Drainage network extracted from Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation map available from NASA Reverb, LP DAAC Global Data Explorer and J-space systems ASTER G DEM are very useful for hydrological analysis and water resource management. In this paper, ASTER G DEM 1 arc-secondre solution was processed by the D8_deterministic eight-node method to derive stream network. The algorithm used is D8_model where the flow direction and flow accumulation algorithm for estimation of drainage were used to carry out the drainage network of Kagunuzi sub-basin from gridded digital elevation. This paper presents an approach to drainage direction assignment and it is often necessary to assign drainage directions over flats-that is, for each cell in which direction water is moving. The delineation and extraction of stream network are carried out with various ArcGIS hydrology tools. The results of drainage network showed: 1st order 319.085 km, 2nd order: 154.383 km, 3rd order: 94.562 km, 4th order: 25.638 km and 5th order: 38.934 km of length respectively: 50%, 24%, 15%,4% and 6% of the study area. The automatic drainage network extraction using DEM after sinks filled with the threshold value >500 for flow accumulation, were compared with the natural drainage network provided by topographic maps. The results demonstrate that the drainage network extracted from Digital Elevation map 1 arc-second resolution (2011) using D8_model is not very well matching with the natural network provided by Toposheet scale of 1:50000 (1973) for the whole sub-basin of Kaganuzi. However, this method can calculate the flow direction from high-resolution DEMs efficiently and extract high-precision continuous drainage networks. For drainage network of Kagunuzi sub-basin in Kibira National Park North-Western part of Burundi, it will be useful especially in the case of water resource management for hydropower study in downstream of Rwegura, flood risk and irrigation plan in Bubanza district and surroundings.
