Comparison of artificial neural network and combined models in estimating spatial distribution of snow depth and snow water equivalent in Samsami basin of Iran

Snow water equivalent (SWE) is a key parameter in hydrological cycle, and information on regional SWE is required for various hydrological and meteorological applications, as well as for hydropower production and flood forecasting. This study compares the snow depth and SWE estimated by multivariate linear regression (MLR), discriminant function analysis, ordinary kriging, ordinary kriging-multivariate linear regression, ordinary kriging-discriminant function analysis, artificial neural network (ANN) and neural network-genetic algorithm (NNGA) models. The analysis was performed in the 5.2 km² area of Samsami basin, located in the southwest of Iran. Statistical criteria were used to measure the models’ performances. The results indicated that NNGA, ANN and MLR methods were able to predict SWE at the desirable level of accuracy. However, the NNGA model with the highest coefficient of determination (R ² = 0.70, P value < 0.05) and minimum root mean square error (RMSE = 0.202 cm) provided the best results among the other models. The lower SWE values were registered in the east of study area and higher SWE values appeared in the west of study area where altitude was higher.     

Precipitation forecasting by using wavelet-support vector machine conjunction model

A new wavelet-support vector machine conjunction model for daily precipitation forecast is proposed in this study. The conjunction method combining two methods, discrete wavelet transform and support vector machine, is compared with the single support vector machine for one-day-ahead precipitation forecasting. Daily precipitation data from Izmir and Afyon stations in Turkey are used in the study. The root mean square errors (RMSE), mean absolute errors (MAE), and correlation coefficient (R) statistics are used for the comparing criteria. The comparison results indicate that the conjunction method could increase the forecast accuracy and perform better than the single support vector machine. For the Izmir and Afyon stations, it is found that the conjunction models with RMSE=46.5 mm, MAE=13.6 mm, R=0.782 and RMSE=21.4 mm, MAE=9.0 mm, R=0.815 in test period is superior in forecasting daily precipitations than the best accurate support vector regression models with RMSE=71.6 mm, MAE=19.6 mm, R=0.276 and RMSE=38.7 mm, MAE=14.2 mm, R=0.103, respectively. The ANN method was also employed for the same data set and found that there is a slight difference between ANN and SVR methods.     

Applications of soft computing techniques for prediction of energy dissipation on stepped spillways

In this study, numbers type of soft computing including artificial neural network (ANN), support vector machine (SVM), multivariate adaptive regression splines (MARS), and group method of data handling (GMDH) were applied to model and predict energy dissipation of flow over stepped spillways. Results of ANN indicated that this model including hyperbolic tangent sigmoid as transfer function obtained coefficient of determination (R ² = 0.917) and root-mean-square error (RMSE = 6.927) in testing stage. Results of development of SVM showed that developed model consists of radial basis function as kernel function achieved R ² = 0.98 and RMSE = 2.61 in validation stage. Developed MARS model with R ² = 0.99 and RMSE = 0.65 has suitable performance for predicating the energy dissipation. Results of developed GMDH model show with R ² = 0.95 and RMSE = 5.4 has suitable performance for modeling energy dispersion. Reviewing of results of prepared models showed that all of them have suitable performance to predict the energy dissipation. However, MARS and SVM are more accurate than the others. Attention to structures of GMDH and MARS models declared that Froude number, drop number, and ratio of critical depth to height of step are the most important parameters for modeling energy dissipation. The best radial basis function was found that as best kernel function in developing the SVM.

     

Comparison of modelling ANN and ELM to estimate solar radiation over Turkey using NOAA satellite data

In this study, solar radiation (SR) is estimated at 61 locations with varying climatic conditions using the artificial neural network (ANN) and extreme learning machine (ELM). While the ANN and ELM methods are trained with data for the years 2002 and 2003, the accuracy of these methods was tested with data for 2004. The values for month, altitude, latitude, longitude, and land-surface temperature (LST) obtained from the data of the National Oceanic and Atmospheric Administration Advanced Very High Resolution Radiometer (NOAA-AVHRR) satellite are chosen as input in developing the ANN and ELM models. SR is found to be the output in modelling of the methods. Results are then compared with meteorological values by statistical methods. Using ANN, the determination coefficient (R²), mean bias error (MBE), root mean square error (RMSE), and Willmott’s index (WI) values were calculated as 0.943, ?0.148 MJ m^?2, 1.604 MJ m^?2, and 0.996, respectively. While R² was 0.961, MBE, RMSE, and WI were found to be in the order 0.045 MJ m^?2, 0.672 MJ m^?2, and 0.997 by ELM. As can be understood from the statistics, ELM is clearly more successful than ANN in SR estimation.     

Predicting discharge coefficient of triangular labyrinth weir using extreme learning machine,artificial neural network and genetic programming

Weirs are a type of hydraulic structure used to direct and transfer water flows in the canals and overflows in the dams. The important index in computing flow discharge over the weir is discharge coefficient (C _d). The aim of this study is accurate determination of the C _d in triangular labyrinth side weirs by applying three intelligence models [i.e., artificial neural network (ANN), genetic programming (GP) and extreme learning machine (ELM)]. The calculated discharge coefficients were then compared with some experimental results. In order to examine the accuracy of C _d predictions by ANN, GP and ELM methods, five statistical indices including coefficient of determination (R ²), root-mean-square error (RMSE), mean absolute percentage error (MAPE), SI and δ have been used. Results showed that R ² values in the ELM, ANN and GP methods were 0.993, 0.886 and 0.884, respectively, at training stage and 0.971, 0.965 and 0.963, respectively, at test stage. The ELM method, having MAPE, RMSE, SI and δ values of 0.81, 0.0059, 0.0082 and 0.81, respectively, at the training stage and 0.89, 0.0063, 0.0089 and 0.88, respectively, at the test stage, was superior to ANN and GP methods. The ANN model ranked next to the ELM model.

     

ANN based simulation and experimental verification of analytical four- and five-parameters models of PV modules

In this article, artificial neural network (ANN) is adopted to predict photovoltaic (PV) panel behaviors under realistic weather conditions. ANN results are compared with analytical four and five parameter models of PV module. The inputs of the models are the daily total irradiation, air temperature and module voltage, while the outputs are the current and power generated by the panel. Analytical models of PV modules, based on the manufacturer datasheet values, are simulated through Matlab/Simulink environment. Multilayer perceptron is used to predict the operating current and power of the PV module. The best network configuration to predict panel current had a 3–7–4–1 topology. So, this two hidden layer topology was selected as the best model for predicting panel current with similar conditions. Results obtained from the PV module simulation and the optimal ANN model has been validated experimentally. Results showed that ANN model provide a better prediction of the current and power of the PV module than the analytical models. The coefficient of determination (R²), mean square error (MSE) and the mean absolute percentage error (MAPE) values for the optimal ANN model were 0.971, 0.002 and 0.107, respectively. A comparative study among ANN and analytical models was also carried out. Among the analytical models, the five-parameter model, with MAPE = 0.112, MSE = 0.0026 and R² = 0.919, gave better prediction than the four-parameter model (with MAPE = 0.152, MSE = 0.0052 and R² = 0.905). Overall, the 3–7–4–1 ANN model outperformed four-parameter model, and was marginally better than the five-parameter model.     

High performance concrete compressive strength forecasting using ensemble models based on discrete wavelet transform

This paper investigates the use of wavelet ensemble models for high performance concrete (HPC) compressive strength forecasting. More specifically, we incorporate bagging and gradient boosting methods in building artificial neural networks (ANN) ensembles (bagged artificial neural networks (BANN) and gradient boosted artificial neural networks (GBANN)), first. Coefficient of determination (R²), mean absolute error (MAE) and the root mean squared error (RMSE) statics are used for performance evaluation of proposed predictive models. Empirical results show that ensemble models (R²_BANN=0.9278, R²_GBANN=0.9270) are superior to a conventional ANN model (R²_ANN=0.9088). Then, we use the coupling of discrete wavelet transform (DWT) and ANN ensembles for enhancing the prediction accuracy. The study concludes that DWT is an effective tool for increasing the accuracy of the ANN ensembles (R²_WBANN=0.9397, R²_WGBANN=0.9528).     

Feasibility of PSO–ANFIS model to estimate rock fragmentation produced by mine blasting

Desired rock fragmentation is the main goal of the blasting operation in surface mines, civil and tunneling works. Therefore, precise prediction of rock fragmentation is very important to achieve an economically successful outcome. The primary objective of this article is to propose a new model for forecasting the rock fragmentation using adaptive neuro-fuzzy inference system (ANFIS) in combination with particle swarm optimization (PSO). The proposed PSO–ANFIS model has been compared with support vector machines (SVM), ANFIS and nonlinear multiple regression (MR) models. To construct the predictive models, 72 blasting events were investigated, and the values of rock fragmentation as well as five effective parameters on rock fragmentation, i.e., specific charge, stemming, spacing, burden and maximum charge used per delay were measured. Based on several statistical functions [e.g., coefficient of correlation (R ²) and root-mean-square error (RMSE)], it was found that the PSO–ANFIS (with R ² = 0.89 and RMSE = 1.31) performs better than the SVM (with R ² = 0.83 and RMSE = 1.66), ANFIS (with R ² = 0.81 and RMSE = 1.78) and nonlinear MR (with R ² = 0.57 and RMSE = 3.93) models. Finally, the sensitivity analysis shows that the burden and maximum charge used per delay have the least and the most effects on the rock fragmentation, respectively.

     

Time-series modeling of fishery landings using ARIMA models and Fuzzy Expected Intervals software

Forecasting, using historic time-series data, has become an important tool for fisheries management. ARIMA modeling, Modeling for Optimal Forecasting techniques and Decision Support Systems based on fuzzy mathematics may be used to predict the general trend of a given fish landings time-series with increased reliability and accuracy. The present paper applies these three modeling methods to forecast anchovy fish catches landed in a given port (Thessaloniki, Greece) during 1979–2000 and hake and bonito total fish catches during 1982–2000. The paper attempts to assess the model's accuracy by comparing model results to the actual monthly fish catches of the year 2000. According to the measures of forecasting accuracy established, the best forecasting performance for anchovy was shown by the DSS model (MAPE = 28.06%, RMSE = 76.56, U-statistic = 0.67 and R² = 0.69). The optimal forecasting technique of genetic modeling improved significantly the forecasting values obtained by the selected ARIMA model. Similarly, the DSS model showed a noteworthy forecasting efficiency for the prediction of hake landings, during the year 2000 (MAPE = 2.88%, RMSE = 13.75, U-statistic = 0.19 and R² = 0.98), as compared to the other two modeling techniques. Optimal forecasting produced by combined modeling scored better than application of the simple ARIMA model. Overall, DSS results showed that the Fuzzy Expected Intervals methodology could be used as a very reliable tool for short-term predictions of fishery landings.     

A new design of evolutionary hybrid optimization of SVR model in predicting the blast-induced ground vibration

This study aims to identify the suitability of hybridizing the firefly algorithm (FA), genetic algorithm (GA), and particle swarm optimization (PSO) with two well-known data-driven models of support vector regression (SVR) and artificial neural network (ANN) to predict blast-induced ground vibration. Here, these combinations are abbreviated using FA–SVR, PSO–SVR, GA–SVR, FA–ANN, PSO–ANN, and GA–ANN models. In addition, a modified FA (MFA) combined with SVR model is also proposed in this study, namely, MFA–SVR. The feasibility of the proposed models is examined using a case study, located in Johor, Malaysia. Then, to provide an objective assessment of performances of the predictive models, their results were compared based on several well known and popular statistical criteria. According to the results, the MFA–SVR with the coefficient of determination (R²) of 0.984 and root mean square error (RMSE) of 0.614 was more accurate model to predict PPV than the PSO–SVR with R² = 0.977 and RMSE = 0.725, the FA–SVR with R² = 0.964 and RMSE = 0.923, the GA–SVR with R² = 0.957 and RMSE = 1.016, the GA–ANN with R² = 0.936 and RMSE = 1.252, the FA–ANN with R² = 0.925 and RMSE = 1.368, and the PSO–ANN with R² = 0.924 and RMSE = 1.366. Consequently, the MFA–SVR model can be sufficiently employed in estimating the ground vibration, and has the capacity to generalize.

     

Assessment of composite beam performance using GWO–ELM metaheuristic algorithm

Composite beams (CBs) include concrete slabs jointed to the steel parts by the shear connectors, which highly popular in modern structures such as high rise buildings and bridges. This study has investigated the structural behavior of simply supported CBs in which a concrete slab is jointed to a steel beam by headed stud shear connector. Determining the behavior of CB through empirical study except its costly process can also lead to inaccurate results. In this case, AI models as metaheuristic algorithms could be effectively used for solving difficult optimization problems, such as Genetic algorithm, Differential evolution, Firefly algorithm, Cuckoo search algorithm, etc. This research has used hybrid Extreme machine learning (ELM)–Grey wolf optimizer (GWO) to determine the general behavior of CB. Two models (ELM and GWO) and a hybrid algorithm (GWO–ELM) were developed and the results were compared through the regression parameters of determination coefficient (R²) and root mean square (RMSE). In testing phase, GWO with the RMSE value of 2.5057 and R² value of 1.2510, ELM with the RMSE value of 4.52 and R² value of 1.927, and GWO–ELM with the RMSE value of 0.9340 and R² value of 0.9504 have demonstrated that the hybrid of GWO–ELM could indicate better performance compared to solo ELM and GWO models. In this case, GWO–ELM could determine the general behavior of CB faster, more accurate and with the least error percentages, so the hybrid of GWO–ELM is more reliable model than ELM and GWO in this study.

     

Generalization performance of support vector machines and neural networks in runoff modeling

Effective one-day lead runoff prediction is one of the significant aspects of successful water resources management in arid region. For instance, reservoir and hydropower systems call for real-time or on-line site-specific forecasting of the runoff. In this research, we present a new data-driven model called support vector machines (SVMs) based on structural risk minimization principle, which minimizes a bound on a generalized risk (error), as opposed to the empirical risk minimization principle exploited by conventional regression techniques (e.g. ANNs). Thus, this stat-of-the-art methodology for prediction combines excellent generalization property and sparse representation that lead SVMs to be a very promising forecasting method. Further, SVM makes use of a convex quadratic optimization problem; hence, the solution is always unique and globally optimal. To demonstrate the aforementioned forecasting capability of SVM, one-day lead stream flow of Bakhtiyari River in Iran was predicted using the local climate and rainfall data. Moreover, the results were compared with those of ANN and ANN integrated with genetic algorithms (ANN-GA) models. The improvements in root mean squared error (RMSE) and squared correlation coefficient (R²) by SVM over both ANN models indicate that the prediction accuracy of SVM is at least as good as that of those models, yet in some cases actually better, as well as forecasting of high-value discharges.     

Comparison of Landsat 8 OLI and Landsat 7 ETM+ for estimating grassland LAI using model inversion and spectral indices: case study of Mpumalanga,South Africa

The leaf area index (LAI) is the key biophysical indicator used to assess the condition of rangeland. In this study, we investigated the implications of narrow spectral response, high radiometric resolution (12 bits), and higher signal-to-noise ratio of the Landsat 8 Operational Land Imager (OLI) sensor for the estimation of LAI. The Landsat 8 LAI estimates were compared to that of its predecessors, namely Landsat 7 Enhanced Thematic Mapper Plus (ETM+) (8 bits). Furthermore, we compared the radiative transfer model (RTM) and spectral indices approaches for estimating LAI on rangeland systems in South Africa. The RTM was inverted using artificial neural network (ANN) and lookup table (LUT) algorithms. The accuracy of the models was higher for Landsat 8 OLI, where ANN (root mean squared error, RMSE = 0. 13; R² = 0. 89), LUT (RMSE = 0. 25; R² = 0. 50), compared to Landsat 7 ETM+, where ANN (RMSE = 0. 35; R² = 0. 60), LUT (RMSE = 0. 38; R² = 0. 50). Compared to an empirical approach, the RTM provided higher accuracy. In conclusion, Landsat 8 OLI provides an improvement for the estimation of LAI over Landsat 7 ETM+. This is useful for rangeland monitoring.     

A comparison of the performance of some extreme learning machine empirical models for predicting daily horizontal diffuse solar radiation in a region of southern Iran

This study uses the empirical models of extreme learning machine (ELM) method to predict daily horizontal diffuse solar radiation (HDSR). As a possibility for modification, the recent hybrid ELM methods such as complex ELM (C-ELM), self-adaptive evolutionary ELM (SaE-ELM), and online sequential ELM (OS-ELM) have been developed for the prediction of the daily HDSR. The empirical model of ELM predicts the HDSR using clearness index as the sole predictor. For this aim, two types of correlations are evaluated: (1) the diffuse fraction-clearness index and (2) the diffuse coefficient-clearness index. The measured diffuse and global solar radiation data sets of southern Iranian cities (Yazd, Shiraz, Bandar Abbas, Bushehr, and Zahedan) are utilized to evaluate the models. The precision of the C-ELM, SaE-ELM, OS-ELM, and ELM models is evaluated for different regions on the basis of five statistical performance evaluation parameters. The results confirm that the performance of hybrid ELM is pretty accurate and trustworthy. Fully complex ELM exhibits the best performance among these hybrid methods, having values of 0.87 and 0.30 for R² and RMSE measures, respectively, in the testing phase. Therefore, it is able to be recognized as an appropriate tool for daily solar radiation forecasting issues.     

Reliability estimation using a genetic algorithm-based artificial neural network: An application to a load-haul-dump machine

In this study, a neural network-based model for forecasting reliability was developed. A genetic algorithm was applied for selecting neural network parameters like learning rate (η) and momentum (μ). The input variables of the neural network model were selected by maximizing the mean entropy value. The developed model was validated by applying two benchmark data sets. A comparative study reveals that the proposed method performs better than existing methods on benchmark data sets. A case study was conducted on a load-haul-dump (LHD) machine operated at a coal mine in Alaska, USA. Past time-to-failure data for the LHD machine were collected, and cumulative time-to-failure was calculated for reliability modeling. The results demonstrate that the developed model performs well with high accuracy (R² = 0.94) in the failure prediction of a LHD machine.     

Performance of radial basis and LM-feed forward artificial neural networks for predicting daily watershed runoff

This study investigated the effects of upstream stations’ flow records on the performance of artificial neural network (ANN) models for predicting daily watershed runoff. As a comparison, a multiple linear regression (MLR) analysis was also examined using various statistical indices. Five streamflow measuring stations on the Cahaba River, Alabama, were selected as case studies. Two different ANN models, multi layer feed forward neural network using Levenberg–Marquardt learning algorithm (LMFF) and radial basis function (RBF), were introduced in this paper. These models were then used to forecast one day ahead streamflows. The correlation analysis was applied for determining the architecture of each ANN model in terms of input variables. Several statistical criteria (RMSE, MAE and coefficient of correlation) were used to check the model accuracy in comparison with the observed data by means of K-fold cross validation method. Additionally, residual analysis was applied for the model results. The comparison results revealed that using upstream records could significantly increase the accuracy of ANN and MLR models in predicting daily stream flows (by around 30%). The comparison of the prediction accuracy of both ANN models (LMFF and RBF) and linear regression method indicated that the ANN approaches were more accurate than the MLR in predicting streamflow dynamics. The LMFF model was able to improve the average of root mean square error (RMSE_ave) and average of mean absolute percentage error (MAPE_ave) values of the multiple linear regression forecasts by about 18% and 21%, respectively. In spite of the fact that the RBF model acted better for predicting the highest range of flow rate (flood events, RMSE_ave/RBF = 26.8 m³/s vs. RMSE_ave/LMFF = 40.2 m³/s), in general, the results suggested that the LMFF method was somehow superior to the RBF method in predicting watershed runoff (RMSE/LMFF = 18.8 m³/s vs. RMSE/RBF = 19.2 m³/s). Eventually, statistical differences between measured and predicted medians were evaluated using Mann-Whitney test, and differences in variances were evaluated using the Levene's test.     

Performance evaluation of hybrid Wavelet-ANN and Wavelet-ANFIS models for estimating evapotranspiration in arid regions of India

This paper evaluates the ability of wavelet transform in improving the accuracy of artificial neural network (ANN) and adaptive neuro-fuzzy interface systems (ANFIS) models. In this study, the performance of hybrid Wavelet-ANN and Wavelet-ANFIS models for estimating daily evapotranspiration in arid regions was evaluated. Prior to the development of models, gamma test was used to identify the best input combinations that could be used under limited data scenario. Performance of the proposed hybrid models was compared to ANN, ANFIS, and conventionally used Hargreaves equation. The results revealed that use of wavelet transform as data preprocessing technique enhanced the efficiency of ANN and ANFIS models. Wavelet-ANN and Wavelet-ANFIS performed reasonably better than other models. Better handling of wavelet-decomposed input variables enabled Wavelet-ANN models to perform slightly better than the Wavelet-ANFIS models. W-ANN2 (RMSE = 0.632 mm/day and R = 0.96) was found to be the best model for estimating daily evapotranspiration in arid regions. The proposed W-ANN2 model used second-level db3 wavelet-decomposed subseries of temperature and previous day evapotranspiration values as inputs. The study concludes that hybrid Wavelet-ANN and Wavelet-ANFIS models can be effectively used for modeling evapotranspiration.

     

Estimation of soil dispersivity using soft computing approaches

The accurate estimation of soil dispersivity (α) is required for characterizing the transport of contaminants in soil. The in situ measurement of α is costly and time-consuming. Hence, in this study, three soft computing methods, namely adaptive neuro-fuzzy inference system (ANFIS), artificial neural network (ANN), and gene expression programming (GEP), are used to estimate α from more readily measurable physical soil variables, including travel distance from source of pollutant (L), mean grain size (D ₅₀), soil bulk density (ρ _b), and contaminant velocity (V _c). Based on three statistical metrics [i.e., mean absolute error, root-mean-square error (RMSE), and coefficient of determination (R ²)], it is found that all approaches (ANN, ANFIS, and GEP) can accurately estimate α. Results also show that the ANN model (with RMSE = 0.00050 m and R ² = 0.977) performs better than the ANFIS model (with RMSE = 0.00062 m and R ² = 0.956), and the estimates from GEP are almost as accurate as those from ANFIS. The performance of ANN, ANFIS, and GEP models is also compared with the traditional multiple linear regression (MLR) method. The comparison indicates that all of the soft computing methods outperform the MLR model. Finally, the sensitivity analysis shows that the travel distance from source of pollution (L) and bulk density (ρ _b) have, respectively, the most and the least effect on the soil dispersivity.

     

A novel solution for simulating air overpressure resulting from blasting using an efficient cascaded forward neural network

Air overpressure (AOp) is a hazardous effect induced by the blasting method in surface mines. Therefore, it needs to be predicted to reduce the potential risk of damage. The aim of this study is to offer an efficient method to predict AOp using a cascaded forward neural network (CFNN) trained by Levenberg–Marquardt (LM) algorithm, called the CFNN-LM model. Additionally, a generalized regression neural network (GRNN) and extreme learning machine (ELM) were employed to demonstrate the accuracy level of the proposed CFNN-LM model. To conduct the CFNN-LM, GRNN, and ELM models, an extensive database, related to four quarry sites in Malaysia, was used including 62 sets of dependent and independent parameters. Next, the performances of the aforementioned models were checked and discussed through statistical criteria and efficient graphical tools. Finally, the results showed the superiority of CFNN-LM (R² = 0.9263 and RMSE = 3.0444) over GRNN (R² = 0.7787 and RMSE = 5.1211) and ELM (R² = 0.6984 and RMSE = 6.2537) models in terms of prediction accuracy. Furthermore, three different regression analysis metrics were used to perform the sensitivity analysis, and according to the obtained results, the maximum charge per delay (\(\beta\) = 0.475, SE = 0.115, t-test = 4.125) was considered as the most influential feature in modeling the AOp.

     

Assessing the efficacy of Landsat-8 OLI imagery derived models for remotely estimating chlorophyll-a concentration in the Upper Ganga River,India

ABSTRACT

Chlorophyll-a (chl-a) serves as an indicator of productivity in surface water. Estimating chl-a concentration is pivotal for monitoring and subsequent conservation of surface water quality. Artificial neural network (ANN) based models were validated and tested for their efficacy against various regression models to determine the chl-a concentration in the Upper Ganga river. Landsat-8 Operational Land Imager (OLI) surface reflectance (SR) imagery for May and October along with in-situ data over a period of 2 years (2016–2017) was used to develop and validated models. Regression model performance was acceptable with a coefficient of determination (R²) of 0.57, 0.63, 0.66 and 0.68 for linear, exponential, logarithmic and power model, respectively. However, there was a significant improvement in the efficacy of chl-a determination using ANN model performance having a root mean square error (RMSE) of 1.52 µg l^–1 and R² = 0.97 in comparison to the best-performing regression model (power) with RMSE = 9.86 µg l^–1 and R² = 0.68. ANN exhibited comparatively more precise spatial and seasonal variability with mean absolute error (MAE) of 1.26 µg l^–1 as compared to the best regression model (power) MAE = 7.98 µg l^–1 suggesting the applicability of ANN for large-scale spatial and temporal monitoring river stretches using Landsat-8 OLI SR images.