In tropical pastures, weeds are a significant threat and their control is an important issue depending in their ecological and economic impacts. Weed management requires information on their identification, their biology, their behavior at different scales (from field to region). Part of this knowledge is present in the stakeholder’s heads (farmers, extension people, researchers, trainers), but its compilation and its diffusion present constraints of a regulatory, human and technological order. Management of information is now facilitated by digital infrastructures, but requires knowledge to be shared, updated, and disseminated. The WIKTROP collaborative portal (http://portal.wiktrop.org) allows all users in the tropical agricultural world, including grazing land, to share, capitalize and disseminate their knowledge and questioning about tropical weeds and their management for more than 700 species. The portal includes also a species identification tool and an observtion module. This bilingual (English/French) tool is available to multisystem applications (PC on- and offline, mobile for Android and iOS). Plants that appear spontaneously in pastures can be native or exotic, can be harmful or/and beneficial depending on the point of view (agricultural, medicinal, food...) and its abundance. How much information can or should be disseminated to help farmers manage weeds as best as possible? Is it necessary to eradicate or reduce the population of a particular species? Is it important that farmers take their decision with as much knowledge as possible about the disadvantages and benefits of each species? That is why WIKTROP is based on a free access to information. Sharing data, information or knowledge for the benefit of all is still poorly perceived but is fortunately developing rapidly, even in research. This becomes a recommendation or even an obligation in many national and international calls for projects and is becoming a new strategic issue. Wiktrop contributes tothis approach in tropical weed science

La production agricole tropicale est en pleine mutation et la gestion des enherbements est un element majeur de la strategie de l'agriculteur. L'elaboration d'itineraires de gestion des adventices requiert de plus en plus de connaissances, sur la biologie des especes, leur comportement dans les milieux cultives et l'efficacite des pratiques de desherbage. L'acces a l'information est maintenant facilite par les infrastructures numeriques en constante evolution, mais necessite que la connaissance soit partagee, mise a jour et diffusee de facon continue. Une part de connaissance est presente chez chacun des acteurs (producteurs, vulgarisateurs, chercheurs, enseignants), mais sa compilation et sa diffusion presentent differentes contraintes d'ordre reglementaire, humain et technologique. Le portail collaboratif WIKTROP permet a tous les acteurs du monde agricole tropical de partager et diffuser leurs connaissances et leurs questionnements sur les adventices tropicales et leur gestion. Il est associe a des applications mobiles facilitant son utilisation au bureau, sur le terrain ou pour des formations.

Preserving tropical vegetation biomes stems with an accurate identification of constituent plant species. In Asia, identifying mangrove plant species is particularly required in biodiversity assessment, restoration or preservation programs, which are pivotal components of the ecosystem's preservation from unsustainable anthropogenic activities. The necessary rigour coming from botanical expertise, however, remains both insufficiently considered and unavailable for non-botanists including scientists, local communities and stakeholders of the coastal zone management. In this work, we will present a mobile application, working on both iOS and Android, that allows for the identification of 51 Asian mangrove species at your fingertips. We have based our approach on generic drawings covering the full range of botanical characters that guarantee the identification of mangrove species at both propagule and adult growth stages. The tool has been evaluated in situ by botanist and non-botanist scientists from different Asian countries. In the first instance, we will discuss difficulties in correctly identifying Asian mangrove species, especially, those from the same family, and the technical approach we chose to finally solve them. Then, subsequently, we will demonstrate how, with this app, we will be able to map mangrove species in tropical vegetation biomes with the objective of sharing accurate survey data on existing mangrove flora all around the world, year after year.

The Mangrove ecosystem offers a range of benefits and opportunities for local and national economic development such as fisheries, shoreline stability, and resource sustainability hence they should be protected and conserved. In Kenya Mangroves were declared government reserve forests by the Proclamation No. 44 of 30th April 1932, and later by Legal Notice No. 174 of 20th May 1964. Under this “Gazette Notification for Mangrove Forests in Kenya” all land between high water and low water marks (ordinary spring tides) are described as mangrove areas. These forests cover about 61,279ha representing approximately 3% of the natural forest cover or less than 1% of the national land area. Mangrove forests are found in tidal estuaries, creeks, and protected bays along the 536 km long Kenyan coastline that extends from the Kenya-Tanzania border in the south to the Kenya-Somalia border in the north; between latitudes 1°40΄S and 4°25΄S and longitudes 41°34΄E and 39°17΄E. Mikoko (Mangrove in Swahili) is the first project specifically developed for mangrove areas in Kenya. Driven by the Kenya Forest Service, the project takes into account the recommendations of the National Mangrove Ecosystem Management Plan 2017-2027. For effective management of mangroves in the country, the Plan is organised around six programmes including; forest conservation and utilization; fisheries development and management; community; tourism development; research and education; and human resource and operations. These programmes prescribe measures for rehabilitation, conservation, and sustainable management of mangrove ecosystems in Kenya. Implementations of this Plan will be led by Kenya Forest Service (KFS); but will also include other key actors such as Kenya Wildlife Service, State Department of Fisheries, Research Institutions, Academia and Community Forest Associations through a specially constituted National Mangrove Advisory Committee. The project is for a 2-year period during which it shall develop an effective system for management of mangrove resources in Kenya. This will be done through the development of management prescriptions, skills improvement and provision of adequate infrastructure as recommended by the mangrove plan. Biodiversity information on mangroves, covering species descriptions and distribution, is essential for understanding the ecology and distribution of this endangered ecosystem and its management. Nowadays, very few is done in that domain. If the mangroves location are well know, it's not the case regarding the species distribution. Citizen science is a way to address this important issue. MIKOKO, a French FSPI funded project, is focused on Kenya to develop species information systems on mangrove species. Mikoko aims to assimilate knowledge through its participatory portal to strengthen a science and technology network of stakeholders such as scientists and parataxonomists in the African region. The portal platform will contain a Citizen Science module, a graphic driven species identification module, a species pages module conforming to Species Profile Model and a spatial module. Subscribing to open data paradigm, all the data on the portal will be covered by Creative Commons license framework. Mobile applications for identifying 60 plant species and contributing to citizen science module will be deployed. The oral presentation will expose the portal features and related apps seeking participation from the environmentalists from the African region.

Conventionally, species traits concepts have been conceived from an ecological perspective after grouping them as functional traits, response traits or effect traits: attributes of individual organisms that express phenotypes in response to the environment and its effects on the organism. From an informatics perspective, traits may be conceived to encompass a broader vocabulary that can capture any species attribute including, but not limited to those concerning its morphology, taxonomy, functional role, habitat, ecological interactions, trophic strategies, genetics, evolution, conservation status, anthropological uses, ecosystem services etc. The evolution of such a vocabulary and its standardisation across disciplines and taxa is a challenge, but one that needs imminent attention as the field develops. Furthermore, traits can have values that vary within and across individuals and species. The ability to associate traits with levels of a taxonomic hierarchy, aggregate species traits from individual records, flexibility to attribute categorical text, numeric, temporal and spatial values; associate them with ontologies; and conform to standards, can evolve traits as a flexible framework to structure descriptive, numeric and tabular data on species. Such a framework for structuring descriptive species data will, allow better discoverability and navigation of the information and has potential for developing further applications such as polyclave identification keys and analytical aids for big data. The open source Biodiversity Informatics Platform that powers three international initiatives across Asia and Africa has been evolving as an effective platform to aggregate and build open access databases for varied biodiversity data types. It has ability to handle varied data types such as descriptive data, occurrences, maps and documents. The platform has recently added a traits infrastructure that is participatory and can aggregate traits from curated databases as well as by crowdsourcing from observation and collection data. It is flexible in building vocabularies to structure descriptive species information and media, evolving into a framework which allows flexible yet efficient navigation of species information in an information system. Here, we discuss this model, its application within the applied initiatives, its potential use in classifying multimedia data for species characterization in a complex context and in facilitating trait analysis. We also cover potential applications of the trait framework for developing into a comprehensive and effective infrastructure for aggregating and structuring species information.

Invasive plant species are a major threat to the biodiversity of protected areas and South Africa's Kruger National Park (KNP), which covers about 20 000 km2, is no exception. The extensive river network draining highly invaded and transformed external catchments make KNP highly susceptible to alien plant invasion. Efficient control thereof requires early detection of alien species, increased awareness, effective eradication and dissemination of information amongst all relevant stakeholders. Pl@ntInvasive-Kruger was initiated to provide a platform and a range of tools in support of these actions locally by aiming to develop a suite of science-based, computer-driven tools for use by KNP managers, researchers and field teams involved in alien plant control. Four applications were developed, each supported by the global, multi-user PL@NTMANAGER database: 1) PUBLISH returns synthesised species information; 2) IDAO executes computer-aided plant identification; 3) IDENTIFY provides image recognition; and 4) CYBERTRAKER acts as an interface for mobile data collection applications. This toolkit facilitates the sharing of information between stakeholders, including scientists, field officers and citizens, and disseminates synthesised information to managers of invasive plants, eradication teams and conservationists. All tools are usable, or can be downloaded, from an internet based platform where members can also share information and documents and engage in discussions on alien invasive plants. Correct identification is an essential aspect of alien plant control programs, but is difficult and time consuming where large numbers of alien and indigenous plant species co-occur. The Pl@ntInvasive-Kruger database currently contains information on ~237 alien plant species, with the identification tools focussing on the 113 most important species. By facilitating the identification of invasive plants and supporting knowledge-sharing and the use of shared tools for a common goal, Pl@ntInvasive-Kruger promotes invasive alien plant control and thereby biodiversity conservation in KNP. (Resume d'auteur)

A significant challenge to agricultural productivity is weed management. This is particularly impor-tant in tropical areas where weeds can cause at least 20% loss in the productivity of food and cash crops. To help stakeholders of the Western Indian Ocean in improving weed management, the WIK-WIO (Weed Identification and Knowledge in the Western Indian Ocean) project has developed a par-ticular combination of scientific approach and ICT tools. This project is a EU-ACP Science & Technology II program-funded action. The WIKWIO initiative contributes to the establishment and strengthening of a community of stake-holders from research, education, extension and production around integrated weed management, by promoting knowledge sharing and collaboration. The action is aimed at enhancing capacities through an interdisciplinary and participatory approach, by building an ICT knowledge base of major weeds of food and cash cropping systems in the region. The approach consists in sharing data and knowledge among stakeholders, helping people in weed identification, and providing information on weeds and their management. All these aspects are made available through the web 2.0 participatory bilingual (Fr/En) WIKWIO portal, http://portal.wikwio.org. From this platform, any visitor can identify a weed with the IDAO module and get information on a specific weed and its management from the Species module. Several working group modules provide weed management information for different cropping systems. Portal-regis-tered members can provide knowledge on species and contribute to the species information DB, and also post field observations in order to contribute to weed distribution mapping. They can post obser-vations of unidentified weeds to request help from the community for identification and give informa-tion on its management. Scientific or technical documents can be shared on the portal. Members can comment on any piece of information present in the different modules of the portal. This comment can afterward become the starting point of a fruitful discussion. The WIKWIO portal is linked to a couple of free mobile applications available for Android and IOS systems. The application WIKWIO IDAO is dedicated to weed identification, while WIKWIO Citizen Science concerns the posting and browsing of observations in the field. After two years of work, the portal comprises a weed database of 350 species with a network of 466 reg-istered members involved in weed management in the Western Indian Ocean region. We expect to extend the area of interest of the portal to the Sub-Saharan African region with the collaboration of the FAO.

The Wikwio agricultural project aims to build a community of stakeholders including extension agents, students, farmers and scientists with a focus on weed species of food and cash crop systems in southern African region. Interactions between the citizen and scientific communities, facilitated by a technology platform is a promising approach. After almost three years in existence and steady growth, the platform hosts information on about 420 weed species with a community of around 650 members from 19 countries. A citizen science module on the Wikwio web platform enables people to contribute their observations about weeds. The module is interconnected with expert-curated species pages, spatial and documents modules that fosters interactions between citizen and scientific communities. These interactions improve the aggregation of scientifically validated data. For example, the citizen science module can inform scientists about the nature and level of occurrence of different weeds in different crop systems, their seasonality and distribution across the southern African region. In turn, the species pages module can inform extension agents and farmers about appropriate control measures for each species. Platform developers have built custom fields in the citizen science module to allow specific surveys to be commissioned by the scientific community. The comment stream in the same module can enable dialogue between the citizen and scientific communities allowing people to pose questions directly to experts to which experts can respond with additional details and clarifications. Active, wide and large-scale participation from diverse stakeholders in agricultural research and practice is building a vibrant Wikwio community around a facilitating technology platform and a shared scientific practice. (Texte integral)

We developed an interactive electronic weed identification tool, AFROweeds, and an online network, Weedsbook, for agricultural change agents to aid communication and offer assistance to rice farmers with specific weed problems. We collected quantitative and qualitative data to assess effectiveness and usefulness of these products with potential users. With the online version of AFROweeds, used on an electronic tablet, average weed identification time in the field was 7 min 6 s with 44% successful identifications. Poor mobile network coverage and slow internet were the main reasons for the relative long identification time and low success rate. A second trial was done with the offline version, pre-installed on a tablet. The average identification time was 6 min 34 s, with a success rate of 75%. The online network Weedsbook, established alongside AFROweeds, was assessed by the test users as a useful additional aid, enabling agricultural change agents and agronomists to exchange information or request assistance on all aspects of weeds and weed management. The potential improvements of both products are discussed.

Biodiversity information on weeds, covering species descriptions and distribution, is essential for understanding the ecology and distribution of weeds and their management. WIKWIO (Weed Identification and Knowledge in the Western Indian Ocean http://www.wikwio.org), a European Union African Caribbean Pacific Science & Technology II programme funded project, is focused on the Western Indian Ocean region to develop species information systems on weeds in food and cash crops. WIKWIO aims to assimilate weed knowledge through its participatory portal (http://portal.wikwio.org) to strengthen a science and technology network of stakeholders such as scientists, extension agents, and farmers in the African region. During the computer demonstration, this project's two major citizen science applications will be showcased. The graphically driven weed species identification system (http://www.wikwio.org/idao) helps users to identify about 345 weed species, without requiring prior knowledge in taxonomy. The citizen science application (http://portal.wikwio.org/observation/list/) facilitates uploading images of weed species, with a time and location. Both of these applications are available on web and mobile (https://play.google.com/store/apps/developer?id=Project+Wikwio) platforms. (Texte integral)

Invasive plant species pose significant biodiversity threats to protected areas, including South Africa's Kruger National Park (KNP). Habitat diversity and a river network draining invaded, exterior catchments make KNP highly susceptible to invasion. Efficient alien control requires early detection, effective eradication and increased awareness. Pl@ntInvasive-Kruger aimed to develop science-based, computer-driven tools for use by managers, researchers and teams involved in alien plant control. Three tools, supported by an online, multi-user database (DataManager), result: i. PUBLISH, which returns synthesised species information; ii. IDAO, a computer-aided plant identification platform; iii. IDENTIFY, an image recognition system. DataManager allows data control during field campaigns and facilitates collections management. Automatically synthesised data are available through PUBLISH as HTML pages, which detail descriptions and imagery of species and include information on ecology, biology and management, and support the identification tools. IDAO constructs unknown species in a step-wise manner based on characteristics selected from schematic, multiple-choice menus, and is compatible across multiple mobile electronic devices. IDENTIFY uses image recognition algorithms to guide the identification of images submitted to DataManager through a web interface. In both cases the suggested identity is expressed as the similarity of the unknown specimen to type specimen information in the database. Identification can be confirmed using the PUBLISH tool. All applications operate from a collaborative web platform (Pl@ntNet), where members also share information and documents and join discussions (http://community.plantnet-project.org/pg/groups/561/plntinvasivekruger/). Correct identification is difficult and time consuming where large numbers of alien and indigenous species co-occur. Pl@ntInvasive-Kruger currently houses information on 400 alien plant species, with the identification tools focussed on ~113 priority species. By assisting with the identification of invasive plant species and facilitating the sharing of information between stakeholders, Pl@ntInvasive-Kruger will promote biodiversity conservation in KNP. This project is a case study of the Pl@ntNet project funded by the Agropolis Foundation. (Resume d'auteur)

Increasing agricultural productivity faces challenges from multiple factors, significant of them being weed infestations. Weeds are responsible for a minimum of 20% loss of harvests in food and cash crops. Efficient control of weeds requires identification, effective control methods, and dissemination of current data to increase awareness among all actors. A collaborative project has been initiated to provide a participatory platform and a range of tools to support these actions. WIKWIO: Weed Identification and Knowledge in the Western Indian Ocean aims to develop a suite of science-based, computer-driven tools. A combination of applications will be developed: - The participatory portal from which 1) The public can consult species information, and upload observations; 3) Members can share and manage species data;- IDAO application which utilises computer-aided plant identification; This toolkit will facilitate the sharing of information from actors such as scientists, teachers, agronomists, extension staff, field workers and farmers and to disseminate compiled information to stakeholders in charge of the management of weeds. All tools can be used from the portal (http://portal.wikwio.org). WIKWIO covers 320 weeds of 14 food or cash crops of the area. Since no single actor or institution can study everything on weeds, the contribution of all to the knowledge of species is essential. By assisting with the identification of weeds and facilitatinf the sharing of information between interested parties, WIKWIO will promote sustainable agriculture and will contribute to capacity strengthening of local institutions. This project is funded by ACP S&T programme of the European Union. (Texte integral)

Objectives • To enrich the existing knowledge base (BIOTIK project*) on a major hot spot of biodiversity: the rain forests of Annamite Mountain range of Lao PDR, in the framework of the Inventory of biodiversity of forest canopies conducted in 2012 in Lao PDR under F. Halle scientific direction, and also to participate to broader initiatives such as Pl@ntNet project**. • To translate the identification tool in the Lao language in order to contribute to the capacity-building in plant taxonomy in the country. • To transfer the identification system to touch pads, enabling plant identification and entering data directly in the field.