Application of spectral diversity index for rangeland degradation assessment in North Khorasan province
This study aimed to use a remotely-sensed spectral diversity index called Moving Standard Deviation Index (MSDI) for identifying and assessment of degraded and non-degraded rangelands based on landscape heterogeneity. According to the utilization of rangelands, three sites including reference site (Sarygol National Park), medium grazed site, and over-grazed site with a sampling area of 100 pixels and with similar vegetation, soil type and landform were selected in North Khorasan province. The MSDI index was derived by passing a 3×3 moving standard deviation filter across the 2009 Landsat TM band 3 (red reflectance). To assess the performance of MSDI technique, the widely-used normalized difference vegetation index (NDVI) was also calculated using band 3(red reflectance) and band 4 (near-infrared reflectance). Results showed that over-grazed rangeland had higher mean MSDI and lower NDVI values than medium-grazed and reference site. MSDI difference value was highest in over-grazed rangeland (11), followed by medium-grazed (8) and reference site (6). The results for NDVI index were 0.053, 0.066, and 0.054, respectively. According to T-test statistical analysis, MSDI showed significant differences (p<0.001) between all rangeland sites but NDVI was unable to separate medium-grazed and over-grazed rangelands (p>0.01) due to soil background effects. The results of this study indicate that the MSDI in contrast with NDVI is more sensitive to rangeland landscape degradation and heterogeneity; therefore, it can be used as an appropriate method for assessment of rangeland degradation in Iran.