Soil salt content (SSC) is normally featured with obvious spatiotemporal variations in arid and semi-arid regions. Space factors such as elevation, temperature, and spatial locations are usually used as input variables for a model to estimate the SSC. However, whether temporal patterns of salt-affected soils (identified as temporal spectral patterns) can indicate the SSC level and be applied as a covariate in a model to estimate the SSC remains unclear. Hence, temporal changes in soil spectral patterns need to be characterized and explored as to their use as an input variable to improve SSC estimates. In this study, a total of 54 field samples and a time-series of Sentinel-2 multispectral images taken at monthly intervals (from October 2017 to April 2018) were collected in the Yinbei area of western China. Then, two-date satellite images were used to quantify significant spectral changes over time using spectral change vector analysis, and four two-date-based index methods were used to characterize soil spectral changes. Lastly, the optimal two-date-based spectral indices and multispectral bands were used as input variables to build the estimation models using a random forest algorithm. Results showed that the two-date-based spectral index could be applied as an input variable to improve the accuracy of SSC estimation at a regional scale. Temporal changes in salt-induced spectral patterns can be indicated by the band difference in the wavelength range from 400 nm to 900 nm. Three two-date-based indices designated as D_28a (i.e., the band difference between band 2 from an image acquired in April 2018 and band 8a from an image acquired in December 2017), D₂₂, and D₂₈ were the optimal parameters for characterizing salt-induced spectral changes, which were dominated by the total brightness, chloride, and sulfate accumulation of the soils. The model did not yield satisfactory estimation results (RPD = 1.49) when multispectral bands were used as the input variables. Multispectral bands coupled with two two-date-based indices (D₂₂ and D_28a) used as the input variables produced the best estimation result (R² = 0.92, RPD = 3.27). Incorporating multispectral bands and two-date-based indices into the random forest model provides a remotely-sensed strategy that effectively supports the monitoring of soil salt content.

A suitable calibration sample set is extremely important to acquire an accurate spectral-based model for estimating soil organic matter (SOM). However, an unrepresentative calibration sample set was frequently collected due to the inappropriate samplings pattern caused by problematic transportation logistics and complex geographic conditions, which resulted in fairly poor generalization and low accuracy of the spectroscopic model. Thus, we hypothesized that a soil sample dataset equivalent to natural soil samples could be prepared under controlled laboratory conditions, and increase the accuracy of spectroscopic estimation of SOM content by use of a coverage assessment method that added laboratory-simulated near-natural samples to the natural samples set in order to enhance the representative sample size and variability of the calibration set. The results showed that the near-natural samples enhanced (NSE) calibration set contained 42 natural soil samples and 28 near-natural soil samples. This set exhibited sufficient coverage and better information integrity within estimators space than the initial calibration set that included 43 natural soil samples. Random forest model based on the NSE calibration set (R2=0.90; RPIQ=4.17) more accurately estimated SOM content than the spectral-based model built with the initial calibration set (R2=0.73; RPIQ=2.32); the SOM chemical compositions (e.g., lipids, polysaccharides, and lignin) and their relative abundance from the laboratory-simulated near-natural soil samples were basically consistent with those of natural soil samples. The inclusion of near-natural soil samples in the calibration set improved the SOM spectral-based estimation model, and was observed to be a practical method. Our results provided a calibration set enhancement strategy that effectively supports spectroscopic estimation model of SOM contents in the case of pattern-biased field samplings at the local scale.

Soil function degradation has impaired global work in the implementation of sustainable development goals (SDGs), and soil organic matter (SOM) is a basic and the most important indicator. The deep learning neural network (i.e., DL network) has become a popular tool for mapping SOM content at a regional scale. However, outlier sample data caused by environmental factors (e.g., moisture and vegetation) and uncertain noise (e.g., random noise and instability effects) have interfered with the determination of the function mapping relationship between target soil properties and spectral features, leading to DL networks with low generalization capability. Therefore, we introduced a spatial association module into a deep neural network to remove outlier sample data in order to construct an optimal sample set for calibrating an improved deep learning (IDL) network for SOM mapping. A total of 707 soil samples and a Sentinel-2B multispectral image were acquired during the bare soil period in Weibei, China. The variable importance in the projection approach was used to select the SOM-responsive spectral features for model inputs. Measured SOM contents were taken as the dependent variable, and the IDL network was constructed and applied to map SOM at the regional scale. The results showed that Band 11 was the most important band for SOM prediction. The band difference transformation method was able to integrate multiple-band information and enhance the absorption signal of SOM. The optimal SOM-responsive spectral features included B11, B1–B11, B2−B11, B3−B11, B4−B11, and B1−B12. The IDL network exhibited better performance (R² = 0.92; RPIQ = 4.57) regarding SOM estimation compared with the DL model performance (R² = 0.84; RPIQ = 2.84), being improved by 9.52% (for R²) and 60.92% (for RPIQ). After introducing the spatial association module, the DL network generalization capability was enhanced. SOM distribution showed a high-value (>20 g kg⁻¹) area in the south, and a low-value (<6 g kg⁻¹) area in the north of the study area (the area affected by seawater intrusion). These results provide a strategy based on an IDL network and satellite data for effectively and accurately mapping SOM at the regional scale.

Identifying the source effect on heavy metals to human health risk is essential for devising and implementing restoration policies for polluted soils. For this purpose, eight heavy metals (As, Cd, Hg, Cr, Cu, Ni, Pb, and Zn) in soil profile samples (0–10, 10–20, 20–30, and 30–40 cm) collected in the area around aluminum-plastic manufacturing facilities (APMF) were determined. An absolute principal component score multiple linear regression (APCS-MLR) model supported by a health risk assessment (HRA) model was developed to determine the source apportionment of soil heavy metals and contribution rate of pollution sources to human health risk. Results showed significant accumulations of eight metals in the topsoil (0–20 cm), parent material, transportation, APMF, and agricultural practices were the four major contributing sources for heavy metals in soils, with average contribution percentages of 21.69%, 24.99%, 29.60%, and 14.25%, respectively. Carcinogenic risk factors for adults (1.23E-04) and children (1.32E-04) were found to be above the acceptable level (1E-06 to 1E-04). The health risk quantification results indicated that parent material, APMF, transportation, agricultural practices, and unidentified factors accounted for 55.76%, 14.48%, 12.09%, 10.13%, and 7.54% of the carcinogenic risk for children and adults. The adverse impacts of Cd, Zn, and Pb accumulations in soil coming from APMF activities were significant and need to receive more attention.