| Abstrakt: |
Timely intervention, management and control of erosion and sliding in southeastern Nigeria, necessitated an investigation of early warning signs of gully propagation using erosion prediction potential models. Field mapping, topographical, geomorphological, geo-statistical and slope stability models were adopted. Geotechnical-cum-slope parameters were examined while principal component analysis (PCA) and artificial neural network (ANN) were utilized to establish underlying factors of gully propagation. Gully erosion classification and slope stability analysis were carried out using the detachability-liquefaction prediction index (DLPI) and Morgenstern and Price's general limit equilibrium (GLE) method respectively. Topographic measurements in the study area were between 6 and 591 m. Laboratory data showed that soils' geotechnical properties ranged from sand (65–95%), silt/clay (4–30%), LL (4–38%), PI (0–16%), ϕ (25–40°), c (0–12 kPa), MDD (1.61–2.12 g/cm3) and OMC (10.50–20.98%). Hydrological characteristics ranged from n (30–70%), k (1.97 × 10–7–4.02 × 10–4), S (30–97%) and NMC (2.50–16.87%). These values coupled with geomorphological models explained that gullies developed mostly on loose, porous, saturated sandy uplands. Factor analysis revealed high PCA values for geo-mechanical and hydrogeological properties—sand (0.840), silt/clay (− 0.826), LL (0.710), PI (− 0.672), c (− 0.795) and ϕ (− 0.738), k (0.842), S (0.822) and n (0.773). Soil properties with negative (−) signs are factors that enhance soil stability against erosion, which unfortunately are either lacking or present in little amounts in the erodible soils. This suggests that low values of geo-mechanical properties (fines content, PI, c, and ϕ) and high hydrogeological characteristics (k, S and n) are early pointers to soils prone to erosion. The soils' SGR and SFR were generally less than 1, with values ranging between 0.00 and 0.086 and 0.042–0.462 respectively, signifying high detachability and liquefaction potential and strong affinity to erosion forces. Values from DLPI spanned from 83.86 and 204.48 implying low to considerable erodibility potential. The ANN model displayed a coefficient of determination (R2) value of 0.966, signifying high accuracy in the DLPI modelling, while presenting sand content, n, S and PI as among the five most influential soil parameters triggering erosion in the study area. Slope stability models confirmed that the factor of safety (FOS) of the gully against slope failure ranged from 1 > FOS ≤ 1.2, indicating instability to critical stability. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink