Assessment of gully erosion in eastern Ethiopia using photogrammetric techniques

Abstract

Eastern Ethiopia is severely affected by soil erosion as it is revealed by the commonly occurring deep and wide gullies. These gullies are dynamically expanding into agricultural lands at an alarming rate. They are also the chief sources of sediment causing siltation of lakes that are valuable sources of both drinking and irrigation water in the Alemaya and Harar areas. A quantitative assessment of gully dynamics and the associated soil loss and a prognosis of areas susceptible to further gullying were deemed crucial. Aerial photographs from two different dates (1966 and 1996) with approximate scales of 1:45 000 were used to study the development and dynamics of gullies as indicated by volume changes between the two dates. They were digitized at a pixel size of 20 μm. Digital elevation models (DEM) with grid widths of 20 and 2 m were derived from stereo image pairs separately. Gully breaklines and borderlines were measured for the identified gully systems. Gully profiles were positioned at regular intervals of 20 m along the gully length, giving rise to gully sections that are portions of a gully between neighbouring profiles. Gully volume was computed for every gully section. Soil loss was estimated from the computed gully volumes using soil bulk density. It showed that soil loss per unit of gully surface is very high (1.7 tons/m 2 of gully surface over the 30 years time period) indicating the severity of gully erosion in the subcatchment. There was significant agreement between breakline measured and field measured gully depth and width for selected points. This is an indication of the potential of the method for reliable gully volume estimation and hence quantification of soil loss due to gully erosion. The possibility of predicting hill-slope areas susceptible to further gullying based on the topographic attributes such as specific contributing area and local slope from an interpolated detailed DEM with 2-m resolution was tested. It was found that wetness index alone or a combination of the wetness index and the compound topographic index at the threshold values established in literature would give better results as viewed from better connectivity of flow lines and higher conformity of the predicted susceptible areas to natural gully headcut and expansion in a landscape (reality).