Worldwide, observers report changes in the tree-grass matrix of arid and semi-arid grasslands insofar as woody species seem to be encroaching and crowding out grass species. Research suggests this widespread phenomenon of woody encroachment has occurred during the past 50-300 years in parts of Africa, South America, North America, and Australia, coincident with introductions of exotic species and/or changes in disturbance regimes. Considered a rapid phenomenon, encroachment seems to be irreversible on time scales that are relevant to ecosystem management. This article reviews the characteristics and causes of woody encroachment with a particular emphasis on the southwestern United States.
Background on Woody Encroachment
Most woody encroachment problems are located in savanna systems. The savanna landscape consists of a matrix of trees and grasses in a wet-dry climate where most precipitation falls during a single time period. Usually occurring in subtropical and tropical regions, the savanna climate is characterized by seasonal droughts and hot conditions. The wet season occurs during the high-sun period, and the dry season during the low-sun period. The very hot season occurs before the onset of the rainy season. The soil water regime undergoes a seasonal decline in storage to a point where it will not support a closed-canopy forest.
"Raingreen" vegetation reflects the seasonality of the moisture regime: grasses dry out and trees lose foliage during the dry season but green up quickly with the onset of precipitation. Some tree species are xerophytic with small leaves and thorns; others are broad-leaved deciduous, shedding leaves in the dry season. Savanna vegetation classes range from grassland to woodland, according to the relative proportions of these two plant forms. Widely scattered trees or shrubs define a savanna grassland, and their densities may change so that the grassland grades into a tree savanna, shrub savanna, or savanna woodland. The savanna woodland occurs where widely dispersed trees grow within a grassland matrix. A savanna parkland consists of a two-phase mosaic landscape in which circular clumps, groves, or mottes of woody plants (the discrete phase) are dispersed throughout a grassy matrix (the continuous phase). Savanna landscapes may host other types of vegetation, including riparian or gallery forest, patches of woodland, swamps, or marshes. In general, savanna structure is dictated by interactions among climate, topography, soils, geomorphology, herbivory, fire, and human activities. Studying the interactions among these factors helps to understand the woody encroachment problem.
The change in ecosystem structure precipitated by woody encroachment affects hydrology, biogeochemical cycling, landscape evolution, biodiversity, and future land-use options. Increases in Prosopis glandulosa (honey mesquite) density and distribution decrease grass production. This decreases forage for livestock and makes management of cattle more difficult due to the blockage of sight lines and corridors through which to move herds. The change in vegetation cover may reduce the quantity and quality of water within watersheds. This shift in land-cover is also of interest to global change research because of its effects on nitrogen (N) and carbon (C) cycling and other gas fluxes. The increase in woody biomass might also store a substantial amount of carbon. Thus, woody encroachment might be important to the global change community's research emphasis on the C budget. On a smaller scale, shifts in available soil water may result from changes in species composition; this may be particularly important in these arid and semi-arid ecosystems.
Woody Encroachment in the Southwestern U.S.
Woody encroachment is one component of vegetation dynamics in the southwestern United States. In southern and northern Texas, the invasion of P. glandulosa into grassland has initiated woody cluster development and increased the presence of woody species overall. In southern New Mexico, a historical study found that the range of P. glandulosa, Flourensia cernua (tarbush), and Larrea tridentata (creosote bush) increased between 1858 and 1963 and that no open ground remained. In southern Arizona, research has focused on shifts in the ecotone between woodland and semi-desert grassland. In this system, conclusions about the direction of change have varied from upslope shifts to stable woodland distribution to downslope shifts. One carbon isotope study in the southwestern U.S. and northwestern Mexico revealed Quercus emoryi (Emory oak) and P. juliflora (velvet mesquite) as recent components of former grasslands. However, grasslands immediately below the existing woodland boundaries have been stable for at least 1700 years.
The southwestern U.S. was not pure grassland before the advent of modern land-use. Early reports of the area described grassland and savanna landscapes, but also noted the presence of Prosopis and other trees and shrubs to varying degrees. Many historic accounts mention open prairies with woody plants confined to areas adjacent to watercourses; others suggest a lateral spread from wooded bottoms to upland prairies. Woody vegetation has probably been present throughout the Holocene, perhaps restricted to escarpments, caliche ridges, drainages and riparian zones. The woody species present in southern Texas have not changed their historic ranges in the last 300-500 years, but rather have increased in stature and abundance.
Besides vegetation change, land use has changed substantially in this region over the past few hundred years. In southwest Arizona and northern Mexico, the Madrean Quercus (oak) woodlands have been used by settlers, miners, explorers, and pastoralists for more than 100 years as mine timbers and smelter fuelwood (1890-1910), pastures for grazing (heavy in 1880s, continues today at lower stocking rates), and locations of fire suppression.
Woody Encroachment Unknowns
Despite the observations and studies of this phenomenon, information is lacking about the rates, dynamics, patterns, and successional processes involved. The amount or significance of change is hard to judge because there are no quantitative reports before the 1900s. Certainly, woody species were mentioned in early reports—some mentioned little cover while others noticed thickets. Settlers did not happen upon pure grasslands. The reports of early travelers and settlers are descriptive, and the objectiveness of these reports is difficult to judge. One means of quantifying historical cover beyond the information provided in historical documents involves the isotopic studies previously mentioned.
Woody encroachment is non-linear and has been accentuated by extreme climatic events such as drought. Associated causes of encroachment include heavy livestock grazing, fire suppression, elimination of browsers, and the influx of unpalatable, stress-tolerant, nitrogen-fixing plants. Shifts in vegetation composition and structure involve shifts in ecotones, cluster development of woody species, and establishment of woodland thickets at the expense of grasslands.
Causes of Woody Encroachment
The causes of woody encroachment are unclear, or at least not uniform. The balance between woody plants and grass species results from dynamic, non-linear, non-directional processes which must be teased apart. The relative proportions of various species are influenced by climate, competition, chance events, edaphic properties, seasonality, amount of rainfall, and the magnitude/frequency of disturbances such as grazing, browsing, and fire. Many causes and their interactions have been identified, but no definitive story nor general causal theory has been elucidated. In order to address this, the varieties of study sites and research questions that address this problem must be synthesized and analyzed to discover what is common among them. What is known are the two change vectors through which woody encroachment can occur: (1) increase in stature and abundance; or (2) increase in extension of range. For the time scale of concern—approximately 300-500 years—most researchers consider woody encroachment to be the result of changes in natural and anthropogenic disturbance regimes, rather than shifts in the geographic range of woody species. Causal explanations of woody encroachment must explain both how the landscape conversion occurs and what allows woody species to increase their presence.
In southern Texas, it is proposed that landscape conversion follows the establishment by P. glandulosa of a woody cluster. So, in this area, P. glandulosa plays a key role as the colonizing species. Once established, it serves as a recruitment focus for bird-disseminated seeds of other woody species which in turn permits the development of a woody cluster with a P. glandulosa nucleus. The coalescence of clusters imposes a gradual shift from grassland to savanna to woodland. The scientists conclude that their study site in southern Texas consists of a chronosequence of woody assemblages which are merging toward more widespread woodlands on a landscape level.
Seedling Dispersal versus Establishment
Seedlings must establish and persist at a grassland/woodland boundary in order for any kind of vegetation change to occur. Different views on the causes for seedling survival—establishment or dispersal opportunities—offer various insights on the woody encroachment problem. For some well-adapted species, the limiting factor seems to be dispersal. For example, P. glandulosa invades well: it produces abundant, long-lived seed. Germination and establishment of this species can occur on a wide range of soil types (with a variety of chemical, physical, moisture, and light properties); it can fix nitrogen as both a seedling and adult. Each sapling quickly develops extensive tap and lateral roots, and about 80 percent of 2-3 year old seedlings can survive hot fires. P. glandulosa may have been dispersal-, rather than establishment-, limited. Whereas P. glandulosa was restricted previously to riparian zones, drainages, and escarpments, modern land-use abets the wider distribution of the species.
On the other hand, Quercus spp. seedlings rarely establish in grasslands below the current lower treeline. Several factors—such as competition from other plants for above and belowground resources, herbivores (both invertebrate and vertebrate), edaphic constraints, and climatic limitations—interact and limit the establishment of Quercus in southern Arizona. Invertebrate herbivores defoliate seedlings during the summer and constitute the most common source of mortality in low-elevation Quercus of southern Arizona. Vertebrates kill Quercus seedlings in autumn and winter. Above-ground interference from herbs reduces seedling survival, but does not affect growth. Below-ground interference underlines the importance of accessing soil resources, and rapid early root growth allows Quercus seedlings to access soil moisture not available to herbs within one year after germination. Whereas these establishment constraints are important, woody plant recruitment occurs even when these factors exist: Quercus has overcome these constraints to expand. Experiments also show that Quercus can establish in the presence of herbaceous vegetation and invertebrates under existing climate conditions.
Influences of Soil Resources and Competition
Edaphic conditions also determine seedling establishment. Under various vegetative covers, the microenvironment can change and so the edaphic constraints alter through time. For instance, as a nurse plant, P. glandulosa alters its immediate surroundings and so makes it more possible for other woody species to establish in proximity. Especially crucial in these arid environments may be changes in the soil water regime.
A study of intraspecific competition in high- and low-density stands of P. glandulosa found that the resources necessary for individual growth were limiting under increased stand density, and that the limiting factor appeared to be soil water. Comparisons were made between high-density stands (300 trees/ha), representing a woodland thicket, and low-density stands (80 trees/ha), representing a savanna. Limitations were seen at the whole plant level via modification of tree size and leaf area per tree; adjustments of leaf physiological processes were not significant. The hypothesized competition occurred among lateral roots in the spaces between trees. The findings that intraspecific competition occurs in P. glandulosa stands supports Archer's theory of facilitation in the landscape conversion process.
Interspecific competition most likely plays a role in the variable abundance of grass and woody species. Grasses might influence negatively the size and density of woody vegetation by affecting establishment, growth, and development directly through interference or indirectly through processes such as fire. Woody vegetation can have negative, neutral, or positive effects on grass, depending on the species, its growth form, canopy architecture, rooting pattern, size, and density.
Exogenous Causal Mechanisms
Drought, grazing, and fire suppression receive the most attention as the causes of woody encroachment. Also, the influence of climate change may influence any one of these factors or directly affect the woody encroachment process. The mechanisms of possible change and their interactions are explored in the following sections.
Weather Regime and Climate Change
Since soil available moisture is one of the limiting factors in arid woodlands and savannas, drought comprises one of the principal factors limiting seedling establishment. Meso-climatic, edaphic, and geomorphic factors may be important for woody plants. Some studies suggest that low soil moisture does not constrain Quercus germination and early survival and that establishment temperature alone does not constrain seedling establishment since it occurs at many elevations including below current treeline. More likely, the critical influence is that of temperature on the soil water balance.
Quite likely, climate change does or will affect vegetation distribution. For studies which consider long-term vegetation dynamics, the climate imprint must be accounted for when considering the variables responsible for past vegetation change. Studies seeking to predict future change face just as many challenges in deciding the climate influence.
Projections suggest that global warming and drying will cause woody plants to shift upslope in general, yet studies of North American savannas converting to woodlands are contrary to this. Spatial distribution will change due to changes in relative competitive ability. Physiological differences between life forms distinguish competitive ability of plants. Under increased atmospheric concentrations of carbon dioxide (CO2), studies indicate that plants:
- increase net photosynthesis;
- reduce photorespiration;
- change dark respiration; and
- reduce stomatal conductance (decreasing transpiration and increasing water use efficiency [WUE]).
A widespread assumption states that an increase in CO2 will lead to enhanced establishment and growth of plants with the C3 photosynthetic pathway (generally, shrubs and trees) in grasslands currently dominated by herbs with the C4 photosynthetic pathway. In experimental studies, C3 plants exposed to high CO2 exhibit greater increases in growth and photosynthesis than C4; growth and photosynthesis of the latter are limited to cool temperatures. Water use efficiency of C3 plants is usually affected more than C4 under high [[carbon dioxide|CO2. Interacting effects of temperature change and alteration in soil moisture need also be considered when projecting plant response to increased CO2. In controlled experiments, high CO2 and constant temperature favors C3 over C4. These experimental findings, however, are inconclusive; some studies suggest that C4 plants will increase soil water availability through reduced stomatal conductance and thereby facilitate net carbon uptake. Others maintain the ecophysiological advantage of C4 plants under increasing CO2.
Commercial Grazing and Fire History
The increased density of cattle has allowed woody encroachment to occur through a number of vectors. The advent of the grazing industry provided a vector for dispersal of P. glandulosa seed. Many seeds usually germinate in one fecal site (from cow dispersal), but only one seedling at the most establishes even if 20-30 seedlings emerge. It has been argued that this supports the theory of intraspecific competition in P. glandulosa. In addition, livestock's preferential consumption of herbaceous species reduces the competition for woody plant establishment. In contrast to cattle grazing, wildlife herbivory aids the suppression of woody species. Browsers prevent woody plant domination by reducing tree growth, consuming resources, and keeping the tree species within the flame zone of grass fires.
Grazing also affects woody plant establishment indirectly by decreasing the amount of fine fuel available for fire. Cattle grazing has reduced biomass enough during the past 150 years to limit fire spread. Prior to modern settlement of the southwest, fire frequency ranged from 10 to 20 years. Today, the most common constraint on fire is inadequate fuel, given that ignition sources are still plentiful and that significant drying occurs every year. Reduced fire frequency resulting from decreased fuel load and active fire suppression has allowed more woody plants to establish whereas young seedlings probably would have been killed in a natural fire regime.
A management plan advocating fire as a means to reduce the density of P. glandulosa thickets in northern Texas notes that heat from low intensity fires inhibits flower and bean development of the woody species during the growing season following fire. The low-intensity fire regime can also kill the lower branches of mature trees and so "trim" the tree into a more aborescent form. Consequently, low-intensity fires maintain a tree-like savanna with less seed production than multi-stemmed regrowth thickets.
Another part of the fire dynamic stems from the herbaceous layer itself: Exotic herbaceous species are spreading and produce more fine fuel than native species. Increased fuel may lead to more fire, and a positive feedback may result as shown by the increase in one exotic grass after fire. Fragmentation and suppression, though, might maintain the limited fire regime despite this feedback. Changes in community structure caused by climate change or other external factors would also affect the fire regime. The interaction between the fire regime and climate change would depend on the specifics of the climate change: increases in summer precipitation would increase fine fuel loading and so increase fire frequency, but drier conditions would reduce the amount of fine fuels.
Modern land management influences vegetation dynamics in favor of woody species. For example, in areas with P. glandulosa thickets, woody plant density has been exacerbated by nonlethal chemical or mechanical treatment techniques. These treatments kill above-ground biomass but do not affect the roots. This stimulates regrowth from stem bases and increases stem number per plant and per land area. Dense thickets of P. glandulosa reduce grass production and plant species diversity.
Frameworks for Research
The problem of identifying the cause or interacting causes of woody encroachment is tied to the problem of understanding the process of vegetation dynamics. In this case, the key is to find out whether encroachment represents dynamic oscillation between different cover types or directional change from grasslands to woodlands. The traditional emphasis on equilibrium and successional processes often leads researchers to design studies that evaluate directional change. Scientific understanding, however, may be limited. The process of woody encroachment may be a dynamically stable one that does not coincide with human time scales of perception. In other words, the noticed change may be an artifact of spatial and temporal perspectives that have limited human observations from the first reports through the present time period. On the other hand, encroachment might constitute landscape conversion from one system to the next. Certainly, carbon isotope studies indicate that present land cover differs from past land cover in certain areas. At the same time, woody encroachment does not constitute directional change. Woody species invade grasslands, but grasslands also currently cover areas once dominated by trees. One study concludes that this shows “redistribution of trees over time within these savannas, and … periodic disturbances and soil resource partitioning by trees and grasses contribute to long-term persistence of savannas”. Even those that study the cause of woody encroachment acknowledge the variability of the phenomenon: for instance, Archer states "savanna vegetation is often in a state of disequilibrium, such that neither grasses nor woody plants can exclude the other and dominate the site. Given the dynamic nature of processes interacting at various spatial and temporal scales, it may be difficult to distinguish fluctuation from directional (successional) change in savanna ecosystems”.
Such concessions allow other pertinent research questions to develop. In particular, the link between structure and function needs to be made. By understanding how the structural changes of woody encroachment translate into changes in the functional aspects of the ecosystem, the consequences of woody encroachment can be identified. Process studies can occur at the plot level in traditional field campaigns as well as at larger scales. Structural variables using larger areas can be extracted from satellite imagery and used to derive functional variables. These, in turn, can be used in ecosystem process models to understand the functional aspects of woody encroachment. Measuring the functional changes in the system will provide direct information about alterations in biogeochemical and hydrological cycles. In addition, data about process changes might provide insight to the causes of encroachment through increasing the knowledge about mechanisms involved. Perhaps these kinds of studies can improve scientific thought about the causes and consequences of woody encroachment as well as theoretical ideas about vegetation dynamics.
- Ansley, R.J. and J.F. Cadenhead III. 1996. Mesquite Savanna: A Brush Management Option. The Cattleman (April):10-12.
- Ansley, R.J., B.A. Trevino, and P.W. Jacoby. 1998. Intraspecific competition in honey mesquite: Leaf and whole plant responses. Journal of Range Management 51:345-352.
- Archer, S. 1995. Tree-grass dynamics in a Prosopis-thornscrub savanna parkland: Reconstructing the past and predicting the future. Ecoscience 2:83-99.
- Archer, S., D.S. Schimel, and E.A. Holland. 1995. Mechanisms of shrubland expansion: land use, climate, or CO2? Climate Change 29:91-99.
- Archer, S., C. Scifres, C.R. Bassham, and R. Maggio. 1988. Autogenic Succession in a Subtropical Savanna: Conversion of Grassland to Thorn Woodland. Ecological Monographs 58:111-127.
- Barnes, P.W. and S. Archer. 1996. Influence of an overstorey tree (Prosopis glandulosa) on associated shrubs in a savanna parkland: implications for patch dynamics. Oecologia 105:493-500.
- Boutton, T.W., S.R. Archer, A.J. Midwoody, S.F. Zitzer, and R. Bol. 1998. *13 values of soil organic carbon and their use in documenting vegetation change in a subtropical savanna ecosystem. Geoderma 82:5-41.
- Buffington, L.C. and C.H. Herbel. 1965. Vegetational Changes on a Semidesert Grassland Range from 1858 to 1963. Ecological Monographs 35:139-164.
- Glendening, G. 1952. Some Quantitative Data on the Increase of Mesquite and Cactus on a Desert Grassland Range in Southern Arizona. Ecology 33:319-328.
- Hibbard, K.A., S. Archer, D.S. Schimel, and D.W. Valentine. 2001. Biogeochemical changes accompanying woody plant encroachment in a subtropical savanna. Ecology 82:1999-2011.
- Humphrey, R.R. and L.A. Mehrhoff. 1958. Vegetation Changes on a Southern Arizona Grassland Range. Ecology 39:720-726.
- Johnsen, T.N., Jr. 1962. One-seed juniper invasion of northern Arizona grasslands. Ecological Monographs 32:187-207.
- Johnston, M.C. 1963. Past and present grasslands of southern Texas and northeastern Mexico. Ecology 44:456-466.
- Lundell, C.L. 1967. Flora of Texas. Texas Research Foundation: Renner.
- Patton, D.R. and P.F. Ffolliott. 1975. Selected bibliography of wildlife and habitats for the Southwest. U.S. Dept. of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station: Fort Collins.
- Scholes, R.J. and S.R. Archer. 1997. Tree-grass interactions in savannas. Annual Review of Ecology and Systematics 28:517-544.
- Strahler, A.N. and A.H. Strahler. 1989. Elements of physical geography. John Wiley & Sons: New York. ISBN: 0471616478
- Swetnam, T.W. and J.L. Betancourt. 1998. Mesoscale disturbance and ecological response to decadal climatic variability in the American Southwest. Journal of Climate 11:3128-3147.
- Walker, B. and W. Steffen. 1997. An overview of the implications of global change for natural and managed terrestrial ecosystems. Conservation Ecology [online]1:2.
- Weltzin, J.F. and G.R. McPherson. 1994. Potential effects of climate change on lower treelines in the southwestern United States. USDA Forest Service Technical Report RM-GTR-264: 180-193.