SEARCH

SEARCH BY CITATION

Keywords:

  • conservation;
  • heterogeneity;
  • hierarchy;
  • patch burn;
  • pyric herbivory;
  • rangeland;
  • scale

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

1. Patterns of landscape heterogeneity are crucial to the maintenance of biodiversity in shrublands and grasslands, yet management practices in these ecosystems typically seek to homogenize landscapes. Furthermore, there is limited understanding of how the interaction of ecological processes, such as fire and grazing, affects patterns of heterogeneity at different spatial scales.

2. We conducted research in Artemisia filifolia (Asteraceae) shrublands located in the southern Great Plains of North America to determine the effect of restoring the fire–grazing interaction on vegetation structure. Data were collected for 3 years in replicated pastures grazed by cattle Bos taurus where the fire–grazing interaction had been restored (fire and grazing = treatment pastures) and in pastures that were grazed but remained unburned (grazing only, no fire = control pastures). The effect of the fire–grazing interaction on heterogeneity (variance) of vegetation structure was assessed at scales from 12·5 m2 to 609 ha.

3. Most measurements of vegetation structure within treatment pastures differed from control pastures for 1–3 years after being burned but were thereafter similar to the values found in unburned control pastures.

4. Treatment pastures were characterized by a lower amount of total heterogeneity and a lower amount of heterogeneity through time.

5. Heterogeneity of vegetation structure tended to decrease as the scale of measurement increased in both treatment and control pastures. There was deviation from this trend, however, in the treatment pastures that exhibited much higher heterogeneity at the patch scale (mean patch size = 202 ha) of measurement, the scale at which patch fires were conducted.

6.Synthesis and applications. Vegetation structure in A. filifolia shrublands of our study was readily altered by the fire–grazing interaction but also demonstrated substantial resilience to these effects. The fire–grazing interaction also changed the total amount of heterogeneity characterizing this system, the scale at which heterogeneity in this system was expressed and the amount of heterogeneity expressed through time. Land managers seeking to impose a shifting mosaic of heterogeneity on this vegetation type can do so by restoring the fire–grazing interaction with potential conservation benefits similar to what has been achieved in other ecosystems where historic cycles of disturbance and rest have been restored.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Understanding the effects of heterogeneity on the structure and function of ecological communities and landscapes is a critical component of biodiversity conservation (Christensen 1997; Wiens 1997). The alteration of landscape and community heterogeneity by large herbivores is a global phenomenon (Augustine & Frank 2001; Vandvik et al. 2005; Collins & Smith 2006; Mouissie et al. 2008), whereby the spatial pattern of grazing can interact with pre-existing vegetation heterogeneity resulting in either an increase or a decrease in heterogeneity of vegetation communities (Adler, Raff & Lauenroth 2001). Additionally, fires have the potential to alter landscape heterogeneity with concomitant effects on biodiversity (Parr & Andersen 2006). In contrast to the recognized importance of heterogeneity, a primary objective of range management has been the uniform distribution of grazing animals in space and time (Williams 1954; Bailey 2004; Holecheck, Pieper & Herbel 2004), which may in fact homogenize rangeland landscapes (Knopf & Samson 1997; Fuhlendorf & Engle 2001). An alternative rangeland management practice known as patch burning is based on the evolutionary interaction of fire and grazing known as pyric herbivory with the goal of manipulating animal distribution through the application of discrete fires that attract animals to different locations. This approach is intended to approximate historic cycles of vegetation disturbance and rest across multiple scales (Fuhlendorf & Engle 2001; Fuhlendorf et al. 2009). The interaction between fire and large grazers is described by a model in which both positive and negative feedbacks create a shifting mosaic of out-of-phase landscape patches that differ in vegetation structure and composition, the amount of herbaceous biomass and levels of forage quality (Fuhlendorf & Engle 2004; Fuhlendorf et al. 2009). High levels of forage utilization in recently burned patches, and concomitantly low levels of forage utilization in adjacent patches that have not burned recently, result in a landscape mosaic of herbaceous biomass (fuel) that determines the location and behaviour of subsequent fires within a landscape (Kerby, Fuhlendorf & Engle 2007; Savadogo et al. 2007; Fuhlendorf et al. 2009; Kirkpatrick, Marsden-Smedley & Leonard 2011). The attraction of grazing and browsing herbivores to recently burned areas has been documented with multiple taxa in ecosystems around the globe (Moe & Wegge 1997; Murphy & Bowman 2007; Sensenig, Demnent & Laca 2010). The shifting mosaic of heterogeneity in vegetation structure arising from the fire–grazing interaction is likely to be crucial for the maintenance of biodiversity in rangeland ecosystems (Fuhlendorf et al. 2009).

Furthermore, the expression and understanding of ecological phenomena, such as the interaction of disturbances, depend on the scale of observation (Wiens 1989). For example, mechanisms that influence habitat selection by wildlife can differ depending on the scale of assessment (Cornell & Donovan 2010; Crook & Chamberlain 2010), and food web dynamics may be better understood by considering the scales at which consumers function within their environment relative to the different scales at which their resources may be found (Van de Koppel et al. 2005). In rangelands, the effects of grazing on vegetation and the expression of temporal patterns of vegetation dynamics vary with the spatial scale of observation (Fuhlendorf & Smeins 1996, 1999). Thus, a comprehensive understanding of the fire–grazing interaction necessitates an examination of response variables at multiple scales.

Much research on restoration of the fire–grazing interaction has been conducted in mesic grasslands (Fuhlendorf & Engle 2004; Fuhlendorf et al. 2006; Coppedge et al. 2008; Pillsbury et al. 2011). Our research, however, was conducted in a shrub-dominated region characterized by a drier climate where information on the fire–grazing interaction is limited (but see Vermeire et al. 2004). The objectives of our research were to: (i) determine the response of vegetation structural characteristics (bare ground, litter, live and dead vegetation, live and dead graminoids, live and dead forbs, live and dead shrubs, vegetation height and vegetation visual obstruction) to increasing time since being burned; (ii) determine the relationship between heterogeneity in vegetation structural characteristics and scale-of-observation within pastures managed in a traditional manner (control pastures: moderate grazing without patch burning) and pastures where the fire–grazing interaction had been restored (treatment pastures: moderate grazing with patch burning); and (iii) determine the amount of patch-scale heterogeneity in vegetation structural characteristics in control pastures and treatment pastures.

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study Site

The study site was the Hal and Fern Cooper Wildlife Management Area (Cooper WMA) in Woodward County, Oklahoma, USA (36°32′10′′N, 99°30′05′′W). The long-term (1940–2008) average annual precipitation at the National Oceanic and Atmospheric Administration Fort Supply cooperative weather station (http://www.ncdc.noaa.gov) was 59·9 cm. The annual total precipitation and per cent of the long-term average for 2005, 2006, 2007 and 2008 was 72·5 cm (121%), 40·5 cm (68%), 77·0 cm (129%) and 55·3 cm (92%), respectively. The majority of the study site, 63%, was characterized by soils in the Eda-Tivoli soil complex (loamy fine sands and fine sands; USDA-NRCS, 2009), and all sampling occurred in areas occupied by this soil complex. Vegetation of the study region is considered an Artemisia shrubland with the dominant species being the shrub A. filifolia Torr. (Asteraceae; Collins, Bradford & Sims 1987; Gillen & Sims 2004). Herbaceous vegetation was a diverse mixture of grasses such as Andropogon hallii Hack., Schizachyrium scoparium (Michx.) Nash, Eragrostis trichodes (Nutt.) Wood, Paspalum setaceum Michx. and Bouteloua gracilis (Willd. ex Kunth) Lag. ex Griffiths. Common forbs included Ambrosia psilostachya DC., Commelina erecta L., Croton texensis (Klotzsch) Muell.-Arg. and Eriogonum annuum Nutt.

Prior to and during this study, all study pastures were grazed by yearling steers (Bos taurus L.) from 1 April to 15 September. Stocking level in all pastures was c. 6·85 ha per animal unit (1 steer = 0·6 animal unit), and cattle had free access to all areas of each pasture. During 1999–2001, prescribed fires were used to create 14 separate 4-ha patches within the study pastures during research on the effects of patch burning on the distribution of grazing cattle (Vermeire et al. 2004). Prior to the prescribed fires conducted during 1999–2001 and in this study, no fires had occurred in the study pastures at least since the property was purchased by the state of Oklahoma in 1992.

Study design

The study was conducted in five pastures of 406–842 ha (mean = 608 ha). During 2003–2008, three of the pastures (hereafter treatment pastures) were treated with spring (March–May) prescribed fires such that approximately one-third of each pasture was burned. Mean size of the patches burned during 2003–2008 was 195 ha and ranged from 83 to 415 ha. The remaining two pastures had no fires from 2003 to 2008 and were considered control pastures.

Sampling

For sampling purposes, each pasture was divided into three approximately equal-sized patches; patch size was thus proportional to pasture size. Patch boundaries in treatment pastures corresponded with firebreaks delineating individual burn units. Four 100-m transects were randomly located in Eda-Tivoli soils within each patch (= 12 transects/pasture); the 4-ha patches that had been burned during 1999–2001 (Vermeire et al. 2004) were visible on aerial photographs, and all transects were located so that they did not occur in the 4-ha burns. From 21 May to 16 June in 2006–2008, we quantified vegetation structure variables to the nearest 5% in a 0·10-m2 rectangular plot (0·20 × 0·50 m) placed on the ground at each 10-m interval along each transect (= 10 plots per transect). Vegetation structure variables were as follows: per cent bare ground; per cent cover of litter; live and dead vegetation; live and dead graminoids; live and dead forbs; and live and dead shrubs. Litter was considered to be any dead or senesced plant material that was horizontally arranged and in contact with the ground or in contact with other litter that was itself in contact with the ground. Dead vegetation, dead graminoids, dead forbs and dead shrubs were considered to be any dead or senesced plant material in each respective category that was not horizontally arranged and in contact with the ground, i.e. standing dead plant biomass not in the litter category. Dead vegetation variables were measured because they can be an important wildlife habitat variable (Fisher & Davis 2010) and they influence the behaviour of fires (Savadogo et al. 2007).

We also measured vegetation height and visual obstruction at 10-m intervals along each transect using a visual obstruction pole modified from Robel et al. (1970; n = 10 placements of the pole per transect). The visual obstruction pole was marked in 1-cm increments, and observations were made two metres from the pole, one metre above the ground surface. One observation was made from each of the four cardinal directions at each placement of the pole (= 4 observations per placement of the pole; n = 40 observations per transect). Vegetation height was determined by recording the highest point at which vegetation crossed between the observer and the pole. Visual obstruction was determined by recording the lowest point at which the pole was visible.

Analysis

Generalized linear mixed-model analyses were conducted using the SAS/GLIMMIX procedure in sas (sas version 9.2; SAS Institute 2007). We treated all vegetation structure measurements (per cent bare ground, per cent cover of litter, live and dead vegetation, live and dead graminoids, live and dead forbs, live and dead shrubs, vegetation height and visual obstruction) as response variables. A hierarchical model identifying the random effects of year, pastures, patches within pastures and transects within patches was used in the analysis. Burn history (unburned control pastures, unburned patches within the treatment pastures, and burned patches within treatment pastures with various time-since-fire classifications) was the only fixed effect included in the model. Throughout the study, burn history (i.e. the fixed effect) of patches within the treatment pastures changed every year, and thus, a traditional repeated measures analysis was not appropriate. Additionally, most response variables did not assume a normal distribution, further invalidating the use of anova methods. Following a significant test of fixed effects (i.e. burn history), pairwise comparisons of means in each time-since-fire category (0·5, 1, 2, 3, 4 and 5 years) and means from unburned patches within treatment pastures were compared with means from the unburned control pastures using Dunnett’s method for multiple comparisons (Hsu 1996).

Response variables were also modelled as either linear or quadratic functions of time-since-fire. All per cent cover response variables were modelled using a beta distribution with a logit link function, vegetation height was modelled with a normal distribution, and vegetation visual obstruction was modelled with a gamma distribution and a log link function.

A second hierarchical model was used to compute variance component estimates across all years (2006–2008) for all variance components associated with spatial scale (quarter-point, point, transect, patch and pasture) and temporal (2006–2008) variables for vegetation height and vegetation visual obstruction data. Restricted maximum likelihood (REML) variance components were estimated for vegetation height data that assumed a normal distribution, and residual pseudo likelihood (REPL) variance components were estimated for vegetation visual obstruction data that assumed a gamma distribution (Littell et al. 2006). The quarter-point, the smallest scale of measurement in our study, represented a scale of c. 12·5 m2 (the area circumscribed by the four readings around each placement of the visual obstruction pole). The point scale accounted for the data from the ten pole placements along each 100-m transect, while the transect scale accounted for the data from the four transects in each patch. The patch scale (mean patch size = 202 ha) accounted for the data from the three patches in each pasture, and the pasture scale (mean pasture size = 608 ha) accounted for the data from the pastures in the treatment (= 3) and control (= 2) categories. The sum of all scale and temporal variance estimates for vegetation height and vegetation visual obstruction provided the total amount of variance for each variable within treatment and control categories. Finally, to assess heterogeneity at the patch scale during each year of the study, we calculated REML and REPL variance component estimates at the patch level for each of the 3 years (2006, 2007 and 2008) of the study in the control and treatment pastures. Our variance component estimates for vegetation height and vegetation visual obstruction are sound, but we lacked sufficient observations (i.e. replicates) for reliable testing of differences between treatment and control categories at the larger spatial scales.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Multiple comparisons of response variables at each time-since-fire category (0·5, 1, 2, 3, 4 and 5 years) from the treatment pastures with the same variables from unburned control pastures indicated that all measurements of vegetation structure returned to levels that were not significantly different from those of unburned control pastures 4 years after being burned (Table 1). Per cent cover of dead shrubs and vegetation visual obstruction in treatment pastures differed from control pastures for 3 years post-fire. Per cent bare ground, per cent cover of litter and dead vegetation, and vegetation height in treatment pastures differed from control pastures for 2 years post-fire. Per cent cover dead graminoids, dead forbs and live shrubs in treatment pastures differed from control pastures for one-half of a year post-fire.

Table 1.   Mean ± SE of vegetation structure response variables for time-since-fire categories of 0·5, 1, 2, 3, 4 and 5 years post-burn in patches that had been burned within treatment pastures, as well as unburned patches in treatment pastures, and in control pastures at Cooper Wildlife Management Area in Oklahoma, USA
Response variableTime-since-fire (years)Unburned treatmentUnburned control
0·512345
  1. P-values are from Dunnett’s multiple comparisons of each time-since-fire category with control pastures. Bold-face font indicates significance at the α = 0·05 level.

Bare ground47 ± 4 < 0·0127 ± 4 = 0·0331 ± 4 < 0·0121 ± 3 = 0·5018 ± 3 = 0·9818 ± 5 = 1·0015 ± 2 = 1·0016 ± 2
Litter25 ± 4 < 0·0130 ± 5 = 0·1625 ± 5 = 0·0235 ± 5 = 0·5838 ± 5 = 0·9645 ± 10 = 1·0040 ± 3 = 0·9743 ± 2
Live vegetation43 ± 3 = 0·3349 ± 4 = 1·0045 ± 4 = 0·8656 ± 3 = 0·6361 ± 4 = 0·1053 ± 7 = 1·0048 ± 3 = 0·9550 ± 2
Dead vegetation16 ± 2 < 0·0132 ± 3 < 0·0137 ± 4 = 0·0342 ± 3 = 0·3149 ± 4 = 1·0049 ± 6 = 1·0050 ± 2 = 1·0048 ± 2
Live graminoids24 ± 3 = 0·1832 ± 4 = 1·0030 ± 4 = 1·0039 ± 4 = 0·5243 ± 4 = 0·1341 ± 8 = 0·8330 ± 3 = 0·9732 ± 2
Dead graminoids11 ± 2 < 0·0126 ± 3 = 0·0727 ± 3 = 0·1235 ± 3 = 1·0040 ± 4 = 0·7340 ± 6 = 0·9836 ± 2 = 1·0035 ± 1
Live forbs18 ± 3 = 0·8716 ± 3 = 1·0011 ± 3 = 0·8811 ± 2 = 0·7811 ± 3 = 0·7911 ± 4 = 0·9815 ± 2 = 1·0015 ± 1
Dead forbs3 ± 1 < 0·018 ± 1 = 1·007 ± 1 = 0·996 ± 1 = 0·585 ± 1 = 0·308 ± 2 = 1·008 ± 1 = 0·997 ± 1
Live shrubs5 ± 1 < 0·019 ± 2 = 0·3413 ± 3 = 1·0018 ± 3 = 0·5219 ± 3 = 0·4519 ± 5 = 0·9114 ± 2 = 1·0014 ± 1
Dead shrubs3 ± 1 < 0·013 ± 1 < 0·015 ± 1 < 0·017 ± 1 < 0·0110 ± 2 = 0·2110 ± 3 = 0·7113 ± 1 = 0·8415 ± 1
Vegetation height14 ± 7 < 0·0121 ± 8 < 0·0137 ± 8 < 0·0152 ± 8 = 0·7755 ± 8 = 0·9958 ± 10 = 1·0060 ± 7 = 0·9958 ± 7
Vegetation visual obstruction2 ± 1 < 0·014 ± 1 < 0·016 ± 1 < 0·019 ± 3 < 0·0411 ± 3 = 0·3913 ± 5 = 1·0016 ± 5 = 1·0014 ± 4

Modelling of response variables as either linear or quadratic functions of time-since-fire within the treatment pastures indicated relationships ( 0·04) between all structural characteristics and time-since-fire except dead forbs (= 0·28). For per cent bare ground and per cent cover of live forbs, the relationship with time-since-fire was negative ( 0·04).

Using the data for vegetation height and vegetation visual obstruction, estimates of the REML (for normal data) and REPL (non-normal data) variance components for all spatial scale (quarter-point, point, transect, patch and pasture) and temporal (2006–2008) variables, as well as the total amount of variance, distinctions between the treatment and control pastures became evident (Fig. 1). The total variance in vegetation height was 1151 in the treatment pastures and 1289 in the control pastures. The total variance in vegetation visual obstruction was 260 in the treatment pastures and 572 in the control pastures. In both treatment and control pastures, variation in vegetation height (Fig. 1a) and vegetation visual obstruction (Fig. 1b) tended to decrease as the scale of measurement increased. In the treatment pastures, however, that trend was interrupted by a substantial amount of variation at the patch scale. Treatment pastures also were characterized by less variance through time than the control pastures. Within the treatment pastures, time contributed 13% and 22% of the total variance in vegetation height and vegetation visual obstruction, respectively. This contrasts with the control pastures, where time contributed 1% of the total variance in vegetation height and 8% of the total variance in vegetation visual obstruction.

image

Figure 1.  Proportion of total variation contributed by all spatial scale (quarter-point, point, transect, patch and pasture) and temporal (2006–2008) variance components for (a) vegetation height and (b) vegetation visual obstruction in treatment pastures (closed circles) and control pastures (open circles) at Cooper Wildlife Management Area in Oklahoma, USA.

Download figure to PowerPoint

For most measurements of vegetation structure (litter, live and dead vegetation, live and dead graminoids, dead forbs, live and dead shrubs, and vegetation height), variance at the patch level was relatively constant throughout the 3 years of the study in the control pastures (Figs 2–4; note that values on the vertical axes are variances, not means). Bare ground, live forbs and vegetation visual obstruction displayed slight changes in variance through the 3 years within the control pastures. Conversely, variance at the patch level in the treatment pastures tended to be greater than in the control pastures and to increase through the 3 years of the study.

image

Figure 2.  Patch-scale variation in (a) per cent bare ground, and per cent cover of (b) litter, (c) live vegetation and (d) dead vegetation for treatment pastures (closed circles) and control pastures (open circles) during each year of the study at Cooper Wildlife Management Area in Oklahoma, USA.

Download figure to PowerPoint

image

Figure 3.  Patch-scale variation in per cent cover of (a) live graminoids, (b) dead graminoids, (c) live forbs and (d) dead forbs for treatment pastures (closed circles) and control pastures (open circles) during each year of the study at Cooper Wildlife Management Area in Oklahoma, USA.

Download figure to PowerPoint

image

Figure 4.  Patch-scale variation in per cent cover of (a) live shrubs and (b) dead shrubs, (c) vegetation height and (d) vegetation visual obstruction for treatment pastures (closed circles) and control pastures (open circles) during each year of the study at Cooper Wildlife Management Area in Oklahoma, USA.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Resiliency has been defined as the amount of time required to return to a state, following disturbance, which approximates the pre-disturbance state (Pimm 1984). Our results demonstrate that vegetation structure of A. filifolia shrublands is resilient to the interacting disturbances of fire and grazing. Nearly all vegetation structural measurements in the treatment pastures were readily altered by the fire–grazing interaction. Within treatment pastures, the first 3 years after a fire were characterized by increased amounts of bare ground and reduced amounts of litter, dead vegetation, vegetation height and vegetation visual obstruction. Some measures of vegetation structure were only affected by fire in the treatment for a very brief time, through 6 months to 1 year, and all measures that were affected recovered to levels similar to those characteristics of the unburned control pastures within 1–4 years. These results are fitting when these shrublands are considered within their environmental context. Temperate grasslands and shrublands are part of Earth’s most extensive fire-dependent ecosystems (Bond & Keeley 2005; Bond, Woodward & Midgley 2005). Use of anthropogenic fire as a means of creating or enhancing landscape heterogeneity for conservation purposes has been described and critiqued (Brockett, Biggs & van Wilgen 2001; Parr & Andersen 2006), but in areas where fire and large grazers coevolved, the heterogeneity that results from the interaction of fire with large herbivore grazing is likely to be of critical importance to biodiversity conservation (Fuhlendorf & Engle 2001; Hamilton 2007; Fuhlendorf et al. 2009).

In our study, components of vegetation structure represented by non-living biomass (litter, dead vegetation, dead graminoids, dead forbs and dead shrubs) tended to be affected by the fire–grazing interaction to a greater extent than components represented by living biomass. Litter and dead vegetation can be an important wildlife habitat variable (Fisher & Davis 2010). Levels of non-living biomass can also be an important influence on fire behaviour (Savadogo et al. 2007; Leonard, Kirkpatrick & Marsden-Smedley 2010). Our results suggest that restoration of the fire–grazing interaction in Artemisia shrubland may readily alter both wildlife habitat and the probability and behaviour of fire, but these effects are relatively transient.

Previous research on the fire–grazing interaction has documented an increase in the per cent cover or standing crop of forbs in patches that have been recently burned and heavily grazed (Coppedge et al. 1998; Fuhlendorf & Engle 2004; Vermeire et al. 2004). In mesic grasslands, grazing of perennial grasses results in higher soil temperatures and higher levels of light available to associated forbs (Fahnestock & Knapp 1993, 1994). The greater availability of resources available to forbs when their neighbouring grasses are grazed is thought to explain the increased growth, reproduction and abundance of forbs in grazed patches (Fahnestock & Knapp 1993, 1994; Hartnett, Hickman & Walter 1996; Damhoureyeh & Hartnett 1997; Vermeire & Gillen 2000).

In more arid regions, however, competition for below-ground resources such as soil moisture may drive plant community dynamics to a greater extent than competition for the above-ground resource of light which can be of critical importance in tallgrass prairie (Scheintaub, Derner & Knapp 2009). Meek et al. (2008) did not record a change in per cent cover of forbs following summer patch burns in a semi-arid rangeland. Their research period was characterized by drought conditions, and they hypothesized that climatic variability may play a role in determining vegetation responses to patch burning in arid and semi-arid regions (Meek et al. 2008). Our results from a region where water is more limited, relative to mesic grassland, demonstrated a negative relationship between per cent cover of live forbs and time-since-fire in the treatment pasture, with the highest cover of forbs occurring during the growing season immediately after a spring burn (Table 1). The negative relationship between per cent cover of live forbs and time-since-fire in our study contrasted the positive relationship we found between per cent cover of live graminoids and time-since-fire. This suggests that forbs in the A. filifolia shrubland of our study site may indeed be competing with grasses for resources and that the fire–grazing interaction allows forbs a period of release from such competition.

Patterns of landscape heterogeneity are important because they influence ecosystem, community and population processes (Turner 1989; Augustine & Frank 2001; Fryxell et al. 2005; Vandvik et al. 2005; Godfree et al. 2011). For instance, variable patterns of herbaceous biomass have been shown, both theoretically and empirically, to influence the processes of fire (Kerby, Fuhlendorf & Engle 2007; Savadogo et al. 2007) and herbivory (Archibald et al. 2005; Mouissie et al. 2008) across landscapes. A relationship between heterogeneity (i.e. variance) and scale, whereby heterogeneity decreases as the scale of measurement, or grain size, increases, has been described by Wiens (1989) and subsequently demonstrated by Fuhlendorf & Smeins (1999). Large herbivore distribution and foraging activities occur within a hierarchy of spatial scales and consumption of plant matter typically occurs at the smallest scale in the hierarchy, the micropatch (Senft 1989). Selective and repeated grazing of micropatches, which may be driven by the positive feedback of enhanced forage quality within the micropatch, can create persistent patterns of heterogeneity in grazed ecosystems (Bakker, de Leeuw & van Wieren 1983; Ring, Nicholson & Launchbaugh 1985; Hobbs et al. 1991). Ungrazed ecosystems may have an inherent level of abiotic heterogeneity that contributes to small-scale dynamics associated with plant populations and communities, and the imposition of grazing-induced heterogeneity on top of this may alter the amount or scale of heterogeneity (Fuhlendorf & Smeins 1999; Mouissie et al. 2008).

Our results demonstrate that the fire–grazing interaction altered the scale at which heterogeneity occurred within A. filifolia shrublands. In the absence of a fire–grazing interaction (i.e. in the control pastures), most of the heterogeneity characteristics of this ecosystem (74–79%; see Fig. 1) were found at the smallest scales we measured, the quarter-point and point, while a minimal amount of heterogeneity was found at the patch scale (≤2%). Conversely, restoration of the fire–grazing interaction in the treatment pastures of our study site resulted in the amount of heterogeneity at the quarter-point and point scales decreasing to 65–66% of the total, while patch-scale heterogeneity increased to 18–26% of the total. Our results also suggested that, while the treatment pastures were characterized by an altered scale of spatial heterogeneity, they also were characterized by less total heterogeneity and less heterogeneity through time than the control pastures. This reduced level of heterogeneity through time could indicate greater temporal stability associated with increased spatial heterogeneity at some scales. A relationship between spatial and temporal heterogeneity has been described for aquatic systems where greater community stability in stream insects through time was associated with greater variability in stream-bottom substrate (Brown 2003).

The patchwork of contrasting vegetation structure at the patch scale resulting from the restoration of the fire–grazing interaction at our study site has been shown to have a substantial influence on the composition of passerine communities at this site (Doxon 2009), similar to what has been found in North American tallgrass prairie (Fuhlendorf et al. 2006; Coppedge et al. 2008) and Serengeti grasslands in East Africa (Nkwabi et al. 2010). The influence of the fire–grazing interaction on the heterogeneity of primary production across landscapes also has been shown to influence other secondary consumer trophic guilds (Yarnell et al. 2007; Engle et al. 2008; Fuhlendorf et al. 2010).

Our study demonstrated that land managers can readily alter vegetation structure in A. filifolia shrublands by restoring the fire–grazing interaction. Our results also demonstrated that this vegetation type is resilient to the interactive effects of fire and grazing. There have been numerous calls for the implementation of heterogeneity-based management as a means of conserving biodiversity in the Great Plains (Knopf & Samson 1997; Fuhlendorf et al. 2006, 2009; Toombs & Roberts 2009). The imposition of a shifting mosaic of heterogeneity in A. filifolia shrublands would likely have conservation benefits similar to those that have been elucidated in other ecosystems (Fuhlendorf et al. 2006; Cummings, Fuhlendorf & Engle 2007). Additionally, we demonstrated that restoration of the fire–grazing interaction changed the scale of heterogeneity within this system, which has important implications for population and community dynamics of higher trophic levels. Restoring the fire–grazing interaction and understanding the patterns of heterogeneity in space and time may be critical in understanding these processes across large and complex landscapes. Yet, our limited understanding of critical scales for conservation and management may be a hindrance to broad-scale application of these practices because there may be no single appropriate scale or fire–return interval.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This research was funded from State Wildlife Grant Project T-30-P of the Oklahoma Department of Wildlife Conservation and Oklahoma State University, administered through the Oklahoma Cooperative Fish and Wildlife Research Unit (Oklahoma Department of Wildlife Conservation, Oklahoma State University, United States Geological Survey, United States Fish and Wildlife Service and Wildlife Management Institute cooperating), and the National Research Initiative of the USDA Cooperative State Research, Education and Extension Service, grant number 2003-35101-12928. Assistance with field work was provided by S. Robertson, A. Ainsworth, M. Zendah, E. Doxon, J. Bryant, J. Burton, M. Cunningham, J. Richards, K. Spears and C. Waldin, and the USDA Southern Plains Research Range provided housing and logistical support during all field seasons. The manuscript benefitted from helpful comments by M. Fishbein, D. Shoup, the editor and two anonymous reviewers.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Adler, P.B., Raff, D.A. & Lauenroth, W.K. (2001) The effect of grazing on the spatial heterogeneity of vegetation. Oecologia, 128, 465479.
  • Archibald, S., Bond, W.J., Stock, W.D. & Fairbanks, D.H.K. (2005) Shaping the landscape: fire-grazer interactions in an African savanna. Ecological Applications, 15, 96109.
  • Augustine, D.J. & Frank, D.A. (2001) Effects of migratory grazers on spatial heterogeneity of soil nitrogen properties in a grassland ecosystem. Ecology, 82, 31493162.
  • Bailey, D.W. (2004) Management strategies for optimal grazing distribution and use of arid rangelands. Journal of Animal Science, 82, E147E153.
  • Bakker, J.P., de Leeuw, J. & van Wieren, S.E. (1983) Micro-patterns in grassland vegetation created and sustained by sheep-grazing. Vegetation, 55, 153161.
  • Bond, W.J. & Keeley, J.E. (2005) Fire as a global ‘herbivore’: the ecology and evolution of flammable ecosystems. TRENDS in Ecology & Evolution, 20, 387394.
  • Bond, W.J., Woodward, F.I. & Midgley, G.F. (2005) The global distribution of ecosystems in a world without fire. New Phytologist, 165, 525538.
  • Brockett, B.H., Biggs, H.C. & van Wilgen, B.W. (2001) A patch mosaic burning system for conservation areas in southern African savannas. International Journal of Wildland Fire, 10, 169183.
  • Brown, B.L. (2003) Spatial heterogeneity reduces temporal variability in stream insect communities. Ecology Letters, 6, 316325.
  • Christensen, N.L. (1997) Managing for heterogeneity and complexity on dynamic landscapes. The Ecological Basis of Conservation: Heterogeneity, Ecosystems, and Biodiversity (eds S.T.A. Pickett, R.S. Ostfeld, M. Shachak & G.E. Likens), pp. 167186. Chapman and Hall, New York.
  • Collins, S.L., Bradford, J.A. & Sims, P.L. (1987) Succession and fluctuation in Artemisia dominated grassland. Vegetation, 73, 8999.
  • Collins, S.L. & Smith, M.D. (2006) Scale-dependent interaction of fire and grazing on community heterogeneity in tallgrass prairie. Ecology, 87, 20582067.
  • Coppedge, B.R., Engle, D.M., Toepfer, C.S. & Shaw, J.H. (1998) Effects of seasonal fire, bison grazing and climatic variation on tallgrass prairie vegetation. Plant Ecology, 139, 235246.
  • Coppedge, B.R., Fuhlendorf, S.D., Harrell, W.C. & Engle, D.M. (2008) Avian community response to vegetation and structural features in grasslands managed with fire and grazing. Biological Conservation, 141, 11961203.
  • Cornell, K.L. & Donovan, T.M. (2010) Scale-dependent mechanisms of habitat selection for a migratory passerine: an experimental approach. Auk, 127, 899908.
  • Crook, A.C. & Chamberlain, M.J. (2010) A multiscale assessment of den selection by black bears in Louisiana. Journal of Wildlife Management, 74, 16391647.
  • Cummings, D.C., Fuhlendorf, S.D. & Engle, D.M. (2007) Is altered grazing selectivity of invasive forage species with patch burning more effective than herbicide treatments? Rangeland Ecology & Management, 60, 253260.
  • Damhoureyeh, S.A. & Hartnett, D.C. (1997) Effects of bison and cattle on growth, reproduction, and abundances of five tallgrass prairie forbs. American Journal of Botany, 84, 17191728.
  • Doxon, E. D. (2009) Nesting and feeding ecology of grassland birds in mixed-prairie managed with patch-burn techniques. PhD Dissertation, Oklahoma State University.
  • Engle, D.M., Fuhlendorf, S.D., Roper, A. & Leslie, D.M. Jr (2008) Invertebrate community response to a shifting mosaic of habitat. Rangeland Ecology & Management, 61, 5562.
  • Fahnestock, J.T. & Knapp, A.K. (1993) Water relations and growth of tallgrass prairie forbs in response to selective grass herbivory by bison. International Journal of Plant Sciences, 154, 432440.
  • Fahnestock, J.T. & Knapp, A.K. (1994) Plant responses to selective grazing by bison - interactions between light, herbivory and water-stress. Vegetation, 115, 123131.
  • Fisher, R.J. & Davis, S.K. (2010) From Wiens to Robel: a review of grassland-bird habitat selection. Journal of Wildlife Management, 74, 265273.
  • Fryxell, J.M., Wilmshurst, J.F., Sinclair, A.R.E., Haydon, D.T., Holt, R.D. & Abrams, P.A. (2005) Landscape scale, heterogeneity, and the viability of Serengeti grazers. Ecology Letters, 8, 328335.
  • Fuhlendorf, S.D. & Engle, D.M. (2001) Restoring heterogeneity on rangelands: ecosystem management based on evolutionary grazing patterns. BioScience, 51, 625632.
  • Fuhlendorf, S.D. & Engle, D.M. (2004) Application of the fire-grazing interaction to restore a shifting mosaic on tallgrass prairie. Journal of Applied Ecology, 41, 604614.
  • Fuhlendorf, S.D. & Smeins, F.E. (1996) Spatial scale influence on longterm temporal patterns of a semi-arid grassland. Landscape Ecology, 11, 107113.
  • Fuhlendorf, S.D. & Smeins, F.E. (1999) Scaling effects of grazing in a semi-arid grassland. Journal of Vegetation Science, 10, 731738.
  • Fuhlendorf, S.D., Harrell, W.C., Engle, D.M., Hamilton, R.G., Davis, C.A. & Leslie, D.M. Jr (2006) Should heterogeneity be the basis for conservation? Grassland bird response to fire and grazing. Ecological Applications, 16, 17061716.
  • Fuhlendorf, S.D., Engle, D.M., Kerby, J. & Hamilton, R. (2009) Pyric herbivory: rewilding landscapes through the recoupling of fire and grazing. Conservation Biology, 23, 588598.
  • Fuhlendorf, S.D., Townsend, D.E. II, Elmore, R.D. & Engle, D.M. (2010) Pyric-herbivory to promote rangeland heterogeneity: evidence from small mammal communities. Rangeland Ecology & Management, 63, 670678.
  • Gillen, R.L. & Sims, P.L. (2004) Stocking rate, precipitation, and herbage production on sand sagebrush-grassland. Journal of Range Management, 57, 148152.
  • Godfree, R., Lepschi, B., Reside, A., Bolger, T., Robertson, B., Marshall, D. & Carnegie, M. (2011) Multiscale topoedaphic heterogeneity increases resilience and resistance of a dominant grassland species to extreme drought and climate change. Global Change Biology, 17, 943958.
  • Hamilton, R.G. (2007) Restoring heterogeneity on the Tallgrass Prairie Preserve: applying the fire-grazing interaction model. Proceedings of the 23rd Tall Timbers Fire Ecology Conference: Fire in Grassland and Shrubland Ecosystems (eds R.E. Masters & K.E.M. Galley), pp. 163169. Tall Timbers Research Station, Tallahassee, Florida.
  • Hartnett, D.C., Hickman, K.R. & Walter, L.E.F. (1996) Effects of bison grazing, fire, and topography on floristic diversity in tallgrass prairie. Journal of Range Management, 49, 413420.
  • Hobbs, N.T., Schimel, D.S., Owensby, C.E. & Ojima, D.S. (1991) Fire and grazing in the tallgrass prairie: contingent effects on nitrogen budgets. Ecology, 72, 13741382.
  • Holecheck, J.L., Pieper, R.D. & Herbel, C.H. (2004) Range Management Principles and Practices, 5th edn. Pearson-Prentice-Hall, Upper Saddle River, NJ.
  • Hsu, J. (1996) Multiple Comparisons: Theory and Methods. Chapman and Hall, London.
  • Kerby, J.D., Fuhlendorf, S.D. & Engle, D.M. (2007) Landscape heterogeneity and fire behavior: scale-dependent feedback between fire and grazing processes. Landscape Ecology, 22, 507516.
  • Kirkpatrick, J.B., Marsden-Smedley, J.B. & Leonard, S.W.J. (2011) Influence of grazing and vegetation type on post-fire flammability. Journal of Applied Ecology, 48, 642649.
  • Knopf, F.L. & Samson, F.B. (1997) Conservation of grassland vertebrates. Ecology and Conservation of Great Plains Vertebrates (eds F.L. Knopf & F.B. Samson), pp. 237289. Springer-Verlag, New York.
  • Leonard, S., Kirkpatrick, J. & Marsden-Smedley, J. (2010) Variation in the effects of vertebrate grazing on fire potential between grassland structural types. Journal of Applied Ecology, 47, 876883.
  • Littell, R.C., Milliken, G.A., Stroup, W.W., Wolfinger, R.D. & Schabenberger, O. (2006) SAS for Mixed Models, 2nd edn. SAS Institute Inc., Cary, NC, USA.
  • Meek, M.G., Cooper, S.M., Owens, M.K., Cooper, R.M. & Wappel, A.L. (2008) White-tailed deer distribution in response to patch burning on rangeland. Journal of Arid Environments, 72, 20262033.
  • Moe, S.R. & Wegge, P. (1997) The effects of cutting and burning on grass quality and axis deer (Axis axis) use of grassland in lowland Nepal. Journal of Tropical Ecology, 13, 279292.
  • Mouissie, A.M., Apol, M.E.F., Heil, G.W. & van Diggelen, R. (2008) Creation and preservation of vegetation patterns by grazing. Ecological Modelling, 218, 6072.
  • Murphy, B.P. & Bowman, D.M.J.S. (2007) The interdependence of fire, grass, kangaroos and Australian Aborigines: a case study from central Arnhem Land, northern Australia. Journal of Biogeography, 34, 237250.
  • Nkwabi, A.K., Sinclair, A.R.E., Metzger, K.L. & Mduma, S.A.R. (2010) Disturbance, species loss and compensation: wildfire and grazing effects on the avian community and its food supply in the Serengeti Ecosystem, Tanzania. Austral Ecology, 36, 403412.
  • Parr, C.L. & Andersen, A.N. (2006) Patch mosaic burning for biodiversity conservation: a critique of the pyrodiversity paradigm. Conservation Biology, 20, 16101619.
  • Pillsbury, F.C., Miller, J.R., Debinski, D.M. & Engle, D.M. (2011) Another tool in the toolbox? Using fire and grazing to promote bird diversity in highly fragmented landscapes. Ecosphere, 2, art. 28.
  • Pimm, S.L. (1984) The complexity and stability of ecosystems. Nature, 307, 321326.
  • Ring, C.B., Nicholson, R.A. & Launchbaugh, J.L. (1985) Vegetational traits of patch-grazed rangeland in west-central Kansas. Journal of Range Management, 38, 5155.
  • Robel, R.J., Briggs, J.N., Dayton, A.D. & Hulbert, L.C. (1970) Relationships between visual obstruction measurements and weight of grassland vegetation. Journal of Range Management, 23, 295297.
  • SAS Institute (2007) SAS Version 9.2. SAS Institute, Cary, NC, USA.
  • Savadogo, P., Zida, D., Sawadogo, L., Tiveau, D., Tigabu, M. & Odén, P.C. (2007) Fuel and fire characteristics in savanna-woodland of West Africa in relation to grazing and dominant grass type. International Journal of Wildland Fire, 16, 531539.
  • Scheintaub, M.R., Derner, J.D. & Knapp, A.K. (2009) Response of the shortgrass steppe plant community to fire. Journal of Arid Environments, 73, 11361143.
  • Senft, R.L. (1989) Hierarchical foraging models: effects of stocking and landscape composition on simulated resource use by cattle. Ecological Modelling, 46, 283303.
  • Sensenig, R.L., Demnent, M.W. & Laca, E.A. (2010) Allometric scaling predicts preferences for burned patches in a guild of East African grazers. Ecology, 91, 28982907.
  • Toombs, T.P. & Roberts, M.G. (2009) Are Natural Resources Conservation Service range management investments working at cross-purposes with wildlife habitat goals on Western United States rangelands? Rangeland Ecology & Management, 62, 351355.
  • Turner, M.G. (1989) Landscape ecology: the effect of pattern on process. Annual Review of Ecology & Systematics, 20, 171197.
  • USDA-NRCS. (2009) Web Soil Survey. Available at: http://websoilsurvey.nrcs.usda.gov. Accessed December 21, 2009 2009.
  • Van de Koppel, J., Bardgett, R.D., Bengtsson, J., Rodriguez-Barrueco, C., Rietkerk, M., Wassen, M.J. & Wolters, V. (2005) The effects of spatial scale on trophic interactions. Ecosystems, 8, 801807.
  • Vandvik, V., Heegaard, E., Maren, I.E. & Aarrestad, P.A. (2005) Managing heterogeneity: the importance of grazing and environmental variation on post-fire succession in heathlands. Journal of Applied Ecology, 42, 139149.
  • Vermeire, L.T. & Gillen, R.L. (2000) Western ragweed effects on herbaceous standing crop in Great Plains grasslands. Journal of Range Management, 53, 335341.
  • Vermeire, L.T., Mitchell, R.B., Fuhlendorf, S.D. & Gillen, R.L. (2004) Patch burning effects on grazing distribution. Journal of Range Management, 57, 248252.
  • Wiens, J.A. (1989) Spatial scaling in ecology. Functional Ecology, 3, 385397.
  • Wiens, J.A. (1997) The emerging role of patchiness in conservation biology. The Ecological Basis of Conservation: Heterogeneity, Ecosystems, and Biodiversity (eds S.T.A. Pickett, R.S. Osterfeld, M. Shachak & G.E. Likens), pp. 93107. Chapman and Hall, New York.
  • Williams, R.E. (1954) Modern methods of getting uniform use of ranges. Journal of Range Management, 7, 7781.
  • Yarnell, R.W., Scott, D.M., Chimimba, C.T. & Metcalfe, D.J. (2007) Untangling the roles of fire, grazing and rainfall on small mammal communities in grassland ecosystems. Oecologia, 154, 387402.