Combined use of environmental and spectral variables with vegetation archives for large-scale modeling of grassland habitats

Abstract

Grassland habitats provide many ecosystem services but are threatened by agricultural intensification and urbanization. While the lack of accurate and comprehensive inventories at the national scale makes them difficult to manage, advances in spatial modeling using open remote sensing data and open-source software, as well as the increasing use of ecological archives, provide new perspectives for mapping natural habitats. In this context, this study evaluated the contribution of spectral and environmental variables to discriminate and then map grassland habitats throughout France. To this end, 92 spectral variables derived from moderate-resolution imaging spectroradiometer data, 19 bioclimatic variables derived from WorldClim data, 4 topographic variables derived from the European Union Digital Elevation Model (DEM), and 8 soil variables derived from SoilGrids data were combined at a spatial resolution of 250 m. Reference plots that characterized 6 and 21 grassland ecosystems at European Nature Information System (EUNIS) levels 2 (broad habitats) and 3 (habitats), respectively, were collected from vegetation archives. We first performed descriptive analysis that included habitat description, ordination, and pairwise separability. We then performed predictive analysis of grassland habitats using a cross-validated random forest model that included a spatial constraint. While environmental and spectral variables characterized most grassland habitats well and consistently, some confusion occurred between habitats with similar abiotic conditions. The main grassland habitat types were correctly mapped at EUNIS level 2 ( F1 score = 0.68), but not at EUNIS level 3 ( F1 score = 0.52). In addition, the two variables that contributed most to the model were the near-infrared spectral band in spring and the minimum temperature of the coldest month. The model’s prediction at EUNIS level 2 for mainland France provides the map of grassland habitats at a new spatial scale.

Large‐scale and fine‐grained mapping of heathland habitats using open‐source remote sensing data

“…(2018). Recently, the value of using SoilGrids data to map grassland (Rapinel et al., 2021) and forest (Agrillo et al., 2021) habitats was highlighted. Our study shows that these data can be relevant for identifying low‐carbon (ORCDRC) and acidic (PHIHOX) soils favorable to heathland habitats.…”

“…We used an OCC instead of a multiclass classifier to map heathland habitats without knowing the other habitats present in the landscape (Rapinel & Hubert‐Moy, 2021). Indeed, mapping of some habitats of interest using a multiclass classifier would have required masking areas where those habitats are expected to occur (Agrillo et al., 2021; Rapinel et al., 2021; Schindler et al., 2021). However, providing an accurate and comprehensive ‘heathland’ mask would have been challenging given (i) the one‐to‐many correspondences between heathland habitats and LULC classes (shrub, sward, forest, grassland, etc.)…”

“…In parallel, the EU Directive INSPIRE (Bartha & Kocsis, 2011) led to the diffusion of open‐access field databases that support calibration and validation of RS‐based classifications (Rocchini et al., 2017), especially at a national scale, for which the cost of collecting field data would be too high. Although the quality of open data is considered heterogeneous, whether derived from RS (Moudrý et al., 2018) or field data (Suarez‐Seoane et al., 2020), open data have recently been used to map grasslands (Rapinel et al., 2021) and forest habitats (Agrillo et al., 2021) at a national scale. However, these two studies used a multi‐classifier, which requires first masking the area of interest (e.g.…”

Prediction of Grassland Biodiversity Using Measures of Spectral Variance: A Meta-Analytical Review

“…While examining the literature on the Spectral Variation Hypothesis, we noted emerging approaches that expand on the traditional definition, which relates to the spectral variation in space. For example, some authors have looked at the spectral variance of a pixel or cluster of pixels over time [109][110][111]. This is based on the idea that plant species or community-specific responses to temperature, rainfall, day length and soil conditions can be exploited for diversity estimations.…”