You have no items in your shopping cart.
ABSTRACT
In geotechnical engineering analysis and design, one of the problems that generally arise involves uncertainty in properties of soils due to spatial variability and how to establish characteristic or representative soil properties for design. For gullies, estimation of erosion of the gully bed and slope stability of the gully walls are important design requirements. Therefore, it is necessary that any design for gullies consider the influence of soil spatial variability which are expected to be significant.
This thesis describes the spatial variability of soil properties on two gully sites by collecting samples from the gull’s left and right walls and beds. Soil tests such as; specific gravity, moisture content, atterberg limits, particle size distribution, undrained triaxial tests, organic carbon content test and organic matter content test that are required to characterize the soil for the determination of erosion rate and slope stability of the soils were performed. Statistical and geostatistical analyses were also performed to determine the descriptive statistics and develop spatial distribution and probability distribution models for each of the soil property for both gullies. The Kolmogorov–Smirnov goodness-of-fit test was used to determine the best probability distribution model for each soil property needed to generate random values. A combination of deterministic and probabilistic (Monte Carlo simulation method) approaches were used in analyzing the probability of exceeding specified tolerable soil erosion rates and the probability of having a stable slope using the randomly generated soil properties.
Results showed that the soils in both gullies are predominantly sandy and silty clayey. The spatial distribution maps indicate similar trend in variation of the soil properties along both gully walls and bed of both gullies. Slope stability reliability analysis showed a very high probability of failure indicating that the walls of both gullies are very unstable and likely to fail. The associated reliability index is approximately 0.84 for both walls of both gullies which is lower than 3.8 recommended by Eurocode 0. This suggest that the gullies in its present state has a high probability to continue to expand and therefore calls for immediate remedial action to be taken. The factor of safety of both gully walls were found to follow a lognormal distribution. Soil erodibility analysis indicates that the annual rate of soil loss for both gullies follow a normal distribution and that for both gullies, the probability of getting moderate to severe annual rate of soil loss is very low and that the likely soil erosion rate will be in the region of very low to the lower bound of moderate soil loss (0 – 14 tonnes/hectare/year) for Ekhagure gully and very low (0 – 7 tonnes/hectare/year) for Iguosa gully. To obtain a reliability index greater than the target value of 3.8 in respect of gully bed erosion, the probability of obtaining or exceeding the very low to low soil loss rate has to be reduced by putting in place serious control measures to limit the influence of agents of soil loss and erosion.
