655 D.A. McGranahan, G.M. Raicovich, W.N. Wilson, and C.K. Smith 22001133 SOUSToHuEthAeSaTstEeRrnN N NaAtuTrUaRlisAtLIST 12V(o3l). :1625,5 N–6o6. 03 Preliminary Evidence that Patch Burn-Grazing Creates Spatially Heterogeneous Habitat Structure in Old-field Grassland Devan Allen McGranahan1,*, Gina M. Raicovich2, W. Nathan Wilson3, and C. Kenneth Smith4 Abstract - Heterogeneity created through patchy disturbance is an important component of grassland conservation, but little research has focused on patch burn-grazing in grassland of the eastern United States. To test the viability of patch burn-grazing in a Tall Fescue-invaded old-field grassland on the Cumberland Plateau, we conducted a prescribed patch burn, stocked cattle, and measured vegetation structure and plant functional group composition. We found that patch burn-grazing creates spatial heterogeneity (patch contrast) for grass height, litter cover, bare ground, and canopy cover of native and exotic grasses. These results suggest that patch burn-grazing is a viable tool for heterogeneitybased grassland management in the region. Introduction Heterogeneity is an important component of conservation, and in grassland primarily results from spatially patchy disturbance (Fuhlendorf and Engle 2001). In many rangeland ecosystems, fire and grazing interact to create a unique disturbance—pyric-herbivory (Fuhlendorf et al. 2009)—which in turn creates heterogeneity in habitat structure along with wildlife and plant diversity (Derner et al. 2009). Relatively little research has contributed to an understanding of the fire-grazing interaction in eastern North America. In the southeastern United States, specifically, many grasslands are characterized by small tract sizes and are frequently invaded by Schedonorus phoenix (Scop.) Holub (Tall Fescue), an Eurasian coolseason grass that occurs across 14 million ha in the region (Fribourg et al. 1991). As an invader in native warm-season grassland and an abundant species in oldfields, Tall Fescue has been implicated in reduced fire spread and associ ated with decreased native plant species richness (McGranahan et al. 2012a, 2013a). We established a pilot project to determine whether the fire-grazing interaction will create spatial heterogeneity in a pasture stand common in the southeastern United States. We implemented a patch burn-grazing management scheme in old-field grassland dominated by Tall Fescue and Andropogon virginicus L. (Broomsedge Bluestem) and compared habitat variables related to vegetation structure (grass height, bare ground, and litter cover) in addition to plant 1Mellon Environmental Fellow, Department of Environmental Studies, The University of the South, Sewanee, TN 37375.2Farm Manager, The University of the South, Sewanee, TN 37375. 3Domain Manager, The University of the South, Sewanee, TN 37375. 4University Forester and Professor, Department of Forestry and Geology, The University of the South, Sewanee, TN 37375. *Corresponding author - mcgranah@alumni.grinnell.edu. D.A. McGranahan, G.M. Raicovich, W.N. Wilson, and C.K. Smith 2013 Southeastern Naturalist Vol. 12, No. 3 656 functional group composition. We predicted that bare ground, native grass cover, and forb cover would be greater in the burned patch than the unburned area, while grass height, litter, and Tall Fescue cover would be lower in the burned patch. We discuss our results with respect to fuel load and fire-behavior characteristics. Field-site Description and Methods Our inquiry was conducted at the University of the South in Sewanee, TN, located on the Cumberland Plateau. We selected a 16-ha old-field not treated with grazing, fire, mowing, soil amendments or herbicide for >10 yrs aside from periodic tree removal to maintain the grassland state. The majority of the grassland vegetation was dominated by Tall Fescue and Broomsedge Bluestem, with less than 10% canopy cover of upland hardwood species (Acer rubrum L. [Red Maple], Liquidambar styraciflua L. [Sweetgum], and Quercus spp. [oak]) and Juniperus virginiana L. (Eastern Red Cedar) in addition to small patches of blackberry (Rubus spp.). We conducted a prescribed patch burn on 3 ha in mid-March 2012. Prior to the burn, we measured fuel moisture and fuel load for both live and dead fuel components using the constituent differential method (Gillen and Tate 1993). Fuel moisture (dry-weight basis) and fuel-load data were combined with weather data from the day of the fire to estimate flame length and rate of spread with BehavePlus fire-behavior modeling software (Andrews et al. 2008). Twenty 227–318-kg steers were stocked in mid-April 2012. Cattle had free access to water in stock ponds without interior fencing. Beyond basic mineral lick, no feed supplements were provided during the period of this study. Total rainfall for the study period—mid-March–late May—was 8.7 cm. To determine the effect of patch burn-grazing on spatial heterogeneity, we measured habitat structure variables (grass height, bare ground, and litter cover) and plant functional group composition (canopy cover of forbs, native grasses, and exotic grasses) in late May 2012. We randomly located 25-m transects in the burned patch (n = 3) and across the unburned areas (n = 5) along which we measured the height of two randomly selected grass tillers at 1-m intervals. For bare ground, litter, and plant functional composition, we randomly located five 0.25- m2 quadrats along the transect within 5 m of either side and visually estimated percent cover using the Daubenmire (1959) cover-class index. While wildlife response to patch burn-grazing was not assessed in this study, vegetation variables reported here are associated with habitat for grassland fauna (McGranahan et al. 2013b), and sampling was timed to fall within the grassland bird nesting season in the Southern US (Conover et al. 2011). To inform whether observed responses were due to the combined effect of fire and grazing or due to fire alone, we compared aboveground biomass from three grazer exclosures erected in the burned patch to aboveground biomass from three points in the burned area accessible to grazers. All aboveground plant biomass was collected from 0.25-m2 quadrats and dried for 48 hrs at 45 °C. The lack of replicate pastures in our pilot study precluded the use of analysis of variance in data analysis. Instead, we calculated effect sizes using Cohen’s d (Cohen 1977) and associated 95% confidence intervals for each measured 657 D.A. McGranahan, G.M. Raicovich, W.N. Wilson, and C.K. Smith 2013 Southeastern Naturalist Vol. 12, No. 3 variable. Cohen’s d compares the mean and standard deviation of two groups to determine a treatment effect, which is indicated when effect sizes are non-zero and 95% confidence intervals do not overlap zero. Although we expected patch burn-grazing to decrease some values in the burned patch relative to unburned areas (e.g., grass height, litter cover) but increase others (bare ground), we report the absolute value of all effect sizes to focus on the hypothesis that patch burngrazing creates spatial heterogeneity (contrast among patches) across burned and unburned areas (sensu McGranahan et al. 2013b). Results and Discussion At the time of our prescribed fire, mean total herbaceous fuel load was 4448 kg/ha (1.8 t/ac). Live fuel load, comprised primarily of Tall Fescue, averaged 346 kg/ha (0.14 t/ac) and dead fuel load, comprised of both Broomsedge Bluestem and Tall Fescue, averaged 4102 kg/ha (1.66 t/ac). Field-measured fuel moistures averaged ≈300% and 17% for live and dead fuels, respectively. Eye-level wind speed ranged 3–8 km/hr, and relative humidity was approximately 48%. BehavePlus predicted flame lengths of 0.5–1.1 m and rates of spread 1.4–7.9 m/min across the grass-dominated portions of our relatively flat fuelbed, which corresponded to typical fire behavior observed during the burn. The spatial pattern of fire drove the spatial pattern of grazing in this oldfield pasture, concordant with patch burn-grazing theory (Fuhlendorf and Engle 2004, Fuhlendorf et al. 2009, McGranahan et al. 2012b). Nine weeks after the prescribed burn, mean aboveground plant biomass within burned-but-ungrazed exclosures was an order of magnitude greater than that of the surrounding grazed burned patch (756 kg/ha ± 307 SE vs. 56 kg/ha ± 2 SE). Furthermore, grazing frequency and cattle occupancy time, measured by proportion of grass tillers grazed and fecal pat density, respectively, were both greater in the burned patch than unburned areas (J.S. Hill et al., The University of the South, Sewanee, TN, unpubl. data). Taken together, these data indicate that cattle removed substantial herbage through grazing that was spatially determined by the patch burn. Mean values varied between the burned patch and unburned areas for most response variables with the exception of forb canopy cover (Table 1). Patch burn-grazing created spatial heterogeneity between the burned patch and unburned areas for five of the six measured variables: grass height, litter cover, Table 1. Mean (± standard error) values for six habitat variables across the burned and unburned patches in a single patch burn-grazing demonstration pasture in an old-field grassland on the Cumberland Plateau, Sewanee, TN. Values for grass height given in cm, all other values in percent canopy cover following Daubenmire (1959). Response variable Patch Grass height Litter cover Bare ground Forbs Native Grasses Exotic grasses Burned 14 ± 2 18 ± 4 41 ± 7 16 ± 3 34 ± 6 28 ± 6 Unburned 29 ± 3 57 ± 4 1 ± 1 18 ± 4 56 ± 6 44 ± 6 D.A. McGranahan, G.M. Raicovich, W.N. Wilson, and C.K. Smith 2013 Southeastern Naturalist Vol. 12, No. 3 658 bare ground, native grass canopy cover, and exotic grass canopy cover (Fig. 1). Although the effect size for forb canopy cover was positive, the 95% confidence interval included zero. These results suggest patch burn-grazing can create heterogeneity among patches within old-field grassland on the Cumberland Plateau. Combined with our data and observations on cattle response to the spatial pattern of fire, the response of patch burn-grazing on the reported variables (Fig. 1) suggest that patch burn-grazing creates a disturbance unique from fire or grazing alone (Fuhlendorf et al. 2009). As an example of the conservation value of these effects, reduced vegetation cover has been associated with Grasshopper Sparrow nest survival in Tall Fescue-invaded eastern grassland managed with fire and grazing (Hovick et al. 2012). Despite low wind speed and the high live-fuel moisture content of the Tall Fescue, which decreases fire spread in invaded grassland (McGranahan et al. 2013a), our prescribed fire achieved sufficient spread and intensity to consume litter, create bare ground, and reduce grass cover in the herbaceous-dominated Figure 1. Effect sizes and associated 95% confidence intervals demonstrate the response of six variables to patch burn-grazing in a single patch burn-grazing demonstration pasture in an old-field grassland on the Cumberland Plateau, Sewanee, TN. Effect size, measured as Cohen’s d (McGranahan et al. 2012b), compares the mean and standard deviation of three transects in the burned patch to five transects in the unburned area. Absolute values of effect sizes are reported to focus on the effect of patch-burn grazing on creating spatial heterogeneity (patch contrast) in each variable (McGranahan et al. 2013b). For actual values and direction (increase or decrease) in each value across burned and unburned areas of the pasture, see Table 1. 659 D.A. McGranahan, G.M. Raicovich, W.N. Wilson, and C.K. Smith 2013 Southeastern Naturalist Vol. 12, No. 3 areas of the burned patch. Patch contrast in the variables studied here should be easily achieved in similar old-field grassland fuel beds with at least 4448 kg/ ha (1.8 t/ac) total fuel load under typical prescribed-fire conditions, although decreased total fuel load, increased live-fuel load, and/or increased live-fuel moisture might require greater wind speed and/or decreased relative humidity for sufficient fire spread (McGranahan et al. 2013a). While long-term vegetation responses to patch burn-grazing in this ecosystem remain unknown, heterogeneity in plant functional group abundance at the tract scale is likely important to maintain forage production through temporal cycles in plant productivity (Fynn 2012). Additionally, other research has shown lower abundance of invasive species under patch burn-grazing compared to grazing with homogeneous fire regimes (Cummings et al. 2007). Continued research should monitor changes in relative abundance of exotic (i.e., Tall Fescue) versus native (i.e., Broomsedge Bluestem) grasses. Ecologically sound grazing systems—low-input systems that rely on perennial vegetation and enhance environmental outcomes (Franzluebbers et al. 2012)—have the potential to contribute to biodiversity conservation and ecosystem protection in the eastern United States. Both fire and grazing have a long tradition in the region, with the best economic returns shown under moderate stocking rates (Grelen 1978, Pearson and Whitaker 1973). Fire has been shown to increase forage quality of native Andropogon spp. stands similar to the oldfield pasture community described here (Grelen and Whitaker 1973, Lewis et al. 1982). The results of this pilot project indicate that patch burn-grazing might represent a viable tool to manage diverse, perennial grassland for both biodiversity conservation and livestock production on the Cumberland Plateau. We recommend replicated research on the vegetation and wildlife ecology of patch burn-grazing using moderate stocking rates in grassland in the eastern US. Acknowledgments We appreciate the support of the Treasurer and Office of Domain Management of The University of the South, the Sewanee Environmental Institute, and the contributions of D. Carter, S. Gilliam, R. Petropoulos, J. Hill, C. Henderson, and the 2012 Sewanee Summer Farm Team. Literature Cited Andrews, P.L., C. Bevins, and R. Seli. 2008. BehavePlus fire modeling system. US Forest Service, United States Department of Agriculture, Rocky Mountain Research Station, Ogden, UT. Cohen, J. 1977. Statistical Power Analysis for the Behavioral Sciences. Academic Press, New York, NY. Conover, R.R., S.J. Dinsmore, and L. Burger, Jr. 2011. Effects of conservation practices on bird nest density and survival in intensive agriculture. Agriculture, Ecosystems, and Environment 141:126–132. Cummings, D.C., S.D. Fuhlendorf, and D.M. Engle. 2007. Is altering grazing selectivity of invasive forage species with patch burning more effective than herbicide treatments? Rangeland Ecology and Management 60:253–260. D.A. McGranahan, G.M. Raicovich, W.N. Wilson, and C.K. Smith 2013 Southeastern Naturalist Vol. 12, No. 3 660 Daubenmire, R. 1959. A canopy-coverage method of vegetational analysis. Northwest Science 33:43–64. Derner, J.D., W.K. Lauenroth, P. Stapp, and D.J. Augustine. 2009. Livestock as ecosystem engineers for grassland bird habitat in the western Great Plains of North America. Rangeland Ecology and Management 62:111–118. Franzluebbers, A.J., L.K. Paine, J.R. Winsten, M. Krome, M.A. Sanderson, K. Ogles, and D. Thompson. 2012. Well-managed grazing systems: A forgotten hero of conservation. Journal of Soil and Water Conservation 67:100A–104A. Fribourg, H.A., C.S. Hoveland, and K.D. Gwinn. 1991. Tall fescue and the fungal endophyte: A review of current knowledge. Tennessee Farm and Home Science 160:30–37. Fuhlendorf, S.D., and D.M. Engle. 2001. Restoring heterogeneity on rangelands: Ecosystem management based on evolutionary grazing patterns. BioScience 51:625–632. Fuhlendorf, S.D., and D.M. Engle. 2004. Application of the fire-grazing interaction to restore a shifting mosaic on tallgrass prairie. Journal of Applied Ecology 41:604–614. Fuhlendorf, S.D., D.M. Engle, J. Kerby, and R. Hamilton. 2009. Pyric herbivory: Rewilding landscapes through the recoupling of fire and grazing. Conservation Biology 23:588–598. Fynn, R.W.S. 2012. Functional resource heterogeneity increases livestock and rangeland productivity. Rangeland Ecology and Management 65:319–329. Gillen, R.L., and K.W. Tate. 1993. The constituent differential method for determining live and dead herbage. Journal of Range Management 46:142–147. Grelen, H.E. 1978. Forest grazing in the South. Journal of Range Management 31:244–250. Grelen, H.E., and L.B. Whitaker. 1973. Prescribed burning rotations on a pine-bluestem range. Journal of Range Management 26:152–153. Hovick, T.J., J.R. Miller, S.J. Dinsmore, D.M. Engle, D.M. Debinski, and S.D. Fuhlendorf. 2012. Effects of fire and grazing on Grasshopper Sparrow nest survival. The Journal of Wildlife Management 76:19–27. Lewis, C. E., H. E. Grelen, and G. E. Probasco. 1982. Prescribed burning in southern forest and rangeland improves forage and its use. Southern Journal of Applied Forestry 6:19–25. McGranahan, D.A., D.M. Engle, B.J. Wilsey, S.D. Fuhlendorf, J.R. Miller, and D.M. Debinski. 2012a. Grazing and an invasive grass confound spatial pattern of exotic and native grassland plant species richness. Basic and Applied Ecology 13:654–662. McGranahan, D.A., D.M. Engle, S.D. Fuhlendorf, S.J. Winter, J.R. Miller, and D.M. Debinski. 2012b. Spatial heterogeneity across five rangelands managed with pyricherbivory. Journal of Applied Ecology 49:903–910. McGranahan, D.A., D.M. Engle, J.R. Miller, and D.M. Debinski. 2013a. An invasive grass increases live fuel proportion and reduces fire spread in a simulated grassland. Ecosystems 16:158–169. McGranahan, D.A., D.M. Engle, S.D. Fuhlendorf, S.L. Winter, J.R. Miller, and D.M. Debinski. 2013b. Inconsistent outcomes of heterogeneity-based management underscore importance of matching evaluation to conservation objectives. Environmental Science and Policy 31:53–60. Pearson, H.A., and L.B. Whitaker. 1973. Returns from southern forest grazing. Journal of Range Management 26:85–87.