Estimation Of Sediment Load Research Articles

Sediment assessment for a watershed in arid region via neural networks

In the present study, the estimation of suspended sediment load is computed by four Artificial Neural Networks (ANNs) algorithms, Cascade Forward Back Propagation (CFBP), Feed Forward Back Propagation (FFBP), Radial Basis Function (RBF), and Recurrent Neural Network (RNN). Five cases of model input are calibrated to establish the relationship among precipitation, discharge and suspended sediment load. While discharge and rainfall up to four previous days as employed for input, model gives pre-eminent performance. Sensitivity of all models is appraised concerning Nash-Sutcliffe coefficient (ENS) and coefficient of determination (R2) for predicting sediment load. Among all ANNs, MMF (Morgan-Morgan-Finney) model when trained with stream flow as the input in RNN, gives best result with coefficient of determination, R2 as 0.9474, while the values for FFBP, CFBP and RBF are 0.9115, 0.8766 and 0.8511, respectively. Performance of all results show that an MMF model is superior to conventional SRC (Sediment Rating Curve) and MLR (Multiple Linear Regression) models in determining the complex relationship between discharge and suspended sediment load.

Uncertainty Analysis on Hybrid Double Feedforward Neural Network Model for Sediment Load Estimation with LUBE Method

The assessment of uncertainty prediction has become a necessity for most modeling studies within the hydrology community. This paper addresses uncertainty analysis on a novel hybrid double feedforward neural network (HDFNN) model for generating the sediment load prediction interval (PI). By using the Lower Upper Bound Estimation (LUBE) method, the lower and upper bounds are directly generated as outputs of neural network based models. Coverage Width-based Criterion (CWC) is employed as an objective function for searching high quality PIs. The LUBE-based model is then applied to estimate sediment loads of Muddy Creek in Montana of USA. Results demonstrate the suitability of HDFNN-LUBE model in producing PI in both 90% and 95% confidence levels (CL). It is capable of generating appropriate lower bounds of PIs with narrow intervals. Partitioning analysis reveals consistently excellent performances of HDFNN model in constructing PI in terms of low, medium and high loads. These results therefore verify the reliability and potentiality of the HDFNN model for sediment load estimation with uncertainty. LUBE shows its efficiency in uncertainty prediction as well, which could be used to quantify total uncertainty of data-driven models.

Development of the FCM-SVR Hybrid Model for Estimating the Suspended Sediment Load

The accurate estimation of suspended sediment load (SSL) carried by a river is one of the primary issues in river engineering, water resources, and environment projects. Boundary condition and simplification of some important parameters lead to some limitation on the empirical sediment equations which are based on flow and sediment properties. In this study, the potential of a developed Fuzzy C-mean clustering-support vector regression model, as a FCM-SVR hybrid model, was investigated in comparison with sediment rating curve (SRC), artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), and support vector regression (SVR) models for estimating the daily SSL in Sistan River in Iran. Root mean square error (RMSE), mean absolute error (MAE), determination coefficient (R2), and Nash-Sutcliffe efficiency coefficient (NSE) statistics were utilized to evaluate the accuracy of the models. The result showed that FCM-SVR model estimates SSL (with RMSE, MAE, R2, and NSE equal to 34,415.52 ton/day, 12,256.28 ton/day, 0.922, and 0.918 respectively in testing period) more accurate than other models. The RMSE value, which was 50% lower compared to other models, reveals that this model possesses the lowest error than the other models. Also, the obtained results indicated that the estimated SSLs, using the best FCM-SVR, were in good agreement and linear dependence with observed values. Unlike other models, FCM-SVR model appropriately estimates extreme values of SSL.

ASSESSING SUSPENDED SEDIMENT YIELD IN THE SAF SAF GAUGED CATCHMENT, NORTHEASTERN ALGERIA

Many studies on sediment transport have been carried out on Algerian rivers but few studies have been undertaken in catchments of the Northeast of Algeria. The scarcity or discontinuity on sediment transport measurements reduces knowledge about soil loss. In some cases, researchers find often difficulties to apply the most suitable methods to estimate sediment load. The present work represents an assessment of suspended sediment yield from the Saf Saf catchment (322 km2) over 39 years. Long-term annual suspended sediment loads are estimated using non-linear power model, developed on mean discharge class technique as a sediment rating curve. There is a challenge to estimate suspended sediment load in the Saf Saf catchment, which is distinguished by rapid discharge variation. A second aim is to examine monthly and annual variations in discharge, suspended concentration, rainfall and load in this river and to find causes for these variations. The results show that the mean annual sediment yield is equal to 477 T km-2 yr-1 during the study period. Moreover, the long term variability analysis of sediment load seems to be very high from year to year depending on climatic conditions. The analysis of annual sediment load shows a decreasing trend along 39 years, mainly from 1997. Most sediment loads are transported during the winter season, which represents 78% of the total sediment load. The understanding of sediment transport relationships gained from this study should provide a good starting point for researchers and policymakers to begin addressing sediment issues within the catchment.

Integrated sediment transport process modeling by coupling Soil and Water Assessment Tool and Environmental Fluid Dynamics Code

Reservoir sedimentation is a major environmental issue, and integrated sediment modeling provides a holistic picture of the sediment transport processes in watershed management planning. This study proposes a sediment modeling framework by coupling a watershed loading model, Soil and Water Assessment Tool (SWAT), with a receiving water-body model, Environmental Fluid Dynamics Code (EFDC). In the proposed framework, upstream sediment load estimates made by SWAT are dynamically linked to EFDC to predict sediment deposition in a reservoir, facilitating multiple model runs required for what-if scenario analyses in developing sediment management plans and for optimization using sampling-based search algorithms. The modeling framework was applied to predict the spatial distributions of sediment depositions in an agricultural reservoir and quantify their uncertainty. This article introduces how the two models, SWAT and EFDC, are linked for integrated sediment modeling and demonstrates how the proposed framework can be applied in a case study.

Assessment of potential soil erosion and sediment yield in the semi-arid N′fis basin (High Atlas, Morocco) using the SWAT model

Arid and semi-arid regions face multiple problems in terms of water management, particularly where surface water is a primary resource. On the northern flank of the High Atlas Mountains (Morocco), deforestation has resulted in soil erosion and siltation of reservoirs. Better characterization of sediment yield is important for reducing the impact of siltation and prolonging the lifespan of dams. The Soil and Water Assessment Tool was used for modeling the N′fis basin in the southern Tensift watershed, leading to a better understanding of the rate of siltation behind Lalla Takerkoust dam. Runoff and sediment yield simulations were evaluated using graphical and statistical methods. The SWAT model performed well in estimating sediment load during the calibration period from 1990 to 2015 (Nash–Sutcliffe efficiency = 0.5–0.62, R2 = 0.5–0.61). The model enabled the determination of soil loss within each hydrological response unit in the watershed. The overall rate was approximately 123 t Ha−1 for an average annual rainfall of 315 mm yr−1. This high yield has to be taken into account for effective water-resources management in the N′fis basin.

A performance evaluation of neuro-fuzzy and regression methods in estimation of sediment load of selective rivers

Sediment rating curves (SRCs) have been recognized as the most popular method for estimating sediment in the hydrology of river sediments and in watersheds. In this regard, in order to compare and correct estimation methods of river sediment load, estimated rates of several univariate types of SRCs and a multivariate type of SRCs (MSRCs) were studied using the neuro-fuzzy and tree regression models in five selective hydrometric stations of different climatic zones of Iran and with various indexes of the accuracy (AI) and the precision (PI). The results of the data analysis showed that the mean of the AI of neuro-fuzzy and tree regression models in selective stations is 151 and 536%, respectively, which shows the low efficiency compared with SRCs. Also according to the results, the best rate of the AI of the MSRCs belongs to the Glink station with the rate of 1.12. Also, the average value of the AI of MSRCs is 1.15 which is an acceptable amount of the other considered various methods.

Developing nonlinear models for sediment load estimation in an irrigation canal

The study was performed to estimate the weekly sediment load in Thal canal located in Mianwali district Punjab, Pakistan. Past records of sediments and discharge have been considered as the input parameters. The best input combinations have been identified with the help of advanced algorithms including full, sequential and increasing embedding, genetic algorithm and hill climbing in combination with the gamma test. Model training has been carried out using two artificial neural network-based algorithms, namely Broyden–Fletcher–Goldfarb–Shanno (BFGS), back-propagation and a local linear regression technique. A variety of statistical parameters including R square, root mean squared error, mean square error and mean bias error (MBE) has been calculated in order to evaluate the best models. The results strongly suggested that BFGS-based model performed better than all other models with remarkably low values of MBE. Significantly high values of correlation coefficient (R square) in both training and testing evidenced a close similarity between actual and predicted sediment load values for the same model.

Estimation of suspended sediment load using three neural network algorithms in Ramganga River catchment of Ganga Basin, India

The information on suspended sediments of river is considered to be crucial for issues concerning water management and the environment. The abrupt quantity and nature of sediment loads can be best studied by simultaneously considering the governing variables contributing towards this physical phenomenon. Artificial Neural Network (ANN) is one of the suitable data-mining technique which helps in carrying out the modelling of this phenomenon. In this study, ANNs are employed to approximate the monthly mean suspended sediment load for Ramganga River. Three simulations with rainfall and water discharge data were carried out to predict the suspended sediment load. In terms of the selected performance criteria, three algorithms were evaluated and the results so obtained are presented. It has been found that rainfall values were not sufficient to correctly predict the suspended sediment load. However, considering water discharge values as input improves the performance of all the three considered algorithms.

The impact of persistent dynamics on suspended sediment load estimation

Fluvial suspended sediment concentration and discharge is most commonly estimated by applying rating curves developed over a given base period of sediment monitoring to longer, continuous records of water discharge. This practice invokes the tacit assumption of stationary suspended sediment concentration–water discharge relationships (i.e. suspended sediment behavior). Here the ramifications of this assumption are explored through an investigation into the temporal dependence of suspended sediment concentration–discharge relationships for all hydrologic stations on small rivers draining the west coast of the continental US that have been monitored for suspended sediment over a period of 10 years or more. Persistent suspended sediment behavior was found to be the norm, with 19 of 24 stations expressing apparent monotonic trends over their base periods, and 21 stations with finer scale (subannual to interdecadal) persistent patterns. The identified persistent periods of suspended sediment behavior corresponded to distinct rating curves that differed significantly from other periods in terms of offset, slope, or both. Assuming stationary suspended sediment behavior resulted in 1% to 63% biasing of suspended sediment load. Applying rating curves developed for single persistent periods to broader discharge records resulted in up to 10× potential error in estimated sediment load. The risk of falsely characterizing the long-term (interannual-scale) suspended sediment behavior of a river through a single persistent period of elevated or depressed behavior may be much higher than expected. The average duration of identified persistent periods was 8.6 years, which is longer than the base period of suspended sediment monitoring for most river stations in the US. Peak discharge (high discharge events 20 to 1000 times mean discharge) was usually followed by periods of increased sediment concentration across US Pacific Coast. Seven rivers in Oregon and northern California displayed the same pattern of decadal- to interdecadal-scale elevated suspended sediment concentrations following the Great Christmas Floods of 1964.

Estimation of Sediment Deposition at Nausehri Reservoir by Using Multiple Linear Regression and Assessment of its Effect on Hydropower Generation Capacity

This article presents investigation of the inflow of sediments depending upon the stream flow, density and kinematic viscosity of water. The process involves Multiple Linear Regression (MLR) technique for estimation of sediment load and volume, by using UNISTAT software (Version 5.6). A general equation is developed for a specific flow rate while considering density and kinematic viscosity of water as other parameters. The developed equation can be applied in future for determining the sediment concentration when the flow rate is known. The daily stream flow and suspended sediment data belonging to the station Nausehri operated by the Water and Power Development Authority, Pakistan is used as case study. The results indicate that the dam site experiences a mean annual sediment load of 3.74mst. With this much incoming sediment load the reservoir life comes out to be 45 years when no flushing is done. The impact of this sediment deposition on hydro-electric energy generation is also determined, which shows a decrease of 113MW (12%) in the production capacity of the power plant over the life of the reservoir.

Enhancing SWAT simulation of forest ecosystems for water resource assessment: A case study in the St. Croix River basin

Forests often cover a significant portion of a watershed and play an important role in water, carbon, and nutrient cycling. The widely-used Soil and Water Assessment Tool (SWAT) has not been thoroughly tested for simulating forest ecosystems and associated impacts on watershed-scale water resource assessment. Using the St. Croix River basin as a testbed, we examined SWAT simulations of forest biomass, net primary productivity (NPP), evapotranspiration (ET), and water yield with default and improved model settings. Model simulations with the default parameterization significantly underestimated forest biomass, ET, and NPP relative to benchmark data. With adjusted forest growth and development parameters, including radiation use efficiency, temperature control on plant growth, and leaf to biomass fraction, as well as a new phosphorus supply module, the SWAT model’s performance with respect to the above variables was substantially improved. These improvements also enhanced SWAT estimates of water yield, streamflow, and sediment loads at the watershed scale. Results of this study show the important role of forest ecosystems in watershed scale water balance and the connections between terrestrial and aquatic processes via lateral water and sediment fluxes. We suggested that watershed models, such as SWAT, should be assessed for both terrestrial and aquatic processes in order to provide credible water quantity and quality simulations at the watershed scale. Calibrated forest parameters and the phosphorus input algorithms will be applicable to future SWAT simulations of watershed hydrology.

A comparison of artificial intelligence models for the estimation of daily suspended sediment load: a case study on the Telar and Kasilian rivers in Iran

Abstract The estimation of the suspended sediment load in rivers is one of the main issues in hydraulic engineering. Different traditional methods such as the sediment rating curve (SRC) can be used to estimate the suspended sediment load of rivers. The main problem with this method is its low accuracy and uncertainty. In this study, the ability of three intelligence models namely: gene expression programming (GEP), artificial neural networks (ANN) and adaptive neuro fuzzy inference system (ANFIS) were compared with the SRC method. The daily flow discharge and sediment discharge at two hydrometric stations of the Kasilian and Telar rivers in the period of 1964–2014 were used to develop intelligence models. The performance of these methods indicated that all intelligence models give reliable results in the estimation of the suspended sediment load and their performance was better than the SRC method. Moreover, results showed that the GEP model with a high coefficient of determination (R2) and a low mean absolute error (MAE) was better than both the ANN and ANFIS models for the estimation of daily suspended sediment load of the two sub-basins of the Kasilian and Telar rivers.

Case Study of the Integrated Model for Estimation of Sediment Load in Artificial River Channel

AbstractSedimentation is one of the most important factors affecting stream channel stability. A proposed model was developed to estimate the sediment load of an artificial channel by the integrati...

Identifying a reliable method for estimating suspended sediment load in a temporary river system

Sampling strategies and methods used for estimating load can lead to large uncertainties in suspended sediment transport quantification, especially in rivers with a high variability in streamflow. The aim of this paper is to evaluate suspended sediment load, using a number of direct estimation techniques, in order to find a suitable method for temporary river systems, and to assess the uncertainty associated with load estimation, due to the specific method applied. One year of continuous measurement of flow, and discrete sampling (n = 216) of suspended sediment concentrations, taken from 2010 to 2011 in the Celone River (SE, Italy), were used to estimate annual load. Averaging, ratio, and regression estimator methods were applied to the entire dataset, and to subsets of data, to calculate load. The results show a wide range of values, from 220 to 1123 t km−2 yr−1, with respect to the applied suspended sediment load estimation techniques. Averaging methods resulted biased. Sediment rating curves underestimated load, while, if the back-transformation bias correction was used, load was overestimated. The ratio methods generally overestimate load. Increased precision and accuracy was achieved through applying data stratification, based on flow regime and seasonality. After applying three different flow regime stratifications, the annual load ranged from 240 to 606 t km−2 yr−1 and, using seasonal stratification, from 258 to 974 t km−2 yr−1. It seems that ratio estimator methods, and the regression equations applied to the stratification on a flow regime basis, are more suitable for estimating load in temporary, flashy streams.

Comparison of River Suspended Sediment Load Estimation, using Regression and GA Methods

Comparison of River Suspended Sediment Load Estimation, using Regression and GA Methods

Assessing uncertainty in annual nitrogen, phosphorus, and suspended sediment load estimates in three agricultural streams using a 21-year dataset.

Accurate estimation of constituent loads is important for studies of ecosystem mass balance or total maximum daily loads. In response, there has been an effort to develop methods to increase both accuracy and precision of constituent load estimates. The relationship between constituent concentration and stream discharge is often complicated, potentially leading to high uncertainty in load estimates for certain constituents, especially at longer-term (annual) scales. We used the loadflex R package to compare uncertainty in annual load estimates from concentration vs. discharge relationships in constituents of interest in agricultural systems, including ammonium as nitrogen (NH4-N), nitrate as nitrogen (NO3-N), soluble reactive phosphorus (SRP), and suspended sediments (SS). We predicted that uncertainty would be greatest in NO3-N and SS due to complex relationships between constituent concentration and discharge. We also predicted lower uncertainty with a composite method compared to regression or interpolation methods. Contrary to predictions, we observed the lowest uncertainty in annual NO3-N load estimates (relative error 1.5-23%); however, uncertainty was greatest in SS load estimates, consistent with predictions (relative error 19-96%). For all constituents, we also generally observed reductions in uncertainty by up to 34% using the composite method compared to regression and interpolation approaches, as predicted. These results highlight differences in uncertainty among different constituents and will aid in model selection for future studies requiring accurate and precise estimates of constituent load.

Evaluation of suspended sediment concentration using descent neural networks

This study includes three methods having considerable differences with each other and with experimental observations, because the sediment measures have certain limits. The equations relating to sediment transport are used in estimating sediment load. In the present study, black box models, ANN (Artificial Neural Network) are used for the simulation of the suspended sediment load. Hence, models which give the lowest RMSE and highest R2 are considered to be the best model for this study. The lowest values of RMSE based on normalized data for Feed forward back propagation, Cascade forward back propagation and neural network fitting are 0.00873, 0.00834 and 0.01193 respectively. The corresponding values of R2 are 0.9304, 0.9713 and 0.9831 respectively for the cited MSE. The study shows Neural Network Fitting model is superior to the other models. However, a drawback of neural network fitting is that it produces few negative estimates, which is not at all tolerable in the field of estimation of sediment load and hence these models are not crowned in this study.

Estimation of suspended sediment load in different time steps using hybrid wavelet-ANFIS

The aim of this study was to predict suspended sediment load for one and two months ahead using hybrid wavelet-ANFIS model. For this purpose, in the first step, the raw data were imposed to the ANFIS model and modelling was carried out. Afterwards, the data were decomposed at different levels and by different mother wavelets and the obtained coefficients were imported to the ANFIS model. The results indicate the significant impact of one-dimensional wavelet analysis on the performance of ANFIS model and the acceptable performance of hybrid wavelet-ANFIS in modelling sediment for one and two months ahead. The results also suggest reducing the accuracy of the model by increasing the time step from one to two months. Results showed that the hybrid wavelet-ANFIS model had the best performance for predicting sediment on month ahead with a modified correlation coefficient of 0.97 and RMSE of 0.71.

Estimation of suspended sediment load in different time steps using hybrid wavelet-ANFIS

Estimation of suspended sediment load in different time steps using hybrid wavelet-ANFIS