Farmland soil environmental quality is important for farmland management. To precisely classify the environmental quality grades of farmland soil, additional samples may be required for multistage sampling or supplementary investigations. Compared with the sampling optimization methods used for mapping or estimating global means, environmental quality grade classifications are primarily focused on estimating the relationships between the values of unsampled locations and the thresholds that classify the environment quality grades. Such classifications must use a sampling layout optimization method to distribute additional sampling units into areas with a high risk of misclassification. To resolve such problems, this paper provides an additional sampling layout optimization method that initially develops a classification error index by building a multi-Gaussian model with the predicted values and error variances of unsampled locations and then calculates the probability of a threshold value occurring in the standardized Gaussian distribution. The average error indexes of all locations in the study area are then set as the objectivity function of the additional sampling layout optimization, and the spatial simulated annealing is adopted to obtain the optimized sampling layout by minimizing the objectivity function. The performance of the error index sampling layout optimization method was demonstrated in a case study using chromium concentration data for Hunan Province, China. The results showed that the additional samples generated by the proposed method produce lower and more stable classification error rates than the minimization of the mean of the shortest distances and spatially random sample methods. The proposed method can be used to improve the efficiency of additional sampling for environmental quality grade classifications of farmland soil.

Vegetation is essential for the provision of ecosystem services. To ensure that ecosystem services are provided, urban planning must consider vegetation cover in both urban core areas and in peripheral areas. This study explored changes in vegetation cover in both urban core and peripheral areas (buffers of 5, 10 and 15 km from the urban core areas) and the relationships between these changes and urban development. To compare the differences in vegetation cover between the urban core and peripheral areas, we developed two vegetation indices, the vegetation influence strength index (VISI) and the vegetation restoration index (VRI), using normalized difference vegetation index datasets and nighttime stable light data from 50 cities in southern China. Using these two indices, we clustered the cities into four types based on their vegetation cover characteristics. The city types varied among the different regions of the study area and cities of the same type tended to occur together geographically. In addition, the two indices showed temporal changes during the study period: A downtrend from 1992 to 2000, an uptrend from 2000 to 2005, and stability from 2005 to 2013. Furthermore, in all cities, the range of the VISI was larger in the peripheral areas than in the urban core areas. However, the range of the VRI was smaller in the peripheral areas than in the urban core areas. As the distance to the urban core areas increased, the fluctuation range of both the VISI and VRI decreased, suggesting that there was less vegetation disturbance farther from the urban core areas. We also showed that changes in vegetation cover were related to differences in urbanization rates, gross domestic products, population densities, and stages of urban development among the cities. These results provide scientific support for urban planning and sustainable development in rapid urbanization areas.

Complex geographical spatial sampling usually encounters various multi-objective optimization problems, for which effective multi-objective optimization algorithms are much needed to help advance the field. To improve the computational efficiency of the multi-objective optimization process, the archived multi-objective simulated annealing (AMOSA)-II method is proposed as an improved parallelized multi-objective optimization method for complex geographical spatial sampling. Based on the AMOSA method, multiple Markov chains are used to extend the traditional single Markov chain; multi-core parallelization technology is employed based on multi-Markov chains. The tabu-archive constraint is designed to avoid repeated searches for optimal solutions. Two cases were investigated: one with six typical traditional test problems, and the other for soil spatial sampling optimization applications. Six performance indices of the two cases were analyzed—computational time, convergence, purity, spacing, min-spacing and displacement. The results revealed that AMOSA-II performed better which was more effective in obtaining preferable optimal solutions compared with AMOSA and NSGA-II. AMOSA-II can be treated as a feasible means to apply in other complex geographical spatial sampling optimizations.

Identifying the current status of the heavy metal(loid) input of agricultural soils is vital for the soil ecological environment of agricultural-producing areas. Most previous studies have typically carried been out in small regions with limited sampling sites, which is insufficient to reveal the overall status of China. This study reviewed publications from over the past 20 years and calculated the input fluxes of heavy metal(loid)s in agricultural soil via atmospheric deposition, fertilizer, manure, and irrigation in different regions of China based on spatiotemporal heterogeneity using a meta-analysis, providing more accurate and reliable results. It was found that the heavy metal(loid) input flux of atmospheric deposition in China is large, while that of fertilizer and manure is relatively low compared to Europe. The major sources of As, Cd, Cr, Ni, and Pb entering the soil was atmospheric deposition, which accounted for 12% to 92% of the total input. Manure was responsible for 19% to 75% of the Cu and Zn input. Cd is the element presenting the most significant risk to the environment of agricultural soils in China and its safety limit will be reached within 100 years for most regions. The region we need to be concerned about is Huang-Huai-Hai due to its comprehensive pollution.

Machine learning-based digital soil mapping often struggles with spatial heterogeneity and long-range dependencies. To address these limitations, this study proposes Multi-Attention Convolutional Neural Networks (MACNN). This deep learning algorithm integrates multiple attention mechanisms to improve mapping accuracy. First, environmental covariates are determined from the soil-landscape model. These are then fed as structured input to the Convolutional Neural Network. Next, by incorporating Transformer self-attention and multi-head attention mechanisms, this study effectively models the long-range dependencies between soil types and features. Concurrently, the Convolutional Block Attention Module (CBAM) is introduced. CBAM features both channel and spatial dual attention, enabling adaptive weighting of crucial feature channels and spatial locations. This significantly enhances the algorithm’s sensitivity to discriminative information. To validate its effectiveness, the proposed MACNN algorithm was used for soil type mapping in Heilongjiang Province. Compared to Random Forest, Decision Tree, and One-Dimensional Convolutional Neural Network algorithms, MACNN demonstrated superior classification performance. It achieved an overall classification accuracy of 81.27%. An ablation study was conducted to investigate the importance of individual modules within the proposed algorithm. The findings indicate that progressively integrating Transformer and CBAM modules into the 1D-CNN baseline significantly enhances algorithm performance through synergistic gains. Therefore, this integrated algorithm offers a feasible solution to improve digital soil mapping accuracy, providing significant reference value for future research and applications.