University of Twente Student Theses


Hydrologic impacts of Landuse change in the Upper Gilgel Abay River Basin, Ethiopia; TOPMODEL Application.

Gumindoga, Webster (2010) Hydrologic impacts of Landuse change in the Upper Gilgel Abay River Basin, Ethiopia; TOPMODEL Application.

[img] PDF
Abstract:Landuse and landcover change affect the different hydrological components like interception, infiltration and evaporation thereby influencing runoff generation (both process and volume) and streamflow regimes. Comparatively, little is known about factors that affect runoff behaviour and their relation to landuse in a data poor catchment like the Upper Gilgel Abay basin. Remote sensing was therefore used in this study to observe catchment characteristics and to estimate the model parameters that reflect on the land surface characteristics. Firstly, the TOPMODEL approach was applied to simulate streamflow for this basin. An ASTER 30m DEM was used to compute the topographic Index, critical for the simulation of streamflow in the basin. Results of calibration gave a Nash-Sutcliffe model efficiency (NS) of 0.81 and a Relative Volume Error (RVE) of 6.1%. Sensitivity analysis of the model showed that the parameters most critical for accurately simulating runoff were: the exponential transmissivity function (m), the soil transmissivity at saturation (To) and the root zone available water capacity (SRmax). The model was validated using a 2003 meteorological dataset and a satisfactory model performance was obtained (NS=0.75, RVE= -4.0 %). GIS and remote sensing were further used for the quantification of vegetation indices such as SAVI and LAI. Rainfall interception as a function of LAI from different vegetation types was determined. The implementation of landuse in TOPMODEL was done by treatment of each vegetation/landuse type as a ‘subcatchment’ through a GIS overlay of landuse types thus creating a topographic index distribution for each landuse type. These were run separately with specific landuse parameters. The areally weighted results were summed to get a total output imitating having multiple subcatchments with different topographic index distributions. Results showed that the maximum peakflow from agricultural land increased by 51% from 1973-1986 and by 44% between 1986 and 2001. Annual runoff volume increased by 12% between 1986 and 2001 which corresponds to increases in agricultural land from 1973 to 2001. From 1973-1986 and from 1986-2001, forest and shrubland decreased in maximum peakflow by same amount (29%). The annual runoff volume also decreased by 36% from 1973-1986 and by 34% from 1986-2001. This could be attributed to decreases in forests between the years 1973, 1986 and 2001. Finally for each year, a comparison was made between the sum of all landuse simulated discharge and the observed discharge at the outlet. The following satisfactory model efficiencies were obtained: 1973 (NS=0.81, RVE=5.82 %); 1986 (NS=0.72, RVE=29.72 %) and 2001 (NS=0.73, RVE= 18.50 %). These results prove that in data poor basins, a promising way to analyse hydrological impacts of land-use change is by combining remote sensing for land surface parameterization and a semi distributed rainfall-runoff model. The findings also provide useful support for land use planning and management. Key words: Upper Gilgel Abay, Land use, Remote sensing, TOPMODEL, Nash–Sutcliffe, Streamflow.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page