Savannas and Barrens

[ Home ]

Slides of the more frequently discussed North American range savannas were arranged under the biome headings as follows:

Examples of five North American barrens were arranged under the Grassland Biome:

1. Chert Glade (= Chert Barrens= Chert Bald)
2. Dolomite Glade (= Dolomite Bald= Dolomite Barrens= "Bald Knobs")
3. Sandstone Barrens or Balds (Palo Pinto Cross Timbers)
4. Serpentine Barrens, including rock outcrop
5. Stoney Rock Land & Granite Escarpment.

Three examples of North American barrens were treated under the Shrubland Biome:

1. Gypsum Dune
2. Malpais (= Lava Flow).
3. Black Mangrove

One example of North American barrens was included under the Forest and Woodland Biome:

1. Pacific Pygmy Forest is the Western Pine and Cypress Barrens.

Based on physiological or ecological dominance (dominant species are those having the most control over, most important reaction in succession of, the community environment and greatest competitive ability in the climax vegetation) a number of the above range cover types shown under Forest and Woodland Biome and Shrublands Biome should have been arranged under the Grassland Biome. The author arbitrarily arranged these "misplaced" range types (eg. big sagebrush shrub-steppe, Rio Grande Plains shrub savanna belong more to the grassland biome than to the shrubland biome) based on aspect dominance and where most students would be most likely to look for them based on names of types given in standard texts and references (eg. sagebrush-grass type, South Texas brush country). In some instances, range cover types were arranged by and for species consistency rather than by ecological role. For example, all range types with quaking aspen were arranged under the Forest and Woodland Biome even though the quakie scrub is shrubland and aspen scrub parkland is more grassland than woodland.

The remaining rationale for this arrangement was included in the following explanation.

Savannas are those plant communities which— in the strictest, most specific, and also most general definition as used in Range Management— are transition zones between herbaceous formations, especially grassland, and woody formations (forests, woodlands, and shrublands). In this strict sense of a transition zone, a zone of competitive tension, between two of the broadest units of vegetation (formation or biome if the combined plant and animal or biotic community is viewed) a savanna is a mixture of herbaceous species (grasses are usually dominant) with trees and/or shrubs.

In ecological theory this transition zone is an ecotone and, again, an ecotone between grassland and forest, woodland, or scrub. Savannahs, therefore, can be either tree-savanna or shrub-savannah. These floristic and physiogonomic characteristics are specified in the standard definitions of savanna as used in Range Management and Forestry:

"A grassland with scattered trees, either as individuals or clumps; often a transitional type between true grassland and
forest", Society for Range Management (Huss, 1964; Kothmann, 1974; Jacoby, 1989), and by the Society of
American Foresters (to which savanna was both more variable and more restrictive in scope of vegetation), "an area of
grassland devoid of trees or containing only scattered trees, but occurring in a generally forestered area" (Munns,
1950), "essentially lowland, tropical or subtropical grassland, general with a scattering of trees or shrubs—note: if woody
growth is absent, such an area is termed grass savanna; with shrubs and no trees, shrub savanna; with shrubs and widely,
irregularly scatterd trees, tree savanna" (Ford-Robertson, 1971; Helms, 1998, except shrub savanna was omitted).

Definitions by the two societies devoted to native or natural vegetation were quite similar and conveyed the same mental picture. The SAF definition(s), however, were more of a Plant Geography concept in which savanna implied both low elevation and low latitude geographic distribution whereas the SRM definition was not restricted by geographic parameters but described strictly by floristics and physiogonomy and by the nature of transition (ie. being ecotonal) between general kinds of vegetation. The SAF definitions did not ascribe a transitional or ecotonal nature to savanna even though its glossaries included ecotone which it defined as "the transition zone between two adjoining communities" (Ford-Robertson, 1971; Helms, 1998).

The SRM definition is beyond question more consistent with cover (= dominance) types for both forest and rangeland vegetation (both SAFand SRM cover types) because cover types and all major large spatial—scale units of vegetation in all major vegetation classification systems have been and still are based on floristics and physiogonomy. The concept of transition— and tension —zone between large-scale plant communities conveys the importance of such biotic interactions as competition and reaction (= facilitation) which are central to plant succession.

For the above reasons the author of this publication, which is devoted to range cover types, followed the SRM definition (the Plant Ecology rather than the Plant Geography definition) in description of the range cover types and their arrangement within the biomes (ie. inclusion under Grassland vs. Shrubland vs. Forest and Woodland slides).

In context of cover (= dominance) types a unique feature of savannahs is the fact that the community dominants are not those species in the highest layer(s) of vegetation. In forests, woodlands, and shrublands it is the superior (highest or tallest) layer which 1) exerts most control over community environment (eg. most moderates climate, contributes most to soil formation) and 2) whose species are the most competitive and dominate species of lower layers of the vegetation. More than that, in climax vegetation the dominant species are superior competitors in all strata or layers of the plant community (eg. in forests they are the most tolerant species). In savannas "the real control is not in the scattered trees but in the grass-dominated field layer" (Cain and Castro, 1959, p. 181). The larger and taller individual woody plants (trees or shrubs) are not the community dominants in savannahs. That is why the author arranged so many tree- and shrub- containing savannahs (eg. post oak savanna, live oak savanna) under the Grassland Biome slides rather than under the Forest and Woodland Biome. It is also the reason why, as stated above, even more range types should have been included in the Grassland Biome.

McPherson (1997, p. 3) defined savannas as "ecosystems with a continuous grass layer and scattered trees or shrubs". He then observed that ecology and management of savannas were "incompletely defined compared to those of other physiognomic types in North America (eg. forest, desert, grassland)" as well as being "generally under-studied compared to other physiognomic types" (McPherson, 1999, p. 4). Anderson et al. (1999, p. 1) remarked that savanna had "no precise definition" and its meaning varied region-to-region. In Anderson et al. (1999) several North American savanna communities were described but no overall, all-encompassing definition of savanna was offered. Instead individual authors gave their own definition as it fit the vegetation with which they dealt.

Both McPherson (1997, p. 4) and Anderson et al. (1999, p. 2) discussed origin of the term savanna. The former credited Dyksterhuis (1957) for creating acceptance of the term as used in Vegetation Science and particularly as applied in the agricultural disciplines. While Dyksterhuis did not give a definition of savanna per se his detailed description was that which was incorporated into the SRM definition, as would be expected. (Note once again the prominence of the Clements-Tansley-Weaver-Dyksterhuis-Braun lineage of the Anglo-American Tradition.)

It has been accepted generally among rangemen and wildlifers that savannas are biomes (formations, if consideration is restricted to plant community) on par with grasslands, forests, shrublands (scrub like deserts and chaparral), tundra, alpine, marshes, etc. (see definition of rangeland in SRM glossaries). These vegetation transition zones or ecotones have biotic and abiotic features of adjoining plant communities. They also have, regardless of attributes shared with neighboring units of vegetation, unique and distinctive features "all their own". Savannahs are a biome unto themselves. They are almost invariably richer in plant and animal species and differ in ecosystem structure and function as well as components therein such as soils and even internal microclimate. The "edge effect" of ecotonal communities (whether local as in fencerows or regional as in large climax savannahs) usually offer the best of wildlife habitat. Proper plant, animal, and ecosystem management practices routinely differ between savannas and the grasslands, deserts, forests, etc. that are contiguous to them.

Logically then there could— perhaps, should —have been an inclusion of slides of savanna vegetation under this Savanna heading as with grassland, scrub, tundra , alpine, and forest vegettion under the headings of Grasslands, Shrublands, Tundra, Alpine, and Forest and Woodlands.

That would be both consistent and inconsistent because the "pigeon-holing" of range savannahs inevitably becomes problematic, subjective, and judgmental on part of the one who arranges the representative examples of savanna communities. This is due to the continuum nature of vegetation in going from grassland to savanna to forest. "Where to draw the line" is the Gordonian knot. Between a Bouteloua-Buchloe-Hilaria plains grassland and a Juniperus woodland there are apt to be communities ranging from that of shortgrass grassland with a juniper tree "now and again" to the same grassland but with "enough trees" and "close enough together" that even a committee of vegetation scientists would agree that it is a shortgrass-juniper savanna. Problems arise when there are "umpteen" intermediate gradations in between these "in-between" categories. And this may occur within the same pasture, a small one at that. Which, if any, are natural, potential, climax, etc.? Between grasslands of the Palouse Prairie and Artemesia deserts of the Great Basin there is the Artemesia shrub steppe, a shrub savanna. What criteria or even approximate guidelines are to be used just to separate these three once vegetation moves beyond the endpoints of the "pure" types? What density or cover of Artemesia filifolia distinguishes a Southern High Plains mixed prairie, a grassland, with an occasional sand sagebrush plant to a sand sagebrush-mixed prairie savanna?

What are the famed Cross Timbers of Oklahoma and Texas? Is this a unique forest formation, an outlier of the Eastern North American Deciduous Forest Formation, or a contiunously competitive "tug of war" zone between the last outpost of oak-hickory forests and the leading edge of the vast (once vast) continential grasslands? What about the various forms of Cross Timbers that vary from open woodlands with grassy understories to "tangles" from which a ghost could not extract himself? Are the Cross Timbers the southern terminus of the Prairie Peninsula or are they distinct from this immense (once immense) savanna? In Anderson et al. (1999), which is probably the most comprehensive treatise on savannahs in North America, these two were recognized as separate units of vegetation and as savannas and not forests.

Various combinations of California oak species and naturalized annual Mediterranean grasses and forbs can be alternatively: 1) California annual grassland with widely scattered oak trees, 2) annual grass-oak tree savanna, 3) annual grass-oak woodland (the most common form), or 4) closed canopy oak forest with a sparse herbaceous understory. These four forms can exist for the same floristic community (identical flora or species present) and vary only in proportions of the same spcecis based on density, frequency, or cover (and not uncommonly enclosed by the same perimeter fence). The same could be said of the Oregon oak cover types in the Coast Ranges and Willamette Valley.

Then there is the Tamulipan brushland, the Rio Grande Plains or "brush country". What is the climax or the pre-Columbian mix of leguminous shrubs to short, mid- and tallgrass species? Is it potential mixed prairie grassland or subtropical savanna that comports perfectly with the Society of American Foresters definition? Yes, current consensus agreed on savanna, but there are both tree and shrub savannahs. Maybe the climatic climax is a subtropical shrubland and fires kept a grassy understory as a pyric climax. So is the Rio Grande Plains range type a grassland or a shrubland?

In considerations of savannas and their key attributes it is the woody canopy cover that is of most importance in defining and distinguishing savannahs from neighboring grasslands, shrublands, and forests or woodlands. Anderson et al. (1999, p. 2) set the lower limit of canopy coverage at 5-10% and cited work that defined the upper limit of woody cover at 25-80%. McPherson (1997, p. 3-4) set the proportion of woody cover for savannas at "less than 1% to about 30%". Obviously these values were arbitrary even if "reasonable" or "agreeable" to students of savanna vegetation. They brought no solution the problem examples discussed above. Most likely there is no definitive definition or canopy characteristic (eg. cover percentage) adequate to "cover" all savanna communities or cover types for all purposes.

In conclusion, it seemed expeditious to the author to discuss the problems inherit in designation of the savanna biome and to arrange each example of range ecotone under the heading of the non-transitional biome it mostly closely resembled to him. The photographer/author is a grasslander, a prairieman, and interpretations of savanna vegetation were biased toward that perspective. For example, the savanna of Andropogon-Sorgastrum-Panicum understory and Quercus-Carya woodland was included under the Grasslands not Forests and Woodlands heading.

Major ecotones (eg. Prairie Peninsula, quaking aspen parkland, Rio Grande Plains savanna) were those recognized by leading authorities such as Shelford (1963, ps. 306-326), Garrison et al. (1977, ps. 39-43), Vankat (1979, ps. 220-223), Barbour et al. 1999, ps.599-601), McPherson (1997), and Anderson et al. (1999). The latter two are the most definitive contemporary works on savannahs in North America with Anderson et al. (1999) being the more comprehensive and authoritative as it included many authors.

"Barrens" is a unique category of land with species (and, especially, ecotypes) uniquely adapted to it. "Barrens" is also relatively unique as a term there apparently being no accepted definition even among specialists working with the various lands and vegetation described as "barrens". For instance, numerous chapter authors in Anderson et al. (1999) used the term barrens to describe various habitats and plant communities but all said they used the term only in the context of environment and vegetation described by them therein. According to several authors in Anderson et al. (1999) "barrens" was used by frontiersmen and settlers generally to describe areas with shallow, stoney soil capable of supporting sparse grass cover and widely dispersed, short or stunted trees (eg. Anderson et al. 1999, ps. 1-2, 135, 140, 171, 220-221, 309).

Morris (1992) defined "barrens" as "an area of poor, often sandy soil supporting scant vegetation". Allaby (1998) described "barren" as "a community of few and scattered plants that occupy less than half the available ground area" (but he went on to use "barrens" suggesting that the "s" was mistakenly omitted from the entry). "Barrens" as a plant community or habitat was not in either the SRM or SAF glossaries of terms.

While the consensus definition/description of "barrens" states or implies a habitat which is harsh or severe for plant life due primarily to unfavorable edaphic conditions at least one form of barrens, the malpais, often supports plant communities which are more mesic than surrounding vegetation. Even though the malpais has immature (even, infant) soils the lava flows absorb and channel water such as to support greater plant production and species diversity than that of contiguous communities.

Anderson et al. (1999) is also the comprehensive authority on barrens. Kruckeberg (1984) is the ultimate reference for California Serpentine Barrens.

[ Home ]