University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln
Papers in Natural Resources Natural Resources, School of
7-2013
Seasonal Fires, Bison Grazing, and the Tallgrass Prairie Forb Arnoglossum plantagineum Raf.
Stephen L. Winter [email protected]
Karen R. Hickman Oklahoma State University
Carla L. Goad Oklahoma State University
Samuel D. Fuhlendorf Oklahoma State University, [email protected]
Mark S. Gregory Oklahoma State University
Follow this and additional works at: https://digitalcommons.unl.edu/natrespapers
Part of the Natural Resources and Conservation Commons, Natural Resources Management and Policy Commons, Other Environmental Sciences Commons, and the Other Plant Sciences Commons
Winter, Stephen L.; Hickman, Karen R.; Goad, Carla L.; Fuhlendorf, Samuel D.; and Gregory, Mark S., "Seasonal Fires, Bison Grazing, and the Tallgrass Prairie Forb Arnoglossum plantagineum Raf." (2013). Papers in Natural Resources. 449. https://digitalcommons.unl.edu/natrespapers/449
This Article is brought to you for free and open access by the Natural Resources, School of at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Papers in Natural Resources by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. R E S E A R C H N O T E ABSTRACT: Fire and grazing can interact to affect the structure and composition of vegetation com- munities in a manner that may differ from the effects of fire or grazing that occurs in isolation of the other. In order to better understand the effects of a fire-grazing interaction at the level of an individual plant species, we studied the response of a perennial tallgrass prairie forb, Arnoglossum plantagineum Raf., to the interaction of spring and summer fires with grazing by bison (Bison bison L.). During one field season (2006), we collected data in areas that had been treated with summer fires while in a • subsequent field season (2007) we collected data in areas that had been treated with spring fires. Many measures of plant growth (plant height, vegetative biomass, and total biomass) and reproductive effort (reproductive biomass, indices of flowering plant density) suggested greater resource availability for Seasonal Fires, individuals of A. plantagineum growing in areas that had been recently burned and were being heavily grazed by bison. However, the response of these variables to the fire-grazing interaction often varied Bison Grazing, and among differing topographical positions. Our results demonstrate that the interaction of fire and bison grazing can further interact with topographical position in tallgrass prairie to affect the growth and the Tallgrass Prairie reproductive effort of the perennial forb A. plantagineum. Forb Arnoglossum Index terms: heterogeneity, patch burn grazing, pyric herbivory, shifting mosaic
plantagineum Raf. INTRODUCTION richness of this ecosystem (Freeman 1998; Palmer 2007). Fire may affect the diversity A variety of management tools and of tallgrass prairies by altering competitive techniques are available to natural area Stephen L. Winter1,4 relationships between dominant grasses managers tasked with the preservation, and subdominant forbs; spring fires en- 1 Winona, MN 55987 maintenance, and restoration of tallgrass hance the growth of dominant warm-season prairie ecosystems. Prescribed fire has a grasses, which may then preempt resources Karen R. Hickman2 long history of application in the restora- available to subdominant forbs, resulting tion and management of tallgrass prairies Carla L. Goad3 in reduced forb growth or reproductive ef- (Kollmorgen and Simonett 1965; Hoy fort (Knapp 1984; Hartnett 1990; Hartnett 2 Samuel D. Fuhlendorf 1989), particularly in the context of natural 1991). Frequent spring fires can contrib- Mark S. Gregory2 area management and restoration where ute to greater dominance of grasses with its use is widespread (Curtis and Partch concomitant declines in the abundance of 1948; Helzer 2010; Rowe 2010). The use forbs (Kucera and Koelling 1964; Collins 2 Department of Natural Resource of grazing as a restoration and management and Steinauer 1998). Ecology and Management tool in tallgrass prairies has not been as Oklahoma State University ubiquitous as the use of fire. Certainly, in Grazing in tallgrass prairie can modulate Stillwater, OK 74078 western portions of the tallgrass prairie, community diversity by reducing the there is a long history of cattle (Bos taurus competitive ability of dominant grasses 3 Department of Statistics L.) grazing as an economic activity that when they are grazed and by altering the Oklahoma State University has shaped the ecosystem in areas where environmental conditions at the microsite Stillwater, OK 74078 it has occurred (Malin 1942; Kollmorgen occupied by an individual forb (Collins et and Simonett 1965). In other areas, perhaps al. 1998). Selective grazing of grasses can especially in the eastern portion of this create a favorable growing environment • ecosystem’s range, there has been debate around a prairie forb as a result of greater over the value of large herbivore grazing as levels of light and higher soil temperatures a conservation tool (Williams 1997; Har- (Fahnestock and Knapp 1993, 1994). Prai- rington 1998; Henderson 1998; Davison rie forbs growing among grazed grasses and Kindscher 1999; Howe 1999; Knapp can be characterized by greater biomass 4 Corresponding author: et al. 1999; Leach et al. 1999). than conspecifics in ungrazed areas, exem- [email protected]; 402-310- plifying the relationship between resource 5460 Management of tallgrass prairies is often availability and biomass accumulation focused on manipulating the richness, (Fahnestock and Knapp 1993, 1994; Dam- diversity, and composition of vegetation houreyeh and Hartnett 1997). Somewhat communities (Brudvig et al. 2007). While counterintuitively, prairie forbs growing in vegetation productivity of tallgrass prairies grazed areas can be shorter than conspecif- is driven largely by a few species of warm ics growing in ungrazed areas (Fahnestock Natural Areas Journal 33:327–338 season grasses (Knapp et al. 1998), forbs and Knapp 1993, 1994; Damhoureyeh make the greatest contribution to species and Hartnett 1997), suggesting that forbs
Volume 33 (3), 2013 Natural Areas Journal 327 growing among ungrazed grasses need to selective grazing of grasses within recently multiple species of tallgrass prairie forbs, be taller to obtain sufficient light resources. burned patches, contrasting with the limited but we were constrained by limitations on When prairie forbs are able to capitalize on grazing of grasses within patches that have the time available for field activities. increased resource availability as a result not burned recently (Coppedge et al. 1998; of altered competitive interactions with Winter et al. 2012). We conducted our study to gain insight neighboring grasses, they may allocate into whether a prairie forb experiences biomass in a manner that differs among spe- However, grassland landscapes are charac- different levels of resource availability in cies: forbs that are primarily asexual may terized by topoedaphic variability that also the different patches characterizing the increase total biomass while forbs that are affects vegetation communities (Barnes and shifting mosaic of a fire-grazing interac- primarily sexual may increase reproductive Harrison 1982; Knapp et al. 1993; Dodd tion. While we did not explicitly measure biomass, number of florets, or number of et al. 2002; Winter et al. 2011). Slope and resource availability (e.g., levels of light, seeds (Fahnestock and Knapp 1993, 1994; aspect are topoedaphic features that influ- water, or nutrients) in patches with different Damhoureyeh and Hartnett 1997). ence the microsite conditions experienced burn histories, we assume our measures by individual plants. Contrasting slope of plant growth and reproductive effort A management practice gaining increased and aspect features can be characterized infer resource availability. Specific null attention in the last decade is known vari- by contrasting levels of soil temperature hypotheses that we tested in our research ously as patch burn grazing, heterogeneity and soil moisture content (Ayyad and Dix were: (1) measures of plant growth (height, management, the fire grazing interaction, 1964), contrasting plant leaf temperatures vegetative biomass, and total biomass) or pyric herbivory (Fuhlendorf and Engle (Smith et al. 1983), and contrasting level does not differ between plants growing in 2001; Fuhlendorf et al. 2006; Fuhlendorf et of response by grasses and forbs to altered patches that were recently burned and were al. 2009). This practice capitalizes on the levels of nutrient availability (Benning and being heavily grazed and plants growing profound influence of fire on the distribu- Seastedt 1995). Thus, any shifting mosaic in patches that were not recently burned tion and behavior of grazing animals to alter of altered competitive dynamics between and were being minimally grazed; and (2) vegetation heterogeneity across landscapes dominant grasses and subdominant forbs measures of reproductive effort (number of (Fuhlendorf et al. 2004; Vermeire et al. arising from a fire-grazing interaction may floral structures, reproductive biomass, and 2004; Fuhlendorf et al. 2006; Coppedge et be constrained or modulated by a pre-ex- density of flowering individuals) does not al. 2008; Allred et al. 2011; McGranahan isting template of heterogeneity imposed differ between plants growing in patches et al. 2012; Winter et al. 2012). Because by topoedaphic variability (Winter et al. that were recently burned and were be- application of the fire grazing interaction 2011). ing heavily grazed and plants growing in has been the subject of focused research patches that were not recently burned and for a relatively short time, land managers To better understand the effect of a fire- were being minimally grazed. accustomed to using fire alone or grazing grazing interaction on individual prairie alone may have uncertainty regarding its forbs, we collected data on a common METHODS use. perennial forb of tallgrass prairies, Arno- glossum plantagineum Raf. Furthermore, Fire and grazing can interact in landscapes we quantified the topographical setting Species Description to create heterogeneity when fire concen- of individual plants that were sampled to trates grazing activity within a burned patch determine if there was an influence of slope Arnoglossum plantagineum (Indian plan- of a landscape, thereby reducing grazing and aspect on the plant variables that were tain) is a perennial herb arising from a activity in other landscape patches that measured. Our research was conducted at weakly rhizomatous caudex with fleshy- haven’t burned for an extended period of The Nature Conservancy’s Tallgrass Prairie fibrous roots (Barkley 1986; Anderson time; a subsequent fire that occurs in an- Preserve in northeast Oklahoma, a natural 2006). Spring growth at our study site was other patch of the same landscape causes area where the interacting processes of fire initiated with the production of a basal grazing activity to shift to the newly burned and bison (Bison bison L.) grazing are used rosette of long-petioled leaves. Anthesis patch (Fuhlendorf and Engle 2004; Allred to restore historic patterns of disturbance is realized with the production of a stem et al. 2011). The resulting dynamics of and rest (Hamilton 2007). We selected A. terminating in a broad, corymbiform disturbance and rest through space and plantagineum as a study species because cyme, and 50–100 cm in height. The cyme time create a patchwork of contrasting its level of abundance at the study site was consists of numerous capitula, each con- vegetation structure and composition sufficient to permit sampling, it had an taining five florets. Flowering individuals (Fuhlendorf and Engle 2004; Vermeire et architecture that permitted differentiation of A. plantagineum at our study site were al. 2004; Coppedge et al. 2008; McGrana- of vegetative and reproductive biomass, it typically single-stemmed. Non-flowering han et al. 2012; Winter et al. 2012). The did not appear to be palatable to bison, and individuals (i.e., presence of only a basal compositional changes that characterize it has not been the subject of previous aut- rosette) were present at our study site this shifting mosaic are presumed to arise ecological research in tallgrass prairie. We into July but anemochory of seeds from from the altered competitive dynamics acknowledge that more information would individuals that did flower appeared to between grasses and forbs driven by the have been generated by an examination of be well-synchronized, occurring in mid-
328 Natural Areas Journal Volume 33 (3), 2013 July. In the northern part of its range, A. fire return interval for most areas of the our sampling to flowering individuals. We plantagineum occurs from South Dakota pasture (Hamilton 2007). In 2006, the further restricted our sampling to individu- to Ontario, while in the south it ranges bison pasture was 8517 ha and contained als that were single-stemmed, the typical from Texas through Alabama (Anderson approximately 2400 animals during the condition for this species at this study site. 2006). Typical habitats of A. plantagineum summer. In 2007, the pasture was expanded Additionally, we never observed evidence are prairies and pastures where it can be to 9532 ha and contained approximately of herbivory on this species by bison, but a common species (Smeins and Diamond 2600 animals during the summer. With we did occasionally observe, and did not 1983; Foti 1989; Leidolf and McDaniel the exception of an annual fall roundup sample, plants that appeared to have been 1998; Hickman and Derner 2007; Polley in which surplus animals are culled and trampled. Finally, to reduce the confound- et al. 2007). remaining animals receive veterinary main- ing effect of intra-specific competition, tenance treatments, bison are free to move we did not sample individuals that were throughout the pasture during the year and rooted within 5 cm of a neighboring A. Study Site receive no supplemental feed. Salt and plantagineum. Within each sub-popula- trace minerals are provided to bison on a tion, plants suitable for sampling were The study site was The Nature Conservan- free-choice basis at multiple locations, and located visually until a total of 30 plants cy’s 15,700-ha Tallgrass Prairie Preserve streams and stock ponds provide water at had been sampled. The geographic coor- in Osage County, Oklahoma (36°50’N, scattered locations throughout the pasture dinates of each sampled individual were 96°25’W). Long-term (1945 – 2007) av- (Hamilton 2007). Bison distribution within recorded using a Garmin GPS Map60CSX. erage annual precipitation at the nearby the pasture is strongly correlated with The Map60CSX was used with the Wide Foraker weather station in Osage County the location of burned patches with the Area Augmentation System enabled, and was 89.21 cm (www.ncdc.noaa.gov; www. most recently-burned patches being areas accuracy of each sampled plant location mesonet.org). In 2005, 2006, and 2007, characterized by the highest level of bison was enhanced through use of the averaging the annual total precipitation and percent activity (Coppedge and Shaw 1998; Schuler function; estimated accuracy, as indicated deviation from the long-term average were et al. 2006; Allred et al. 2011). by the unit during averaging, was consis- 77.75 cm (87%), 61.09 cm (68%), and tently < 10 m. Sampled individuals were 133.25 cm (149%), respectively. Soils are harvested at the ground surface and all tis- derived from sandstone, shale, limestone, 2006 Field Sampling sues (basal leaves, stem, and inflorescence) and chert of Permian and Pennsylvanian were individually-bagged and labeled for age (USDA-NRCS 2009). Specific areas There has been minimal work examin- further processing in an enclosed, climate- sampled during this study were character- ing the response of prairie vegetation to controlled laboratory. All sampling in 2006 ized by soils in the Ustoll (very shallow prescribed fires conducted during seasons occurred during June 21–26. to very deep; moderately well-drained to other than spring (Engle and Bidwell 2001; well-drained; slopes of 0–30% slope), Towne and Kemp 2008; Howe 2011), and Udert (deep; moderately well-drained; we elected to conduct our study in a man- 2007 Field Sampling 0–5% slopes), and Ustalf (moderately ner that would contribute to this limited deep; moderately to well-drained; 0–30% body of knowledge. Thus, we compared Because a statewide burn ban was in effect slopes) suborders (USDA-NRCS 2009). plants in recent summer-burned patches during the summer and fall of 2006, no The majority of the Preserve is character- to plants in older summer burn patches burns occurred within the bison pasture ized by tallgrass prairie vegetation with the within the Preserve’s bison pasture. Three during the period of July – September dominant grass species being Andropogon sites of paired patches were chosen in 2006 2006. Consequently, it was not possible gerardii Vitman., Sorghastrum nutans (L.) where a patch that was burned during the in 2007 to replicate the manner in which Nash, Sporobolus compositus (Poir.) Merr., late growing season (July – September) in summer burns were sampled in 2006. Panicum virgatum L., and Schizachyrium 2005 was adjacent to another patch that Instead, in 2007 we located three sites scoparius (Michx.) Nash (Palmer 2007). had been burned during the same season of paired patches where a patch that was in 2003 or earlier. Recent summer burn burned during the spring (March – April) Beginning in 1993, management of the patches (burned in 2005) were 85–114 ha in 2007 was adjacent to another patch that Preserve has included the restoration of (mean = 114 ha), and older summer burned had been burned during the same season in historic cycles of rest and disturbance patches (burned in 2001, 2002 or 2003) 2005 or earlier. Thus, we compared plants driven by the interaction between fire were 81–206 ha (mean = 161 ha). in recent spring-burned patches to plants and bison grazing. Within the Preserve’s in older spring-burned patches within bison pasture, seasonal fires are applied At each site of paired patches, a population the bison pasture. Recent spring-burned in a spatially-random manner such that of A. plantagineum was located in which patches (burned in 2007) were 88–366 ha approximately 40% of the area burned the population was subdivided by the (mean = 194 ha) and older burned patches in a single year is burned during March boundary between the burn patches. Non- (burned in 2004 or 2005) were 113–293 – April, 20% during July – September, flowering individuals (i.e., basal rosettes) ha (mean = 184 ha). At each site of paired and 40% during November – December, were readily obscured by other vegetation patches, a population of A. plantagineum resulting in an approximately three-year and difficult to locate, and we restricted was located in which the population was
Volume 33 (3), 2013 Natural Areas Journal 329 subdivided by the boundary between the slope, and plant location data layers were Statistical Analysis burn patches. The methods used to locate, used as input features in the Intersect tool determine the sampling suitability, and of the Analysis toolbox. Statistical analyses were conducted using eventually sample individual plants was the SAS/GLIMMIX procedure (version identical to the methods utilized in 2006 Aspect and slope values, originally ex- 9.2; SAS Institute 2007) with Kenward- except that 40 individuals were sampled pressed as degrees, were transformed us- Roger degrees of freedom adjustments. within each sub-population of each patch- ing trigonometric functions (Guisan et al. Plant variables (height, number of capitula, pair during 2007. All sampling in 2007 1999) and the product of the transformed biomass) measured in 2006 and 2007 were occurred during June 17–24. values resulted in values of “northness” and analyzed using a linear mixed model where “eastness” (Kariuki et al. 2006) that could patch-pair identity, plant sub-population identity, and individual plant identity were be associated with each sampled plant: Laboratory Procedures treated as random effects, while burn his- tory (recent burn, older burn) was treated as northness = [sine(slope)]*[cosine(aspe During both years, immediately after field a fixed effect. Northness and eastness were ct)] sampling, all plants were transported to included as covariates in the mixed model a climate-controlled laboratory where analyses. Density indices (NND, MCP, plant height was measured from the cut eastness = [sine(slope)]*[sine(aspect)] density of flowering individuals) were portion of the stem to the farthest point analyzed by sample year using a general- that an individual’s inflorescence tissues Thus, each sampled plant was associated ized linear mixed model with a fixed effect could be extended along a flat surface. The with a value of northness, ranging from -1 of burn history. For all analyses, statistical number of capitula on each individual was through 1, indicating if it was on a southern significance was set at α = 0.05. counted. Vegetative tissue (basal leaves, (northness = -1) or northern (northness stem, and stem leaves) of each individual = 1) aspect (northness = 0 indicated no was separated from reproductive tissue orientation with regard to south or north). RESULTS (inflorescence consisting of stem, pedicels, Likewise, each sampled plant was asso- and capitula) by cutting the stem at the base ciated with a value of eastness, ranging of the lowest stem leaf that had a pedicel 2006 Sampling of Summer Burns with at least one capitula originating from from -1 through 1, indicating if it was on its axis. Reproductive and vegetative tis- a western (eastness = -1) or eastern (east- Mixed model analyses identified a signifi- sue of each individual were bagged and ness = 1) aspect (eastness = 0 indicated cant effect of burn history on vegetative labeled separately and oven dried at 60 °C no orientation with regard to east or west). biomass (P = 0.04). The aspect variables for 48 hours. Following drying, samples Subsequent analyses were done using eight were adopted as covariates in the analysis were weighed to obtain measurements of of the possible combinations of northness of plant height (P ≤ 0.014), total biomass reproductive biomass, vegetative biomass, and eastness values (north, northeast, east, (P = 0.045), and reproductive biomass (P ≤ 0.046). After adopting the aspect variables and total (reproductive + vegetative) southeast, south, southwest, west, and as covariates, post-hoc analyses comparing biomass. northwest) as well as the condition of no the two burn histories at selected values aspect (slope = 0). of the covariates were performed. Height GIS Procedures of plants within recent burns was lower Analyzing the data with ArcMap, we ob- (P < 0.001) than those in older burns on A Geographic Information System (GIS) tained multiple indices of the density of southwestern and western aspects while was constructed using the software pack- flowering individuals. First, we used the vegetative biomass of plants in recent age ArcMap 9.3 (www.esri.com). A 30-m Average Nearest Neighbor tool, within summer burns was higher (P ≤ 0.04) than grid Digital Elevation Model (DEM) was the Spatial Statistics toolset of ArcMap, those in older summer burns on northern, obtained from the National Elevation to compute nearest neighbor distances northeastern, eastern, southeastern, and southern aspects as well as on areas with Dataset (Gesch 2007) and combined with (NND) for all sampled plants within each no slope or aspect (Figure 1). In 2006, the GPS coordinates of each sampled sub-population of each patch-pair during MCP values were higher (P = 0.028) in plant from each year of sampling within each year of sampling. Additionally, we older summer burns and density of flower- the GIS. Within the Spatial Analyst tool- used the Hawth’s Tools extension (Beyer ing individuals was higher (P = 0.029) in box of ArcMap, the DEM was used as an 2004) to construct minimum convex poly- recent summer burns (Figure 2). input raster in the Aspect and Slope tools, gons (MCP) enclosing each sub-population creating aspect and slope datasets. Aspect of each patch-pair during each year, to 2007 Sampling of Spring Burns and slope datasets were subsequently con- obtain two additional indices of the density verted to polygon layers using the Raster of flowering individuals: the area (ha) of The aspect variables were adopted as co- to Polygon tool in the Conversion toolbox. each MCP and the density (flowering in- variates in the analysis of plant height (P In order to associate each sampled plant dividuals/ha) of sampled plants contained = 0.026), vegetative biomass (P ≤ 0.027), with an aspect and slope value, the aspect, within each MCP. total biomass (P ≤ 0.029), reproductive
330 Natural Areas Journal Volume 33 (3), 2013 Figure 1. Differences between the estimated least square means for Arnoglossum plantagineum plant variables in recent (2005) and older (2001 – 2003) sum- mer burns for each aspect category in 2006. A positive value indicates the least square means estimate for the recent burn was greater than the estimate for the older burn; a negative value indicates the estimate for the recent burn was less than the estimate for the older burn. Significance levels for tests of recent burn means versus older burn means are as follows: * indicates P < 0.05; ** indicates P < 0.01; *** indicates P < 0.001). biomass (P = 0.027), and number of ca- spring burns on southern aspects (Figure was higher (P ≤ 0.031) than within older pitula (P ≤ 0.049). After adopting the aspect 3). Vegetative biomass within recent spring burns on northwestern, northeastern, variables as covariates, post-hoc analyses spring burns was higher (P ≤ 0.026) than eastern, southeastern, southwestern, and comparing the two burn histories at selected within older spring burns on northwestern, western aspects. The number of capitula values of the covariates were performed. northeastern, eastern, southeastern, south- per plant in recent spring burns was higher Height of plants within recent spring burns western, and western aspects. Likewise, (P = 0.017) on northeastern aspects but was lower (P = 0.010) than those in older total biomass within recent spring burns this variable was lower (P ≤ 0.039) than
Volume 33 (3), 2013 Natural Areas Journal 331 the number of capitula per plant in older spring burns on southern, southwestern, and western aspects. In 2007, NND and MCP values were higher (P = 0.038 and P = 0.032, respectively) in older spring burns while flowering plant density was higher (P = 0.033) in recent spring burns (Figure 4).
DISCUSSION
Our results demonstrate that fire, bison grazing, and topographic position interact to influence the growth and reproductive effort of A. plantagineum. In most instances where A. plantagineum responded to this interaction, the response suggested greater resource availability in recently burned patches that were being heavily grazed relative to patches where a longer period of time had elapsed since burning and where minimal grazing was occurring. Further- more, this response occurred when the fire-grazing interaction was driven by both summer fires and spring fires. During both years of our study, plant height was usually similar between patches with differing burn histories for most aspect categories, but there were instances where plant height was significantly lower in recently burned patches that were being heavily grazed compared to plants that were in patches that weren’t recently burned and were being minimally grazed. This suggests that, on some topographical aspects, grazing in a re- cently burned patch allowed sufficient light to reach A. plantagineum plants such that this forb could allocate carbon resources to life history attributes other than height, such as leaves or reproductive structures. A similar response of plant height in areas grazed by bison compared to ungrazed areas was reported for the prairie forbs Am- brosia psilostachya DC., Symphyotrichum ericoides (L.) G.L. Nesom var. ericoides, and Vernonia baldwinii Torr. (Fahnestock and Knapp 1993, 1994). A caveat of the research done by Fahnestock and Knapp (1993, 1994) is that their use of the term grazed implied that sampling occurred in Figure 2. Estimates of least square means (± SE) for Arnoglossum plantagineum density indices in recent areas where the vegetation matrix sur- (2005) and older (2001 – 2003) summer burns measured in 2006. Different letters above bars indicate significant differences at the α = 0.05 level. rounding a sampled forb was primarily grasses that had been recently grazed. Conversely, their use of the term ungrazed implied that sampling occurred in an area
332 Natural Areas Journal Volume 33 (3), 2013 Figure 3. Differences between the estimated least square means for Arnoglossum plantagineum plant variables in recent (2007) and older (2004 – 2005) spring burns for each aspect category in 2007. A positive value indicates the least square means estimate for the recent burn was greater than the estimate for the older burn; a negative value indicates the estimate for the recent burn was less than the estimate for the older burn. Significance levels for tests of recent burn means versus older burn means are as follows: * indicates P < 0.05; ** indicates P < 0.01; *** indicates P < 0.001). where the vegetation matrix surrounding Damhoureyeh and Hartnett (1997) found forbs (Baptisia bracteata Muhl. ex Elliott, a sampled forb was primarily grasses that that height of the prairie forb Oenothera Salvia azurea Michx. ex Lam., and V. had not been grazed recently, even though speciosa Nutt. was less in bison-grazed baldwinii) when bison-grazed areas were areas relative to ungrazed areas (the un- compared to ungrazed areas. Damhoureyeh that area was actually in a pasture where grazed areas being pastures that were not and Hartnett (1997) also reported an effect grazing animals were present and grazing being grazed by bison or cattle). However, of fire frequency on the response of height may have occurred earlier in the season or Damhoureyeh and Hartnett (1997) found in the prairie forb Solidago missouriensis in previous seasons. no differences in the height of three prairie Nutt. to bison grazing; in annually burned
Volume 33 (3), 2013 Natural Areas Journal 333 watersheds there was no difference in height, in watersheds burned every fourth year plants in ungrazed areas were taller than plants in grazed areas. When Dam- houreyeh and Hartnett (1997) compared areas grazed by cattle to ungrazed areas, however, they found no differences in plant height for B. bracteata, O. speciosa, S. missouriensis, and V. baldwinii, regard- less of fire frequency. Damhoureyeh and Hartnett (1997) did find that height of S. azure was greater in ungrazed areas than cattle grazed areas in watersheds that were annually burned, but there were no differ- ences in height of S. azure in watersheds that were burned every fourth year.
During our study in 2006, total biomass did not differ between patches that had been recently burned and were heavily grazed compared to patches that had not recently burned and were minimally grazed, regard- less of aspect category. In 2007, however, total biomass of plants in recently burned patches was higher for most aspect catego- ries compared to total biomass of plants in patches that had not recently burned. During both years, vegetative biomass of plants in recently burned patches was higher for most aspect categories compared to total biomass of plants in patches that had not recently burned. Greater total bio- mass in areas grazed by bison compared to ungrazed areas was also reported for A. psilostachya (Fahnestock and Knapp 1993, 1994), while Damhoureyeh and Hartnett (1997) reported that B. bracteata had greater total biomass in bison-grazed areas compared to ungrazed areas within watersheds burned every fourth year, but not in watersheds that were burned an- nually. Total biomass of V. baldwinii was reported by Fahnestock and Knapp (1993, 1994) to not differ between areas grazed by bison compared to ungrazed areas; to not differ in annually burned watersheds grazed by bison compared to annually burned wa- tersheds that were ungrazed (Damhoureyeh and Hartnett 1997); but to be greater in watersheds that were burned every fourth year and grazed by bison compared to watersheds that were burned every fourth Figure 4. Estimates of least square means (± SE) for Arnoglossum plantagineum density indices in recent year and were ungrazed (Damhoureyeh (2007) and older (2004 – 2005) spring burns measured in 2007. Different letters above bars indicate and Hartnett 1997). significant differences at the α = 0.05 level. Contrasting with much of the reported
334 Natural Areas Journal Volume 33 (3), 2013 results for forb biomass in bison grazed During both years of our study, we found 1984; Hartnett 1991). and ungrazed areas, total biomass of S. higher indices of flowering plant density ericoides was reported by Fahnestock and in recent burns compared to older burns. The fire grazing interaction has been Knapp (1993) to be greater in ungrazed We acknowledge these aren’t indices of demonstrated to alter vegetation structure areas compared to areas grazed by bison. actual population density because we and composition (Fuhlendorf and Engle However, S. ericoides is a species known only recorded the presence of individuals 2004; Vermeire et al. 2004; Coppedge et al. to be consumed by cattle, and a relation- that were flowering. Flowering individu- 2008; Winter et al. 2012). For natural area ship between increasing cattle stocking als of A. plantagineum, which can have managers, application of the fire grazing rates and decreasing biomass of S. ericoi- an inflorescence positioned up to 0.10 m interaction may be especially useful for des has been documented (Hickman and above the ground, were readily visible the achievement of management objectives Hartnett 2002). Indeed, Fahnestock and during sampling while individuals that that seek to increase the heterogeneity or Knapp (1993) indicated that some of the weren’t flowering at the time of sampling diversity of vegetation in prairies (Helzer S. ericoides plants they sampled in grazed were most often represented by only a and Steuter 2005; Fuhlendorf et al. 2006). areas appeared to have had some of their basal rosette of leaves and were much Previous research has shown that forb plant parts removed by grazing bison. In less visible. This was especially the case communities or functional groups respond areas grazed by cattle, Damhoureyeh and within patches that hadn’t been burned positively to the fire grazing interaction Hartnett (1997) reported that total biomass for multiple years and a rank canopy of (Fuhlendorf and Engle 2004; Vermeire et of B. bracteata and O. speciosa was greater grasses was present. Nonetheless, our al. 2004; Coppedge et al. 2008; Winter et when compared to ungrazed areas, while density indices provide further evidence al. 2012). Previous research on individual total biomass of S. azurea, S. missouriensis, that A. plantagineum individuals within a prairie forb species has highlighted the role and V. baldwinii did not differ between recently burned patch were responding to of altered competitive dynamics between grazed and ungrazed areas, regardless of greater resource availability. If there is an dominant grasses and subdominant forbs in burn frequency. assumption that actual population density driving the response of forbs at the level of of this perennial plant was equal in all an individual plant (Knapp 1984; Hartnett In 2006, we found no differences in repro- patches, the higher density of flowering 1991; Fahnestock and Knapp 1993, 1994; ductive biomass or the number of capitula individuals within recently burned patches Damhoureyeh and Hartnett 1997). Our re- when recent summer burns were compared implies that a greater proportion of indi- sults for a single prairie forb provide further to older summer burns, regardless of topo- viduals within recently burned patches had evidence of the presumed mechanism for graphic position. In 2007 when spring burns access to sufficient resources permitting all of these responses – forbs growing in were sampled, there were no differences in sexual reproduction. areas that have been recently burned, and reproductive biomass at any topographic where their neighboring grasses are being In the landscape we studied, where pyric heavily grazed, can be characterized by position. There was variability, however, herbivory and topography interacted to enhanced growth and reproductive effort, in the response of number of capitula create a shifting mosaic of heterogeneity, regardless of fire season. This implies the that year. This variable was greater on A. plantagineum may have been responding fire grazing interaction provides prairie some topographic positions, less on other to alterations of environmental conditions forbs with a period of enhanced resource topographic positions, and did not differ (light levels, soil temperature) experienced availability. on the remaining topographic positions at the microsite of an individual plant as when recent spring burns were compared well as to altered competitive relationships The response of individual prairie forbs to to older spring burns. Vernonia baldwinii with neighboring grasses. The occurrence altered competitive dynamics, as mediated was shown to have greater reproductive of fire alone or grazing alone could poten- by fire and grazing, has been shown to vary biomass in bison-grazed areas compared tially result in these alterations as well, but depending on the forb being studied, the to ungrazed areas (Fahnestock and Knapp it is possible that the interaction of fire and plant variables being measured, fire fre- 1993, 1994). The number of florets and grazing that arises from pyric herbivory quency, and the species of herbivore; but flowering heads of V. baldwinii were facilitates an alteration of environmental the majority of results indicate that grazing greater in bison-grazed areas compared conditions and community interactions that of neighboring grasses around an indi- to ungrazed areas (Fahnestock and Knapp are unique, relative to the occurrence of vidual forb has neutral to positive effects 1993, 1994). Damhoureyeh and Hartnett either fire alone or grazing alone. This is on the forb (Fahnestock and Knapp 1993, (1997) measured multiple variables rep- supported by research on V. baldwinii and 1994; Damhoureyeh and Hartnett 1997). resenting reproductive structures (flowers, Ratiba columnifera (Nutt.) Woot. & Standl. Our results provide evidence of another flower heads, capsules, seeds, etc.); and in in landscapes that were only burned and variable that influences the response of a all instances when there were significant were not grazed, where multiple measures prairie forb to altered competitive dynamics differences between grazed areas and un- of plant growth and reproductive effort – the response of A. plantagineum to the grazed areas, those variables were greater had their greatest values in unburned wa- fire grazing interaction varied depending in grazed areas, regardless of whether the tersheds, or watersheds that had gone the on topographical position. This indicates grazing was done by bison or cattle. longest time without being burned (Knapp that the effects of land management ac-
Volume 33 (3), 2013 Natural Areas Journal 335 tions, such as application of the fire graz- Samuel Fuhlendorf is a Professor in the of diversity by grazing and mowing in native ing interaction, may vary among differing Department of Natural Resource Ecol- tallgrass prairie. Science 280:745-747. locations of a landscape. ogy and Management at Oklahoma State Collins, S.L., and E.M. Steinauer. 1998. Distur- University. His research interests include bance, diversity, and species interactions in tallgrass prairie. Pp. 140–156 in A.K. Knapp, landscape ecology and the interaction of J.M. Briggs, D.C. Hartnett, and S.L. Col- fire and grazing animals. ACKNOWLEDGMENTS lins, eds., Grassland Dynamics: Long-Term Ecological Research in Tallgrass Prairie. Funding was provided by a State Wildlife Mark Gregory is an Assistant Researcher Oxford University Press, New York. Grant under Project T-30-P of the Okla- in the Department of Natural Resource Coppedge, B.R., D.M. Engle, C.S. Toepfer, and J.H. Shaw. 1998. Effects of seasonal homa Department of Wildlife Conservation Ecology and Management at Oklahoma State University whose research interests fire, bison grazing and climatic variation on and Oklahoma State University and admin- tallgrass prairie vegetation. Plant Ecology istered through the Oklahoma Cooperative include the application of spatial tech- 139:235-246. nologies (GIS, GPS, remote sensing) to Fish and Wildlife Research Unit; and the Coppedge, B.R., S.D. Fuhlendorf, W.C. Harrell, National Research Initiative of the USDA the evaluation and management of natural and D.M. Engle. 2008. Avian community Cooperative State Research, Education and resources. response to vegetation and structural features Extension Service, grant number 2003- in grasslands managed with fire and grazing. Biological Conservation 141:1196-1203. 35101-12928. The Oklahoma Chapter of Coppedge, B.R., and J.H. Shaw. 1998. Bison The Nature Conservancy provided logisti- LITERATURE CITED grazing patterns on seasonally burned tall- cal assistance and lodging during both field grass prairie. Journal of Range Management seasons. Assistance with field sampling Allred, B.W., S.D. Fuhlendorf, D.M. Engle, 51:258-264. and laboratory procedures was provided and R.D. Elmore. 2011. Ungulate prefer- Curtis, J.T., and M.L. Partch. 1948. Effect of by A. Ainsworth, L. Haynes, J. Lofton, S. ence for burned patches reveals strength fire on the competition between blue grass Robertson, K. Spears, and J. Worthington. of fire-grazing interaction. Ecology and and certain prairie plants. American Midland Evolution 1:132-144. Additional assistance in the laboratory Naturalist 39:437-443. was provided by M. Howe, D. Varney, C. Anderson, L.C. 2006. Arnoglossum. Pp. Damhoureyeh, S.A., and D.C. Hartnett. 1997. 622–625 in F.O.N.A. Committee, ed., Flora Walden, and L. Wilkerson. The manuscript Effects of bison and cattle on growth, re- of North America. Oxford University Press, production, and abundances of five tallgrass benefitted from comments and suggestions New York. prairie forbs. American Journal of Botany by two anonymous reviewers. Ayyad, M.A.G., and R.L. Dix. 1964. An analysis 84:1719-1728. of a vegetation-microenvironmental complex Davison, C., and K. Kindscher. 1999. Tools on prairie slopes in Saskatchewan. Ecologi- for diversity: fire, grazing and mowing on cal Monographs 34:421-442. tallgrass prairies. Ecological Restoration Stephen Winter obtained a PhD in Range- Barkley, T.M. 1986. Cacalia. Pg. 895 in G.P.F. 17:136-143. land Ecology and Management from Okla- Association, ed., Flora of the Great Plains. Dodd, M.B., W.K. Lauenroth, I.C. Burke, and homa State University. While at OSU, his University Press of Kansas, Lawrence. P.L. Chapman. 2002. Associations between research examined the interaction of fire vegetation patterns and soil texture in the Barnes, P.W., and A.T. Harrison. 1982. Species shortgrass steppe. Plant Ecology 158:127- and large grazers (both cattle and bison) distribution and community organization in 137. in Artemisia shrublands and tallgrass a Nebraska Sandhills prairie influenced by prairies of Oklahoma as well as tallgrass plant/soil water relationships. Oecologia Engle, D.M., and T.G. Bidwell. 2001. The prairies of Nebraska. He is currently a 52:192-201. response of central North American prairies to seasonal fire. Journal of Range Manage- Wildlife Biologist with the U.S. Fish and Benning, T.L., and T.R. Seastedt. 1995. ment 54:2-10. Wildlife Service. Landscape-level interactions between Fahnestock, J.T., and A.K. Knapp. 1993. Water topoedaphic features and nitrogen limita- relations and growth of tallgrass prairie forbs Karen Hickman is a Professor in the tion in tallgrass prairie. Landscape Ecology in response to selective grass herbivory by Department of Natural Resource Ecology 10:337-348. bison. International Journal of Plant Sciences and Management at Oklahoma State Uni- Beyer, H.L. 2004. Hawth’s Analysis Tools for 154:432-440. versity. Her research focuses on popula- ArcGIS. Available online
336 Natural Areas Journal Volume 33 (3), 2013 Hartnett, and S.L. Collins, eds., Grassland Hickman, K.R., and J.D. Derner. 2007. Black- Leach, M.K., R.A. Henderson, and T.J. Givnish. Dynamics: Long-Term Ecological Research land tallgrass prairie vegetation dynamics 1999. A caution against grazing. BioScience in Tallgrass Prairie. Oxford University Press, following cessation of herbicide applica- 49:599-600. New York. tion. Rangeland Ecology & Management Leidolf, A., and S. McDaniel. 1998. A floristic Fuhlendorf, S.D., and D.M. Engle. 2001. Restor- 60:186-190. study of Black Prairie plant communities at ing heterogeneity on rangelands: ecosystem Hickman, K.R., and D.C. Hartnett. 2002. Effects Sixteen Section Prairie, Oktibbeha County, management based on evolutionary grazing patterns. BioScience 51:625-632. of grazing intensity on growth, reproduction, Mississippi. Castanea 63:51-62. and abundance of three palatable forbs in Fuhlendorf, S.D., and D.M. Engle. 2004. Ap- Malin, J.C. 1942. An introduction to the history Kansas tallgrass prairie. Plant Ecology plication of the fire-grazing interaction to of the bluestem-pasture region of Kansas. 159:23-33. restore a shifting mosaic on tallgrass prairie. Kansas Historical Quarterly 11:3-28. Journal of Applied Ecology 41:604-614. Howe, H.F. 1999. Dominance, diversity and McGranahan, D.A., D.M. Engle, S.D. Fuhlen- Fuhlendorf, S.D., D.M. Engle, J. Kerby, and grazing in tallgrass restoration. Ecological dorf, S.L. Winter, J.R. Miller, and D.M. R. Hamilton. 2009. Pyric herbivory: re- Restoration 17:59-66. Debinski. 2012. Spatial heterogeneity wilding landscapes through the recoupling Howe, H.F. 2011. Fire season and prairie forb across five rangelands managed with pyric of fire and grazing. Conservation Biology richness in a 21-year experiment. Ecosci- herbivory. Journal of Applied Ecology doi: 23:588-598. ence 18:317-328. 10.1111/j.1365-2664.2012.02168.x Fuhlendorf, S.D., W.C. Harrell, D.M. Engle, Hoy, J. 1989. Controlled pasture burning in Palmer, M.W. 2007. The vascular flora of the R.G. Hamilton, C.A. Davis, and D.M. Leslie, the folklife of the Kansas Flint Hills. Great Tallgrass Prairie Preserve, Osage County, Jr. 2006. Should heterogeneity be the basis for conservation? Grassland bird response Plains Quarterly 9:231-238. Oklahoma. Castanea 72:235-246. to fire and grazing. Ecological Applications Kariuki, M., M. Rolfe, R.G.B. Smith, J.K. Van- Polley, H.W., B.J. Wilsey, and J.D. Derner. 16:1706-1716. clay, and R.M. Kooyman. 2006. Diameter 2007. Dominant species constrain effects Gesch, D.B. 2007. The National Elevation growth performance varies with species of species diversity on temporal variability Dataset. Pp. 99–118 in D. Maune, ed., functional-group and habitat characteristics in biomass production of tallgrass prairie. Digital Elevation Model Technologies and in subtropical rainforests. Forest Ecology Oikos 116:2044-2052. Applications: The DEM Users Manual, 2nd and Management 225:1-14. Rowe, H.I. 2010. Tricks of the trade: techniques ed. American Society for Photogrammetry Knapp, A.K. 1984. Effect of fire in tallgrass and opinions from 38 experts in tallgrass and Remote Sensing, Bethesda, Md. prairie on seed production of Vernonia prairie restoration. Restoration Ecology Guisan, A., S.B. Weiss, and A.D. Weiss. 1999. baldwinii Torr. (Compositae). Southwestern 18:253-262. GLM versus CCA spatial modeling of Naturalist 29:242-243. plant species distribution. Plant Ecology SAS Institute. 2007. SAS version 9.2. SAS 143:107-122. Knapp, A.K., J.M. Briggs, J.M. Blair, and Institute, Cary, N.C. C.L. Turner. 1998. Patterns and controls Hamilton, R.G. 2007. Restoring heterogeneity Schuler, K.L., D.M. Leslie, Jr., J.H. Shaw, and of aboveground net primary production in on the Tallgrass Prairie Preserve: applying E.J. Maichak. 2006. Temporal-spatial dis- tallgrass prairie. Pp. 193–221 in A.K. Knapp, the fire-grazing interaction model. Pp. 163– tribution of American bison (Bison bison) J.M. Briggs, D.C. Hartnett, and S.L. Col- 169 in Proceedings of the 23rd Tall Timbers in a tallgrass prairie fire mosaic. Journal of Fire Ecology Conference: Fire in Grassland lins, eds., Grassland Dynamics: Long-Term Mammalogy 87:539-544. and Shrubland Ecosystems. Tall Timbers Ecological Research in Tallgrass Prairie. Research Station, Tallahassee, Fla. Oxford University Press, New York. Smeins, F.E., and D.D. Diamond. 1983. Rem- nant grasslands of the Fayette Prairie, Texas. Harrington, J. 1998. Grazing prairies. Restora- Knapp, A.K., J.T. Fahnestock, S.P. Hamburg, American Midland Naturalist 110:1-13. tion and Management Notes 16:5-6. L.B. Statland, T.R. Seastedt, and D.S. Hartnett, D.C. 1990. Size-dependent allocation Schimel. 1993. Landscape patterns in soil- Smith, W.K., A.K. Knapp, T.A. Pearson, J.H. to sexual and vegetative reproduction in four water relations and primary production in Varmon, J.B. Yavitt, and D.R. Young. 1983. clonal composites. Oecologia 84:254-259. tallgrass prairie. Ecology 74:549-560. Influence of microclimate and growth form on plant temperatures of early spring spe- Hartnett, D.C. 1991. Effects of fire in tallgrass Knapp, A.K., D.C. Hartnett, S.L. Collins, and cies in a high-elevation prairie. American prairie on growth and reproduction of prairie J.M. Briggs. 1999. A caution against graz- coneflower. American Journal of Botany ing: response from Knapp et al. BioScience Midland Naturalist 109:380-389. 78:429-435. 49:600-601. Towne, E.G., and K.E. Kemp. 2008. Long-term Helzer, C.J. 2010. The ecology and manage- Kollmorgen, W.M., and D.S. Simonett. 1965. response patterns of tallgrass prairie to fre- ment of prairies in the Central United States. Grazing operations in the Flint Hills-blue- quent summer burning. Rangeland Ecology University of Iowa Press, Iowa City. stem pastures of Chase County, Kansas. & Management 61:509-520. Helzer, C.J., and A.A. Steuter. 2005. Preliminary Annals of the Association of American [USDA-NRCS] U.S. Department of Agri- effects of patch-burn grazing on a high- Geographers 55:260-290. culture, Natural Resources Conservation diversity prairie restoration. Ecological Kucera, C.L., and M. Koelling. 1964. The Service. 2009. Web Soil Survey. USDA Restoration 23:167-171. influence of fire on composition of central Natural Resources Conservation Service. Henderson, R.A. 1998. To graze or not? Restora- Missouri prairie. American Midland Natural- Available online
Volume 33 (3), 2013 Natural Areas Journal 337 Vermeire, L.T., R.B. Mitchell, S.D. Fuhlendorf, Winter, S.L., S.D. Fuhlendorf, C.L. Goad, Winter, S.L., S.D. Fuhlendorf, C.L. Goad, C.A. and R.L. Gillen. 2004. Patch burning effects C.A. Davis, and K.R. Hickman. 2011. Davis, K.R. Hickman, and D.M. Leslie. on grazing distribution. Journal of Range Topoedaphic variability and patch burning Management 57:248-252. 2012. Restoration of the fire-grazing interac- in sand sagebrush shrubland. Rangeland Williams, A.H. 1997. In praise of grazing. tion in Artemisia filifolia shrubland. Journal Restoration and Management Notes 15:116- Ecology & Management 64:633-640. of Applied Ecology 49:242-250. 118.
338 Natural Areas Journal Volume 33 (3), 2013