Rorke B. Bryan

Soil erodibility and processes of water erosion on hillslope

Abstract The importance of the inherent resistance of soil to erosional processes, or soil erodibility, is generally recognized in hillslope and fluvial geomorphology, but the full implications of the dynamic soil properties that affect erodibility are seldom considered. In Canada, a wide spectrum of soils and erosional processes has stimulated much research related to soil erodibility. This paper aims to place this work in an international framework of research on water erosion processes, and to identify critical emerging research questions. It focuses particularly on experimental research on rill and interrill erosion using simulated rainfall and recently developed techniques that provide data at appropriate temporal and spatial scales, essential for event-based soil erosion prediction. Results show that many components of erosional response, such as partitioning between rill and interrill or surface and subsurface processes, threshold hydraulic conditions for rill incision, rill network configuration and hillslope sediment delivery, are strongly affected by spatially variable and temporally dynamic soil properties. This agrees with other recent studies, but contrasts markedly with long-held concepts of soil credibility as an essentially constant property for any soil type. Properties that determine erodibility, such as soil aggregation and shear strength, are strongly affected by climatic factors such as rainfall distribution and frost action, and show systematic seasonal variation. They can also change significantly over much shorter time scales with subtle variations in soil water conditions, organic composition, microbiological activity, age-hardening and the structural effect of applied stresses. Property changes between and during rainstorms can dramatically affect the incidence and intensity of rill and interrill erosion and, therefore, both short and long-term hillslope erosional response. Similar property changes, linked to climatic conditions, may also significantly influence the stability and resilience of plant species and vegetation systems. Full understanding of such changes is essential if current event-based soil erosion models such as WEPP and EUROSEM are to attain their full potential predictive precision. The complexity of the interacting processes involved may, however, ultimately make stochastic modelling more effective than physically based modelling in predicting hillslope response to erodibility dynamics.

The relationship of soil loss by interrill erosion to slope gradient

Abstract The influence of slope gradient on erosion rate differs for rill and interrill conditions. Rill erosion increases substantially more with increasing slope gradient than interrill erosion. Combining the two erosion processes into single data sets led to the development of regression equations (e.g., USLE) that overestimated the effect of slope gradient on erosion rate for low slope gradients and short slopes. This study investigated the change in interrill erosion rate with slope gradient and examined its relationship to runoff velocity. A sandy loam (grey brown luvisol) was packed in 100×40×10 cm3 soil trays and subjected to simulated rainfall for a period of 75 min. Rain-impacted flow erosion and downslope splash were monitored, and runoff velocity measurements were made at three positions within the flume. Downslope splash erosion never accounted for more than 20% of the total erosion. Rain-impacted flow erosion peaked early in the simulation then decreased to a constant rate; erosion rate was therefore probably detachment-limited. For a constant runoff rate, rain-impacted flow erosion increased roughly with the square root of slope gradient, as did the runoff velocity. Soil loss was correlated (0.81) with runoff velocity under the experimental conditions.

Interception loss and rainfall redistribution by three semi-arid growing shrubs in northeastern Mexico

Abstract Interception loss and rainfall redistribution were measured in individual shrubs of Diospyrus texana, Acacia farnesiana and Prosopis laevigata from a semi-arid vegetal community in northeastern Mexico in the summer of 1987. In this period 230 mm of precipitation were recorded from 17 storms. Net precipitation averaged 167.6 mm, of which throughfall formed 160.5 and stemflow 7.1 mm. Interception loss was 27.2% of the total gross precipitation. Significant differences in stemflow were noted both among species and within the species D. texana . Stemflow inputs averaged 321 and 115 ml min −1 for a 40 mm h −1 simulated storm for the species D. texana and A. farnesiana-P. laevigata , respectively. The theoretical areas and distances over which stemflow spread averaged 0.320 m 2 and 0.115 m 2 and 0.30 and 0.15 m, respectively. These areas were calculated to receive 3030 and 2650 mm as annual precipitation.

The influence of slope angle on final infiltration rate for interrill conditions

Abstract Despite numerous studies, the relationship between slope angle and infiltration rate remains unclear. Under different experimental conditions, new processes may be introduced which influence the relationship. The presence of rills, for example, may significantly affect the relationship between slope angle and infiltration rate, so distinct relationships may exist for rill and interrill areas. The objective of this study was to investigate the influence of slope angle on infiltration rate for an interrill area with a surface crust. The experiment was carried out in a laboratory using simulated rainfall and soil trays designed to limit the processes to those found in interrill areas. Slope angles ranged from 1.5° to 21.5°: infiltration was calculated from the overland flow rate, and microtensiometers and micromorphological analysis were used to characterise the seals. Infiltration rate decreased with increasing slope angle; image analysis of pore characteristics in the affected layer and subseal pressure head measurements indicated there were no differences in seal characteristics between slope angles. The results showed that sealing intensity did not vary with slope angle, and the dominant influence of slope angle on infiltration rate resulted from changes in overland flow depth and surface storage. The results also suggest that small changes in seal hydraulic conductivity with microrelief play an important role in the infiltration process.

The use of rainfall simulation tests to assess the influence of vegetation density on soil loss on degraded rangelands in the Baringo District, Kenya

The relationship between cover density and soil loss under simulated rainstorms of 30 and 60-minute duration and 33 mm h−1 intensity was investigated. Soil loss varied from 0–7.3 g m−2 for cover of 55–95% and reached maximum values of over 80.0 g m−2 (30-minute storms) and 140.0 g m−2 (60-minute storms) for cover of 25% or less. A critical threshold occurred at 55% cover below which erosion rates rapidly increased to over 15.0 g m−2 (30-minute storms) and 30.0 g m−2 (60-minute storms) during single rainstorm events. Storm duration and frequency were important determinants of erosion over more prolonged time periods. Total calculated soil loss for an 18-year period under 55% cover for frequent 30-minute storms (frequency: 3–4 times a year) was 6 times that for infrequent 60-minute storms (5-year return period).

Rill network development and sediment budgets

Threshold conditions for rill incision are well known, but few studies have examined the effect of network geometry on water and sediment fluxes within an evolving rill system. This paper reports the first in a series of studies designed to identify the influence of soil properties on rill network and confluence geometry, and on water and sediment fluxes in rill systems. The object of this study was to prepare detailed rill network water and sediment budgets identifying important sources and sinks. Simulated rainfall experiments were carried out on a Canadian silt loam soil in a 7·1 m × 2·4 m flume on a 5° slope. Rill networks of varying complexity developed, which were ultimately constrained by flume boundaries. Sediment and water fluxes and hydraulic conditions were measured within networks and at a terminal weir. Networks evolved by initial knickpoint incision in the lower flume, as flow shear velocities reached critical levels of 4·5–5 cm s−1 followed by headward migration and tributary development. Microtopography determined tributary location, but the timing and intensity of development were controlled by the incision and migration rate of the main channel, which changed local thalwegs and base levels, raising shear velocities on side slopes above critical levels. Sediment discharge at the weir broadly reflected rill incision intensity and transport-limited conditions, but as active incision moved headward, the linkage became attenuated and identification of discrete erosional incidents (e.g. local bank collapse) in the signal of the weir record became very difficult. Detailed water and sediment budgets showed much more complex patterns of localized incision and deposition within networks, strongly influenced by local changes in thalweg and in water discharge due to seepage or return flow, and by the effect of confluences on hydraulic conditions. Results indicate the value of detailed sediment budgets in interpreting weir water and sediment flux records, and the necessity of linking such measurements for erosion plots and hillslope segments to rill network characterization. Experimental results are consistent with a simple model of rill system evolution based on channel incision, headward migration, and the critical shear velocities for rill initiation. Copyright

Impact of afforestation on hydrological response and sediment production in a small Calabrian catchment

Extensive afforestation using Pinus and Eucalyptus has taken place in Calabria since the early 1960s to control expansion of calanchi and biancane. In 1978 three small catchments were established near Crotone to monitor the effect of afforestation on hydrological response and sediment yield. In 1992, rainfall simulation experiments were carried out on plots in these catchments to determine more precisely the effect of tree and ground vegetation on surface runoff and erosional response. Most experiments were carried out in a logged catchment with slopes ranging from 20 to 30° and aspects from SW to NE. Results showed complex runoff generation and sediment production, reflecting the effect of microclimate and subtle variations in vegetation on infiltration characteristics. On south-facing slopes with little ground vegetation runoff generation was rapid with runoff coefficients from 27 to 37% and peak sediment concentrations reaching 83.7 g · 1−1. On north-facing slopes with good tree cover, little grass, but continuous leaf litter, runoff coefficients reached 21%, but peak sediment concentration was only 3.6 g · 1−1, while on recently logged north-facing slopes with dense grass cover the highest runoff coefficient was only 5.5% runoff coefficient, and there was virtually no sediment production. Implications of results for forest management and soil conservation are discussed.

Particle size distribution of sediment transported by shallow flow

Abstract A small laboratory flume on a 0.035 m m−1 slope was used to examine particle selection in shallow flows. Initial experiments were carried out on smooth surfaces to measure travel rates of different sediment sizes. Grain size fractions were introduced by hand into 1.37 × 10−4 to 3.96 × 10−4 m2 s−1 flows, and clear selective mobility occurred with slowest transport in the 63–90 μm range. Following those experiments, a sediment mixture was introduced a selectively from a vibrating overhead hopper into 1.39 × 10−4 to 8.16 × 10−4 m2 s−1 flows on roughened surfaces. The highest discharge had sufficient energy to transport all particle sizes but 355–595 μm were preferentially transported. None of the lower discharges were competent to transport the full range of particles injected and 45–125 μm grains were always preferentially transported. Sediment finer than the median bed grain diameter was entrapped in micro bed pockets. The progressively coarsening of transported material in time, observed for all flow conditions, may be related to progressive filling of micro bed pockets by fines. Both transport and entrapment selectivity is strongly related to the ability of flow to transport the full range of grains present, as well as to the relationship of the bed roughness to the particles transported.

Critical conditions for rill initiation on sandy loam Brunisols: laboratory and field experiments in southern Ontario, Canada

Abstract In rainfall simulation experiments on 0.8 m × 10 m plots with agricultural soils, knickpoint development and incipient rill erosion was associated with distinct ranges of flow intensity, expressed by shear velocity, unit stream power, and unit discharge. Critical threshold shear velocity and unit stream power values were considerably higher than those identified in other studies, probably due to differences in experimental methodology. Rill initiation was found to be a function of local flow hydraulic conditions and independent of total plot discharge. This complicates prediction of this process for effective modelling as some index which links soil microtopography and flow concentration must also be incorporated in addition to runoff volume. Rill initiation could not be successfully linked to soil shear strength variations measured by a torsional shear vane. The vane shear strength values measured cannot be linked sufficiently precisely to a thin surface soil layer to provide a reliable indicator of rill and interrill erodibility.

Fitting the analytical model of rainfall interception of Gash to individual shrubs of semi-arid vegetation in northeastern México

Abstract This study was conducted to estimate the components of interception loss and to fit the analytical model of rainfall interception of Gash to a semi-arid plant community in northeastern Mexico. The interception and components of interception loss were estimated from measurements of gross rainfall, throughfall and stemflow in individual shrubs of Diospyrus texana, Acacia farnesiana and Prosopis laevigata . The components of the interception loss did not deviate from other estimates in other ecosystems, except for evaporation rate during storms. In general, the analytical model performed well, although estimated total interception loss was highly sensitive to changes in both the magnitude of evaporation rate during storms and the proportion of rainfall that is stemflow.