Chapter 5

Ecosystem - Level Processes

Steve Archer and Fred E. Smeins

Home Page



Organization of Ecological Systems
       -     Hierarchy Theory
       -     Spatial Scale of Disturbance

Role of Climate

Grazing Effects on Ecosystems
       -     Domestic Livestock
       -     Other Herbivores

Mediation of Ecosystem Processes and Characteristics
       -     Microenvironment and Hydrology
       -     Nutrient Cycling
       -     Plant Community Dynamics

Condition and Trend
       -     Retrogression
       -     Succession and Resilience
       -     Transition Thresholds

Risk Management


Plant Species Cited

List of Figures



Plant species composition and productivity within a region largely reflect the prevailing climate, whereas seasonal and annual variability in rainfall and temperature play a central role in dictating the dynamics of populations over time. However, substantial spatial variability occurs across landscapes, and broad-scale climatic variables cannot account for the spatial patterns which shape vegetation form and function on a local scale. Soils and topography exert a strong influence on patterns of plant distribution, growth, and abundance over the landscape through regulation of the availability of moisture from precipitation, which also affects nutrient availability. Grazing influences are superimposed on this background of topo-edaphic heterogeneity and climatic variability to further influence community-level processes. As a result, species whose adaptations to the prevailing climate and soils would make them the competitive dominants of the community under conditions of light grazing may assume subordinate roles or even face local extinction as grazing pressure increases.

Grazing animals affect plants directly and indirectly. Direct effects of grazing are those associated with alterations in plant physiology and morphology resulting from defoliation and trampling (see Chapter 4). Grazing also influences plant performance indirectly by altering microclimate, soil properties (see Chapter 6), and plant competitive interactions. These indirect affects accentuate plant response to defoliation in ways not readily simulated by clipping experiments.

Over time, the combined direct and indirect effects of grazing on plant growth and reproduction are manifested in plant population dynamics. Herbivores affect the productivity, composition, and stability of plant assemblages through mediation of plant natality, recruitment, and mortality and may cause directional changes in com-munity structure and function. A community may be relatively stable and resistant to changes produced by grazing up to a certain threshold point(s). Beyond these thresholds, chan es are rapid and often augmented by climatic events. The pathway of succession following relaxation or removal of grazers may differ substantially from the pathway of retrogression, depending on the mobility and availability of propagules, soil conditions, and climatic variables. In addition, the probability of ecosystem recovery to previous states may be greatly reduced beyond certain critical threshold levels of disturbance or change. The goal of grazing management for sustained yield is to identify these critical thresholds and manage landscapes so as not to exceed them.

Management and manipulation of plant communities for sustained livestock and wildlife production (see Chapter 7 and Chapter 8) requires the seasonal integration of information on plant species composition and production across expansive, often heterogeneous areas (landscapes) and over extended planning horizons (decades). If the structural and functional aspects of grazed ecosystems are to be understood at spatial and temporal scales appropriate for long-term sustainability, key plant and population processes must be identified and linked across time to community and landscape levels of integration. Chapter 4 summarizes direct effects of grazing (defoliation) on plant and population level processes developed from short-term, small-scale controlled experiments. In this chapter we review long-term, large-scale changes in plant communities on grazed landscapes. At this level of organization, herbivore effects on microclimate, hydrology, energy flow, nutrient transformation and translocation, and soil physical/chemical properties operate against a backdrop of climatic variability to influence plant species interactions and cause fluctuation, retrogression and succession in communities.


Organization of Ecological Systems

Ecosystems are dynamic, complex, and difficult to define or delimit in space and time. Hierarchical ordering has been applied to multilevel ecological systems to provide a conceptual framework for practical definition. Systems, including ecological systems, are groups of interacting, interdependent parts operating together for some purpose. Systems have unique characteristics or emergent (nonreducible) properties which are manifest only when sub-components interact to produce larger functional wholes. Hence the axioms "the whole is greater than the sum of the parts" and "a forest is more than just a collection of trees' " A system's principal attribute is that we can only understand it fully if we view it as a whole. Water, for example, might be considered a system comprised of hydrogen and oxygen subcomponents. However, the physical/chemical attributes of hydrogen and oxygen by
themselves could not be used to predict the unique physical characteristics of H20. In ecological systems, the study of individual organisms does not reveal the unique properties of higher levels of organization that emerge from interactions of organisms with each other and their environment.

Hierarchy Theory

Because ecological systems are complex and composed of many interacting parts, it is useful to view their organization as a hierarchy, or a graded series with several levels of organization, for example, organisms, populations, communities, ecosystems, and landscapes (Rowe 1961; MacMahon et al. 1978; Allen and Starr 1982). Any level of organization in the hierarchy can be represented as a system, and interactions with the physical environment at each level produces a characteristic, functional system. The components of ecosystems (plants, animals, microbes, geologic substrates, soils, climate) interact and are dependent upon one another for the flow of energy and cycling of nutrients. Each level of organization has characteristic processes that operate at prescribed spatial and temporal scales (Woodmansee and Adamsen 1983; Woodmansee 1988) (Fig. 5.1).

Conceptual schemes such as those depicted in Figure 5.1 are important in that they explicitly identify levels of organization and hence give concrete meaning to the abstract concept of "ecosystem". To minimize misleadin& confusing, or confounding comparisons, the same level of organization and the same processes, inputs, and outputs should be compared in ecological studies. At different levels of organization other processes, inputs, and outputs should be evaluated. The goal of research is to understand the behavior of ecological systems at various levels of organization and ascertain the properties emerging at each level. Management can then focus on the processes and inputs that regulate these key properties.

Fig. 5.1

In contrast to reductionism, hierarchy theory permits evaluation of a complex system without reducing it to a series of simple, disconnected subsystems. No single level in the hierarchy of an ecological system should be considered fundamental. Understanding a system at one level of organization requires knowledge of the levels both above and below the targeted level (Webster 1979; Allen et al. 1984). Interpreting the behavior of a system at one level of organization without consideration of adjacent levels may generate misleading results. Plant response to grazing illustrates this point (Archer and Tieszen 1986). Controlled defoliation studies have increased our understanding of adaptations that confer tolerance to leaf removal at the organism level (see Chapter 4). However, such studies do not account for other factors which come into play at higher levels of organization. For example, at the organism level, species tolerant to leaf removal in a controlled environment may disappear from the community if they are grazed more frequently or intensely than neighboring plants which are perhaps less tolerant of defoliation. Thus, the infrequently grazed or ungrazed plants may come to dominate the site by virtue of their competitive advantage over the species used more frequently. Studies at the individual plant level often do not take into account key processes operating at the community level of organization, for example, differential grazing of competing species under conditions of limiting resources.

Systems (holistic) and reductionist approaches to analyzing and managing ecological systems should not be viewed as mutually exclusive. Each provides a unique perspective. The reductionist approach dissects lower levels of organization, and provides mechanistic explanations and insights as to how systems work The holistic approach, on the other hand, views a system in the context of the higher levels in which it is embedded, and provides insight into the significance of phenomena at lower levels. The search for mechanisms should therefore be balanced by concern for significance (Passioura 1979, Allen et al. 1984, Lidicker 1988).

Spatial Scale of Disturbance

Grazing influences interact with climatic variability and other variables to cause changes in plant communities at various spatial and temporal scales. On a large scale, precipitation and temperature regulate vegetation dynamics in and and semiarid systems (MacMahon 1980; Austin et al. 1981; Sala et al. 1988). However, most plant communities and landscapes are extremely patchy (Belsky 1983), and broad-scale climatic factors cannot account for the existence of these small-scale patterns. Frequent, small-scale perturbations such as ant mounds, small animal activities, patch grazing, and dung and urine deposition (Coffin and Lauenroth 1988) occur within the context of larger-scale, less frequent disturbances such as fire and drought to produce a complex disturbance regime (Collins 1987). Thus, as spatial and temporal frames of observation are diminished, and resolution increased, edaphic heterogeneity and biotic processes (such as grazing) assume greater importance in determining community structure and function.

Herbivore contributions to patchiness include localized defoliation, urine and dung deposition, altered competitive interactions resulting from the differential utilization of plants variously tolerant of defoliation, trampling and the transformation and redistribution of nutrients. These frequent, small-scale perturbations associated with grazing contribute to the development of fine-grained mosaics of varying successional age-states across landscapes. Interpreting community composition and productivity is thus contingent upon our ability to understand the interactive role of concurrent, multiple-scale disturbances (Collins and Barber 1985; Loucks et al. 1985; Collins 1987). To fully appreciate the response of various levels of ecological organization to grazing impacts, it is essential to identify levels of landscape organization, the key processes that occur at each level, their interrelationships, and the influences of various disturbance factors and regimes.

Role of Climate

The type, magnitude, duration, frequency, and season of climatic change plays an important role in regulating the rate and direction of plant community changes. On grazed landscapes it is often difficult to assess the extent to which herbivory influences ecosystem processes relative to abiotic factors (McNaughton 1983a; Foran 1986). Despite numerous studies of secondary succession in grasslands, few generalizations have emerged, perhaps because vegetation dynamics in and and semiarid systems is influenced so strongly by climate (MacMahon 1980). Infrequent but extreme climatic events may be especially important in masking or confounding patterns of vegetation change (Chew 1982).

The influence of grazing on species composition and productivity can be minor relative to the changes caused by variations in rainfall (Fig. 5.2). For example, had the data in Figure 5.2 been collected only from 1932-1937 (a long-term study by most standards!) and only from the heavily grazed site, one would have documented retrogression (Fig. 5.3) without knowing whether climate (drought) or grazing was

Figure 5.2

Figure 5.3

In some cases, effects of climate have been erroneously ascribed to grazing (see Hastings and Turner 1965, Western and van Praet 1973, Chew 1982, Branson 1985). In other instances, vegetation established under past climatic regimes may persist under the present regime in a vegetative state. Such communities may exhibit a great deal of "biological inertia", but may not be able to reestablish following disturbance and would be disposed to change, regardless of grazing pressure. However, herbivory would likely influence the rate, direction and magnitude of change in such systems. The apparent displacement of grasses by woody plants in desert grasslands of North America over the past 100 years may exemplify this phenomena (Neilson 1986). Climate-based simulations predict present day grasslands will become increasingly susceptible to desertification and woody plant encroachment if changes in temperature and precipitation associated with the greenhouse (warming) effect occur (Emmanuel et al. 1985a, b).

The extent to which shifts in vegetation structure lag behind climatic changes which drive them and the extent to which vegetation can ever be said to be in equilibrium with climate are largely unknown. This makes it difficult to form a baseline from which to judge the effects of grazing on community structure and function or range condition and trend.

Grazing Effects on Ecosystems

Domestic Livestock

The impact of livestock grazing on ecosystems varies in relation to the evolutionary history of the site and the level of grazing pressure (Stebbins 1981; Milchunas et al. 1988). Intermountain grasslands of North America evolved with light grazing and have changed markedly since the introduction of livestock (Mack and Thompson 1982). In contrast tall-, mixed- and shortgrass prairies of North America, which evolved with bison, pronghom, and prairie dogs, have been relatively resistant to stresses associated with livestock grazing. Although plant species in ecosystems that evolved with grazing are well adapted to defoliation, domestic livestock can substantially impact their growth and persistence in numerous ways (Pieper and Heitschmidt 1988):

• In contrast to wild herbivores, whose numbers or patterns of grazing may vary substantially from year to year, concentrations of domestic livestock can be artificially maintained at consistently high levels.
• Fences prevent the emigration of livestock to new areas when the abundance of desired forages decreases, resulting in higher frequencies and intensities of defoliation than would occur otherwise.
•  Supplemental feeding is used to
Minimize animal mortality that would otherwise reduce grazing pressure when over-utilization of forage occurs; and
Maintain grazing pressure over a greater portion of the year and over a higher frequency of years than would have occurred with native herbivores, both present and prehistoric.
• In grassland or savanna systems that occur in areas climatically and edaphically capable of supporting trees and shrubs, prolonged grazing may decrease the capacity of grasses to competitively exclude woody plants, while at the same time reducing fire frequency and (usually) intensity by preventing the accumulation of fine fuels.
Thus, while grazing has been an important selection pressure in many ecosystems, man has substantially changed its frequency, intensity, extent, and magnitude with the introduction of livestock. The result has been rapid and widespread changes in species composition and productivity of plant communities.
Other Herbivores

Herbivores other than livestock also influence plant community structure and productivity, and their effects on vegetation must be considered when setting livestock carrying capacity and interpreting the effects of livestock grazing on community structure. Activities of root-feeding nematodes, leaf-chewing grasshoppers, termites, herbivorous rodents, lagomorphs, large mammals, and granivores interact with livestock to affect rangeland vegetation. Seasonal and annual fluctuations in the abundance of wildlife populations are often marked, and accurate census data are difficult to obtain. As a result, estimates of numbers or biomass and the relative role of these organisms in regulating energy flow and nutrient cycling are seldom available.

Wild herbivores are often regarded as pests because they may compete with livestock for forage, consume seeds of desirable plants, impair restoration efforts, or disperse seed of undesirable weed and woody species. However, regarding these organisms as wholly detrimental oversimplifies their role. Beneficial activities include consumption of undesirable plants and their seeds, dispersal of desirable species, consumption of insect pests, loosening and aerating soils, and enhancing nutrient cycling (Huntly and Inouye 1988).

High densities of rodents and lagomorphs are perhaps indicators of degraded rangeland. Although the literature on the interactions among rodents, lagomorphs, livestock and vegetation is sparse and diffuse, it generally appears that rodent and lagomorph densities are correlated with factors such as a high incidence of eroded ground, a high diversity of herbaceous or woody dicots, and low grass cover (Fogden 1978). These organisms may therefore represent symptoms of rangeland deterioration rather than its cause. Potential competition between rodents or lagomorphs and livestock for grasses may be more than offset if these wildlife species retard invasions of undesirable woody plants or weeds.

Although most research has focused on the effects of above-ground grazers on vegetation, below-ground herbivores may actually consume more plant material (Coleman et al. 1976). In addition, below-ground grazers may have a proportionally greater impact on total primary production than would be predicted on the basis of their consumption rates. For example, root-feeding nematodes, grubs, and scarabaeid larvae reduce above-ground productivity by adversely affecting plant growth, metabolism, and nutrient and water uptake (Ridsdill Smith 1977; Ueckert 1979; Detling et al. 1980). Some evidence also suggests aboveground herbivory may change plant-soil-microbial interactions in a manner that benefits soil herbivores (Seastedt et. al. 1988). The role of these important but unseen herbivores in causing vegetation change relative to conspicuous above-ground grazers is not well understood (Anderson 1987).

Meditation of Ecosystem Processes and Characteristics

Direct effects of defoliation on plant growth and development are addressed in Chapter 4. Indirect effects associated with grazing activities further influence plant growth and community composition. These include:

• Alteration of energy flow through the detrital pathway.
• Redistribution and transformation of nutrients.
• Modification of microclimate.
• Alteration of edaphic physical and hydrologic properties.
• Destabilization of plant competitive interactions.
Acting in concert, these factors associated with the grazing process dictate the net impact of plant defoliation on community processes. When direct and indirect effects are considered together, it is easy to understand why activities of grazing animals may exaggerate variations within the community in ways not easily simulated by clipping experiments on individual plants.
Microenvironment and Hydrology

Modifications of site microchmatic and hydrologic properties by grazing constitute potentially important feedbacks that regulate plant responses to defoliation, community composition, and productivity. Grazing and related activities reduce litter accumulation and decrease plant cover which results in increased bare ground. The result may be a warmer, drier, and more extreme microenvironment facilitating an increase in short-lived perennials, annuals, or more xerophytic plants adapted to such conditions.

Microclimatic comparisons of grazed and ungrazed sites are few, but they generally indicate that air and soil temperatures are higher and ground-level wind speeds greater on grazed sites (Whitman 1971). Such changes affect primary productivity and species composition over time. For example, in and and semiarid systems where water is commonly a limiting factor, herbivore alteration of microclimate may affect availability and utilization of water by primary producers. When the leaf area index is high, grazing may remove transpiring leaf tissue and reduce canopy interception losses of precipitation (see Chapter 6), thereby enhancing soil moisture and enabling plants to sustain growth over longer periods. Defoliation may also increase the water potential of remaining plant parts and contribute to increased rates of leaf expansion (Hodgkinson 1976; Wolf and Perry 1982).

On the other hand, higher rates of transpiration and evaporation on grazed sites resulting from higher radiant heat loads, soil temperatures, and wind speeds may contribute to depletion of soil moisture. Defoliation during periods of low soil water availability may also accentuate plant water stress by reducing root initiation, extension, and activity (see Chapter 4). However, removal of leaf tissue can increase & ratio of root: leaf area, improve the water relations of remaining tissues, and result in conservation of soil water (Archer and Detling 1986; Svejcar and Christiansen 1987).

The net result of grazing-induced modifications of microenvironment on community structure and function has been estimated with the ELM ecosystem-level grassland simulation model on tallgrass prairie. The model showed that cattle weight gain per head, above- and below-ground plant production, transpirational water loss, and standing dead biomass decreased, while soil temperature, water content, and soil water loss increased with increased grazing intensity (Parton and Risser 1980). Results of the simulation are consistent with field observations.

Where grazing has reduced plant and fitter cover, sealing of soil surfaces via raindrop impact and hoof compaction may reduce infiltration and increase erosion and runoff (see Chapter 6). In addition, germination and survival of perennial grasses may be greatly reduced on such sites and recovery of surface soil properties following cessation of grazing may require decades (Braunack and Walker 1985; Salihi and Norton 1987). The rate and direction of plant succession following relaxation of grazing may therefore depend upon the degree to which soil properties have been altered in addition to the climatic factors discussed previously.

Nutrient Cycling

Grazers influence nutrient inputs, outputs, and transformations. Consumption of foliage diverts above-ground biomass from the litter component and modifies microclimate, both of which affect activity of soil microbes. In addition, defoliation affects the below-ground nutrient exchange pool by reducing root initiation and extension and increasing root mortality (see Chapter 4). Nutrients are also exported from the system when livestock are moved to other pastures or sold. Grazing- induced changes in microbial activity and the local distribution, form, and abundance of nutrients may then feed back and intensify plant response to defoliation and contribute to changes in species composition. Over the long term, changes in plant species composition or diversity additionally affects litter quality, mass, and seasonal dynamics of decomposition such that a positive feedback loop develops. Under conditions where erosion and runoff increase because of grazing (Chapter 6), nutrient losses from a site may be greatly accelerated.

Nutrient cycling via grazing animals can be important in enhancing or maintaining soil fertility (Floate 1981). Cycling of nutrients through grazers may help keep a pool of readily mineralizable organic nutrients near the soil surface where they are more accessible to plants and microbes (Botkin and Wu 1981). Consumption of vegetation and subsequent defecation could also increase the turnover and availability of various elements that would otherwise remain in recalcitrant organic forms. The fact that shoots of plants on grazed areas may have higher nutrient concentrations than plants from comparable ungrazed areas (Coppock et al. 1983; McNaughton 1984) may be a consequence of:

• Enhanced nutrient uptake by defoliated plants (Ruess 1984).
• Increased nutrient availability resulting from: 9 Input from dung and urine;
• Enhanced microbial activity associated with higher soil temperature and moisture (Parton and Risser 1980);
••  Decreases in root biomass which limits carbon availability to decomposers causing decreased microbial biomass and increased net mineralization (Holland and Detling 1990).
••  Elevated levels of root exudation resulting from defoliation. Leakage of high quality organic compounds from roots stimulates microbial activity and speeds nutrient cycling (Ingham et al. 1985; Klein et al. 1988).
Plants require substantial amounts of nitrogen (N) for photosynthesis. When adequate water is available, nitrogen is typically the resource that most limits plant production in and and semiarid systems. As a result, factors affecting its availability and retention in forms used by plants are important determinants of ecosystem processes. Cattle may return 80-85% of the N ingested with plant tissue to the soil system via urine and feces. Urine generally contains 50-80% of the N excreted by cattle, and vegetation on urine spots may have foliar concentrations of nitrogen 2-3% greater than that of plants on non-urine spots over a 40-60 day period (Stillwell and Woodmansee 1981, Thomas et al. 1988, Day and Detling 1990). Inorganic N levels in soils of urine spots may remain elevated for up to 90 days and above-ground standing crop increased 3- to 7-fold.

Cattle defecate as many as 14 times in a 24-hour period (Weeda 1967), with each defecation impacting an average of 225-600 cm2 of ground surface (Welch 1985; Brown and Archer 1987). Urination patches of 0.28 m2 can affect grass growth over an area of 1 m2 (Wilkinson and Lowrey 1973). Thus, substantial portions of a landscape can be impacted, depending upon the number of animals and their temporal and spatial patterns of movement. Fences, water, shade, and topography dictate animal distribution patterns (see Chapter 3) and hence the pattern of nutrient translocation and redistribution (Senft et al. 1985). For example, Hilder (1964) found 33% of the total feces are concentrated in less than 5% of the area grazed by free-roaming sheep. Productivity may therefore be enhanced on some portions of the landscape by rapid cycling and nutrient import. Conversely, long-term net removal of nutrients may contribute to a reduction of productivity and shifts in species composition in other locales.

Nutrients ingested by grazing animals not returned via excreta and mortality may be lost from the system via harvesting of animal products, animal emigration, and transformation to labile or volatile forms. At stocking rates typical of tallgrass prairie, removal of N in animal biomass is small, amounting to about 0.08 g N/m2/y [assuming 60 kg weight gain per animal (Parton and Risser 1980, Svejcar 1989) and an N concentration in gained tissue of 0.014 g N/g wet body mass (Berg and Butterfield 1976)]. By contrast, net annual deposition inputs of N to tallgrass prairie have been estimated at 0.44 g/m2 (wet fall only; NADP 1987) to 1.7 g N/m2 (Seastedt 1985).

Losses of nutrients consumed in foliage is probably significant, but the magnitude is difficult to quantify. Approximately 12% of the N consumed by ruminants escape as gaseous products of digestion (Church 1969). Of the N excreted by ungulates, 50% may be lost via ammonia volatilization (Woodmansee 1978). However, subsequent analyses by Schimel et al. (1986) suggest such losses may be much lower and spatially variable. In addition, volatilization of N from urine and dung may not represent a net loss, because volatilization from unconsurned vegetation may exceed losses from animal excreta (Detling 1988). In modeling primary and secondary productivity of a tallgrass prairie, Parton and Risser (1980) assessed the various inputs and outputs of nitrogen and predicted that volatilized losses of nitrogen from cattle urine and feces would decrease the net nitrogen balance of a site as grazing pressure increased. Urinary N not taken up by plants or volatilized can be quickly converted to nitrates which are vulnerable to loss by leaching where precipitation is high (Stillwell and Woodmansee 1981).

Grazing may also effect a decrease in the input of "new" plant-available N by impacting lichens and blue-green algae. These nitrogen-fixing organisms constitute potentially important sources of plant-available N in and and semiarid systems, although their input is highly variable and pulsed. Cryptogams also play a significant role in enhancing surface soil stability and water infiltration. Trampling associated with grazing reduced the number of moss and lichen species by 50% and reduced moss, lichen, and algal cover by 90%, 60% and 50%, respectively, in desert shrub communities (Anderson et al. 1982).

Over the long term, excessive levels of grazing can potentially reduce nitrogen fixation; increase ammonia volatilization, leaching and erosional losses; and cause a net directional transport of nutrients to localized portions of the landscape. The result can be a decrease in overall site fertility and increased heterogeneity of primary production. Reductions in site fertility constitute a positive feedback accentuating defoliation stresses, augmenting shifts in species composition, and determining the rate and direction of succession following relaxation of grazing.

Plant Community Dynamics

Competition. Wooton (1908), reporting on the status of grazing lands in New Mexico, observed that "Stock eat the valuable forage plants and leave the poor ones, thus giving the latter undue advantages in the struggle for existence.' Thus, when plants are defoliated in a community rather than as isolated individuals, it becomes difficult to tell whether the plants' response is to loss of leaf tissue or to changes in interference from surrounding vegetation (Jameson 1963). In this regard, Mueggler (1972, 1975) and Archer and Detling (1984) found that regardless of severity of defoliation, plants under competition had greater reductions in biomass and flower production and were slower to recover from defoliation than were plants clipped under conditions of reduced competition.

At the community level of organization, differences in animal forage preferences (see Chapter 3) often result in differential frequency and intensity of defoliation of plants on a site and cause a shift in competitive interactions. Plants grazed less frequently gain an advantage over plants that are utilized at greater frequency or intensity. Thus, grazing-induced shifts in competitive interactions contribute to changes in plant community composition over time. On a local scale, herbivores may therefore play a key role in mediating species composition in the community through the differential utilization of plants variously tolerant to defoliation.

The effects of grazing on plant distribution and abundance vary. Figure 5.4 illustrates plant distributions in 1982 along a topo-sequence in pastures stocked at different rates with cattle since 1939. Plant responses to grazing varied with stocking rate and topographic position. Some species increased with increasing grazing pressure at some locations and decreased on others (e.g., buffalograss and sun sedge). Species tolerant of heavy grazing at some locations on the topo-sequence were less tolerant at other locations (e.g, buffalograss). Blue grama was able to maintain its status in the community regardless of grazing pressure, except in the lowlands under conditions of heavy grazing. Soil physical properties, nutrient content, and nitrogen mineralization rates differed significantly between swale, mid-slope, and ridge locations on this site (Schimel et al. 1985). Seasonal patterns of relative cattle grazing preference (ratio of percent of time grazing at a given topographic location to the percent of pasture area occupied by that topographic feature) ranged from 1.4 for swales to 0.4 for ridgetops on the site (Senft et al. 1985).

Differences in plant distribution and abundance along hill-slope gradients in these pastures subjected to different grazing regimes may be explained in several ways:

• Energy budgets, water relations and nutrient availability varied along the gradient and affected intrinsic plant tolerance to herbivory.
• The probability of being grazed at a given frequency and intensity by a given class of herbivore (e.g., nematodes, insects, rodents, ungulates) varied across the topo-sequence along with patterns of grazing (e.g., patchy versus uniform).
• The competitive neighborhood of a given plant species changed along the gradient.
Figure 5.4
Thus, at any given location, some species may be more or less tolerant to defoliation than others and they may be more or less likely to be grazed than others. As summarized in Chapter 4, plant adaptations to herbivory may fan into two general categories: grazing avoidance and grazing tolerance. Grazing tolerant plants have characteristics that facilitate the reestablishment of foliage following grazing, whereas plants that avoid grazing have characteristics that minimize the probability of defoliation (Archer and Tieszen 1980; Mooney and Gulmon 1982; Briske 1986). Although plants tolerant of defoliation would seem to have an advantage in grazed systems, such plants will eventually be disadvantaged when competing with avoidance-type species which, although less tolerant of grazing, are defoliated at a lower frequency or intensity. Where grazing intensity is high, the grazing avoidance-type plants will inevitably dominate the site because of the overriding influence of utilization. Evidence for this premise is most striking from southwestern hot deserts, the Great Basin and the Southern Plains, where extensive areas of grasslands characterized by grazing tolerant-type plants (perennial grasses) have given way to shrublands dominated by unpalatable grazing avoidance-type plants. In California, grazing-tolerant perennial grasses appear to have given way to annual grasses which could be viewed as employing an avoidance or escape-type strategy.

Diversity. In terrestrial plant communities, much of the theory of community organization has stressed the role of competitive interactions among plant species. Although plant species may occupy distinct niches in terms of their resource requirements, herbivores mediate species abundance and diversity through differentially utilizing plants variously susceptible to defoliation. Certain levels and combinations of grazing or disturbance increase overall plant species diversity by decreasing the capacity of competitive dominants to exclude other species and by creating gaps available for occupation by other species (Huston 1979; Archer et al. 1987; Collins 1987; Collins et al. 1987). Above certain frequencies or intensities, disturbance typically lowers diversity. This phenomenon of increased diversity at moderate levels of disturbance has been termed the intermediate disturbance hypothesis (Connell 1978).

Grazing can stimulate diversity by reducing the capacity of competitive dominants to exclude other species via defoliation, trampling, and dung deposition. The effects of grazing on plant diversity depends upon grazing intensity, the evolutionary history of the site, and climatic regimes (Milchunas et al. 1988). In semiarid grasslands with an evolutionary history of grazing, vertebrate herbivory appears to have a relatively small effect on community composition (Fig. 5.5). In these systems, defoliation enhances tillering and spread by rhizomes and stolons and seems not to affect competition for resources (McNaughton 1983a, 1984). In contrast, climatically similar grasslands with a short evolutionary history of grazing lose diversity at much lower grazing intensities. In regions with higher rainfall, a few tall species will be the competitive dominants of the community when grazing pressure is low. Moderate grazing increases diversity in these systems by creating mosaics of short grasses and forbs which occur on heavily grazed patches, mixtures of tall and short grasses on moderately grazed patches, and tall grasses on lightly grazed patches. Heavier grazing eventually causes diversity to decline as short grasses dominate an increasingly greater proportion of the community.

Community diversity has important implications for grazing management. Growth of each species in a community is limited by a different combination of environmental factors. Fluctuations in weather cause production of individual species to vary substantially from year to year. However, production of the whole community is more stable, because years favorable for growth of some species ccaausee aa compensatory decrease in growth of other species (McNaughton 1977; Chapin and Shaver 1985; Collins et al. 1987). Conversely, in stressful years, the loss of productivity of some species is compensated for by growth of others. As a result, changes in relative growth rates and abundances of co-occurring species tend to stabilize ecosystem processes such as primary production (Fig. 5.6). For example, C3 grasses and shrubs may be physiologically active early in the spring, whereas C4 species maintain growth for a greater proportion of the warmest, driest portions of the growing season. The result, at certain latitudes or elevations, is that a diverse mixture Of C3 and C4 plants may give more stable and sustained productivity than a monocul-ture of either.

Figure 5.5

Figure 5.6

Certain levels of grazing may enhance above-ground net primary production (ANPP) (McNaughton 1979) if diversity increases. In other cases ANPP is relatively unaffected over a wide range of grazing intensities and durations (Whicker and Detling 1988). Above certain grazing intensities, the level of ANPP is reduced and its variability increased by either reducing productivity of dominant species or by increasing the proportion of short grasses, forbs, and annuals in the community. Species-rich plant communities are potentially more resilient (i.e., better able to regain functional characteristics after disturbance) to repeated defoliation than species-poor communities (Brown and Ewel 1988), perhaps reflecting the ability of species mixtures to use resources more fully than monocultures (Harper 1977).

Species Composition and Population Dynamics. Sustained productivity and long-term survival of plants in grazed systems depends upon successful reproduction in parental generations and the recruitment of new individuals. By mediating plant natality, recruitment, and mortality, herbivores affect the productivity, composition, and stability of plant communities. Changes in basal area, relative abundance, and species composition in grazed systems inevitably reflect differential recruitment, longevity, and survival of individuals comprising the community. Population param-eters such as these provide a tangible, quantifiable link between individual plant and community-level processes and should forecast impending changes in community composition.
Proper grazing management aids recruitment and persistence of desired species, whereas poor management hastens the demise of preferred species and leads to their replacement by other species (Jones and Mott 1980).

Traditionally, plants have been classified as decreasers, increasers, or invaders with respect to their response to grazing (Dyksterhuis 1949) (Figure 5.7). However, the functional response of a given species to grazing can vary from site to site and across topographic and edaphic gradients (Figure 5.4). Species that initially make up a large proportion of the community, but which decline with increased grazing, are categorized as decreasers. Their decline with grazing is related both to defoliation tolerance and level of utilization relative to other species in the community. Species which increase in grazed communities do so because they are more tolerant of defoliation and/or they are less frequently grazed (less preferred, less palatable). Where annual rainfall supports tall- and midgrass species, there is often a decrease in the contribution of the competitively dominant taller species and an increase in more defoliation-tolerant, short-statured, or prostrate species or growth forms with grazing (Clarke et al. 1947; Herbel and Anderson 1959; Detling and Painter 1983; Archer et al. 1987; Smeins and Merrill 1988; Noy-Meir et al. 1989). Reductions in basal area, root biomass, mulch levels, and leaf area of dominants by grazing creates establishment gaps which benefit other growth and lifeforms. As a result, an influx of species which were absent or of minor importance at lesser levels of grazing occurs. These invaders, typically annuals or unpalatable (and sometimes toxic) perennial herbaceous or woody plants, are often undesirable for livestock production because they displace more palatable species, are of lower nutritive value, or have low, erratic, or highly seasonal productivity.

Figure 5.7

Rates of change which accompany grazing are not well quantified. Shifts in species composition can be abrupt, non-linear (Fig. 5.3), and punctuated by fluctuations in rainfall (Fig. 5.2). Changes in physiognomy from tall to short grasses can occur within a few years (Archer et al. 1987; Smeins and Merrill 1988; Thurow et al. 1988b). Woody species pose special problems in many grasslands and savannas around the world. Significant physiognomic changes from grass to woody plant domination can occur within decades.

The Role of Woody Plants. Woody plants are a common component of most rangelands around the world. However, trees and shrubs have traditionally been viewed negatively, because they are presumed to reduce herbaceous production and because their presence increases the difficulty of livestock manipulation. For these reasons, shrubs and trees are frequently the targets of vegetation manipulation technologies aimed at improving livestock carrying capacity or handling. However, when assessing whether to invest in efforts to reduce woody plant cover or density, the following points should be considered:

•  In many regions or landscapes within a region, the shrub or tree lifeform is best adapted to the prevailing biotic and abiotic conditions, and woody plants play a key role in primary production and nutrient cycling while stabilizing soils and providing habitat for wildlife. On sites such as these, it may not be realistic to expect management or manipulation to enhance herbaceous production.
•  Are woody plants reducing herbaceous production or is the increased abundance of woody species the result rather than the cause of decreased herbaceous production associated with improper grazing management? Effects of woody plants on herbaceous production varies from (a) stimulation beneath tree or shrub canopies relative to interstitial zones (Christie [1975] for poplar box; Barth and Klemmedson [1978] for algarrobo; Scifres et al. [1982] for huisache; Stuart-Hill et al. [1987] for Acacia karroo; Ludwig et al. [1988] for various hot desert shrubs; Belsky et al. [1989] and Weltzin and Coughenour [1990] for umbrella thom and baobab); to (b) no effect (Harrington [19731 for Acacia hockii); to (c) decreased standing crop (Jameson [19671 for pines; Walker et al. [19861 for poplar box and narrow-leaved iron bark; Clary [19871 for pinyon-juniper woodlands; Scanlan [19881 for subtropical thorn shrubs).
•  The effect of shrubs on herbaceous production depends upon the site (topo-edaphic features), the climate (especially the temperature and rainfall regime), the species of woody plant and its growth-form (i.e., single- vs. multiple-stemmed), canopy architecture, size, and density (e.g., Frischknecht 1963; Scifres et al. 1982; Aucamp et al. 1983; Walker et al. 1986; Belsky et al. 1989). Herbaceous species respond differently to the presence of woody plants (Fig. 5.8), and manipulation of woody species may adversely affect key forbs and grasses. For example, mesquite savannas in the Rolling Plains of Texas were characterized by C3 grasses beneath mesquite canopies versus C4 grasses in zones between trees. Following herbicide application C4 grasses displaced C3 grasses beneath mesquite plants (Heitschmidt et al. 1986). As a result forage production in spring was delayed and greatly reduced, necessitating extended supplemental feeding.
Figure 5.8
•  The presence of woody plants may:
••  Reduce grazing pressure on grasses and provide protection for heavily utilized herbaceous species (Welsh and Beck 1976; Davis and Bonham 1979; Jaksic and Fuentes 1980);
••  Reduce supplemental feed requirements during cold or dry periods (Cook 1971);
••  Aid in meeting nutritional requirements of herbivores; Enhance soil nutrient status and water infiltration;
••  Provide shade; and
••  Improve habitat for game and non-game wildlife.
• Will woody manipulation stimulate herbaceous production and increase livestock carrying capacity sufficiently to offset treatment costs? If so, how much time will be required before a follow-up treatment? Will treating one problem perhaps create another (i.e., single- vs. multiple-stemmed growth habit; replacement of non-sprouting species [e.g, big sagebrush] with sprouting species [e.g, rubber rabbitbrush])? Answers to these questions vary between species and sites (Scifres et al. 1988) and the type, timing, and sequencing of vegetation manipulation technologies (Scifres et al. 1983).
• Desirable woody plants that provide cover and palatable forage for livestock and wildlife may have been reduced in abundance or eliminated by excessive grazing (Orodho et al. 1990) or by application of non-selective brush manipulation practices and replaced by other woody species (Fulbright and Beasom 1987). When woody plants are sprayed with herbicides, forbs utilized by wildlife (for example, deer and quail) are often adversely affected.
•  Aesthetic value.

Shifts in Grass and Woody Plant Abundance. Grazing animals in concert with human activities have caused the degradation of woodlands in Africa, Asia, and India (Tothill and Mott 1985). On the other hand, grazing has also been implicated in the spread of bush in Africa, desert and thorn scrub in North and South America, and acacia and eucalyptus woodlands in Australia, at the expense of grasslands and savannas. In the latter instances, increased grazing intensity may favor woody plants by decreasing herbaceous standing crop, reducing fire frequency and intensity, and enhancing the dispersal and germination of woody plant seeds. Alternative hypotheses regarding the balance between contrasting life-forms (grasses versus shrubs and trees) are centered on climatic (frequency, amount, and seasonality of rainfall) and edaphic factors (soil texture, nutrient status, moisture).

Quantitative and historical assessments indicate woody plant abundance has increased substantially in grasslands during the last century in Africa (Kelly and Walker 1976; van Vegten 1983), Australia (Harrington et al. 1984), India (Singh and Joshi 1979), North America (Smeins 1984), and South America (Schofield and Bucher 1986; Bucher 1987). Remaining grasslands and savannas may become increasingly susceptible to woody plant encroachment in response to anticipated global changes that may generate warmer, drier climates characterized by greater variability (Emmanuel et al. 1985a, 1985b). Although encroachment of woody plants into grasslands has been widely recognized, the rates, patterns, and dynamics of the process have seldom been quantified. However, in western and southwestern North America the transformation of grasslands and savannas to shrublands or woodlands appears to have proceeded exponentially within the last 200 years (Blackburn and Tueller 1970; Herbel et al. 1972; Young and Evans 1981; Madany and West 1983; Williams et al. 1987; Archer 1989).

Factors regulating the balance between graminoid and woody plant life-forms include climate, soils, disturbance (e.g., grazing, fire), and their interaction. Changes in one or more of these factors may enable woody plants to increase in abundance. A shift from grassland or savanna to shrub or woodland may result if:

•  Climate or disturbance regimes change to enable native woody species to
••  Extend their geographic range (e.g., creosote bush [Hunziker et al. 1977]);
••  Increase in stature and density within their historic ranges (for example, honey mesquite Dohnston 1963]; one-seeded juniper Uohnsen 1962]; baccharis [Williams et al. 1987]).
•  Introduced woody species successfully establish and reproduce (e.g., McCart-ney Rose [Scifres 1980]; mimosa [Lonsdale and Braithwaite 1988]).
Characteristics common to many woody species that increase in grazed environments include:
•  High seed production
•  Seeds that persist in soil for many years
•  Ability to disperse over long distances
•  Ability to sprout following top removal
•  Tolerance to low levels of water and nutrients
•  Low palatability.
Evidence suggests climate may have changed in recent history to favor woody plants over grasses in portions of North America (Hastings and Turner 1965). Grasslands and savannas which established under previous climatic regimes may have been only marginally supported by the recent climate and were perhaps prone to woody plant invasion (Neilson 1986, 198 7). In addition, oscillations between different climatic regimes in recent history (Mitchell 1980) may have effected shifts in plant recruitment patterns to promote episodes of woody plant seed production and seedling establishment, but not necessarily their local extinction (Neilson and Wullstein 1985). However, changes or fluctuations in broad-scale climatic regimes cannot explain how grasslands and savannas have persisted on some sites within a climatic zone but not others.

On a local scale, fire, grazing, and soil properties interact within a variable cli-mate to determine the balance between grasses and woody plants. Madany and West (1983) have documented a case in which a savanna protected from cattle grazing was maintained despite low fire frequency (and possible climatic change), whereas nearby edaphically similar sites subjected to cattle grazing had substantially higher densities of woody plants which established after the introduction of livestock in the late 1800s (Fig. 5.9). Such data suggest climatic fluctuation in recent history may have been necessary, but was not sufficient, to have caused a shift from savanna to woodland.

Figure 5.9

Woody Plant Establishment. Numerous studies in North America indicate increased densities of woody plants in grasslands and savannas coincident with the introduction of domestic livestock (Chew and Chew 1965; Blackburn and Tueller 1970; Young and Evans 1981; McPherson et al. 1988; Archer 1989). Large numbers and high concentrations of livestock potentially favor establishment of woody plants in numerous ways:

•Increased dispersal of viable, germinable seed, either directly (Table 5.1) or indirectly, by creating changes leading to increased numbers of other seed-dispersing organisms such as rodents and birds.

• Compaction of surface soils may favor recruitment of woody plants over grasses (Braunack and Walker 1985).
• Defoliation of grasses can increase the probability of woody plant seedling establishment and subsequent rate of growth and development by:
••  Adversely affecting the capacity of grasses to preempt resources above-and below-ground (Caldwell et al. 1987);
••  Causing shifts in herbaceous composition to assemblages less effective in garnering site resources (Eissenstat and Caldwell 1988); and
••  Reducing biomass and continuity of fine fuels and hence fire frequency.
• Time to reproductive maturity may be decreased (McPherson and Wright 1987) and the frequency and magnitude of seed production ruay be increased among woody plants on grazed sites.
Grazing management strategies for limiting woody plant encroachment should focus on reducing seed production and dispersal, enhancing the capacity of grasses to competitively exclude seedlings of woody plants and enabling periodic use of fire to induce mortality and minimize the number of woody plants that attain reproductive status.

Table 5.1

Although numerous studies have cited the importance of the relationship between grazing, graminoid biomass and woody plant invasion, these parameters have seldom been quantified. Herbaceous retrogression that accompanies grazing has been established in numerous studies. Yet few guidelines have been developed to predict how grazing at different stages of herbaceous retrogression and levels of stocking alters the susceptibility of sites to invasion by woody plants. It is often assumed that well-developed stands of grass can exclude woody seedlings and that sites with a history of heavy grazing are most likely to have woody invasion, other factors held constant.

To test these assumptions, honey mesquite (hereafter called mesquite) establishment was quantified on edaphically similar areas with different grazing histories (long-term heavily grazed vs. protected from livestock grazing for 40 years) and with various levels of herbaceous defoliation (none, moderate, and heavy). Mesquite seedling emergence and survival after two years was comparable and high (>75%) on plots subjected to moderate and heavy defoliation, regardless of site grazing history (Brown and Archer 1989). Above-ground growth, photosynthesis, conductance, and water potentials were comparable among one-year-old mesquite seedlings on grazed and protected sites, even though differences in herbaceous species composition and above- and below-ground biomass on the two sites were substantial. These data suggest competition for soil resources between grasses and woody plants such as mesquite may be minimal early in the life cycle of mesquite. The ability of woody plants such as mesquite to establish in grass-dominated stands may be related to the rapid development of a root system (Table 5.2) which enables the plant to access soil moisture at depths not effectively utilized by grasses (Table 5.3).

Table 5.2

Table 5.3

Establishment would be particularly enhanced when seeds germinate during periods when competition for soil moisture is minimal. These results, together with observations of Meyer and Bovey (1982) and Smith and Schmutz (1975), indicate grasses may not be effective at competitively excluding mesquite. The relatively unpalatable shrub big sagebrush also has great potential to increase in density regardless of ecological or management conditions short of periodic fires (Tisdale and Hironaka 1981; Johnson 1987).

Thus, in many instances, thresholds of herbaceous biomass production or composition required to limit establishment of woody seedlings may be exceeded even at low levels of grazing. Management focused on regulating levels of herbaceous utilization may slow the rate but may not curtail the invasion of aggressive woody species. Grazing systems which allow for the regular use of fire seem to be required to successfully regulate woody plant density, stature, and seed production on many sites. For woody legumes whose seeds are effectively distributed by livestock (e.g., mesquite and acacia), management directed at reducing seed dispersal might retard rates of encroachment. Pastures containing seed-bearing trees might be deferred during times of pod production. Livestock grazed on areas coincident with pod production should be detained long enough for seeds to pass through their digestive system (several days) before moving them to pastures where these plants are not a problem (Fig. 5.10).

Figure 5.10

Condition and Trend

Within a climatic area, differences in vegetation across a landscape reflect intrinsic variations in soils and topography. Resource managers have divided landscapes into range sites, which are comprised of one to several soil series potentially capable of producing the same kind, proportion, and abundance of late successional plant species (Shiflet 1973). The estimated climax community on a given range site has been used as a basis for determining range condition which is defined as the composition of the current plant community relative to the known, presumed, or preferred climax. Additional details on rangeland condition in relation to plant succession can be found in Lauenroth and Laycock (1989).

Vegetation dynamics on range sites include fluctuation, retrogression, and succession (Fig. 5.3). Fluctuation represents reversible changes in dominance within a stable species composition, whereas succession and retrogression are directional changes in composition and dominance (Rabotnov 1974). As species composition and the contribution of various species to site production deviates from the idealized climax, range condition declines through the phenomenon of retrogression or retrogressive succession (Barbour et al. 1987). Grazing or changes in environmental conditions can cause retrogression which eventually leads to a loss of diversity net primary production, and ground cover. As a result, site processes become increasingly coupled to and regulated by abiotic factors (Fig. 5.3). This, in turn, may accentuate fluctuation. Progressive directional changes, termed succession, occur in the opposite direction and represent the recovery of ecosystem structure following biotic or abiotic disturbance. As plant diversity, production, and ground cover increase through time, the plants themselves exert substantial control over microclimate, energy flow, nutrient cycling, and species interactions, thus dampening fluctuation associated with oscillation of weather and abiotic factors. Changes in range condition through time, termed trend, indicate whether succession or retrogression predominate on a site.

"Proper" management is assumed to be that which (1) minimizes the likelihood of a site retrogressing or being degraded to the point where primary and secondary productivity are adversely affected and soil resources are endangered; and (2) facilitates succession. As illustrated in Figure 5.2 and Figure 5.3, determination of trend requires data over sufficient periods of time to distinguish fluctuation related to infrequent climatic events from directional change. Few such data exist. Collins et al. (1987) analyzed 39 years of vegetation data from two Oklahoma sand sagebrush-little bluestem-blue grama sites grazed by cattle. Each site contained an exclosure. The vegetation in the grazed portions of each pasture exhibited shifting patterns of abundance rather than sequential species replacement (Le, fluctuation), as did one of the exclosures. The other exclosure exhibited a directional change from dominance by annuals to dominance by perennial grasses.

Terminology traditionally associated with condition classes (excellent, good, fair, poor) and species (desirable, undesirable) represent situation- or user-specific value systems. Terms such as "poor" and "undesirable" are relative and the reference point may vary substantially among different people. Species assemblages regarded as desirable for sheep may be undesirable for cattle, and vice-versa. Plants regarded as desirable for livestock may be undesirable for wildlife such as deer and quail, and vice-versa. A site in poor condition for livestock may be in excellent condition for wildlife. Because the connotation of these terms varies between user groups, they should be well qualified when used.

The idea underlying the range condition concept was to provide a base line from which to evaluate changes in ecosystem attributes through time (Dyksterhuis 1949). It does not necessarily follow that management for climax is necessary, desirable, or achievable for several reasons:

• The original state of vegetation and natural fire and grazing regimes prior to settlement is seldom known. In most cases it is not clear what constitutes climax vegetation.
• Sub-climax or introduced plant species assemblages may be desirable (i.e., palatable, productive, seasonally available) when managing for certain herbivores.
• Maximum plant productivity is not necessarily higher at later stages of succession.
• Productivity of many large herbivores (e.g., white-tailed deer, moose, cattle, sheep) may be greatest in vegetation states not considered "climax" (see Chapter 1).
• Late-successional or climax communities are not necessarily more stable than mid- or even early-successional communities (McGinty et al. 1978).
• The concept of single equilibrium communities that progress steadily toward or away from climax depending on grazing pressure does not apply in many and and semiarid systems. Examples of the importance of stochastic events in shaping the path of succession, alternative steady states, and discontinuous and irreversible transitions are abundant (see Westoby et al. 1989).
• The rate and path of succession varies depending on the extent to which soil properties have been impacted. Return to the presumed climax or late successional communities may not occur over time frames relevant to management.
• Alternative steady states are possible on a given site. If disturbance thresholds are exceeded, positive feedbacks can cause systems to shift to a new steady state, regardless of grazing practices.

If climate becomes warmer and drier or if the frequency of drought increases, plant productivity, cover, and diversity decreases and species composition shifts toward an increasing proportion of annuals and xeromorphic perennials. This represents retrogressive succession (Barbour et al. 1987). Grazing can also induce retrogressive succession, as palatable grasses, forbs, and shrubs succumb to repeated defoliation (see Chapter 3 and Chapter 4) and are eventually replaced by other growth and life-forms. Grasslands are able to tolerate a moderate degree of grazing intensity before changing in composition, diversity, or productivity. However, as grazing intensity is increased or becomes continuous, tall and mid-grasses eventually give way to short-statured perennial grasses, which, in turn, give way to annuals and unpalatable perennials with a concomitant loss of primary and secondary productivity, diversity, cover, and soil. As discussed previously, the level of grazing required to cause a change from one of these states to another depends upon the type and numbers of herbivores, the plant species involved, and the evolutionary history of the site with respect to grazing. Retrogression associated with livestock grazing may be mitigated when growing season conditions are favorable or magnified in unfavorable years.

Retrogression implies degradation, and in range management it is typically used to describe the replacement of perennial species by annuals or the replacement of palatable species with unpalatable species. The latter case does not necessarily constitute ecological degradation, however. Instances in which grasslands and savannas become shrublands or woodlands may represent succession in that plant diversity, primary productivity, soil fertility, etc., are maintained or even enhanced on the site. This example is viewed as retrogression only in that forage plants valued for certain classes of livestock are lost. However, the change in species composition simply reflects the fact that the new assemblage of species on the site is better adapted to the prevailing environmental conditions which include livestock grazing.

The process of retrogression appears to be step-wise rather than linear. Thus, a community may be rather stable and resistant to change up to a certain threshold. Beyond certain threshold levels, changes can be rapid, dramatic, and potentially irreversible over reasonable time frames (Fig. 5.11). The goal of sustainable grazing management is to anticipate these critical thresholds and manipulate livestock so as not to exceed them. Once a threshold is exceeded, it may not be possible for the system to return to the previous condition, even with large inputs of energy and nutrients.

Figure 5.11

Succession, Stability and Resilience

Grazing management offers an opportunity to maintain or enhance species composition, diversity, and production, but hinges upon an understanding of processes regulating community succession, stability, and resilience. Stability is a measure of persistence in the face of disturbance. Two components of stability are resistance and resilience. Resistance describes the ability of the community to avoid displacement when a given type, frequency or magnitude of disturbance occurs (Begon et al. 1986). Resilience describes the speed with which a community returns to its former state after it has been displaced from that state. We know little of these attributes. How much stress can a particular community endure before significant changes in structure and function occur? Once disturbed, will the community return to its previous state? If so, what factors affect the rate of recovery? The concepts of resistance and resilience presume the existence of disturbance and transition thresholds. If thresholds exist, the magnitude of stress a system can absorb before changing to a less desirable configuration must be determined. Plant and population attributes which may forecast impending changes must also be quantified so that management can be adjusted to avert undesirable shifts.

Generally, as grazing pressure increases, a site assumes a new configuration or condition class (Fig. 5.7 and Fig. 5.11). In managing for livestock and certain classes of wildlife, changes in composition may be acceptable or even desirable, as long as site productivity and soil stability are maintained. When grazing pressure is reduced or removed, the rate of succession back to a previous configuration (higher condition class) depends on the extent to which soils, seedbank, and vegetative regeneration potential of the previous vegetation has been modified. It should not be assumed that the pathway of succession following reduction in grazing intensity will be a simple reversal of the pathway of retrogression. Reasons for this will be discussed later.

Rates of succession on sites released from grazing are highly variable, but generally proceed much more slowly than desired for management purposes. Some studies have reported significant, quantitative changes in vegetation (i.e., succession) to occur rapidly in the absence of grazing (Cooper 1953; Penfound 1964; Austin et al. 1981; Collins and Adams 1983; Potvin and Harrison 1984; Biondini et al. 1985). In other cases, the species present on retrogressed sites released from grazing may represent a new steady state and persist nearly unchanged for long periods (>30-50 y) following protection from grazing (Robertson 1971; Smith and Schmutz 1975, West et al. 1979; Glenn-Lewin 1980; Holechek and Stephenson 1983). For example, grazing within a live oak savannah caused mid-grasses (sideoats grama, Texas winter-grass, Texas cupgrass, and cane bluestem) to give way to short grasses, primarily the stoloniferous common curlymesquite (Smeins and Merrill 1988). Once established as a dominant, curlymesquite has persisted >25 years in the absence of grazing. Even when seed sources were present, midgrass species failed to re-establish in the curlymesquite community (Kinucan 1987). Thus, a relatively stable community has been maintained on a site previously supporting a higher successional class of herbaceous vegetation.

As discussed earlier, grazing, with the natural consequence of reduced fire frequency and intensity, can precipitate a physiognomic shift from grassland or savanna to shrubland or woodland. Once woody plants are established, new successional processes and positive feedbacks can cause a rapid conversion in physiognomy (Fig. 5.11). In many instances, cattle and sheep, the principal herbivores in most ranching enterprises, consume negligible amounts of woody vegetation and do little to inhibit its recruitment and growth. In some cases, livestock serve to increase the dissemination of seeds of woody plants (Table 5.1, Fig. 5.10). Over time, the capacity of defoliated grasses to carry fire or competitively exclude woody plants diminishes, after which a new, potentially stable community dominated by woody plants develops. Increases in the woody plant seedbank coupled with high vegetative regeneration potential and the deterioration of the herbaceous seedbank (Koniak and Everett 1982; Hobbs and Mooney 1986) and vegetative reserve make it highly unlikely this new plant community will revert to grassland, even if livestock are removed from the site (Niering and Goodwin 1974; Walker et al. 1981; Holechek and Stephenson 1983; West et al. 1984; Wester and Wright 1987). Once in the shrub- or woodland domain, these new communities may be highly resistant or resilient to fire or anthropogenic manipulation (e.g., herbicides, mechanical treatments). Apparent examples of regional shifts in North American vegetation to alternate steady states in recent history include the following:

• The Great Basin-from perennial bunchgrasses and open stands of sagebrush to dense sagebrush and annuals such as cheatgrass and medusahead.
• Southwestern Desert Grasslands-from tobosa and black grama grasslands to creosotebush, tarbush, or mesquite shrublands.
• Southern Grasslands and Savannas-from tallgrass prairies and savannas to oak, juniper, mesquite, or thorn scrub woodlands.
• California Mediterranean Grasslands-from perennial bunchgrasses to annual grasses.
As the preceeding examples illustrate, the pathway(s) of succession following removal of grazers are often not simply a reversal of the pathway of retrogression. As a result of this hysteresis phenomenon, large inputs of cultural energy (seeding, fertilization, mechanical and chemical manipulation, etc.) would be required to regain the desired composition. These management-imposed manipulations may increase the grass to woody plant ratio on a site, but the conversions may be short-lived and periodic follow-up treatments will be required (Scifres et al. 1983).
Transition Thresholds

The concept of community resistance and resilience portrayed in Figure 5.11 suggests the existence of transition thresholds. Output from the SPUR (Simulation of Production and Utilization on Rangelands) model (Wight and Skiles 1987) predicted the occurrence of a threshold in sagebrush-crested wheatgrass systems (Torell 1984). Simulated production of both crested wheatgrass and big sagebrush remained fairly constant over time when stocking rates varied from 1.12 to 0.56 ha/AUM. However, a slight increase in stocking rates from 0.56 to 0.50 ha/AUM resulted in a marked decrease in crested wheatgrass production accompanied by a sharp increase in sagebrush production (Fig. 5.12). The existence of such thresholds also account for the abrupt, non-linear development of woody plant communities in areas formerly dominated by grasses (Fig. 5.13) (Buffington and Herbel 1965; Herbel et al. 1972; Archer et al. 1988). Several studies suggest periods of drought followed by rainfall may trigger shrub encroachment on areas subjected to livestock grazing. If thresholds exist in the retrogressive process, we must determine the extent to which a system can be grazed before changing from one herbaceous state to another or from an herbaceous to a woody state.

Figure 5.12

Figure 5.13

Risk Management

The notion of resilience emphasizes the importance of maintaining community structure by maintaining diversity and variability. However, in agroecosystems, stable, sustained high yields are sought by employing intensive management schemes to minimize variability (see Chapter 10), but at the expense of resilience (Table 5.4). Over the short-term, yields are increased and planning is facilitated. However, problems may be generated over the long-term, for as a system loses its resilience, it may become increasingly sensitive to errors in management and to exogenous events. Reducing the biological variability of a system reduces its resilience and increases the probability that chance or rare events previously "absorbed" by the system will cause dramatic change. Relative to intensive management, extensive practices are less effective in reducing variability. As a result they typically generate fewer livestock products. However, over the long-term, integrity of the extensively managed system may be maintained in the face of perturbations in whatever unexpected form they might take. Given the long time spans required for natural recovery of many rangelands to desired states, and the high monetary and energy costs and often low probability of successful restoration following disturbance, this more conservative strategy is likely the most profitable and sustainable.

Table 5.4


Most grazed systems in and and semiarid regions are expansive and heterogeneous. Plant production is limited by climatic and edaphic factors. Because of the numerous, interactive, and to a large extent stochastic processes that regulate species composition and productivity in natural systems, it is difficult and misleading to propose standard prescriptions for vegetation management. Agronomic approaches, with their heavy emphasis on expensive cultural inputs and treatments, have a high rate of failure in and and semiarid situations and are typically not economically feasible even when successfully implemented. It is therefore essential to acquire a functional understanding of the basic ecological processes that drive natural systems and develop flexible management strategies that work within constraints dictated by soils and a highly variable and unpredictable climate.

The traditional concept of single equilibrium communities that progress steadily toward or away from climax depending on grazing pressure seems not to apply in many and and semiarid systems. Examples of alternative steady states, abrupt thresholds, and discontinuous and irreversible transitions are becoming increasingly abundant for both succession and retrogression. When one group of plants has been displaced by another as a result of altered climate-grazing-fire interactions, the new assemblage may be long-lived and persistent, despite progressive grazing management practices. The adage "an ounce of prevention is worth a pound of cure" thus has substantial application to vegetation management, particularly in situations where the probability of re-establishment of desired species composition and soil cover may be quite low once a change has occurred.

Abrupt transitions between various states of vegetation composition may be triggered by stochastic events related to the vagaries of climate, seed dispersal, and seedling establishment. Managers should seek to identify circumstances whereby desirable transitions can be augmented and facilitated and undesirable transitions mitigated or avoided. Westoby et al. (1989) liken grazing management to a con-tinuous game where the object is to seize opportunities and avoid hazards. Such a philosophy is based on timing and flexibility rather than fixed policy. In systems where climatic variability is the rule rather than the exception, situations conducive to vegetation improvement or deterioration may arise infrequently and unexpectedly. Failure to recognize and respond to either situation constitutes missed opportunity. If the potential for transition to undesirable states is ignored, long-lasting, potentially irreversible impacts can result. Conversely, progressive and flexible management schemes which can capitalize on infrequent windows of opportunity for vegetation improvement or livestock production may realize long-term benefits in livestock and wildlife productivity.


Plant Species Cited
Acacia  Acacia Mill. spp.
------  Acacia karroo Hayne
------  Acacia hockii DeWild
Algarrobo  Prosopis juliflora (= glandulosa Torr. var Torreyana [L. Benson, M. C. Johnst.])
Baccharis  Baccharis pilularis var. consanguinea (DC.) C. B. Wolf
Baobab  Adansonia digitata L.
Big Sagebrush  Artemisia tridentata Nutt.
Black Grama  Bouteloua eriopoda (Torr.) Torr.
Blue Grama  Boutelous gracilis (H. B. K.) Lag. ex Steud.
Buffalograss  Buchloe dactyloides (Nutt.) Engelm.
Cane Bluestern  Bothriochloa barbinodis (Lag.) Herter
Cheatgrass  Bromus tectorum L.
Common Curlymesquite  Hilaria belangeri (Steud.) Nash
Creosote Bush  Larrea tridentata (DC.) Cov.
Crested Wheatgrass  Agropyron cristatum (L.) Gaertn.
Eucalyptus Eucalyptus L'Heritier spp.
Honey Mesquite; Mesquite  Prosopis glandulosa Torr. var. glandulosa
Hooded Windmillgrass  Chloris cucullata Bisch.
Huisache  Acacia farnesiana (L.) Willd.
Juniper  Juniperus L. spp.
Little Bluestern  Schizachyrium scoparium (Michx.) Nash
Live Oak  Quercus virginiana Mill.
McCartney Rose  Rosa bracteata Wendl.
Medusahead  Taeniatherum caput-medusa (= asperum) Simonkai) Nerski
Mimosa  Mimosa pigra L
Narrow-leaved Ironbark  Eucalyptus crebra F. Muell.
One-seeded Juniper  Juniperus monosperma (Engelm.) Sarg.
Pines  Pinus Lindl. spp.
Pinyon Pine  Pinus monophylla Torr. and Frem.
Ponderosa Pine  Pinus ponderosa Laws
Poplar Box  Eucalyptus populnea F. Muell.
Prickly Acacia  Acacia Mill. sp.
Rubber Rabbitbrush  Chrysothamnus nauseosus (Pall.) Brit.
Sagebrush  Artemisia L. spp.
Sand Sagebrush  Artemisia filifolia Torr.
Sideoats Grama  Boutelous curtipendula (Michx.) Torr.
Sun Sedge  Carex eleocharis Bailey
Tarbush  Flourensia cernua DC.
Texas Cupgrass  Eriochloa sericea (Scheele) Munro
Texas Wintergrass  Stipa leucotricha Trin and Rupr.
Tobosa  Hilaria mutica (Buckl.) Benth.
Umbrella Thom  Acacia tortilis (Forsk.) Hayne subsp. spirocarpa
Western Wheatgrass  Agropyron smithii Rydb.

List of Figures

Figure 5.1. Conceptual integration of concepts from biological and pedological hierarchies (from Woodmansee and Adamsen 1983). Each level of organization in the hierarchy is characterized by key processes, inputs and outputs (Table 5.1). Energy and matter are exchanged between components within a level and between levels. In this scheme, an "ecosystem" is a biotic community (plants, microbes, herbivores, carnivores) in association with a soil polypedon. A "catena" is comprised of linked soil/plant associations. The landscape is a hierarchical level whose components are catenas.

Figure 5.2. Changes in percentage basal cover in the shortgrass steppe vegetation type during and after drought (from Branson 1985). The effects of grazing on plant basal cover were relatively minor compared to changes caused by variations in rainfall.

Figure 5.3. Hypothetical changes in species composition over a number of climatic cycles (adapted from Walker et al. 1986). All lines illustrate the role of climatic variability in causing community composition to fluctuate through time (Fig. 5.2). The upward trend represents succession, whereas the downward trend represents retrogression.

Figure 5.4. Mean importance values (relative cover plus relative frequency) of four graminoids along shortgrass steppe toposequences in pastures subjected to different levels of cattle grazing since 1939 (N=none; L=light; M=moderate; H=heavy) (from Archer and Tieszen 1986).

Figure 5.5. Hypothesized relationships between relative plant diversity in grassland communities in relation to grazing intensity along gradients of moisture and evolutionary history of grazing (from Milchunas et al. 1988).

Figure 5.6. Categories of plant response to grazing at the community level of organization (from Dyksterhuis 1949). Plants which are intolerant to defoliation or are more heavily grazed than others will decrease with increasing grazing (decreasers). Plants less preferred or more tolerant to defoliation will subsequently increase. If grazing pressure is maintained gaps created by the loss of perennials will be occupied by unpalatable and/or ephemeral plants (including shrubs in many systems) previously absent or of limited abundance (invaders).

Figure 5.7. Illustration of how hypothetical differences in species abundance resulting from grazing (a) might affect the amount and variability of aboveground net primary production in two different seasons (b) and (c). Vertical bars represent rainfall (from Walker et al. 1986).

Figure 5.8. Generalized array of herbaceous/woody vegetation interactions along spatial/temporal gradients in a south Texas savanna parkland (from Scanlan 1988). A classification analysis of 19 herbaceous species generated six response groups. Spatial gradients extend from the center and edge (dripline) of woody clusters into neighboring interstitial zones. Temporal gradients reflect increased density and diversity of woody plants in clusters (P = pioneer mesquite (Prosopis glandulosa and 2 to 4 other woody species; D = developing woody cluster [8-10 woody species]; and M = mature cluster [12-15 woody species]).

Figure 5.9. Conceptual model of changes in community structure as a function of grazing pressure (from Archer 1989). Within the grassland domain, grazing alters herbaceous composition while decreasing fire frequency. If grazing pressure is relaxed prior to some critical threshold(s), succession toward higher condition grasslands can occur. In some cases such changes require decades. If sufficient numbers of woody plants become established, new successional processes may drive the system to a new steady state. Once in the shrub- or woodland domain seed bank and vegetative regeneration potentials are altered and the site may not revert to grassland or savanna, even following removal of grazing. Herbicides and grazing management that allows for subsequent use of fire at regular intervals may be required to reduce woody cover and enhance herbaceous production (Scifres et al. 1983).

Figure 5.10. Predicted trends in crested wheatgrass (Agropyroncristatum) and big sagebrush (Artemisia tridentata) production under different stocking rates (SPUR model output, courtesy of Ross Wight, unpublished).

Figure 5.11. Changes in percent woody plant cover on three sites in a southern Texas savanna parkland between 1941 and 1983 (from Archer et al. 1988). The 1941-1960 period included a major drought in the 1950's. Annual rainfall during the 1960-1983 period was normal to above-normal. Cattle grazing was heavy and continuous through most of the 42 year period

Figure 5.12    Predicted trends in crested wheatgrass and big sagebrush production under different stocking rates(SPUR model output, after Torell 1984).

Figure 5.13    Changes in woody plant cover on three sites in a southern Texas savanna parkland between 1941 and 1983.