Burning and Grazing to Promote Persistence of Warm-Season Grasses Sown into a Cool-Season Pasture

Abstract

We combined burning and rotational grazing in an effort to promote persistence of recently established native grasses. The experiment took place on a farm in south-central Wisconsin on a cool-season grass pasture that was drill seeded with native warm-season grasses: big bluestem (Andropogon gerardii), Indiangrass (Sorghastrum nutans), and switchgrass (Panicum virgatum). We used a split-plot experimental design to assess native grass persistence under varying disturbance treatments (burned, burned-grazed, and grazed). We used a paired t-test to determine if the difference between 2006 and 2007 native grass density was significantly different from zero. Native grass tiller density increased under the burned (202%) and grazed (186%) treatments, but not the burned-grazed (29%) treatment. However, the actual native grass tiller numbers in 2007 were much higher in the burned-only than the grazed-only treatment (80 ± 10 tillers/m² and 2 ± 1 tillers/m², respectively). We found no loss to native grass tiller density when rotational grazing was applied to plots in the first year after two years of grazing exclusion with burning. In addition, we found that native grass cover was greatest in the burned treatment but not significantly different in the burned-grazed and grazed treatments. Our results suggest that the combined use of burning and grazing as a management tool for native grass persistence in pastures may be possible with deferred grazing during the establishment phase, but alternative timing, intensity, and types of grazing animals should be tested.

Introduction

The effects of burning on the tall-grass prairie have been widely studied (Vinton et al. 1993, Fuhlendorf and Engle 2004), but very little is known about burning and grazing cool-season pasture to promote reintroduced tallgrass prairie species. Prior to European settlement, prairie ecosystems dominated much of the upper Midwest (Transeau 1935). In Wisconsin, about 850,000 ha of prairie existed at the time of European settlement (Curtis 1959); less than one percent remains (ca. 8,000 ha). While there are numerous relatively large- and small-scale efforts to restore native prairie in this region, the total area in native vegetation remains very small. For significant landscape-level benefits to accrue, native-plant restoration will need to be integrated with agricultural production. As restoration and conservation efforts in managed grasslands grow (Jackson 1999, Tracy et al. 2004, Doll et al. 2009), it is infeasible to convert most pastures to fully restored native communities because of their high quantity and quality of biomass for livestock feed in early spring and fall months when dominated by introduced, cool-season forage grasses and legumes (Casler et al. 1998).

Ecological restoration often means "starting over" with respect to the plant community, but Williams and colleagues (2007) argue that it may be beneficial to keep the established vegetation and focus on enhancing species composition. This approach falls under the rubric of managing "novel ecosystems" (Hobbs et al. 2006, Seastedt et al. 2008) and meets the Society for Ecological Restoration's definition of restoration as "intentional activity that initiates or accelerates the recovery of an ecosystem with respect to its health, integrity, and sustainability" (SERI 2004). Moreover, warm-season grasses possess the C₄ photosynthetic pathway, affording them greater water-use efficiency (Lambers et al. 1998, Anderson et al. 2000) and therefore less water stress during dry periods. The coexistence of non-native, cool-season and native, warm-season grass species may improve the distribution of forage production in pasture systems, especially during the hot summer months when cool-season grass production is low (Anderson et al. 2000).

Burning is considered a necessary tool in the establishment and restoration of the tallgrass prairie to warm the soil and reduce light-suppressing plant litter accumulating on the soil surface (Howe 1995, Howe 2000, Knapp and Seastedt 1986). Fire influences both above- and belowground net primary productivity via reduced nitrogen availability and increased root production (Johnson and Matchett 2001) and can also promote plant community heterogeneity (Vermeire et al. 2004). While spring burning unequivocally favors warm-season grasses, cattle may be useful as a management tool in prairie restoration because of their selective grazing (Helzer and Steuter 2005). In many restored prairies, the dominance of C4 grasses favored by burning is undesirable, and the selectivity of grazing may promote coexistence of shorter-stature native forbs. That said, Jackson and colleagues (Forthcoming) showed that high-intensity grazing (< 10 cm residual stubble height) in summer months resulted in precipitous warm-season grass decline.

Past research on pastures has shown that burning alone promoted greater native, warm-season grass establishment and root production than rotational grazing, while grazing alone promoted C₃ grasses (Woodis 2008). Here, we addressed the following question: Will native, warm-season grasses that have recruited into cool-season pastures increase or decrease under burned, burned-grazed, or grazed treatments? Our goal was to successfully combine a conservation tool (burning) with an agricultural management tool (rotational grazing) in order to promote native plant persistence.

Methods

This study occurred at the Cates Family Farm in Spring Green, Wisconsin, during the 2007 growing season. Soils from the study site are classified as Fayette silt loam (fine-silty, mixed, superactive, mesic Typic Hapludalfs) and have a 12% to 20% slope (USDA-NRCS 2008). The study site is in a temperate climate with average annual precipitation of about 900 mm and average temperatures of -6°C in January 2007 and 20°C in July 2007. During the 2007 growing season, the site received about 1,400 mm of precipitation and daily average temperatures were -4°C in January 2007 and 21°C in July 2007.

Figure 1.

Experimental design showing one of three blocks where grazing and burning were first randomly applied, then grazing was applied to half of the burned area in Spring Green, Wisconsin.

The farm produces about 50 head of cattle of the Jersey and Angus breeds. The cattle are grazed using a management-intensive rotational grazing system on pastures dominated by non-native, cool-season grasses such as meadow fescue (Lolium pratense), smooth brome (Bromus inermis), orchardgrass (Dactylis glomerata), and quackgrass (Elymus repens). Dominant forbs are white clover (Trifolium repens), red clover (Trifolium pratense), and common dandelion (Taraxacum officinale). Native warm-season grass seed was collected from several restored prairies in southern Wisconsin in fall 2002 and stored outdoors over winter in a sealed container. In fall 2003 and spring 2004, warm-season grasses were drill seeded into the existing cool-season sod of a roughly 2 ha pasture at a rate of about one gram pure live seed per square meter with a mix by weight of 70% big bluestem (Andropogon gerardii), 15% switchgrass (Panicum virgatum), and 15% Indiangrass (Sorghastrum nutans) (Woodis 2008). The seeded pasture was then used for the experimental treatments described below.

An establishment-phase experiment was conducted from 2004 to 2006 comparing burning and grazing separately. In the experiment discussed here, we reintroduced grazing to one-half of the area of burned plots in 2007 but continued spring burning. The burned treatment was the control in our experiment because we were interested in testing burning and grazing together to promote native grasses of the tallgrass prairie. Three disturbance treatments—burned, burned-grazed, and grazed—were applied during 2007 in a complete randomized block design using three 0.42 ha pastures as experimental blocks (Figure 1). Grazed only and burned only treatments were randomly chosen within each block, whereas burned-grazed treatment was not. Burning occurred in April of 2005, 2006, and 2007 and removed 100% of the aboveground biomass. Management-intensive rotational grazing systems use short-duration (< 12 h to 3 d) grazing with high stocking densities and two- to five-week rest periods (Paine et al. 1999) resulting in 5 to 10 cm residual stubble height. The experiment area has been under this type of grazing for more than 15 years and grazing treatments followed this system. About 20 steers were rotated through our three graze-only treatments for about two days on a monthly basis. When applying grazing to the smaller burn-graze experimental units, the same number of animals was rotated through but for only one day. At the request of the farmer, 2007 fall grazing ceased in early October to accumulate biomass for a spring burn the following year.

Figure 2.

Mean (± SE, n = 6 for 2006 and n = 3 for 2007) native grass tiller density for burned, burned-grazed, and grazed treatments for 2006 and 2007. In 2007, grazing began on half of the 2006 burned plots to include a new treatment: burned-grazed. Different lowercase letters over each treatment indicate that differences between 2006 and 2007 were significantly different among treatments as determined with Bonferroni multiple comparison test (p < 0.05).

Native grass tiller density was estimated in late August of 2006 and 2007. Using quadrats 1 m², we counted native grass tillers every five meters along a 50 m permanent transect in each experimental unit. We then calculated the average number of native grass tillers per plot to estimate native grass density. In addition to native grass density, we estimated native grass cover. Species composition was assessed in August 2007 with the line-point method (Heady et al. 1959). Each plot contained a 45 m transect, and species were recorded every ten meters using a 20 cm × 50 cm quadrat. Native grass cover was then calculated per experimental unit.

Soil temperature, soil moisture, and nitrogen availability were ancillary data. Soil temperature was recorded monthly on the burned plots and before and after grazing events (June through November 2007) by inserting a soil thermometer into the ground until a constant reading was observed. Gravimetric water content and net nitrogen mineralization were measured in June, August, and October of 2007 by taking 12 soil cores (1.91 cm diameter, 10 cm deep) per experimental unit. A portion of the soil samples was used to find gravimetric water content (g water/g dry soil) by recording the soil weight before and after drying at 65°C for 48 hours. Another portion of the soil samples was used to find net nitrogen mineralization, based on seven-day aerobic lab incubation (Rhine et al. 1998).

Figure 3.

Mean (± SE, n = 3) percent cover of native grass species for burned, burned-grazed, and grazed treatments during August 2007. Means with different lowercase letters were significantly different as determined with Bonferroni multiple comparison test (p < 0.05).

For each treatment combination, we used a paired t-test to determine if the difference between 2006 and 2007 native grass tiller density was significantly different from zero. We then used split-plot ANOVA to evaluate disturbance treatment effects on these differences in native grass tiller density between 2006 and 2007 and on native grass cover for 2007. Significance was determined at the α = 0.05 level for all tests. Bonferroni's multiple comparison tests were used to separate treatment means after ANOVA. Ancillary data were assessed graphically, since we were using them to interpret native grass results rather than to make statistical inferences about these parameters.

Results

Native grass tiller density increased significantly from 2006 to 2007 under the burned-only (p = 0.006) and grazed-only (p = 0.005) treatments, but not the burned-grazed treatment (p = 0.18). While two of three treatments were determined to be statistically different from zero using the paired t-test, it is important to note the disparity in absolute native grass tiller density. Native grasses in burned plots increased from 19.8 ± 6.5 to 59.6 ± 7.3 tillers/m², while grazed plots had only 1.0 ± 0.24 to 2.8 ± 0.11 tillers/m².

When comparing the differences in native grass tiller density between 2006 and 2007 using split-plot ANOVA to assess whether tillers were expanding over time, we found that density in the burned-only treatment was significantly greater than in the burned-grazed and grazed-only treatments, while there was no significant difference between burned-grazed and grazed treatments (Figure 2).

Native grass cover in the burned treatment was significantly higher than in the burned-grazed and grazed treatments (Figure 3). There was no significant difference in native grass cover between burned-grazed and grazed treatments.

Visual assessment of ancillary data indicated little variability among treatments in soil temperature or gravimetric water content for 2007. However, net nitrogen mineralization appeared to be lower in the burned-only plots than the grazed-only plots for June and October of 2007 (Figure 4).

Discussion

Many studies have examined the effects of burning (Collins et al. 1998) and grazing (Seastedt et al. 1994, Knapp et al. 1999, Fuhlendorf and Engle 2001, Fuhlendorf and Engle 2004) on native grass establishment in western tallgrass prairie; however, there has been less research on the use of burning and grazing to encourage native grass persistence in the eastern tallgrass prairie. The work described here focused on the first year of combined burning and grazing after a three-year establishment phase under burning alone and grazing alone. The lesson from the establishment phase was that burning greatly facilitated native grass recruitment into an existing stand of cool-season grasses (Woodis 2008). Grazing during the establishment phase did not preclude recruitment, but neither did it allow native grasses to recruit into the stand in an agronomically significant way.

In high-productivity grasslands, many have shown that fire enhances native grass establishment and abundance and reduces non-native species (Howe 2000, Camill et al. 2004, Prober et al. 2005, Klopfenstein et al. 2007). Camill and others (2004) found a significant shift to warm-season grasses in prairie ecosystems around year 3 of restoration, and Howe (2000) found that spring burning promoted five to six times greater warm-season grass production than fall burning and no burning treatments. Burning increases soil temperature and reduces soil moisture and nutrient availability, which allows for greater dominance of native, warm-season grasses (Collins et al. 1998). However, we observed no consistent effects of management on soil temperature or moisture for 2007. We did observe that inorganic nitrogen concentrations were somewhat lower in the burned treatment during June and October, which may have given the native grasses in the burned plots a competitive advantage.

Figure 4.

Mean (± SE, n = 3) net N mineralization for burned, burned-grazed, and grazed treatments during June, August, and October 2007.

In the western tallgrass prairie, grazing regimes increase species diversity (Collins et al. 1998, Knapp et al. 1999) and promote patch diversity (Fuhlendorf and Engle 2001). However, one study examined bison grazing on native, warm-season grasses east of the Mississippi and found that grazing caused a decline in native grass persistence (Jackson et al. forthcoming). Similarly, we found that native grass establishment was limited under the grazed treatment. While we found a significant increase in native grass density from 2006 to 2007, the establishment was limited under grazing alone (an increase of roughly 1 to 3 native grass tillers/m² under graze only). Studies have demonstrated that less intense or less frequent grazing promotes persistence of big bluestem (Mousel et al. 2003, Mousel et al. 2005). The apical meristem of cool-season grasses is closer to the ground than for most native, warm-season grasses, so grazing native grasses too closely to the soil surface could suppress regrowth after defoliation.

Studies have found that combining grazing and burning increases species diversity and heterogeneity (Collins and Gibson 1990), and enhances wildlife habitat (Fuhlendorf and Engle 2001). We found that a combination of burning and grazing in the first year resulted in little to no change in native grass density, suggesting that once native grasses are established, rotational grazing regimes may be applied without immediate detriment to native grass persistence. However, it is also important to note that by combining grazing and burning we did not observe the significant expansion of native grasses that was evident in the burned-only plots. Moreover, these results must be viewed with caution as longer-term rotational grazing may reduce native grass persistence. Jackson and others (Forthcoming) documented a precipitous decline of warm-season grasses (though rates of decline were species-specific) under summer rotational bison grazing. While stocking rates were quite high, and the timing of grazing may not have been ideal, native grass cover on their plots decline from around 90% to about 30% over a six-year period. Hence more work with other types, intensities, and timings of grazing livestock is needed.

When burning grazed pastures, farmers will need to take into consideration the amount of accumulated biomass remaining from the previous season. Farmers may need to reduce or eliminate grazing in the fall to allow biomass to accumulate for the following spring burn. One alternative is to incorporate "rest paddocks" (Jackson 1999, Hickman et al. 2004, Firn 2007) for one to two years to facilitate burning and then reintroduce rotational grazing, but with follow-up monitoring that can inform adaptive management.

While combining burning and grazing did not result in native grass loss, there was significantly less native grass cover in these treatments compared to burning alone. This shows how different vegetation metrics might lead to alternate conclusions. Density is calculated as the mean number of individuals per unit area, while species cover reflects how much ground area is covered by the vertical projection of the plant (Packard and Mutel 1997). Native grasses were present in significant numbers (density), but occupying much less space (cover) under the burned-grazed treatments. Under both the burned and burned-grazed treatments, native grasses were present, but when grazing was included, native grasses were not apparent to the observer. The tallgrass prairie species were not allowed to fully and visually express themselves when grazing was applied. This shows the importance of sampling method selection and interpretation of results. Farmers and land managers will need to carefully quantify native grass presence when grazing is applied.

Conclusions

Our land management approach was to blend biodiversity and conservation efforts with agricultural and pasture production by reintroducing native, warm-season grasses into a cool-season grass pasture. The coexistence of native, warm-season and non-native, cool-season grasses offers many potential ecological, agronomic, and social benefits, including improved forage production, restored native species, carbon sequestration, wildlife habitat, and weed exclusion (Firn 2007). Introducing native grasses as part of the pasture community also offers flexibility to the management system (Moore et al. 2004), especially when production of cool-season grasses is low during hot summer months. One goal in the field of restoration ecology is to enhance conservation values in productive landscapes through integrating production and conservation values (Hobbs and Norton 1996). Combining burning and grazing as a management tool for native grass persistence in the eastern tallgrass prairie region may be difficult to implement initially, but is feasible with deferred grazing during the establishment phase. Further research into "rest paddocks" or rotating paddocks in and out of rotational grazing cycles to allow for spring burning and native grass recruitment is needed.