Control

Management and Control

Once established, knotweed populations can be difficult to impossible to eradicate (Forman and Kesseli 2003). Whatever methods are chosen to control knotweed, persistence over several years or more is required to eliminate even the smallest patches. Even aggressive treatment is likely to fail if not repeated for several years.

The best way to prevent the spread of invasive knotweeds is to avoid planting them. Two horticultural varieties of Japanese knotweed, variegated knotweed and dwarf knotweed, are often advertised as being "non-invasive" but are quite capable of crossing with "standard" knotweeds and producing viable seeds (Forman and Kesseli 2003). Both forms also spread vigorously via rhizomes, and can establish new patches vegetatively.

Manual Control

Manual control methods can be used to weaken patches, in preparation for other methods. If applied diligently over several years, these methods can even eliminate small patches (Soll 2004). The goal of mechanical control methods is to starve and weaken the roots. Cutting can also prevent any seed set that might otherwise occur.

It is best to destroy the rhizomes and shoots left on the site after cutting or digging, as these plant parts can reroot and form new shoots. On sites where live plant parts are not likely to be spread by human or natural forces, drying and/or burning the litter is a good strategy. Turning the plant parts with a rake will hasten the drying process. Extreme care should be taken to keep plant parts out of floodplains or other high-risk habitats.

Pulling or digging: Digging can be effective for small patches, especially in soft soil or sand (Soll 2004). The patch should be checked weekly and new shoots redug. Exposed rhizomes should be dried out or otherwise destroyed. The Japanese Knotweed Manual (Child and Wade 2000, in Soll 2004) reports control of a small patch, after uprooting plants in August for 3 consecutive years. Digging is very labor-intensive, though, and tends to breakup and spread the rhizome fragments (Seiger 1991). Damaging the rhizomes also causes new buds to develop.

Cutting or mowing: Cutting or mowing patches can be an effective means of control, but only if repeated diligently over several growing seasons (Soll 2004). An experiment by Nature Conservancy personnel showed some control over a small patch with 17 monthly cuttings spread out over 3 seasons (Soll 2004). In general, though, patches must be cut at least twice a month throughout the season (or April through August at a minimum) to be effective (Soll 2004). Shoots should be cut as close to the ground as possible, and should never be allowed to exceed 15 cm (6 in) in height (Soll 2004). At least three cuts are needed through the growing season simply to offset rhizome production (Seiger 1991). Any lapses in treatment will allow the knotweed patch to quickly recover.

Covering: Soll (2004) reports several attempts at controlling Japanese knotweed by covering patches with black plastic or several layers of cardboard. All these attempts were unsuccessful, including a US Forest Service attempt where the patch was extensively dug before covering. Part of the problem with covering the patches is that knotweed often grows through or around the cover. The cover should be applied at the beginning of the growing season and stay in place well into the second season to have any hope of killing the rhizomes (Soll 2004).

Chemical Control

Seiger (1991) states that "At present, the only method to control large stands is with repeated application of herbicides". As always, follow the label directions when using herbicides.

A variety of herbicides have proven useful in controlling knotweed, including glyphosate (Rodeo, Aquamaster, Gly Star, Round-up among others), triclopyr (Garlon 3a and many “shrub-killers”), 2,4-D, and Imazapyr (Arsenal) (Soll 2004). Each herbicide offers benefits and potential risks. The Nature Conservancy has had success with mixes of both 3-4% (or about 1 oz/gallon) tryclopyr (Garlon), and 3-5 % glyphosate (Rodeo). (Illustrating the fact that more is not necessarily better, 5% Garlon kills the shoots but does a poor job of killing the root system, probably because the shoots die before they can transport enough herbicide to the roots to kill them.) In the Nature Conservancy's trials on small patches, 2-4 applications of Garlon over two years eliminated about half the patches, while Rodeo took 3 years for similar control.

While the best time to spray knotweeds is when they are in flower, it is probably more practical to spray them earlier, when they are still only 1-2 m tall (Soll 2004). Taller plants can be cut down to 1.5 m or so tall before spraying. Spraying very early in the season (April or May) is less effective.

Though more labor intensive than simply spraying the foliage, injecting the hollow stems with herbicide has proven quite successful (Soll 2004). Trials by The Nature Conservancy and the Clark County Weed Management Agency of Washington State (under an experimental permit) have found that injecting 5 ml of 100% Aquamaster or Round Up Pro into each stem of a given clump gave 100% control (Soll 2004). This method also has the advantage of virtually eliminating unwanted herbicide drift. At this time, unfortunately, stem injection of undiluted glyphosate is currently allowed at only 1 ml per 2 inches of stem diameter, a level which is generally ineffective (Soll 2004). Further testing will likely soon lead to approval of these higher dose levels for injection treatments.

A related method that also works well is to cut the stems about 2 inches above the ground (between the lowest nodes), and carefully pour about 5 ml of undiluted herbicide into the stem cavity (Soll 2004). The EPA has recently approved this method for the herbicide Aquamaster, and approval for Rodeo and other herbicides is likely to follow. Please check the label to make sure your herbicide has been approved for use in this way.

Whenever herbicides are used, care should be taken to avoid spraying surrounding (and competing) vegetation and to minimize human contact. Any attempt to control knotweeds in wet habitats must be done using herbicides formulated for use over water. Permits are required for herbicide application over water in many states, including Wisconsin, Michigan and Minnesota. Consult your local resource agency for advice prior to treating knotweed infestations with herbicide.

Cultural Control

These methods may suppress knotweed to some degree but, except for reforestation, generally will NOT eliminate the stands.

Fire: Fire appears to be ineffective in controlling Japanese (Beerling et al. 1994) and (presumably) giant and Bohemian knotweeds.

Grazing: Goats will feed on knotweed shoots, and intensive grazing may suppress the stands. It generally will not eliminate knotweed, though (Soll 2004).

Reforestation: Knotweeds are not shade-tolerant (Beerling 1994). Therefore regeneration of forest should suppress and eventually eliminate patches of knotweed. Large, dense knotweed patches will greatly suppress tree seedling establishment and growth, though and have been reported to persist within and maintain openings inside regenerating forest in the northeast US for decades.

Biological Control

When knotweeds were imported from the far east, their natural pests and pathogens were left behind, giving them an advantage over the native plants. In North America, stands of knotweed rarely exhibit significant damage from insects and pathogens. Within its native range, however, a wide variety of insects and fungi attack Japanese knotweed, often defoliating the plants by the end of the growing season (Shaw and Seiger 2002). Knotweed plants grown in the laboratory have even been killed by their natural pests (JKA 1999).

Scientists from the Centre for Agriculture and Bioscience International (CABI) carried out an initial survey for knotweed pests in Japan in August 2000 (CABI 2005). With the addition of several funding partners, including Cornell University in New York, research into the biology of insects and disease organisms infesting Japanese knotweed began in May 2003. Three insects - a psyllid (or a "jumping plant louse"), a weevil and a chrysomelid beetle - have passed preliminary screening and are being studied further. A rust fungus and a leafspot organism have also shown promise. The goal is to find and release insects and pathogens that are host-specific and that will effectively control these knotweeds outside their native range.

For More Information

TNC Wildland Weed Management and Research Library