Assessment of Natural Vegetation Clearing and Re-Growth in Southern Gadarif (Sudan) Using Change Vector Analysis Based on Remote Sensing and Field Data

There is a global increase in the recognition of environmental, social and economic values of native vegetation, particularly in terms of both sustainability of agricultural production and maintenance of natural resources. The rapid growth of the human population in Sudan (2.6 % per year) stimulated the evolution of mechanized agriculture in the Gedarif Area from 500 ha in the 1940s to about 2.3 million ha in 2003. Nearly one third of Sorghum (Sorghum bicolor) and Sesame (Sasemum indicum) produced in Sudan is cultivated in this area. Destruction of natural vegetation to provide agricultural land, associated with poor agricultural practices has resulted in a continuous degradation of the natural resources. A significant amount of agricultural land is now abandoned. Within this context, the objectives of the present study are to analyze the historical changes of natural vegetation due to agricultural expansion and to assess the present condition of natural regeneration on the abandoned agricultural land. Multi-temporal Landsat (MSS and ETM) data has been utilized to detect the historical vegetation changes using Change Vector Analysis (CVA). Image transforms (NDVI and TCT), supervised classification and field data have been used to quantify different land-use/land cover-classes and for assessing the present condition of the natural vegetation on abandoned agricultural land in the study site. Field survey has been conducted using stratified random sampling. All sample plots have been registered using GPS. Number and composition of trees/shrubs and above-ground herbaceous biomass were recorded. The field data has been combined with the satellite imagery using regression technique. The results demonstrate the capacity of the CVA to stratify different historical land-use/land-cover dynamics with a measurable direction and magnitude. Results showed a fast process of deforestation within critical levels. The remaining natural vegetation of 2003 represented approximately one fifth of the total natural vegetation of 1972. Field data has proven to be important to increase classification accuracy and to assess the vegetation attributes which otherwise could not be estimated using the Landsat imagery only.