Search

Deep learning excels at managing spatial and temporal time series with variable patterns for streamflow forecasting, but traditional machine learning algorithms may struggle with complicated data, including non-linear and multidimensional complexity. Empirical heterogeneity within watersheds and limitations inherent to each estimation methodology pose challenges in effectively measuring and appraising hydrological statistical frameworks of spatial and temporal variables. This study emphasizes streamflow forecasting in the region of Johor, a coastal state in Peninsular Malaysia, utilizing a 28-year streamflow-pattern dataset from Malaysia’s Department of Irrigation and Drainage for the Johor River and its tropical rainforest environment. For this dataset, wavelet transformation significantly improves the resolution of lag noise when historical streamflow data are used as lagged input variables, producing a 6% reduction in the root-mean-square error. A comparative analysis of convolutional neural networks and artificial neural networks reveals these models’ distinct behavioral patterns. Convolutional neural networks exhibit lower stochasticity than artificial neural networks when dealing with complex time series data and with data transformed into a format suitable for modeling. However, convolutional neural networks may suffer from overfitting, particularly in cases in which the structure of the time series is overly simplified. Using Bayesian neural networks, we modeled network weights and biases as probability distributions to assess aleatoric and epistemic variability, employing Markov chain Monte Carlo and bootstrap resampling techniques. This modeling allowed us to quantify uncertainty, providing confidence intervals and metrics for a robust quantitative assessment of model prediction variability.

At Klang Valley, ground-level ozone is a significant source of air pollution. Ozone (O3) concentration is affected by meteorological conditions and air pollutants. Linear Regression Models (LRM), Regression Trees (RT), Support Vector Machines (SVM), Ensembles of Trees (ET), Gaussian Process Regression (GPR), and Neural Networks (NN) are utilized in a thorough analysis to determine the accuracy of various machine learning in forecasting the ground level O3 concentration. The primary associated contributions from this research are comparisons of regression statistical model performance based on indicators of root mean squared error (RMSE), coefficient of determination (R2), mean squared error (MSE), mean absolute error (MAE), prediction speed, and training time of regression models. Overall, exponential GPR outperformed other regression models in scenario 1 (S-1), scenario 2 (S-2), scenario (S-3), and scenario 4 (S-4) by incorporating multiple number of lags into respective scenarios and new method of testing ''re-substitution'' performed more reliable and consistent than applying identical datasets to 20 % of model testing. The findings showed that GPR performed accurate results with R2 = 0.98, 0.95, 0.96, and 0.96 for S-1, S-2, S-3 and S-4 respectively.

Drought has been the main environmental issue in Peninsular Malaysia. Hence, this study undertook a thorough evaluation of drought assessment methodologies and focused on the temporal analysis of multiple drought indices, namely, the standardised precipitation index (SPI), deciles index (DI), percent of normal precipitation index (PNPI), rainfall anomaly index (RAI) and Z-score index (ZSI) – across timescales of one-, six- and 12-month durations. This assessment incorporates the average moving range (AMR), Mann–Kendall (MK) test and Sen's slope estimator in temporal analysis and the results showed that shorter timescales lead to higher fluctuation in AMR values, indicating short-term droughts are best assessed using drought indices of shorter timescale. It was found that most drought indices exhibited a similar trend and trend magnitude in all timescales. SPI is utilised as the standard model for the accuracy evaluation of drought indices using root mean square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE). The results showed that ZSI has the highest accuracy of all indices. The novelty of this study lies in evaluating the accuracy and temporal characteristics of precipitation-based drought indices in tropical areas, particularly in Peninsular Malaysia. HIGHLIGHTS Drought indices have higher fluctuations on shorter timescales.; Northwest, northeast and southwest regions of Peninsular Malaysia experienced an increasing trend in drought indices.; Temporal analysis suggested that significant trends could be detected in drought indices computed on larger timescales.; Most drought indices showcased a similar trend in assessing drought.; SPI and ZSI performed similarly in drought monitoring.;

Using a comparison of three different major types, the best predictive model was determined. Statistical models and machine learning algorithms automatically learn and improve based on data. Deep learning uses neural networks to learn complex data patterns and relationships. A combination of satellite imagery, radar data, and ground-based observations are used and using aircraft or satellites, and remote sensing (RS) collects data on distant objects or locations. Satellites and radar are used to gather regional precipitation data for hybrid models. An algorithm trained on historical rainfall measurements would then process the data. Using remote monitoring instrument input features, the machine-learning model can predict precipitation. Evaluation of machine learning regression methods is based on the degree of agreement between predicted and observed values. The RMSE, R2, and MAE statistical measures check on the precision of a prediction or forecasting model. Machine learning excels at rainfall prediction regardless of climate or timescale. As one of the more popular models for predicting rainfall, the LSTM models demonstrate their superiority. Remote sensing and hybrid predictive models should be investigated further due to their scarcity.

Abstract Natural calamities like droughts have harmed not just humanity throughout history but also the economy, food, agricultural production, flora, animal habitat, etc. A drought monitoring system must incorporate a study of the geographical and temporal fluctuation of the drought characteristics to function effectively. This study investigated the space–time heterogeneity of drought features across Sabah and Sarawak, East Malaysia. The Standardized Precipitation Index (SPIs) at timescales of 1-month, 3-months, and 6-months was selected to determine the spatial distribution of drought characteristics. Rainfall hydrographs for the area for 30 years between 1988 and 2017 have been used in this study. A total of six five-year sub-periods were studied, with an emphasis on the lowest and highest drought occurrence. The sub-periods were a division of the 30 years over an arbitrary continual division for convenience. The results showed that the sub-periods 1993–1997 and 2008–2012 had the highest and lowest comparative drought events. The drought conditions were particularly severe in Central and Eastern parts of East Malaysia, owing to El Nino events and the country's hilly terrain. Understanding how and when drought occurs can aid in establishing and developing drought mitigation strategies for the region.

Abstract Due to excessive streamflow (SF), Peninsular Malaysia has historically experienced floods and droughts. Forecasting streamflow to mitigate municipal and environmental damage is therefore crucial. Streamflow prediction has been extensively demonstrated in the literature to estimate the continuous values of streamflow level. Prediction of continuous values of streamflow is not necessary in several applications and at the same time it is very challenging task because of uncertainty. A streamflow category prediction is more advantageous for addressing the uncertainty in numerical point forecasting, considering that its predictions are linked to a propensity to belong to the pre-defined classes. Here, we formulate streamflow prediction as a time series classification with discrete ranges of values, each representing a class to classify streamflow into five or ten, respectively, using machine learning approaches in various rivers in Malaysia. The findings reveal that several models, specifically LSTM, outperform others in predicting the following n-time steps of streamflow because LSTM is able to learn the mapping between streamflow time series of 2 or 3 days ahead more than support vector machine (SVM) and gradient boosting (GB). LSTM produces higher F1 score in various rivers (by 5% in Johor, 2% in Kelantan and Melaka and Selangor, 4% in Perlis) in 2 days ahead scenario. Furthermore, the ensemble stacking of the SVM and GB achieves high performance in terms of F1 score and quadratic weighted kappa. Ensemble stacking gives 3% higher F1 score in Perak river compared to SVM and gradient boosting.

AbstractIt is critical to monitor water quality to keep water bodies ecologically healthy and facilitate the sustainable development of Kenyir Lake. Water quality differs temporally and spatially and is affected by several factors. Typically, water quality inspection systems are cost- and labour-intensive depending on water quality indicator count and sampling frequency. Optimising the frequency and location of water quality sampling is crucial. This study focused on collecting water samples from 22 locations in Kenyir Lake during different seasons (normal, dry, and wet). The study aimed to assess the spatial and temporal variations in the water quality of Kenyir Lake based on multivariate statistical methods. In this study, the following water quality parameters were selected for analysis: temperature, dissolved oxygen (DO), pH, biochemical oxygen demand (BOD), chemical oxygen demand (COD), total suspended solids (TSS), and ammoniacal nitrogen (NH3-N). In addition, a water quality index was also calculated. GIS software was used to assess water quality data, and various multivariate statistical methods like cluster analysis (CA), discriminant analysis (DA), and principal component analysis (PCA) were employed. The outcome shows minor spatial differences concerning Kenyir Lake; however, the temporal variations were noteworthy during this study duration. Cluster analysis divided the locations into 3 clusters with TSS being key parameter affecting the spatial differences in water quality. Stepwise discriminant analysis based on three parameters, pH, temperature, and TSS, produced the associated classification matrix that correctly estimated 69.7% of the input. NH3-N and TSS were found to be the two critical aspects that affect water quality during dry, wet, or normal climatic conditions.

Sensitivity analysis (SA) intends to identify the key meteorological variables that affect the performance of reference crop evapotranspiration (ET0) models. It is of importance in assessing the variability of meteorological variables and ET0, especially in the face of increasing climate uncertainties. However, the surging of inconsistencies resulting from global changes in meteorological conditions due to climate change have impacted the ET0 model estimation in different regions, with detrimental effects on water resources and crop production. Therefore, efficient SA is necessary to evaluate the impact of changes in meteorological variables that influence ET0 model estimation. This mini review analyses the various SA methods applied in the field of ET0, based on a comprehensive and comparative analysis of existing SA methods from all around the world. The study discusses the advantages and disadvantages of each SA method, as well as the factors affecting the SA of ET0. The study also provides future prospects that may contribute to more solid and powerful analysis for ET0 model estimations and conclusions. HIGHLIGHTS The local and global sensitivity analysis for ET0 estimation were reviewed.; The categories, methods, and limitations for SA were examined.; The factors affecting the SA of ET0 were discussed.; The new possible research directions for SA in ET0 estimations are outlined.;

In consideration of the distinct behavior of machine learning (ML) algorithms, six well-defined ML used were carried out in this study for predicting sea level on a day-to-day basis. Data compiled from 1985 to 2018 was utilized for training and testing the developed models. An assessment of the multiple statistics-driven regression algorithms resulted such that each tested location was associated with a particular preferred model. The following were the developed best models for their respective study areas: In Peninsular Malaysia, the interactions linear regression model was the best at Pulau Langkawi (RMSE = 19.066), the Matern 5/2 gaussian process regression model at Geting (RMSE = 49.891), and the trilayered artificial neural network at Pulau Pinang (RMSE = 20.026), while the linear regression model was the best at Sandakan in Sabah, East Malaysia (RMSE = 14.054). Other metrics, such as MAE and R-square, were also at their best values, each providing its best values, further substantiating the RMSE respectively, at each of the study areas. These empirical statistics (or metrics) also revealed that despite employing sea level as the sole parameter, results obtained were exceptional better when utilizing a 7-day lag, regardless of the model used. Notably, lag variables with less than a 7-day lag could degrade the model's accuracy in representing ground reality. The study emphasizes the importance of thorough training and testing of ML to aid decision-makers in developing mitigation actions for the climate change phenomena of sea level rise through reliable ML.

Abstract For decision-making in farming, the operation of dams and irrigation systems, as well as other fields of water resource management and hydrology, evaporation, as a key activity throughout the universal hydrological processes, entails efficient techniques for measuring its variation. The main challenge in creating accurate and dependable predictive models is the evaporation procedure's non-stationarity, nonlinearity, and stochastic characteristics. This work examines, for the first time, a transformer-based deep learning architecture for evaporation prediction in four different Malaysian regions. The effectiveness of the proposed deep learning (DL) model, signified as TNN, is evaluated against two competitive reference DL models, namely Convolutional Neural Network and Long Short-Term Memory, and with regards to various statistical indices using the monthly-scale dataset collected from four Malaysian meteorological stations in the 2000–2019 period. Using a variety of input variable combinations, the impact of every meteorological data on the E p forecast is also examined. The performance assessment metrics demonstrate that compared to the other benchmark frameworks examined in this work, the developed TNN technique was more precise in modelling monthly water loss owing to evaporation. In terms of predictive effectiveness, the proposed TNN model, enhanced with the self-attention mechanism, outperforms the benchmark models, demonstrating its potential use in the forecasting of evaporation. Relating to application, the predictive model created for E p projection offers a precise estimate of water loss due to evaporation and can thus be used in irrigation management, agriculture planning based on irrigation, and the decrease in fiscal and economic losses in farming and related industries where consistent supervision and estimation of water are considered necessary for viable living and economy.

Abstract Hydrologists rely extensively on anticipating river streamflow (SF) to monitor and regulate flood management and water demand for people. Only a few simulation systems, where previous techniques failed to anticipate SF data quickly, let alone cost-effectively, and took a long time to execute. The bat algorithm (BA), a meta-heuristic approach, was used in this study to optimize the weights and biases of the artificial neural network (ANN) model. The proposed hybrid work was validated in five different study areas in Malaysia. The statistical tests analysis of the preliminary results revealed that hybrid BA-ANN was superior to forecasting the SF at all five selected study areas, with average RMSE values of 0.103 m3/s for training and 0.143 m3/s for testing as compared to ANN standalone training and testing yielding 0.091 m3/s and 0.116 m3/s, respectively. This finding signifies that the implementation of BA into the ANN model resulted in a 20% improvement. In addition, with an R 2 score of 0.951, the proposed model showed a better correlation than the 0.937 value of R 2 of standard ANN. Nonetheless, while the proposed work outperformed the conventional ANN, the Taylor diagram, violin plot, relative error, and scatter plot findings confirmed the disparities in the proposed work’s performance throughout the research regions. The findings of these evaluations highlighted that the adaptability of the proposed works would need detailed investigation because its performance differed from case to case.

A reliable model to predict the changes in the water levels in a river is crucial for better planning to mitigate any risk associated with flooding. In this study, six different Machine Learning (ML) algorithms were developed to predict the river’s water level, on a daily basis based on collected data from 1990 to 2019 which were used to train and test the proposed models. Different input combinations were explored to improve the accuracy of the model. Statistical indicators were calculated to examine the reliability of the proposed models with other models. The comparison of several data-driven regression methods indicate that the exponential Gaussian Process Regression (GPR) model offered better accuracy in predicting daily water levels with respect to different assessment criteria. The GPR model was then used to predict the water level after sorting the data based on 10 days maximum and minimum values of the water level, and the results proved the success of this model in catching the extremes of the water levels. In addition to that, based on two uncertainty indicators, it was concluded that the proposed model, the GPR, was capable of predicting the water level of the river with high precision and less uncertainty where the computed using the 95% prediction uncertainty (95PPU) and the d-factor were found to be equal to 98.276 and 0.000525, respectively. The findings of this study show the efficacy of the GPR model in capturing the changes in the water level in a river. Due to the importance of the water level of a river being an parameter for flood monitoring, this technique is likely beneficial to the design of the mitigation strategies for future flooding events.

Modeling wind speed has a significant impact on wind energy systems and has attracted attention from numerous researchers. The prediction of wind speed is considered a challenging task because of its natural nonlinear and random characteristics. Therefore, machine learning models have gained popularity in this field. In this paper, three machine learning approaches – Gaussian process regression (GPR), bagged regression trees (BTs) and support vector regression (SVR) – were applied for prediction of the weekly wind speed (maximum, mean, minimum) of the target station using other stations, which were specified as reference stations. Daily wind speed data, gathered via the Malaysian Meteorological Department at 14 measuring stations in Malaysia covering the period between 2000 and 2019, were used. The results showed that the average weekly wind speed had superior performance to the maximum and minimum wind speed prediction. In general, the GPR model could effectively predict the weekly wind speed of the target station using the measured data of other stations. Errors found in this model were within acceptable limits. The findings of this model were compared with the measured data, and only Kota Kinabalu station showed an unacceptable range of prediction. To investigate the prediction performance of the proposed model, two models were used as the comparison models: the BTs model and SVR model. Although the comparison of GPR with the BTs model at Kuching station showed slightly better performance for the BTs model in maximum and minimum wind speed prediction, the prediction outcomes of the other 13 stations showed better performance for the proposed GPR model. Moreover, the proposed model generated smaller prediction errors than the SVR model at all stations.

Abstract The optimization of dam reservoir operations is of the utmost importance, as operators strive to maximize revenue while minimizing expenses, risks, and deficiencies. Metaheuristics have recently been investigated extensively by researchers in the management of dam reservoirs. But the animal-concept-based metaheuristic algorithm with Lévy flight integration approach has not been used at Karun-4. This paper investigates the optimization of dam reservoir operation using three unexplored metaheuristics: the whale optimization algorithm (WOA), the Levy-flight WOA (LFWOA), and the Harris hawks optimization algorithm (HHO). Utilizing a time series data set on the hydrological and climatic characteristics of the Karun-4 hydroelectric reservoir in Iran, an analysis was conducted. The objective functions and constraints of the Karun-4 hydropower reservoir were examined throughout the optimization procedure. HHO produces the best optimal value, the least-worst optimal value, the best average optimal value, and the best standard deviation (SD) with scores of 0.000026, 0.001735, 0.000520, and 0.000614, respectively, resulting in the best overall ranking mean (RM) with a score of 1.5 at Karun-4. Throughout the duration of the test, the optimized trends of water release and water storage indicate that HHO is superior to the other investigated metaheuristics. WOA has the best correlation of variation (CV) with a score of 0.090195, while LFWOA has the best convergence rate (3.208 s) and best CPU time. Overall, it can be concluded that HHO has the most desirable performance in terms of optimization. Yet, current studies indicate that both WOA and LFWOA generate positive and comparable outcomes.

Cities are growing geographically in response to the enormous increase in urban population; consequently, comprehending growth and environmental changes is critical for long-term planning. Urbanization transforms naturally permeable surfaces into impermeable surfaces, causing an increase in urban land surface temperature, leading to the phenomenon known as urban heat islands. The urban heat islands are noticeable across Malaysia's rural communities and villages, particularly in Kuala Lumpur. These effects must be addressed to slow, if not halt, climate change and meet the Paris Agreement's 2030 goal. The study posits an application of thermal remote sensing utilizing a space-borne satellite-based technique to demonstrate urban evolution for urban heat island analysis and its relationship to land surface temperature. The urban heat island (UHI) was analyzed by converting infrared radiation into visible thermal images utilizing thermal imaging from remote sensing techniques. The heat island is validated by reference to the characteristics of the normalized difference vegetation index (NDVI), which define the land surface temperature (LST) of distinct locations. Based on the digital information from the satellite, the highest temperature difference between urban and rural regions for a few chosen cities in 2013 varied from 10.8 to 25.5 °C, while in 2021, it ranged from 16.1 to 26.73 °C, highlighting crucial temperature changes. The results from ANOVA test has substantially strengthened the credibility of the significant temperature changes. Some notable reveals are as follows: The Sungai Batu area, due to its rapid development and industry growth, was more vulnerable to elevated urban heat due to reduced vegetation cover; therefore, higher relative vulnerability. Contrary, the Bukit Ketumbar area, which region lies in the woodland region, experienced the lowest, with urban heat islands reading from 2013 at −0.3044 and 0.0154 in 2021. It shows that despite having urban heat islands increase two-fold from 2013 to 2021, increasing the amount of vegetation coverage is a simple and effective way of reducing the urban heat island effect, as evidenced by the low urban heat islands in the Bukit Ketumbar woodland region. The study findings are critical for advising municipal officials and urban planners to decrease urban heat islands by investing in open green spaces.

Abstract Evaporation is the primary aspect causing water loss in the hydrological cycle; therefore, water loss must be precisely measured. Evaporation is an intricate nonlinear process occurring as a result of several climatic aspects. The purpose of this research is to assess the feasibility of using Random Forest (RF) and two deep learning techniques, namely convolutional neural network (CNN), and deep neural network (DNN) to accurately estimate monthly pan evaporation rates. Month-based weather data gathered from four Malaysian weather stations during the 2000–2019 timeframe was used to train and evaluate the models. Several input attributes (predictor variables) were investigated to select the most suitable variables for machine learning models. Every approach was tested with several models, each with a different set of model aspects and input parameter combinations. The formulated ML approaches were benchmarked against two commonly used empirical methods: Stephens & Stewart and Thornthwaite. Model outcomes were assessed using standard statistical measures to determine their effectiveness in predicting evaporation. The results indicated that the three ML models developed in the study performed better than empirical models and could significantly improve the precision of monthly Ep estimates even with the identical input sets. The performance assessment metrics also show that the formulated CNN approach was acceptable for modelling monthly water loss due to evaporation with a higher degree of accuracy than other ML frameworks explored in this study. In addition, the CNN framework outperformed other AI techniques evaluated for the same areas using identical data inputs. The investigation’s findings in relation to the various performance criteria show that the proposed CNN model is capable of capturing the highly non-linearity of evaporation and could be regarded as an effective tool to predict evaporation.

Iraq is facing a dire water crisis due to the decrease in water quantities flow in Tigris and Euphrates Rivers. Due to population growth, several studies estimated the water shortage in 2035 to be 44 Billion Cubic Meter (BCM). Thus, Water Budget-Salt Balance Model (WBSBM) has been developed, applied and examined for the Euphrates River basin to compute the net water saving from Non-Conventional Water Resources (NCWRs). WBSBM includes 4-stages; the first is to identify the required data correspond to the conventional water resources in the study-area. The second stage is demonstrating the water-users activities. Thirdly, develop model through the proposed NCWR projects that reflect the required data. The final stage involves net water saving computation while applying all the NCWR projects simultaneously. The results obtained the optimal potential net water saving amount, which are 6.823 and 6.626 BCM/year in 2025 and 2035, respectively. In conclusion, the proposed WBSBM model has comprehensively examined different scenarios of utilizing NCWRs and has determined the optimal potential the net water saving amounts.

Machine learning models have been effectively applied to predict certain variable in several engineering applications where the variable is highly stochastic in nature and complex to identify utilizing the classical mathematical models. Therefore, this study investigates the capability of various machine learning algorithms in predicting the power production of a reservoir located in China using data from 1979 to 2016. In this study, different supervised and unsupervised machine learning algorithms are proposed: artificial neural network (ANN), AutoRegressive Integrated Moving Aveage (ARIMA) and support vector machine (SVM). Three different scenarios are examined, such as scenario1 (SC1): used to predict daily power generation, scenario 2 (SC2): used to predict power generation for monthly prediction and scenario 3 (SC3): used to predict hydropower generation (HPG) seasonally. The statistical analysis and pre-processing techniques were applied to the raw data before developing the models. Five statistical indexes are employed to evaluate the performances of various models developed. The results indicate that the proposed models can be used to predict HPG efficiently and could be an effective method for energy decision-makers. The sensitivity analyses found the most effective models for predicting HPG for three scenarios using graphical distribution data (Taylor diagram). Regarding the uncertainty analysis, 95PPU and d-factors were adopted to measure the uncertainties of the best models for ANN and SVM. The results presented that the value of 95PPU for all models falls into the range between 80% and 100%. As for the d-factor, all values in all scenarios are less than one.

Abstract Floods and droughts are environmental phenomena that occur in Peninsular Malaysia due to extreme values of streamflow (SF). Due to this, the study of SF prediction is highly significant for the purpose of municipal and environmental damage mitigation. In the present study, machine learning (ML) models based on the support vector machine (SVM), artificial neural network (ANN), and long short-term memory (LSTM), are tested and developed to predict SF for 11 different rivers throughout Peninsular Malaysia. SF data sets for the rivers were collected from the Malaysian Department of Irrigation and Drainage. The main objective of the present study is to propose a universal model that is most capable of predicting SFs for rivers within Peninsular Malaysia. Based on the findings, the ANN3 model which was developed using the ANN algorithm and input scenario 3 (inputs consisting of previous 3 days SF) is deduced as the best overall ML model for SF prediction as it outperformed all the other models in 4 out of 11 of the tested data sets; and obtained among the highest average RMs with a score of 3.27, hence indicating that the model is very adaptable and reliable in accurately predicting SF based on different data sets and river case studies. Therefore, the ANN3 model is proposed as a universal model for SF prediction within Peninsular Malaysia.

Potential Evapotranspiration (PET) functions as an indicator to estimate the amount of water loss to the atmosphere. Over the years, global climate change has eventually led to a change of PET capacity and this has affected the agricultural sector and water resource management. The objective of this study was to determine the best PET estimation method as well as to carry out a trend analysis and stationarity test of PET in Peninsular Malaysia. The Mann–Kendall (MK) test and Sen's slope estimator were applied for the trend analysis while the Augmented Dickey–Fuller (ADF) test and Kwiatkowski–Phillips–Schmidt–Shin (KPSS) test were applied for the stationarity test. The findings showed that Pulau Langkawi and Kuantan stations exhibited increasing trends while Bayan Lepas station exhibited decreasing trends for the daily, monthly, and annual PET time series. The daily, monthly, and annual PET time series at Bayan Lepas, Ipoh, Subang and Muadzam Shah stations were found to be stationary. Overall, the PET trend was found to be higher in the coastal regions and stationary in the mountainous region. HIGHLIGHTS The Mann–Kendall test and Sen's slope estimator were used to analyse the trends of potential evapotranspiration (PET) series.; The stationarity of PET series was evaluated using the Kwiatknowski–Phillips–Schmidt (KPSS) test.; The Turc method gave the best performances in estimating PET.; The coastal regions exhibited increasing PET trend.; The daily, monthly and annual PET trends in the mountainous region were stationary.;

Machine learning models hybridized with optimization algorithms have been applied to many real-life applications, including the prediction of water quality. However, the emergence of newly developed advanced algorithms can provide new scopes and possibilities for further enhancements. In this study, the least-square support vector machine (LSSVM) integrated with advanced optimization algorithms is presented, for the first time, in the prediction of the water quality index (WQI) at the Klang River of Malaysia. Thereafter, the LSSVM model using RBF kernel was optimized using the hybrid particle swarm optimization and genetic algorithm (HPSOGA), whale optimization based on self-adapting parameter adjustment and mix mutation strategy (SMWOA) as well as ameliorative moth-flame optimization (AMFO) separately. It was found that the SMWOA-LSSVM model had the better performance for WQI prediction by having the best achievement root mean square error (RMSE), mean absolute error (MAE), coefficient of determination (R2) and mean absolute percentage error (MAPE). Comprehensive comparison was done using the global performance indicator (GPI), whereby the SMWOA-LSSVM had the highest average score of 0.31. This could be attributed to the internal architecture of the SMWOA, which was catered to avoid local optima within short optimization period. HIGHLIGHTS Advanced optimization algorithms were applied, for the first time, in WQI prediction.; LSSVM using RBF as kernel function was found to be the best model.; All the hybrid LSSVM integrated with optimization algorithms had improved in accuracy, against the base models.; SMWOA-LSSVM was found to be the most suitable hybrid model for WQI prediction at the Klang River.;

Ever since the first introduction of Artificial Intelligence into the field of hydrology, it has further generated immense interest in researching aspects for further improvements to hydrology. This can be seen in the rising number of related works published. This culminated further with the combination of pioneering optimization techniques. Who would have thought that the birds and the bees can offer advances in the mathematical sciences and so have the ants too? The ingenuity of humans is spelled out in the algorithms that mimic many natural activities, like pack hunting by the wolves! This review paper serves to broadcast more of the intriguing interest in newfound procedures in optimal forecasting. Reservoirs are the main and most efficient water storage facilities for managing uneven water distribution. However, due to the major global climate changes which affect rainfall trend and weather, it has been a necessity to find an alternative solution for effective conventional water balance. A multifunctional reservoir operation appears to require the operator to make wise decisions to achieve an optimal reservoir operation. One of the most important aspects of all this is the forecasting of streamflows. For this, Artificial Intelligence (AI) seems to be the best alternative solution; as in the past three decades, there has been a drastic increase in building and developing AI models for forecasting and modelling unstable patterns in various hydrological fields. Nevertheless, AI models are also required to be optimized in tandem to achieve the best result, leading thus to the desirous forming of hybrid models between a standalone AI model and optimization techniques. This comprehensive study categorizes machine learning into three main categories, together with the optimization techniques, and will next explore the various AI model used for different hydrology fields along with the most common optimization techniques. Summarization of findings under every section is provided. Some advantages and disadvantages found through literature reviews are summarized for ease of reference. Finally, future recommendations and overall conclusions drawn from the results of researchers are included. This current review focuses on papers from high-impact factor publications based on 10 years starting from (2009 to 2020).

Abstract High loads of suspended sediments in rivers are known to cause detrimental effects to potable water sources, river water quality, irrigation activities, and dam or reservoir operations. For this reason, the study of suspended sediment load (SSL) prediction is important for monitoring and damage mitigation purposes. The present study tests and develops machine learning (ML) models, based on the support vector machine (SVM), artificial neural network (ANN) and long short-term memory (LSTM) algorithms, to predict SSL based on 11 different river data sets comprising of streamflow (SF) and SSL data obtained from the Malaysian Department of Irrigation and Drainage. The main objective of the present study is to propose a single model that is capable of accurately predicting SSLs for any river data set within Peninsular Malaysia. The ANN3 model, based on the ANN algorithm and input scenario 3 (inputs consisting of current-day SF, previous-day SF, and previous-day SSL), is determined as the best model in the present study as it produced the best predictive performance for 5 out of 11 of the tested data sets and obtained the highest average RM with a score of 2.64 when compared to the other tested models, indicating that it has the highest reliability to produce relatively high-accuracy SSL predictions for different data sets. Therefore, the ANN3 model is proposed as a universal model for the prediction of SSL within Peninsular Malaysia.

Sediment deposition causes the reduction of aquatic habitats and increase of water velocities within rivers, which negatively impacts the environment and the surrounding ecology. This makes the prediction of river sediment deposition a key factor for the protection of river environments. The prediction of sediment deposition in rivers through the integration of satellite imagery and unsupervised machine learning is beneficial and convenient, as it is less resource-intensive due to not requiring ground-truth data. The Terengganu River in Malaysia is used as a case study in this research. This study aims to discuss satellite imagery's key preparation processes, namely image correction and identification of determinant image bands through a correlation analysis. Satellite imagery of the Terengganu River between 1989 and 2019 is obtained from the United States Geological Survey (USGS). Image correction is successfully implemented on the available satellite imagery with the results shown in this study. Through the performed correlation analysis, the study finds that the determinant image bands for river sediment deposition prediction using unsupervised machine learning are the NST spectral bands, which consist of the NIR, SWIR, and TIR bands. This is due to the NST spectral bands exhibiting low correlations with respect to the RGB bands. It is found that correlation coefficients between the NIR band and red, green, and blue bands are generally the lowest, especially in 2009 with values of 0.1087, 0.2085, and 0.1252, respectively. This indicates that the NIR band is the most important determinant image band in predicting river sediment deposition. This study also identifies k-means, clustering large application (Clara), and hierarchical agglomerative clustering (HAC) as suitable unsupervised machine learning algorithms to be utilized in predicting river sediment deposition. Studies on the application of unsupervised machine learning algorithms on satellite imagery in the field of river sediment deposition prediction are currently scarce, possibly due to the gap of knowledge on the initial steps required for such application. Therefore, this study's novelty is the introduction and discussion on critical preliminary processes, specifically image correction and identification of determinant image bands, that are required for the successful implementation of unsupervised machine learning algorithms on satellite imagery for the prediction of river sediment deposition.

Potential evapotranspiration (PET) is an important parameter for the operation of irrigation projects and water resources management. The globally recognized PET estimation model, the FAO-56 Penman–Monteith (FAO-56 PM) model, had been criticized for its requirement of many detailed meteorological variables, but nevertheless has been accepted as the baseline model in many worldwide studies. The performances of different PET models can be found to be excellent for a specific location but may not be representative in other regions. The aim of this study is to select the most suitable PET model to estimate PET in Malaysia. Three radiation-based models and four temperature-based models were compared with the FAO-56 PM model at seven selected meteorological stations in Peninsular Malaysia. The mean bias error, relative error (Re) and normalized root-mean-square error (NRMSE) and coefficient of determination (R2) were used to evaluate the performances of the PET models. The Re values of Turc models were below 0.2 at all stations, while Priestly–Taylor, Thornthwaite, Thornthwaite-corrected and Blaney–Criddle models were above 0.2. The Makkink and Hargreaves–Samani models were below 0.2 at most of the stations. Thus, the Turc model was recommended as the best model to estimate PET in Peninsular Malaysia. HIGHLIGHTS This study applied various empirical potential evapotranspiration (PET) models to estimate the PET.; Radiation-based models and temperature-based models were compared with the FAO Penman–Monteith model.; MBE, Re, NRMSE and R2 were used to evaluate the performances of PET models.; The Turc model was found to give superior performance.; The outcomes can be used as a reference for water resources-related project design.;

Abstract Evaporation is a key element for water resource management, hydrological modelling, and irrigation system designing. Monthly evaporation (Ep) was projected by deploying three machine learning (ML) models included Extreme Gradient Boosting, ElasticNet Linear Regression, and Long Short-Term Memory; and two empirical techniques namely Stephens-Stewart and Thornthwaite. The aim of this study is to develop a reliable generalised model to predict evaporation throughout Malaysia. In this context, monthly meteorological statistics from two weather stations in Malaysia were utilised for training and testing the models on the basis of climatic aspects such as maximum temperature, mean temperature, minimum temperature, wind speed, relative humidity, and solar radiation for the period of 2000–2019. For every approach, multiple models were formulated by utilising various combinations of input parameters and other model factors. The performance of models was assessed by utilising standard statistical measures. The outcomes indicated that the three machine learning models formulated outclassed empirical models and could considerably enhance the precision of monthly Ep estimate even with the same combinations of inputs. In addition, the performance assessment showed that Long Short-Term Memory Neural Network (LSTM) offered the most precise monthly Ep estimations from all the studied models for both stations. The LSTM-10 model performance measures were (R2 = 0.970, MAE = 0.135, MSE = 0.027, RMSE = 0.166, RAE = 0.173, RSE = 0.029) for Alor Setar and (R2 = 0.986, MAE = 0.058, MSE = 0.005, RMSE = 0.074, RAE = 0.120, RSE = 0.013) for Kota Bharu.

Predicting crop yield is an important issue for farmers. Food security is important for decision-makers. The agriculture industry can more accurately supply human demand for food if the crop yield is predicted accurately. Tomato is one of the most important crops so that 160 million tonnes of tomatoes are produced annually around the world. In this study, tomato yield based on data of 40 cities of Iran country including annual average temperature (T), relative humidity (RH), effective rainfall (R), wind speed (WS), and Evapotranspiration (EV) for the period of 1968–2018 was predicted using a new Bayesian model averaging (BMA). The paper's main innovation is the use of the new BMA so that it allows the modellers to quantify the uncertainty of model parameters and inputs simultaneously. For this aim, first, the multiple Adaptive neuro-fuzzy interface system (ANFIS) and multi-layer perceptron (MLP) were used for predicting crop yield. To train the ANFIS and MLP model, a new algorithm, namely, multi verse optimization algorithm (MOA) was used. Also, the ability of MOA was benchmarked against the particle swarm optimization (PSO), and firefly algorithm (FFA). In the next level, the new BMA used the outputs of the ANFIS-MOA, MLP-MOA, ANFIS, FFA, MLP-FFA, ANFIS-PSO, MLP-PSO, ANFIS, and MLP for predicting tomato yield in an ensemble framework. The five- input combination of RH, T, and R, WS, and EV gave the best result. The mean absolute error (MAE) of the BMA in the testing level was 20.12 (Ton/ha) while it was 24.12, 24.45, 24.67, 25.12, 29.12, 30.12, 31.12, and 33.45 for the ANFIS-MOA, MLP-MOA, ANFIS-FFA, MLP-FFA, ANFIS-PSO, MLP-PSO, ANFIS, and MLP models. Regarding the results of uncertainty analysis, the uncertainty of BMA was lower than those of the ANFIS-MOA, MLP-MOA, ANFIS-FFA, MLP-FFA, ANFIS-PSO, MLP-PSO, ANFIS, and MLP models while the MLP model provided the highest uncertainty. The results of this study indicated that BMA using multiple MLP and ANFIS model was useful for predicting tomato yield.

Groundwater levels have been declining recently in Malaysia. This is why, the current study was aimed to propose an accurate groundwater levels prediction model using machine learning algorithms in highly populated towns in Selangor, Malaysia. The models developed used 11 months of previously recorded data of rainfall, temperature and evaporation to predict groundwater levels. Three machine learning models have been tested and evaluated; Xgboost, Artificial Neural Network, and Support Vector Regression. The results showed that for the first scenario, which had combinations of 1,2 and 3 days delayed of rainfall data only considered as an input, the models’ performance was the worst. while in the second scenario the proposed Xgboost model outperformed both the Artificial Neural Network and Support Vector Regression models for all different input combinations. A significant increase in performance was achieved in the third scenario, when using 1 day delayed of groundwater levels as an input as well where R2 equal to 0.92 in the Xgboost model in scenario 3 and 0.16, 0.11 in scenarios 2 and 1 respectively. The results obtained in this study serves as a great benchmark for future groundwater levels prediction using Xgboost algorithm.

The infiltration process during irrigation is an essential variable for better water management and hence there is a need to develop an accurate model to estimate the amount infiltration water during irrigation. However, the fact that the infiltration process is a highly non-linear procedure and hence required special modeling approach to accurately mimic the infiltration procedure. Therefore, the ability of Adaptive Neuro-Fuzzy Interface System (ANFIS) models in estimating infiltrated water during irrigation in the furrow for sustainable management is proposed. The main innovation of current research is the first attempt to employ the ANFIS model for predicating infiltration rates, in addition, integrate the ANFIS model with three new optimization algorithms. Three optimizing algorithms, viz. Sine Cosine Algorithm (SCA), Particle Swarm Optimization (PSO), and Firefly Algorithm (FFA) were used to tune the ANFIS-parameters. Experimental data from six different studies in different countries have been used in this study to validate the proposed model. The inflow rate, furrow length, infiltration opportunity time, cross-sectional area, and waterfront advance time have been utilized as the input parameters. The results indicated that the ANFIS-SCA could provide a better estimation for the infiltration rate compared to ANFIS-PSO. The Mean Absolute Error (MAE) and Percent Bias (PBIAS) errors computed for the ANIFS-SCA (0.007 m3/m and 0.12) was significantly better than those achieved from the ANFIS-FFA and the ANFIS-PSO In addition to that, ANIFS-SCA model outperformed ANFIS-FFA with high level of accuracy. The proposed Hybrid ANFIS-SCA showed outstanding performance over the other optimizer algorithms in estimating the infiltration rate and could be applied in different irrigation systems for better sustainable irrigation management.

Precise reference crop evapotranspiration (ET0) estimation plays a key role in agricultural fields as it aids in the proper operation and management of irrigation scheduling. However, reliable ET0 estimation poses a challenge when there is insufficient or incomplete long-term meteorological data at the East Coast Economic Region (ECER), Malaysia, where the economy is highly dependent on agricultural crop production. This study evaluated the performances of different standalone machine learning (ML) models, namely, the light gradient boosting machine (LGBM), decision forest regression (DFR), and artificial neural network (ANN) models using four different combinations of meteorological variables. The incorporation of solar radiation enhanced the accuracy of the standalone ML models, demonstrating the role of energetic factors in the evapotranspiration mechanism. Additionally, both the ANN and LGBM models showed overall satisfactory performances, and were thus recommended them as alternate models for ET0 estimation. This was owing to their good capability in capturing the non-linearity and interaction process among the meteorological variables. The outcomes of this study will be advantageous to farmers and policymakers in determining the actual crop water demands to maximize crop productivity in data-scarce tropical regions.

This study proposes two techniques: Deep Learning (DL) and Ensemble Deep Learning (EDL) to predict groundwater level (GWL) for five wells in Malaysia. Two scenarios were proposed, scenario-1 (S1): GWL from 4 wells was used as inputs to predict the GWL in the fifth well and scenario-2 (S2): time series with lag time up to 20 days for all five wells. The results from S1 prove that the ensemble EDL generally performs superior to the DL in the estimation of GWL of each station using data of remaining four wells except the Paya Indah Wetland in which the DL method provide better estimates compared to EDL. Regarding S2, the EDL also exhibits superior performance in predicting daily GWL in all five stations compared to the DL model. Implementing EDL decreased the RMSE, NAE and RRMSE by 11.6%, 27.3% and 22.3% and increased the R, Spearman rho and Kendall tau by 0.4%, 1.1% and 3.5%, respectively. Moreover, EDL for S2 shows a high level of precision within less time lag, ranging between 2 and 4 compared to DL. Therefore, the EDL model has the potential in managing the sustainability of groundwater in Malaysia.

Spatial interpolation is important for visualizing changes of parameters over space. Interpolation methods for the spatiotemporal analysis of rainfall pattern changes of Peninsular Malaysia due to climate change were evaluated. Inverse Distance Weighting (IDW) and Ordinary Kriging (OK), Geographical Weighted Regression (GWR) and Multi-scale Geographical Weighted Regression (MGWR) methods were used. Based on the statistic results of RMSE, MAE and R2, the MGWR was the best performing model. To investigate the spatial interpolations with the MGWR, the period considered was arbitrarily and conveniently divided into six 5-year sub-periods for spatiotemporal analysis. Monthly Rainfall, Number of Wet Days and Maximum Daily Rainfall were parameters identified for the analyses. The results showed that the Peninsular Malaysia generally receives relatively higher rainfall amounts and intensities over the northern part of the East Coast region, while lower values were commonly received at the Central region. Changes in spatial pattern were also observed in the maps generated for the onset and withdrawal months of both the Northeast Monsoons and the Southwest Monsoons. The sub-period analyses also showed that Peninsular Malaysia has gradual increase of rainfall intensity due to climate change and is susceptible and vulnerable to El Nino/La Nina events. Hence, in the future Peninsular Malaysia is likely to have increasing occurrence of extreme rainfall events.

Nanotechnologies present a promising application in the production of water to obtain potable water from natural sources by treatment. Clinoptilolite is an abundant and low-cost natural zeolite, which has adsorbent properties. Moreover, this study introduces a simple method for non-harmful natural adsorbent preparation and modification using different acids and bases with convenient and accessible desorption. The characterization methods used in this study are X-ray diffraction (XRD), Fourier-transform infrared spectroscopy (FT-IR), scanning electron microscopy (SEM), and transmission electron microscopy (TEM). Response Surface Method (RSM) has been utilized to reflect the influence of operations variables in Design Expert 7.0. Other Variables, including the Effect of Modifications, contact time, the percentage of the dosage of adsorbent to the beginning (initial) pollutants concentration (D/C), and pH on total iron, were researched during the research procedure. The characterization results indicated that clinoptilolite nanoparticles were well prepared, and the preparation method was effective. Best modification Performance was shown in Z10(0.1 M Hydrochloric Acid dried by Air) at 24 h Setting time. The achieved results from Design Expert showed that at D/C and pH equal to 81.53 and 6.5, respectively, the optimal percentage of the iron removal is obtained with a value equal to 97.97%. In addition, Freundlich isothermal model best described the total iron adsorption process, and it has been observed that the maximum adsorption capacity is 50 mg/g for the total iron removal.

Streamflow prediction help the modelers to manage water resources in watersheds. It gives essential information for flood control and reservoir operation. This study uses the copula-based -Bayesian model averaging (CBMA) as an improved version of the BMA model for predicting streamflow in the Golok River, the Kelantan River, the Lanas River, and the Nenggiri River of Malaysia. The CBMA corrected the assumption of the utilization of Gaussian distortion in the BMA. While the BMA used normal distribution for the variables, the CBMA uses different distribution and copula functions for the variables. This study works on the Archimedes optimization algorithm (AOA) to train the mutlilayer perceptron (MLP) model. The ability of the MLP-AOA model was benchmarked against the MLP-bat algorithm (BA), MLP-particle swarm optimization (MLP-PSO), and the MLP-firefly algorithm (MLP-FFA). The models used significant climate signals, namely, the southern oscillation index (SOI), El NiÑo–Southern Oscillation (ENSO), North Atlantic oscillation (NAO), and the pacific decadal oscillation (PDO) as the inputs to the models. The Gamma test (GT) was coupled with the AOA to provide the hybrid GT for choosing the best inputs. The gamma test was used to determine the suitable lag times of the Nino 3.4, PDO, NAO, and SOI as the inputs. The novelty of the current paper includes introducing new hybrid MLP models, new gamma test for choosing the best input combination, the comprehensive uncertainty analysis of outputs, and the use of an advanced ensemble CBMA model for predicting streamflow. First, the outputs of the MLP-AOA, MLP-BA, MLP-FFA, MLP-PSO, and MLP were obtained, following which, the CBMA as an ensemble framework based on outputs of the hybrid and standalone MLP models was used to predict monthly streamflow. The CBMA at the training level, decreased the root mean square error (RMSE) of BMA, MLP-AOA, MLP-BA, MLP-FFA, MLP-PSO, and MLP models by 28%, 32%, 52%, 53 53%, and 55%, respectively. The CBMA at the training level of another station decreased the mean absolute error (MAE) of BMA, MLP-AOA, MLP-BA, MLP-FFA, MLP-PSO, and MLP models by 6.04, 29%,42%, 49%, 52%, and 52%, respectively. The Nash Sutcliff efficiency (NSE) of the CBMA at the training level was 0.94 while it was 0.92, 0.90, 0.85, 0.84, 0.82, and 0.80 for the BMA, MLP-AOA, MLP-BA, MLP-FFA, MLP-PSO, and MLP models. The RMSE of the MLP-AOA was reported 4.3%, 12%, 14%, and 17% lower than those of the MLP-BA, MLP-FFA, MLP-PSO, and MLP models, respectively. The current research showed the CBMA and the BMA models had high abilities for predicting monthly streamflow. The results of this current study indicated that the CBMA and BMA provided lower uncertainty the standalone MLP models. The general results indicated that the streamflows in the hotter months decreased and flood control is of higher priority during the other months.

Difficulties are faced when formulating hydrological processes, including that of evapotranspiration (ET). Conventional empirical methods for formulating these possess some shortcomings. The artificial intelligence approach emerges as the best possible solution to map the relationships between climatic parameters and ET, even with limited knowledge of the interactions between variables. This review presents the state-of-the-art application of artificial intelligence models in ET estimation, along with different types and sources of data. This paper discovers the most significant climatic parameters for different climate patterns. The characteristics of the basic artificial intelligence models are also explored in this review. To overcome the pitfalls of the individual models, hybrid models which use techniques such as data fusion and ensemble modeling, data decomposition as well as remote sensing-based hybridization, are introduced. In particular, the principles and applications of the hybridization techniques, as well as their combinations with basic models, are explained. The review covers most of the related and excellent papers published from 2011 to 2019 to keep its relevancy in terms of time frame and field of study. Guidelines for the future prospects of ET estimation in research are advocated. It is anticipated that such work could contribute to the development of agriculture-based economy.

A study was undertaken to detect long-term trends in the annual and seasonal series of maximum and minimum temperatures. Measurements were taken at 11 meteorological stations located in the Langat River Basin in Malaysia. The rainfall and maximum and minimum temperature data were obtained from the Malaysia Meteorological Department (MMD) and the Department of Irrigation and Drainage (DID) Malaysia. The procedures used included the Mann-Kendall test, the Mann-Kendall rank statistic test, and the Theil-Sen’s slope method. The analytical results indicated that when there were increasing and decreasing trends in the annual and seasonal precipitation and temperature, only the increasing trends were significant at the 95% confidence level. The Theil-Sen’s slope method showed that the rate of increment in the annual precipitation is greater than the seasonal precipitation. A bootstrap technique was applied to explore uncertainty about significant slope values for rainfall, as well as the maximum and minimum temperatures. The Mann-Kendall rank statistics test indicated that most of the trends in the annual and seasonal time series started in the year 2000. All of the annual and seasonal significant trends were obtained at the stations located in the north, east, and northeast portions of the Langat River Basin.