Hedychium gardnerianum (kahili ginger)

The spread of H. gardnerianum is mainly furthered by its seeds, which are produced in very large numbers, and eaten by frugivorous exotic birds and rats (e.g. Acridotheres tristis (mynah bird) and tūī (Prosthemadera novaeseelandiae) in New Zealand; Turdus merula azorensis in the Azores). This mechanism allows both short- and long-distance transport of the seeds. Seeds collected by birds may be eaten where they are found or transported and the seeds will be ejected some distance from the parent tree and will not impair germination. Such long-distance seed dispersal by animals appears to be the most effective method for establishing new points ('foci') of invasion. Each new colony also increases the size of the expanding margin of the invasion of H. gardnerianum by means of rhizomes, which can also be broken off and spread by water. Dispersal may also occur by hydrochory (rain water flushing of seeds and corms) over short distances (Cordeiro and Silva, 2003).

New plants may establish from rhizome fragments spread in dumped vegetation and fill, movement of soil, flooding and contaminated machinery (Weedbusters, 2016).

Hedychium gardnerianum is used for ornamental purposes; several horticulturists sell rhizomes or plants. Such introductions are encouraged by the availability of seed from the horticultural industry via mail-order catalogues, and as such, deliberate introduction of H. gardnerianum is quite likely.

Chromosome numbers of 2n = 34 (Sharma and Bhattacharya, 1959) and 2n = 54 (Raghavan and Venkatasubban, 1943) have been reported for H. gardnerianum. A spontaneous naturalizing hybrid (H. coronarium × H. gardnerianum = H. sadlerianum) is thought to have been found in Medeiros, Haleakala National Park Maui (Hawaii), which favours the simultaneous seed production of H. gardnerianum (P Bily, The Nature Conservancy, USA, personal communication, 2008). Other interspecific hybridization may be possible and a natural hybrid of ‘Tara’ (form of H. coccineum) and H. gardnerianum is postulated as the two species coexist in Nepal (PIER, 2007).

Hedychium gardnerianum reproduces by seed and vegetatively by rhizomes. Flowers are produced in attractive spikes. The scent of the flowers and the conspicuous petals and petaloid staminodes indicate insect pollination, while the long, exerted stamen and style suggest that hovering insects in particular may be involved. Wing scales of Lepidoptera have been found adhering to stigmas of some Hedychium spp. in their native habitats and it is likely that long-tongued butterflies in search of nectar produced by the prominent gland at the base of the style are likely to brush the stigma followed by the anther-cells with their wings and carry out cross pollination.

In its native range, H. gardnerianum flowers and fruits from July to December. Flowering occurs from December to March in La Réunion and after fructification in April to August, the above-ground parts of the plant senesce and die. A new stem will grow from an axillary bud present in the rhizome. In New Zealand, it has been recorded flowering from February to April. In Queensland (Australia), it has been observed producing flowers all year round. In South Africa, flowers appear from January to March. In Hawaii, bud initiation begins from May, peaking in June-July, flowering peaks in July-August, immature fruits in August-October, and ripe fruit and seed are present from October to December (Medeiros, 2004).

Hedychium gardnerianum is a non-woody herbaceous perennial.

In the Azores, a study of seed production and vegetative growth on São Miguel revealed that the number of seeds per spike ranged from 300 to 500 and concurred with the findings of Byrne (1992), who reported 20 to 600 depending on light conditions. H. gardnerianum seeds do not have dormancy (Cordeiro, 2001) and only remain viable in the soil for a short period, relying on high annual seed production as a mechanism for efficient dispersal. Each new corm usually produces new leaves and eventually a spike although many act as storage organs and lack leaves. Generally, the pseudo-stem, leaves and spike collapse in winter, leaving a scar at the top of the corm. Vegetative growth is characterized by the development of new corms at the rhizome ends, creating a dense cover and crowding the soil, which is how it impairs regeneration of native species.

Hedychium gardnerianum prefers moist, wet and well-drained soil, but it is ecologically very versatile, thriving in bright light or dense shade, good or poor drainage, high or low fertility, thick humus or very little soil, even sprouting in tree forks above the ground and on tree trunks, like epiphytes in the native rainforest in La Réunion. This species is also adapted to grow in wetlands with annual rainfall of 2000-5000 mm, and with an annual average temperature between 11° and 17°C (Cadet, 1977). It generally occurs at altitudes of 800-2000 m in native wet forests and disturbed areas. In Hawaii, populations are found on all islands between sea level and 1700 m (Smith, 1985).

There are no records of direct impact on crops, but mechanical and chemical control of H. gardnerianum is very costly. A combined assessment of the estimated annual expenditure on H. gardnerianum control in the Hawaiian Islands of Kauai, Oahu, Maui and Hawaii comes close to US $1 million, with approximately 50,000 acres estimated to be infested, although this is certainly an underestimate that does not include the full range of invading outlier populations (P Bily, The Nature Conservancy, USA, personal communication, 2008).

Although rhizome beds may become very deep, H. gardnerianum roots are quite shallow and maintain poor purchase in the soil. In steep areas, prolonged rainfall causes these rhizome beds to become heavy with absorbed water and the soil to become slip-prone. The combined effect of added weight, slip-prone soil and weak roots often leads to erosion, with entire hillsides of H. gardnerianum disappearing at once. Erosion also downgrades water quality and causes siltation of rivers and harbours (Northland Regional Council, 2000). H. gardnerianum produces thick beds of rhizomes, forming a dense ground cover which prevents regeneration of other species. It spreads into shaded areas on native or managed forests and displaces lower tier plants of the native communities. This greatly impairs the regeneration of forests and proper functioning of ecosystems: H. gardnerianum threatens the biodiversity of native, undisturbed ecosystems. Studies by Minden et al. (2010) show that with removal of H. gardnerianum, the native Hawaiian forest is likely to regenerate and regain its natural structure.

In Hawaii, areas between sea level and 2300 m where the annual rainfall exceeds 1500 mm, the land below 1350 m has been greatly altered by agriculture. The gullies at these lower altitudes are forested by a tropical weed flora and between 1000 and 1650 m, H. gardnerianum creates significant problems, and able to invade and take over native ecosystems without any apparent disturbance (Smith, 1985), preventing regeneration and threatening viability. An aircraft-based analysis has found it to reduce the amount of nitrogen in the native Metrosideros polymorpha rainforests which have an impact on the natural ecosystem processes and can alter the type of fauna able to inhabit such habitats (ISSG, 2006).

Hedychium gardnerianum competes with native flora and disturbed native ecosystems and threatens biodiversity (Macdonald et al., 1991). It out-competes other species for light, space, nutrients and moisture and its shade tolerance enables it to thrive in forests. Other invasive characteristics include rapid growth, long-term persistence of plants and ability to recover from removal of stems, making slashing worthless as a control measure. H. gardnerianum produces thick beds of rhizomes, forming a dense ground cover, which prevents seedlings of other plants from growing through them. The stems and leaves also form dense thickets, preventing other species from germinating below or growing amongst them.

Hedychium gardnerianum threatens remnants of primary forests of La Réunion. In the Azores, continuous expansion also threatens several fragments of endemic vegetation, leading to the prediction that several communities of lichens, vascular plants, molluscs and arthropods native and endemic to the Azores might be endangered, particularly on the islands of São Miguel, Santa Maria and Flores. Recent studies (Borges et al., unpublished observation) suggest that for several endemic species of arthropods with a wide distribution in the Azores, the smallest population densities are found in fragments disturbed by exotic plants.

Of concern also is the present situation of H. gardnerianum on Terceira Island, since it is now present in small gaps in the middle of large fragments of otherwise pristine native forest. Infestations on São Miguel Island also threaten the Azores bullfinch (Pyrrhula murina) which is restricted to threatened natural laurel scrub and forest. The Madeira archipelago supports remnants of a type of laurel forest (laurisilva or lauraceas madeirense) which is a protected area towards the higher altitudes that was once widespread throughout southern Europe and north-western Africa. The flora and fauna of this relict forest is quite unique; it has many endemic species including the Madeiran long-toed wood pigeon (Columba trocaz) and Zino's petrel (Pterodroma madeira). Having suffered extensive clearance after the islands were settled, the native vegetation is currently threatened by invasive species and the laurisilva forest, recently declared World Nature Heritage under the aegis of UNESCO is under serious threat due in particular to H. gardnerianum, which recently went into its colonization phase in new and extensive areas and can also impede the natural expansion of laurisilva on abandoned rural land.

Bird density and vegetation characteristics were studied in wet montane forest in native Hawaiian forests dominated by Metrosideros polymorpha at the summit of Kilauea Volcano, on plots containing greater than 90% H. gardnerianum invasion and on plots from which the species was eradicated. Results supported the hypothesis that species less dependent on native fruiting plants in the understorey will remain unaffected or possibly augmented (apapane (Himatione sanguinea) and Japanese white-eye (Zosterops japonicus)), but did not support the hypothesis that H. gardnerianum invasion decreases the density of understorey feeding birds. H. gardnerianum is a threat to Labordia tinifolia var. lanaiensis and Clermontia samuelii in Hawaii (PIER, 2007).

H. gardnerianum competes with native flora and disturbed native ecosystems and threatens biodiversity (Macdonald et al., 1991). It out-competes other species for light, space, nutrients and moisture and its shade tolerance makes it able to thrive in forests. Other invasive characteristics include rapid growth, long-term persistence of plants and ability to recover from removal of stems, making slashing worthless as a control measure. H. gardnerianum produces thick beds of rhizomes, forming a dense ground cover, which prevents seedlings of other plants from growing through them. The stems and leaves also form dense thickets, preventing other species from germinating below or growing amongst them. H. gardnerianum threatens remnants of primary forests of La Réunion. In the Azores, continuous expansion also threatens several fragments of endemic vegetation, leading to the prediction that several communities of lichens, vascular plants, molluscs, and arthropods native and endemic to the Azores
might be endangered, particularly on the islands of São Miguel, Santa Maria and Flores. Recent studies (Borges et al., unpublished data) suggest that for several endemic species of arthropods with a wide distribution in the Azores the smallest population densities are found in fragments disturbed by exotic plants. Also of concern is the present situation of H. gardnerianum in Terceira island, since it is now present in small gaps in the middle of large fragments of otherwise pristine native forest. Infestations on Sao Miguel Island also threaten the Azores bullfinch which is restricted to threatened natural laurel scrub and forest. The Madeira archipelago supports remnants of a type of laurel forest ("laurisilva" or "lauraceas madeirense") which is a protected area towards the higher altitudes that was once widespread throughout southern Europe and north-western Africa. The flora and fauna of this relict forest is quite unique; it has many endemic species including the Madeiran long-toed wood pigeon (Columba trocaz) and Zino's petrel (Pterodroma madeira). Having suffered extensive clearance after the islands were settled, the native vegetation is currently threatened by invasive species and the laurisilva forest, recently declared World Nature Heritage under the aegis of UNESCO is under serious threat due in particular to H. gardnerianum, which recently went into its colonization phase in new and extensive areas and can also impedethe natural expansion of laurisilva on abandoned rural land. Bird density and vegetation characteristics were studied in wet montane forest in native Hawaiian forests dominated by Metrosideros polymorpha at the summit of Kilauea Volcano, on plots containing greater than 90% ginger invasion and on plots from which ginger was eradicated. Results supported the hypothesis that species less dependent on native fruiting plants in the understorey will remain unaffected or possibly augmented (Apapane and Japanese white-eye), but did not support the hypothesis that ginger invasion decreases the density of understorey feeding birds.

H. gardnerianum is a threat to Labordia tinifolia var. lanaiensis and Clermontia samuelii in Hawaii (PIER, 2007).

Hand pulling can be effective on young H. gardnerianum seedlings. Small populations may be dug out and all underground parts must be removed. In La Réunion, mechanical cutting is frequently used to clear it from the forest understorey, forest margins and ravines or roadsides. This method provides no long-term control but slows down its propagation and spread. Cutting before flowering (December to April, in La Réunion) is more effective in reducing seed production. H. gardnerianum inflorescences can be cut and dropped on the ground prior to the seeds being formed: this removal will not kill the plant but it will stop it seeding. The National Biological Invasions research programme INVABIO in La Réunion was launched in 2000 by the French Ministry of Environment, and as part of this, a three-year research project began in 2002 with the aim of studying the impacts of mechanical control measures on native biodiversity and dynamics of native communities. The results clearly showed that the removal of an invasive alien species causes modifications to the biodiversity and a quick return to the initial situation. Contrary to the results expected, the negative impacts on the ecosystem were numerous. From an economic point of view, the intervention was both very expensive (the removal of 70 t/ha of H. gardnerianum was estimated at 24,062 euros) and constrained (specialized workers, steep slopes, etc.). By contrast, non-intervention caused few disturbances. In the case of the very low-invaded native ecosystems, the recommendation for non-intervention and the least disturbance as possible is made when the ecosystem is little invaded, or invasion is recent, or the invasive alien species occupies a small area; then early detection and rapid eradication are of interest (preventive action) (Lavergne, 2005).

A number of herbicides have been investigated and used in the past 20 years for control of H. gardnerianum (Harris et al., 1996). Currently, the most effective herbicide reported for its control is metsulfuron, sprayed on the leaves, stems and root system. The effects are noticeable after 3 months and the weed will die and its rhizomes will rot after 12 to 15 months (Auckland Regional Council, 1999). Glyphosate and amitrole are also used in New Zealand (Timmins and Mackenzie, 1995). Probably the most effective and cost-efficient control method is stump treatment with herbicide. The stems are cut close to the ground, just above the pink-coloured swelling at the stem base, and concentrated herbicide is applied to the freshly cut stump; glyphosate is the common herbicide for this in La Réunion. Due to the widespread distribution of H. gardnerianum, chemical control is more cost effective than manual or mechanical control. In Hawaii, unpublished research found that metsulfuron at 0.04 lb/acre, imazapyr at 0.7 lb/acre, and amitrole at 0.7 lb/acre applied to visible rhizomes after mechanical clearing of top growth was very effective. A commercial mixture of 2,4-D + triclopyr was ineffective but 20% triclopyr ester product applied in straight stream to the stems and surface rhizomes provided 100% kill on Maui (Motooka et al., 2003).

Chemical and mechanical control of H. gardnerianum is labour intensive, expensive, time consuming and often ineffective, and environmental concerns such as soil leaching and ground water contamination often limit the usefulness of herbicides in some areas. Biological control is now considered the only practical approach for the long-term management of large H. gardnerianum infestations in native forests (Harris et al., 1996).

Ralstonia solanacearum, a naturally occurring bacterium in Hawaii, was tested for its efficacy as a bioherbicide in Hawaiian forests. The ability of this bacterium to cause severe disease in H. gardnerianum, together with its apparent lack of virulence in other ginger species seemed to indicate high potential as a biological control agent (Anderson and Gardner, 1999). Plants were inoculated with an aqueous suspension of the H. gardnerianum-infecting strain of the bacterium by stem injection or root wounding and all inoculated plants developed irreversible chlorosis and severe wilting 3-4 weeks following inoculation. Systemic infection also caused death and decay of rhizomes. Subsequently, further experiments were carried out (Paret et al., 2008) on species of ginger to evaluate their susceptibility to R. solanacearum (Rs) race 4 (ginger strains) by several methods of inoculation, including tests to simulate natural infection. Contrary to previous reports that Rs strains from H. gardnerianum were non-pathogenic on ornamental gingers, the H. gardnerianum strain wilted both ornamental and edible ginger (Zingiber officinale) species within 21 days. The ability of Rs race 4 to infect many ginger species without wounding, and survive long periods indicates that high risks will be incurred if the H. gardnerianum strain is inadvertently introduced from the forest reserves into ginger production areas. Conflicts of interest can often limit which plant species can be targeted for biocontrol however, and ornamentals, such as ginger (Hedychium spp.), cannot be targeted for biocontrol in Hawaii because they are seen as being of high value to the Hawaiian tourist industry, despite their status as serious invaders of Hawaiian forests (Gardner et al., 1995).

In 2008, a scoping study to investigate the potential for classical biological control of invasive Hedychium species was funded by a consortium of sponsors from Hawaii and New Zealand. Surveys in the native range (Eastern Himalayan foothills of India) identified a large guild of natural enemies exerting considerable pressure on H. gardnerianum populations. The biological control project led by CABI scientists in collaboration with Government organizations in India is ongoing and research is now focused on prioritizing natural enemies for host specificity testing in the UK (D Djeddour, CABI, personal commmunication, 2011).