The peel of pomegranate (Punica granatum) is rich in bioactive compounds, specifically phenolic compounds and tannin compounds. However, there is still a lot of difficulty dealing with the extraction of these substances due to the limitations of traditional methods. Microwave-assisted extraction (MAE) has shown promise, but optimizing it for maximum efficiency and yield remains a challenge. In this work, a microwave-assisted extraction improved using machine learning approaches was used to extract tannins and phenolic compounds from pomegranate peel. The experimental design consisted of four independent variables: microwave power (100–300 W), extraction time (10–40 min), temperature (35–50 °C), and food-to-solvent ratio (0.25–0.5 g/10 mL). The evaluated response variables were total phenolic (mg GAE/g), total tannin (mg CE/g), and antioxidant activity (DPPH scavenging activity). Thirty experiments were conducted using the microwave extraction system. Two machine learning models, LSBoost with Random Forest (LSBoost/RF) and LSBoost with K-Nearest Neighbors Neural Network (LSBoost/KNN-NN), were developed and compared for predicting extraction outcomes. The LSBoost/RF model demonstrated superior performance, achieving correlation coefficients (R²) of 0.9998, 0.9018, and 0.9269 for total phenolic, total tannin, and DPPH %, respectively. Feature importance analysis revealed microwave power as the most influential parameter, particularly for tannin content and antioxidant potency. The findings indicate that the combination of microwave-assisted extraction with machine learning provides an effective and accurate approach for the extraction and prediction of phenolic and tannin compounds in natural sources.

This study investigates the presence and adsorption of heavy metals (HMs) on microplastics (MPs) in the drinking water and water resources of Zabol, Iran. Sampling was conducted at five stations in Zabol and two Chah-Nimeh reservoirs (CHWRs), the primary drinking water sources, using glass samplers and polyvinylidene fluoride membrane filters. Fourier-transform infrared spectroscopy, scanning electron microscopy, and optical microscopy identified MPs, while inductively coupled plasma optical emission spectrometry quantified heavy metals. Results revealed that polyethylene, polystyrene, polyamide, and polypropylene were the predominant polymers, with Fe, Mn, Zn, Cu, and Ni adsorbed on their surfaces. Iron (Fe) exhibited the highest concentration, reaching 84 µg.L⁻¹ in CHWRs and 85 µg.L⁻¹ in the distribution network. In CHWRs the highest percentage (61%) of size of MPs were between 500 and 1000 μm, while in water distribution network maximum range of obtained MPs were between 50 and 100 μm. The average concentration of the other detected heavy metals were as: zinc (Zn) 43.9 µg.L^− 1 and 7.75 µg.L^− 1, manganese (Mn) 11.48 µg.L^− 1 and 6.5 µg.L^− 1, arsenic (As) 20.22 µg.L^− 1 and 0.26 µg.L^− 1, cadmium (Cd) 3.16 µg.L^− 1 and 2.8 µg.L^− 1, copper (Cu) 15.18 µg.L^− 1 and 0.03 µg.L^− 1, and nickel (Ni) 0.10 µg.L^− 1 and 2.25 µg.L^− 1 in distribution network and CHWRs, respectively. World Health Organization recommended the permissible limit of these cations in drinking water as 10 µg.L^− 1 for As, 5 µg.L^− 1 for Cd, 50 µg.L^− 1 for Cr, 20 µg.L^− 1 for Ni, 10 µg.L^− 1 for Mn, 500 µg.L^− 1 for Zn, 300 µg.L^− 1, 2000 µg.L^− 1 for Cu and 300 µg.L^− 1 for Fe. The study highlights MPs as carriers of toxic heavy metals, presenting significant health and environmental risks. This novel research emphasizes the impact of secondary pollution and water treatment processes on MP fragmentation and HM contamination. Recommendations include adopting enhanced water treatment protocols to mitigate MP and HM risks, implementing stricter quality monitoring at all stages of water distribution, and promoting public awareness of plastic pollution. Future studies should explore the health effects of MPs and HMs, optimize sampling methods, and focus on long-term monitoring under diverse environmental conditions to address this emerging issue comprehensively.

Reliable estimations of sediment yields are very important for investigations of river morphology and water resources management. Nowadays, soft computing methods are very helpful and famous regarding the accurate estimation of sediment loads. The present study checked the applicability of the radial M5 tree (RM5Tree) model to accurately estimate sediment yields using daily inputs of the snow cover fraction, air temperature, evapotranspiration and effective rainfall, in addition to the flow, in the Gilgit River, Upper Indus Basin (UIB) tributary, Pakistan. The results of the RM5Tree model were compared with support vector regression (SVR), artificial neural network (ANN), multivariate adaptive regression spline (MARS), M5Tree, sediment rating curve (SRC) and response surface method (RSM) models. The resulting accuracy of the models was assessed using Pearson’s correlation coefficient (R²), the root-mean-square error (RMSE) and the mean absolute percentage error (MAPE). The prediction accuracy of the RM5Tree model during the testing period was superior to the ANN, MARS, SVR, M5Tree, RSM and SRC models with the R², RMSE and MAPE being 0.72, 0.51 tons/day and 11.99%, respectively. The RM5Tree model predicted suspended sediment peaks better, with 84.10% relative accuracy, in comparison to the MARS, ANN, SVR, M5Tree, RSM and SRC models, with 80.62, 77.86, 81.90, 80.20, 74.58 and 62.49% relative accuracies, respectively.

Urban thermal comfort analysis in hot-arid climates requires robust methodological frameworks to address intensifying heat stress. This study introduces a reliability-based thermal comfort framework using Monte Carlo simulation to evaluate six urban design interventions in Zahedan, Iran. Field measurements encompassed reference conditions, concrete gazebos with varying vegetation covers, and tree shade configurations. Monte Carlo simulation (n = 10,000) computed prediction uncertainties and probabilistic thermal comfort evaluations for Physiological Equivalent Temperature (PET) and Universal Thermal Climate Index (UTCI). A novel Thermal Comfort Reliability Index (TCRI) was developed integrating comfort probability and system reliability. The approach demonstrated excellent convergence (CV < 1%) and predictive capability (R² > 0.999, RMSE < 1.0 °C). Results revealed critical deficiencies across all interventions. The best-performing configuration (concrete gazebo with full tree shade) achieved only 20.56% comfortable conditions. Hot failure probabilities ranged from 29.85% to 65.25%. UTCI demonstrated superior reliability versus PET, with 3.14 × higher performance and improvement factors of 1.7 × to 58.2 × . Sensitivity analysis identified mean radiant temperature (r = 0.942) and air temperature (r = 0.896) as dominant variables. Extremely low TCRI values (< 5%) indicate systematic failures requiring radical climate adaptation measures in hot-arid urban areas.

In this study, the viability of radial M5 model tree (RM5Tree) is investigated in prediction and estimation of daily streamflow in a cold climate. The RM5Tree model is compared with the M5 model tree (M5Tree), artificial neural networks (ANN), radial basis function neural networks (RBFNN), and multivariate adaptive regression spline (MARS) using data of two stations from Sweden. The accuracy of the methods is assessed based on root mean square errors (RMSE), mean absolute errors (MAE), mean absolute percentage errors (MAPE), and Nash Sutcliffe Efficiency (NSE) and the methods are graphically compared using time variation and scatter graphs. The benchmark results show that the RM5Tree offers better accuracy in predicting daily streamflow compared to other four models by respectively improving the accuracy of M5Tree with respect to RMSE, MAE, MAPE, and NSE by 26.5, 17.9, 5.9, and 10.9%. The RM5Tree also acts better than the M5Tree, ANN, RBFNN, and MARS in estimating streamflow of downstream station using only upstream data.