The Significant Roles of Landscape Pattern and Spatial Scale in Assessing Aeolian Desertification

Kang, W. (Kangwon National University, Chuncheon, South Korea);

Kang, S. (Kangwon National University, Chuncheon, South Korea)

One of research focus in combating desertification is monitoring and assessing aeolian desertification by temporal satellite remote sensing imageries. Landscape heterogeneity hampers robust interpretation from the satellite imagery on the status and process of aoelian desertification in arid regions, for which scale effect on satellite-imagery interpretation must be investigated. In this study, area and degree of aeolian desertification were assessed at different spatial scales from 30m to 1km by using TM and Landsat8 imageries, respectively. Then, landscape pattern index was utilized to explain effect of spatial resolution on the satellite-based assessment of aeolian desertification in the middle of Mu Us and the west of Otindag Sandy Lands in Inner Mongolia of China. Our results indicate: (1) The middle of Mu Us was rehabilitated from 2000 to 2014 by reduction of desertified area from 12,861 km2 to 11,700 km2. Whereas, the desertified area of the west of Otindag fluctuated but rebounded during the last five years; (2) scale effects on the area and degree of desertified land were not significant, along the spatial resolution, the desertified area fluctuated from12,962km2 to 12,861km2 for Mu Us hinterland and from 7,752km2to 7,700km2 for Otindag west, which corresponds 0.78% and 0.67 % of relative variations, respectively. The aeolian desertification degree index(ADI) fluctuated from 0.36 to 0.35 for Mu Us hinterland and from 0.17 to 0.16 for Otindag west, which corresponds 2.78% and 5.88 % of relative variations, respectively. (3) scale effects on landscape pattern indices were significant, particular in landscape fragment; the correlation of spatial resolution and landscape fragment are positive(RMu Us2=0.76-0.77, ROtindag2=0.73-0.78). It was found substantial uncertainty in satellite-based assessment on aeolian desertification in Mu Us and Otindag Sandy Lands and the relevance of the uncertainty with the landscape pattern indices. Our study proposes more attentions on relations between landscape patterns and spatial resolution of satellite imagery in assessing aeolian desertification.

Publication: American Geophysical Union, Fall Meeting 2015, abstract id. NH43A-1868 

Pub Date: December 2015 

Bibcode: 2015AGUFMNH43A1868K

Author: Willem Van Cotthem

Honorary Professor of Botany, University of Ghent (Belgium). Scientific Consultant for Desertification and Sustainable Development.