Investigating and Predicting the Extension of Dunes Using Land Change Modeler (LCM) in the North West of Yazd, Iran

Abstract

In the present study, in order to calculate the movement of sand dunes in the period between 2001 and 2010, the ASTER (Advanced Space-borne Thermal Emission and Reflection Radiometer) images were used. The training samples were obtained from the field, and the images of the years 2001 and 2010 were classified using maximum likelihood algorithm and decision making tree. The study area was classified into four classes, including vegetated areas, urban areas, sand dunes, and bare lands. Accuracy of the created land cover maps was assessed using five hundred ground control points, Google Earth, and Landsat 7 satellite images. Based on the results, the overall accuracy and kappa coefficient for the maps of 2001 and 2010 were estimated 96%, 0.94 and 98%, 0.98, respectively. In the next step, the resulted land cover maps were used as input for Land Change Modeler (LCM) and Markov calculation. Statistical outputs and change map of this model show the extensive changes in the dunes during the study period. In addition, the area of sand dunes increased from 12,103 to 14,355 hectares. However, 6 hectares of vegetated areas around the Ashkezar and Zarch cities in the North West of Yazd changed to sand dunes. The highest probability of change in the dunes is the change of dunes to bare lands and the lowest probability is the change of dunes to vegetation lands. These results show the movement and change of the dunes in this arid zone.

Keywords

• aster image
• dune
• land change modeler
• markov
• prediction