Soil erosion induced by rainfall is a critical problem in many regions in the world, particularly in tropical areas where the annual rainfall amount often exceeds 2000 mm. Predicting soil erosion is a challenging task, subjecting to variation of soil characteristics, slope, vegetation cover, land management, and weather condition. Conventional models based on the mechanism of soil erosion processes generally provide good results but are time-consuming due to calibration and validation. The goal of this study is to develop a machine learning model based on support vector machine (SVM) for soil erosion prediction. The SVM serves as the main prediction machinery establishing a nonlinear function that maps considered influencing factors to accurate predictions. In addition, in order to improve the accuracy of the model, the history-based adaptive differential evolution with linear population size reduction and population-wide inertia term (L-SHADE-PWI) is employed to find an optimal set of parameters for SVM. Thus, the proposed method, named L-SHADE-PWI-SVM, is an integration of machine learning and metaheuristic optimization. For the purpose of training and testing the method, a dataset consisting of 236 samples of soil erosion in Northwest Vietnam is collected with 10 influencing factors. The training set includes 90% of the original dataset; the rest of the dataset is reserved for assessing the generalization capability of the model. The experimental results indicate that the newly developed L-SHADE-PWI-SVM method is a competitive soil erosion predictor with superior performance statistics. Most importantly, L-SHADE-PWI-SVM can achieve a high classification accuracy rate of 92%, which is much better than that of backpropagation artificial neural network (87%) and radial basis function artificial neural network (78%).

Soil erosion induced by rainfall under prevailing conditions is a prominent problem to farmers in tropical sloping lands of Northeast Vietnam. This study evaluates possibility of predicting erosion status by machine learning models, including fuzzy k-nearest neighbor (FKNN), artificial neural network (ANN), support vector machine (SVM), least squares support vector machine (LSSVM), and relevance vector machine (RVM). Model evaluation employed a historical dataset consisting of ten explanatory variables and soil erosion featured four different land use managements on hillslopes in Northwest Vietnam. All 236 data samples representing soil erosion/nonerosion events were randomly prepared (80% for training and 20% for testing) to assess the robustness of the five models. This subsampling process was repeatedly carried out by 30 rounds to eliminate the issue of randomness in data selection. Classification accuracy rate (CAR) and area under receiver operating characteristic (AUC) were used to evaluate performance of the five models. Significant difference between different algorithms was verified by the Wilcoxon test. Results of the study showed that RVM model achieves the best outcomes in both training (CAR = 92.22% and AUC = 0.98) and testing phases (CAR = 91.94% and AUC = 0.97). Four other learning algorithms also demonstrated good performance as indicated by their CAR values surpassing 80% and AUC values greater than 0.9. Hence, these results strongly confirm the efficacy of applying machine learning models for soil erosion prediction.

Efficient nutrient management is vital to sustaining rice production in the sandy soils of Northeast Thailand, where zinc (Zn) deficiency and low organic matter often constrain yield. This study evaluated the effects of Zn fertilization on the growth, yield, and Zn use efficiency (ZUE) of rice (Oryza sativa L. cv. Chai Nat 1) grown under greenhouse conditions in contrasting soil textures (loamy sand and clay). Four Zn rates were applied: 0, 5, 10, and 15 kg ZnSO₄·7H₂O ha⁻¹ (0, 0.013, 0.026, and 0.039 g ZnSO₄·7H₂O pot⁻¹). Clay soil, with higher organic matter, nitrogen, and available Zn, supported greater vegetative growth, biomass, and yield than loamy sand. Zinc fertilization significantly increased plant height, tiller number, chlorophyll content, biomass, panicle number, grain number, and filled grain weight. Yield improvement in loamy sand was associated mainly with reproductive efficiency, whereas in clay it was driven by vegetative vigor, biomass accumulation, and Zn uptake. Thousand-grain weight was not affected by Zn. ZUE peaked at 5 kg ha⁻¹ in loamy sand and 10 kg ha⁻¹ in clay, with clay showing a greater overall increase in ZUE across Zn rates and loamy sand exhibiting diminishing returns at higher application rates, reflecting differences in Zn availability and retention capacity. Correlation, PCA, and SEM analyses confirmed soil-specific yield mechanisms. Overall, Zn fertilization improved rice productivity and tissue Zn concentration, with optimal rates differing by soil texture. These findings highlight the importance of site-specific Zn management in enhancing yield, nutrient efficiency, and biofortification in rice-based systems of Northeast Thailand.

Nitrogen (N) management is critical for improving rice productivity, but increasing N inputs may reduce N-use productivity when grain production does not increase proportionally. This study evaluated the effects of urea rate and application method on grain weight and partial factor productivity of N (PFPN) of rice (Oryza sativa L. cv. RD49) grown in low-fertility sandy loam soil under controlled greenhouse pot conditions. Three urea rates (0.53, 1.07, and 1.60 g pot⁻¹) and two application methods—basal application (BD) and basal plus topdressing (BTD)—were evaluated across two growing seasons. Grain weight was significantly affected by urea rate, application method, and their interaction. The lowest urea rate (N1; 0.53 g pot⁻¹) maintained grain weight comparable to the highest rate (N3; 1.60 g pot⁻¹) while producing the highest harvest index and substantially greater PFPN. Grain weight exhibited a significant quadratic response to urea rate, whereas PFPN, calculated on the basis of total fertilizer N applied, showed significant linear and quadratic responses and declined markedly with increasing urea rate. BTD resulted in greater mean grain weight and PFPN than BD. These findings indicate that the lowest tested urea rate provided a favorable balance between grain production and fertilizer-N input under the conditions evaluated, while the greater grain weight and PFPN under BTD highlighted the importance of coordinating N rate and application timing. These findings provide a basis for targeted field validation of N rate and application timing for RD49 rice grown in low-fertility, coarse-textured soils.

In recent years, the demand for electric power in Vietnam has increased at annual growth rates of 10% to 12%, and the challenge is to promote renewable energy sector. One of these sustainable energy sources is to harness energy from the wind through wind turbines (WTs). In fact, more wind power plants in Vietnam are continuously to be built due to the rapidly growing demand of country's industrialization and modernization process. However, a significant hindrance preventing the widespread use of WTs in Vietnam is the noise they produce, which significantly contributes to the annoyance experienced by residents living near wind farms. The prediction of noise impacts for new wind farms is one of the many aspects of the environmental impact assessment process in Vietnam as well as in the world. In addition, the determination of the 45 dBA noise contour-line is very important because it is the basis for determining the scope of the project impact according to the IFC/WB performance standards and the number of households to be relocated from the project site. The article’s main focus is therefore on the estimation and simulation of the acoustical noise produced by 18 WTs during the operation phase of Nexif Energy Ben Tre wind power plant and the background noise levels at the project site have been performed by using a combination of specific study methods such as environmental modeling (iNoise Pro modeling software), mapping and geographic information systems. The obtained results show the importance of using modeling method in quantifying the noise levels generated from 18 wind turbines of the Nexif Energy Ben Tre wind power plant met IFC standard and Vietnamese regulation on noise during day-time, but did not meet IFC standard on noise during night-time. The level of background noise measured during night-time in the project area also did not meet IFC standard. Therefore, the overall cumulative noise level exceeds the IFC standard for residential area (45 dBA only). In addition, the appropriate solutions to reduce noise levels from WTs are also proposed.

Cultivar-dependent responses of Jerusalem artichoke (Helianthus tuberosus L.) to combined potassium (K) and zinc (Zn) fertilization remain poorly understood, particularly under low-fertility sandy loam conditions in tropical environments. This study evaluated growth, yield, and tuber-quality responses to K and Zn fertilization in a greenhouse. A 2 × 4 × 3 factorial experiment included two cultivars [CN 52867 × JA6-23 (V1) and JA37 × HEL65 (V2)], four supplemental K application rates (0, 0.61, 1.22, and 1.83 g K pot⁻¹) and three supplemental Zn application rates (0, 0.05, and 0.10 g Zn pot⁻¹). Potassium exerted a more consistent influence on growth and tuber production than Zn, whereas Zn responses varied with cultivar, trait, and nutrient combination. Significant cultivar × K × Zn interactions were detected for tuber number (p < 0.001) and tuber dry weight (p = 0.016). At the highest K rate (1.83 g K pot⁻¹), increasing Zn reduced tuber number in V1 from 38.00 to 19.00 tubers plant⁻¹, whereas in V2 it increased from 17.33 to 26.67 tubers plant⁻¹. Across cultivars and Zn treatments, tuber dry weight increased from 38.53 g plant⁻¹ without supplemental K to 45.91 g plant⁻¹ at 1.83 g K pot⁻¹ and was positively correlated with aboveground biomass (r = 0.77) and stem diameter (r = 0.76). Multivariate analysis further illustrated associations among growth, yield, physiological, and quality traits. From a practical perspective, these results indicate that K and Zn management in Jerusalem artichoke should consider cultivar-dependent nutrient responses rather than assuming a uniform fertilizer response. However, these findings are specific to the tested fertilizer rates and controlled pot conditions and require field validation before cultivar-specific fertilizer recommendations can be established.