Supporting USLE‐MM reliability by analyzing soil loss measurement errors

Sampling the collected suspension in a storage tank is a common procedure to obtain soil loss data. A calibration curve of the tank has to be used to obtain actual concentration values from those measured by sampling. However, literature suggests that using a tank calibration curve was not a common procedure in the past. For the clay soil of the Sparacia (Italy) experimental station, this investigation aimed to establish a link between the relative performances of the USLE‐M and USLE‐MM models, usable to predict plot soil loss at the event temporal scale, and soil loss measurement errors. Using all available soil loss data, lower soil loss prediction errors were obtained with the USLE‐MM (exponent of the erosivity term, b₁ > 1) than the USLE‐M (b₁ = 1). A systematic error of the soil loss data is unexpected for the Sparacia soil because the calibration curve does not depend on the water level in the tank. In any case, this type of error does not have any effect on the b₁ exponent. Instead, this exponent decreases as the level of underestimation increases for increasing soil loss values. This type of error can occur at Sparacia if it is assumed that a soil loss measurement can be obtained by a bottle sampler dipped close to the bottom of the tank after mixing the suspension and assuming that the measured concentration coincides with the actual one. In this case, the risk is to obtain a lower b₁ value than the actual one. In conclusion, additional investigations on the factors determining errors in soil loss data collected by a sampling procedure are advisable because these errors can have a noticeable effect on the calibrated empirical models for soil loss prediction.     

Comparing theoretically supported rainfall‐runoff erosivity factors at the Sparacia (South Italy) experimental site

Interpreting rainfall‐runoff erosivity by a process‐oriented scheme allows to conjugate the physical approach to soil loss estimate with the empirical one. Including the effect of runoff in the model permits to distinguish between detachment and transport in the soil erosion process. In this paper, at first, a general definition of the rainfall‐runoff erosivity factor REF_e including the power of both event runoff coefficient Q_R and event rainfall erosivity index EI₃₀ of the Universal Soil Loss Equation (USLE) is proposed. The REF_e factor is applicable to all USLE‐based models (USLE, Modified USLE [USLE‐M] and Modified USLE‐M [USLE‐MM]) and it allows to distinguish between purely empirical models (e.g., Modified USLE‐M [USLE‐MM]) and those supported by applying theoretical dimensional analysis and self‐similarity to Wischmeier and Smith scheme. This last model category includes USLE, USLE‐M, and a new model, named USLE‐M based (USLE‐MB), that uses a rainfall‐runoff erosivity factor in which a power of runoff coefficient multiplies EI₃₀. Using the database of Sparacia experimental site, the USLE‐MB is parameterized and a comparison with soil loss data is carried out. The developed analysis shows that USLE‐MB (characterized by a Nash–Sutcliffe Efficiency Index NSEI equal to 0.73 and a root mean square error RMSE = 11.7 Mg ha^?1) has very similar soil loss estimate performances as compared with the USLE‐M (NSEI = 0.72 and RMSE = 12.0 Mg ha^?1). However, the USLE‐MB yields a maximum discrepancy factor between predicted and measured soil loss values (176) that is much lower than that of USLE‐M (291). In conclusion, the USLE‐MB should be preferred in the context of theoretically supported USLE type models.     

Plot‐scale measurement of soil erosion at the experimental area of Sparacia (southern Italy)

Obtaining good quality soil loss data from plots requires knowledge of the factors that affect natural and measurement data variability and of the erosion processes that occur on plots of different sizes. Data variability was investigated in southern Italy by collecting runoff and soil loss from four universal soil‐loss equation (USLE) plots of 176 m², 20 ‘large’ microplots (0·16 m²) and 40 ‘small’ microplots (0·04 m²). For the four most erosive events (event erosivity index, R_e ≥ 139 MJ mm ha^?1 h^?1), mean soil loss from the USLE plots was significantly correlated with R_e. Variability of soil loss measurements from microplots was five to ten times greater than that of runoff measurements. Doubling the linear size of the microplots reduced mean runoff and soil loss measurements by a factor of 2·6–2·8 and increased data variability. Using sieved soil instead of natural soil increased runoff and soil loss by a factor of 1·3–1·5. Interrill erosion was a minor part (0·1–7·1%) of rill plus interrill erosion. The developed analysis showed that the USLE scheme was usable to predict mean soil loss at plot scale in Mediterranean areas. A microplot of 0·04 m² could be used in practice to obtain field measurements of interrill soil erodibility in areas having steep slopes. Copyright © 2003 John Wiley & Sons, Ltd.     

Testing the Universal Soil Loss Equation‐MB equation in plots in Central and South Italy

Planning soil conservation strategies requires predictive techniques at event scale because a large percentage of soil loss over a long‐time period is due to relatively few large storms. Considering runoff is expected to improve soil loss predictions and allows relation of the process‐oriented approach with the empirical one, furthermore, the effects of detachment and transport on soil erosion processes can be distinguished by a runoff component. In this paper, the empirical model USLE‐MB (USLE‐M based), including a rainfall‐runoff erosivity factor in which the event rainfall erosivity index EI₃₀ of the Universal Soil Loss Equation (USLE) multiplies the runoff coefficient Q_R raised to an exponent b₁ > 1 is tested by the measurements carried out for the Masse (10 plots) and Sparacia (22 plots) experimental stations in Italy. For the Masse experimental station, an exponent b₁ > 1 was also estimated by tests carried out by a nozzle‐type rainfall simulator. For each experimental site in fallow conditions, the effect of the sample size of the plot soil loss measurements on the estimate of the b₁ coefficient was also studied by the extraction of a fixed number N of randomly obtained pairs of the normalized soil loss and runoff coefficient. The analysis showed that the variability of b₁ with N is low and that 350 pairs are sufficient to obtain a stable estimate of b₁. A total of 1,262 soil loss data were used to parameterize the model both locally and considering the two sites simultaneously. The b₁ exponent varied between the two sites (1.298–1.520), but using a common exponent (1.386) was possible. Using a common b₁ exponent for the two experimental areas increases the practical interest for the model and allows the estimation of a baseline component of the soil erodibility factor, which is representative of the at‐site soil intrinsic and quasi‐static properties. Development of a single USLE‐MB model appears possible, and sampling other sites is advisable to develop a single USLE‐MB model for general use.     

A comprehensive analysis of Universal Soil Loss Equation-based models at the Sparacia experimental area

Improving Universal Soil Loss Equation (USLE)-based models has large interest because simple and reliable analytical tools are necessary in the perspective of a sustainable land management. At first, in this paper, a general definition of the event rainfall- runoff erosivity factor for the USLE-based models, REF_e = (Q_R)^b1(EI₃₀)^b2, in which Q_R is the event runoff coefficient, EI₃₀ is the single-storm erosion index, and b₁ and b₂ are coefficients, was introduced. The rainfall-runoff erosivity factors of the USLE (b₁ = 0 and b₂ = 1), USLE-M (b₁ = b₂ = 1), USLE-MB (b₁ ≠ 1 and b₂ = 1), USLE-MR (b₁ = 1 and b₂ ≠ 1), USLE-MM (b₁ = b₂ ≠ 1), and USLE-M2 (b₁ ≠ b₂ ≠ 1) can be defined using REF_e. Then the different expressions of REF_e were simultaneously tested against a data set of normalized bare plot soil losses, A_eN, collected at the Sparacia (south Italy) site. As expected, the poorest A_eN predictions were obtained with the USLE. The observed tendency of this model to overestimate small A_eN values and underestimate high A_eN values was reduced by introducing in the soil loss prediction model both Q_R and an exponent for the erosivity term. The fitting to the data was poor with the USLE-MR as compared with the USLE-MB and the USLE-MM. Estimating two distinct exponents (USLE-M2) instead of a single exponent (USLE-MB, USLE-MR, and USLE-MM) did not appreciably improve soil loss prediction. The USLE-MB and the USLE-MM were recognized to be the best performing models among the possible alternatives, and they performed similarly with reference to both the complete data set and different sub-data sets, only including small, intermediate, and severe erosion events. In conclusion, including the runoff coefficient in the soil loss prediction model is important to improve the quality of the predictions, but a great importance has to be paid to the mathematical structure of the model.     

Variation in curve numbers derived from plot runoff data for New South Wales (Australia)

The curve number method is a simple one parameter (the curve number) rainfall runoff model. While its theoretical underpinning has been questioned it remains a powerful hydrological tool in the absence of detailed data and is therefore used extensively in hydrological models. This study aims to characterize the variation in maximum retention values (S), which underlie curve numbers, for a range of agricultural treatments across a large spatial area in New South Wales (NSW), Australia. The data used for the analysis spans several decades of rainfall runoff observations. A range of different derivation methods result in variation in mean and variance of S. In particular, methods that emphasize the larger storms result in greater S and thus lower runoff. For larger spatial scales, emphasis on larger storms gives more reliable estimates of S. Systematic variation in S arises from variations in treatment, pre‐runoff soil moisture, rainfall depth, and variations in cover. On the basis of the analysis, a table of curve number values for different land uses found in NSW is presented. The resulting distributions of S and curve numbers provide guidance for rainfall runoff modelling studies in the agricultural important areas of NSW. Copyright © 2011 John Wiley & Sons, Ltd.     

Using a transient infiltrometric technique for intensively sampling field‐saturated hydraulic conductivity of a clay soil in two runoff plots

The point measurement of soil properties allows to explain and simulate plot scale hydrological processes. An intensive sampling was carried out at the surface of an unsaturated clay soil to measure, on two adjacent plots of 4 × 11 m² and two different dates (May 2007 and February–March 2008), dry soil bulk density, ρ_b, and antecedent soil water content, θ_i, at 88 points. Field‐saturated soil hydraulic conductivity, K_fs, was also measured at 176 points by the transient Simplified Falling Head technique to determine the soil water permeability characteristics at the beginning of a possible rainfall event yielding measurable runoff. The ρ_b values did not differ significantly between the two dates, but wetter soil conditions (by 31%) and lower conductivities (1.95 times) were detected on the second date as compared with the first one. Significantly higher (by a factor of 1.8) K_fs values were obtained with the 0.30‐m‐diameter ring compared with the 0.15‐m‐diameter ring. A high K_fs (> 100 mm h^?1) was generally obtained for low θ_i values (< 0.3 m³m^?3), whereas a high θ_i yielded an increased percentage of low K_fs data (1–100 mm h^?1). The median of K_fs for each plot/sampling date combination was not lower than 600 mm h^?1, and rainfall intensities rarely exceeded 100 mm h^?1 at the site. The occurrence of runoff at the base of the plot needs a substantial reduction of the surface soil permeability characteristics during the event, probably promoted by a higher water content than the one of this investigation (saturation degree = 0.44–0.62) and some soil compaction due to rainfall impact. An intensive soil sampling reduces the risk of an erroneous interpretation of hydrological processes. In an unstable clay soil, changes in K_fs during the event seem to have a noticeable effect on runoff generation, and they should be considered for modeling hydrological processes. Copyright © 2012 John Wiley & Sons, Ltd.     

Testing the use of an image‐based technique to measure gully erosion at Sparacia experimental area

The first part of this investigation was aimed at testing the use of a three‐dimensional (3D) digital terrain model and a quasi‐tridimensional (2.5D) digital elevation model obtained by a large series of oblique images of eroded channels taken from consumer un‐calibrated and non‐metric cameras. For two closed earth channels having a different sinuosity, the ground measurement of some cross sections by a profilometer (P) was carried out and their real volume was also measured. The comparison among the three methods (3D, 2.5D, and P) pointed out that a limited underestimation of the total volume always occurs and that the 3D method is characterized by the minimum difference between measured and real volume. For this reason, 3D model can be used as benchmark. In the subsequent part of the investigation, the three ground measurement methods were applied for surveying of an ephemeral gully (EG) channel at the Sparacia area. The morphological and hydraulic variable values of the 24 surveyed cross sections determined by both 2.5D model and profilometer were compared. This comparison showed that the estimate error is generally less than ±10%. The EG measurements carried out by the three methods supported the applicability both of the empirical relationship between EG length and its eroded volume and the theoretical dimensionless relationship among the morphological variables describing the channelized erosion process. Finally, it was demonstrated that the effect of the distance interval on the EG volume measurement by 3D and 2.5D models is negligible for the investigated EG.     

Spatio‐temporal distribution of near‐surface and root zone soil moisture at the catchment scale

Soil moisture is highly variable both spatially and temporally. It is widely recognized that improving the knowledge and understanding of soil moisture and the processes underpinning its spatial and temporal distribution is critical. This paper addresses the relationship between near‐surface and root zone soil moisture, the way in which they vary spatially and temporally, and the effect of sampling design for determining catchment scale soil moisture dynamics. In this study, catchment scale near‐surface (0–50 mm) and root zone (0–300 mm) soil moisture were monitored over a four‐week period. Measurements of near‐surface soil moisture were recorded at various resolutions, and near‐surface and root zone soil moisture data were also monitored continuously within a network of recording sensors. Catchment average near‐surface soil moisture derived from detailed spatial measurements and continuous observations at fixed points were found to be significantly correlated (r² = 0·96; P = 0·0063; n = 4). Root zone soil moisture was also found to be highly correlated with catchment average near‐surface, continuously monitored (r² = 0·81; P < 0·0001; n = 26) and with detailed spatial measurements of near‐surface soil moisture (r² = 0·84). The weaker relationship observed between near‐surface and root zone soil moisture is considered to be caused by the different responses to rainfall and the different factors controlling soil moisture for the soil depths of 0–50 mm and 0–300 mm. Aspect is considered to be the main factor influencing the spatial and temporal distribution of near‐surface soil moisture, while topography and soil type are considered important for root zone soil moisture. The ability of a limited number of monitoring stations to provide accurate estimates of catchment scale average soil moisture for both near‐surface and root zone is thus demonstrated, as opposed to high resolution spatial measurements. Similarly, the use of near‐surface soil moisture measurements to obtain a reliable estimate of deeper soil moisture levels at the small catchment scale was demonstrated. Copyright © 2007 John Wiley & Sons, Ltd.     

GIS‐based modelling of soil erosion processes using the modified‐MMF (MMMF) model in a large watershed having vast agro‐climatological differences

The present study demonstrates a spatially distributed application of a field‐scale annual soil loss model, the modified‐MMF (MMMF), to a large watershed using hydrological routing techniques, remote sensing data and geospatial technologies. In this study, the MMMF model is implemented after incorporating the corrections suggested in recent literature along with appropriate modifications of the model to suit the agro‐climatological conditions prevailing in most parts of India. Sensitivity analysis carried out through an Average Linear Sensitivity approach indicates that the model outputs are highly sensitive to soil moisture (MS), bulk density (BD), effective hydraulic depth (EHD), ground cover (GC) and settling velocity for clay (VS_c). During calibration and validation, the performance evaluation statistics are mostly in the range of very good to satisfactory for both runoff and soil loss at the watershed outlet. Even spatial validation of the results of intermediate processes in the water phase and the sediment phase, although qualitative, seems to be reasonable and rational. Furthermore, the soil erosion severity analysis for different land‐uses existing in the watershed indicates that about 90% of the watershed area, especially that occupied by agricultural lands, is vulnerable to the long‐term effects of soil erosion. Copyright © 2018 John Wiley & Sons, Ltd.     

Evaluating the effect of land use changes on soil erosion and sediment yield using a grid‐based distributed modelling approach

Processes of soil erosion and sediment transport are strongly influenced by land use changes so the modelling of land use changes is important with respect to the simulation of soil degradation and its on‐site and off‐site consequences. The reliability of simulation results from erosion models is circumscribed by considerable spatial variation in many parameters. However, most of the currently widely used erosion models at the mesoscale are semidistributed, which leads to difficulties in incorporating a high degree of spatial information, especially land use information, so that the effects of land use changes on soil erosion have hitherto not been investigated in detail using these models. In this article, a grid‐based distributed erosion and sediment transport model is introduced, which simulates the spatial pattern of erosion and deposition rates and sediment transport processes in river channels. In this model, land use affects soil erosion through altering soil loss and influencing sediment delivery. Simulated soil erosion for events recorded in 1989 and 1996 in the Lushi basin in China was analyzed by comparing it with historical land use maps. The results indicated that even relatively minor land use changes had a significant effect on regional soil erosion rates and sediment transport to rivers. The average erosion rate increased from 1989 to 1996, after the transformation of forest to farmland. The results of the study suggest that the proposed soil erosion model can be applied in similar river basins. Copyright © 2011 John Wiley & Sons, Ltd.     

Semi‐empirical model for site effects on acceleration time histories at soft‐soil sites. Part 2: calibration

A previously developed simplified model of ground motion amplification is applied to the simulation of acceleration time histories at several soft‐soil sites in the Valley of Mexico, on the basis of the corresponding records on firm ground. The main objective is to assess the ability of the model to reproduce characteristics such as effective duration, frequency content and instantaneous intensity. The model is based on the identification of a number of parameters that characterize the complex firm‐ground to soft‐soil transfer function, and on the adjustment of these parameters in order to account for non‐linear soil behavior. Once the adjusted model parameters are introduced, the statistical properties of the simulated and the recorded ground motions agree reasonably well. For the sites and for the seismic events considered in this study, it is concluded that non‐linear soil behavior may have a significant effect on the amplification of ground motion. The non‐linear soil behavior significantly affects the effective ground motion duration for the components with the higher intensities, but it does not have any noticeable influence on the lengthening of the dominant ground period. Copyright © 2004 John Wiley & Sons, Ltd.     

Semi‐empirical model for site effects on acceleration time histories at soft‐soil sites. Part 1: formulation and development

A criterion is developed for the simulation of realistic artificial ground motion histories at soft‐soil sites, corresponding to a detailed ground motion record at a reference firm‐ground site. A complex transfer function is defined as the Fourier transform of the ground acceleration time history at the soft‐soil site divided by the Fourier transform of the acceleration record at the firm‐ground site. Working with both the real and the imaginary components of the transfer function, and not only with its modulus, serves to keep the statistical information about the wave phases (and, therefore, about the time variation of amplitudes and frequencies) in the algorithm used to generate the artificial records. Samples of these transfer functions, associated with a given pair of soft‐soil and firm‐ground sites, are empirically determined from the corresponding pairs of simultaneous records. Each function included in a sample is represented as the superposition of the transfer functions of the responses of a number of oscillators. This formulation is intended to account for the contributions of trains of waves following different patterns in the vicinity of both sites. The properties of the oscillators play the role of parameters of the transfer functions. They vary from one seismic event to another. Part of the variation is systematic, and can be explained in terms of the influence of ground motion intensity on the effective values of stiffness and damping of the artificial oscillators. Another part has random nature; it reflects the random characteristics of the wave propagation patterns associated with the different events. The semi‐empirical model proposed recognizes both types of variation. The influence of intensity is estimated by means of a conventional one‐dimensional shear wave propagation model. This model is used to derive an intensity‐dependent modification of the values of the empirically determined model parameters in those cases when the firm‐ground earthquake intensity used to determine these parameters differs from that corresponding to the seismic event for which the simulated records are to be obtained. Copyright © 2004 John Wiley & Sons, Ltd.     

Water repellency as run‐off and soil detachment controlling factor in a dry‐Mediterranean hillslope (South of Spain)

This study examines the effect of water repellency on controlling temporal variability of runoff generation mechanisms and soil detachment on metamorphic derived soils under dry‐Mediterranean climate. The research is carried out in an unburnt Mediterranean hillslope in souther Spain characterized by a patchy vegetation pattern and shallow soils. The Water Drop Penetration Time test (WDPT) is applied to measure water repellency at the end of summer (Sep‐2008), mid autumn (Nov‐2008) and mid winter (Feb‐2009). Rainfall simulations were used to obtain runoff generation and soil detachment in the same periods of time. The main shrub specie is Cistus monspeliensis which leaves a load of litter during the summer due to the lack of water. This great amount of organic material is accumulated under the shrubs triggering an extreme water repellence (WDPT > 6,000 s) that limits infiltration processes. This process is enforced due to the low soil water content at the end of dry season. Certain water repellency (WDPT > 1,500 s) is also observed on bare soil as consequence of their sandier texture and the accumulation of annual plants which die at the end of the wet season. Soil moisture increases during the autumn and water repellency disappears in both shrub and bare soil at the middle of the wet season (WDPT < 5 s). The main consequence is that the temporal trend of water repellency controls the mechanism and frequency of runoff generation and, hence, soil detachment. At the end of the summer, Hortonian mechanisms predominates when water repellency is extreme, even in soils under Cistus monspeliensis where runoff generation can reach higher peaks of overland flow and sediment concentration. Conversely, only the saturation of soil could generate runoff during the wet season being this quite less frequent in bare soil and absent in shrub. Copyright © 2010 John Wiley & Sons, Ltd.     

A macro‐element model for non‐linear soil–shallow foundation–structure interaction under seismic loads: theoretical development and experimental validation on large scale tests

In this paper, different formulations of a macro‐element model for non‐linear dynamic soil‐structure interaction analyses of structures lying on shallow foundations are first reviewed, and secondly, a novel formulation is introduced, which combines some of the characteristics of previous approaches with several additional features. This macro‐element allows one to model soil‐footing geometric (uplift) and material (soil plasticity) non‐linearities that are coupled through a stiffness degradation model. Footing uplift is introduced by a simple non‐linear elastic model based on the concept of effective foundation width, whereas soil plasticity is treated by means of a bounding surface approach in which a vertical load mapping rule is implemented. This mapping is particularly suited for the seismic loading case for which the proposed model has been conceived. The new macro‐element is subsequently validated using cyclic and dynamic large‐scale laboratory tests of shallow foundations on dense sand, namely: the TRISEE cyclic tests, the Public Works Research Institute and CAMUS IV shaking table tests. Based on this comprehensive validation process against a set of independent experimental results, a unique set of macro‐element parameters for shallow foundations on dense sand is proposed, which can be used to perform predictive analyses by means of the present model. Copyright © 2011 John Wiley & Sons, Ltd.     

A detailed post-IR IRSL chronology for the last interglacial soil at the Jingbian loess site (northern China)

The chronology of dust deposition and climate during the last interglacial is poorly known on the Chinese Loess Plateau. Here, 51 samples were taken from the ∼5 m S1 palaeosol (MIS5) at the desert marginal Jingbian site to develop what is currently the most detailed S1 chronology on the Plateau using instrumental dating techniques. We use the post-IR IRSL signal from sand-sized grains of K-rich feldspar. Signal resetting in the agricultural layer shows that it is possible to almost completely zero this signal in nature. First IR stimulation plateau measurements show that there is no clear dependence of D_e on first IR stimulation temperature between 50 and 260 °C suggesting negligible signal fading. Resultant ages are consistent with a last interglacial age (∼130 to ∼75 ka) and are also consistent within errors with continuous linear sedimentation rates. The average mass accumulation rate for S1 is ∼150 g m⁻² a⁻¹, considerably higher than at many other sites but within the overall range of Loess Plateau estimates. The remarkably stable sediment accumulation at the site contrasts with a more complex record of environmental and monsoonal change recorded in grain-size and magnetic susceptibility.     

Using data‐driven methods to explore the predictability of surface soil moisture with FLUXNET site data

Soil moisture (SM) is a key variable of land surface‐atmosphere interactions. Data‐driven methods have been used to predict SM, but the predictability of SM has not been well evaluated. This study investigated what variables and methods can be used to better predict SM for leading times of 7 days or longer with a global coverage of FLUXNET site data for the first time. Three machine‐learning models, that is, Bayesian linear regression, random forest, and gradient boosting regression tree, are used for the prediction. Variables including atmospheric forcing, surface soil temperature, time variables (year, day of year, and hour), the Fourier transformation of time variables, and lagged SM (7‐ to 14‐day lagged) were sequentially added into models. A framework with five experiments is designed for factorial exploration of SM predictability. A stepwise method was used to build the best models for each site. The performance of regression models became better when adding more explaining variables in most cases. The results showed that from 50 to 95% of variation of the best models can be explained. The important explaining variables are lagged surface SM, followed by day of year, year, soil temperature, and atmospheric forcing. The predictability of SM depends highly on SM memory characteristics and the persistence of seasonality. The effect of SM memory characteristics on SM prediction as an initial condition question has been widely discussed in this paper. Our results also provide an insight that mechanisms of seasonality effects on SM should be also paid more attention to.     

SCALES: a large‐scale assessment model of soil erosion hazard in Basse‐Normandie (northern‐western France)

The cartography of erosion risk is mainly based on the development of models, which evaluate in a qualitative and quantitative manner the physical reproduction of the erosion processes (CORINE, EHU, INRA). These models are mainly semi‐quantitative but can be physically based and spatially distributed (the Pan‐European Soil Erosion Risk Assessment, PESERA). They are characterized by their simplicity and their applicability potential at large temporal and spatial scales. In developing our model SCALES (Spatialisation d'éChelle fine de l'ALéa Erosion des Sols/large‐scale assessment and mapping model of soil erosion hazard), we had in mind several objectives: (1) to map soil erosion at a regional scale with the guarantee of a large accuracy on the local level, (2) to envisage an applicability of the model in European oceanic areas, (3) to focus the erosion hazard estimation on the level of source areas (on‐site erosion), which are the agricultural parcels, (4) to take into account the weight of the temporality of agricultural practices (land‐use concept). Because of these objectives, the nature of variables, which characterize the erosion factors and because of its structure, SCALES differs from other models. Tested in Basse‐Normandie (Calvados 5500 km²) SCALES reveals a strong predisposition of the study area to the soil erosion which should require to be expressed in a wet year. Apart from an internal validation, we tried an intermediate one by comparing our results with those from INRA and PESERA. It appeared that these models under estimate medium erosion levels and differ in the spatial localization of areas with the highest erosion risks. SCALES underlines here the limitations in the use of pedo‐transfer functions and the interpolation of input data with a low resolution. One must not forget however that these models are mainly focused on an interregional comparative approach. Therefore the comparison of SCALES data with those of the INRA and PESERA models cannot result on a convincing validation of our model. For the moment the validation is based on the opinion of local experts, who agree with the qualitative indications delivered by our cartography. An external validation of SCALES is foreseen, which will be based on a thorough inventory of erosion signals in areas with different hazard levels. Copyright © 2010 John Wiley & Sons, Ltd.     

On the use of multi‐algorithm,genetically adaptive multi‐objective method for multi‐site calibration of the SWAT model

With the availability of spatially distributed data, distributed hydrologic models are increasingly used for simulation of spatially varied hydrologic processes to understand and manage natural and human activities that affect watershed systems. Multi‐objective optimization methods have been applied to calibrate distributed hydrologic models using observed data from multiple sites. As the time consumed by running these complex models is increasing substantially, selecting efficient and effective multi‐objective optimization algorithms is becoming a nontrivial issue. In this study, we evaluated a multi‐algorithm, genetically adaptive multi‐objective method (AMALGAM) for multi‐site calibration of a distributed hydrologic model—Soil and Water Assessment Tool (SWAT), and compared its performance with two widely used evolutionary multi‐objective optimization (EMO) algorithms (i.e. Strength Pareto Evolutionary Algorithm 2 (SPEA2) and Non‐dominated Sorted Genetic Algorithm II (NSGA‐II)). In order to provide insights into each method's overall performance, these three methods were tested in four watersheds with various characteristics. The test results indicate that the AMALGAM can consistently provide competitive or superior results compared with the other two methods. The multi‐method search framework of AMALGAM, which can flexibly and adaptively utilize multiple optimization algorithms, makes it a promising tool for multi‐site calibration of the distributed SWAT. For practical use of AMALGAM, it is suggested to implement this method in multiple trials with relatively small number of model runs rather than run it once with long iterations. In addition, incorporating different multi‐objective optimization algorithms and multi‐mode search operators into AMALGAM deserves further research. Copyright © 2009 John Wiley & Sons, Ltd.