Measured Sediment Yield Research Articles

Modelling sediment supply to rivers and reservoirs in Eastern Europe during and after the collectivisation period

At present, the landscapes in the former communist countries of Eastern Europe are undergoing a second major change in 50 years. After World War II, collective agricultural systems, such as the United Agricultural Cooperatives and the State Farms, were established. Following the example of the Soviet Union, individual fields were put together in order to enable mass production. This, however, led to an enormous acceleration of erosion and sediment transport processes on arable land. Since the fall of the communist regimes in 1990, farmers, or their successors, can claim back their original property. The large collective fields are split up in smaller spatial units. The land that is not claimed back is abandoned. This new transition has again a huge impact on soil erosion and sediment export to rivers and reservoirs. In order to control and optimize the transition from collective to private farming systems, land arrangement administrations need modeling tools to simulate the geomorphic impact of possible future land use scenarios. In this paper, SEDEM, a spatially distributed sediment delivery model, is calibrated and validated with measured sediment yield data from Czech drainage basins. The results suggest that SEDEM is able to predict the impact of land use changes on the mean annual sediment supply to rivers. Next, SEDEM is applied to assess the impact of a range of possible future landscape scenarios such as the splitting up of large fields in smaller spatial units and conversions from arable land to pasture.

Sediment yields and soil loss rates from different land uses on Triassic shales in western Sydney, NSW

Sedimentation surveys of small dams demonstrate that land use is the dominant factor generating high sediment yields in the ungullied shale catchments of western Sydney where rainfall erosivity and soil erodibility are relatively constant. A single urban catchment produced 6.5 t/ha.year and cropped catchments an average of 6.7 ± 1.99 t/ha.year, whereas grazed woodland/forest and grazed pasture exported averages of only 2.5 ± 0.57 and 2.9 ± 1.02 t/ha.year, respectively. These yields are high by Australian standards and the farm dam sediments are enriched in both clay and phosphorus, in comparison to catchment topsoils. Empirical soil loss equations based on the Universal Soil Loss Equation (USLE) [Modified Universal Soil Loss Equation (MUSLE), Soiloss and Revised Universal Soil Loss Equation (RUSLE)] accurately predicted the measured sediment yields, with Soiloss being the most accurate. Although Soiloss is the only empirical equation to use Australian data, it is only marginally better than MUSLE, a simplified version of the USLE used for teaching. RUSLE predictions of soil loss rates were closely correlated with measured sediment yields but required inputs for poorly defined parameters. European land uses in the South Creek catchment, the largest shale catchment in western Sydney, have probably increased mean annual sediment yield by 4.4 times over that in 1788. Further increases are likely with increasing urbanisation.

Role of landscape parameters in riverine run-off, and sediment and organic matter yield on Disko Island, West Greenland

Measurements of discharge and suspended sediment and organic matter yield from nine different drainage basins on the island Disko in central West Greenland were carried out in the period 19–30 July 1997. A series of landscape properties (glacier cover, altitude, slope, aspect and vegetation cover) were measured for each drainage basin from a digital elevation model and a satellite image. Principal Component Analysis and regression statistics have been used to examine which landscape parameters dominate the measured discharges and yields. It is concluded that differences in suspended sediment and organic matter yield between the drainage basins can be explained by the measured morphometric properties while differences in water run-off can not. Glacier cover percentage and percent area above 800 m a.s.l. were the most important parameters influencing suspended sediment and organic matter yield in the drainage basins on Disko during the study period.

CHANNEL EVOLUTION AND EROSION IN PAM–TREATED AND UNTREATED EXPERIMENTAL WATERWAYS

Unprotected earthen waterways (e.g., grassed waterways before vegetation) and ephemeral gullies are prone to severe erosion. Previous research has suggested that polyacrylamide (PAM) may reduce erosion in areas of concentrated flow. This research tested the hypothesis that a PAM–treated channel would result in significantly less erosion than untreated soil in a pre–formed, trapezoidal channel measuring 0.6 m at the top, 0.1 m at the bottom, 0.13 m deep, and 15.2 m long. Anionic PAM (30% charge density, 18 Mg mol–1 molecular weight) was applied in solution at a rate of 80 kg ha–1. The soil used in these experiments was red clay loam (37% sand, 35% silt, 28% clay). Channel geometry and sediment concentration were measured for each of four inflow rates (0.0016, 0.0032, 0.0063, and 0.0126 m3 s–1). A secondary objective was to measure the influence of PAM on headcut rate advance. Measured sediment yield rate was significantly less from PAM–treated channels than from the control. Reductions in sediment yield rate ranged from 93% to 98%. Channel incision depth was not different between the two treatments; however, effective flow widths (assuming rectangular channel geometry) were significantly greater for the untreated control channel. Headcut advance rates were greatly reduced in PAM–treated channels (0.06 to 0.6 m h–1) compared to the untreated channel (17.8 m h–1) in our limited data. These results show that PAM was an effective means of controlling erosion in experimental earth channels.

Land use effects on sediment yields and soil loss rates in small basins of Triassic sandstone near Sydney, NSW, Australia

Sedimentation surveys of dams in small sandstone drainage basins near Sydney, Australia, demonstrate that land use is the dominant factor determining sediment yields and soil loss rates. Cultivated basins produce an average sediment yield of 7.1 t/ha/year whereas grazed pasture and forest/woodland basins export averages of only 3.3 and 3.1 t/ha/year, respectively. Nevertheless, these yields are high by Australian standards. Sediment exports from grazed pasture and forest/woodland basins are similar because the forest/woodland basins are also grazed. Dam sediments are enriched in clay and organic matter in comparison to topsoils. Gullies and bank erosion are not active geomorphic processes in the drainage basins investigated so that the measured sediment yields could be validly compared to soil loss rates determined by empirical soil loss equations, Modified Universal Soil Loss Equation (MUSLE), Soiloss and Revised Universal Soil Loss Equation (RUSLE), which do not account for gully and channel erosion. These equations accurately predicted the measured sediment yields, with MUSLE being the most accurate. Although Soiloss is the only empirical equation to use Australian data, MUSLE performed marginally better, despite being a simplified version of the Universal Soil Loss Equation (USLE) that is used for teaching. RUSLE predictions of soil loss rates were also closely correlated with measured sediment yields.

Sediment yields and soil loss rates from native forest, pasture and cultivated land in the Bathurst area, New South Wales

Summary Sedimentation surveys of small dams in granite catchments in the Bathurst area of New South Wales (NSW) demonstrate that land use is the dominant factor determining soil loss rates and sediment yields. We found that cultivated catchments produced, on average, 3.1 t ha−1y−1 whereas grazed pasture and forest/woodland catchments exported only 2.2 and 0.8 t ha−1y−1, respectively. The yields for cultivated and grazed catchments are high by Australian standards. Dam sediments were enriched in clay and organic matter in comparison to catchment topsoils. The dams were deliberately selected so that gully and bank erosion were not active geomorphic processes in the catchments and so the measured sediment yields could be validly compared to soil loss rates predicted by various versions of the Universal Soil Loss Equation. The Modified Universal Soil Loss Equation, Soiloss and Revised Universal Soil Loss Equation did not closely predict the measured sediment yields, as we have found ensewhere in NSW. Although Soiloss is the only empirical equation to use Australian data, it is not as accurate in the Bathurst area as for areas near Sydney. Further research is needed to test and develop soil loss prediction equations under Australian conditions.

Vegetation influence on runoff and sediment yield in the Andes region: observation and modelling

The Precordillera of the Andes Mountains (Mendoza, Argentina) is prone to severe flash floods, caused by heavy rainfall events of short duration and high intensities. Two catchments were instrumented in order to study the rainfall–runoff process and soil management impact on runoff and/or sediment yield. In the first catchment (Divisadero Largo, DL, 5.47 km 2), characterized by a large heterogeneity of surface geology, a data set of about 50 rainfall–runoff events covering the 1983–1994 period was available. Vegetation cover changed significantly after the catchment was enclosed in 1989–1990. This change was successfully mapped using Landsat TM image analysis. The second catchment (Cuenca Aluvional Piloto, CAP, 35 ha), the soil of which was homogeneous, was instrumented in 1992 for total runoff and sediment yield measurements. Three small plots of 3×10 m 2 (bare soil, 42 and 60% vegetation cover) and three sub catchments (2–4.5 ha) were delimited with different average vegetation cover. Data analysis showed the difficulty in relating runoff volume and sediment yield to simple descriptors of the catchments such as the average slope and/or the average vegetation cover. The DL and CAP catchments were modelled using the Areal Non Point Source Watershed Environment Response Simulation (ANSWERS) model with contrasting results. Good agreement between model and observation could be achieved after calibration on the 3×10 m 2 plots, but the model failed to correctly reproduce runoff on the three 2–4.5 ha CAP sub-catchments using the values calibrated on the small plots. Better results were obtained on the larger and heterogeneous DL basin, where surface geology variations and rainfall variability seemed to be the most influential factors. In this case, no sensitivity to vegetation coverage changes, induced when enclosing the catchment, was found. On the other hand, the model proved sensitive to differences in vegetation cover at smaller scales when the geology was homogeneous.

Mountain erosion over 10 yr, 10 k.y., and 10 m.y. time scales

We used cosmogenic 10 Be to measure erosion rates over 10 k.y. time scales at 32 Idaho mountain catchments, ranging from small experimental watersheds (0.2 km 2 )t o large river basins (35 000 km 2 ). These long-term sediment yields are, on average, 17 times higher than stream sediment fluxes measured over 10‐84 yr, but are consistent with 10 m.y. erosion rates measured by apatite fission tracks. Our results imply that conventional sediment-yield measurements—even those made over decades—can greatly underestimate long-term average rates of sediment delivery and thus overestimate the life spans of engineered reservoirs. Our observations also suggest that sediment delivery from mountainous terrain is extremely episodic, sporadically subjecting mountain stream ecosystems to extensive disturbance.

Calibration and validation of a non-point source pollution model

Surface runoff and sediment transport influence soil quality and the quality of receiving waters. Simulation models are useful tools to describe drainage behavior and sediment yield to develop management strategies for agricultural waters. The objective of this study was to investigate the performance of the Agricultural Non-Point Source Pollution Model (AGNPS) and modified versions comparing measured and predicted surface runoff and sediment yield at the drainage outlet. The study was carried out using 52 rainfall-runoff events from two small watersheds in Bavaria, 22 for calibration and 30 for validation. Evaluation of model outputs was based on graphical displays contrasting measured and predicted values for each rainfall-runoff event, and standard statistics such as coefficient of efficiency. A comparison between three different surface runoff methods Q 1 (uncalibrated curve number (CN) method), Q 2 (calibrated CN method), and Q 3 (Lutz method) showed that the uncalibrated CN method underestimated measured surface runoff considerable, while the Lutz method outperformed the calibrated CN method and was better than the uncalibrated CN method. The modifications made to sediment yield calculations encompassed: (i) replacement of the Universal Soil Loss Equation LS factor algorithm by one based on stream power theory (variant S 2), and (ii) linkage of channel erosion by individual categories of particle-size to runoff velocity (variant S 3). The sediment yield predictions by S 2 and S 3 outperformed the predictions of S 1 (USLE). Calculations based on S 3 gave the best agreement when compared to measured sediment yield values at the drainage outlet.

Length Slope Factors for applying the Revised Universal Soil Loss Equation at Basin Scale in Southern Italy

In this paper, for a basin divided into morphological units, a distributed model based on the Revised Universal Soil Loss Equation (RUSLE), with different expressions for the topographic factors, and on the sediment delivery ratio of each morphological unit is used. At first, the caesium-137 data available from a Sicilian basin are calibrated with two different models [the Proportional Method (PM), the Simplified Mass Balance (SMB) model] to provide net soil loss data for each morphological unit. Then, for a selected expression of the topographic factor, the slope length exponent is calculated for each morphological unit, equating the calculated sediment yield with the net soil loss. The analysis showed that the median value of the slope length exponent is dependent on the mesh size of the Digital Elevation Model used for calculating the slope of each morphological unit. The analysis also showed that the slope of the linear relationship between slope length exponent and mesh size is independent of the calibration method used for the caesium-137 measurements. Finally, the sediment yield measurements carried out in three small Calabrian basins and a Monte Carlo technique are used for evaluating the effect of the knowledge uncertainty and the stochastic variability of the model parameters on calculated sediment yield.

Sediment yield on a devastated hill in southern China: effects of microbiotic crust on surface erosion process

In southern China, most of the forests consist of Pinus massoniana, but they are usually poor, thin, and deteriorated due to human intervention such as cutting, removal of litter for fuel and over-grazing. Therefore, exposed bare soil is found in many devastated hills. As a consequence, both the direct runoff percentage and the sediment yield are very high. In order to understand the surface erosion processes in this environment, we measured sediment yield and surface erosion depth in terms of catchment scale, and carried out in situ sprinkling experiments, focusing on the spatial variability of surface conditions and sediment sources. The results of sediment yield measurements at the end of headwater catchment suggest the following characteristic. (1) The surface water is capable of transporting more material than is supplied by detachment. That is, the sediment discharge process is detachment-limited. In situ sprinkling experiments indicate that: (2) In the lower part of the hillslope where the surface is covered by microbiotic crust, the ratio of surface runoff to rainfall is high, but sediment yield is small. (3) In the upper part of the hillslope where there is no microbiotic crust, the ratio of surface runoff to rainfall is low, but sediment yield is large. Measurements of surface erosion depths indicate that: (4) Microbiotic crusts provide a high degree of protection against soil erosion, even though the surface runoff ratio is large at the lower hillslope. Consequently, the surface erosion depth is almost zero at the lower part of hillslope. (5) The main source of sediment was the upper part of hillslope. Based on these indications, it can be concluded that when the sediment discharge process is detachment-limited, the effect of the surface crust and/or the vegetation cover on the sediment discharge process is great. That is, the spatial variation of the surface crust and/or the vegetation cover controls the source of sediment.

Measured and predicted sediment yield from a subtropical, heavy rainfall, steep-sided river basin: Hanalei, Kauai, Hawaiian Islands

To determine the sediment yield of the 54.4 km 2 Hanalei River basin, we employ three methods: (1) the Universal Soil Loss Equation (USLE), which uses natural characteristics of the basin such as the amount of rain, slope steepness and length values, and soil types to predict sediment erosion in a basin; (2) the thickness and calibrated radiocarbon age of fluvial deposits cored from the coastal plain; and (3) field measurements of suspended sediment in the river. Method 1 (USLE) provides a model prediction of sediment yield that we test with observational data of methods 2 and 3. Several curves, including one by the US Soil Conservation Service, predict a sediment delivery ratio (measured sediment yield: gross erosion) between approximately 15% and 50%. With 5260±2210 Mg (metric tons) yr −1 of suspended sediment in the Hanalei River and 2300±700 Mg yr −1 deposited on the coastal plain, however, the delivery of sediment in the Hanalei basin ranged between 45% and 101% of the maximum predicted USLE value (88±103 Mg km −2 yr −1). This higher than predicted yield may be the result of mass movement. We are not able to differentiate, however, between erosion and mass movement as the principle agent of denudation. Our measurements indicate a total sediment yield of 140±55 Mg km −2 yr −1 for the Hanalei Valley.

Scales and processes of water and sediment redistribution in drylands: results from the Rambla Honda field site in Southeast Spain

Arid lands are characterised by a combination of high temporal variability of rainfall and spatial heterogeneity of soil surface properties. In response to these environmental conditions, sources and sinks of runoff water and sediments tend to be organised in mosaics with distinct spatial attributes. These patterns can be identified at several scales, each with a predominance of a different set of processes. The dynamic relationships between these patterns and processes are an essential aspect of spatial connectivity in arid landscapes. During the last six years, part of the research at Rambla Honda, a field site in Southeast Spain operating under the MEDALUS project, has been concerned with this subject. This paper reviews the results obtained up to date at the patch and the hillslope scales. The research at the patch scale focused on the role of vegetation as a source of spatial heterogeneity that affects short-range redistribution patterns of water and sediments. The approach has been to identify the dynamic relationships between plant clumps and bare ground in sparse vegetation mosaics, using field observations, experiments and simulation models. Field observations included runoff and sediment yield measurements on bounded plots and hillslope sectors, analysis of spatial correlation structures, as well as physiological and architectural properties of plant functional types. Experiments included rainfall simulation and runoff exclusion in the field, and soil fertility bioassays both in the field and the laboratory. A cellular automata model was built to explore the interactions between plant clumps and sediment movement. The research at the hillslope scale was concerned with the long-range transference of water and sediments from rocky upperslopes to their footslope sediment fill. The approach was based on an analysis of the available information about spatial patterns of soil moisture and discharge of runoff and sediments from plots and stream gauges in a first order catchment. Results show that, at the patch scale, in sparse vegetation, a range of positive feedback mechanisms lead to nucleation, or to the increase of spatial heterogeneity, by concentrating resources in the soil beneath plant clumps at the expense of the neighbouring bare ground. This spatial heterogeneity arises dynamically through the interaction between plant growth and hillslope fluxes of water and sediments. Within specific boundary conditions, this interaction is `tuned' towards the formation of mosaics of bare and vegetated patches with patterns that minimise redistribution lengths of water and sediments. The boundary conditions that affect the `tuning' process include factors that determine the potential distance and transport capacity of runoff, such as temporal variability of rainfall, slope angle, slope length, among others, and plant specific factors that affect the efficiency of plant clumps in trapping the resources that are redistributed on the hillslope. At the hillslope scale, the transference of sediment and water between hillslope elements requires very specific within-event temporal distributions of rainfall that allow for the widespread formation of a saturated layer at shallow depth and overland flow to reach first order channels. During most rainfall events these conditions are not met and, therefore, in most seasons, mean values of soil moisture do not increase downhill, and rather reflect variation in local soil properties than the effects of lateral redistribution processes. As a consequence, it may be expected that small changes of the frequency distribution of rainfall characteristics, in terms of within-storm temporal distribution of intensities, could lead to significant changes in soil moisture patterns and hydrologic connectivity between hillslope elements.

“R1‐R4” AND “BOISED” SEDIMENT PREDICTION MODEL TESTS USING FOREST ROADS IN GRANITICS1

ABSTRACT: Erosion and sedimentation data from research watersheds in the Silver Creek Study Area in central Idaho were used to test the prediction of logging road erosion using the R1‐R4 sediment yield model, and sediment delivery using the “BOISED” sediment yield prediction model. Three small watersheds were instrumented and monitored such that erosion from newly constructed roads and sediment delivery to the mouths of the watersheds could be measured for four years following road construction. The errors for annual surface erosion predictions for the two standard road tests ranged from +31.2 t/ha/yr (+15 percent) to ‐30.3 t/ha/yr (‐63 percent) with an average of zero t/ha/yr and a standard deviation of the differences of 18.7 t/ha/yr. The annual prediction errors for the three watershed scale tests had a greater range from ‐40.8 t/ha/yr (‐70 percent) to +65.3 t/ha/yr (+38 percent) with a mean of ‐1.9 t/ha/yr and a standard deviation of the differences of 25.2 t/ha/yr. Sediment yields predicted by BOISED (watershed scale tests) were consistently greater (average of 2.5 times) than measured sediment yields. Hillslope sediment delivery coefficients in BOISED appear to be overly conservative to account for average site conditions and road locations, and thus over‐predict sediment delivery. Mass erosion predictions from BOISED appear to predict volume well (465 tonnes actual versus 710 tonnes predicted, or a 35 percent difference) over 15 to 20 years, however mass wasting is more episodic than the model predicts.

USING GIS FOR FACILITATING EROSION ESTIMATION

Geographic Information System (GIS) combined with soil loss models can enhance the evaluation of soilerosion estimation. ARC/INFO geographic information system with the Revised Universal Soil Loss Equation (RUSLE)was used to estimate soil erosion on a portion of the Walnut Gulch experimental watershed in southeast Arizona. Spatialdata from different sources provided input for four alternate GIS based procedures in computing the combined slopelength and steepness factor in RUSLE for determining soil erosion estimates. Results of GIS based RUSLE erosionestimates from the four procedures are compared with actual sediment yield observed on the experimental watershed forthe period 1973 through 1989. Results indicate GIS based RUSLE predicted soil erosion estimates are less than theobserved measured sediment yield in most years. Application of a sediment delivery ratio which varies with watershedarea is addressed as possible explanation for the differences in estimated erosion and measured sediment yield. GIS canbe used with RUSLE to get a good estimate of soil erosion but care has to be taken in interpreting the result andcomparing it to measured sediment yield. The results from this study clearly show the need for more work in using GISand RUSLE for soil erosion estimation.

Erosion and sediment transport on steep marly hillslopes, Draix, Haute-Provence, France: an experimental field study

Despite a thin regolith layer, measured sediment yields indicate denudation rates of more than 10 to 20 mm yr −1 on strongly denuded slopes in black marl (terres noires), near Draix, Alpes de Haute-Provence. Overland-flow and rainfall-simulation experiments were conducted on three slopes in the Draix Erosional Research Basin to determine the nature and intensity of hillslope erosion processes. The experiments showed that during a single storm, sediment transport includes water–solid mixtures with a wide range of concentrations, from highly charged rill flow (up to 400 g l −1) to hyperconcentrated flow (up to 800 g l −1) to debris flows (up to 1400 g l −1). Miniature debris flows (MDFs), strongly controlled sediment transport as long as the sediment store was not depleted. Slope surfaces were completely reworked in the winter and the regolith layer was rebuilt. Temperature records from nine recent winter periods indicate that conditions are favourable for soil frost processes.

Testing a distributed approach for modelling sediment delivery

Both the theoretical basis of a distributed approach to sediment delivery and its agreement with basin sediment yield measurements are tested. At first, by using morphological data of five Apulian and four Calabrian basins, the applicability of a theoretically-based relationship for evaluating the sediment delivery ratio of each morphological unit SDR h, into which a basin is divided, is verified. Using the morphological data of the nine investigated basins, the sediment delivery relationship, i.e. the relationship of the basin sediment delivery ratio SDR Wand SDR h, is tested. The analysis showed that the relationships, proposed by Ferro (1997), establishing the dependence of the a coefficient of the sediment delivery relationship on the median travel time and on a drainage density index are yet applicable. The influence of the scale of the basin topographic map on the sediment delivery distributed approach is also studied. The analysis showed that the ratio between the a coefficient and the median travel time is scale-invariable while the link between the coefficient a and the drainage density index has to be re-calibrated because the drainage density index is dependent on the map scale used. Finally, the predictive capability of the sediment delivery distributed model is validated using the sediment yield measurements carried out in three small Calabrian basins and in the Apulian Triolo basin. The analysis showed that the model accuracy is dependent on the temporal scale used and the model reliability increases from the event to the annual scale.

Anti-erosive effectiveness of Eucalyptus coppices through the cover management factor estimate

In this paper sediment yield data, measured from 1978 to 1994 in two small Calabrian basins (W2 and W3) reafforested with eucalyptus trees (Eucalyptus occidentalis Engl.), and the Modified Universal Soil Loss Equation (MUSLE) applied in a distributed form are used to evaluate the anti-erosive effects of eucalyptus cover. At first step the sediment yield measurements observed in W2 basin are used to estimate a single cover management factor representative of the eucalyptus coppice and equal to the median value (0·164) of the cover management factor values calculated for each runoff event. Then, the reliability of the selected representative cover management factor is verified on W3 by comparing the cumulative distribution function (CDF) of the measured sediment yield with the CDF of the calculated one. Finally, the temporal analysis of the crop and management factor is developed searching, at monthly and annual scale, the correlation between crop anti-erosive effectiveness and rainfall erosivity index. © 1998 John Wiley & Sons, Ltd.

Chemical and physical denudation in the Amazon River Basin

We present major and trace element data on the suspended and dissolved phases of the Amazon River and its main tributaries. The Sr isotopic composition of the dissolved load is also reported. Special attention is paid to the abundances of REE and to their fractionation between the dissolved and suspended phase. The rivers of the Amazon Basin are among the richest in dissolved REE and are similar to the rivers of the Congo system. However a greater range of fractionation between LREE and HREE is reported here. At a global scale the rivers have intermediate patterns between those of the Congo system and those of high pH rivers such as the Indus and Mississippi rivers. Only few elements (Rb, U, Ba, K, Na, Sr and Ca) are mobilized by silicate weathering. These elements are strongly depleted in the suspended phase with respect to upper continental crust. In the dissolved load, these elements are controlled by atmospheric inputs and the weathering of the main lithologies. We propose a model based on mass budget equations, that allow the proportions derived from the different sources to be calculated. As a consequence silicate, carbonate and evaporite weathering rates can be estimated as well as the consumption of CO2 by weathering of each of these lithologies. Physical weathering rates can be estimated by two complementary approaches. On the one hand, the multi-year average of suspended sediments yields can be used to estimate physical denudation. On the other hand, we have developed a steady-state model of erosion that allows us to calculate physical erosion rates on the basis of the dissolved load of rivers. A mean crustal composition is assumed in this model for the rock sources of the drainage basins. Comparison of the rates predicted by the model to the observed rates shows good agreement for the lowland rivers, but a strong discrepancy for the rivers derived from the Andes. Andean rivers (Solimoes, Madeira and Amazon) have observed sediment yields much greater than those predicted according to the steady-state model of chemical and physical weathering. Two interpretations can account for this inconsistency. The first is that these rivers are not in steady state and hence that the soils are being destroyed. The second requires that the local continental crust is different from the average continental crust of Taylor and McLennan, and contains a large proportion of sedimentary rocks. Using the measured sediment yields, and assuming a steady state, we can estimate the amount of sediment recycling for each drainage basin. For the Amazon at Santarem, we find that at least 25% of the mass of the upper continental crust of the Amazon drainage basin is constituted of recycled material.

Sediment yield and seasonal soil erodibility for semiarid cropland in western Canada

Despite its importance to the degradation of the land resource, few estimates of soil loss or soil erodibility for water erosion from cropland have been made for the semiarid Brown soil zone of western Canada. To estimate these, we calibrated the event-based Modified Universal Soil Loss Equation with measured sediment yield for three 5-ha fields near Swift Current, Saskatchewan. The calibration used predetermined C, LS, P, and warm-season K factors with all erositivities and the K factors for other seasons optimized. Using this calibrated equation with 31 yr of measured runoff data, mean annual sediment yield for a conventional-tillage spring wheat-fallow cropping system on an undulating landscape (0 to 5% slopes) was estimated to be 0.8 Mg ha−1 yr−1. Estimated erosion from fallow land from 16 March to 15 April constituted 86% of the mean annual sediment yield. Erosion occurring during rainfall-runoff were relatively unimportant, accounting for about 4% of the estimated total sediment yield. An exceptionally severe rainstorm was predicted to have caused erosion of many Mg ha−1but the frequency of such events on a given field is too low to greatly affect mean annual erosion. Apparent field-scale soil erodibility was least in the winter (November to 15 March) when the soil would typically be frozen to the surface and was greatest in April when the soil would typically be partially frozen. Apparent erodibility in the summer and in late March was intermediate between those values. Key words: Erosion, runoff, snow, Universal Soil Loss Equation