The Cartovege project proposes to develop a decision support tool for the conservation of flora and the preservation of habitats on the Crozet and Kerguelen archipelagos, by combining vegetation mapping and predictive modeling of changes that may affect it via satellite and field data.


After defining a vegetation classification for the French sub-Antarctic islands, we will combine remote sensing vegetation mapping approaches with modeling of potential plant species occurrence. The mapping of vegetation and habitats, and their future trajectories, will be used to support the management of these islands.

Temperatures are increasing at an unprecedented rate in many parts of the world, particularly in alpine and polar regions. These changes are inducing a redistribution of species and modifications of natural and semi-natural habitats, with impacts on communities and ecosystem services. In addition, the characteristics of many habitats are being altered by anthropogenic activities and invasive alien plant and animal species. By reshaping species distribution and habitat composition and functionality, these changes affect biodiversity health and ecosystem functioning. At the level of plant communities, the processes of homogenization of biocenoses and loss of native species that result from these changes are also amplified by the incursion of exotic species. Faced with these challenges, it is urgent to have maps of biodiversity at different scales, associated with a modeling of the various factors governing the distribution of this diversity in time and space. The acquisition of this knowledge is all the more important for the conservation and management of protected areas in France, such as nature reserves.

The Crozet and Kerguelen archipelagos, included in the French Southern Territories National Nature Reserve (RNN TAF), are of particular interest because of their original landscapes (Figure 1) and the number of endemic species with high heritage value that they host. These characteristics led to the classification of these islands as UNESCO World Heritage Sites in 2019. Since the mid-1950s, these islands have undergone significant climatic changes, marked by an increase in annual temperature, a reduction in the number of days of winter frost, and increasingly frequent water deficits (rainfall) during the summer periods. These changes, coupled with anthropogenic disturbances, have facilitated the establishment and expansion of non-native plants whose proliferation amplifies the threats to native biodiversity.

Figure 1: (gauche) Complexe de végétation de la Vallée des Branloires (Ile de la Possession , IIes Crozet) (droite) Complexe de végétation au pied du Doigt de Sainte Anne (Iles Kerguelen) - (c) R. Poncet (UMS PatriNat).

Figure 1: (left) Vegetation complex in the Vallée des Branloires (Ile de la Possession, Crozet Islands) (right) Vegetation complex at the foot of the Doigt de Sainte Anne (Kerguelen Islands) ©  R. Poncet (UMS PatriNat)

In this context, the conservation of the flora and the preservation of the habitats constitute major challenges for the manager of the RNN TAF, who must have relevant decision-making tools. Knowledge of the spatial distribution of species, coupled with information on the nature of the environments they prefer to occupy (pedology, topography, local climatology) is thus indispensable. Such synoptic information on plant species and threatened habitats remains fragmentary at the scale of the RNN territories, for which there is no complete mapping of the vegetation, although they are home to many rare and threatened species and habitats.

Machine-learning map modeling based on satellite remote sensing data, coupled with field surveys, is the most efficient way to map vegetation on a large scale and to apply homogeneous and reproducible methodologies that allow for cost effective updates that are temporally and spatially comparable. Methodologies that rely on supervised machine learning image classification algorithms provide highly reliable results. In order to fully exploit the potential of remote sensing in a decision support tool, it is crucial to add a predictive dimension, by correlating large-scale time series of species distribution to the biophysical characteristics of the environment they occupy.

In this project, we propose to develop a cartographic decision support tool, producing information at different spatio-temporal scales and combining remote sensing mapping of current or past vegetation and modeling of potential occurrence of key plant species (Figure 2). This tool will improve the effectiveness of actions implemented within the perimeter of the RNN TAF, in particular the second management plan of the RNN TAF (2018-2027). In a first approach, we will define a classification of the vegetation of these islands. This classification (or typology) is a necessary prerequisite for the classification of satellite images but will also be a support for information on the distribution of species, especially those that should be subject to management measures. In a second step, we will define the potential distribution areas of plant species, and particularly of emblematic species (with high conservation stakes) of the French subantarctic islands.

Figure 2 : Cartographie de la végétation de trois îles du Golfe du Morbihan (Iles Kerguelen, Terres Australes et Antarctiques Françaises)

Figure 2 : Vegetation map of three islands in the Gulf of Morbihan (Kerguelen Islands, French Southern and Antarctic Territories) © D. Fourcy - INRAE

The multiscale decision support tool (different spatial, temporal and biological scales) that we propose will be based on our floristic monitoring of the Crozet and Kerguelen islands in the field, on different satellite or aerial imagery, on an adequate typology of the vegetation and on image analysis and predictive modeling methodologies. This approach, transferable to other territories, is new and will allow to make reliable the way areas will be prioritized during the implementation of management plans. This project is also a unique opportunity to increase our understanding of the interactive effects of climate change and alien species on biodiversity dynamics. This tool will offer two types of complementary outputs concerning vegetation and habitats: (a) on the one hand, an inventory carried out according to a reproducible methodology and with confidence metrics on the results, and on the other hand (b) a simulation of potential changes under various scenarios. This simulation will be fed by the previous inventory, and will be evaluated a posteriori by a subsequent inventory. This dual descriptive and predictive capacity, combined with a capacity for self-assessment, is what makes such a tool so powerful for managers of protected areas. This tool will provide significant advances in terms of redefining the parameters of predictive models aimed at assessing the future impacts of climate change on plant communities.


French Southern and Antarctic Lands: Crozet Islands and Kerguelen Islands



  • Satellite data and imagery with high spatial and/or spectral resolution (Pleiades, Spot 6/7, WorldView, QuickBird, ALOS, etc.).
  • Radar data (Sentinel 1): Approach to try to convert these data into indices on vegetation and environments (NDVI, NDWI, etc.) in order to have regular monitoring data covering all territories.
  • Dinamis (Data Terra) and Copernicus infrastructures.


  • Meteorological data

            - Time series of in-situ microclimatic data (SoilTemp

            - High resolution climate data (current and future) (French Polar Institute project IPEV136, WorldClim, MerraClim, CHELSA, TerraClimate, Météo-France)

  • Statistical tools

           - Species distribution models: correlative, hybrid and mechanistic models (MaxEnt, RandomForest...)

           - Statistical interpolation (GWR 14 model) and topoclimate prediction.

  • Field data

           - Habitats-Flora-Invertebrates database co-managed by the RNN TAF and the SUBANTECO IPEV/136 project; this database compiles a description of the habitats, and geo-referenced data in time and space of the flora and invertebrates of the Crozet and Kerguelen Islands. These surveys will be completed during the project to reinforce the machine learning datasets: typological refinements, expertise on already available data, acquisition of material adapted to modeling, validation sampling, etc.

  • Data from the literature

           - Knowledge from the literature (biological invasions, autoecology and synecology of native and non-native species, phytosociology).


The project will produce the following final deliverables / products / actions:

  • Banking of data on the distribution (georeferenced plants) in the Terres Australes National Nature Reserve (RNN TAF), satellite images and microclimatic data used to build the habitat and vegetation mapping,
  • Production of a first typology of plant formations in the French Southern Territories (matching of available naturalist data with the habitat typology),
  • Mapping by modeling of the habitats and vegetation of the RNN TAF, based on the available material (naturalist data and explanatory variables),
  • Comparative mapping of habitat modeling of the Crozet Archipelago islands: Ile de l'Est and Ile aux Cochons, both in strict reserve, vs. Ile de la Possession in order to evaluate the effect of climate change in the absence of anthropogenic disturbance,
  • Integration of habitat typology information into the HABREF database, revision of learning materials and explanatory variables,
  • Feeding, with the habitat distribution models we will produce, the INPN knowledge base on habitat distribution knowledge,
  • Prediction of plant distribution changes in the RNN TAF according to different scenarios; transmission of information to the RNN TAF for an optimized management of the biodiversity of these islands,
  • Production of a multi-purpose repository for the classification of plant formations in the TAF National Park Reserve.


  • Project of the French Polar Institute Paul-Emile Victor IPEV 136 'SUBANTECO' (Subantarctic biodiversity, effects of climate change and biological invasions on the terrestrial biota', 2018-2021, then 2022-2025)
  • H2020 BiodivERsa call 2019-2020 Biodiversity and Climate Change 'ASICS' project (ASsessing and mitigating the effects of climate change and biological Invasions on the spatial redistribution of biodiversity in Cold environmentS)
  • SoilTemp: a global database of soil microclimate and soil surface data
  • InEE-CNRS Antarctic and Southern Territories Workshop Zone: long-term monitoring of Antarctic and sub-Antarctic biodiversity and ecosystems.



By ticking this box, I acknowledge that I have read and accept the Legal Notice.