Multi‐resolution dune morphology using Shuttle Radar Topography Mission (SRTM) and dune mobility from fuzzy inference systems using SRTM and altimetric data

Abstract

The complex morphology of large sand dunes of the world's great deserts have significant importance on conservation and climate change and hence are of interest to a wide variety of scientific and environmental applications including studies on aeolian processes, paleoclimate, civilian infrastructure management, and design of blown‐sand control systems. Scientific studies on dune formation and dynamics have been limited to desert margins due to inaccessibility of the desert interior by conventional surveying and mapping techniques. Thus, dune morphology in the deep desert interiors is not well studied and much about the driving forces controlling dune activity and dynamics are still poorly understood.

We demonstrate the utility of space‐based observations to characterize dune morphology. Specifically, we used the Shuttle Radar Topography Mission (SRTM) C‐band data to investigate and compare morphologic attributes of dune fields of the Taklimakan and the Namib Deserts. Cross‐sectional amplitude roughness estimates have similar magnitude but stoss slopes are shallower in the Taklimakan Desert than in the Namib Desert. The high Height‐Width (H:W) ratio of 0.09 for the linear dunes in the Taklimakan Desert is indicative of its equilibrium with aeolian shear stress whereas the Namib Desert dunes are unstable. Multi‐resolution planimetric properties from SRTM Digital Elevation Model (DEM) using B‐spline wavelet decomposition reveal linear dunes in the Taklimakan Desert are superimposed on a dome‐like substrate whereas the linear dune in the Namib Desert are constructed on a knoll‐like submorphology. These long‐wavelength features, taken as paleotopography, may be a major controlling factor on wind patterns and dune sinuosity.

We demonstrate the utility of ANFIS to assess seasonal dune changes in the Namib Desert using ICESat observations and SRTM. ANFIS is a data‐driven prediction scheme. Predicted topography along ICESat tracks have low rms of 3.5m, a 30% increase over SRTM accuracy. Seasonal track comparison from August 2003 to January 2005 shows that most changes to dune topography occur at crestal deformation of isolated dunes. The results show seasonal waning and waxing of dune crests.

Acknowledgments

ICESat data were provided by the National Snow and Ice Data Center (NSIDC), Boulder, CO and processed by NASA Goddard Space Flight Center and the ICESat project. We thank Rainer Mautz for providing Matlab code for wavelets. We thank Dr. Juan Crespo and an anonymous reviewer for their comments.