September 2013 Vol. 1 No. 2

Full Length Research Paper

Assessment of vegetation indices for estimating plant coverage and plant density in the Northern Sarawat Mountains, Saudi Arabia

*^1,2K. H. Dewidar and ²B. Al Rehili

¹Department of Environmental Science, Faculty of Science, Damietta University, New Damietta City, Box 34517, Egypt.
²Department of Geography, Faculty of Social Science, Imam Mohamed Ben Saud Islamic University, B.O. 5701, Al Riyadh, 11432.

*Corresponding Author E-mail:khdewidar@yahoo.com

Accepted September 06, 2013

Abstract

Vegetation indices from remote sensing data have been widely used for more than three decades for the quantitative assessment of the biophysical characteristics of plants. Because the Sarawat Mountains are dry and semi-dry, many vegetation indices are less sensitive to plants but highly sensitive to the soil in this region. This study evaluated various vegetation indices (22) that estimate plant coverage and plant density and compared the indices to the method of hybrid classification in order to identify the best vegetation indices for estimating plant coverage in the region. The study found exaggerated values in the vegetation indices MNDVI, TSAVI2, WDRVI, PVI1, IPVI, NDVI, NRVI, and TNDVI in the dry period and in the vegetation indices TSAVI2, MNDVI, MGNDVI, TNDVI, PVI2, and WDRVI in the wet period. The indices gave poor estimates of plant coverage in the study area and were unable to separate the spectral reflectance of soil from that of plants. We used two SPOT satellite images from two dates, one in 2010 and the other in 2011. Image analysis was performed using various programs including ERDAS IMAGINE9.1, IDRISI TAIGA16 and ARCGIS10. This study indicated that vegetation indices are not very capable of separating categories of plant density and that vegetation indices are better than the hybrid classification method. To validate the results of different classes of plant coverage and plant density, an accuracy assessment technique was used. According to the assessment, the vegetation indices MSAVI1, MSAVI2, WDVI, GESAVI are best in the dry period, and GESAVI, MSAVI1, SAVI, PVI1, OSAVI, and WDVI are best in the wet period for the study area, while MNDVI and MGNDVI never performed well in the study area.

Keywords: Remote sensing, vegetation indices, plant coverage, plant density, hybrid classification, Satellite data.