SEARCH

SEARCH BY CITATION

Keywords:

  • Arthropod;
  • floodplain;
  • Loire River;
  • meadow;
  • remote sensing;
  • surrogate taxa

Abstract

  1. Remotely sensed data are frequently employed for monitoring vegetation and for estimating herbivore diversity. Their use for predicting predator arthropod species abundance and richness has also been investigated with success for ants and beetles in forests using normalised difference vegetation index (NDVI) and for beetles in mountain forests using light detection and ranging data.
  2. We investigated whether vegetation indices, derived from multispectral SPOT imagery could predict abundance and species richness of ground active spiders and ground beetles in a new ecological context, the floodplain meadows of the Loire River in Western Europe. Using pitfall traps, we collected carabids and spiders in the field.
  3. Maximum vegetation height, litter-depth and plant species richness best explained species assemblages of both groups (multivariate analyses). NDVI and enhanced vegetation index (EVI 2) were strongly related to activity-density and species richness for ground beetles only, EVI 2 being the best surrogate. Relationships between vegetation indices and spider assemblage patterns were either non-significant or weak.
  4. We demonstrated that EVI 2 is a good surrogate of the abundance and richness of carabid species in a temperate floodplain, and has potential as a low cost method for mapping arthropod assemblages at large spatial scales.
  5. Our approach provides a tool which contributes to biodiversity assessment at large spatial scales. It can also contribute to the prioritisation of conservation areas and early change detection, as carabids are keystone indicators.