Rapid acquisition of the spatial distribution of soil nutrients holds great implications for farmland soil productivity safety, food security and agricultural management. To this end, we collected 1297 soil samples and measured the content of soil total nitrogen (TN), soil available phosphorus (AP) and soil available potassium (AK) in Zengcheng, north of the Pearl River Delta, China. Hyperspectral remote sensing images (115 bands) of the Chinese Environmental 1A satellite were used as auxiliary variables and dimensionality reduction was performed using Pearson correlation analysis and principal component analysis. The TN, AP and AK of soil were predicted in the study area based on auxiliary variables after dimensionality reduction, along with stepwise linear regression (SLR), support vector machine (SVM), random forest (RF) and back-propagation neural network (BPNN) models; 324 independent points were used to verify the predictive performance. The BPNN model, which demonstrated the best predictive accuracy among all methods, combined ordinary kriging (OK) with mapping the spatial variations of soil nutrients. Results show that the BPNN model with double hidden layers had better predictive accuracy for soil TN (root mean square error (RMSE) = 0.409 mg kg⁻¹, R² = 44.24%), soil AP (RMSE = 40.808 mg kg⁻¹, R² = 42.91%) and soil AK (RMSE = 67.464 mg kg⁻¹, R² = 48.53%) compared with the SLR, SVM and RF models. The back propagation neural network-ordinary kriging (BPNNOK) model showed the best predictive results of soil TN (RMSE = 0.292 mg kg⁻¹, R² = 68.51%), soil AP (RMSE = 29.62 mg kg⁻¹, R² = 69.30%) and soil AK (RMSE = 49.67 mg kg⁻¹ and R² = 70.55%), indicating the best fitting ability between hyperspectral remote sensing bands and soil nutrients. According to the spatial mapping results of the BPNNOK model, concentrations of soil TN (north-central), soil AP (central and southwest) and soil AK (central and southeast) were respectively higher in the study area. The most important bands (464–517 nm) for soil TN (b10, b14, b20 and b21), soil AP (b3, b19 and b22) and soil AK (b4, b11, b12 and b25) exhibited the best response and sensitivity according to the SLR, SVM, RF and BPNN models. It was concluded that the application of hyperspectral images (visible-near-infrared data) with BPNNOK model was found to be an efficient method for mapping and monitoring soil nutrients at the regional scale.

An accurate estimation of soil organic matter (SOM) content for spatial non-point prediction is an important driving force for the agricultural carbon cycle and sustainable productivity. This study proposed a hybrid geostatistical method of extreme learning machine-ordinary kriging (ELMOK), to predict the spatial variability of the SOM content. To assess the feasibility of ELMOK, a case study was conducted in a regional scale study area in Shaanxi Province, China. A total of 472 topsoil (0–20 cm) samples were collected. A total of 14 auxiliary variables (predictors) were obtained from remote sensing data and environmental factors. The proposed method was compared with the ability of traditional geostatistical methods such as simple kriging (SK) and ordinary kriging (OK), in addition to hybrid geostatistical methods such as regression-ordinary kriging (ROK) and artificial neural network-ordinary kriging (ANNOK). The results showed that the extreme learning machines (ELM) model used principal components (PCs) as input variables, and performed better than both multiple linear regression (MLR) and artificial neural network (ANN) models. Compared with geostatistical and hybrid geostatistical prediction methods of SOM spatial distribution, the ELMOK model had the highest coefficient of determination (R² = 0.671) and ratio of performance to deviation (RPD = 2.05), as well as the lowest root mean square error (RMSE = 1.402 g kg⁻¹). In conclusion, the application of remote sensing imagery and environmental factors has a deeper driven significance of a non-linear and multi-dimensional hierarchy relationship for explaining the spatial variability of SOM, tracing local carbon sink and high quality SOM maps. More importantly, it is possibly concluded that the sustainable monitoring of SOM is a significant process through the pixel-based revisit sampling, an analysis of the mapping results of SOM, and methodological integration, which is the primary step in spatial variations and time series. The proposed ELMOK methodology is a promising and effective approach which can play a vital role in predicting the spatial variability of SOM at a regional scale.

Toxic trace elements in farmland soils are potential threats to human health. In this study, we collected soil samples from the farmlands of southern Guangzhou. We used a sequential indicator simulation (SIS) to deal with the problem of skewed distribution in the sample data. We assessed the human health risks, as well as the uncertainties, of five toxic trace elements: arsenic (As), cadmium (Cd), chromium (Cr), lead (Pb), and mercury (Hg). The results were as follows: (1) The risk indices of two trace elements (Cd and Hg) were less than the standard threshold, which means that there was no human health risk due to Cd and Hg in the study area. However, the maximum risk indices of As, Cr, and Pb exceeded the standard threshold. In particular, the maximum risk index of Pb was twice the standard threshold; (2) The risk probabilities of As and Cr were less than 25% in most areas, and only a few parcels of farmland have a 100% risk probability. The risk map of Pb was used to identify contiguous areas of high-risk probability (i.e., 75%–100%) in the center of the study area. (3) E-type estimation by the SIS method overestimates the risk when the number of samples with concentrations above the threshold have a large proportion of total samples. Our conclusions are as follows: (1) The simulation results show that areas with high-risk indices were concentrated in the Panyu District, which is close to the Pearl River and the core urban area of Guangzhou; (2) Except for Pb, these trace elements are not likely to pose health risks in southern Guangzhou; (3) This study considers the risk probability found with the SIS method to be more reliable for visualizing regional risk.