Neuro-fuzzy models employing wavelet analysis for suspended sediment concentration prediction in rivers

Abstract

The study of sediment load is important for its implications to the environment and water resources engineering. Four models were considered in the study of suspended sediment concentration prediction: artificial neural networks (ANNs), neuro-fuzzy model (NF), conjunction of wavelet analysis and neuro-fuzzy (WNF) model, and the conventional sediment rating curve (SRC) method. Using data from a US Geological Survey gauging station, the suspended sediment concentration predicted by the WNF model was in satisfactory agreement with the measured data. Also the proposed WNF model generated reasonable predictions for the extreme values. The cumulative suspended sediment load estimated by this model was much higher than that predicted by the other models, and is close to the observed data. However, in the current modelling, the ANN, NF and SRC models underestimated sediment load. The WNF model was successful in reproducing the hysteresis phenomenon, but the SRC method was not able to model this behaviour. In general, the results showed that the NF model performed better than the ANN and SRC models.

Citation Mirbagheri, S. A., Nourani, V., Rajaee, T. & Alikhani, A. (2010) Neuro-fuzzy models employing wavelet analysis for suspended sediment concentration prediction in rivers. Hydrol. Sci. J. 55(7), 1175–1189.     

Long-range forecasting of suspended sediment

This study examines the employment of two methods, multiple linear regression (MLR) and an artificial neural network (ANN), for multistep ahead forecasting of suspended sediment. The autoregressive integrated moving average (ARIMA) model is considered for one-step ahead forecasting of sediment series in order to provide a comparison with the MLR and ANN methods. For one- and two-step ahead forecasting, the ANN model performance is superior to that of the MLR model. For longer ranges, MLR models provide better accuracy, but there is an important assumption violation. The Durbin-Watson statistics of the MLR models show a noticeable decrease from 1.3 to 0.5, indicating that the residuals are not dependent over time. The scatterplots of the three methods (MLR, ARIMA and ANN) for one-step ahead forecasting for the validation period illustrate close fits with the regression line, with the ANN configuration having a slightly higher R² value.     

Artificial neural networks for sheet sediment transport

Abstract

Sheet sediment transport was modelled by artificial neural networks (ANNs). A three-layer feed-forward artificial neural network structure was constructed and a back-propagation algorithm was used for the training of ANNs. Event-based, runoff-driven experimental sediment data were used for the training and testing of the ANNs. In training, data on slope and rainfall intensity were fed into the network as inputs and data on sediment discharge were used as target outputs. The performance of the ANNs was tested against that of the most commonly used physically-based models, whose transport capacity was based on one of the dominant variables—flow velocity (V), shear stress (SS), stream power (SP), and unit stream power (USP). The comparison results revealed that the ANNs performed as well as the physically-based models for simulating nonsteady-state sediment loads from different slopes. The performances of the ANNs and the physically-based models were also quantitatively investigated to estimate mean sediment discharges from experimental runs. The investigation results indicated that better estimations were obtained for V over mild and steep slopes, under low rainfall intensity; for USP over mild and steep slopes, under high rainfall intensity; for SP and SS over very steep slopes, under high rainfall intensity; and for ANNs over steep and very steep slopes, under very high rainfall intensities.     

分季节的太湖悬浮物遥感估测模型研究

根据1996-2002年无锡太湖监测站的水质资料分析,太湖悬浮物具有季节性特征,因而分季节的悬浮物估测模型比单一的模型可能更加适合用来估测太湖全年的悬浮物浓度.在分析太湖水体光谱特征的基础上,根据太湖悬浮物的季节性分布特征,使用春夏秋冬四季的Landsat TM/ETM图像和准同步的水质采样数据,建立了太湖分季节的悬浮物估算模型.结果表明:估测因子(B2 B3)/(B2/B3)在春、秋、冬三季都能很好地估测出悬浮物的浓度(R2＞0.52).夏季由于叶绿素的干扰性较大,悬浮物的估测效果不理想.冬季的估测效果最好(R2=0.81),模型为lnSS=14.656×(B2 B3)/(B2/B3) 1.661,其中,ln SS表示悬浮物取自然对数后的值,B2、B3为TM/ETM图像经过6S大气校正、3×3低通滤波后第2、3波段的反射率值.     

Application of newly developed ensemble machine learning models for daily suspended sediment load prediction and related uncertainty analysis

ABSTRACT

Ensemble machine learning models have been widely used in hydro-systems modeling as robust prediction tools that combine multiple decision trees. In this study, three newly developed ensemble machine learning models, namely gradient boost regression (GBR), AdaBoost regression (ABR) and random forest regression (RFR) are proposed for prediction of suspended sediment load (SSL), and their prediction performance and related uncertainty are assessed. The SSL of the Mississippi River, which is one of the major world rivers and is significantly affected by sedimentation, is predicted based on daily values of river discharge (Q) and suspended sediment concentration (SSC). Based on performance metrics and visualization, the RFR model shows a slight lead in prediction performance. The uncertainty analysis also indicates that the input variable combination has more impact on the obtained predictions than the model structure selection.     

Estimation of suspended sediment concentration and yield using linear models,random forests and quantile regression forests

For sediment yield estimation, intermittent measurements of suspended sediment concentration (SSC) have to be interpolated to derive a continuous sedigraph. Traditionally, sediment rating curves (SRCs) based on univariate linear regression of discharge and SSC (or the logarithms thereof) are used but alternative approaches (e.g. fuzzy logic, artificial neural networks, etc.) exist. This paper presents a comparison of the applicability of traditional SRCs, generalized linear models (GLMs) and non‐parametric regression using Random Forests (RF) and Quantile Regression Forests (QRF) applied to a dataset of SSC obtained for four subcatchments (0·08, 41, 145 and 445 km²) in the Central Spanish Pyrenees. The observed SSCs are highly variable and range over six orders of magnitude. For these data, traditional SRCs performed inadequately due to the over‐simplification of relating SSC solely to discharge. Instead, the multitude of acting processes required more flexibility to model these nonlinear relationships. Thus, alternative advanced machine learning techniques that have been successfully applied in other disciplines were tested. GLMs provide the option of including other relevant process variables (e.g. rainfall intensities and temporal information) but require the selection of the most appropriate predictors. For the given datasets, the investigated variable selection methods produced inconsistent results. All proposed GLMs showed an inferior performance, whereas RF and QRF proved to be very robust and performed favourably for reproducing sediment dynamics. QRF additionally provides estimates on the accuracy of the predictions and thus allows the assessment of uncertainties in the estimated sediment yield that is not commonly found in other methods. The capabilities of RF and QRF concerning the interpretation of predictor effects are also outlined. Copyright © 2008 John Wiley & Sons, Ltd.     

An evaluation of sediment rating curves for estimating suspended sediment concentrations for subsequent flux calculations

In the absence of actual suspended sediment concentration (SSC) measurements, hydrologists have used sediment rating (sediment transport) curves to estimate (predict) SSCs for subsequent flux calculations. Various evaluations of the sediment rating‐curve method were made using data from long‐term, daily sediment‐measuring sites within large (>1 000 000 km²), medium (<1 000 000 to >1000 km²), and small (<1000 km²) river basins in the USA and Europe relative to the estimation of suspended sediment fluxes. The evaluations address such issues as the accuracy of flux estimations for various levels of temporal resolution as well as the impact of sampling frequency on the magnitude of flux estimation errors. The sediment rating‐curve method tends to underpredict high, and overpredict low SSCs. As such, the range of errors associated with concomitant flux estimates for relatively short time‐frames (e.g. daily, weekly) are likely to be substantially larger than those associated with longer time‐frames (e.g. quarterly, annually) because the over‐ and underpredictions do not have sufficient time to balance each other. Hence, when error limits must be kept under ±20%, temporal resolution probably should be limited to quarterly or greater. The evaluations indicate that over periods of 20 or more years, errors of <1% can be achieved using a single sediment rating curve based on data spanning the entire period. However, somewhat better estimates for the entire period, and markedly better annual estimates within the period, can be obtained if individual annual sediment rating curves are used instead. Relatively accurate (errors <±20%) annual suspended sediment fluxes can be obtained from hydrologically based monthly measurements/samples. For 5‐year periods or longer, similar results can be obtained from measurements/samples collected once every 2 months. In either case, hydrologically based sampling, as opposed to calendar‐based sampling is likely to limit the magnitude of flux estimation errors. Published in 2003 John Wiley & Sons, Ltd.     

Estimating suspended sediment concentrations in areas with limited hydrological data using a mixed‐effects model

Sediment rating curves are commonly used to estimate the suspended sediment load in rivers and streams under the assumption of a constant relation between discharge (Q) and suspended sediment concentrations (SSC) over time. However, temporal variation in the sediment supply of a watershed results in shifts in this relation by increasing variability and by introducing nonlinearities in the form of hysteresis or a path‐dependent relation. In this study, we used a mixed‐effects linear model to estimate an average SSC–Q relation for different periods of time within the hydrologic cycle while accounting for seasonality and hysteresis. We tested the performance of the mixed‐effects model against the standard rating curve, represented by a generalized least squares regression, by comparing observed and predicted sediment loads for a test case on the Chilliwack River, British Columbia, Canada. In our analyses, the mixed‐effects model reflected more accurate patterns of interpolated SSC from Q data than the rating curve, especially for the low‐flow summer months when the SSC–Q relation is less clear. Akaike information criterion scores were lower for the mixed‐effects model than for the standard model, and the mixed‐effects model explained nearly twice as much variance as the standard model (52% vs 27%). The improved performance was achieved by accounting for variability in the SSC–Q relation within each month and across years for the same month using fixed and random effects, respectively, a characteristic disregarded in the sediment rating curve. Copyright © 2012 John Wiley & Sons, Ltd.     

Modelling suspended sediment concentration using artificial neural networks for Gangotri glacier

The dynamics of suspended sediment involves inherent non‐linearity and complexity because of existence of both spatial variability of the basin characteristics and temporal climatic patterns. This complexity, therefore, leads to inaccurate prediction by the conventional sediment rating curve (SRC) and other empirical methods. Over past few decades, artificial neural networks (ANNs) have emerged as one of the advanced modelling techniques capable of addressing inherent non‐linearity in the hydrological processes. In the present study, feed‐forward back propagation (FFBP) algorithm of ANNs is used to model stage–discharge–suspended sediment relationship for ablation season (May–September) for melt runoff released from Gangotri glacier, one of the largest glaciers in Himalaya. The simulations have been carried out on primary data of suspended sediment concentration (SSC) discharge and stage for ablation season of 11‐year period (1999–2009). Combinations of different input vectors (viz. stage, discharge and SSC) for present and previous days are considered for development of the ANN models and examining the effects of input vectors. Further, based on model performance indices for training and testing phase, a suitable modelling approach with appropriate model input structure is suggested. The conventional SRC method is also used for modelling discharge–sediment relationship and performance of developed models is evaluated by statistical indices, namely; root mean square error (RMSE), mean absolute error (MAE) and coefficient of determination (R²). Statistically, the performance of ANN‐based models is found to be superior as compared to SRC method in terms of the selected performance indices in simulating the daily SSC. The study reveals suitability of ANN approach for simulation and estimation of daily SSC in glacier melt runoff and, therefore, opens new avenues of research for application of hybrid soft computing models in glacier hydrology. Copyright © 2015 John Wiley & Sons, Ltd.     

Prediction of hourly actual evapotranspiration using neural networks,genetic programming,and statistical models

The complexity of the evapotranspiration process and its variability in time and space have imposed some limitations on previously developed evapotranspiration models. In this study, two data‐driven models: genetic programming (GP) and artificial neural networks (ANNs), and statistical regression models were developed and compared for estimating the hourly eddy covariance (EC)‐measured actual evapotranspiration (AET) using meteorological variables. The utility of the investigated data‐driven models was also compared with that of HYDRUS‐1D model, which makes use of conventional Penman–Monteith (PM) model for the prediction of AET. The latent heat (LE), which is measured using the EC method, is modelled as a function of five climatic variables: net radiation, ground temperature, air temperature, relative humidity, and wind speed in a reconstructed landscape located in Northern Alberta, Canada. Several ANN models were evaluated using two training algorithms of Levenberg–Marquardt and Bayesian regularization. The GP technique was used to generate mathematical equations correlating AET to the five climatic variables. Furthermore, the climatic variables, as well as their two‐factor interactions, were statistically analysed to obtain a regression equation and to indicate the climatic factors having significant effect on the evapotranspiration process. HYDRUS‐1D model as an available physically based model was examined for estimating AET using climatic variables, leaf area index (LAI), and soil moisture information. The results indicated that all three proposed data‐driven models were able to approximate the AET reasonably well; however, GP and regression models had better generalization ability than the ANN model. The results of HYDRUS‐1D model exhibited that a physically based model, such as HYDRUS‐1D, might be comparable or even inferior to the data‐driven models in terms of the overall prediction accuracy. Based on the developed GP and regression models, net radiation and ground temperature had larger contribution to the AET process than other variables. Copyright © 2010 John Wiley & Sons, Ltd.     

Fluvial suspended sediment transport from cold and warm-based glaciers in Svalbard

An analysis of temporal variability in proglacial suspended sediment concentration is undertaken using time series data collected from three Svalbard basins which include one largely cold-based glacier (Austre Brøggerbreen), one largely warm-based glacier (Finsterwalderbreen) and one intermediate polythermal glacier (Erdmannbreen). The temporal variability in proglacial suspended sediment concentration is analysed using multiple regression techniques in which discharge is supplemented by other predictors acting as surrogates for variability in sediment supply at diurnal, medium-term and seasonal timescales. These multiple regression models improve upon the statistical explanation of suspended sediment concentration produced by simple sediment rating curves but need to account for additional stochastic elements within the time series before they may be considered successful. An interpretation of the physical processes which are responsible for the regression model characteristics is offered as a basis for comparing the different arctic glaciofluvial suspended sediment transport systems with that of their better known temperate glaciofluvial counterparts. It is inferred that the largely warm-based glacier is dominated by sediment supply from subglacial reservoirs which evolve in a similar manner to temperate glaciers and which cause a pronounced seasonal exhaustion of suspended sediment supply. The largely cold-based glacier, however, is dominated by sediment supply from marginal sources which generate a responsive system at short time scales but no significant seasonal pattern. The intermediate polythermal glacier basin, which was anticipated to be similar to the warm-based glacier, instead shows a highly significant seasonal increase in suspended sediment supply from an unusual subglacial reservoir emerging under pressure in the glacier foreland. The temperate model of glaciofluvial suspended sediment transport is therefore found to be of limited use in an arctic context. Copyright © 1999 John Wiley & Sons, Ltd.     

Probabilistic and ensemble simulation approaches for input uncertainty quantification of artificial neural network hydrological models

Artificial neural network (ANN) has been demonstrated to be a promising modelling tool for the improved prediction/forecasting of hydrological variables. However, the quantification of uncertainty in ANN is a major issue, as high uncertainty would hinder the reliable application of these models. While several sources have been ascribed, the quantification of input uncertainty in ANN has received little attention. The reason is that each measured input quantity is likely to vary uniquely, which prevents quantification of a reliable prediction uncertainty. In this paper, an optimization method, which integrates probabilistic and ensemble simulation approaches, is proposed for the quantification of input uncertainty of ANN models. The proposed approach is demonstrated through rainfall-runoff modelling for the Leaf River watershed, USA. The results suggest that ignoring explicit quantification of input uncertainty leads to under/over estimation of model prediction uncertainty. It also facilitates identification of appropriate model parameters for better characterizing the hydrological processes.     

A discrete Bayesian network to investigate suspended sediment concentrations in an Alpine proglacial zone

In this study, a Bayesian Network (BN) is used to model the suspended sediment concentrations (SSC) in the catchments of the glaciers Noir and Blanc in the Ecrins National Park, France, and at the distal end of the proglacial zone into which both torrents drain. Relationships between air temperature, glacier discharge and SSC are represented as random variables; thereby taking the natural next step from proposed modified rating curve methods which increasingly approximate random variable approaches. Hydrological relationships are propagated through the network via conditional probability distributions computed from 980 field records obtained at three monitoring sites during July 2005. Rainfall affected data are removed from the modelling process. A two‐sample Kolmogorov–Smirnov goodness‐of‐fit (two‐sample KS) test (n = 5) shows good agreement between the probability distributions of SSC predicted by the BN, and those recorded in the field at the outflow of the proglacial zone over an air temperature range of 5–25 °C. The BN performs poorly for air temperatures between 25 and 30 °C and this is attributed to limited field records covering this temperature range. Discussion of the significant limitations surrounding the widespread application of BNs in hydrological modelling are offered with a focus on data volume and temporal limitations. Copyright © 2008 John Wiley & Sons, Ltd.     

Artificial neural network simulation for prediction of suspended sediment concentration in the River Ramganga,Ganges Basin,India

The relation between the water discharge (Q) and suspended sediment concentration (SSC) of the River Ramganga at Bareilly, Uttar Pradesh, in the Himalayas, has been modeled using Artificial Neural Networks (ANNs). The current study validates the practical capability and usefulness of this tool for simulating complex nonlinear, real world, river system processes in the Himalayan scenario. The modeling approach is based on the time series data collected from January to December (2008-2010) for Q and SSC. Three ANNs (T1-T3) with different network configurations have been developed and trained using the Levenberg Marquardt Back Propagation Algorithm in the Matlab routines. Networks were optimized using the enumeration technique, and, finally, the best network is used to predict the SSC values for the year 2011. The values thus obtained through the ANN model are compared with the observed values of SSC. The coefficient of determination (R2), for the optimal network was found to be 0.99. The study not only provides insight into ANN modeling in the Himalayan river scenario, but it also focuses on the importance of understanding a river basin and the factors that affect the SSC, before attempting to model it. Despite the temporal variations in the study area, it is possible to model and successfully predict the SSC values with very simplistic ANN models.     

Estimation of suspended sediment concentration in the Saint John River using rating curves and a machine learning approach

ABSTRACT

Sedimentation in navigable waterways and harbours is of concern for many water and port managers. One potential source of variability in sedimentation is the annual sediment load of the river that empties in the harbour. The main objective of this study was to use some of the regularly monitored hydro-meteorological variables to compare estimates of hourly suspended sediment concentration in the Saint John River using a sediment rating curve and a model tree (M5?) with different combinations of predictors. Estimated suspended sediment concentrations were multiplied by measured flows to estimate suspended sediment loads. Best results were obtained using M5? with four predictors, returning an R² of 0.72 on calibration data and an R² of 0.46 on validation data. Total load was underestimated by 1.41% for the calibration period and overestimated by 2.38% for the validation period. Overall, the model tree approach is suggested for its relative ease of implementation and constant performance.

EDITOR M.C. Acreman; ASSOCIATE EDITOR B. Touaibia     

River suspended sediment load prediction based on river discharge information: application of newly developed data mining models

ABSTRACT

Suspended sediment load (SSL) is one of the essential hydrological processes that affects river engineering sustainability. Sediment has a major influence on the operation of dams and reservoir capacity. This investigation is aimed at exploring a new version of machine learning models (i.e. data mining), including M5P, attribute selected classifier (AS M5P), M5Rule (M5R), and K Star (KS) models for SSL prediction at the Trenton meteorological station on the Delaware River, USA. Different input scenarios were examined based on the river flow discharge and sediment load database. The performance of the applied data mining models was evaluated using various statistical metrics and graphical presentation. Among the applied data mining models, the M5P model gave a superior prediction result. The current and one-day lead time river flow and sediment load were the influential predictors for one-day-ahead SSL prediction. Overall, the applied data mining models achieved excellent predictions of the SSL process.     

Characterization of suspended sediment in Meltwater from Glaciers of Garhwal Himalaya

Glacierised basins are significant sources of sediments generated by glacial retreat. Estimation of suspended sediment transfer from glacierised basins is very important in reservoir planning for hydropower projects in Himalaya. The present study indicates that storage and release of sediment in proglacial streams may categorise the pattern of suspended sediment transfer from these basins. Assessment of suspended sediment concentration (SSC), suspended sediment load (SSL) and yield has been undertaken for Dunagiri Glacier basin located in Garhwal Himalaya (30^o33'20”N, 79^o53'36”E), and its results are compared with the Gangotri and Dokriani glaciers sharing close proximity. Out of the total drainage basin area, about 14.3 % of the area is glacierised. Data were collected for five ablation seasons (1984–1989, barring 1986). The mean daily SSCs for July, August and September were 333.9, 286.0 and 147.15 mg/l, respectively, indicating highest concentration of mean daily suspended sediment in July followed by August. SSL trends were estimated to be 93.0, 57.0 and 21.3 tonnes. About 59% of the total SSL of the melt period was transported during the months of August and September. Sediment yield for the study basin was computed to be 296.3 t km^?2 yr ^?1. It is observed that the cumulative proportion of SSC precedes the discharge throughout the melt season except in the year 1987. Release of SSL in terms of total load is less in the early part of melt season than in the later stage as compared to that of discharge. Diurnal variations in SSC reach their maximum at 2400 h, and therefore, SSC was found to be high during night (2000–0400 h). There was a good relationship between SSC and SSL with discharge for the ablation seasons (1988 and 1989). Mean monthly SSC and mean monthly SSL provide a good exponentional relationship with mean monthly temperature. Copyright © 2012 John Wiley & Sons, Ltd.     

Turbidity and suspended sediment dynamics in small catchments in the Nepal Middle Hills

The Himalayan environment has, until recently, been perceived to be in a critical state of environmental decline, resulting from rapid population growth and associated land‐use change. Recent research, however, has emphasized the difficulty of developing an objective appraisal of the state of the environment in a region where empirical data are scarce and unstructured and where an understanding of the spatial and temporal dynamics of natural environmental processes remains highly uncertain. This paper presents results from an intensive three‐year project designed to help address the regional empirical deficit, establish detailed baseline environmental data and to gain an insight into storm period and seasonal suspended sediment dynamics. The instrumentation, calibration and analysis of high‐frequency infrared turbidimetric records from a number of small subcatchments in the Nepal Middle Hills are reported. Storm period and seasonal variation in turbidity and suspended sediment are examined and hysteresis patterns explored and explained. A variety of methods to estimate seasonal suspended sediment yield in a mixed land‐use catchment are examined, and found to vary by up to a factor of five. Despite the inherent uncertainty, all estimates of catchment sediment yield are found to be high with respect to erosion plot studies from the local area, and this suggests the importance of riparian and channel erosion as major sediment sources, a finding consistent with other regional studies. Copyright © 2000 John Wiley & Sons, Ltd.     

Interannual variation of suspended sediment load and sediment yield in an alpine catchment

Abstract

The results are described of 16 years operation of a measuring station for the automatic recording of water discharge, bed load and suspended sediment transport in the Rio Cordon catchment, a small alpine basin (5 km²) located in northeastern Italy. Hillslope erosion processes were investigated by surveying individual sediment sources repeatedly. Annual and seasonal variations of suspended sediment load during the period 1986–2001 are analysed along with their contribution to the total sediment yield. The results show that suspended load accounted for 76% of total load and that most of the suspended sediment transport occurred during two flood events: an extreme summer flash flood in September 1994 (27% of the 16-years total suspended load) and a snowmelt-induced event in May 2001 accompanied by a mud flow which fed the stream with sediments. The role of active sediment source areas is discussed in relation to the changes in flood peak—suspended load trends which became apparent after both the 1994 and the 2001 events.     

Analysis of suspended sediment yields after low impact forest harvesting

Disturbances to forest catchments have profound effects on the environment of headwater streams and have an impact on suspended sediment (SS) management. Forest harvesting is a dominant factor in increasing SS yields. Road construction, skidder activity and ploughing associated with harvesting cause serious soil disturbance that results in SS increases. However, few studies have shown whether harvesting itself increases SS yields. This study examined how harvesting influenced SS yields in a steep forested area. During harvesting, soil surface disturbance was prevented as much as possible by using skyline logging treatments and piling branches and leaves at selected locations in the watershed. Using these methods, the representative SS rating curve did not change significantly after harvesting. The results also show that the characteristics of SS transport were related to the SS source area, and reveal that the riparian zone/stream bank was a dominant SS source area at the study site. Annual SS yields did not increase despite increasing annual water yields after harvesting. The limited water capacity of the soil at the study site likely led to only slight differences in pre‐ and post‐harvest water discharge from heavy rainfall events. Most SS was transported during heavy rainfall events, and increases in SS yields were not detected after harvesting. We concluded that it is possible to prevent post‐harvest SS increases by performing careful, low‐impact harvesting procedures. Copyright © 2007 John Wiley & Sons, Ltd.