Peter B. Hairsine

Determining the sources of suspended sediment in a forested catchment in southeastern Australia

[1]xa0Knowledge is scarce regarding the relative contributions from harvested areas and unsealed roads to sediments in forested catchments. We investigate the source composition of suspended sediment using sediment tracers and an improved multivariate mixing model. Sediment samples were collected during six rainfall events. Geochemical and radiometric tracer properties were corrected for particle size and organic content as well as conservativeness during erosion and sediment delivery. The mixing model incorporates variability of the tracer properties, using a Monte Carlo simulation technique. Mean sediment contribution from the undisturbed forest was 50–70%; harvested areas, gravel-surfaced roads, and ungravelled roads contributed 5–15%, 6–14%, and 12–25%, respectively. The unsealed roads contribute 20 to 60 times more sediment than the undisturbed forest and about 10 times more sediment than the harvested areas on a per unit area basis. Harvested areas contribute 1 to 5 times greater sediment than the undisturbed forest. These results support other studies that identify unsealed roads as important sediment sources in forested catchments.

Runoff generation and re-distribution in logged eucalyptus forests, south-eastern Australia

Abstract While pristine forests are traditionally regarded as environments with low runoff and low erosion potential, timber harvesting can dramatically affect surface runoff production, with some important consequences for in-stream water quality. A series of large-scale (300xa0m 2 ) rainfall simulator experiments on disturbed forest hillslopes, incorporating both snig track (skid trail) and general harvesting areas (GHA), examined runoff generation and redistribution during three sequential simulated storms on 13 sites. The simulated storms represented 30-min rainfall events with recurrence intervals of 2, 10, and 100xa0y respectively. The 13 sites were selected to represent dominant soil types and ages since disturbance. The snig tracks and GHA are characterised by significantly different soil hydraulic and vegetation properties as reflected in the nature and magnitude of runoff production. Infiltration-excess runoff dominated the snig track areas, while on recently disturbed GHA, runoff was distinctly patchy because of the high degree of spatial variability in saturated hydraulic conductivity ( K s ) and degrees of disturbance. Surface runoff from recently constructed snig tracks was an order of magnitude higher than the neighbouring GHA. Relative differences in runoff production between the two areas declined during extreme rainfall events. Infiltration tests also indicated that a change in the processes and rates of runoff persists for at least 5xa0y after disturbance. The practice of redistributing concentrated snig track runoff at cross banks was most effective for small storms. However, during more extreme events, the volume of snig track runoff increased and flow velocities and transport distances increased as the hillslope became increasingly saturated. The varying ability of the GHA to generate runoff and absorb concentrated flow from the snig track is likely to be critical in predicting both the initial hydrologic response and the recovery of a small catchment from forest disturbance.

Evaluating a multiclass net deposition equation in overland flow conditions

[1]xa0The movement of sediment and related pollutants in overland flow from the point of erosion to a stream normally occurs through a series of deposition zones. Here we present the evaluation of a new multiclass model of sediment transport through these zones. This model simulates size-selective net deposition as a result of simultaneous deposition by settling and nonselective reentrainment of previously deposited sediment [Hairsine et al., 2002; Sander et al., 2002]. Multiclass sediment deposition models provide a useful basis for simulating off-site effects of soil erosion, but information on the grain-size composition of deposited and exported sediment is crucial for calibrating and evaluating such models. Therefore the multiclass sediment deposition model was tested using detailed experimental data. Information on the amount of sediment deposition, on the grain-size distribution of the deposited and exported sediment, and on the spatial pattern of the deposits was collected during sediment deposition experiments. Calibration of the multiclass model, using one part of these data, was performed by adjusting the fraction of stream power available for sediment reentrainment by minimizing errors between both simulated and observed total outflow sediment concentration and grain-size distribution of the exported sediment. Evaluation was conducted, using the other part of the data, over a wide range of conditions. The evaluation shows that the deposition equation is capable of reasonably simulating the observed size selectivity of the deposition process. However, the model is not capable of simulating correctly the spatial pattern of the deposited sediment because of simplifications used in the hydraulic model.

Sediment transport through an area of net deposition

[1]xa0When sediment-laden overland flow passes across an area of reducing surface slope or increased hydraulic roughness, there is often a reduction in the sediment load carried by that flow. This paper provides a new model for the description of processes and sediment fluxes through such zones. The model describes simultaneous deposition and reentrainment of a multiclass sediment mix and includes explicit description of the mass and size composition of the deposit. A range of solutions to the developed equation for time-steady inputs is provided. These solutions are the multisediment class case when no reentrainment occurs and the single-sediment class solution when reentrainment is significant. The more general case of multiclass reentrainment and deposition is developed here, but a solution is provided in a subsequent paper. A preliminary evaluation of the model using a laboratory data set is presented. The model is shown to predict well trends of sediment concentration leaving an area of net deposition with incoming sediment concentration and unit discharge. Advantages of this new model over current models include a physically based sorting mechanism and no reliance on an externally invoked transport capacity concept.

Steady state sediment transport through an area of net deposition: Multisize class solutions

[1]xa0In this paper, analytical solutions are developed for the multisize class equations derived by Hairsine et al. [2002] for steady state sediment transport over areas of net deposition. Under conditions of a constant flow velocity we are able to decouple the equations and reduce the solution to a straightforward integral for the largest size class. All other size class solutions are then given by an algebraic equation. For practical applications we also derive simple but accurate approximations to this integral, resulting in solutions for all size classes being known explicitly as a function of distance. The effect of infiltration within a net deposition zone is then considered; however, decoupling of the governing equations is no longer possible as the flow velocities are now spatially dependent, and a full numerical solution is required.

Nutrient movement due to overland flow in managed native Eucalyptus forests, southeastern Australia

A rainfall simulator study of total nutrient movement due to overland flow on forest skid trails and general harvesting areas (GHA) in the native eucalypt forests of southeastern Australia, confirmed that the disturbed and compacted areas were the dominant sources of sediment-bound pollutants. Low runoff production and low sediment availability on the rough surfaces of the GHA reduced total nutrient yields. On the more recently disturbed and burnt sites, the solution pathway may have been more dominant due to runoff coming in contact with surface ash and charcoal. We also recorded significant reductions in the total nutrient yields as runoff from the skid trail was delivered to the adjacent hillslope via a cross bank (water bar). This was due primarily to coarse-grained sediment deposition at the base of the banks. Nutrient concentrations in the eroded sediment at the cross bank outlets, however, remained 9 times higher than the in-situ surface soils, due to the preferential delivery of fine aggregate material to the adjacent hillslope. Further reductions in total nutrient yields occurred as overland flow passed through the GHA, illustrating the effectiveness of vegetative filtering in reducing runoff volumes and sediment fluxes. Temporal trends in the data highlighted significant reductions in total nutrient movement on both areas of disturbance as sites regenerated over a period of ∼ 5 years. Water pollution abatement practices in these environments should aim to locate skid trails to maximise the redistribution of runoff and sediment from highly disturbed areas to more vegetated parts of the hillslope, and by appropriate cross bank spacing, reduce the volume of overland flow carrying sediment and nutrients.

An Alternative Approach to Modelling Sediment Deposition and Related Sorting

When sediment-laden overland flow passes across an area of reducing surface slope or increased nhydraulic roughness, it often results in a reduction in the sediment load carried by that flow. We present a nnew model for describing the formation of such a deposit as the outcome of deposition by settling and nresuspension occurring simultaneously. Consideration of size selectivity of the individual processes leads nto predictions of moderate reduction of the proportion of coarse sediment passing an area of deposition nwhile most fine sediment remains in suspension. We evaluate the model performance against results from a nseries of experiments for slope reductions using two sediment mixtures and a range of water flux, surface ngradient and incoming sediment concentration. The model is found to perform well in predicting the gross nsediment delivery ratio and the size composition of the sediment remaining in suspension. It is concluded nthat the model is a viable alternative to algorithms of Foster currently used in a wide range of sediment ndelivery models.