Removing the Invasive Tree Ailanthus Altissima and Restoring Natural Cover

HISTORY OF AILANTHUS ALTISSIMA

Ailanthus altissima (tree-of-heaven) is a native of China. For the past 250 years, Ailanthus has rapidly spread through many regions of the western world. Ailanthus was first introduced into England and France in 1751 by a missionary who mistook it for a Japan varnish tree (Feret 1985). Ailanthus was brought to North America from two different sources. In the east, gardeners first brought the seeds from England in the late 1700s and planted them as ornamentals (Feret 1985). The second route was through the West Coast by Chinese immigrants who probably planted the Ailanthus seeds because of its cultural significance and medicinal uses (Hoshovsky 1995). By the mid 1800s, Ailanthus was a common nursery species in the eastern United States (Hoshovsky 1995). It began on urban and rural home sites and spread along transportation rights-of-way. Today the species can be found invading grasslands and disturbed forests throughout North America.

METHODS AND MATERIALS

Test plots were established and treated in June 1997. The plots were located along Skyline Drive at the US 33 interchange and across from mile marker 101 in Shenandoah National Park. Troublesome Ailanthus altissima sprouts were located along roadsides in the park for the study. Nine treatments were applied: eight low-volume basal herbicide treatments plus a manual cutting (Table 1). Aqumix, Inc. (Cloverdale, VA), Eco-Pak LLC (Selma, IN), and Arborchem Products Co. (Mechanicsburg, PA) provided ready-to-use mixtures for this study. Four replicate plots were established for each treatment.

Each plot consisted of a 5 by 5 m square, containing a targeted minimum of ten established Ailanthus saplings or trees. The plots were delineated with tagged PVC stakes. A buffer was left between plots to avoid effects from surrounding plots. Herbaceous cover was visually estimated pre- and posttreatment over the entire plot for the three most dominant herbaceous species as well as total species presence or absence. Sequentially numbered tags identified each Ailanthus tree within the 5 m plot, and diameter at breast height (dbh) was measured for each tagged tree. Treatments were randomly assigned to each numbered plot for a randomized complete split-plot design.

Low-volume basal herbicide treatments were applied with a Solo^® backpack sprayer using a solid cone nozzle at approximately 30 psi. Herbicide mixtures were applied to the lower 30 to 45 cm of each stem. Solutions were applied to wet the bark but not to the point of surface runoff. Complete coverage constituted wrapping of the solution around all sides of the stem surface. Approximately 3 mL of mixture was applied per inch of stem diameter.

Manual cutting was made with brush saws approximately 3 to 6 in. above ground level at the time of herbicide application.

Six weeks after treatment, visual estimations of defoliation were made to quantify short-term treatment effects. One year after treatment, in September 1998, diameters were measured of remaining live trees to determine mortality and control. Dead trees and manually cut trees were examined for stump sprouting. The percentage of control was determined using the change in diameter of live trees from 1997 to 1998 for chemical treatments according to the following:

Dead trees were assigned 100% control. The percentage of control for manual cutting treatment was based on the number of stump sprouts emanating from the cut stump above groundline. Each stump sprout accounted for (–)100% control according to the following:

In January 2000, a tally was taken of the number of Ailanthus stems in each plot and converted to the number of stems per acre.

Native grass and herbaceous species seed mixtures were considered but not applied after determining that posttreatment effects included a shift from non-native species to those common in the Shenandoah region. Data were analyzed using ANOVA, and means were separated by LSD at the a = 0.05 significance level. Stem counts 2 years after treatment were analyzed by box plot because the treatment variance precluded use of ANOVA.

RESULTS AND DISCUSSION

Defoliation 6 weeks after treatment indicated variable effects for all treatments except manual cutting (Table 2). Defoliation ranged from 75% for the low rate (15% v/v) of Garlon 4 combined with a high Stalker rate (3% v/v), to 100% defoliation for Garlon 4 alone and in Aqumix RTU. All other treatments were intermediate in defoliation.

Herbicide treatments to Ailanthus altissima greatly exceeded control and mortality over that induced by manual cutting of stems. Manual cutting averaged 1.6 new stump sprouts per stem, while none of the chemical treatments produced any stump sprouts. Only 21% of cut stump trees failed to resprout, compared to 79% or greater mortality using chemical control methods.

Herbicide combinations that induced the highest mortality 1 year after treatment were Garlon 4 at 20% v/v, 20% Garlon 4 plus 5% Tordon K (in either the Aqumix RTU or Eco-Pak RTU carrier), and 9% Stalker in HY-Grade EC (Table 2). Similarly, 20% Garlon 4 plus 1% Stalker and 20% Garlon 4 plus 5% Tordon K (Arborchem RTU) approached 100% mortality. While Arborchem RTU and 20/1 Garlon/Stalker mixtures did not result in significantly lower mortality than the aforementioned treatments, persistent live trees could be the result of incomplete application rather than chemical ineffectiveness. Chemical combinations that did not control Ailanthus as well as other treatments were 15% Garlon 4 plus 3% Stalker and 20% Garlon 4 plus 5% Stalker.

Garlon 4 at 20% and 9% Stalker, when applied independently, successfully controlled Ailanthus. When combined, the control was slightly decreased. Statistically, there were no differences among any of the RTU mixes. Garlon or Stalker alone or in mixtures with low rates (1%) of Stalker, and any of the Garlon/Tordon RTUs, provided adequate control of Ailanthus over manual cutting.

Overall stem reduction was assessed 2 years after treatment. ANOVA was not used due to the lack of treatment variance homogeneity. Data are presented using box plot analysis. There was no difference among the RTU mixes, but the differences among the treatments containing picloram and the others were observed with the reduction in treatment variance being very evident. While the Garlon and Garlon plus Stalker treatments were similar, all treatments reduced the overall stem count when compared to cutting alone (Figure 1). Stalker alone at 9% provided intermediate control between the RTU mixes and the Garlon + Stalker mixes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Box plot of stem count (number of stems per acre) as affected by treatments. Untreated is hand-cut only.

Herbaceous species present on the plots prior to treatment consisted largely of garlic mustard (Alliaria petiolata) along with native herbaceous plants (Table 3). Either as a result of the removal of Ailanthus, or due to natural species shifts, garlic mustard, a non-native plant, was absent 1 year later and was replaced by predominantly native species. Dominant species present over all four blocks largely consisted of native wildflowers, including Joe Pye weeds (Eupatorium spp.), Dutchman’s pipe (Aristolochia durior), and jewelweed (Impatiens capensis), among others. Once Ailanthus was removed, reseeding with native herbaceous species was not necessary.

One interesting note regarding herbaceous species composition was the dominant vegetation surrounding the rock wall that lies along Skyline Drive near the US 33 interchange between the road and treatment plots. Prior to treatment (June 1997), roadside herbaceous vegetation was dominated by bouncing bet (Saponaria officinalis), introduced from Europe. At the time of post-treatment measurements (early September 1998), bouncing bet was either absent altogether or only a minor component of the roadside vegetation (Table 4). This finding suggests natural shifts in herbaceous vegetation independent of Ailanthus control.

SUMMARY AND CONCLUSION

Manual cutting of Ailanthus stimulated resprouting and increased overall stand density. Chemical control not only removed existing trees but also prevented resprouting by eliminating the existing root system. Optimal herbicide treatments for control of Ailanthus should include combination with 5% Tordon K. Garlon 4 at 20% v/v alone, Garlon 4 combined with Stalker, or Stalker alone controlled Ailanthus better than hand cutting but were not as effective as treatments containing picloram.

Removal of Ailanthus resulted in a shift in herbaceous species to native species of the region without having to reseed with naturally occurring herbs.

Manual control of Ailanthus should be avoided in order to prevent proliferation throughout the park. Herbicide control of Ailanthus would be the preferred method of control because it not only successfully kills the trees, but it also prevents resprouting. Because major Ailanthus infestations occur near roadways, access with a backpack sprayer should not be a major concern.