Monitoring wetland cover changes during the past two decades within the protected areas of the Sudd wetland in South Sudan using MODIS data

Abstract: The Sudd wetland, situated along the White Nile in the Republic of South Sudan, is recognized internationally for its unique environmental features. The distinctive ecological characteristics have resulted in the creation of numerous nature reserves in efforts to conserve the habitats and ecosystem services provided. Monitoring changes within the nature reserves usually contributes to their protection. However, the political instabilities within the Republic of South Sudan makes obtaining in-situ data measurements difficult. Remote sensing technology therefore provides a solution to monitoring areas otherwise restricted by inaccessibility or regional conflicts. The aim of this thesis project was to monitor potential wetland vegetation cover changes within the protected areas of the Sudd wetland in March 2000-2020. Possible variations in wetland cover were related to water level changes in Lake Victoria, Lake Albert, and Lake Kyoga, in addition to average precipitation from the years preceding the obtain wetland vegetation cover. Furthermore, frequency and magnitude of armed conflicts associated with herder-pastoralist communities surrounding the protected areas was related to changes in vegetation dynamics and the hydrological variables. The wetland vegetation cover was categorized into three groups, open water, seasonally-flooded grassland, and marsh, using previously reported classification methods of applying historical MODIS NDVI imagery data. This thesis project found the water levels of Lake Albert to significantly influence the spatial and temporal variations in wetland vegetation cover extent, corresponding with previously reported results. Continued elevated water basins levels may have potentially sustained a high value of mean NDVI within the Sudd wetland during years of nationwide droughts. The water levels of Lake Victoria and Lake Kyoga, in addition to the average yearly precipitation appeared to have no statistically significant association with the mean NDVI and vegetation cover area of the Sudd wetland. This study found the yearly number of reported conflicts to be associated with an elevated mean NDVI, and to be predominantly distributed in close proximity to ethnic pastoral boarders and within larger cities and towns. No statistically significant correlation was found between the magnitude of conflicts and the vegetational and hydrological variables. The analysis further highlighted the spatial variations in vegetation covers between the protected areas of the Sudd wetland, indicating a high natural variability within the areas. Monitoring variations in vegetation cover as performed in this study contributes to our understanding how the wetland is affected by different mechanisms and processes over time. Furthermore, in a region expected to receive an increased frequency of climate extremes, the understanding of vegetation responses to hydrological fluctuations increases. The usage of satellite remote sensing for monitoring variations in wetland vegetation cover have in this thesis project shown to be an effective method for observing large inaccessible areas. Future studies should further focus on monitoring local changes within the wetland and their potential hydroclimatic drivers.