Evaluating the Performance of Different Artificial Intelligence Techniques for Forecasting: Rainfall and Runoff Prospective

Abstract

The forecasting plays key role for the water resources planning. Most suitable technique is Artificial intelligence techniques (AITs) for different parameters of weather forecasting and generated runoff. The study compared AITs (RBF-SVM and M5 model tree) to understand the rainfall runoff process in Jhelum River Basin, Pakistan. The rainfall and runoff of Jhelum river used from 1981 to 2012. The Different rainfall and runoff dataset combinations were used to train and test AITs. The data record for the period 1981–2001 used for training and then testing. After training and testing, modeled runoff and observed data was evaluated using R2, NRMSE, COE and MSE. During the training, the dataset C2 and C3 were found to be 0.71 for both datasets using M5 model. Similar results were found for dataset of C3 using RBF-SVM. Over all, C3 and C7 were performed best among all the dataset. The M5 model tree was performed better than other applied techniques. GEP has also exhibited good results to understand rainfall runoff process. The RBF-SVM performed less accurate as compare to other applied techniques. Flow duration curve (FDCs) were used to compare the modeled and observed dataset of Jhelum River basin. For High flow and medium high flows, GEP exhibited well. M5 model tree displayed the better results for medium low and low percentile flows. RBF-SVM exhibited better for low percentile flows. GEP were found the accurate and highly efficient DDM among the AITs applied techniques. This study will help understand the complex rainfall runoff process, which is stochastic process. Weather forecasting play key role in water resources management and planning.

Runoff Estimation Using Advanced Soft Computing Techniques: A Case Study of Mangla Watershed Pakistan

“…(c) depicts the overall RMSE and MAE values. Figure 8a,b demonstrates a significant reduction in error values from input combinations up to the input combination, which is consistent with the R 2 and NSE values (4). According to the findings, it is possible to assert that the precipitation P(t), P(t-1), P(t-2), and P(t-3) contain complete data about the watershed's hydrological signature and that there is no vital information hidden in the lag precipitation data.…”

“…So, to address these drought and flood issues, the estimation of runoff generated from the rainfall event is vital. In the transition of precipitation into a runoff, precipitation finally transforms into a runoff after fulfilling various losses such as interception, depression storage, infiltration, and evaporation [3,4]. Runoff is a complex and nonlinear outcome of rainfall and watershed properties.…”

“…An artificial neural network was used successfully to model daily and monthly runoff computations in several hydrological studies [3,9,[19][20][21][22][23][24][25][26][27]. SVM is the statistical learning technique used for water table depth estimation and streamflow forecasting [4,[28][29][30][31][32][33]. Gene expression programming (GEP) and ANFIS examine hydrological concerns such as flood and river flow forecasting [34,35].…”

Hydrological model parameter regionalization: Runoff estimation using machine learning techniques in the Tha Chin River Basin, Thailand

“…Accurately predicting streamflow requires advanced modeling approaches due to the inherent complexity of hydrological systems, further exacerbated by anthropogenic effects and climate unpredictability [ 5 ]. Over recent decades, hydrological models have emerged as the predominant method for predicting runoff [ 6 , 7 ]. Nevertheless, these models impose calibration of free parameters using observed discharge data before forecasting runoff hydrographs, data requirements unmet in numerous catchment areas [ 8 ].…”

Modeling of Monthly Rainfall–Runoff Using Various Machine Learning Techniques in Wadi Ouahrane Basin, Algeria

“…Waqas et al [19] developed radial basis function (RBF)-SVM and M5 models to model the rainfall-runoff process in the Jhelum River Basin, Pakistan. The models were trained and tested using various combinations of datasets.…”