Version classiqueVersion mobile

Les espèces envahissantes dans l’archipel néo-calédonien

Marie-Laure Beauvais
Alain Coléno
Hervé Jourdan

Part one. Synthesis and recommendations

Control strategies

Texte intégral

1Despite the often considerable resources devoted to combating an invasive species already present, success is by no means certain, whether the means employed are mechanical, chemical, biological or ecological (i.e. overall ecosystem management, favouring endemic species and disadvantaging invasive species; wildfire control is one example). The best protection is certainly prevention. This can be done in a number of ways. The most radical form of prevention is to exercise rigorous control over points of access to the territory to avoid any introduction of species already known elsewhere for their invasive potential. Should such a species be already present without its invasive potential being known, its development and possible spread must be closely monitored.

2One of the difficulties is that there is no generic biological description of invasive species by which to identify them with certainty and so keep ahead of the situation. However, there are some plausible correlations between invasive potential and particular characteristics that justify prudence. Apart from their ability to adapt to disturbance of the environment where they are established, invasive species only demonstrate their invasive potential in the field. Taking account of the problem only when an invasion has occurred leads to tragic situations on the biodiversity side but also in economic terms. The cost of eradication can be enormous, and sometimes out of all proportion to the land area affected.


3The earlier an invasive species is detected, the less costly will be eradication or control. Once a species has been introduced the best that can be done is often a defensive strategy to control it, which is often very costly. However, there is no reason to stop quarantine measures once the presence of a noxious species has been detected. Genetic variability within a species could produce a new and even more damaging variety. But it is vain to hope to detect all pests (weeds included) everywhere in the territory. Detection should be targeted, and based on reference lists like that of the IUCN. It should also be based on better cooperation between the public and private sectors. For example, in Australia, industry biosecurity plans have been drawn up. For each noxious or potentially dangerous species these plans detail the different stages of the measures to take, including quarantine protocols, controlling movement and eradication.

4Detection at ports of entry and along transport routes within the territory is decisive. This requires qualified staff and cooperation between the public and the authorities. In Australia the costs are shared, which encourages farmers to react quickly. Vigilance must be particularly constant in areas of high population density, and it is useful to pool efforts regionwide. All those concerned must be alerted to the issue. For example, nearly 75% of new taxons naturalised on Australian soil have been intentionally introduced, mostly for ornamental purposes, so gardeners’ clubs and botanical societies need to be alerted. Invertebrates are usually introduced accidentally, and can be detected using existing pest lists (for arthropods, various trapping techniques seem to be effective). Detection of new diseases requires regional networks capable of promoting awareness among the communities and economic actors concerned.

5More generally, raising awareness by disseminating pictures of suspect species, e.g. on posters, is relatively effective. Images of and information about noxious species are available from various online databases (see paper by A. Sheppard et al.). Australia, New Zealand and South Africa have been pioneers in programmes of public education designed to provide early warning of the presence of high-risk species. The French have developed a simple online plant identification system.

6As regards active surveillance it will help to review the state of knowledge on the biology, epidemiology etc. of potentially harmful organisms. Since only a small proportion of non-native species prove to be invasive (see the rule of tens described in the introduction), priorities must be set. This is the aim of a surveillance system. For most introduced invertebrate pests and crop diseases, strategies of eradication or at least containment should be almost immediate. For weeds, reactions are generally slower as the process of establishment and naturalisation is usually far slower. But this type of management is damaging because in the long run economic losses due to weeds are significantly higher.

7The example of weeds sheds light on some rules that internal risk assessment systems must obey, as decisions are made at an increasingly local level. These systems are based on border assessment systems. The selection of internal weed-risk assessment systems must consider the spread stage(s) of all pests being compared, impacts on the affected systems, the likely benefits (and beneficiaries) of control efforts and the quality of the available information (see Table 11). These factors vary between countries. Sometimes a country uses several systems combined. A single system applicable on a large scale cannot be recommended until the goals have been clearly defined (see paper by A. Sheppard et al.).

Table 11 – Main systems used for internal weed risk assessment and prioritizing

Table 11 – Main systems used for internal weed risk assessment and prioritizing

8There are various factors to take into account in monitoring a species. It is worth examining the behaviour of related species if such information is available. It is also useful to have an approximate estimate of the species’ rate of spread (a determining factor for a possible decision to attempt eradication). Biological characteristics are another factor: certain biological features may have proven favourable to invasion in other parts of the world. The timeliness of detection, the mode and rate of reproduction and the dispersal potential are also factors to consider. Climate is also important, since this can be a barrier to the spread of a plant imported from other latitudes. The behaviour of the human population also plays an important part, especially when an otherwise noxious species seems to be useful for certain industries. Lastly, the choice of priorities must be based on an assessment of potential impacts, which can be done from historical data from other regions.

9To set priorities, i.e. identify the species it is most important to keep under control, a scoring system is used analogous to that used for border surveillance. Classification systems differ in the information they require and the structure of their internal rules, the simplest being to attribute numerical values to a set of criteria1.

10The effectiveness of surveillance depends on a practical, quantitative response following detection. On the island of Maui, Hawaii, an early detection project was introduced under a cooperation scheme with the US Geological Survey and the Maui Invasive Species Committee (MISC). Under this project a list was drawn up of some hundred potentially invasive weeds, either weeds known to be invasive on other Hawaiian islands or locally cultivated plants known to be invasive elsewhere. Nearly 2000 km of road were inspected and over 16,000 observations of target species recorded. For 79 species, some 1000 off-road sites were added with the help of field botanists. This action was an essential starting point for the MISC’s eradication efforts.

11A new approach, using “sentinel sites”, is being applied experimentally in Australia. It consists of selecting urban areas and national parks where there are numerous alien plant species, most of which are potentially invasive. Assessments of density and biomass were made, then repeated three years later to classify the species according to the trends in these parameters. In general the assessment efforts made in Australia have increased the detection rate. This is important, because early detection makes it possible to rapidly implement eradication or containment strategies and so reduce economic losses and environmental impact (see box).

The value of early detection: the red imported fire ant in Australia
The Red Imported Fire Ant (RIFA) was discovered on Fishermans’ Island near Brisbane, Australia, in February 2001; 40,000 ha were already infested. This ant had been declared a pest under the Plant Protection Act in 1989. It is now mandatory to declare its presence. In September 2001, a fire ant control centre was set up and a major nationally funded eradication programme was launched. This programme was set up by the Queensland Department of Primary Industries and Fisheries, with the suport of the Agricultural Resource Management Council of Australia and New Zealand. The first eradication campaign took place in October 2001. Six months later, regulations were introduced to restrict movements of materials liable to spread the fire ant. At the same time massive public awareness and education campaigns were undertaken. In April 2004, the programme received an additional 37.5 million Australian dollars to extend the eradication campaign by one year and treat a new area. The total budget for a six-year period has been AU$ 175.4 million.
The red imported fire ant can cause a wide range of damage to animals, human health, industrial infrastructures, farming and horticulture etc. (see paper by H. Jourdan). The Australian Bureau of Agricultural and Resource Economics undertook a cost-benefit analysis in 2001 to estimate the usefulness of an eradication, programme. This analysis showed that the cost to society would reach AU$ 8.9 billion over the next thirty years unless the ant was brought under control. The cost-benefit ratio of the programme was estimated at 25 to 1 (based on the sum of AU$ 124 million for a five year programme).
If nothing was done, it was variously estimated that the ants would colonise 600,000 km2 to 4 million km2 by 2035. It is thought that they had been present for ten years before they were discovered. The delay in detection explains how they were able to spread from Southeast Queensland. Today, treatment is being carried out on limited sites covering 30,000 hectares and surveillance continues elsewhere. At this stage, no one can tell how long eradication may take.


12Eradication is highly controversial issue. Some consider that it is rarely achievable, often entails exorbitant cost and can cause substantial damage to non-target organisms. Others think the choice between eradication and control is like the choice between paying outright and paying in instalments. In this view control is the worse choice because it involves investing public funds over a long period and with no closure date set in advance. The same people argue that if eradication is undertaken early enough it has every chance of succeeding, and that subsequent control amounts to only routine surveillance. Nonetheless, some objections (like the problem of collateral damage) may be so serious that government action is required.

13When should eradication be undertaken? There is no doubt that early detection of an invasive species is decisive. Some authors suggest a maximum infested surface area as a criterion, the logarithm of the cost increasing linearly and rapidly with the logarithm of the area. In California, for example, it would appear that out of 50 infestations by 16 weeds, few eradication projects involving areas of more than 1,000 hectares have been successful. However, other writers dispute this upper limit. Recently six criteria for success eradication have been suggested (see box).

The six factors for successful eradiation (from Myers et al., 2000)
sufficient resources to finance the programme to completion (costs can be high when damage is already widespread);
a clear authority, necessary for authorising a person or administration to take all necessary steps including access to private land;
use of a biological approach to decide which species should be treated in this way (dispersal capacity, mode of reproduction, historical data etc.);
vigilance with regard to any risk of re-invasion;
the possibility of detecting the species at low densities;
vigilance with regard to any negative reaction by the ecosystem (e.g. proliferation of another noxious weed).

14Certain species characteristics can facilitate the work. It is easier to detect large mammals than small insects, large plants than small. With plants, some biological characteristics such as a high rate of reproduction and dormant seed banks are major drawbacks. In any case scientific analysis of the results must be encouraged, so that decisions are made on the basis of scientific rather than political criteria. For reasons unrelated to scientific rationality, governments tend to favour eradication programmes and to continue them even if complete eradication seems impossible.

15The economic equation is complex. It is often pointed out that the cost of eliminating 99% of a target population is less than the cost of efforts to totally eliminate it (see paper by A. Sheppard et al.). This may lead the authorities to reduce their funding, as happened in Florida with management of hydrilla (hydrilla verticillata). Ideally, a full cost-benefit analysis should be performed before deciding to attempt eradication.

16Cost-benefit analysis in the natural resources field is a complex undertaking (see p. 199). In the case of invasions, it is difficult to predict their progress or the effects of different management measures. In 2004, a conceptual model for agricultural and natural ecosystems was put forward. It takes account of the intrinsic rate of spread of the invader’s range, the cost per unit area of controlling the invasion, and the damage per unit area caused by infestation. It identifies a threshold (to be determined) beyond which the colonised area is too large for successful eradication. However, economic analyses of this kind should include the risk of spread to other areas.

17Eradication is more likely to succeed on islands, small islands especially. Many successes have been recorded on islands around the world. With weeds, the most successful eradication campaigns on Pacific islands have been achieved where detection occurred early on. Among the species eradicated are the octopus tree (Schefflera actinophylla) in Palau, Antigonon leptopus on Niue, and in Hawaii the ivy gourd (Coccinia grandis), Jerusalem thorn (Parkinsonia aculeata) and fountain grass (Pennisetum setaceum). Other successes have been achieved in New Zealand. Commensal rodents (Pacific rat, brown rat, black rat) have been eliminated on more than 90 islands over the past forty years (see other examples in the paper by A. Sheppard et al.).

18Not all attempts are successful. A classic case is that of common crupina (Crupina vulgaris), first detected in an 18-hectare area of Idaho, USA, in 1969. Twelve years later it was established on 9,000 hectares and was classed as a noxious weed at the federal level. An eradication feasibility study was conducted and concluded that eradication was possible. But the project’s joint federal and state planning group was not formed until 1991, by which time common crupina had spread to California, Oregon and Washington and affects 25,000 hectares. If action had been taken in the 1970s perhaps an immediate solution would have been found. In fact the planning group decided against immediate action on the grounds that the recommended herbicide would have a negative impact on salmon. Other examples of failure are the fire ant in the south-eastern United States and skeleton weed or rush skeleton weed, Chondrilla juncea, in Western Australia.

19According to Judith H. Myers there are three alternatives to eradication: containment to slow the spread of an organism, maintaining it at a tolerable density, and biological control. Containment leaves the way open for the possible emergence of new control methods such as better pesticides/weedkillers and genetic modification. Density control calls for conventional mechanical, chemical and biological control methods, and biological control has sometimes been successful, sometimes not. The precautions taken at present, based on relevant scientific knowledge, suggest that these methods will play a key role in future. In any event, there is no simple solution to infestation problems. Even where eradication works, another alien species whose spread has been kept in check by the first species may then proliferate rapidly.

Table des illustrations

Titre Table 11 – Main systems used for internal weed risk assessment and prioritizing
Fichier image/jpeg, 213k

© IRD Éditions, 2006

Conditions d’utilisation :

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search