Dry prairies are large treeless areas endemic to Florida.  Resembling pine flatwoods without the overstory, the treeless stature of dry prairies has been attributed to a fire regime of frequent fires.  Dry prairies were thought to have developed with a fire regime of every 1-4 years, although lower fire return intervals  (1-2 years) may have been common in regions devoid of major natural fire barriers.  Historically, most of these burns likely occurred during the late spring and early summer months (fire season) and were caused by lightening.  Anthropogenic fires expanded the season of fires in dry prairies to all seasons of the year.  Along with flatwoods, ranchers typically burned dry prairie rangeland annually or biennially during the winter or early spring months to stimulate forage growth.

Fires in dry prairie are typically intense due to the high flammability of saw palmetto and wiregrass, both of which are species that are tolerant of fire.  Dormant season burns in dry prairies favor graminoids over forb species, while burns during the growing season tend to favor forbs. Frequent growing season burns also tend to keep saw palmetto small in stature and sparsely distributed.  Fauna endemic to dry prairies have adapted to growing season burning and many ground nesting species of birds (Florida grasshopper sparrow, Bachman’s sparrow, and eastern meadowlark) will re-nest if their nest if ruined by fire.

Due to landscape fragmentation and fire suppression, the extent and condition of dry prairie has been altered considerably making this community type a priority for restoration.  Long periods (35 years) of fire exclusion can cause a loss of ground cover species, changes in pine density and recruitment, invasion of non-constituent oaks, and excessive growth of shrubs and palmetto in dry prairies. Prescribed fire is an important management tool for the restoration and maintenance of dry prairies.  Land managers can manipulate fire effects by deciding season of burn (growing vs. dormant season burns).  Prescribed burning is also used to maintain habitat for endangered animal species in dry prairies, such as the federally endangered Florida grasshopper sparrow.  In some cases where fire has been suppressed from the ecosystem, mechanical methods such as roller chopping are used in combination with prescribed burning to jump-start the restoration process.    

Dry prairies are large treeless areas located in Central Florida and are considered endemic to the state. Over 240 plant species have been documented in dry prairie (Orzell and Bridges 1999), but they are primarily dominated by low shrubs (e.g. saw palmento (Serenoa repens), runner oak (Querus minima), fetterbush (Lyonia lucida), grasses (e.g. wiregrass (Aristida beyrichiana) and Schizachryium stoloniferum, and other herbaceous species (e.g. Rynchospora plumosa, Syngonanthus flavidulus). Dry prairies closely resemble pine flatwoods without the pine overstory (Abrahamson and Harnett 1990). In fact, dry prairie can be thought of as the endpoint along a forested to treeless continuum of flatwoods/savanna landscapes, in response to variation in the natural fire regime.

Once covering over 830,000 hectares of central Florida, currently there is an estimated 156,000 hectares of dry prairie left in the state (Shriver and Vickery 1999). Florida dry prairie is ranked as a globally imperiled community type (FNAI and FDNR 1990, Grossman et al. 1994, Weakley et al. 1996). Based upon DeSelm and Murdock (1993), Noss et al.(1995) considered ungrazed dry prairie of Florida as an endangered ecosystem (greater than 98 percent habitat loss and continued threat). Presently, dry prairies are restricted to the south central and west-central parts of Florida, primarily along the Kissimmee River and west of Lake Okeechobee. Two large areas of dry prairies can be found at Myakka River State Park (6500 ha) and Kissimmee Prairie State Preserve (12,000 ha) (Fitzgerald and Tanner 1992, Shriver and Vickery 1999).

Burrowing Owl

Dry prairies occur on low, flat topography with poorly drained acidic soil that is subject to flooding for a short duration after heavy summer rains. Dry prairies provide habitat for variety of threatened fauna including the Florida panther (Puma (=Felis) concolor coryi), the Audubon?s crested caracara (Polyborus plancus audubonii), Florida sandhill crane (Grus canadensis pratensis), Florida grasshopper sparrow (Ammodramus savannarum floridanus), and burrowing owl (Speotyto cunicularia floridana). No federally listed plants are currently known for dry prairie, although Calypogon multiflorus, an orchid, is currently listed a state endangered species.

Although the reason for treelessness remains subject to debate, dry prairies are adjacent to pine flatwoods and scrubby hardwood ecotypes. If fire is kept out of dry prairie ecosystems trees from adjacent ecosystems will invade a dry prairie. The persistence of dry prairie has also been closely tied to cattle ranching that has existed in central and South Florida since European settlement.

For a complete description on five different natural community types of dry prairie, see the US Fish and Wildlife Service?s South Florida Multi-Species Recovery Plan: Dry Prairie.

History of Livestock Grazing in Florida

The persistence of dry prairie is closely tied to cattle ranching that has existed in central and South Florida since European settlement, and has continued to the present. The central Florida region was once sparsely populated (in comparison to other regions of Florida) and alternative land use options were scarce, thereby supporting management of native rangeland (DeSelm and Murdock 1993). Livestock have been present in dry prairie, and throughout central and southern peninsular Florida, for several hundred years (Yarlett 1985). Livestock grazing was historically (Harper 1921, Davis 1943) the primary economic benefit derived from the dry prairie/ flatwoods landscapes of central and south-central Florida (Moore 1974, Sievers 1985, Sullivan 1994). Harper (1927) reported that prior to the division of Desoto County into five counties the U. S. Census of 1900 indicated that Desoto County had 82,183 free-ranging cattle on 9,713 km² (3,750 mi²), about half of which was prairie. In 1920 there were 53,192 cattle in the county (over 95 percent beef cattle) and 70,459 cattle in 1925 (Harper 1927). Open, unfenced range existed in Florida until 1949 (Yarlett 1985). Between 1940 and 1960, many pastures of bahia grass (Paspalum notatum), digit grass (Digitaria sp.) and Bermuda grass (Cynodon dactylon) were established (Mozley 1985), as a result of land clearing and drainage alterations of native rangeland. By 1985, Florida had 1,215,000 ha (3 million acres) of perennial tame pasture and 405,000 ha (million acres) of annual tame pasture (Mozley 1985), an extensive area of which had formerly been dry prairie and flatwoods. In the early 1970s, production costs caused a shift towards native forage resources, although many ranchers had already recognized the value of native rangeland over improved pasture in the 1960s (Yarlett 1985). In 1984 there was a total cattle and calf inventory of 2.2 million head statewide (Yarlett 1985), with commercial cow-calf operations representing the major livestock production in Florida (Mozley 1985).

Forage Production

One reason Florida attracted such a large cattle industry is a growing season longer than 270 days and ample rainfall, ideal for growing forage and year-round use of native and improved pastures (Mozley 1985, DeSelm and Murdock 1993). Forage production for native rangeland (i.e. dry prairie and flatwoods) can range from 1,688 kg/ha (1,500 lbs/acre) on in-poor condition ranges to in excess of 8,438 kg/ha (7,500 lbs/acre) on in-excellent condition ranges, with the average production being from 4,500 to 6,750 kg/ha (4,000 to 6,000 lbs/acre) (Fults 1991, Sievers 1985, Mozley 1985). Hilmon and Hughes (1965) found that native rangelands burned in March or May produced a two-to-fourfold increase in forage, compared to those burned in October or November, with May burns causing Aristida beyrichiana to flower. However, native rangeland vegetation is not particularly nutritious for cattle, except immediately following a burn, with nutritional values rapidly declining thereafter (Hilmon and Hughes 1965).

Using Fire in Florida Rangelands

Native central and South Florida rangelands (i.e. dry prairies and flatwoods) are typically burned by ranchers annually or biennially during the winter or early spring months to stimulate forage growth, nutrition and palatability during the lean winter months (Abrahamson and Harnett 1990, Frost et al. 1986, Sullivan 1994). Ranchers also burn native pastures to maintain openness, reduce shrub cover, reduce fuel accumulations, and improve wildlife habitat (Abrahamson and Harnett 1990, Frost et al. 1986). However, frequent winter burning, when coupled with continuous grazing pressure, can lead to a decline in Schizachyrium stoloniferum (creeping bluestem, formerly Andropogon stolonifera) (White and Terry 1979), Sorghastrum secundum (lopsided Indiangrass), Amphicarpum muhlenbergianum (goobergrass), Andropogon virginicus var. glaucus (little chalky bluestem, formerly Andropogon capillipes), all of which are major preferred native forage grasses (Sievers 1985) that occur in dry prairie. Under such a management regime, Aristida beyrichiana (previously Aristida stricta, in part) (Hilmon and Hughes 1965) and saw palmetto are reported to increase (Sievers 1985), as well as other brush cover (Fults 1991). To lessen the impacts of grazing on rangelands with a previous history of frequent burning, some range managers have recommended burning on a 3-year cycle (Sievers 1985, Penfield 1985), which allows more opportunity for recovery in cover of preferentially grazed grasses.

Effects of Grazing on Prairie/Flatwoods Communities

Although the effects of fire and cattle grazing on the central Florida prairie/flatwoods communities have been examined (Hilmon and Hughes 1965, Sievers 1985, Fults 1991, Mozley 1985), there is a lack of knowledge on the effects of livestock grazing on various components of native prairie biodiversity. This is particularly true of most prairies outside of the Tallgrass Prairie region of the central United States. Herbivores preferentially feed on different plant species, consequently affecting dominance and relative abundance of plant species, thereby altering species composition (Howe 1994), with the potential to increase or decrease plant species diversity (Huntly 1991). Prolonged or severe preferential grazing can lead to colonization by ruderal plant taxa (Weaver 1968). In contrast, moderate grazing can result in greater species diversity relative to ungrazed sites, where a few species may dominate a community (Milchunas et al. 1988). Grazing activity may change the structure of the community by disrupting soil properties (Abrahamson and Harnett 1990) or by eliminating some life forms (Howe 1994). Furthermore, grazers may interact with other processes operating at different spatial and temporal scales, such as fire, drought, or other species interactions, resulting in scale-dependent effects (Harnett et al. 1996). Differences in livestock grazing management (i.e. year-round grazing, high intensity-short duration grazing, etc.) contribute to differences in vegetation responses (Harnett et al. 1996). Finally, historical ungulate assemblages were not confined, adding stochasticity to grassland development that is less evident in ungrazed prairies, small remnants, and fenced pastures (Howe 1994). All of these factors suggest that domestic livestock grazing, managed in different ways can produce significantly different effects on prairie community structure, species composition and prairie biodiversity (Hartnett et al. 1996). The effects of domestic livestock grazing on components of dry prairie biodiversity therefore have significant implications for management.

Dry prairies were thought to have developed with a fire regime of every 1-4 years (Abrahamson and Harnett 1990), although some sources suggest an even higher fire return frequency (1-2 years) in regions devoid of major natural fire barriers (Orzell and Bridges 1999).  Frequent burning kept pines and oaks from invading dry prairie, turning them into pine flatwood or scrubby hardwood sites.  These burns would have been very intense due to the high flammability of saw palmetto and wiregrass, both of which are species that are tolerant of fire.  Dry prairie burns more intensely and completely compared to other Florida ecosystems such as scrub because scrub contains lower biomass litter and patches of bare sand (Abrahamson and Harnett 1990).

Historically, most of these burns likely occurred during the late spring and early summer months (fire season) and were caused by lightening.  Florida has a high frequency of lightening strikes, more than any other region in the United States (Abrahamson et al. 1984). When burned in the summer this ecosystem tended to burn completely, covering thousands of acres from a single ignition source.  Before construction of roads, power lines and canals, there were few firebreaks to slow down the speed of a fire in a dry prairie, so given the right conditions a single lightening strike could have burned large areas.

Although most lightning-ignited fires occurred during the late spring and early summer months, it is likely that under the right conditions dry prairies could have burned during any time of the year. Fires that burned at different times of the year would have likely been patchier because of moister conditions in the dry prairie.  Anthropogenic fires expanded the season of fires in dry prairies to all seasons of the year.  Along with flatwoods, ranchers typically burned dry prairies rangeland annually or biennially during the winter or early spring months to stimulate forage growth.

Different burning seasons tend to favor different groups of species in Florida dry prairies.  In dry prairies, dormant season burns favor graminoids over forb species, while burns during the growing season tend to favor forbs.  Calopogon mutiflorus has been shown to flower after dormant season burns, but rarely flowers in years that have not burned, or burned during the summer (Goldman and Orzell 2000).  Cutthroat, beard grasses (Andropogon spp.), and wiregrass are more likely to flower when burned in the early growing season (Abrahamson 1984).  Burning during the growing season also delays the flowering of certain species and synchronizes flowering peaks of groundcover species (Platt et al. 1988).

Fires play an important role in nutrient cycling of dry prairie ecosystems.  Frequent, low-intensity fires increase the rate of nutrient turnover, raise pH and stimulate nitrogen fixation (Abrahamson and Hartnett 1990).  Phosphorus and cations in biomass are mineralized in these conditions as well (Abrahamson and Hartnett 1990).

Effects of Fire Suppression

Deviation in fire intensity, fire return interval, and seasonality from the natural fire regime of frequent growing-season burns have modified vegetation structure and composition of dry prairie communities (Perry 1997, Dye 1997, Bridges 1997). Loss of ground cover species, changes in pine density and recruitment, invasion of non-constituent oaks, and excessive shrub growth has been documented from dry prairies with long periods (ca. 35 years) of fire exclusion (Dye 1997, Perry 1997). When dry prairie is frequently burned, saw palmetto is typically of small stature and sparsely distributed, but it tends to increase in stature and density when fire is absent or infrequent (Cole et al. 1994a).

Parts of Myakka River State Park, located in the southwestern peninsula of Florida in Sarasota and Manatee Counties, have not burned for 35 years.  With fire suppression, more woody shrubs and oaks have invaded dry prairies.  In the Wilderness Preserve Series at Myakka State Park, areas of dry prairie that were adjacent to oaks saw open areas decrease from 66 percent to 26 percent in 37 years (Huffman and Blanchard 1991).  Saw palmettos that are small and sparse in frequently burned dry prairies become larger and tend to overtake wiregrass if the area is fire suppressed (Orzell and Bridges 1999).  If there is an available seed source pines will creep into dry prairies, it is thought that a burning frequency of 1-5 years will keep pines from invading dry prairies and turning them into pine flatwoods (Orzell and Bridges 1999).

After years of fire suppression, recovery to a state that was present pre-suppression is not likely after one burn.  After years of fire suppression, woody shrubs that invade prairie develop extensive root systems that will send up many sprouts if burned.  Herbicides and roller chopping are two methods that are used often in combination with burning in the restoration of dry prairies to a pre fire-suppressed state.

In contrast to the effects of fire on vegetation, the effects of fire on the faunal component of dry prairie is largely unknown. However, most of the faunal taxa should be well adapted to the natural fire regime. In particular, those which are endemic to the dry prairie region of south-central Florida have evolved within a high fire frequency ecosystem. In fact, some if not most prairie faunal taxa that are dependent on dry prairie also appear to be dependent on recently burned sites (i.e., Florida grasshopper sparrow, see Walsh et al. 1995), or prefer recently burned sites over fire-suppressed dry prairies.

Florida Grasshopper Sparrow

It had been thought that ground-nesting birds might be susceptible to mortality or lower reproductive success due to growing season prescribed burning (T. Dean, Department of Forestry and Wildlife, University of Massachusetts, personal communication 1998). The Florida grasshopper sparrow, Bachman?s sparrow and the eastern meadowlark are all ground-nesting birds of dry prairie. All initiate breeding in mid-March through April (Perkins et al. 1998), and have relatively short reproductive cycles (30 to 40 days from egg-laying through fledging). It had been hypothesized that natural frequent late spring and early summer lightning fires, now simulated with prescribed burns during the growing season from late March through early June, would destroy some nests and young birds, and that this would result in significant reduction in reproductive success and population levels. However, all three of these birds regularly re-nest if nests are destroyed (T. Dean, Department of Forestry and Wildlife, University of Massachusetts, personal communication 1998). In addition, multiple successful clutches are possible for all three taxa (T. Dean, Department of Forestry and Wildlife, University of Massachusetts, personal communication 1998). Recent studies have shown that adult Florida grasshopper sparrows and other ground-nesting avian taxa appear to be able to successfully avoid fires, and can occupy recently burned areas within 1 week after fire (Shriver et al. 1994, Shriver 1996). Although some nests and young birds could be lost to fire, growing season burning does not preclude successful reproduction in that season for any of these birds. Recent studies have shown that Florida grasshopper sparrows successfully fledge young after June fires (Vickery and Perkins 1997), and increase the length of the breeding season by as much as 8 weeks, extending as late as September, after these fires (Shriver 1996, Vickery et al. 1997). However it is still unknown whether reproductive success differs between summer-burned and winter-burned areas (FWS 1997).

Several factors suggest that growing season burns may benefit the Florida grasshopper sparrow. These sparrows preferentially nest in areas less than 1.5 years post-burn (Walsh et al. 1995, Shriver 1996). If such habitat patches were available within a mosaic of areas with differing post-burn times, they would most likely be the areas utilized for nesting. Prescribed burning during the nesting season would most likely be conducted on areas of 2 years or more post-burn, therefore the areas least likely to be currently utilized for nesting. Therefore, as long as a sufficient area less than 1.5 years post-burn is present within the population area, nesting season prescribed burning of adjacent areas should destroy relatively few nests. Those disrupted should renest and the extension of the breeding season by prescribed burning during the nesting season may have important consequences for a species that appears to have low reproductive success (Shriver 1996). Lengthening the breeding season by provision of newly burned areas effectively doubles the number of nesting opportunities, may increase the number of pairs that attempt additional clutches, and may provide sub-dominant males the opportunity to establish territories and breed for the first time (Shriver 1996). Further research is needed to test these new hypotheses.

Crested Caracara

The long-term effects of growing season fires on the overall productivity of these ground-nesting sparrows are largely unknown (T. Dean, Department of Forestry and Wildlife, University of Massachusetts, personal communication 1998). However, neither is it known how the traditional winter burning of dry prairie by the cattle ranchers and some current land managers may have affected the long-term population levels of such ground-nesting birds. Larger prairie birds, including sandhill cranes and caracaras, would possibly have greater potential to be affected by growing-season burns, since they require longer periods to produce young, 30 days incubation in cranes (Walkinshaw 1981) and 30 to 33 days for caracaras (Layne 1996, Morrison 1996). However re-nesting is still common in both species (Nesbitt 1988, Morrison 1996), and loss of nests would not preclude nesting in either bird. In addition, the breeding cycles of both species are in the dry season, before the peak time for natural growing season fires. Since caracaras nest in trees and cranes nest in wetlands, nest losses due to fire may be minimal. However, the indirect effects of fire on overall productivity are unknown.

The effects of fire on other prairie occupants are also largely unknown. Armadillo and gopher tortoise burrows and the numerous depressional ponds characteristic of the dry prairie landscape likely provide refuge from fires for herpetofauna and small mammals (T. Dean, Department of Forestry and Wildlife, University of Massachusetts, personal communication 1998).

The uncertainty of the effects of growing season burning on the Florida grasshopper sparrow has caused land managers to shift prescribed burns from optimal late spring and early summer burns (especially April and May). The acceptable burning seasons therefore are pre-avian breeding, dormant season (i.e. winter) burns (January to early March), or postponement of prescribed burning until post-avian breeding (July through September). Winter or dormant season burns generally result in less overall reduction of aboveground woody biomass, and reduction in flowering of grasses and other native forbs. Prescribed burning conducted during the post-breeding season (July through September), after the onset of summer convective thunderstorms, has the potential to cause significant shifts in species composition in dry prairie. Heavy rainfall events, which are typically more frequent from mid-June through the summer months, can produce two dramatic effects on vegetation in areas of post-breeding season burns: (1) more patchy burns (resulting in less overall reduction of woody vegetation) and (2) flooding of postburn regrowth, resulting in submersion of regrowth and stress to perennial groundcover grasses. The result is that annual cyperoids, such as Scleria reticularis (nut-rush), present in the seed bank are favored over perennial grasses and can locally co-dominate summer-burned areas (Orzell and Bridges, personal field observations). Whether this vegetative response is short-term or long-term is unknown, but it needs to be studied.

Prescribed fire in is an important management tool for the restoration and maintenance of dry prairies.  Due to landscape fragmentation and fire suppression, the extent and condition of dry prairie has been altered considerably making this community type a priority for restoration.  Dry prairie requires fire every 1-4 years to maintain coverage of dry prairie flora.  Within this burn frequency, land managers manipulate fire effects by deciding season of burn (growing vs. dormant season burns).  For instance, winter burns are often wetter and will not burn as completely or as hotly as summer burns.  This will lead to a patchier condition in dry prairie, and will also favor many shrub species because of their ability to resprout.  Alternatively, summer burns favor more herbaceous species, and reduce the cover of woody stems in an area. 

Prescribed burning can be used to maintain habitat for endangered animal species in dry prairies.  Most of the faunal research conducted in dry prairie ecosystems has concentrated on the Florida grasshopper sparrow.  This is a federally endangered species that is considered endemic to the dry prairies of Florida.  Higher numbers of the Florida grasshopper sparrows were found in recently burned sites and it is recommended that a burning frequency of <3 years be used to maintain higher populations of this bird (Shriver and Vickery 2001). 

Using Prescribed Fire and Mechanical Treatments for Restoration

In some cases where fire has been suppressed from the ecosystem, additional methods are needed for restoration (Huffman and Blanchard 1991).  Roller chopping in combination with prescribed burning is a method proven to be effective in reducing the woody component of dry prairies (Watts 2002). This method cuts the roots of the woody shrubs so that they do not sprout back as quickly when the prairie is burned.  Roller chopping also has an effect on the herbaceous species in dry prairie.  Pine lily (Lilium catesbaei Walt.), a widely distributed but uncommon species in dry prairie decreases in number with fire suppression.  In a study conducted in Myakka State Park, greater numbers of pine lily were found in plots with roller chopping treatment and fewest found in burn only treatments (Huffman and Werner 2000).

Mechanical treatments (bedding, root-raking, use of a disc, roller chopping, web plowing etc.), by whatever means, typically produce a wide variety of changes in plant community structure and composition in dry prairie. The use of bedding to reduce or remove competing shrubs and other vegetation, to prepare dry prairie for planting with pine or Eucalyptus monocultures (Moore and Swindel 1981), produces distinct microsites in the form of beds, flats, and furrows (Abrahamson and Harnett 1990). The native intact ground cover characteristic of dry prairie is significantly affected by these microsite alterations. Practices such as root-raking and use of a disc displace surface organic and litter layers and mineral soil, and may potentially eliminate native species, allow for the establishment of numerous weedy species, and significantly alter the native species composition and vegetation structure of dry prairie.

The most frequently employed mechanical treatment on dry prairie is the use of roller chopping to control woody shrub height, in particular excessive growth in saw palmetto (Yarlett 1965). The use of roller chopping on areas of intact, diverse native groundcover vegetation is not recommended and is strongly discouraged.  Double chopping has been reported to almost eliminate wiregrass and saw palmetto (Hilmon et al. 1963). Frequencies of some herbaceous plants and species richness can increase following some types of roller chopping. However, the resulting increase in species richness (= species diversity), primarly from changes in overall species composition and relative abundance, is from an increase in undesirable ruderal weedy species.

Impacts to vegetation from chopping vary widely and are largely dependent upon the method and type of chopping, and the soil and moisture conditions when the chopping is conducted. Studies conducted by Fitzgerald et al. (1995) on formerly fire suppressed dry prairie (12 to 15 years of fire suppression) at Myakka River State Park found that application of fire regardless of season was not significant in reducing woody cover. However, repeated chopping and follow-up burning aided attempts to restore dry prairie in shrub-dominated dry prairie (Fitzgerald et al. 1995, Tanner 1997). Therefore, roller chopping on former fire-suppressed or previously disturbed dry prairie may be a useful restoration tool when used in conjunction with frequent burning (Tanner 1997, Fitzgerald and Tanner 1992). Fitzgerald and Tanner’s (1992) study on formerly fire-suppressed (35 years of fire suppression), overgrown dry prairie found shrub control treatments (roller chopping) had an acute effect on bird abundance and species composition, as they had postulated. They noted that the first birds observed in the chopped plots were prairie bird species (eastern meadowlark, Melospiza georgiana; loggerhead shrikes, Lanius ludovicianus; and grasshopper sparrow) perhaps an early sign of prairie restoration (Fitzgerald and Tanner 1992). Roller chopping is now accepted by many land managers as an effective means to begin restoration of overgrown former prairies and flatwoods (Tanner et al. 1988, Moore 1974). This is particularly true where fire has not been sufficient to control woody shrubs, where chopping can be effective in reducing the height and dominance of woody shrubs (Tanner 1997, Fitzgerald et al. 1995). However, the destructive nature of this technique, even in disturbed prairie, has the potential to cause unforeseen changes in natural processes. If applied during the spring, summer, or early fall, roller chopping may interfere with nesting activities of many prairie bird species. In addition, Bridges (1997) has postulated that roller chopping may have been the cause of significant long-term vegetation composition changes at some dry prairie sites.