A geospatial approach to assess habitat diversity and inform conservation efforts using a case study from North Africa

Abstract

AbstractGeographic Information System (GIS) combined with remote sensing and field data was used to identify and map Souss‐Massa National Park (SMNP) habitats, an exceptionally biodiverse coastal area located in Morocco. Habitats in fenced and unfenced areas were compared to assess the fencing effectiveness for conserving semi‐arid/arid environments. Sentinel‐2A level‐2A data acquired on March 25, 2021, were analysed using the Semi‐Automatic Classification Plugin in QGIS software. The supervised classification technique and the maximum likelihood classifier were used. Ground truthing consisted of training data used for the supervised classification and reference data (850 validation points) used for accuracy assessment based on the confusion matrix method. SMNP composed of 11 natural habitat classes (70.91%) and 6 disturbed habitat classes (29.09%). The overall accuracy (>96%) and kappa coefficient (0.95) showed a high classification accuracy. Fenced areas were dominated by natural habitats (>88%), whereas unfenced areas contained more disturbed habitats (>51%), indicating that fencing was an effective conservation tool. Natural habitats in unfenced areas are largely being modified and degraded, and therefore, monitoring and conservation actions should be further strengthened. The presented methodology shows a promising application potential in future and related ecological studies conducted in other regions, notably developing countries.