Structure and Function in Agroecosystem Design and Management - Chapter 14 ppt

Altering the frequency and severity of defoliation can have profound effects on the dynamics of grassland systems Parsons et al., 1988, and dif-ferences in the soil and plant properties

CHAPTER 14

The Effect of Elevated Atmospheric

Paul C.D Newton, Harry Clark, and Grant R Edwards

CONTENTS

Introduction 297

Properties of Grazed Pasture Ecosystems 298

Nutrient Cycling 299

Physical Effects 303

Preference 306

Conclusions 308

References 308

INTRODUCTION

Grasslands cover about a fifth of the terrestrial surface of the world (Hadley, 1993), and the majority of this area is grazed by animals The impact

of an increasing concentration of CO2in the atmosphere on these grasslands has assumed importance, first because of the direct effects on food produc-tion (Gregory et al., 1999), and second because of the influence terrestrial ecosystems can have on the composition of the atmosphere and therefore on our climate (Pielke et al., 1998) In the case of grasslands this includes not only C sequestration, N2O release, and CH4uptake by soils, but also CH4 emissions from ruminants Consequently, many research programs have been developed to explore these impacts, and our knowledge of the likely outcomes is progressing rapidly However, our understanding is based almost exclusively on cut (as opposed to grazed) grassland (e.g., Wolfenden

297

0-8493-0904-2/01/$0.00+$.50

and Diggle, 1995; Casella et al., 1996a,b; Newton et al., 1996; Clark et al., 1997; Hebeisen et al., 1997; Potvin and Vasseur, 1997; Taylor and Potvin, 1997; Clark

et al., 1998; Leadley et al., 1999; Navas et al., 1999, but see Edwards et al., 2000), and grazed swards are very different in their botanical and soil char-acteristics (Watkin and Clements, 1976; Haynes and Williams, 1993) In addi-tion, some of these experiments involved the transfer of previously grazed areas to a cutting management (Newton et al., 1996; Clark et al., 1997; Potvin and Vasseur, 1997; Taylor and Potvin, 1997; Clark et al., 1998; Leadley et al., 1999) and therefore do not necessarily display the responses typical of a cut system but of a system in transition In these examples it is probable that the change in management resulted in a process of succession, one consequence

of which would likely be a loss of early successional species Clearly, any interpretation of a response to elevated CO2 in these transitional systems must be made with this background change in mind

Altering the frequency and severity of defoliation can have profound effects on the dynamics of grassland systems (Parsons et al., 1988), and dif-ferences in the soil and plant properties of cut and grazed swards can often (in part) be attributed to difference in the timing and severity of harvesting However, such comparisons conceal the intrinsic effects introduced by graz-ing animals Consequently, in this chapter we concentrate on compargraz-ing cut-ting with grazing at the same frequency and severity of defoliation We are concerned with identifying characteristics introduced by grazing animals that have the potential to alter how pastures might respond to elevated CO2 While the question of CO2 grazing interactions has been raised previously (Wilsey, 1996), we are not aware of any comprehensive treatment of this sub-ject Without considerations of different responses of cut and grazed swards

to elevated CO2we are not in a position to extrapolate from the considerable bulk of existing experimental data to grazed grasslands—the predominant pastoral land use

Much of what we present is based on temperate pastures; this does not imply any special importance of this type of grassland but simply reflects that these systems have been more extensively examined in terms of CO2 effects than any other grassland type, and there is a long and detailed litera-ture on responses of these ecosystems to grazing from which we can draw general principles

PROPERTIES OF GRAZED PASTURE ECOSYSTEMS

Despite the common practice of using cutting to simulate grazing by animals, there are clear differences in the ecosystem properties which can be directly related to these managements The actions that the grazing animal introduces involve nutrient cycling, physical damage to plants and soils, and selective grazing

Table 14.1 Proportion of Nutrients Returned by Milking

Cows Grazing at Pasture.

From Hutton, J.B., Jury, K.E., and Davies, E.B., NZ J Ag Res.,

10:367 –388, 1967.

Nutrient Cycling

Grazing animals return nutrients to the pasture, and it is in the composi-tion and spatial arrangement of these nutrient returns that lies the major dif-ference between cut and grazed systems (Haynes and Williams, 1993) Animals use only a small proportion of the nutrients they ingest; 60–99% are returned to the pasture as dung and urine (Barrow, 1987) The actual amounts returned are dependent on the species of animal and the stage of its devel-opment (Haynes and Williams, 1993) Some typical values for dairy cows are shown in Table 14.1 There are some differences between animal types in the proportion of nutrients returned; for example, sheep return greater amounts

of N in the urine than cattle (about 70–75% of the excreted N in urine; Sears

et al., 1948) However, as a general principle, the concentration returned depends upon the concentration in the food ingested In the case of the N in urine and the P in dung, the relationship with the feed composition is linear (Barrow and Lambourne, 1961)

Unlike cutting, which removes nutrients from the whole of a paddock and then returns them evenly by fertilizer application, grazing removes nutrients from the whole paddock but returns them heterogeneously in the excreta A sheep may have 18–20 urinations in a day, each event returning nutrients over an area of 0.03–0.05 m2 A typical value for cattle would be 8–12 urinations, each covering an area 0.16–0.49 m2(Haynes and Williams, 1993) In the excretal areas, the nutrients are at very high concentrations (Table 14.2) There are three consequences of this localized return at high con-centrations First, the pasture becomes a mosaic of patches ranging from very high to very low nutrient status The outcome of such a distribution is that pasture growth is very high in a small area of the paddock; for example, Saunders (1984) found that under cattle grazing, 75% of the dry matter was produced from 38% of the pasture area Second, losses of nutrients through gaseous emissions, leaching, and runoff are all exaggerated in the high nutri-ent patches Third, plants in the excretal areas can be damaged or killed (e.g.,

by urine scorch or buried under dung) making immediate recovery of the nutrients less likely (Haynes and Williams, 1993) In addition, animals may

Figure 14.1 Potential N response curves at ambient (solid line) and elevated

(dashed line) CO

Table 14.2 Typical Application Rates of Nutrients (kg ha 1 )

Contained in Single Urination or Defecation Events

Data sources and assumptions given in Haynes, R.J and Williams, P.H., Adv Agon.,

49:119 –190, 1993.

avoid grazing areas close to excreta (Haynes and Williams, 1993) resulting in ungrazed patches of herbage that may be at ceiling yield interspersed with grazed areas in the early stages of regrowth

The long-term effects of excretal return are to increase organic matter (C and N) storage largely because of the return of organic matter as dung (Carran and Theobald, 1998) This outcome, that grazing management can influence the equilibrium organic matter content (Haynes and Williams, 1993), has important implications for C storage and therefore greenhouse gas emissions from pasture One negative consequence of grazing is the lower soil Ca and Mg contents due to the high rate of cycling of K through excretal returns (Carran and Theobald, 1998)

How might we expect pasture response to elevated CO2to be modified

by grazing-mediated changes in biogeochemical cycles? The distribution of nutrients into high and low patches is the characteristic that has the greatest potential to interact with CO2 Let us consider a hypothetical example of the distribution of N which gives the same average application of 240 kg ha1in both cut and grazed swards, but in the cut sward, the N is distributed evenly and in the grazed sward it is at a rate equivalent to 1000 kg ha1in 20% of the area and at 50 kg ha1in 80% of the area If plant responses to N are linear at both ambient and elevated CO2(Figure 14.1a), then it makes no difference to

Figure 14.2 N response curves for Lolium perenne plants grown at 390 or 690 ppm

CO (From Schenk et al., J Pl Nutr 19:1423 –1440, 1996.)

the response whether the N is distributed homo- or heterogeneously Consequently, despite a strong response to elevated CO2, there is no differ-ence in response between a cut (homogeneous N) and grazed (heterogeneous N) management If there is no CO2 effect—despite nonlinear N response curves (Figure 14.1b)—then it makes no difference whether the pastures are cut or grazed However, if the plant/community responses to N are nonlin-ear, and if they are different between CO2levels (Figure 14.1c), then the rela-tive responses to CO2will depend on the nutrient distribution; i.e., they will

be different depending on whether the pasture is cut or grazed and the man-ner of the difference will depend on the relative shape of the curves In fact, nonlinear curves of the kind shown in Figure 14.1c are frequently seen in experimental data for a range of variables, such as dry matter (Schenk et al., 1996; Figure 14.2), photosynthesis (Bowler and Press, 1996), and competitive ability (Navas et al., 1999) Obviously the argument made for N can also apply to other nutrients that are returned in high concentration by animals (e.g., P, K, or S) and for which there are likely to be nonlinear response curves and CO2 nutrient interactions

Soussana and Hartwig (1996) have described the consequences of ele-vated CO2 for N cycling in cut systems but were not able to speculate on aboveground transfer of N by grazing animals at elevated CO2due to a lack

Figure 14.3 Nitrogen content (%) of the leaves of legume species exposed to

ele-vated CO2 for different durations and under different management

regimes Values described as Face are for Trifolium repens and T

sub-terraneum plants sampled after 18 months exposure to ambient (360

ppm) or 475 ppm CO2using free air carbon dioxide enrichment; the plants were in an established permanent pasture and were grazed intermittently by sheep (see Edwards et al., 2000) Short term data are

for T repens plants harvested after exposure to 350, 500, 650, or 800

ppm for a period of 4 weeks in controlled environment rooms; the light level was 700 mol m 2 s1for the 14 h photoperiod and the day/night

temperatures 22/16°C Long term exposure data are for Lotus

uligi-nosus plants growing at different distances from a natural CO2spring and presumed to have been exposed to elevated CO2 for many decades; the CO2concentration the plants experienced was estimated

as the mean of spot measurements taken at canopy height over a period of three years (see Ross et al., 2000); the plants were subject to intermittent cutting.

of experiments under grazing The arguments for changes in N cycling revolve in part around the well-documented decrease in N content of plant leaves at elevated CO2(Poorter et al., 1997) and the increase in the fixed N contribution by legumes (Soussana and Hartwig, 1996) These arguments also apply to grazed swards (although see later section on legume content under grazing), but we must also consider the aboveground return through excreta The reduction in leaf N content at elevated CO2appears to be main-tained over the long term; i.e., over lengths of time during which adaptation could occur (Körner and Miglietta, 1994) and in grazed as well as cut systems (Figure 14.3) Consequently, less N will be returned by animals at elevated

CO2because the N in the urine is directly proportional to the N in the herbage eaten (Barrow and Lambourne, 1961) The lower N in herbage could be com-pensated for by a greater volume of returns but as dry matter yield (and therefore the potential to increase animal numbers) is not markedly increased

at elevated CO2(Hebeisen et al., 1997; Edwards et al., 2000) this compensa-tion is not likely If the N returned in each urinacompensa-tion is reduced we might expect lower losses through leaching and volatilization as these are concen-tration dependent (Haynes and Williams, 1993), and greater uptake by plants which are able to use the lower concentrations more effectively These changes should result in tighter N cycling and greater N efficiency under grazing at elevated CO2

Physical Effects

By the action of their hooves, animals have the potential to physically alter (usually detrimentally) properties of soils and plants The hoof of an ungulate exerts a pressure that can be calculated from the area of contact and the mass of the animal Typical static load values per hoof of domestic ani-mals would be 192 kPa for a cow, 83 kPa for a sheep, and 60 kPa for a goat (Willatt and Pullar, 1983) In practice, the pressure applied is almost always greater than this as the hoof is rarely applied flat to the ground The result of treading can be seen in soil properties; there is a positive relationship between treading intensity and soil bulk density and a negative relationship with hydraulic conductivity (Willatt and Pullar, 1983) In addition, treading alters surface properties, leaving patches (gaps) of bare ground (Watkin and Clements, 1976; Betteridge et al., 1999) Plants are also susceptible to damage from treading, either by crushing or through cutting of plant parts by sharp hooves The consequences of these physical aspects of grazing are not always separated from the effects of other grazing influences However, Edmond conducted a comprehensive series of trials to study the effects of treading alone on pastures (see Brown and Evans, 1973, for a review of this work) Edmond (1970) showed that treading could reduce herbage yields by 30–40%, with the yield reduction being dependent on the plant species

pres-ent (Figure 14.4) Lolium perenne is particularly resistant to treading (Edmond,

1964) and is observed to increase in abundance as treading pressure increases

During the process of biting it is not just that leaves are removed—as they would also be under cutting—but there is also the potential for damage

to meristems and other plant parts resulting in a loss of function, e.g., photo-synthetic capacity, transport of nutrients, or increased susceptibility to pest and diseases Part of the reason for this is that the biting process also involves pulling, which lifts plant parts, such as stolons, leaving meristems vulnera-ble to being eaten Pulling can also lift plant roots from the soil Typically, 9%

of the apical meristems of Trifolium repens are removed during a rotational

grazing event (Hay et al., 1991) The consequences of the different mechanical

Figure 14.4 Sensitivity of some pasture species to treading by sheep, expressed as

yield relative to an untrodden control (From Edmonds, D.B., N.Z J Ag.

Res., 7:1 –16, 1964.)

effects of cutting and grazing are rarely considered and can be compared only

at the same defoliation interval, at the same severity of defoliation, and with the same nutrient returns Sears (1953) conducted a five year study of graz-ing effects on pastures which included a number of subtreatments From these, we can find a comparison of the mechanical effects of grazing; this does

not exclude the treading effects sensu Edmond, but in this case observation

showed the most marked effects were through the biting process In

particu-lar there was a sharp decline in Trifolium pratense under grazing because the

animals were able to remove plant crowns, whereas the cutting process left them intact (Figure 14.5)

We can envisage CO2interacting with many of these physical conse-quences of grazing First, the damaging effects of treading on soil structure— compaction, loss of drainage capacity—may have different effects at elevated

CO2in which greater allocation of C below ground is frequently observed Changes in soil biota have also been reported at elevated CO2 (O’Neill, 1994; Yeates et al., 1999), and these can strongly influence soil structural prop-erties (O’Neill, 1994) Second, the creation of gaps by the grazing animals has important consequences for population processes as these promote both recruitment from seed (Panetta and Wardle, 1992) and vegetative

Figure 14.5 Effects of mechanical damage by grazing animals on the botanical

composition of a pasture (From Sears, P.D., N.Z J Sci Tech.,

35A:1 –29, 1953.)

development—e.g., branching and tillering (Arnthórsdóttir, 1994)—and both

of these regenerative processes have been shown to be influenced by elevated

CO2 Many studies have shown elevated CO2can alter the number of seeds produced (Lawlor and Mitchell, 1991; Farnsworth and Bazzaz, 1995) which can lead to changes in recruitment in seed-limited species (Edwards et al., 2000) It may also be the case that more seed heads are left intact after defoli-ation by grazing rather than cutting, allowing greater expression of any CO2 effects on seed characteristics Other studies have shown changes in seed quality at elevated CO2 (e.g., in C:N ratio and seed mass) which have the potential to alter germination and establishment rates If the likelihood exists

of different seed behavior in gaps at elevated CO2, as shown by Spring et al (1996), then there is a strong possibility that the increased gap frequency under grazing will result in a grazing management  CO2 interaction Increased vegetative propagation has been shown to be an important mech-anism driving changes in species abundance in response to elevated CO2in a wide range of situations, such as temperate pasture (Clark et al., 1997) and alpine meadows (Leadley et al., 1999) It should achieve even greater expres-sion in the presence of more gaps (regeneration niches) and be of more criti-cal importance given the losses of meristems experienced in a grazed system

The botanical composition of pastures (species identity and abundance)

is determined in part by the method of harvesting Even a consistent, uniform process, such as cutting, has a selective effect which reflects the vertical dis-tribution of plant species in the canopy For example, cutting removes

pro-portionally more clover (Trifolium repens) laminae than grass (Lolium perenne)

laminae in a mixed grass/clover sward because the clover leaves are held higher in the canopy (Woledge et al., 1992) It has been argued that grazing animals such as sheep have a selective effect on sward composition simply by the kind of passive selection imposed by cutting (Milne et al., 1982) In fact, Parsons and co-workers have shown that sheep have a preference (i.e., actively select) for white clover; in this case, it is a partial preference, with sheep preferring a diet of 70% clover and 30% grass (Parsons et al., 1994) Note that this also means that animals might select against clover when the clover percentage in the sward exceeds 70% As a consequence, the clover removed under grazing is proportionally larger than the clover removed by the passive selection of a lawnmower (Woledge et al., 1992)

The outcome for a plant species that is a preferred component of the diet

is clearly not favorable in comparison to a nonpreferred species Indeed, ani-mals may reduce the abundance of their preferred species in the sward until

it becomes a small component of their diet (Parsons et al., 1991b)—the

“Paradox of Imprudence” (Slobodkin, 1974) The only way in which a plant species can overcome the deleterious consequences of being preferred (in relation to other plant species) is if it holds some advantage in growth over its companion species By this means, a preferred species can maintain its presence in a grazed sward until a point at which the grazing pressure out-weighs the growth advantage (Parsons et al., 1991b) It has been suggested that the growth advantage held by clover is a greater specific leaf area (Parsons et al., 1991a)

At elevated CO2, there is strong evidence that legumes are advantaged in comparison to grass species (Newton et al., 1994; Clark et al., 1997; Hebeisen

et al., 1997; Leadley et al., 1999); although this evidence comes from cutting experiments, we might anticipate that, in the absence of any change in ani-mal preference, CO2would also result in greater legume growth under graz-ing In a Face experiment grazed by sheep we compared the effects of grazing (Figure 14.6a) and cutting (Figure 14.6b) on pasture responses to elevated

CO2 The ambient values show that clover was deleteriously affected by graz-ing (compare Figure 14.6a and b); however, CO2enrichment gave the clover sufficient advantage to compensate for the grazing effect so that clover growth under grazing at elevated CO2 (Figure 14.6a) was comparable to clover growth under cutting at ambient CO2(Figure 14.6b) In this instance, clover responded positively to elevated CO2only when grazed; under cut-ting, there was only a minimal stimulation of clover, suggesting that in this environment factors other than CO set a limit to the growth of clover