Publications des agents du Cirad


Pl@ntInvasive-Kruger: computer-based identification and information tools to manage alien invasive species in the Kruger national park, South Africa

Le Bourgeois T., Foxcroft L.C., Thompson D.I., Guezou A., Grard P., Taylor R.W., Marshall T., Carrara A.. 2013. In : 12th International Conference Ecology and Management of Alien Plant Invasions (EMAPi 2013), 22-26 September 2013, Pirenopolis, Brazil. s.l. : s.n., 2 p.. International Conference Ecology and Management of Alien Plant Invasions. 12, 2013-09-22/2013-09-26, Pirenopolis (Brésil).

Invasive plant species are a major threat to the biodiversity of protected areas and South Africa's Kruger National Park (KNP), which covers some 20 000 km2, is no exception. Landscape and habitat diversity and a river network that drains highly invaded catchments outside the park make KNP highly susceptible to alien plant invasions. Efficient control of the threat requires early detection of introduced species, effective methods of eradication, and dissemination of current data to increase awareness. A collaborative project was initiated to provide a platform and a range of tools to support these actions. Methods Pl@ntInvasive-Kruger aimed to develop a suite of science-based, computer-driven tools for use by KNP managers, researchers and teams involved in alien plant control. Three applications were developed, each of which is supported by a global, multi-user database: i. The PUBLISH tool returns synthesised species information; ii. IDAO, which utilises computer-aided plant identification ; and iii. IDENTIFY, an image recognition system. Results & Discussion The online DataManager database allows data management by members for field surveys and facilitates collections management and the automated synthesis of species information. These syntheses are available as HTML pages through the PUBLISH tool, which detail descriptions and imagery of both invasive and contained alien plants, and includes information on ecology, biology and management. These data support the two identification tools. PlantInvKruger-IDAO constructs unknown species in a step-wise manner from prominent characteristics selected by the user from schematic multiple-choice menus. The IDAO application is compatible with a range of mobile electronic devices. PlantInvKruger-IDENTIFY assists, through image recognition algorithms, with the specific identification of images of plants and plant parts submitted to the database through a web interface. In both cases the suggested identification is expressed as the similarity of the unknown specimen to the type specimen information housed in the database and can be confirmed by accessing the species HTML pages. All tools are used from a collaborative web platform (Pl@ntNet) where members can also share information and documents and manage discussions ( Conclusions Correct identification is an essential aspect of alien plant control programs, but is difficult and time consuming where large numbers of alien and indigenous species occur. The Pl@ntInvasive-Kruger database currently contains information on almost 400 alien plant species, with the identification tools focussing on the 113 most important species. By assisting with the identification of invasive plant species and facilitating the sharing of information between interested parties, Pl@ntInvasive-Kruger will promote biodiversity conservation in KNP. This project is a case study of the Pl@ntNet project funded by Agropolis Fondation. It is currently in the final testing and refinement stage, with evaluations being conducted with end users....

Documents associés

Communication de congrès

Agents Cirad, auteurs de cette publication :