Southeast Exotic Pest Plant Council

REFORESTATION TECHNIQUES IN COGONGRASS (Imperata cylindrica) INFESTED AREAS. Wilson H. Faircloth1, James H. Miller2, Michael G. Patterson3, and David H. Teem3. 1USDA/Agricultural Research Service, Dawson, GA; 2USDA/Forest Service, Auburn, AL; 3Auburn University, Auburn, AL. (


Cogongrass [Imperata cylindrica (L.) Beauv.] is a non-native, invasive weed that is disrupting economically and ecologically important plant communities in the southeastern United States. Cogongrass is invading forest lands, especially those growing loblolly pine (Pinus teada L.) plantations. These silvicultural systems are prone to cogongrass establishment and growth due to clear-cutting and site preparation treatments that expose soil on large, contiguous areas in a short period of time. Such disturbance may create a mineral seed-bed that favors cogongrass seedling establishment and movement of machinery between infested and un-infested areas may introduce vegetative propagules. Once established, cogongrass promotes fire and is assumed to reduce loblolly pine growth, though no published studies quantify this reduction. Cogongrass also affects secondary uses of plantations such as the leasing of land for hunting and other types of recreation. This could occur as the overall biological diversity of stands is reduced to a few species, including the grass and the pines, eliminating plant species needed by wildlife for food and shelter. Cogongrass has the potential to create vast monocultures in the Southeast, in which rehabilitation and reforestation will by necessity become increasingly intensified and expensive. In addition, property devaluation will become a serious problem where cogongrass is involved due to the high costs of rehabilitation and absence of productive land-use. The primary research objective of this project was to investigate options for the establishment or reforestation of loblolly pine into cogongrass-infested areas. Factorial combinations of herbicide and mechanical site preparation as well as first-year herbicide release treatments were tested. Cogongrass response and loblolly pine survival and production were the measured responses.

The study site was located in Mobile County, near the town of Theodore, AL. Soil at the study area was a Bama fine sandy loam with high site index (90) for loblolly pine. The experiment was a factorial arrangement that tested two herbicide site preparation (SP) treatments, two mechanical SP treatments, and two pine release herbicide treatments. Herbicide SP levels were none and a broadcast-applied tank mixture of glyphosate at 3.3 kg ae/ha, imazapyr at 0.34 kg ae/ha, and nonionic surfactant at 0.5 % v/v. Application was on October 3-4, 2001. Mechanical SP levels were a scalping treatment and none. Scalping consisted of using a bulldozer and fire plow to remove the upper 10 to 15 cm of cogongrass rhizomes and roots, folding these back upon intervening grass to create a furrow in which seedlings were planted. Scalping was performed on December 19, 2001. Release treatment levels were band-applied herbicide and none and were applied after seedling planting. Visual inspection of the test area in the spring of 2002 (3 mo. after planting) indicated that plant species composition, including cogongrass, was greatly influenced by SP methods. Therefore, plots were subdivided for release treatment in order to investigate timing of release application, tank-mix partners for imazapyr, and release applications of varying cost. In addition to the eight plots in the factorial core, a ninth treatment, termed "complete control", was added. Since a non-treated or negative control was part of the factorial arrangement, the intent of "complete control" was to provide a positive control to compare the responses from the factorial experiment. Because of persistent cogongrass regrowth, "complete control" was not actually achieved and risk of tree damage did not permit further treatments, however, a greater level of cogongrass suppression was added to the study by this treatment. The nine treatments were replicated four times in a randomized complete block design. Treatment plots were 100 trees planted on 2.5 x 2.5 m grid, of which the inner 36 comprised the measurement area. Bare-root, improved loblolly pine seedlings were hand-planted on January 15, 2002. Plant biomass was harvested and partitioned into live cogongrass, cogongrass thatch, and other woody and herbaceous species during the first three growing seasons (2002-2004). Trees were measured (height, ground-line diameter, diameter breast height) yearly from planting through 2005. Plant biomass response and tree response from the first and second years are presented. The subdivided plots mentioned previously were analyzed collectively for this reporting to examine the overall effect of release compared to both mechanical and herbicide SP.

Biomass response.

In the year immediately following planting of trees, mechanical SP reduced biomass in each category. However, in the following year, regrowth occurred to an extent that mechanical SP made no treatment differences. The combination treatment of mechanical and herbicide SP reduced live grass by 98.5% compared to the non-treated check in the first year, which suggests greater than an additive effect when combined compared to a single use. Live cogongrass regrew by 8 to 18-fold by the second year on all SP treatments, and all treatments contained live grass at the 2003 harvest (780-7400 kg/ha). As expected, mechanical SP significantly reduced thatch around seedlings compared to herbicide SP; the combination of SP methods was additive and resulted in near complete removal of thatch. Thatch accumulation ranged from 190 to 6410 kg/ha in year two. Other woody and herbaceous plant biomass was least in the non-treated (830 kg/ha) and greatest in the complete control (8310 kg/ha), indicating the participation of other plant species. Herbicide SP and complete control yielded the greatest recruitment of other woody and herbaceous species in both years to assist with the rehabilitation process (8310 kg/ha in year two). Herbicide SP was significantly greater than both the mechanical SP and the non-treated, while mechanical SP yielded woody and herbaceous biomass equivalent to the non-treated. Data analysis revealed few significant differences between treatments or establishment methods for live grass and other species at year two. Despite intensive control efforts, cogongrass remained a significant component of the plant community as live cogongrass comprised at least 15% of the total plant biomass through year two. Herbicide SP resulted in at least 30% other species, increasing the overall plant community diversity in those treatments. The combination treatment of herbicide SP, mechanical SP, and release yielded the least total plant biomass, but yielded also the largest portion of other species (> 80%).

Tree response.

Loblolly pine survival was equal to or exceeded 90% on all treatments including the non-treated check in years one and two after planting. This is widely considered an acceptable level of survival to establish a fully stocked pine plantation. Herbicide SP increased survival greater than 2% in year one, while the interaction of mechanical SP and release significantly decreased survival. On average, mechanical SP plus release and complete control treatments were equivalent to the non-treated and less than the remaining treatments for both years. Survival in the second year was most influenced by SP, which significantly increased survival 4% compared with no SP. Treatments with herbicide SP had greater survival, either with or without mechanical SP or release. Survival appeared to be related to levels of cogongrass biomass. Site preparation, whether herbicide or mechanical, yielded a significant increase in ground-line diameter (GLD) compared with no SP, while the addition of release made no difference in GLD in the first year. On average, herbicide SP and release showed an additive effect for GLD response. Tree height in year one ranged from 46.1 to 66 cm. Herbicide SP yielded 6 cm taller trees than mechanical SP unless combined with mechanical SP, in which case an antagonistic effect was found. The negative interaction of mechanical SP and herbicide SP was significant for both loblolly pine GLD and HT. One possible explanation for this response is that soil, previously treated with SP herbicides, sloughed into the mechanically-created furrows, thus concentrating herbicide around the seedlings. Stunting of the trees resulted along with a visual yellowing of needles during the first growing season. This effect was evident in both treatments that received herbicide SP followed by mechanical SP. Effects were transient, however, and not observed during the second growing season. Site preparation significantly increased tree growth in the second year by all measures compared to no SP. Aside from the non-treated, the release only treatment consistently yielded the smallest trees. For HT only, herbicide SP yielded taller trees than did mechanical SP. These data suggest that two years after application, herbicide SP is positively influencing loblolly pine growth more than the other establishment techniques. Release alone was not as effective as other establishment techniques as trees were generally smaller than SP.


Herbicide SP consistently increased loblolly pine growth, decreased live cogongrass, and increased overall plant community diversity. Release alone was not an effective reforestation technique, however, was generally additive when combined with a SP treatment. The use of herbicides was critical to the recruitment of woody and herbaceous species other than cogongrass. No reforestation technique reduced cogongrass to acceptable levels, however, the establishment of loblolly pine was successful, thus, some productivity restored to the land.

University of Georgia The Bugwood Network Forestry Images   The Bugwood Network - The University of Georgia
College of Agricultural and Environmental Sciences and Warnell School of Forest Resources
Copyright 2002. All rights reserved.       Page last modified: Monday, April 29, 2002
Questions and/or comments to: