Gully development and control on a volcanic ash soil at Olkaria, Kenya

Abstract

A study on two gullies (A and B) on a volcanic ash soil in a semi-arid area at Olkaria, in Hell's Gate National Park, was done to determine the main gully development processes, characteristics that contribute to gullying and to test low cost measures of control. Gullying was common in this study area where geothermal power development was taking place. The study on gully A included its morphometry, development processes, catchment, soil, and rainfall characteristics, and tests on various control measures. Similar investigations were done on gully B except that they were less detailed and did not include rainfall characteristics and tests on control measures. The main development processes observed were bed scouring, undercutting and sidewall collapse. Observations revealed that a rainfall event that caused the most bed scouring of 68.0 em had a 130 of 24.4 mm/hr which yielded runoff of 0.21 ~/s into the gully. This bed scouring was more than that caused by a later rainfall event, with a 130 of 28.8 mm/frr, that produced runoff estimated at 1.40 m3/s which scoured the bed for only 20.·Q cm . It was deduced that erosion of the gully bed depended, to a large extent, on the erodibility of the bed which was in contact with the runoff and the moisture condition or rainfall amount 5 days prior to a runoff event. The erodibility of the volcanic ash soil at Olkaria was rated high and it was established that land use changes, like overgraz ing and unprotected disposal of geothermal development waste water downhill, contributed to gUllying. Results of control measures showed that brushwood and loose stone check dams had an average deposition in their upstream of 51.0 cm and 57.0 cm respectively but the former check dams were more stable than the latter since brushwoods were secured between posts unlike the loose stones. The main problem encountered on brushwood check dams was key failure. Conversion of the channel of gully A to a grassed parabolic waterway did not prevent bed scouring but minimized undercutting and sidewall collapse. A reshaped and grassed gully sidewall got rilled and its loose soil at the toe region was undercut by runoff making grass establishment difficult. Tests done on four grass species, on an adjacent land which had been bulldozed in an unsuccessful attempt to fill the gully, showed that star grass (Cynodon dactylon) had the highest survival rate. Due to the high erodibility of the volcanic ash soils, disturbed or overgrazed lands should immediately be revegetated to avoid gully erosion. Once gullies have started minimization of the main erosion processes can be done through the installation of loose stone and brushwood check dams (to prevent bed scouring) and a grassed parabolic waterway (to minimize undercutting and sidewall collapse). Shaping of the gully wall needs further investigation since the