Willmott Index Research Articles

Process‐Based Modeling of Ecosystem‐Level Monoterpene From a Japanese Larch (Larix kaempferi) Forest

AbstractGlobally, the emission of biogenic volatile organic compounds (BVOC) by plants represents the dominant source of volatile organic compounds emitted to the atmosphere. Monoterpenes, as the major BVOC group, can contribute to forming secondary organic aerosols and influence cloud properties. In this study, we developed a process‐based monoterpene module in the Vegetation Integrative SImulator for Trace gases (VISIT) model by considering the production, storage, and emission of monoterpene as three main processes. We further evaluated the modeled monoterpene emissions against the ecosystem‐level observation data at a half‐hour scale at a Japanese larch (Larix kaempferi) forest site on Mt. Fuji, Japan. The VISIT model performed with relatively higher accuracy with a Willmott's index of agreement at 0.61, a mean bias error (MBE) at 0.29, and a root mean squared error (RMSE) at 0.43, comparable to that of Model of Emissions of Gases and Aerosols from Nature model with a Willmott's index of agreement at 0.63, a MBE at 0.40, and a RMSE at 0.54. In a long‐term simulation under high CO2 emission scenarios, the ratio between monoterpene emission and gross primary production exhibited a stronger correlation with CO2 concentration than temperature. Our study provides a process‐based modeling approach for more accurately simulating monoterpene emissions from Japanese larch.

Enhanced predictive modeling of dissolved oxygen concentrations in riverine systems using novel hybrid temporal pattern attention deep neural networks

Monitoring water quality and river ecosystems is vital for maintaining public health and environmental sustainability. Over the past decade, data-driven methods have been extensively used for river water quality modeling, including dissolved oxygen (DO) concentrations. Despite advancements, challenges persist regarding accuracy, scalability, and adaptability of data-driven models to diverse environmental conditions. Previous studies primarily employed singular models or basic combinations of machine learning techniques, lacking advanced integration of adaptive mechanisms to process complex and evolving datasets. The current study introduces innovative hybrid models that integrate temporal pattern attention (TPA) mechanisms with advanced neural networks, including feed-forward neural networks (FFNNs) and long short-term memory networks (LSTMs). This approach leverages the synergistic strengths of individual models, significantly enhancing the accuracy of DO predictions. The models were rigorously tested against water quality data obtained from two distinct riverine environments, the Illinois River (ILL) and Des Plaines River (DP). Daily measured water quality data, including DO, chlorophyll-a, nitrate plus nitrite, water temperature, specific conductance, and pH, from 2017 to 2024 provided a robust foundation for comprehensive analysis of DO dynamics in these rivers. We conducted 10 scenarios with different model inputs, wherein the hybrid TPACWRNN-LSTM-10 model particularly excelled, achieving coefficient of determination values of 0.993 and 0.965, and root mean squared errors of 0.241 mg/L and 0.450 mg/L for DO predictions at the ILL and DP stations, respectively. The model's reliability was further confirmed by Willmott's index values of 0.998 and 0.992 and Nash-Sutcliffe efficiency values of 0.990 and 0.961 at the ILL and DP stations, respectively. Additionally, Shapley additive explanations (SHAP) values were utilized to interpret each predictor's contribution, revealing key drivers of DO predictions. We believe the novel hybrid modeling approach presented in this study could benefit utilities and water resource management systems for predicting water quality in complex systems.

Calibration and validation of the AquaCrop model for forage cactus production systems under different management interventions in the semi-arid region of Brazil

Simulation models are useful tools for estimating plant response and contributing to the development of management strategies and yield predictions in crops. This study aimed to calibrate and evaluate the AquaCrop-FAO model for the forage cactus imposed on different cropping systems in the semi-arid region of Brazil. Four experiments were conducted: cactus-sorghum intercropping system with five spacings between plants (0.10, 0.20, 0.30, 0.40 and 0.50 m); cactus-sorghum intercropping system with five spacings between rows (1.00, 1.25, 1.50 and 1.75 m); single cactus with four levels of mulch (0, 5, 10 and 15 Mg ha−1); and single forage cactus with four levels of nitrogen fertilisation (50, 150, 300 and 450 kg of N ha−1). The performance of the model was evaluated using the following parameters: root mean squared error (RMSE); normalised RMSE (NRMSE), coefficients of determination (R2), Nash-Sutcliffe model efficiency coefficient (ME) and the Willmott index of agreement (d). During the calibration stage, the statistical indices pointed to the good performance of the AquaCrop model in estimating dry biomass in most of the cropping systems, with 10 < NRMSE <20; R2 > 0.96 and ME > 0.95. The AquaCrop model showed the best performance simulating systems under different levels of nitrogen fertilisation. Among the intercropping systems, the denser arrangements showed the highest values for simulated biomass.

Large Eddy Simulation of Cross‐Shore Hydrodynamics Under Random Waves in the Inner Surf and Swash Zones

AbstractA 3D large eddy simulation coupled with a free surface tracking scheme was used to simulate cross‐shore hydrodynamics as observed in a large wave flume experiment. The primary objective was to enhance the understanding of wave‐backwash interactions and the implications for observed morphodynamics. Two simulation cases were carried out to elucidate key processes of wave‐backwash interactions across two distinct stages: berm erosion and sandbar formation, during the early portion of a modeled storm. The major difference between the two cases was the bathymetry: one featuring a berm without a sandbar (Case I), and the other, featuring a sandbar without a berm (Case II) at similar water depth. Good agreement (overall Willmott's index of agreement greater than 0.8) between simulations and measured data in free surface elevation, wave spectrum, and flow velocities validated the model skill. The findings indicated that the bottom shear stress, represented by the Shields parameter, was significant in both cases, potentially contributing substantial sediment transport. Notably, the occurrence of intense wave‐backwash interactions were more frequent in the absence of a sandbar. These intense wave‐backwash interactions resulted in a pronounced horizontal pressure gradient, quantified by high Sleath parameters, exceeding the criteria for momentary bed failure. Additionally, a more vigorous turbulence‐bed interaction, characterized by near‐bed turbulent kinetic energy, was observed in the case lacking a sandbar, potentially augmenting sediment suspension. These insights are pivotal in understanding the mechanisms underlying berm erosion and how sandbar formation serves to protect further beach erosion.

Stabilized oily-wastewater separation based on superhydrophilic and underwater superoleophobic ceramic membranes: Integrated experimental design and standalone machine learning algorithms

BackgroundReliable computational approaches to evaluate ceramic membrane performance in wastewater treatment mark a transformative step towards optimizing separation processes, ensuring environmental sustainability, and advancing water purification technologies. The current study explores the influential factors using artificial intelligence (AI) tools in the performance evaluation of superhydrophilic and underwater super-oleophobic ceramic membranes for the selective treatment of oily wastewater. MethodsThe chemometrics scenario of the research based on established experimental work employs advanced AI models viz: Random Forest (RF), Support Vector Regression (SVR), and Gaussian Process Regression (GPR) to predict the efficacy of these membranes in terms of rejection and flux. The model predictions were evaluated using the Pearson Correlation Coefficient (PCC), Willmott Index (WI), mean absolute percentage error (MAPE), and mean absolute error (MAE). Significant findingsFrom the results, GPR had shown good agreement with correlations (WI=99.9) during the training and testing phases for flux prediction, indicating an exceptional model fit with negligible error (MAPE=0.001, MAE=0.000 in the testing phase). For rejection modelling, GPR and SVR exhibit similar levels of accuracy, with moderate PCC and WI values, while RF reveals significant limitations with the lowest scores across all statistical metrics. The findings highlight the potential of AI in optimizing wastewater treatment processes, with GPR identified as the most promising model for flux prediction. This study would provide insight into the modelling of the membrane separation process for oily wastewater and integrate AI in the performance evaluation of wastewater reclamation.

Research on methods for estimating reference crop evapotranspiration under incomplete meteorological indicators.

Accurate estimation of reference crop evapotranspiration (ET0) is crucial for farmland hydrology, crop water requirements, and precision irrigation decisions. The Penman-Monteith (PM) model has high accuracy in estimating ET0, but it requires many uncommon meteorological data inputs. Therefore, an ideal method is needed that minimizes the number of input data variables without compromising estimation accuracy. This study aims to analyze the performance of various methods for estimating ET0 in the absence of some meteorological indicators. The Penman-Monteith (PM) model, known for its high accuracy in ET0 estimation, served as the standard value under conditions of adequate meteorological indicators. Comparative analyses were conducted for the Priestley-Taylor (PT), Hargreaves (H-A), McCloud (M-C), and FAO-24 Radiation (F-R) models. The Bayesian estimation method was used to improve the ET estimation model. Results indicate that, compared to the PM model, the F-R model performed best with inadequate meteorological indicators. It demonstrates higher average correlation coefficients (R2) at daily, monthly, and 10-day scales: 0.841, 0.937, and 0.914, respectively. The corresponding root mean square errors (RMSE) are 1.745, 1.329, and 1.423, and mean absolute errors (MAE) are 1.340, 1.159, and 1.196, with Willmott's Index (WI) values of 0.843, 0.862, and 0.859. Following Bayesian correction, R2 values remained unchanged, but significant reductions in RMSE were observed, with average reductions of 15.81%, 29.51%, and 24.66% at daily, monthly, and 10-day scales, respectively. Likewise, MAE decreased significantly, with average reductions of 19.04%, 34.47%, and 28.52%, respectively, and WI showed improvement, with average increases of 5.49%, 8.48%, and 10.78%, respectively. Therefore, the F-R model, enhanced by the Bayesian estimation method, significantly enhances the estimation accuracy of ET0 in the absence of some meteorological indicators.

Prediction of drought hydrological and water scarcity based on optimal artificial intelligence by developing a metaheuristic optimization algorithm

This study presents a novel approach to predicting water stress in Yichang City, China, through an optimized Capsule Neural Network (CapsNN) model enhanced by the Improved Aquila Optimization (IAO) algorithm. The model, applied to monthly water stress data from 2000 to 2022, is gauged using the Streamflow Drought Index (SDI). When compared to traditional metaheuristic models, our CapsNN/IAO framework demonstrates superior accuracy in drought forecasting across various time scales with significant reductions in The Mean Absolute Error is 0.1021, the Root Mean Square Error is 0.1824, and good the Nash-Sutcliffe Efficiency is 0.75, the coefficient of determination R2 is 0.89, and the Willmott's Index is 0.93. Notably, the model predicts short-term water shortages with greater frequency at the SDI3 scale, while identifying the most severe drought conditions at the SDI6 scale. These insights are critical for enhancing water resource management.

Introducing high-order response surface method for improving scour depth prediction downstream of weirs

Scour depth downstream of weirs is considered one of the most important hydraulic problems, which greatly influences the stability of weirs. Recently, artificial intelligence (AI) methods have become increasingly popular in modeling hydraulic variables, especially scour depth, because they can capture nonlinear relationships between input variables and their associated objectives. Despite their importance, these models have problems with hyperparameter tuning in scour depth modeling due to their structures, so algorithms must be used to tune the hyperparameters. Moreover, these algorithms are usually tuned by using the trial-and-error method to select the hyperparameters such as the number of hidden nodes, transfer function, and learning rate, and in this case, the main problem is overfitting during the training phase. To solve these problems, the high-order response surface method (HORSM), an improved version of the response surface method (RSM), is used as an alternative approach for the first time in this study to predict the scour depth. The HORSM model is based on high-order polynomial functions (from two to six) compared with the artificial neural network model (ANN). The findings indicate that the fifth order of the HORSM polynomial function yields the most precise predictions, with a higher coefficient of determination (R2) of 0.912 and Willmott Index (WI) of 0.972 compared to the values obtained using ANN (R2 = 0.886 and WI = 0.927). Moreover, the accuracy of the predictions is represented by a reduction of the mean square error by up to 44.17 and 29.01% compared to the classical RSM and ANN, respectively. The suggested model established an excellent correlation and accuracy with experimental values.

Establishment of critical non-depositing velocity prediction model for sediment in drip irrigation laterals based on PSO-SVM

Accurately determining the critical non-depositing velocity for sediment (CNDVS) in drip irrigation laterals is essential to address issues of sediment deposition and clogging in drip irrigation pipes caused by the use of muddy water. This paper focuses on three main factors: the percentage of intermediate-sized sediment particles (P), pipe diameter (D), and sediment concentration (S), all of which significantly influence the CNDVS in drip irrigation laterals. To predict the CNDVS, a comprehensive gradient test was developed and the Particle Swarm Optimization-Support Vector Machine (PSO-SVM) algorithm was applied to establish a response surface for predicting the CNDVS. Additionally, the PSO algorithm optimized the penalty parameter (C), loss parameter (ε), and kernel parameter (δ) in the SVM, resulting in a prediction model for the CNDVS. The proposed model was trained and tested using experimental data, and its prediction performance and accuracy were compared with the regression fitting model. The results indicate that within a specific range, the CNDVS increases with higher values of the three factors. The PSO-SVM model, with optimal parameters combination values of C = 486.37, ε = 0.04, and δ = 68.49, provided the most prediction accurate. Compared with the regression fitting model, the PSO-SVM model has the smaller values of Mean Absolute Error (MAE) = 0.0049, Root Mean Square Error (RMSE) = 0.0057, and Percent Bias (PBIAS) = 0.0004, and the larger value of Nash–Sutcliffe Efficiency (NSE) = 0.9874, Willmott index (WI) = 0.9401, and Coefficient of Determination (R2) = 0.9875. These results demonstrate that the PSO-SVM model established in this study exhibits high prediction performance and accuracy. The findings of this research can serve as a foundation for developing optimal operating and flushing flow velocity within the laterals in muddy water drip irrigation systems.

Migration time prediction and assessment of toxic fumes under forced ventilation in underground mines

This study aims to predict the migration time of toxic fumes induced by excavation blasting in underground mines. To reduce numerical simulation time and optimize ventilation design, several back propagation neural network (BPNN) models optimized by honey badger algorithm (HBA) with four chaos mapping (CM) functions (i.e., Chebyshev (Che) map, Circle (Cir) map, Logistic (Log) map, and Piecewise (Pie) map) are developed to predict the migration time. 125 simulations by the computational fluid dynamics (CFD) method are used to train and test the developed models. The determination coefficient (R2), the variance accounted for (VAF), the Willmott’s index (WI), the root mean square error (RMSE), the mean absolute percentage error (MAPE), and the sum of squares error (SSE) are utilized to evaluate the model performance. The evaluation results indicate that the CirHBA-BPNN model has achieved the most satisfactory performance by reaching the highest values of R2 (0.9945), WI (0.9986), VAF (99.4811%), and the lowest values of RMSE (15.7600), MAPE (0.0343) and SSE (6209.4), respectively. The wind velocity in roadway (Wv) is the most important feature for predicting the migration time of toxic fumes. Furthermore, the intrinsic response characteristic of the optimal model is implemented to enhance the model interpretability and provide reference for the relationship between features and migration time of toxic fumes in ventilation design.

Comparative study of different machine learning approaches for predicting the compressive strength of palm fuel ash concrete

Nowadays, the challenge of industrialising concrete is to produce economical, ecological and efficient concrete. As a result, cementitious additions are increasingly used in concrete binders with proportions conferring the desired properties, particularly compressive strength. Palm oil fuel ash (POFA) from palm trees is one of these additions and introduced by substitution of cement in concrete. Although laboratory tests applied to concrete samples are expensive and time-consuming, considerable research is directed towards modelling approaches and machine learning (ML) applications. The purpose of this study is to explore different ML techniques, including artificial neural network (ANN), ANN with combined inputs (ANNX), particle swarm optimisation (PSO) and genetic algorithm (GA). These techniques are used to predict the compressive strength of POFA concrete derived from a dataset consisting of 249 samples with regard to 6 input parameters, namely, cement and POFA contents, aggregate ratio, water-to-binder ratio, superplasticiser dosage and curing duration. The performance of these techniques is evaluated using six performance indices: correlation coefficient, Willmott index of agreement, Nash–Sutcliffe efficiency, root mean squared error, mean absolute error and mean squared relative error. The comparison shows that ANN and ANNX have better predictive accuracy than PSO and GA, and ANNX performs relatively better than ANN. Results are presented using a 3D surface response plot which visually depicts the relationship between predicted outcomes and influential parameters. The comparison made between ANNX and other models in the literature demonstrated that the present model is reliable to predict CS of POFA concrete with high precision in a wide range of data.

Predictive modelling of nitrogen dioxide using soft computing techniques in the Agra, Uttar Pradesh, India

Nitrogen dioxide (NO2) is one of the air pollutants which aggravates the human health as well as causes environmental issues. It is more causes respiratory problems due to acid rains. Agra is a major tourist destination spot in India also similarly air pollution also increased growing urbanization and traffic reflux. The current study aims to predicted the NO2 episodes in the Agra city using soft computing models namely, M5P, Random Forest (RF), Group method of data handling (GMDH), Multivariate adaptive regression (MARS), Reduced error pruning tree (REP Tree) and Random tree (RT). The models were generated using 1116 observations, from 2015 to 2020 with input parameters such as Particulate matter (PM2.5), Nitrogen monoxide (NO), Oxides of nitrogen (NOX), Sulphur dioxide (SO2), Carbon monoxide (CO), Ozone (O), Benzene (Be), Toluene, Relative humidity (RH), Wind speed (WS), Wind direction (WD), Solar radiation (SR), Barometric pressure (BP) and Xylene. The performance of each model was evaluated based on the six statistical indices, namely correlation coefficient (CC), root mean square error (RMSE), mean absolute error (MAE), normalized root of mean squared relative error (NRMSE), Willmott's Index (WI), and Legates and McCabe's Index (LMI). The performance evaluation models results of study area, Box plot and Taylor's diagram indicated that M5P is the outperforming model among others with testing CC = 0.9543, RMSE = 5.8006, MAE = 3,9204, NRMSE = 0.1512, WI = 0.9744, and LMI = 0.7549. Based on the sensitivity analysis indicated that NOx is the most influential parameter followed by WD and CO. These results of study area can be helpful to understanding the air pollution causes, health issues, and future NO2 levels around the study area with useful results for air pollution monitoring policy and development.

Neural network-based prediction of auto-ignition temperature of ternary mixed liquids

Auto-ignition temperature (AIT) is one of the crucial exponents in the design of fire and explosion safety measures. Therefore, in this study, quantitative structure-property relationship approach was used to predict the AIT of ternary hybrid liquids based on molecular structure information. The optimal molecular descriptors were calculated and filtered using Mordred software. Twelve mixing rules were proposed for calculating molecular descriptors of mixtures. A prediction model for the AIT value of binary liquid mixtures was developed, validated and evaluated using a back propagation neural network (BPNN) and a one-dimensional convolutional neural network (1DCNN). The relative contribution and positive and negative correlations between individual molecular descriptors and AIT in the model were interpreted using the shapley additive explanations method. The results show that BPNN and 1DCNN models using mixing rule 1 have the best fitting ability, stability and prediction ability. The determination coefficient of the BPNN and 1DCNN models in the training set were 0.996 and 0.992, the root mean square errors were 3.613 °C and 5.284 °C, the mean absolute errors were 2.483 °C and 4.144 °C, the nash efficiency coefficient was 0.996 and 0.992, respectively, the willmott index was 0.999 and 0.998. and the values of the top three molecular descriptors of relative contribution, SssCH2, SsOH and SsCH3, were negatively correlated with the AIT values. The BPNN and 1DCNN models provide an accurate and reliable method for predicting ternary mixing liquid AIT.

Supervised intelligent prediction of shear strength of rockfill materials based on data driven and a case study

The rockfill materials (RFM) are emerging and regarded waste reuse product in construction and mining engineering. In this paper, six distinctive supervised machine learning (SML) models, namely artificial neural network (ANN), extreme learning machine (ELM), random forest (RF), relevance vector regression (RVR), support vector regression (SVR), and extreme gradient boosting (XGBoost), are adopted to predict the RFM shear strength using 165 data cases with 13 features. To improve model performance, an improved physics meta-heuristic algorithm, named snow ablation optimizer with Logistic mapping (LogSAO), is utilized to select the optimal hyperparameter combination of the proposed models. Four popular statistical indices, including coefficient of determination (R2), root mean squared error (RMSE), Willmott’s index (WI), and Scatter index (SI) are employed to quantify the model performance. The model evaluation results show that the LogSAO-SVR model is the best prediction model with the highest values of R2 and WI while the lowest values of RMSE and SI in both training and testing phases (0.9985, 0.0257, 0.9996, and 0.0380; 0.9802, 0.0812, 0.9950, and 0.1335). Normal stress (Ns) is the most important feature with the highest importance score of 0.925 in predicting the RFM shear strength. Interestingly, the negative contribution of Ns to the RFM shear strength prediction will be transformed into the positive contribution when coefficients of curvature (Cu) values are small. The validation results demonstrate that the predicted values of the proposed LogSAO-SVR model are highly consistent with the measured values from the direct shear tests based on a case study.

Suspended sediment load modeling using Hydro-Climate variables and Machine learning

Machine learning (ML) models have the potential to improve the accuracy of Suspended Sediment Load (SSL) predictions. However, their ability to incorporate climate data to enhance SSL predictions remains underexplored. This study investigates the efficacy of ML models in modeling sediment transport using hydro-climate variables as input data. Boosted decision tree models, including Xgboost, LightGBM, and Catboost, were employed to estimate SSL using hydro-climate variables such as streamflow and precipitation, as well as maximum and minimum temperatures. The models were trained, optimized, and evaluated using either climate/streamflow data or a combination of both, encompassing 47 gauges across the contiguous United States (CONUS) between 1950 and 2022. The results demonstrate that streamflow data alone is insufficient for accurate SSL estimation in all rivers. Consequently, integrating climate data and streamflow data proved effective in predicting SSL at the daily scale across diverse climate zones with commendable accuracy. The models achieved an average Nash-Sutcliffe Efficiency (NSE) of > 0.66, Kling-Gupta Efficiency (KGE) of > 0.67, and Willmott Index (WI) of > 0.83 during the unseen (test) phase. These findings highlight the potential of ML models, particularly when combining climate data with streamflow, to enhance SSL predictions and improve our understanding of sediment transport dynamics.

Estimating critical depth and discharge over sloping rough end depth using machine learning

Abstract This study uses machine learning (ML) to predict the end-depth structure's discharge and critical depth (yc). Linear regression, M5P, random forest, random tree, reduced error pruning tree, and Gaussian process (GP) are the ML methods used in this investigation. The findings indicate that the radial kernel function-based GP model is most suitable compared to other applied models with the lowest root-mean-square error = 0.0021, 0.007, normalized root-mean-square error = 0.0361, 0.0516 representing mean absolute error = 0.0015, 0.004 and the highest coefficient of correlation = 0.9912, 0.9916, Legates and McCabe's index = 0.8839, 0.9026 Willmott's index = 0.9956, 0.9956, and Nash Sutcliffe model efficiency = 0.9823, 09830 for yc for the end-depth structure (yc) and discharge (Q) with the testing stage, respectively. Results of the sensitivity study indicate that the friction coefficient is the most significant input variable compared to other parameters for predicting (yc) and flow running via the thickness model's last stage (Q) using this dataset.

Advanced Farming Strategies Using NASA POWER Data in Peanut-Producing Regions without Surface Meteorological Stations

Understanding the impact of climate on peanut growth is crucial, given the importance of temperature in peanut to accumulate Growing Degree Days (GDD). Therefore, our study aimed to compare data sourced from the NASA POWER platform with information from surface weather stations to identify underlying climate variables associated with peanut maturity (PMI). Second, we sought to devise alternative methods for calculating GDD in peanut fields without nearby weather stations. We utilized four peanut production fields in the state of Georgia, USA, using the cultivar Georgia-06G. Weather data from surface stations located near peanut fields were obtained from the University of Georgia’s weather stations. Corresponding data from the NASA POWER platform were downloaded by inputting the geographic coordinates of the weather stations. The climate variables included maximum and minimum temperatures, average temperature, solar radiation, surface pressure, relative humidity, and wind speed. We evaluated the platforms using Pearson correlation (r) analysis (p &lt; 0.05), linear regression analysis, assessing coefficient of determination (R2), root mean square error (RMSE), and Willmott index (d), as well as principal component analysis. Among the climate variables, maximum and minimum temperatures, average temperature, and solar radiation showed the highest R2 values, along with low RMSE values. Conversely, wind speed and relative humidity exhibited lower correlation values with errors higher than those of the other variables. The grid size from the NASA POWER platform contributed to low model adjustments since the grid’s extension is kilometric and can overlap areas. Despite this limitation, NASA POWER proves to be a potential tool for PMI monitoring. It should be especially helpful for growers who do not have surface weather stations near their farms.

Proportional impact prediction model of coating material on nitrate leaching of slow-release Urea Super Granules (USG) using machine learning and RSM technique

An accurate assessment of nitrate leaching is important for efficient fertiliser utilisation and groundwater pollution reduction. However, past studies could not efficiently model nitrate leaching due to utilisation of conventional algorithms. To address the issue, the current research employed advanced machine learning algorithms, viz., Support Vector Machine, Artificial Neural Network, Random Forest, M5 Tree (M5P), Reduced Error Pruning Tree (REPTree) and Response Surface Methodology (RSM) to predict and optimize nitrate leaching. In this study, Urea Super Granules (USG) with three different coatings were used for the experiment in the soil columns, containing 1 kg soil with fertiliser placed in between. Statistical parameters, namely correlation coefficient, Mean Absolute Error, Willmott index, Root Mean Square Error and Nash–Sutcliffe efficiency were used to evaluate the performance of the ML techniques. In addition, a comparison was made in the test set among the machine learning models in which, RSM outperformed the rest of the models irrespective of coating type. Neem oil/ Acacia oil(ml): clay/sulfer (g): age (days) for minimum nitrate leaching was found to be 2.61: 1.67: 2.4 for coating of USG with bentonite clay and neem oil without heating, 2.18: 2: 1 for bentonite clay and neem oil with heating and 1.69: 1.64: 2.18 for coating USG with sulfer and acacia oil. The research would provide guidelines to researchers and policymakers to select the appropriate tool for precise prediction of nitrate leaching, which would optimise the yield and the benefit–cost ratio.

Forest yield prediction under different climate change scenarios using data intelligent models in Pakistan.

This study aimed to develop and evaluate data driven models for prediction of forest yield under different climate change scenarios in the Gallies forest division of district Abbottabad, Pakistan. The Random Forest (RF) and Kernel Ridge Regression (KRR) models were developed and evaluated using yield data of two species (Blue pine and Silver fir) as an objective variable and climate data (temperature, humidity, rainfall and wind speed) as predictive variables. Prediction accuracy of both the models were assessed by means of root mean squared error (RMSE), mean absolute error (MAE), correlation coefficient (r), relative root mean squared error (RRMSE), Legates-McCabe's (LM), Willmott's index (WI) and Nash-Sutcliffe (NSE) metrics. Overall, the RF model outperformed the KRR model due to its higher accuracy in forecasting of forest yield. The study strongly recommends that RF model should be applied in other regions of the country for prediction of forest growth and yield, which may help in the management and future planning of forest productivity in Pakistan.

Evaluation of Infiltration Modeling in the Cisadane Watershed in Indonesia: Existing and New Approach Equation

This study aims to evaluate the accuracy of semi-empirical models for determining infiltration based on the existing equations (Horton, Kostiakov and Kostiakov–Lewis) and a new form equation with optimized parameters which has good accuracy (lowest RMSE and MBE), sufficient linearity (highest R2), and a Willmott index close to 1. The location of this study was the Cisadane watershed, which is a priority watershed in Indonesia. The lowest values of average RMSE and MBE and highest values of R2 and Willmott index are given by Equation 1, which indicated that the model was closest to the field infiltration conditions. The distribution of the RMSE values for all survey points using Equation 1 resulted in small RMSE values ranging from 0.09 to 1.83. The distribution of R2 values for all survey points using Equation 1 yielded R2 values close to 1 (high linearity) with values ranging from 0.96 to 1. Based on these results, it can be concluded that Equation 1 is the most accurate infiltration model for the Cisadane watershed compared to the existing infiltration rate equation. This research can be used as an initial idea to create the infiltration rate model using other forms of equations with optimization parameters to produce a more representative model with field data.