Document Type : Research Paper

Authors

• Fatemeh Hadian ¹
• Hossein Bashari ²
• Reza Jafari ³
• Saeed Soltani ⁴

¹ Master of Science in Range Management, Department of Range Management, Isfahan University of Technology

² Professor, Department of Natural Resources, Isfahan University of Technology, Isfahan

³ Assistant Professor, Department of Natural Resources, Isfahan University of Technology, Isfahan

⁴ Associate Professor, Department of Natural Resources, Isfahan University of Technology, Isfahan

Abstract

Selecting the appropriate satellite images is highly important to achieve acceptable precision and accuracy in rangeland assessment programs. This study aimed to evaluate Landsat5 (TM sensor) and IRS-P6 (AWiFS sensor) satellite images in three rangeland vegetation types in Semirom region, Isfahan. Ten vegetation indices of different groups including slope-based, distance-based and plant-water sensitive indices were calculated and evaluated for all vegetation types. The percentage of canopy cover, litter, rock, gravel, stone and bare ground were determined using step-point method in radial direction (6000 points per rangeland type). Then, the correlations between the measured field components and spectral indices were compared. According to the results, vegetation indices extracted from TM sensor data had higher correlation with field vegetation cover due to its high spatial resolution. Factors such as characteristics of study area, range condition and vegetation types had also vital role in these correlations. The indices were tested against vegetation cover percentage in all vegetation types including Bromus tomentellus, Astragalus spp-Daphne mucronata- and Astragalus spp-Scariola orientalis. The highest determination coefficient was found between live vegetation cover and SSI index (r²=0.85) in Astragalus spp-Daphne mucronata vegetation type. Overall, the results showed that there was an inverse relationship between vegetation cover indices and rangeland condition. TM vegetation indices had minimum of 24 percent of determination coefficient in very poor range condition in comparison with other studied rangeland conditions. Therefore, the performance of a vegetation index highly depends on rangeland condition, vegetation types and also spatial resolution of remote sensing data.