Project Synopsis

‘Mapping the Vegetation of Madagascar’ is a three year project (2003-2006), funded by the Critical Ecosystem Partnership Fund (CEPF) and managed jointly by The Royal Botanic Gardens, Kew (RBG Kew), Missouri Botanical Garden (MOBOT) and Conservation International’s Center for Applied Biodiversity Science (CABS).

The most widely used map of primary vegetation for Madagascar is that of Du Puy and Moat (1996), produced by the Royal Botanic Gardens, Kew (RBG Kew). This map, and in fact the majority of those currently in use in Madagascar, are based on Dr Faramalala’s earlier vegetation map, which in turn was derived from satellite imagery from the late 1970s. The original information from which these maps were derived is wholly out of date, and it has been estimated that 20-30% of the primary vegetation shown has disappeared since. Moreover, these maps classify vegetation using broad categories, the validity of which has often been called into question in the conservation context. An accurate and updated vegetation map is thus imperative for conservation planning and natural resource management in Madagascar. It is also essential that the data on which such a map is based be made freely available, so that conservation organisations, Government departments, academic institutions and other stakeholders can use them as an up to date standard dataset on which to base their activities. In order for a vegetation map to fulfil its intended role it must a) accurately delimit areas with various vegetation types as they currently exist, and b) assign those areas to objective categories that can be easily recognized in the field and that reliably reflect fundamental biological differences (primarily structural features, i.e. physiognomy). This project aims to produce just such a map.

The precedent for the vegetation mapping methodology to be adopted for this project is that used by Frank White in developing his highly regarded vegetation map of Africa . Using remotely sensed and ground data, and a consultative process that drew on a wealth of regional, national and local knowledge, White produced a number of iterations of his map, each more accurate and comprehensive than the last. His map took more than 20 years to produce but thanks to the huge advances in remote sensing and communication technologies over the past two decades, ours will take a fraction of that time.

The recent deforestation map of Madagascar produced by CABS presents us with a useful basis for developing a new vegetation map that accurately depicts the current extent of all major types of primary vegetation throughout the country. This map will be refined into an up to date map based on a completely new classification of Madagascar’s vegetation. The first stratification will be applied by CABS using data from the MODIS sensor on board NASA’s Terra satellite, and this will result in the production of a preliminary vegetation polygon map. MODIS provides daily observations in the visible and thermal spectral regions at 0.5 to 1 km resolutions. These provide estimates of canopy cover, Leaf Area Index and surface temperature. Weekly to monthly ‘composites’ are generated from the daily data to exclude clouds. These composites are then used to provide summary data on seasonal bioclimatology, which provide valuable information about the probable vegetation types. The existing Landsat-based map of forest cover will then be used as a “cookie cutter” to produce finer resolution for the forest-vs-nonforest stratification. This preliminary map will be refined in two ways: through extensive ground survey, and through further interpretation of recent Landsat data already acquired by CABS and RBG Kew.

The preliminary stratification produced by CABS will be followed up with a first technical workshop in Madagascar, where a physiognomic classification of Madagascar’s vegetation will be developed by botanists and other biologists with extensive field experience. Critical regions will also be highlighted where additional fieldwork and/or satellite imagery are needed. This will be a fully integrated, consultative process, involving extensive Malagasy and international botanical expertise. For the first time ever in Madagascar, remotely sensed information will be verified through extensive ground truthing. The large network of field botanists currently active in the country (including those employed by MBG, RBG Kew and CI) will be supplied with a key to recognize the proposed physiognomic types, and will be asked to verify them and suggest possible refinements. In addition, they will be asked to apply their knowledge of the flora to identify key, indicator species in each habitat. The aim will be to produce a simple classification, based on vegetation structure and floristic indicators, which is easy to apply in the field, even by non-specialists. To this end, a standard data form will be designed and used to record vegetation characteristics and note the geo-position using a GPS at field sites visited. A media launch of the project, to which members of the press and representatives from Madagascar’s conservation community will be invited, will be staged at the end of the workshop.

The information generated in the field will be compiled by a central unit based in Madagascar. This will consist of core staff from the Kew and MBG offices. Their task will be to inform people about the project, organise consultative workshops, interact with people going into the field, and collate the resulting field data. These will be made available to the Remote Sensing team at CABS and the Geographical Information System team at RBG Kew, who will use them to update and refine the classification, which in turn will be supplied back to the field teams.
The mapping activities essential to this project will be developed by a Malagasy Research Fellow based at RBG Kew’s GIS Unit. Ideally this Research Fellow will be drawn from a relevant Government organisation (e.g. ANGAP, MEF) or a conservation NGO in Madagascar to make sure that the skills gained return to Madagascar.

A strong feature of this project is the fact that much of the fieldwork will be carried out by botanists going about their normal business. This incremental element builds upon and adds significant value to existing botanical programmes in Madagascar.

Following the collection and collation of the field data, a second technical workshop will be held in Madagascar, involving a focus group of experts, whose task it will be to produce a working vegetation map, based on the synthesis of satellite data and the information collected in the field. This workshop will also identify problem areas and sites with insufficient data. An open day and demonstration of the working map and classification will be staged at the end of the workshop. Staff from conservation planning organisations in Madagascar will be invited to attend this session.
Using the information from this workshop, a volunteer-based project team will be formed, comprising botanists and professionals from participating organisations, who will visit selected sites, including areas that present special problems, and sites which complement the areas visited by the network of field botanists. Special attention may need to be given to poorly known areas, and to areas possessing a wide range of vegetation types, which will allow more detailed ground truthing by the project team.

At all stages of the compilation process, satellite and map data will be made available through the Internet, both as GIS layers and also as general maps which users can view. A website and email list server dedicated solely to the vegetation mapping project will be established on the web to encourage and facilitate input from a wide range of contributors

A final technical workshop, involving all the contributors to the process, will be held in 2005, in Madagascar, to refine the physiognomic vegetation classification for Madagascar, and to assign vegetation types to the interpreted satellite imagery.

The final product, a user-oriented, simple to apply, up to date, accurate vegetation map at a scale of 1:100,000 to 1:250,000 will be presented to the user community in hard copy and digital formats at a meeting to be held in Madagascar. All potential users will be invited, including conservation organisations, Government departments, academic institutions, private companies, etc. The finished map will be presented, and its application demonstrated in training sessions covering fields such as conservation and natural resource planning, environmental impact assessment etc. Publication details will be made available to participants and the press, and the map will be made available to users through the Madagascar Biodiversity Network. The website dedicated to the project will be maintained by RBG Kew after the project is finished so that contributions and updates can continue to be made, and to facilitate the publication of further editions of the map in the future.

The new vegetation map of Madagascar will complement the comprehensive analysis of plant species distribution patterns being conducted under the CEPF-funded Priority Areas of Plant Conservation project, which will produce the country’s first phytochorological map depicting areas with shared floristic composition. Following the successful model developed by White for continental Africa, it will then be possible to superimpose the vegetation and chorological maps to circumscribe areas that contain a single major vegetation type (e.g., low elevation humid evergreen forests) with a common set of constituent species, which represent the two most important elements for a landscape scale mapping of plant diversity.

This proposal has been developed to support the aims of the Malagasy Government’s Environmental Action Plan, and specifically Programme Environnemental 2, with emphasis on gathering data from the field, and Programme Environnemental 3, using that data to manage the environment. In addition, this project perfectly encapsulates the CEPF investment strategy. It represents an all-inclusive collaboration between specialists from botanical institutions, conservationists from international and local NGOs, and land use/natural resource managers. The major outputs of this project are essential tools for assessing, monitoring and managing biological diversity both within and outside the existing protected areas, and one that environmental managers cannot afford to be without.

Major project outputs
1. Digital and hard copy versions of a vegetation map designed and made publicly available for conservation, scientific research and natural resource management purposes.
2. Delivery of all Landsat and MODIS products, all co-registered, to Madagascar conservation-based collaborators, researchers and other stakeholders.
3. A revised vegetation classification scheme for Madagascar, developed, published and made accessible to non-specialists through the Madagascar Biodiversity Network.
4. Malagasy personnel trained in the use of remote sensing and GIS for conservation purposes.
5. A network of botanists, conservationists and other stakeholders working in collaboration throughout Madagascar