Published July 5, 2023 | Version 1
Dataset Open

Presence of perennial halophilous scrubs in 2016 in the former saltworks of Salins de Giraud, random forest classification on WV2

Creators

  • 1. Lammi Biological Station, Department of Forest Sciences, University of Helsinki, Pääjärventie 320, 16900, Lammi, Finland and University of Angers, France.
  • 2. Department of Biology, FI-20014, University of Turku, Finland
  • 3. L@bisen, Institut Supérieur de l'Électronique et du Numérique (ISEN), France
  • 4. Centre for Marine and Environmental Research (CIMA), University of Algarve, Campus of Gambelas, Faro, Portugal
  • 5. Tour du Valat
  • 6. Univ Rennes, CNRS, Géosciences Rennes, UMR 6118, 35000 Rennes, France
  • 7. CIMA- Center for Marine and Environmental Research, CTMA- Department of Earth, Environmental and Marine Sciences, Gambelas Campus, University of Algarve 8005-139 Faro, Portugal
  • 8. Univ Angers, Nantes Université, Le Mans Univ, CNRS, LPG, F-49000 Angers, France
  • 9. LETG UMR CNRS 6554, University of Nantes, CEDEX 3, 44312 Nantes, France

Description

The presence of perennial halophilous scrubs (S. fruticosa  and A. macrostachyum) was mapped using a Worldview 2 (WV2) very-high resolution imagery acquired the 31st of August 2016, strictly cloudless. Multispectral indices, among the most used, were adapted to the bands of WV2. The Random Forest model (RF) package into the R software was used to perform image classification. RF OBB and omission errors for the class “presence” and the class “absence” were less important when the classification was done for each species individually, hence this raster is a combination of both classifications. The OBB error for S.fruticosa is 1.43% and for A. macrostachyum is 2.18%.

Files

halophilous_scrubs.tif

Files (41.1 MB)

Name Size Download all
md5:e8db29cba805f6cd0098f934aa0c68a2
41.1 MB Preview Download