The Urgent Needs of Remotely Sensed Geographic Information for the Biodiversity Research Community
Rousseau, Corentin; Radoux, Julien; Defourny, Pierre
Université catholique de Louvain, BELGIUM

These last two decades, two new paradigms appeared and took huge importance in research: biodiversity decline and ecosystems services. The research about it increased substantially, but these studies suffer from several geographical biases or lack: most of them occur within ecosystems or at the landscape level. The overrepresentation of certain ecosystems is also well known including protected areas, temperate deciduous woodlands, and wealthy countries. Moreover, there is a lack of data about the repartition of ecosystems at large scales. To make up for it, new geographic information need to be made available to the research community.

The European Infrastructure for biodiversity and ecosystem research, "LifeWatch", is a research infrastructure in development. It aims at providing datasets, tools and methods to the biodiversity research community ranging from molecular to landscape biodiversity levels. In this framework, Wallonia-Brussel is funding a project to examine the potential of remote sensing for the consistent description of biotopes at the scale of Europe (EU-38). Biotopes are considered here as an area of uniform environmental conditions providing a living place for a specific biological community.

The needs of the biodiversity research communities in pan-European remotely sensed geographic information have been review . A formal process to collect requirements in two steps has been done.

  • The first step is a review of scientific literature. The key words used for it have been "biodiversity", "remote sensing", "mapping", "ecosystem" and "habitat" during the period 2007-2012.
  • The second step has been a key user requirements consultation for the product specifications. The aim was to provide a detailed user needs assessment and identify specific requirements. Two main mechanisms have been carried out: different meeting participations and direct contact of key users and authors of interesting publications noticed in the first step. Also scientists in important labs working in landscape ecology, biogeography, dispersal, macroecology and population dynamics have been contacted. In total, 54 research groups have been reached.

    The requirements of the biodiversity science community for data about geographic information are broad, the needs are huge. Methods must be developed to fulfill it and they must be consistent and repeatable. Moreover, the requirements can be divided in three groups. The first requirement is the setting of data of trends of biotopes status over time. Also a good diagnostic of the (potential) threats and disturbance sources (some of which may be useful to the biotopes persistence) may be of high interest. The second requirement is to have data for all territory and not just for natural areas: urban areas, including anthropogenic infrastructures, land use including crops etc. must be included. The last requirement is maps of habitats which can be used for series of researches as species distribution, dispersal, ecosystems services, climate change, highlight biodiversity hotspot, etc. These maps must have a legend custom-made for biodiversity (e.g. EUNIS),and with the highest possible thematic precision in terms of land use and plant communities.

    Different sets of global or continental land-cover exist: Global Land Cover for the year 2000 (GLC 2000), the International Geosphere-Biosphere Programme (IGBP), GMES Land Monitoring products, European Corine Land Cover (CLC), etc. But there are not much used by the biodiversity community. Indeed, the classifications of these products are not useful for them. Moreover the changes in land cover are very difficult to highlight.
    To improve the assessment of the biodiversity, future research should focus also on the process of integration different kind of data, from different sensors but also ancillary data from other projects that collect biotopes characteristics (climate, topography, soils, etc.).

    Based on these requirement and our currect state of knowledge about existing products and remote sensing data (including near-future operational missions such as the Sentinels), a realistic custom-made database has been designed. The attributes of this geographic database have been formalized in UML together with tools to derive classifications for the visualisation of the databse. These attributes include distinct information about the land cover and the land use, as well as additional information including soil types and topography. The static and dynamic characteristics of the land cover are obtained based on high resolution data and medium resolution time series, respectively.