GIS and RS soil-vegetation correlations for continental salt-lands habitats in NE Romania

Abstract

Continental saltlands have a high degree of peculiarity amongst European primary habitats and a prominent insular character. The present scientific approach establishes the degree of soil-vegetation correlation in continental slatlands patches as a measure of habitat continuity/fragmentation and soil conservation/degradation. The use of hyperspectral imagery, soil types’ distribution and vegetal associations’ conservation status reveal disturbances in relation with human induced modifications in comparison with normal plant-soil interdependence. Supervised classifications of LANDSAT satellite imagery along with detailed soil maps, ground truth data provided by accurate GPS positioning and field based plants evaluation are used to perform landscape metrics analyses. The landscape metrics approach is meant to find the balance between extent and grain in the case of saltlands habitats analyses and the degree of patches and classes inhomogeneity. These also give an insight of habitats connectivity and/or isolation in relation with land use topology and soil multiplexing. The resulting training sets developed for a representative, protected area in the county of Iaşi enhance the creation of a comprehensive mask to be used for the evaluation of larger areas in the silvan-steppes of North-Eastern Romania. The model is statistically tested to depict the degree of correlation and confidence. The final goal resides in more proper measures elaboration for the mitigation of continental saltland preservation and natural resources exploitation via agricultural and the associated activities.

Keywords

• land-cover
• LANDSAT
• automatic classification
• soil – vegetation
• correlation