Princeton University Press, Princeton, NJ. Kurka, A.M., Starr, M., 1997. Relationship between decomposition of cellulose in the soil and tree stand characteristics in natural boreal forests. Plant Soil 197, 167–175. Lambert, M.G., Clark, D.A., Mackay, A.D., Costall, D.A., 2000. Effects of fertiliser application on nutrient status and organic matter content of hill soils. NZ J. Agric. Res. 43, 127–138. Madritch, M.D., Hunter, M.D., 2002. Phenotypic diversity influences ecosystem functioning in an oak sandhills community. Ecology 83, 2084–2090. Magurran, A.E., 1988. Ecological Diversity and its Measurement. Chapman & Hall, London. Martineau, Y., Leroux, G.D., Seoane, J.R., 1994. Forage quality, productivity and feeding value to beef cattle of quackgrass (Elytrigia repens (L.) Nevski.) compared with timothy (Phleum pratense L.). Anim. Feed Sci. Technol. 47, 53–60. Mitchley, J., 2001. Species diversity in grasslands. In: Shiyomi, M., Koizumi, H. (Eds.), Structure and Function in Agroecosystem Design and Management. CRC Press, Boca Raton, pp. 45–59. Mittelbach, G.G., Steiner, C.F., Scheiner, S.M., Gross, K.L., Reynolds, H.L., Waide, R.B., Willig, M.R., Dodson, S.I., Gough, L., 2001. What is the observed relationship between species richness and productivity? Ecology 82, 2381–2396. Moore, K.J., Barker, D.J., 1999. Hill country biodiversity, production and quality. American Society of Agronomy. Abstracts, pp. 149. Naeem, S., Knops, J.M.H., Tilman, D., Howe, K.M., Kennedy, T., Gale, S., 2000. Plant diversity increases resistance to invasion in the absence of covarying extrinsic factors. Oikos 91, 97– 108. Nelson, C.J., Moser, L.E., 1994. Plant factors affecting forage quality. In: Fahey Jr., G.C. (Ed.), Forage Quality, Evaluation, and Utilization. American Society of Agronomy, Madison, pp. 115–154. Nes, P., 1975. A correction for soil contamination of herbage samples by dry ashing. NZ J. Agr. Res. 18, 69–71. Nicholas, P.K., 1999. Environmental and management factors as determinants of pasture diversity and production of North Island, New Zealand hill pasture systems. Ph.D. Thesis. Massey University, Palmerston North.

Odum, E.P., 1969. The strategy of ecosystem development. Science 164, 262–270. Rijkse, W.C., Daly, B.K., 1972. Soils of Aorangi experimental farm, Manawatu District, New Zealand. NZ J. Agric. Res. 15, 117–136. Sheaffer, C.C., Seguin, P., Cuomo, G.J., 1998. Sward characteristics and management effects on cool-season grass forage quality. In: Cherney, J.H., Cherney, D.J.R. (Eds.), Grass for Dairy Cattle. CAB International, Wallingford, pp. 73–100. Spehn, E.M., Scherer-Lorenzen, M., Schmid, B., et al., 2002. The role of legumes as a component of biodiversity in a cross-European study of grassland biomass nitrogen. Oikos 98, 205–218. Stuth, J., Maraschin, G.E., 2000. Sustainable management of pasture and rangelands. In: Lemaire, G., Hodgson, J., Moraes, A.D., Nabinger, C., Carvalho, P.C.D.F. (Eds.), Grassland Ecophysiology and Grazing Ecology. CAB International, Wallingford, pp. 339–354. Swift, M.J., Heal, O.W., Anderson, J.M., 1979. Decomposition in Terrestrial Ecosystems. Blackwell Scientific Publications, Oxford. Taylor, J.A., Hedges, D.A., Whalley, R.D.B., 1985. Effects of fertilizer and grazing sheep on pasture heterogeneity in a small-scale grazing experiment. Aust. J. Agr. Res. 36, 315–325. Tilman, D., 1982. Resource Competition and Community Structure. Princeton University Press, Princeton, NJ. Tilman, D., 1996. Biodiversity: population versus ecosystem stability. Ecology 77, 350–363. Ulyatt, M.J., 1981. The feeding value of herbage: can it be improved? NZ Agr. Sci. 15, 200–205. Wardle, D.A., 2002. Communities and Ecosystems: Linking the Aboveground and Belowground Components. Princeton University Press, Princeton, NJ. Wardle, D.A., Lavelle, P., 1997. Linkages between soil biota, plant litter quality and decomposition. In: Cadisch, G., Giller, K.E. (Eds.), Driven by Nature—Plant Litter Quality and Decomposition. CAB International, Wallingford, pp. 107–124. Wardle, D.A., Bonner, K.I., Nicholson, K.S., 1997a. Biodiversity and plant litter: experimental evidence which does not support the view that enhanced species richness improves ecosystem function. Oikos 79, 247–258. Wardle, D.A., Zackrisson, O., Hornberg, G., Gallet, C., 1997b. The influence of island area on ecosystem properties. Science 277, 1296–1299. Wardle, D.A., Bonner, K.I., Barker, G.M., 2000. Stability of ecosystem properties in response to above-ground functional group richness and composition. Oikos 89, 11–23. Wardle, D.A., Bonner, K.I., Barker, G.M., 2002. Linkages between plant litter decomposition, litter quality, and vegetation responses to herbivores. Funct. Ecol. 16, 585–595. Wilson, J.R., 1982. Environmental and nutritional factors affecting herbage quality. In: Hacker, J.B. (Ed.), Nutritional Limits to Animal Production from Pastures. Commonwealth Agricultural Bureaux, Slough, pp. 111–131. Zar, J.H., 1996. Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, NJ.