Longleaf Pine

Forests dominated by longleaf pine (Pinus palustris) once covered much of the upland of the Southeastern United States. The species, one of the southern yellow pines, always has been highly valued for its high quality wood and was once the most important source of turpentine in the entire country. Forests dominated by longleaf are some of the most ecologically significant forest ecosystems of the region and support numerous endemic plants and animals.

Frequent fire is an essential ecological process for both longleaf pine trees and longleaf pine forests (Stanturf et al. 2002). Successful recruitment of longleaf seedlings requires germination on bare mineral soil that is exposed following a burn. In addition, without frequent fire, longleaf-dominated forests are invaded by fire-intolerant species (especially hardwood trees and shrubs) from adjoining ecosystems and the open, high-light understory is eventually is eliminated and ground cover species out competed, resulting in the loss of many associated plant and animal species.

Not only does longleaf pine require frequent fire, many characteristics of the species promote spread of fire across the landscape. Many other ecosystems adjoin or are imbedded in the matrix of longleaf pine-dominated forests. Fires that originate in longleaf forests often influence the structure and species composition of ecotones and the nearby ecosystems, such as adjoining hardwood slope forests. Some ecosystems, because of their small size and position in the matrix of a larger longleaf forest, experience a fire regime that is determined by the interaction between the burns initiated in the surrounding forest and the hydrologic characteristics of the embedded community, such a wet seepage savannas.

• Longleaf Pine Community Description and Distribution
• Historic Fire Regimes of Longleaf Pine
• Fire Effects on Longleaf Pine Woodland Vegetation
• Managing Longleaf Pine with Prescribed Fire

Ecosystem Distribution and Extent

Prior to European settlement, the geographic range of longleaf pine forests covered most of the Coastal Plain of the Southeast United States, from southeastern Virginia to eastern Texas and south to south-central Florida (Peet and Allard 1993, Harcombe et al. 1993). Longleaf extended north into the Piedmont providence and mountains of Alabama and Georgia (see map of distribution). Estimates of original acreage of longleaf pine forests range from 60 million (Outcalt and Sheffield 1996) to more than 93 million acres (BROKEN-LINK Frost 1993).

The current geographic distribution of longleaf is similar to the historic one although it has disappeared on the fringes of its original range and is almost extirpated from Virginia (Plocher 1999). Outcalt and Sheffield (1996) documented that longleaf pine currently occupies less than 5% of its original acreage but it is not known how much of the current acreage represents forested areas with native groundcover. In all states except Florida, the majority of remaining stands of longleaf pine are found on private property (Outcalt and Sheffield 1996).

According to Peet and Allard (1993), longleaf pine vegetation is distributed across four physiographic provinces: Coastal Flatlands, Coastal Plain Rolling Hills, Fall-Line Sandhills, the Piedmont and the Interior Uplands. Longleaf woodlands historically comprised the majority of the upland vegetation in the two Coastal Plain provinces and the Sandhills region. In the Piedmont longleaf is, or was, limited to the outer fringes and mountain longleaf is limited to a very few scattered south facing ridge sites in Georgia and Alabama.

Environment

Climate

Longleaf pine forests occur mostly in the Coastal Plain region of the southeastern United States (Peet and Allard 1993, Harcombe et al 1993). Climate of this region is classified as humid subtropical (Trewartha 1968) and is characterized by relatively warm winters, hot summers, abundant rainfall and a long growing season (Twilley et al. 2001, Georgia State Climate Office 1998). Precipitation is distributed throughout the year, but winter and late spring-summer tend to be the wettest periods (Twilley et al. 2001, Chen and Gerber 1990). Winter precipitation is mainly a consequence of frontal cyclonic storms, whereas convection driven thunderstorms account for much of the growing season rainfall (Christensen 1988). In wet years, thunderstorms may occur almost daily from late May through September. Lightning associated with thunderstorms is very common. Indeed the Coastal Plain region has the highest frequency of lightning of any region in North America (Komarek 1968). High lightning frequencies in this region are of fundamental ecological importance since lightning ignitions are thought to have contributed to a historic regime of frequent growing season fires (Komarek 1968, Frost 1998). Within the Coastal Plain, as a whole, there is some geographical variation in climate regime (Christensen 1988, Cook et al. 2001). Areas remote from the coast tend to be cooler and drier. Also the Atlantic and Eastern Gulf Coast sections of the southeastern Coastal Plain tend to be moister than the western Gulf Coast section (Marks and Harcombe 1981, Christensen 1988). The precipitation gradient is especially steep towards the western edge of the range of longleaf pine (Cook et al. 2001).

Soils

Soils of longleaf pine stands are predominantly coarse textured sands, loamy sands or sandy loams that dry rapidly after rain events (Marks and Harcombe 1981, Harcombe et al. 1993, Provencher et al. 2001b). The combination of dry soils, flammable herbaceous dominated groundcover vegetation, and longleaf pine needle litter, creates ideal conditions for fires to start and spread rapidly across the landscape (Frost 1998). Ridges that originated as stranded beach dunes or sandbars near ancient streams are underlain by Entisols, i.e. deep undifferentiated coarse sandy soils. These sites are extremely xeric and are inhabited by sandhill vegetation typified by subcanopy oaks and hickories as described above (Myers 1990). Most other longleaf pine sites occur on Ultisols or Spodosols, soil orders characterized by argillic or spodic horizons, respectively. These subsurface horizons impede the downward flow of water, and, consequently, water accumulates above the subsoil following rainfall events. If the impediment to drainage is located very close to the surface the soil may remain moist for considerable periods since even minor rains produce saturated soils. This will be true to some extent even on flat sites, but moist soil conditions tend to be especially pronounced in seeps or depressions where water accumulates.

Differences in subsurface drainage can produce pronounced differences in longleaf pine ground layer vegetation composition (Walker and Peet 1982, Peet and Allard 1993, Streng et al. 1993, Provencher et al. 2001, Glitzenstein et al. 2001, 2003). Flat sites with shallow subsoils are generally inhabited by wet savannas whereas seeps and shallow depressions support bogs. As the impermeable soil horizon recedes further beneath the surface the surface soil remains drier for longer periods of time since a greater quantity of rain is required before capillary action can enhance surface soil moisture. These drier, better-aerated soils are more conducive to woody plant growth and, even when frequently burned, often support abundant shrubs (Glitzenstein et al. 2003). Since they occupy level topography and incorporate an important shrub component, these infrequently saturated sites have are often referred to as flatwoods (Abrahamson and Hartnett 1990).

Soil texture and drainage are two environmental parameters with important direct effects on species composition of longleaf-associated vegetation (Glitzenstein et al. 2001, Marks and Harcombe 1981, Harcombe et al. 1993, Gilliam et al. 1993, Kirkman et al. 2001, Provencher et al. 2002). These and other environmental parameters affect plant species composition independently of or in addition to any fire regime. Experimental work by Glitzenstein et al. (2001) in South Carolina demonstrated that species characteristic of sandhill or subxeric sites were most prominent on the higher parts of a manipulated gradient with the deepest surface soils. In contrast, species characteristic of moist flatwoods occurred mostly in the lower, poorly drained sections of the gradients. In addition, for any given drainage condition there was a significant effect of soil texture.

Ecosystem Description

Old-growth Condition and Forest Structure

On a landscape-scale, structure of pre-European settlement longleaf pine ecosystems was one dominated by widely spaced large trees that created an open canopy forest. Comments of many early explores of the region included mention of the openness of the forest under the canopy trees. Bartram (1791) applied the term “park-like” to large areas of scattered large canopy longleaf pine trees in the lower Coastal Plain. For a site in North Carolina, Ruffins (1861) described a “thick canopy of green” … “over the open space below” and for a large wet moist he termed a “great savanna”, he noted that “nearly the whole” was “destitute of trees, and nearly so of bushes, and of any shrubs of as much as two feet high”.

Old-growth stands of longleaf pine are rare in the modern landscape. Varner and Kush (2004) estimate that in 2004 old-growth longleaf stands occupy 0.004% of the extant range and 0.00014% of the pre-European settlement extent. These relics provide visual insight into the original longleaf forest. In addition, the USDA Forest Service has endeavored to define old-growth conditions for many longleaf pine and other important forest ecosystems. Landers and Boyer (1999) describe the old-growth condition for upland longleaf pine forests, woodlands and savannas. Harms (1996) does the same for similar but wetter ecosystems. Walker (1999) provides an overview of processes important in maintaining old-growth conditions in longleaf pine forests.

Recent research on some of the remaining old-growth sites, provides quantified descriptions of what may represent original structure of longleaf pine forests (Platt et al.1988, Noel et al. 1998). The quantified descriptions suggest open canopied forests with widely-spaced, single species patches of trees of multiple ages (sizes). The absence of fire and logging alters the habitat structure and species composition of the forest (Gilliam and Platt 1999, Noel et al. 1998). In addition, long periods without fire (e.g., fire suppression) results in large increases in fuel making re-introduction of fire problematic (see Effects of Fire Suppression). Currently there are some attempts at burning long-suppressed stands, in an effort to restore the original forest without killing all of the mature trees (Varner et al. 2003, Kush et al. 2000).

Major Species

In the modern landscape, well-maintained, frequently burned, longleaf pine sites are characterized by a sparse, often nearly monospecific canopy of longleaf pine itself, along with a diverse ground layer vegetation of grasses, forbs, and short, rhizomatous shrubs (Wells 1942, Bridges and Orzell 1989, Taggart 1990, Harcombe et al. 1993, Peet and Allard 1993). This matches the descriptions of Bartram (1791), one of the earliest botanists to comments on regional floral diversity of longleaf pine forests. On a scale of meters to tens of meters, longleaf pine groundcover is the most diverse vegetation in North America (Peet et al. 1983, Peet and Allard 1993, Walker and Peet 1983).

Grass species composition and abundance is especially important in longleaf ecosystems because of this plant family’s relationship to availability of fine fuel that promotes fire in these forests. Species of wiregrass (Aristida), little bluestems (Schizachyrium) and bluestems (Andropogon) are of particular importance. Wiregrass (Aristida stricta and/or Aristida beyrichiana; see Peet 1993) is generally the dominant grass species in the eastern section of the outer Coastal Plain, from southern North Carolina to eastern Mississippi (Peet 1993). However, there is a curious “gap” in the distribution of wiregrass in central South Carolina (Peet 1993). Herbarium records (Peet 1993) and historical observations (Elliott c. 1810 unpublished correspondence to Henry Muhlenberg, Ruffin 1843, BROKEN-LINK Ravenel 1850) suggest that this “wiregrass gap” is not a function of land use history but a long-standing biogeographic feature. Within the wiregrass gap, little bluestem (Schizachyrium scoparium) is generally the prevalent grass species on mesic to dry sites whereas toothache grass (Ctenium aromaticum) or bushybeard (Andropogon glomeratus) are dominant in wet savannas. Bluestem grasses also tend to replace wiregrass in the inner Coastal Plain and piedmont areas of the Carolinas, Georgia and Alabama (BROKEN-LINK Frost 1993). Lastly, little bluestem is also the dominant grass in longleaf pine woodlands west of the range of wiregrass in western Mississippi, Louisiana, and eastern Texas (Bridges and Orzell 1989, Harcombe et al. 1993).

It should be emphasized that these generalizations concerning regional shifts in grass dominance may be true only at a coarse scale. Numerous grass and shrub species are capable of assuming dominance on any given site, and lack or reduced importance of wiregrass and bluestems does not necessarily indicate a history of disturbed soils or agricultural abandonment. This is particularly true of wet savannas in which muhly grass (Muhlenbergia expansa) and various dropseeds (e.g. Sporobolus curtissii, S. teretifolius) may dominate or co-dominate along with wiregrass and toothache grass (Taggart 1990, Hermann 1995, Weakley 1998).

Considering canopy trees only, longleaf pine occurred historically in almost pure stands (Schwarz 1907, Harper 1962, Chapman 1909, 1932, Heyward 1939, Frost 1993, Harcombe et al. 1993) and this is still the case for the best, fire-maintained sites. There are, however, a number of exceptions to this generalization (Heyward 1939). “Sandhill” oaks (Quercus laevis, Q. incana, and Q. margaretta), and, in some areas, sand hickory (Carya pallida) can form a more or less continuous subcanopy on dry ridge sites. While the importance of these subcanopy “scrub” oaks and hickories may have been enhanced by altered fire regimes (BROKEN-LINK Myers 1990), there is good historical evidence that they occurred commonly in presettlement forests (Catesby 1771, Ravenel 1850, Watson 1926, Heyward 1939). Another exception is in wet savannas where pond pine (Pinus serotina), slash pine (Pinus elliottii) or loblolly pine (Pinus taeda) may co-dominate, depending on geography and local site conditions (Watson 1926, Heyward 1939). At the very wettest extreme pond cypress (Taxodium distichum var. nutans) replaces the pines. Relative canopy dominance of longleaf pine also varies geographically. This is particularly evident in the transitional zones between the main longleaf pine areas and the mixed pine-oak-hickory communities prevalent to the north and west (Frost 1993, Harcombe et al. 1993). In eastern Texas there is a region of mixed longleaf-loblolly pine forest transitional to the loblolly pine zone that forms the western limit of southeastern pine dominated woodlands (BROKEN-LINK Frost 1993, Harcombe et al. 1993).

Associated Ecosystems

An interesting and unresolved question concerns the role of fire, and fire regime, in determining the boundary between longleaf pine and hardwood dominated plant communities. If fires are frequent longleaf pine and associated species may move down-slope into at least the drier hardwood dominated habitats. On the other hand, reductions in fire frequency, or fire exclusion, leads to invasions by hardwood trees into longleaf uplands (see Effects of Fire Suppression). Historically, it is probable that the boundary between the two habitats fluctuated depending on recent fire history. It is also likely that upper slope hardwood-dominated communities depended on at least occasional fires to limit encroachment of fire-intolerant trees of lower slope habitats (see: Hardwood Hammocks).

The most conclusive sources of fire history information are tree-ring (especially fire scar) and lake-sediment data, techniques that allow reasonably unambiguous determinations of fire dating (Frost 1998). Unfortunately, published data on longleaf pine fire history are not available for either method. The paucity of fire scar studies is a function of the relative lack of old growth trees, the reluctance to destructively sample those few old trees that remain, and the rarity of fire scars (Frost 1998) since typical low intensity surface fires do not damage the vascular cambium of mature longleaf pine trees. Consequently scarring generally does not occur and the fire goes unrecorded. Despite this problem, fire-dating studies could perhaps be carried out in longleaf pine stands with remnant old growth trees by focusing on the occasional trees that had sustained damage sufficient to cause scarring. Such trees are potentially valuable sources of fire history data since once scarred, e.g. in a locally intense fire, a tree becomes much more susceptible to subsequent scarring.

Forest fires produce airborne charcoal particles that can become incorporated as layers in lake-sediments. Interpretation of such data is somewhat problematic since fires occurring anywhere within the vicinity of a given lake may contribute to charcoal accumulation in the lake. Thus, lake-sediment data provide an integrated picture of fire history around the lake rather than a fire frequency specific to a particular site. Preliminary studies utilizing this approach indicate essentially continuous fire beginning around 5,000 y BP. Prior to that date fires were still very frequent, but slightly less so (J. Porter, personal communication). Increasing fire around 5,000 y BP is hypothesized to result from the onset of anthropogenic burning by Native Americans. Thus the available data suggest an ecosystem adapted to frequent fire managed by humans for millennia with even more frequent fire. However, these conclusions are based on analyses from just a handful of sites and much more work is needed.

Given the absence of truly conclusive information, fire history of the longleaf pine region has been inferred from a variety of sources including historical observations, land form, lightning ignition rates, and remnant vegetation patches (Wright and Bailey 1982, Christensen 1988, Harcombe et al. 1993, Frost 1998). Utilizing similar historical sources, different authors have nevertheless inferred significantly different historic fire frequencies. For example, Christensen (1988) suggested fire return intervals in the 3 to 10 year range, whereas Frost (1998) postulated 1-3 year between fire intervals for the outer Coastal Plain and 4-6 years for the inner Coastal Plain. Harcombe et al. (1993) estimated a median fire return time of 3 years for east Texas longleaf forests, but with longer fire free intervals for smaller longleaf patches. One explanation for the differing estimates produced by Harcombe et al. (1993) and Frost (1998) for eastern Texas concerned the definition of patch size. Both studies assumed that larger patches would burn more frequently, given an equivalent rate of lightning ignitions. Frost (1998), however, based his patches, or “polygons”, on Hammond’s (1964) land form map, whereas the polygon definition in Harcombe et al. (1993) was based on an early US Forest Service map of extant longleaf vegetation. One other point to consider about Frost’s (1998) estimates is that they incorporate a certain degree of circularity, at least where management is concerned. That is, the existence of remnant longleaf groundcover vegetation is used to infer a historic fire frequency that is then used to suggest an appropriate prescribed burn regime for maintaining the observed vegetation. Whereas this may be a valid approach, all things considered, practitioners should at least be aware of the circularity.

Historic fire season has also been inferred from lightning ignition data and climate (Komarek 1968, Frost 1998). Lightning strikes in the southeast USA are primarily associated with summer thunderstorms. Consequently, it has been suggested that most “natural” fires occurred during the growing season. Another related suggestion is that most of the landscape actually burned in late spring, at the beginning of the thunderstorm season, when fuels are driest and fires could burn uninterrupted over large distances.

Given the uncertainties involved with fire history reconstruction in longleaf pine stands readers should keep in mind that history is only one of several factors to consider when planning a prescribed burn program. For example, experimental data on fire responses at the population or community level may allow us to tailor burn regimes to meet particular objectives, e.g. maximizing persistence or population growth of rare and endangered species (Gray et al. 2003). A related point is that historical burn regimes can be inferred somewhat from community responses. If it is correct that the disturbance regime that maximizes species richness is the one to which most plants and animals are adapted (Denslow 1985), it should be possible to infer the historic disturbance regime from management experiments. In the case of longleaf pine plant and animal communities the results of management experiments, including prescribed burn studies, reinforce historical interpretations in suggesting the need for frequent, low intensity fire (Glitzenstein et al. 2003, Gray et al. 2003).

Anthropogenic Fire

There is substantial information indicating that humans have used fire in landscape management of southeastern Coastal Plain ecosystems for at least several thousands of years (Pyne 1982). In addition to the lake sediment data discussed above there are numerous historical observations of Native American fires (Pyne 1982). Lastly, phytolith data suggest that fires set by Native Americans may explain the existence of longleaf pine islands in otherwise scrub dominated areas in central Florida (Kalisz et al. 1986). Native Americans apparently burned to improve habitat for game species, to reduce annoying insects, and perhaps for aesthetics. According to Lawson (1709) Native American burning occurred primarily in autumn when fuels began to dry after the summer thunderstorm season and temperatures became more comfortable. European and African Americans rapidly adopted the Native American practice of woods burning (Pyne 1982). For example, Elliott (1824) and Ruffin (1843) allude to the common practice of burning the woods every spring to improve forage for cattle. These references document a shift in season of anthropogenic fire from earlier fires set by Native Americans. Official industrial and government fire suppression policies began towards the end of the 19th century. However, “woods burning” by rural people continued throughout and probably saved what remains presently of the longleaf pine ecosystem (Pyne 1982). Even today, suppression of arson fires is a major occupation of the US Forest Service in southeastern National Forests. Most current fires, however, are legal prescribed burns set by private landowners or state and federal agencies.

See: Effects of Fire Suppression on Longleaf Pine Woodlands and Savannas

Shifts in Species Composition

The characteristic overstory of widely spaced longleaf pine trees and diverse grass-dominated ground layer is dependent on frequent understory fires (see: Historic Fire Regimes). In the absence of fire a major transformation occurs leading to the loss of longleaf pine itself as well as the numerous fire dependent ground layer species (Heyward 1939, Lemon 1949, Brockway and Lewis 1997, Kush et al. 1998). Rates of species loss with time since fire appear to vary with edaphic conditions, being most rapid on mesic sites and slowest on very dry or wet sites (Streng and Harcombe 1982, Maliakal et al. 2000). The ultimate fate of fire suppressed longleaf pine stands is somewhat debatable. In the initial stages, such stands are likely to be dominated by hardwood tree and shrub species already present as sprouts in the fire maintained woodland (Menges et al. 1993, Abrahamson and Abrahamson 1996, Maliakal et al. 2000). Menges et al. (1993) documented a variety of long-term successional patterns in central Florida longleaf pine forests. The type of succession was dependent on edaphic factors as well as landscape factors, including propagule availability. Menges (1993) suggested that drier upland sites might eventually converge on some type of xeric hammock community. However, long-term data from a fire suppressed subxeric longleaf pine site in eastern Texas demonstrated that even xeric hammock or upper slope oak-hickory species are ultimately dependent on fire at some return interval. Following approximately 50 years of fire exclusion this site is succeeding to a closed stand dominated by upland laurel oak (Quercus hemisphaerica), American Holly (Ilex opaca) and southern magnolia (Magnolia grandiflora).

Effects of fire suppression on fauna are predictable based on habitat relationships. Species dependent on open woodlands and savannas decline with fire exclusion, whereas species that prefer closed woods increase. Generally the latter are common species in the modern landscape whereas the former are often rare and endangered (e.g. red-cockaded woodpecker, gopher tortoise, indigo snake). Thus maintaining open frequently burned pinelands is critical for animal as well as plant conservation.

Altered Fire Behavior

Low intensity surface fires characterize fire behavior in well-maintained longleaf pine stands with good quality groundcover. On the other hand, fire suppressed longleaf pine sites are generally less likely to burn, but are more prone to high intensity fires when they do burn. Streng and Harcombe (1982) discussed effects of fire exclusion on fire behavior in longleaf pine woodlands. Initially pine litter and fine woody fuels accumulate, increasing the probability of higher intensity fires. Long-term fire exclusion, however, results in invasion of hardwood trees, declines in grasses and other fine fuels, decreased fuel height and aeration. The consequence of these changes for predicted fire behavior depends on fuel moisture, which is in turn a function of time since rain and current weather conditions. Over most of the range of fuel moistures fires are less likely to ignite and spread, and flame lengths and estimated fireline intensities are considerably lower than in frequently burned fire maintained sites. However, at low fuel moistures predicted fireline intensities are considerably higher than in frequently burned sites due to high loading of heavy fuels.

Saw palmetto dominated flatwoods represent an exception to the above scenario (Heyward 1939). Due to the unique properties of this terrestrial palm, communities dominated by this species can burn with high intensity within 2-3 years post fire. Furthermore, dense saw palmetto stands can resist invasion of less pyrogenic hardwoods for considerable periods of time. The consequence is an already flammable base fuel that continues to increase in flammability with time since fire. It is not surprising that the most dangerous and economically destructive fires in the southeastern US in recent years have occurred in fire suppressed saw palmetto flatwoods.

The effect of fire on vegetation in high quality longleaf pine land is to maintain the existing plant community. In other words, the effect of a single burn in a habitat accustomed to, and maintained with, frequent fire is remarkably slight (Frost 1998). A few longleaf pine trees may be killed, primarily in the smaller size classes (Glitzenstein et al. 1995). Depending on fire season and intensity a certain percentage of tree-sized scrub oaks and hickories may be top-killed most of which will re-sprout (Glitzenstein et al. 1995). The above ground parts of ground layer plants will be mostly consumed by the fire, but the below ground parts will mostly survive to send up new shoots. Annual or biennial species present prior to the burn will germinate and grow to maturity following the fire. Within a few months post-fire the entire community will be reconstituted, to all appearances the same as before.

In contrast, effects of individual fires in long-unburned longleaf stands can promote meaningful changes, depending on time since fire, accumulated fuels, and burn conditions on the day of the burn (e.g. Plocher 2003). Prolonged periods of fire suppression can lead to invasions of hardwood trees, declines in grasses, and consequent reductions in community flammability (Streng and Harcombe 1982, Maliakal et al. 2000). Prescribed burning under these conditions may be difficult or ineffective, though Provencher et al. (2001a) found that even patchy, low intensity fires promoted increases in ground layer species richness in long unburned west Florida sandhills. On the other hand, accumulated woody fuels can lead to high intensity fires and substantial canopy mortality. For example, thirty years of fire exclusion in relict Virginia longleaf pine stands resulted in prescribed fires with flame heights up to 6 m and post-fire canopy mortality rates ranging from 19 to 46.7% (Plocher 1999).

Fires in long fire-suppressed sites may produce extreme results where the ensuing stand consists of a relatively few surviving longleaf pine trees along with hardwood sprouts and reduced populations of the typical perennial longleaf ground layer plants. In place of the latter, there may be a pronounced increase in “weedy” annuals and short-lived perennials germinating from buried seeds in the soil. These stands are often dominated by such species as pokeweed (Phytolacca americana), pinweed (Hypericum gentianoides), fireweed (Erechtites hieracifolia), Polypremum procumbens, and dogfennels (Eupatorium compositifolium, E. capillifolium). These ruderal (weedy) species are generally not encountered in well-maintained fire frequented sites except in very localized soil disturbances or hotspots (Maliakal 2000, Glitzenstein et al. 2003).

Because effects of individual fires are slight in properly maintained longleaf pine sites it is more productive to discuss effects of burn regimes, i.e. cumulative impacts of repeated fires occurring over long time periods. Most authors recognize three components to a fire regime: fire frequency, fire season, and fire intensity (DeBano et al. 1998).

Effects of Fire Frequency

Some effects of fire frequency on longleaf vegetation are relatively well established whereas others are still uncertain. In general, increases in fire frequency are correlated with increased herbaceous dominance, particularly of grasses, and decreased importance of woody plants (Waldrop et al. 1992, Beckage and Stout 2000, Glitzenstein et al. 2003). Fire frequency effects on vascular plant species richness are less well established. Three different relationships have been documented in the literature:

1. Two studies did not find a significant effect of fire frequency on species richness. Beckage and Stout (2000) did not find a fire frequency effect on species richness in a central Florida sandhill and Brockway and Lewis (1997) found no significant differences among three fire frequency treatments (annual, biennial, triennial burning) in a Georgia flatwoods. Beckage and Stout’s (2000) study was correlational rather than experimental and the lack of a significant effect may have been due to low statistical power. Although Brockway and Lewis’ (1997) reported no significant effect among annual, biennial, and triennial burning treatments, all three of these fire frequencies produced species richness levels (34.9 plant species) that were approximately twice that measured on long unburned (>40 years) forest plots (17.5 plant species).
2. Mehlman (1992) found that species richness in the long-term Stoddard Fire Plots at Tall Timbers Research Station increased as fire frequency increased, but tended to plateau at the shortest fire return intervals. Beckage and Stout (2000) further clarified this mathematical relationship in Mehlman’s (1992) data and termed it the “saturation effect”. The Stoddard plots are located in old-field loblolly stands, and may, historically, have been dominated by shortleaf pine (R. Masters, personal communication), so the relevance of these results to undisturbed longleaf sites may be debatable.
3. Data from long-term studies of longleaf pine flatwoods sites in Francis Marion NF, SC, and Osceola NF, FL indicated a linear relationship between fire frequency and species richness (Glitzenstein et al. 2003). This relationship was strongest at small scales and became marginally non-significant at the largest scale investigated (0.1 ha). If differing relationships between fire frequency and species richness are in fact real properties of different habitats or regions, and are not just artifacts of experimental design and analytical procedures, the implications for management would be significant. More studies are clearly needed.

In addition to the above studies of fire frequency effects on species composition, there is also a recently published study of the influence of fire frequency on population scale parameters of rare species. Gray et al. (2003) investigated effects of fire frequency on extinction, colonization and persistence of 32 rare longleaf ground layer plants at Fort Bragg and Camp Mackall military bases in southeastern North Carolina. They concluded that annual or biennial burning was needed to maintain or increase the numbers of populations of these rare species. Less frequent fires resulted in an unacceptably high rate of local extinctions.

Effects of Fire Season

Season-of-burn effects on longleaf pine vegetation were reviewed by Robbins and Myers (1992) and by Streng et al. (1993). Both reviews concluded that season of burn might be important, but that interpretation of many published studies was complicated by lack of replication, pseudo-replication, and potentially confounding effects of fire behavior and year of burn. Subsequent studies have also produced inconsistent results. The following conclusions may apply.

1. Compared to fires at other seasons, there appears to be a consistent positive effect of growing season fires on flowering and fruiting of dominant grass species and some composites (Biswell and Lemon 1943, Parrot 1967, Streng et al. 1993, Brewer and Platt 1994). Brewer and Platt (1994) also documented a positive effect of growing season fires on clonal proliferation in the dominant longleaf sandhill forb grass-leaved goldenaster (Pityopsis graminifolia).
2. In contrast, detailed studies of legumes have not detected an effect of season of fire on reproductive biology or nitrogen fixation rates (Hiers et al. 2000, 2003).
3. Growing season fires appear to enhance top-kill rates of oak trees and other tree-sized hardwoods in longleaf pine stands (Robbins and Myers 1992, Streng et al. 1993, Boyer 1993, Glitzenstein et al. 1995, Provencher et al. 2001a). Partly this is a function of fire behavior. That is, dry windy conditions promoting more intense fires are especially likely to occur during spring months. Fire behavior alone does not explain the entire effect of season of fire on oak topkill, so there is probably a residual effect of tree physiology as well (Glitzenstein et al. 1995).
4. Long-term annual fires in the growing season can eradicate woody shrubs and hardwood sprouts (Waldrop et al. 1992). However, a similar effect is not evident for fire return intervals of 2 or more years. Two short term studies investigating effects of fire season on shrub stem densities did not find an anticipated strong negative effect of growing season fires, though burning during the growing season did somewhat inhibit in-growth of new woody stems (Olson and Platt 1995, BROKEN-LINK Drewa et al. 2000).
5. Except for the effect of annual summer fires, long-term experiments have not demonstrated a strong impact of fire season on ground layer species composition in longleaf pine savannas (Waldrop et al 1992, Streng et al. 1993, Kush et al. (2000), Glitzenstein et al. 2003b). Kush et al. (2000) demonstrated a rather striking negative effect of spring and, especially, summer biennial burning on legumes. Given the results of Hiers et al. (2000, 2003) discussed above, this finding is difficult to interpret. After 22 years of biennial fires, Kush et al. (2000) also found the highest overall species richness in winter burn plots, but the differences were relatively minor (114 species in the winter burn plots, 104 species in both spring and summer burn plots).

The inconsistent results discussed above provide little guidance for managers. Probably the most important “take home message” is that fires need to occur frequently regardless of season (Glitzenstein et al. 2003, Gray et al. 2003). Given the probable importance of growing season fire as an evolutionary force, and indications that reproductive biology of at least some species is cued to growing season fire, it seems likely that it is important to include some growing season burns as part of a prescribed burn regime. However, it does not seem necessary, or perhaps not even desirable, for all burning to be done during the growing season.

Effects of Fire Intensity

Fire intensity is a third component of fire regime. However, there is relatively little known about effects of this factor in longleaf pine stands. Glitzenstein et al. (1995) demonstrated differences in tree mortality rates for both oaks and pines in response to variation in fire temperatures and various measures of fire intensity. Williamson and Black (1981) and Platt et al. (1991) documented spatial variations in fire temperatures due to needle deposition patterns, with temperatures under or around clusters of longleaf pine trees generally greater than those further away from pines. Rebertus et al. (1989) demonstrated higher rates of sandhill oak topkill nearer to longleaf pine trees, and suggested that this might be a fire temperature effect. Hierro and Menges (2002) experimentally manipulated fuel loadings, and hence fire intensities, in south-central FL pine flatwoods. Shrub responses were monitored pre-fire and four years post-fire. Results indicated similar responses regardless of fuel loading treatment. Densities of some clonal shrubs, particularly Quercus minima and Ximenia americana, increased markedly after fires. Most other species had returned to pre-fire levels by the end of the study period. Overall shrub species richness increased post-fire. Hierro and Menges (2002) concluded that fire intensity effects on ground layer shrub communities in pine flatwoods “are likely only with extremely intense burns”. More such studies are needed, with data collected on herbs as well as shrubs. Also there is a critical need for predictive models of fire intensity effects on vegetation, incorporating such data as plant physiological responses, soil and duff buffering, and fire weather influences.

In general, data suggests that fire effects in longleaf pine woods are almost entirely beneficial, and indeed critical for the continued survival of this ecosystem. However there is much to be learned about effects of different fire regimes in particular habitats or regimes. It is entirely possible that a prescribed burn regime that is suitable for longleaf pine savannas in the outer Coastal Plain of South Carolina or Florida may not be suitable for an inner Coastal Plain site in Alabama or a west Gulf Coast site in eastern Texas.

Frequent fire is essential for maintaining longleaf pine woodlands and savannas in a healthy condition. In the modern landscape of the southeastern US the only practical way of accomplishing this objective is prescribed burning. Because plants and animals will be adapted to the historic fire regime, the specified prescribed burn regime should conform closely to the best guess historic burn regime for the particular location.

Southeastern US habitats dominated by longleaf pine have declined precipitously in area and extent (Frost 1993). Consequently, restoration of such habitats is now a high priority for conservation. Re-initiation of frequent prescribed burning is an absolutely critical first step in restoration of longleaf pine woodlands and savannas. Almost any fire is likely to have some positive effects (Provencher et al. 2001b). However, prescribed fire alone may not be sufficient for restoration (Provencher et al. 2001b). There are two reasons why this may be the case. First, long periods of fire exclusion lead to structural changes in the plant community that may not be reversible with prescribed burning (see: Effects of Fire Suppression). Tree species other than longleaf pine can become established and reach a fire resistant size during long fire free intervals. Overly dense stands of trees decrease light and nutrient availability and consequently inhibit recovery of herbaceous-dominated ground layer plant communities. Hardwood tree species, in particular, can also modify the local environment in such a way as to reduce the likelihood of ignition and subsequent fire spread (Williamson and Black 1981, Streng and Harcombe 1982). It may be necessary to remove undesirable tree species and to thin overly dense pine stands in order to restore a more appropriate stand structure before reinitiating a prescribed burn program.

The second reason why fire alone may be insufficient for restoration is that certain plant and animal species that historically occurred on a site may no longer be present. Loss of species may occur as a consequence of prolonged fire exclusion (Lemon 1949, Abrahamson and Abrahamson 1996, Maliakal et al. 2000) or a history of intensive soil disturbance due to agriculture or forestry. The missing species may include longleaf pine itself as well as many of the typical ground layer plants and arthropods. Selective cutting and reintroduction of fire may create the appropriate environmental framework to support populations of these characteristic longleaf species but this will be of little use if they are not actually present at the site. It may therefore be necessary to reintroduce missing species. A variety of approaches are available to accomplish this objective, including direct seeding, nursery propagation and out-planting, and translocation of mature plants (Glitzenstein et al. 2001). Results to date indicate that most longleaf ground layer plants are relatively easy to propagate. Consequently, re-initiation of new populations is not particularly difficult if a stand is managed appropriately with frequent prescribed fire (Seamon 1998, Glitzenstein et al. 2001) and competition from canopy trees is not too great (Mulligan et al. 2002).

Management Considerations for Wildlife

The surest method for maintaining the characteristic fauna of longleaf pine ecosystems is to maintain the appropriate historical burn regime and overall site integrity (Engstrom 1993, Guyer and Bailey 1993, Folkerts 1993, Means et al. 1996, James et al. 2001, Provencher et al. 2001b). Site integrity includes the micro-topographical, hydrological and floristic components of the site. Activities that destroy these components, e.g. agricultural usage, ditching, bedding, intense mechanical site preparation, and some broadcast herbicides, can permanently damage the habitat for many species of animals and plants. Restoration of some or all of these components will improve habitat suitability and some animal species, particularly mammals and birds, may respond rather quickly (Engstrom et al. 1984, Provencher et al. 2001b). Insects, herpetofauna, and some of the rarer birds tend to perceive the environment on a finer scale, and these groups may be more difficult to restore. Insects and smaller herps may also be dispersal limited, and, like plants will need to be reintroduced. Despite these concerns it is worth noting that restoration science is still in its infancy, and promising studies for a number of groups of organisms are just now being initiated (Colby 2000, Provencher et al. 2001b).

See also: Using Prescribed Fire to Manage Wildlife Habitat

Using Prescribed Fire to Control Insects and Diseases

Although wildfires can increase insect infestations, some authors have suggested that typical low intensity prescribed fires may help to limit populations of damaging insects (Brennan and Hermann 1994). For example fires in early spring may kill bark beetle larvae that are in diapause within the forest litter at that time.

See also: Using Prescribed Fire to Control Insects and Disease

Key Issues in the Implementation of Prescribed Burning

Compared to ecosystems adapted to longer fire return intervals, prescribed burning of longleaf pine is a relatively simple matter. For example, a five-person burn crew with two drip torches and a pumper truck for backup can burn 1000 hectares or more per day of healthy longleaf with an herbaceous dominated understory. Aerial ignition via helicopter, an increasingly common method, allows for prescribed burns this large or larger in a few hours time. Fires administered at the proper high frequency are low in intensity and relatively easy to control.

With some caveats discussed below, implementation issues pertain not so much to the practicalities of how to carry out the fires, but to research questions related to effects of different fire regimes. Pending further work, the historic fire regimes inferred by Frost (1998) may serve as the best guide for prescribed fire as well, with the caveat that fire frequency goals should take precedence over fire season goals when the two are in conflict (Glitzenstein et al. 2003).

Without doubt, keeping prescribed fire as an important land management tool is the most critical issue relating to conservation and management of the longleaf pine ecosystem. Unfortunately, spreading suburbia increasingly hampers use of this critical management tool. According to Wade (1993, p. 352),

?The longleaf pine ecosystem is currently facing another crisis. The land base of this once vast ecosystem is disappearing at an alarming rate (less than 4% of the original 60 to 90 million acres remains in longleaf). Since at least the early 1970?s immigration to Florida has averaged more than 1000 people a day. A seeming insatiable demand for single family homes in the suburbs and the attendant network of highways has displaced the natural landscape. Moreover, fire and smoke generally are not welcomed along the urban-wildland (WUI) interface or along transportation corridors. This, in turn, makes it much more difficult to perpetuate remaining fire-dependent wildland areas.?

Wade (1993) suggests that sensitivity on the part of fire managers to public concerns, along with education about the importance of prescribed fire, may allow managers to continue to use prescribed fire even in WUI areas. One important potential limitation to continued use of prescribed fire is the 1970 Federal Clean Air Act (renewed in 1990)(Achtemeier et al. 1998). This law mandates regional limits on various types of air pollutants including particulate matter and carbon monoxide, two of the main components of smoke. Whereas smoke from woods fires is generally considered a minor source of these pollutants, reducing smoke may be more expedient politically than regulating other sources of these pollutants. Furthermore, heavy local concentrations of prescribed fires can exceed recommended local air quality standards. Recent rule changes by the federal Environmental Protection Agency (EPA), the agency charged with administration of the Clean Air Act, do take into account the need (and legal obligations) for prescribed burns in natural resource management. The rules allow for development of burn plans by land management agencies that are then reviewed by EPA. The latter is then allowed some discretion in not designating the land management agency as ?in violation? during periods when prescribed fires contribute significantly to fine particulate (PM 2.5 and PM 10 loads). Despite these efforts at accommodation, air pollution concerns significantly complicate the lives of burn managers and likely lead to fewer prescribed fires.

While longleaf pine forests are commonly associated with the Coastal Plain, they exist over a wide geographic range that includes mountainous areas of Alabama and Georgia. Fire is a vital agent in montane longleaf pine, just as it is throughout the range this forest type. However, very little has been reported on the use of fire in montane longleaf pine. Most of what is known is from anecdotal accounts that reported that the results from burning in the mountains was not that much different from burning on the Piedmont or Coastal Plain.

The first report in the literature of longleaf pine in the montane region of northeast Alabama and northwest Georgia came from Mohr (1897). He found longleaf pine on isolated ridges, ascending to an elevation of nearly 2,000 feet. North of Hollins, Alabama, Mohr found "the foothills and narrow valleys between them, at an elevation of from 1,400 to 1,500 feet covered with truly magnificent forest of Pinus palustris, yielding to the acre as much merchantable timber as the best class of pine lands in the coastal belt from Alabama to Texas."

Andrews (1917) found the southern slopes of Lavendar Mountain near Rome, Georgia covered by the remains of a great longleaf pine forest. He noted the demise of the Georgia montane longleaf pine communities was tied to European settlement in the 1830’s.

These forests of longleaf pine owed their existence to fire. Reed (1905) may have had the first report on the use of fire in his "Working Plan for Forest Lands in Central Alabama." He noted that surface fires were prevalent in the region, indicating some areas may have burned twice within a year.

Implementing prescribed burns in montane longleaf pine

Based on observations and reports from burning in the Appalachian Mountains, prescribed burns should nearly always be set at higher elevations and allowed to back down. Ridges and spur ridges are often set and the fires are allowed to burn down the slopes. Those lighting the fires need to pay strict attention to conditions, escape routes and safety zones in case there is a change in fire movement. When set low on a slope, the fires preheat ever-increasing amounts of fuel as they ascend, and these fires can be devastating.

The most specific report about prescribed fire in the montane region was done by Laros (1961). He reported that burning in the mountains was similar to burning on the Coastal Plain. He discussed burning slopes of 10-15% for brown-spot control and to prepare a seedbed for a longleaf pine cone crop. After control lines had been established, lines of fire were strung along ridge tops and allowed to back down the slopes. The burning was done 3-5 days following a measurable rainfall. He concluded that fires should be less frequent than on the Coastal Plain because of the danger of soil erosion. Hardwood control could be done on lower slopes with repeated fire due to less danger of soil erosion but control on steeper slopes would be some combination of mechanical and chemical treatments.

The other published report came from work examining longleaf pine regeneration in the mountains and piedmont of Alabama (Croker 1968). This report noted a concern for preventing excessive overstory damage during a late summer burn on steep slopes.