Use of Glyphosate and Imazapyr for Cogongrass (Imperata cylindrica) Management in Southern Pine Forests

Abstract

targeted; thus, systemic herbicides offer one of the most effective tools in an integrated treatment approach. Cogongrass leaves grow directly from buds along rhizomes, without aboveground stems, and may reach 1.5 m in height (Holm et al. 1977, Bryson and Carter 1993). Because shoot meristems are below ground, cogongrass is tolerant of mowing and grazing. Cogongrass is fire tolerant, even though fires are intense during both the growing season and the winter, when the dry thatch remains standing (Dozier et al. 1998). Cogongrass rhizomes have buds at each node that are spaced 1–2 cm apart along the entire length of the rhizome. Rhizomes branch frequently, forming dense mats that can exclude most other vegetation (Ayeni 1985). Rhizome entanglements can fill the upper soil to more than 30 cm deep, they typically make up more than 60% of the total plant biomass, and they are allelopathic to some grass seeds (Sajise 1976, Koger and Bryson 2003). Rapid aboveground regrowth from the rhizomes is stimulated by mowing, disking, burning, or ineffective herbicide treatment (Sajise 1976, Willard et al. 1996, Lippincott 2000). Therefore, rhizomes must be completely killed, leaving no living segments, to achieve eradication. Difficulty in achieving eradication is thought to increase with infestation age as the rhizome mat density and depth increase, although this has not been fully tested. The influence of the soil seed bank on eradication is not a primary concern during local eradication efforts. Prolific numbers of wind-dispersed seeds are produced by cogongrass, but viability is highly variable and declines rapidly after 1 year (Shilling et al. 1997, Dozier et al. 1998). Therefore, seed longevity in the soil seed bank is not a primary concern with regrowth after treatment, whereas spread to nearby mineral soil is a concern. Repeated applications of herbicides, commonly applied in summer to early fall, are required for cogongrass control. Prescribed burning in late winter or early spring preceding treatment is common (Miller 2007b). However, the benefits of burning to eliminate winter thatch have been questioned. Although this approach may allow for more effective herbicide applications early in the spring, burning has been reported to stimulate rhizome initiation and branching, increasing plant density, flowering, and outward spread rates while killing constraining shrubs (Bryson and Carter 1993, Lippincott 2000, Yager et al. 2010). Burning can also produce bare areas favorable for cogongrass seed germination (Yager 2007). Research in Florida and Asia has identified glyphosate, imazapyr, and the combination of these herbicides as the most effective herbicides for controlling cogongrass, although to date, eradication has not been reported even with retreatments (Brook 1989, Willard et al. 1996, 1997). Both glyphosate and imazapyr are readily absorbed by the foliage and translocated to rhizomes (Townson and Butler 1990). Soil-active imazapyr is also absorbed through roots (Little and Shaner 1991), which may foster greater control, although the importance of imazapyr uptake though rhizomes has not been determined. When used alone, imazapyr has proven more effective than glyphosate (Willard et al. 1996, 1997, Ramsey et al. 2003). Willard et al. (1997) found combinations of glyphosate and imazapyr in various proportions equally effective compared with the highest rates tested for either glyphosate or imazapyr applied alone (3.4 kg acid equivalent (ae)/ha glyphosate or 1.1 kg ae/ha imazapyr). Shoot biomass was reduced only 70% and rhizome biomass only 39% in their study. Additional research in Florida has shown that the highest rates tested for glyphosate (up to 9 kg ae/ha) or imazapyr (up to 1.1 kg ae/ha) applied in early winter were the most effective (Shilling et al. 1997, Willard et al. 1997, Ramsey et al. 2003). However, complete control was not achieved, indicating the need for research using higher rates of these herbicides combined with repeated applications. Application timing and carrier volume can also influence herbicide efficacy. November or December applications of glyphosate or imazapyr proved most effective for cogongrass control in Florida (Shilling et al. 1997). However, cogongrass is typically dormant by this time of the year when growing further north, indicating the need for testing earlier application dates for use in areas with a shorter growing season. Few formal studies address the impact of application volume, although imazapyr was more effective at 234 L/ha than 47 L/ha, whereas glyphosate was not responsive to changes in application volume (Willard et al. 1997). Furthermore, information regarding the optimum application volume for use of glyphosate and imazapyr in combination is lacking. Our overall goal was to further refine recommendations for control of cogongrass infestations in states north of Florida. Specific objectives of our research were to (1) test a range of rates of glyphosate or imazapyr applied alone, including higher rates than previously tested, (2) refine late summer and early fall timing for the Gulf Coastal Plain by testing both September and October applications, (3) determine the optimum application volume for the commonly used combination of glyphosate and imazapyr, and (4) test the efficacy of retreating plots with the same treatment 1 year after initial application. Materials and Methods Study Areas Two experiments were installed in 1996 at locations near Bay Minette in Baldwin County, Alabama. These sites are located in the Middle Coastal Plain physiographic province, which contains most of the cogongrass infestations in the region. Bay Minette is approximately 100 km northeast of the first introduction of cogongrass into the southeastern United States in about 1911 (Dickens 1974). The area has a temperate climate, with an average high temperature of 25°C, average low temperature of 13°C, and average annual precipitation of 168 cm. Study locations were upland sites having slopes less than 3%. The initial cogongrass cover (96–100%) was similar between the two study sites, whereas age, dry biomass, and average foliage height differed between the sites. At one study location (30°43.732 N, 87°51.475 W), cogongrass completely covered the understory of a sparse 35-year-old slash pine (Pinus elliottii Engelm.) plantation. According to the landowner, the infestation was more than 20 years old (referred to as the old infestation). The site had historically been burned every other year and was broadcast burned in February 1996, approximately 6 months prior to our first herbicide applications, and again in February 1997, before retreatments. All understory pine seedlings and most shrubs within the study area were killed by these burns. Cogongrass height ranged from 0.3 to 1.3 m in September before the first treatment. Cogongrass foliar dry biomass, as determined by clipping nine randomly located 0.5-m plots and oven drying at 70°C for 72 hours, averaged 3,170 kg/ha (standard error, 139 kg/ha). Soils at this site are classified as Faceville fine sandy loam, very deep, well drained, moderately permeable, Thermic Typic Kandiudults (USDA Natural Resources Conservation Service 2010). The second experiment was established within relatively new cogongrass infestations (referred to as the new infestation) that developed after harvest and site preparation in two young loblolly pine 20 SOUTH. J. APPL. FOR. 36(1) 2012 (Pinus taeda L.) plantations approximately 31 km apart. Multiple circular infestations of various sizes occurred across these plantations. To accommodate each experimental block within a single infestation, one block was established in one plantation (30°32.537 N, 87°39.331 W), whereas the other two blocks were established in two infestations located within the second plantation (30°49.180 N, 87°41.944 W). Both plantations had been site prepared by chopping, followed by broadcast burning before planting, and they were 1 and 2 years old, respectively. Cogongrass height was considerably less than in the old infestation and ranged from 0.15 to 0.3 m. The cogongrass aboveground biomass, determined as described above with three 0.5-m plots per infestation, was 2,861 kg/ha (standard error, 307 kg/ha). Common associated shrub species, sparsely scattered across all blocks, were gallberry (Ilex glabra [L.] A. Gray.), yaupon (Ilex vomitoria Aiton), and waxmyrtle (Morella cerifera [L.] Small). Herbaceous ground cover was largely displaced by cogongrass. Soils within these two plantations are classified as Lakeland loamy fine sand, very deep, excessively drained, Thermic, coated Typic Quartzipsamments (USDA Natural Resources Conservation Service 2010). Treatments and Experimental Design Separate replicated studies were conducted at the new and old infestations. In both studies, the same treatments were assigned in a randomized complete block, split plot design with three replications. Within the new plantations, separate circular infestations were the blocks. Plots measured 6.1 12.2 m and were split lengthwise for testing retreatment. Plots were laid out contiguously within each block to hinder edge reinvasion during the study. An untreated check and 22 herbicide treatments were included in each block. Treatments were structured to provide three separate factorial arrangements of glyphosate [1] rate with application timing, imazapyr [2] rate with application timing, and tank mix spray volume with application timing. Individual glyphosate and imazapyr rates tested are expressed relative to the typical use rate ( ) for cogongrass management. Glyphosate alone was applied at 1.68 (0.5 ), 3.36 (1 ), 6.72 (2 ), and 13.44 kg ae/ha (4 ) in 93.5 L/ha water carrier. Applications of glyphosate with relatively low spray volumes such as this have been shown to improve efficacy on cogongrass (Arif et al. 1986) and other grasses (Ramsdale et al. 2003). Imazapyr alone was applied at 0.275 (0.5 ), 0.55 (1 ), 1