Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Titre : Comparative Study of Vegetation Indices for Assessing Agricultural Land of Gezira Scheme, Sudan, 2000

Auteur : Khadija Ahmed Abdelgafar Ibrahim

Université de soutenance : University of Khartoum

Grade : Master of Science in Soil and Environment Sciences 2019

Résumé
Monitoring of vegetation cover is essential step for estimating crops production, drought, desertification and range management. This study aimed to analyze efficiency of three vegetation indices based on remotely sensed data for mapping land cover of the Gezira scheme in 2000. The study aimed also to generate baseline data for assessing vegetation cover in the study area. A scene acquired by LandSat7 satellite sensor for the path 173 and row 50 was freely downloaded from Global Land Cover Facility in 2000. Geo-processing softwares were used to carry out the digital processing. Visual interpretation technique was used to recognize different vegetation patterns in the scheme. Subset image was produced to cover the scheme. Data transformation was performed to produce ratio vegetation index (RVI), normalized difference vegetation index (NDVI) and soil adjusted vegetation index (SAVI). Different classes were developed to classify the produced RVI, NDVI and SAVI images into vegetated and non-vegetated areas based on the index thresholds. Then the NDVI and SAVI were classified into four classes dense vegetation, sparse vegetation, soil and water or wet soil. Frequency analysis was used to test efficiency of the different vegetation indices. The results indicated that vegetated areas covered the 2877 Km2 , 3231 Km2 and 5986 Km2 by use of NDVI, SAVI and RVI, respectively. On the other hand non-vegetated area covered 8347 km2 , 7910 km2 and 5238 km2 by use of NDVI, SAVI and RVI, respectively. Based on NDVI and SAVI, the areas covered by dense vegetation were 216 Km2 and 894 Km2 , sparse vegetation were 2661 Km2 and 2336 Km2 , soil were 7345 Km2 and 7645 Km2 and water or wet soil were 1002 km2 and265 Km2 . Result indicated that NDVI and SAVI were strongly correlated (r=0.85) and negative correlation between SAVI and RVI (r= -0.87). Frequency analysis showed that IX overall accuracy of classification was 90% for RVI, 92% for NDVI and 88% for SAVI. The study concluded that indices were powerful and can be used for assessing land covers. The study recommended that NDVI and SAVI should be used in assessing agricultural land in the Gezira scheme. The space satellite system ,with high spatial and temporal resolutions are highly recommended for efficient assessment and monitoring of agricultural land in the Gezira Scheme.

Présentation

Version intégrale (2,1 Mb)