Riparian vegetation strips are widely used by river managers to increase streambank stability, among other purposes. However, though the effects of vegetation on bank stability are widely discussed they are rarely quantified, and generally underemphasize the importance of hydrologic processes, some of which may be detrimental. This paper presents results from an experiment in which the hydrologic and mechanical effects of four riparian tree species and two erosion-control grasses were quantified in relation to bank stability. Geotechnical and pore-water pressure data from streambank plots under three riparian covers (mature trees, clump grasses and bare/cropped turf grass) were used to drive the ARS bank stability model, and the resulting factor of safety (F s ) was broken down into its constituent parts to assess the contribution (beneficial or detrimental) of individual hydrologic and mechanical effects (soil moisture modification, root reinforcement and surcharge). Tree roots were found to increase soil strength by 2-8 kPa depending on species, while grass roots contributed 6-18 kPa. Slope stability analysis based on data collected during bank failures in spring 2000 (following a very dry antecedent period) shows that the mechanical effects of the tree cover increased F s by 32 per cent, while the hydrologic effects increased F s by 71 per cent. For grasses the figures were 70 per cent for mechanical effects and a reduction of F s by 10 per cent for the hydrologic effects. However, analysis based on bank failures in spring 2001 (following a wetter than average antecedent period) showed the mechanical effects of the tree cover to increase F s by 46 per cent, while hydrologic effects added 29 per cent. For grasses the figures were 49 per cent and 15 per cent respectively. During several periods in spring 2001 the hydrologic effects of the tree cover reduced bank stability, though this was always offset by the stabilizing mechanical effects. The results demonstrate the importance of hydrologic processes in controlling streambank stability, and highlight the need to select riparian vegetation based on hydrologic as well as mechanical and ecological criteria. Published in

No abstract

No abstract

Step-pools sequences are increasingly used to restore stream channels. This increase corresponds to significant advances in theory for step-pools in recent years. The need for step-pools in stream restoration arises as urban development encroaches into steep terrain in response to population pressures, as stream channels in lower-gradient areas require stabilization due to hydrological alterations associated with land-use changes, and as step-pools are recognized for their potential to enhance stream habitats. Despite an increasingly voluminous literature and great demand for restoration using step-pool sequences, however, the link between theory and practice is limited. In this article, we present four unique cases of stream restoration using step-pools, including the evolution of the approaches, the project designs, and adjustments in the system following restoration. Baxter Creek in El Cerrito, California demonstrates an early application of artificial step-pools in which natural adjustments occurred toward geomorphic stability and ecological improvement. Restoration of East Alamo Creek in a large residential development near San Ramon, California illustrates an example of step-pools increasingly used in locations where such a channel form would not naturally occur. Construction of a step-pool channel in Karnowsky Creek within the Siuslaw National Forest, Oregon overcame constraints posed by access and the type and availability of materials; the placement of logs allowed natural scouring below steps. Dry Canyon Creek on the property of the Mountains Restoration Trust in Calabasas, California afforded a somewhat experimental approach to designing step-pools, allowing observation and learning in the future. These cases demonstrate how theories and relationships developed for step-pool sequences over the past two decades have been applied in real-world settings. The lessons from these examples enable us to develop considerations useful for deriving an appropriate course of design, approval, and construction of artificial step-pool systems. They also raise additional fundamental questions concerning appropriate strategies for restoration of step-pool streams. Outstanding challenges are highlighted as opportunities for continuing theoretical work.

Abstract:Gully head and wall retreat has commonly been attributed to fluvial scour and head collapse as a result of soil saturation, sapping or piping. The empirical evidence to substantiate these conceptual models is sparse, however, and often contradictory. This paper explores the hydrological and mechanical controls on gully head and wall stability by modelling the hydrology, stability and elastic deformation of a marl gully complex in Granada Province, southeast Spain. The hydrological and slope-stability simulations show that saturated conditions can be reached only where preferential fissure flow channels water from tension cracks into the base of the gully head, and that vertical or subvertical heads will be stable unless saturation is achieved. Owing to the high unsaturated strengths of marl measured in this research, failure in unsaturated conditions is possible only where the gully head wall is significantly undercut. Head retreat thus requires the formation of either a tension crack or an undercut hollow. Finite-element stress analysis of eroding slopes reveals a build up of shear stress at the gully head base, and a second stress anomaly just upslope of the head wall. Although tension cracks on gully heads have often been attributed to slope unloading, this research provides strong evidence that the so called 'sapping hollow' commonly found in the gully headwall base is also a function of stress release. Although further research is needed, it seems possible that 'pop out' failures in river channels may be caused by the same process. The hydrological analysis shows that, once a tension crack has developed, throughflow velocity in the gully headwall will increase by an order of magnitude, promoting piping and enlargement of this weakened area. It is, therefore, possible to envisage a cycle of gully expansion in which erosion, channel incision or human action unloads the slope below a gully head, leading to stress patterns that account for the tension crack and a stress-release hollow. The tension crack promotes faster throughflow, encouraging hollow enlargement and piping, which undercut the gully head. The tension crack permits the development of positive porewater pressures behind the gully head, leading either to failure or contributing to toppling. Finally the debris may be eroded by fluvial action, unloading a new section of slope and completing the cycle of gully head retreat. Copyright

The susceptibility of cut slopes to landsliding can be reduced in certain circumstances by the establishment of a vegetation cover. However, the hydrological implications of allowing a cover to develop may offset the mechanical benefits of soil reinforcement by roots. The balance between hydrological and mechanical effects is critical on slopes which are susceptible to the development of an infiltration-induced transitory perched water table, a common cause of landslides in deep, tropical residual soils. This balance is likely to change both between slopes of different types as well as temporally on any given slope. The net effect of a vegetation cover must be predicted either before natural vegetation covers are allowed to encroach on bare slopes, or if engineers are considering the use of trees as a protective measure. This paper presents a method of calculating the impact of a vegetation cover on slope stability. Simulations carried out on a wide range of slope types suggest that where failure is most likely to be triggered by infiltration rather than ground water rise, large-scale vegetation covers may contribute to instability. Whether vegetation had a positive or negative impact on slope stability was controlled by the permeability of the soil matrix, whilst the magnitude of impact was controlled by the soil strength and the slope height.

Erosion of cohesive channel materials is not fully understood, but is assumed to occur largely as a result of hydraulic shear stress. However, field and laboratory observations of pore-water pressures in cohesive streambed materials reveal the presence of positive and negative pore-water pressure effects that may significantly affect the erosion process, as contributing and resisting forces respectively.Measurements of pore-water pressures below cohesive streambeds in the loess area of the midwestern USA were conducted in situ and in undisturbed cores with a digital, miniature tensiometer. Results disclosed matric suction values in the range of 15-50 kPa in eastern Nebraska and northern Mississippi. Repetitive tests in soft materials verified a change from positive pore-water pressures in the upper 10-15 cm, to negative pore-water pressures to depths of at least 50 cm. In firm materials, the entire sampled profile was unsaturated.Laboratory experiments were carried out in which synthetic hydrographs were imposed on undisturbed streambed cores from the same sites. Miniature tensiometers in the cores monitored the resulting pattern of pore-water pressures, and revealed upward directed seepage forces on the recessional limb of the hydrograph. Maximum calculated values of the force ranged from 10 to 275 kN for the materials and heads tested. The maximum value obtained after application and release of a 2Ð5 m head was 119 kN, with 275 kN after a 5Ð0 m head. These results were supported independently by subsequent simulations using a finite-element hydrology model coupled with a stress-deformation model.A numerical scheme was developed to calculate the forces acting on cohesive aggregates in an idealized streambed, and to evaluate the potential for their detachment. The scheme added upward-directed seepage as an additional driving force, and matric suction as an additional resisting force, to the commonly applied factors of particle weight, fluid drag and lift force. Results demonstrate that upward-directed seepage forces of the magnitude measured in the laboratory with 5Ð0 m stages have the potential to detach particles larger than 10 cm in diameter without requiring fluid drag and lift forces. When added to these hydraulic forces, erosion thresholds are lowered, enabling erosion at lower hydraulic stresses.A hypothesis for detachment of chips or blocks of cohesive bed material is proposed: (1) large (>5 m) rises in stage increase pore-water pressures or decrease matric suction dramatically in the region just below the bed surface; (2) a relatively rapid decrease in stage causing a loss of water pressure above the bed, combined with low-rates of excess pore-water pressure dissipation just below the bed surface result in steepened hydraulic gradients; and (3) a resulting net upward seepage force is great enough to contribute to detachment of cohesive bed material, or rupture the bed by exceeding the available strength and confining stress. Published in

Bank-stability concerns along the Missouri River, eastern Montana are heightened by a simulated change in flow releases from Fort Peck Dam to improve habitat conditions for Pallid Sturgeon. The effects of the simulated flow releases on streambank pore-pressures and bank-toe erosion needs to be evaluated to properly model bank-stability. The Bank-Stability Model used incorporates pore-water pressure distributions, layering, confining pressures, reinforcement effects of riparian vegetation and complex bank geometries to solve for the factor of safety. To increase the applicability and accuracy of the model for use in predicting critical conditions, the hydraulic effects of bank-toe erosion have been added. According to the simulated flow-release plan, flows of 216 m 3 /s are increased by 38.3 m 3 /s/day for 12 days to 675 m 3 /s, held for 60 days and decreased for 12 days back to 216 m 3 /s. Results show the important contribution of bank-toe erodibility in controlling mass failure. Banks at River Miles 1624, 1676 and 1716 attain F s < 1.0 indicating imminent failure. These sites contain less resistant sandy-silt material at the bank toe, and experienced simulated undercutting up to 3m. More resistant cohesive, clay bank toes at River Miles 1589 and 1762 were undercut only 0.2 m and remained stable.