Application of Soft Computing Models with Input Vectors of Snow Cover Area in Addition to Hydro-Climatic Data to Predict the Sediment Loads

Abstract

The accurate estimate of sediment load is important for management of the river ecosystem, designing of water infrastructures, and planning of reservoir operations. The direct measurement of sediment is the most credible method to estimate the sediments. However, this requires a lot of time and resources. Because of these two constraints, most often, it is not possible to continuously measure the daily sediments for most of the gauging sites. Nowadays, data-based sediment prediction models are famous for bridging the data gaps in the estimation of sediment loads. In data-driven sediment predictions models, the selection of input vectors is critical in determining the best structure of models for the accurate estimation of sediment yields. In this study, time series inputs of snow cover area, basin effective rainfall, mean basin average temperature, and mean basin evapotranspiration in addition to the flows were assessed for the prediction of sediment loads. The input vectors were assessed with artificial neural network (ANN), adaptive neuro-fuzzy logic inference system with grid partition (ANFIS-GP), adaptive neuro-fuzzy logic inference system with subtractive clustering (ANFIS-SC), adaptive neuro-fuzzy logic inference system with fuzzy c-means clustering (ANFIS-FCM), multiple adaptive regression splines (MARS), and sediment rating curve (SRC) models for the Gilgit River, the tributary of the Indus River in Pakistan. The comparison of different input vectors showed improvements in the prediction of sediments by using the snow cover area in addition to flows, effective rainfall, temperature, and evapotranspiration. Overall, the ANN model performed better than all other models. However, as regards sediment load peak time series, the sediment loads predicted using the ANN, ANFIS-FCM, and MARS models were found to be closer to the measured sediment loads. The ANFIS-FCM performed better in the estimation of peak sediment yields with a relative accuracy of 81.31% in comparison to the ANN and MARS models with 80.17% and 80.16% of relative accuracies, respectively. The developed multiple linear regression equation of all models show an R2 value of 0.85 and 0.74 during the training and testing period, respectively.

Superiority of Hybrid Soft Computing Models in Daily Suspended Sediment Estimation in Highly Dynamic Rivers

“…It can perform well with a large amount of noisy data resulting from the dynamic and non-linear system where the system's fundamental physical relationships are unknown [14][15][16][17]. AI techniques can solve the complex problems of different hydrologic processes [18]. A black box model such as an Artificial Neural Network (ANN) does not need watershed physical characteristics to transform inputs into an output and to recognize any hydrologic process [8,19].…”

Simulating the hydrological regime of the snow fed and glaciarised Gilgit Basin in the Upper Indus using global precipitation products and a data parsimonious precipitation-runoff model

“…The simulations showed JRA-55 has a slightly higher SCA and more intense snow events due to overestimation of precipitation compared with ERA5-Land. These simulations are validated with MODIS satellite derived daily SCA from published data (Hussan et al, 2020) for the Gilgit Basin for 2006-2010 (Fig. 5a).…”

“…LandSat-8 data are the most recent data by the Landsat Data Continuity Mission (LDCM) with 30 m spatial and 16 days temporal resolution. The MODIS snow and ice data were accessed from published data (Hussan et al, 2020) for Gilgit Basin and used for validating the SCA simulations by the precipitation-runoff model. The Distance Distribution Dynamics (DDD) model is a conceptual, semi-distributed, catchment based precipitation-runoff model, scripted in R and Julia programming that can simulate runoff at daily or even smaller time steps (Skaugen and Onof, 2014).…”

A brief review of recent global trends in suspended sediment estimation studies

“…To date, a variety of MLM approaches have been applied to estimate SS. Some examples are Gaussian Processes (GP), Support Vector Machines (SVM), Evolutionary Support Vector Machines (ESVM), ElasticNet Linear Regression (LR), Multi-Layer Perceptron (MLP), Extreme Gradient Boosting, Long Short-Term Memory (LSTM), Adaptive Neuro-Fuzzy Inference System (ANFIS) and Genetic Algorithm (GA) (Amamra et al, 2018;Adnan et al, 2019;Ul Hussan et al, 2020;AlDahoul et al, 2021;Asadi et al, 2021;Kaveh et al, 2021;Nourani et al, 2021;Lund et al, 2022). Although all are capable of predicting sediment with reasonable accuracy, there are instances where they outclass each other at different locations.…”

“…Each method has its advantages and disadvantages. Although, MLMs have grown in popularity (Ul Hussan et al, 2020), their transferability to different regions has become a challenge. In contrast, PNMs are often considered applicable in a wide range of locations.…”