Centaurea diffusa
(diffuse knapweed)

Diffuse knapweed survives in a wide range of soil and environmental conditions, and produces large numbers of viable seeds that are widely dispersed by wind, water, animals and human activity. For much of the growth cycle, plants are hard to graze because they are protected by spines or form low-lying rosettes (Roche and Roche, 1999). It is a major rangeland weed and is considered a noxious or restricted weed in 13 US states and four Canadian provinces (Rice, 2003).

Biennial
Herbaceous
Perennial
Seed propagated

Diffuse knapweed is native to grasslands of western Asia and south-eastern Europe. The native range includes Armenia, Azerbaijan, Georgia, Bulgaria, Greece, Moldova, Romania, Russia, Ukraine and former Yugoslavia (USDA-ARS, 2003). It was introduced into Central Europe and North America, probably through contaminated lucerne shipments (Watson and Renney, 1974). It is a serious invader of lucerne crops in the Crimea (Harris and Cranston, 1979).

The risk of introduction of C. diffusa is high in contaminated lucerne seeds (Zouhar, 2001). It is a restricted noxious weed in 13 US states and four Canadian provinces (Rice, 2003).

Diffuse knapweed is tolerant of a wide range of conditions, but prefers arid or semi-arid habitats. Infestations stop abruptly with an increase in soil moisture near temporary and permanent streams or with flooding and irrigation (Berube and Myers, 1982). It invades disturbed or overgrazed rangeland and dry, open forest sites where it forms dense, often monotypic stands (Lacey et al., 1990). Other areas of invasion include road and railway right of ways, waterways, gravel pits and industrial areas (Roche and Roche, 1988).

Diffuse knapweed is a serious weed of lucerne in its native range (Harris and Cranston, 1979) and will sometimes invade orchards and vineyards. However, it is mainly a rangeland pest (Roche and Roche, 1999).

Diffuse knapweed is a herbaceous annual, biennial, or short-lived perennial that exploits disturbed arid or semi-arid areas. It spreads entirely by seeds, forming monotypic stands that dominate through successful competition for soil moisture and nutrients and through allelopathy (Callaway and Aschehoug, 2001, 2000).

Genetics

Diffuse knapweed is normally diploid (2n=18) but is occasionally tetraploid (2n=36) (Watson and Renney, 1974). It forms rare hybrids with spotted knapweed (Centaurea maculosa) and the hybrid is Centaurea x psammogena (Ochsmann, 2003). Populations have large genetic diversity due to obligate outcrossing and multiple introductions (Harrod and Taylor, 1995; Roche and Roche, 1999).

Physiology and Phenology

Diffuse knapweed reproduces by seed and is generally biennial. Seeds germinate in spring, very late summer or autumn, and over a wide range of environmental conditions. Germination requires more than 55% soil moisture, and seeds germinate best on the soil surface. Emergence rate decreases as seedling depth increases, with little or no emergence below 2.5 cm. More than 80% of seeds germinate at temperatures between 13 and 28°C and optimum moisture (Watson and Renney, 1974).

Physiologically polymorphic seeds give the plant maximum advantage. There are nondormant seeds that germinate in the dark, light-sensitive seeds that germinate after exposure to red light, and dormant seeds that do not respond to exposure to red light. Open canopy conditions provide favourable light conditions for germination, but seeds will germinate in closed canopies (Spears et al., 1980; Nolan and Upadhyaya, 1988).

Most seeds sprout in the first year, producing rosettes that bolt in the second or third year. Root development occurs during the rosette stage, and exposure to cold temperatures primes the plant for bolting. Plants completing juvenile growth by autumn overwinter as rosettes, and usually bolt in early May. Plants that do not complete the juvenile stage by the end of autumn remain as rosettes through the second year and may bolt during the third year. Under conditions of severe crowding, a plant may not flower for 5 or more years (Thompson and Stout, 1991; Zouhar, 2001).

The phenology is highly variable and dependent on the site. Moisture, temperature and plant density are all factors in development. Flower buds are formed in early June. Flowering occurs from July through September or later as permitted by adequate moisture and mild temperatures. Mature seeds are usually formed by mid-August, followed by the death of the plant. Dead plants break off at ground level in the spring and tumble with the wind, spreading seed as they roll (Watson and Renney, 1974; Thompson and Stout, 1991).

Reproductive Biology

Diffuse knapweed has male and female flowers on the same plant (monoecious) and is pollinated by insects, mostly honey bees and bumble bees. Some experiments have shown that the plant is self-compatible, but consensus is that fertilization requires cross-pollination (Watson and Renney, 1974; Harrod and Taylor, 1995).

Monotypic stands of up to 500 plants/m² can produce a massive seed rain of 11,000 to 40,000 seeds/m². However, many seeds remain on the plant until the taproot breaks and a tumbleweed is formed. Plants and seeds are dispersed by wind, water and by vehicles over long distances (Watson and Renney, 1974; Roche and Roche, 1999).

Environmental Requirements

Diffuse knapweed is commonly found on well-drained soils such as sandy or gravelly loams or loamy fine sands containing coarse fragments. It is less competitive on shallow, very coarse textured soils such as sand or loamy coarse sand, although it may thrive on these sites when disturbance removes other vegetation (Zouhar, 2001).

It grows in soil with a pH range of 5.9 to 8.4. In Canada, most infestations are found where mean annual temperatures are between 7.2 and 9.4°C. It grows well where the annual rainfall is between 240 and 420 mm (Watson and Renney, 1974). However, it has been found at locations with mean annual temperatures ranging from 3.6 to 13.8°C (Harris and Cranston 1979; Strang et al., 1979; Roche and Roche, 1999; Climate, 2003). Diffuse knapweed is not competitive in moist microsites such as gullies, depressions or poorly drained soils. It does not grow well in saline soils (Zouhar, 2001).

In British Columbia, Canada, it is commonly found on south-facing slopes below 900 m. In Washington, USA, it is found below 1500 m in 150-900 mm of annual rainfall (Roche and Roche, 1999).

Associations

A number of grasses and herbaceous plants are found with diffuse knapweed. These include bluebunch wheatgrass (Agropyron spicatum [Pseudoroegneria spicata]), western yarrow (Achillea lanulosa [A. millefolium]), Idaho fescue (Festuca idahoensis), sheep sorrel (Rumex acetosella) and downy brome (Bromus tectorum) (Watson and Renney, 1974).

It may also be dominant in bitterbrush (Purshia tridentata), habitat types with or without an overstorey of ponderosa pine (Pinus ponderosa) (Roche and Roche, 1999). In the southern interior of British Columbia, Canada, it may be found with Douglas fir (Pseudotsuga menziesii), lodgepole pine (Pinus contorta), Engelmann spruce (Picea engelmannii), subalpine fir (Abies lasiocarpa) and understorey components such as pinegrass (Calamagrostis rubescens) and fireweed (Epilobium angustifolium) (Hubbard, 1975; Berube and Myers, 1982; Myers and Berube, 1983).

Extensive studies on the natural enemies were made in Europe prior to introductions into North America. A general account of this work was published by Schroeder (1986). There are also papers published on the biology and host specificity of some of the introduced species (see Julien and Griffiths, 1998; Bourchier et al., 2002).

Due to the variable regulations around (de)registration of pesticides, your national list of registered pesticides or relevant authority should be consulted to determine which products are legally allowed for use in your country when considering chemical control. Pesticides should always be used in a lawful manner, consistent with the product's label.

Introduction

Integrated control is most effective against diffuse knapweed. Central to the strategy is the prevention of seed formation, which can be accomplished by grazing, mowing, hand-pulling and biological control. Any control method must be integrated with a programme of revegetation to prevent re-infestation. As knapweeds thrive in direct sunlight, they can be suppressed by shading from competing vegetation (Kennett et al., 1992; Woo et al., 2002).

Cultural Control

Grazing by sheep and goats can suppress knapweeds (Olson and Wallander, 2001). Continual grazing of the tops of young plants can retard plant development, seed formation, and gradually deplete root reserves. As animals usually prefer to eat nearby grasses instead of knapweeds, grazing is most effective against knapweeds when the livestock is enclosed in a fenced-off, weedy area (Beck, 1994; Olson and Lacey, 1994).

Diffuse knapweed should be grazed during the bolting stage for 10 days, and again after 14 days for an additional 10 days to reduce seed production. Although grazing can reduce seed production, it can also cause diffuse knapweed to become a short-lived perennial and when grazing is removed, populations often return to their former levels (Popay and Field, 1996).

In knapweed-infested rangeland, any patches of bare ground can lead to knapweed invasions (Lacey et al., 1990). Thus, reseeding of rangelands with competitive perennial grasses is important. Crested wheatgrass (Agropyron cristatum) has been a successful competitor against knapweeds in dry areas of Canada, reducing the rate of vegetative spread, limiting weed density and reducing weed seed production (Berube and Myers, 1982). Crested wheatgrass is more competitive than Russian wildrye (Elymus junceus [Psathyrostachys juncea]) or lucerne (Maxwell et al., 1992) but orchardgrass (Dactylis glomerata) and thickspike wheatgrass (Agropyron dasystachyum) are more competitive than crested wheatgrass (Larson and McInnis, 1989). Meadow fescue (Festuca pratensis) and rough fescue (F. altaica) are more competitive than bluebunch wheatgrass (Pseudoroegneria spicata) (Lacey et al., 1990; Steinger and Muller-Schärer, 1992).

Manipulation of plant ecology is important for the long-term suppression of knapweed. The kinds and numbers of plants at a site change slowly over time. Initial colonizers, usually seedy, weedy and fast-growing, are replaced successively with other plants, until a final set of plants known as the climax vegetation predominates. Soil analyses have shown that the perennial grasses that dominate climax populations on grazing lands are associated with low available nitrogen. Early colonizers such as knapweed do well in nitrogen-rich soil, but cannot compete with perennial grasses in a low nitrogen environment (Sheley et al., 1996a, b; Herron et al., 2001).

When the soil is rich in nitrogen, aggressive plants such as knapweed, which appear early in the succession, will out-compete perennial grasses. Controlled colonization can be used to force plant succession towards the preferred climax vegetation. This strategy is to plant simultaneously intermediate and late colonizers in the natural plant succession scheme. For instance, Herron et al. (2001) found that plantings of annual rye and bottlebrush squirreltail (Elymus elymoides) can deplete the soil of nutrients, especially nitrogen. As a result, late seral plants such as bluebunch wheatgrass (Pseudoroegneria spicata) will out-compete the early colonizer knapweed. Low nitrogen availability also makes knapweed more susceptible to root-feeding biological control agents (Steinger and Muller-Schärer, 1992).

A more short-term approach is to fertilize or combine fertilization with herbicides. Nitrogen fertilization has been used with mixed results. When fertilizer is applied to areas with a substantial grass understorey, grass yields increase on those sites (Sheley and Jacobs, 1997). Where knapweed has already invaded, nitrogen fertilization with no other treatment may favour knapweed over native perennial grasses (Prather and Callihan, 1991; Velagala et al., 1997; Sheley et al., 1998; Herron et al., 2001).

Effects of fertilization may be complicated by moisture conditions at the site. Fertilization in dry areas may remove moisture from the soil, leading to increased competition of drought-tolerant grasses such as crested wheatgrass (Berube and Myers, 1982). However, Myers and Berube (1983) found that the addition of nitrogen had no effect on biomass of either diffuse knapweed or competing grasses in dry areas of British Columbia, Canada.

Mechanical Control

Control of diffuse knapweed by hand-pulling is feasible for scattered diffuse knapweed plants. It is important to remove the entire taproot with as little soil disturbance as possible. Plants should be pulled in the spring and when the soil is moist, and again in June to remove bolting plants before they flower and set seed. Finally, plants are pulled just before seed dispersal, taking care to remove the plants from the site. Hand-pulling is tedious and not always effective (Roche and Roche, 1999).

Mowing decreases flower and seed production, and in the long-term it may affect knapweed densities. It can also reduce weed competition during a revegetation programme. Mowing diffuse knapweed in British Columbia, Canada, at the bud stage and again at flowering reduced the number of plants producing seed by 77-99%. Mowing treatments also reduced seed germination by 79%. Plants mowed early in the growing season produce few viable seeds; however, mowed plants usually resprout and flower again. In some instances, diffuse knapweed densities increased after a single mowing (Watson and Renney, 1974; SBNM, 1997).

Cultivation and irrigation will kill diffuse knapweed. However, cultivation causes soil disturbance and may not be a management option on rangelands. In a location where cultivation is acceptable, regular deep ploughing can slice knapweed roots and bury seeds, thus weakening the plant. Cultivation in combination with reseeding competitive perennial grasses may minimize re-invasion by the knapweeds (Roche and Roche, 1999).

Chemical Control

In general, herbicides are expensive and do not provide lasting control of diffuse knapweed. They can cause negative impacts on water quality, and select for resistant weeds that may be worse than the original problem (Powles and Shaner, 2001). The persistent herbicides clopyralid and picloram, which are often recommended for knapweed control, are not metabolized by grazing animals and are not destroyed by composting (Fay et al., 1991; Bezdicek et al., 2001; Houck and Burkhart, 2001). If herbicides are used, they should be applied at the rosette or early bolting stage and before seeds are produced. They should always be combined with a revegetation programme (Woo et al., 1999, 2002).

Biological Control

To date, 13 beneficial insect species from Europe have been released in the USA and Canada as biocontrol agents of spotted and diffuse knapweeds. The biocontrol complex includes four moths, four weevils, one beetle and four flies. Eleven species have established, and at least seven are well established and spreading (Smith, 2001). Larval forms do most of the damage and attack either the roots or seedheads (Muller-Schärer and Schroeder, 1993; Rees et al., 1996; Smith, 2001). Knapweed densities are lower in areas where biocontrol agents have been released (Clark et al., 2001a). The weevil Larinus minutus is having a serious impact on plant growth and density at many locations (Story and Piper, 2001). The 13 species are:

Agapeta zoegana, a small, yellow root-boring moth works best where knapweeds are abundant but not yet a monoculture (Rees et al., 1996; Lang, 1997). The moth may preferentially attack older plants, and is more successful in combination with competitive plantings of grasses (Story et al., 2000).

Metzneria paucipunctella, a seedhead moth was released to control spotted knapweed. By 1988 it had spread onto diffuse knapweed. The larvae can destroy up to 90% of the seeds. However, low temperatures restrict the geographical range (Rosenthal et al., 1991; Rees et al., 1996; Good et al., 1997; Lang, 1997). Two other moths, Pterolonche inspersa and Pelochrista medullana were released but did not establish (Rees et al., 1996).

The four weevil and one beetle species released for biocontrol of knapweed include three that attack seedheads and two that attack roots. Bangasternus fausti, a seedhead weevil, attacks diffuse, spotted and squarrose knapweeds and, to a lesser extent, purple and yellow starthistle. The larvae feed inside the seedhead, destroying up to 100% of its contents. B. fausti requires undisturbed release sites with dry summers (Rees et al., 1996; Lang, 1997).

Larinus minutus and L. obtusus are seedhead weevils. L. minutus prefers diffuse knapweed and L. obtusus prefers spotted knapweed (Story and Piper, 2001). The adults are strong flyers and can easily disperse to new patches. The larvae feed on flowers and seeds, reducing seed production by up to 100% in infested seedheads. The weevils require dry areas with some bare ground, such as the outer edges of knapweed patches (Jordan, 1995; Lang et al., 1996; Rees et al., 1996; Lang ,1997; Kashefi and Sobhian, 1998).

Cyphocleonus achates, a root-boring weevil, prefers spotted knapweed, but will attack diffuse knapweed (Story and Piper, 2001). The weevil does not fly. The adults feed on knapweed rosettes. The larvae destroy the interior of the taproot and increase the plants susceptibility to pathogen attack. The insect requires high soil temperatures (Rees et al., 1996; Story et al., 1996; Lang, 1997; Story et al., 1997). Weevil numbers are highest when knapweed cover is 30-70%. Largest establishment rates in Montana, Idaho and Washington occurred at about 900-1500 m (Jacobs et al., 2000; Clark et al., 2001b).

Sphenoptera jugoslavica, a root borer, is a flat, blue-black, copper-coloured beetle which attacks the roots of diffuse knapweed. The larvae feed on taproots, causing gall formation near the root crown. Although larval root feeding can kill rosettes, mature plants survive (Rees et al., 1996; Lang, 1997; Lang et al., 1998).

Urophora affinis and U. quadrifasciata, two seedhead flies, attack both spotted and diffuse knapweeds. Eggs are laid between floral bracts and hatching larvae induce gall formation on the seedhead, causing new flowers to abort. Together, the seedhead flies can cause up to 95% reduction in seed production. Both species prefer open areas with full sun (Maddox, 1979; Rosenthal et al., 1991; Rees et al., 1996; Lang, 1997; Lang et al., 1997). Although galls are eaten by deer mice, extremely cold temperature is the most important factor affecting overwintering survival of the flies (Nowierski et al., 2000; Pearson et al., 2000).

Chaetorellia acrolophi, a seedhead fly, feeds on both diffuse and spotted knapweed. The larvae feed on the flower buds, reducing seed production. However, the effects of larval feeding on diffuse knapweed are unclear (Rees et al., 1996; Lang, 1997).

Terellia virens, a seedhead fly, primarily attacks spotted knapweed. It does not form galls and has two generations a year (Rees et al., 1996; Lang, 1997).

A number of pathogens attack knapweeds. Maculosin, a host-specific toxin produced by the black leaf blight fungus Alternaria alternata can destroy two-thirds of spotted knapweed foliage; however, younger leaves and buds are not affected and the plants are able to resprout (Bobylev et al., 1996). Sclerotinia sclerotiorum, a common soil fungus with a broad host range, is effective in killing juvenile spotted knapweed (Rosenthal et al., 1991; Jacobs et al., 1996).

Integrated Control

Combination treatments of fertilizer and herbicides have given inconsistent results. Sheley and Roche (1982) found knapweed suppression and grass growth was greater at sites receiving both treatments. Hubbard (1975) obtained similar results in dry areas, but not in areas where rainfall was greater. Sheley and Jacobs (1997) found that addition of nitrogen fertilizer had no effect on the growth of knapweed or native grasses in areas treated with the herbicide picloram for knapweed suppression.

It is hard to control diffuse knapweed if the treated area is constantly challenged with seeds. One study with small plots combined burning, cultivation, herbicide, competitive planting and nitrogen fertilizer. At the end of the third year, grass production peaked. For 3 years after that, grasses declined due to drought. Diffuse knapweed had re-established after 8 years. The authors attributed the re-invasion to 3 years of drought and to small plot size, which allowed reseeding from adjacent plots (Roche and Roche, 1999).

Public information systems and bounty programmes can be useful as part of an IPM programme. Bounty programmes have had some success in Montana (Lacey et al., 1988).