Prediction of soil organic carbon in an intensively managed reclamation zone of eastern China: A comparison of multiple linear regressions and the random forest model

Machine learning-driven source identification and ecological risk prediction of heavy metal pollution in cultivated soils

To overcome challenges in assessing the impact of environmental factors on heavy metal accumulation in soil due to limited comprehensive data, our study in Yangxin County, Hubei Province, China, analyzed 577 soil samples in combination with extensive big data. We used machine learning techniques, the potential ecological risk index, and the bivariate local Moran's index (BLMI) to predict Cr, Pb, Cd, As, and Hg concentrations in cultivated soil to assess ecological risks and identify pollution sources. The random forest model was selected for its superior performance among various machine learning models, and results indicated that heavy metal accumulation was substantially influenced by environmental factors such as climate, elevation, industrial activities, soil properties, railways, and population. Our ecological risk assessment highlighted areas of concern, where Cd and Hg were identified as the primary threats. BLMI was used to analyze spatial clustering and autocorrelation patterns between ecological risk and environmental factors, pinpointing areas that require targeted interventions. Additionally, redundancy analysis revealed the dynamics of heavy metal transfer to crops. This detailed approach mapped the spatial distribution of heavy metals, highlighted the ecological risks, identified their sources, and provided essential data for effective land management and pollution mitigation.

Evaluation of the prediction effectiveness for geochemical mapping using machine learning methods: A case study from northern Guangdong Province in China

This study compares seven machine learning models to investigate whether they improve the accuracy of geochemical mapping compared to ordinary kriging (OK). Arsenic is widely present in soil due to human activities and soil parent material, posing significant toxicity. Predicting the spatial distribution of elements in soil has become a current research hotspot. Lianzhou City in northern Guangdong Province, China, was chosen as the study area, collecting a total of 2908 surface soil samples from 0 to 20 cm depth. Seven machine learning models were chosen: Random Forest (RF), Support Vector Machine (SVM), Ridge Regression (Ridge), Gradient Boosting Decision Tree (GBDT), Artificial Neural Network (ANN), K-Nearest Neighbors (KNN), and Gaussian Process Regression (GPR). Exploring the advantages and disadvantages of machine learning and traditional geological statistical models in predicting the spatial distribution of heavy metal elements, this study also analyzes factors affecting the accuracy of element prediction. The two best-performing models in the original model, RF (R2 = 0.445) and GBDT (R2 = 0.414), did not outperform OK (R2 = 0.459) in terms of prediction accuracy. Ridge and GPR, the worst-performing methods, have R2 values of only 0.201 and 0.248, respectively. To improve the models' prediction accuracy, a spatial regionalized (SR) covariate index was added. Improvements varied among different methods, with RF and GBDT increasing their R2 values from 0.4 to 0.78 after enhancement. In contrast, the GPR model showed the least significant improvement, with its R2 value only reaching 0.25 in the improved method. This study concluded that choosing the right machine learning model and considering factors that influence prediction accuracy, such as regional variations, the number of sampling points, and their distribution, are crucial for ensuring the accuracy of predictions. This provides valuable insights for future research in this area.

A comprehensive evaluation of influencing factors of neonicotinoid insecticides (NEOs) in farmland soils across China: First focus on film mulching

Neonicotinoid insecticide (NEO) residues in agricultural soils have concerning and adverse effects on agroecosystems. Previous studies on the effects of farmland type on NEOs are limited to comparing greenhouses with open fields. On the other hand, both NEOs and microplastics (MPs) are commonly found in agricultural fields, but their co-occurrence characteristics under realistic fields have not been reported. This study grouped farmlands into three types according to the covering degree of the film, collected 391 soil samples in mainland China, and found significant differences in NEO residues in the soils of the three different farmlands, with greenhouse having the highest NEO residue, followed by farmland with film mulching and farmland without film mulching (both open fields). Furthermore, this study found that MPs were significantly and positively correlated with NEOs. As far as we know this is the first report to disclose the association of film mulching and MPs with NEOs under realistic fields. Moreover, multiple linear regression and random forest models were used to comprehensively evaluate the factors influencing NEOs (including climatic, soil, and agricultural indicators). The results indicated that the random forest model was more reliable, with MPs, farmland type, and total nitrogen having higher relative contributions.

Response of soil organic carbon to straw return in farmland soil in China: A meta-analysis

Straw return is an effective measure to promote sustainable agriculture by significantly improving soil fertility. At present, few studies have been conducted on the most effective carbon enhancing management measures for various crops. Therefore, we conducted a meta-analysis using data collected from 184 literature sources, comprising 3297 data sets to analyze the carbon increase effects of straw returning in three main crops (rice, maize, and wheat) in China and to explore the influence mechanism of natural factors, soil properties, straw return measures, and cropping systems on the carbon enhancement effect. The study showed that straw return significantly increased soil organic carbon and the rate of increase was higher for wheat at 15.88% (14.74%-17.03%) than for rice at 12.7% (11.5%-13.91%) and maize at 12.42% (11.42%-13.42%), with varying degrees of improvement in other soil physicochemical properties. Natural factors have the greatest impact on the carbon increasing effect of rice fields, reaching 28.8%, especially at temperature between 10 °C and 15 °C, less than 800 mm precipitation, low latitude, and short frost-free period. Maize and wheat are most affected by soil properties, reaching 41% and 34.5% respectively. Furthermore, field management practices also play a pivotal role, organic carbon increasing obviously was observed when the C/N ratio of exogenous nutrients is bigger than 20 with the low initial organic matter. Shallow tillage and less than 7.5 t hm-2 straw returning with 3-10 years to the field are ideal for rice and maize. Crop rotation, especially in drylands, increased soil organic carbon more significantly than continuous. The results of our analysis can provide valuable insights into the effect of straw return on carbon increase. In the future, the soil carbon can be improved by adopting rational cropping patterns and straw return measures with taking into account climate and soil characteristics for different crops.

Long-term fertilization and manuring effects on the nexus between sulphur distribution and SOC in an Inceptisol over five decades under a finger millet-maize cropping system

Our investigation revealed that alterations in sulphur (S) pools are predominantly governed by soil organic carbon (SOC), soil nitrogen (N), microbial biomass, and soil enzyme activities in sandy clay loam (Vertic Ustropept) soil. We employed ten sets of nutrient management techniques, ranging from suboptimal (50% RDF) to super-optimal doses (150% RDF), including NPK + Zn, NP, N alone, S-free NPK fertilizers, NPK + FYM, and control treatments, to examine the interrelation of S with SOC characteristics. Fourier-transform infrared (FT-IR) spectroscopy was utilized to analyze the functional groups present in SOC characterization across four treatments: 100% NPK, 150% NPK, NPK + FYM, and absolute control plots. Principal component analysis (PCA) was then applied to assess 29 minimal datasets, aiming to pinpoint specific soil characteristics influencing S transformation. In an Inceptisol, the application of fertilizers (100% RDF) in conjunction with 10 t ha-1 of FYM resulted in an increase of S pools from the surface to the subsurface stratum (OS > HSS > SO42--S > WSS), along with an increase in soil N and SOC. FT-IR spectroscopy identified cellulose and thiocyanate functional groups in all four plots, with a pronounced presence of carbohydrate-protein polyphenol, sulfoxide (S=O), and nitrate groups specifically observed in the INM plot. The PCA findings indicated that the primary factors influencing soil quality and crop productivity (r2 of 0.69) are SOC, SMBC, SMBN, SMBS, and the enzyme activity of URE, DHA, and AS. According to the study, the combined application of fertilizer and FYM (10 t ha-1) together exert a positive impact on sulphur transformation, SOC accumulation, and maize yield in sandy clay loam soil.

Spatial distribution and driving factors of soil organic carbon in the Northeast China Plain: Insights from latest monitoring data

Soil organic carbon (SOC) is a critical component of soil fertility and plays a crucial role in the global carbon cycle. Despite the widespread reports of a decrease in SOC content and stock in the Northeast China Plain in recent decades, the current status and driving factors of its content and distribution are unclear. In this study, the surface soil (0-20 cm) SOC content data of 1920 sampling points within the Northeast China Plain covering an area of 2.6 × 105 km2 were obtained based on the Land Quality Geochemical Monitoring Network established in 2018. Random forest model and correlation analysis were used to identify the main driving factors of SOC distribution. The results showed that the SOC content, soil organic carbon density (SOCD), and soil organic carbon storage (SOCS) in the Northeast China Plain were 13.48 g·kg-1, 3.45 kg·C·m-2, and 898.95 Tg, respectively. SOC content in paddy land was the highest among different land use types, which reached 18.77 g·kg-1. SOC content showed strong spatial dependence and gradually increased from southwest to northeast in the monitoring area. The results of the random forest analysis showed that the SiO2, mean annual temperature, and Fe2O3 explained 39.4 %, 18.9 %, and 12.8 % of the spatial variation of SOC, respectively. Although the SOCS (0-20 cm) in the Northeast China Plain has decreased by 8.68 % in the last 40 years compared to the Second National Soil Survey (1980), it's important to note that the SOCS has transitioned from a decreasing trend between 1980 and 2006 to an increasing trend from 2006 to 2018.This study provides important information for decision-makers on the spatiotemporal changes of SOC and its driving factors in the Northeast China Plain, which has a great significance for soil carbon sequestration and the development of management strategies to maintain soil fertility.

Evaluating gross primary productivity over 9 ChinaFlux sites based on random forest regression models, remote sensing, and eddy covariance data

Accurate modeling of Gross Primary Productivity (GPP) in terrestrial ecosystems is a major challenge in quantifying the carbon cycle. Many light use efficiency (LUE) models have been developed, but the variables and algorithms used for environmental constraints in different models vary importantly. It is still unclear whether the models can be further improved by machine learning methods and the combination of different variables. Here, we have developed a series of RFR-LUE models, which used the random forest regression (RFR) algorithm based on variables of LUE models, to explore the potential of estimating site-level GPP. Based on remote sensing indices, eddy covariance and meteorological data, we applied RFR-LUE models to evaluate the effects of different variables combined on GPP on daily, 8-day, 16-day and monthly scales, respectively. Cross-validation analyses revealed performances of RFR-LUE models varied significantly among sites with R² of 0.52-0.97. Slopes of the regression relationship between simulated and observed GPP ranged from 0.59 to 0.95. Most models performed better in capturing the temporal changes and magnitude of GPP in mixed forests and evergreen needle-leaf forests than in evergreen broadleaf forests and grasslands. Performances were improved at the longer temporal scale, with the average R² for four-time resolutions of 0.81, 0.87, 0.88, and 0.90, respectively. Additionally, the importance of the variables showed that temperature and vegetation indices were critical variables for RFR-LUE models, followed by radiation and moisture variables. The importance of moisture variables was higher in non-forests than in forests. A comparison with four GPP products indicated that RFR-LUE model predicted GPP better matcher observed GPP across sites. The study provided an approach to deriving GPP fluxes and evaluating the extent to which variables affect GPP estimation. It may be used for predicting vegetation GPP at the regional scales and for calibration and evaluation of land surface process models.

Incorporating agricultural practices in digital mapping improves prediction of cropland soil organic carbon content: The case of the Tuojiang River Basin

Accurate mapping of soil organic carbon (SOC) in cropland is essential for improving soil management in agriculture and assessing the potential of different strategies aiming at climate change mitigation. Cropland management practices have large impacts on agricultural soils, but have rarely been considered in previous SOC mapping work. In this study, cropland management practices including carbon input (CI), length of cultivation (LC), and irrigation (Irri) were incorporated as agricultural management covariates and integrated with natural variables to predict the spatial distribution of SOC using the Extreme Gradient Boosting (XGBoost) model. Additionally, we evaluated the performance of incorporating agricultural management practice variables in the prediction of cropland topsoil SOC. A case study was carried out in a traditional agricultural area in the Tuojiang River Basin, China. We found that CI was the most important environmental covariate for predicting cropland SOC. Adding cropland management practices to natural variables improved prediction accuracy, with the coefficient of determination (R²), the root mean squared error (RMSE) and Lin's Concordance Correlation Coefficient (LCCC) improving by 16.67%, 17.75% and 5.62%, respectively. Our results highlight the effectiveness of incorporating agricultural management practice information into SOC prediction models. We conclude that the construction of spatio-temporal database of agricultural management practices derived from inventories is a research priority to improve the reliability of SOC model prediction.

Analyzing and predicting colour preference of colour palettes

A palette composed of multiple colour patches can express lots of information. This study aimed to explore the factors that influenced colour palette preference, including colour attributes and colour differences between colour patches. A new attribute called Delta_Order for calculating the colour difference of a palette was presented, which fully considered the colour difference and the placement. In order to comprehend colour palette preference intuitively, a prediction model of palette preference was proposed based on lightness, chroma and the new metric Delta_Order. Two psychophysical experiments including analyzing and validating experiments were conducted. Fifty observers were invited to evaluate the colour palette preference. The results indicated that lightness played an important role in colour preference, but colour preference was not related to hue angle. For Delta_Order, there was a significant negative correlation between the new metric and preference score since the Pearson correlation coefficient was -0.801. This meant that observers preferred the palettes with low Delta_Orders. According to the validating test, it confirmed that the proposed prediction model had a good stability. The predicted trends were consistent with the true results, and the scores were similar to each other. These analysis results had a certain guiding significance in design and industry about colour.

Cropland carbon stocks driven by soil characteristics, rainfall and elevation

Soil organic carbon (SOC) can influence atmospheric CO₂ concentration and then the extent to which the climate emergency is mitigated globally. It follows the elucidation of the driving factors of cropland SOC stocks, which is fundamental to reducing soil carbon loss and promoting soil carbon sequestration. Here, we examined the influence of 16 environmental variables on SOC stocks and sequestration based on three machine learning soil mapping methods, i.e. multiple linear regression (MLR), random forest (RF) and extreme gradient boosting (XGBOOST), with 2875 observed soil samples from cropland topsoil across Hunan Province, China in 2010. We employed a structural equation model (SEM) to extricate the driving mechanisms of environmental variables on SOC stocks at the regional scale. Our results show that XGBOOST had the most reliable performance in predicting SOC stocks, explaining 66 % of the total SOC stock variation. Croplands with high SOC stocks were distributed in low-altitude and water-sufficient areas. The partial dependence of SOC on precipitation showed a trend of increasing and then slowly decreasing. In addition, the grid-based SEM results clearly presented the direct and indirect routes of environmental variables' impacts on cropland SOC stocks. Soil properties regulated by elevation, were the most influential natural factor on SOC stocks. Precipitation and elevation drove SOC stocks through direct and indirect effects respectively. Our SEM combined with machine learning approach can provide an effective explanation of the driving mechanism for SOC accumulation. We expect our proposed modelling approach can be applied to other regions and offer new insights, as a reference for mitigating cropland soil carbon loss under climate emergency conditions.

Machine learning-based prediction of toxic metals concentration in an acid mine drainage environment, northern Tunisia

In northern Tunisia, Sidi Driss sulfide ore valorization had produced a large waste amount. The long tailings exposure period and in situ minerals interactions produced an acid mine drainage (AMD) which contributed to a strong increase in the mobility and migration of huge heavy metal (HM) quantities to the surrounding soils. In this work, the soil mineral proportions, grain sizes, physicochemical properties, SO₄^2- and S contents, and Machine Learning (ML) algorithms such as the Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Network (ANN) models were used to predict the soil HM quantities transferred from Sidi-Driss mine drainage to surrounding soils. The results showed that the HM concentrations had significantly increased with the increase of decomposition and oxidation of galena, marcasite, pyrite, and sphalerite-marcasite and Fe-oxide-hydroxides quantities and the sulfate dissolution (marked with SO₄^2- ions increase) that produced the decreased soil pH. Compared to SVM, and ANN models outputs, the RF model that revealed higher R²_val, RPD, RPIQ, and lower error indices had satisfactorily predicted the soil HM accumulation coming from the AMD environment.

The Transformation of 40-Year Coastal Wetland Policies in China: Network Analysis and Text Analysis

The notable improvement of China's wetland management policies over the last four decades prompted this study's goal to quantitatively analyze the transformation of China's coastal wetland policies from 1979 to 2022 by applying an institutional network analysis and policy text analysis. The results of the institutional network analysis revealed an administrative management transformation from a multidepartmental mode to an integrated management framework. Furthermore, the policy text analysis results revealed a change in policy priorities (from exploitation to protection) and management targets (from a single environmental element to a comprehensive ecosystem and further to collaborative governance). In addition, the overall outcome of this study instigated proposals for the improvement of future wetland policies on climate change, integrated planning, natural capital, and public participation. Hence, this study presents an example of wetland policy analysis based on a quantitative review, which we hope will also be valuable for other countries.

Predicting spatial variability of species diversity with the minimum data set of soil properties in an arid desert riparian forest

Species diversity has spatial heterogeneity in ecological systems. Although a large number of studies have demonstrated the influence of soil properties on species diversity, most of them have not considered their spatial variabilities. To remedy the knowledge gap, a 1 ha (100 m × 100 m) plots of arid desert riparian forest was set up in the Ebinur Wetland Nature Reserve (ELWNR) in the NW China. Then, the minimum data set of soil properties (soil MDS) was established using the Principal Component Analysis (PCA) and the Norm Value Determination to represent the total soil property data set (soil TDS). The Geo-statistics and two models (i.e., Random Forest/RF and Multiple Linear Regression/MLR) were used to measure the spatial variability of species diversity, and predict its spatial distribution by the soil MDS, respectively. The results showed that the soil MDS was composed of soil salt content (SSC), soil total phosphorus (STP), soil available phosphorus (SAP), soil organic carbon (SOC) and soil nitrate nitrogen (SNN); which represented the soil TDS perfectly (R^{2 =} 0.62). Three species diversity indices (i.e., Shannon-Wiener, Simpson and Pielou indices) had a high spatial dependence (C₀/(C₀+C)< 25%; 0.72 m ≤ range≤ 0.77 m). Ordinary kriging distribution maps showed that the spatial distribution pattern of species diversity predicted by RF model was closer to its actual distribution compared with MLR model. RF model results suggested that the soil MDS had significant effect on spatial distribution of Shannon-Wiener, Simpson and Pielou indices (Var_ex = 56%, 49% and 36%, respectively). Among all constituents, SSC had the largest contribution on the spatial variability of species diversity (nearly 10%), while STP had least effect (< 5.3%). We concluded that the soil MDS affected spatial variability of species diversity in arid desert riparian forests. Using RF model can predict spatial variability of species diversity through soil properties. Our work provided a new case and insight for studying the spatial relationship between soil properties and plant species diversity.

Estimation of Heavy Metal Content in Soil Based on Machine Learning Models

Heavy metal pollution in soil is threatening the ecological environment and human health. However, field measurement of heavy metal content in soil entails significant costs. Therefore, this study explores the estimation method of soil heavy metals based on remote sensing images and machine learning. To accurately estimate the heavy metal content, we propose a hybrid artificial intelligence model integrating least absolute shrinkage and selection operator (LASSO), genetic algorithm (GA) and error back propagation neural network (BPNN), namely the LASSO-GA-BPNN model. Meanwhile, this study compares the accuracy of the LASSO-GA-BPNN model, SVR (Support Vector Regression), RF (Random Forest) and spatial interpolation methods with Huanghua city as an example. Furthermore, the study uses the LASSO-GA-BPNN model to estimate the content of eight heavy metals (including Ni, Pb, Cr, Hg, Cd, As, Cu, and Zn) in Huanghua and visualize the results in high resolution. In addition, we calculate the Nemerow index based on the estimation results. The results denote that, the simultaneous optimization of BPNN by LASSO and GA can greatly improve the estimation accuracy and generalization ability. The LASSO-GA-BPNN model is a more accurate model for the estimate heavy metal content in soil compared to SVR, RF and spatial interpolation. Moreover, the comprehensive pollution level in Huanghua is mainly low pollution. The overall spatial distribution law of each heavy metal content is very similar, and the local spatial distribution of each heavy metal is different. The results are of great significance for soil pollution estimation.

Mapping of Soil Organic Carbon Stocks Based on Aerial Photography in a Fragmented Desertification Landscape

Northern China’s agropastoral ecotone has been a key area of desertification control for decades, and digital maps of its soil organic carbon (SOC) stocks are needed to reveal the gaps between the actual SOC levels and baseline to support land degradation neutrality (LDN) under the Sustainable Development Goals. However, reliable soil information is scarce, and accurate prediction is hindered by the fragmented landscape, which is a dominant characteristic of desertified land. To improve the patchiness identification and accuracy of SOC prediction, we conducted field surveys and collected low-altitude aerial images along the desertification degrees (severe and extremely severe, moderate, slight) in the Horqin Sandy Land. Linear regressions were performed on the relationships between the normalized difference vegetation index and the fractional vegetation cover (FVC) extracted from aerial images, and regression kriging was applied to predict SOC stocks based on the soil-forming factors (vegetation, climate, and topography). Our prediction and cross-validation showed that the fragmented structure and prediction accuracy of SOC stocks were both greatly improved for desertified land. The FVC (R2c = 0.94) and evapotranspiration (R2c = 0.86) had significant positive effects on SOC stocks, respectively, with indirect and direct causal relationships. Our results could provide soil information with better patchiness and accuracy to help policymakers determine the future LDN status in this fragmented desertification landscape. As drone technology becomes more available, it will fully support digital mapping of soil properties.

Air quality prediction models based on meteorological factors and real-time data of industrial waste gas

With the rapid economic growth, air quality continues to decline. High-intensity pollution emissions and unfavorable weather conditions are the key factors for the formation and development of air heavy pollution processes. Given that research into air quality prediction generally ignore pollutant emission information, in this paper, the random forest supervised learning algorithm is used to construct an air quality prediction model for Zhangdian District with industrial waste gas daily emissions and meteorological factors as variables. The training data include the air quality index (AQI) values, meteorological factors and industrial waste gas daily emission of Zhangdian District from 1st January 2017 to 30th November 2019. The data from 1st to 31th December 2019 is used as the test set to assess the model. The performance of the model is analysed and compared with the backpropagation (BP) neural network, decision tree, and least squares support vector machine (LSSVM) function, which has better overall prediction performance with an RMSE of 22.91 and an MAE of 15.80. Based on meteorological forecasts and expected air quality, a daily emission limit for industrial waste gas can be obtained using model inversion. From 1st to 31th December 2019, if the industrial waste gas daily emission in this area were decreased from 6048.5 million cubic meters of waste gas to 5687.5 million cubic meters, and the daily air quality would be maintained at a good level. This paper deeply explores the dynamic relationship between waste gas daily emissions of industrial enterprises, meteorological factors, and air quality. The meteorological conditions are fully utilized to dynamically adjust the exhaust gas emissions of key polluting enterprises. It not only ensures that the regional air quality is in good condition, but also promotes the in-depth optimization of the procedures of regional industrial enterprises, and reduces the conflict between environmental protection and economic development.

Impact of Flooding on Shallow Groundwater Chemistry in the Taklamakan Desert Hinterland: Remote Sensing Inversion and Geochemical Methods

Understanding the effect of flooding on groundwater quality is imperative for oasis vegetation protection and local ecological environment development. We used geochemical and remote sensing inversion methods to evaluate the effects of flood recharge on the groundwater hydrochemical and geochemical processes in the Daliyaboy Oasis. Groundwater samples were collected from 30 ecological observation wells in the study area before (PRF) and after (POF) the flood. Except for small changes in HCO3− and K+ and a decrease in pH, ion levels were higher POF than PRF, and the water chemistry was essentially unchanged. In the POF groundwater, HCO3− was correlated with Cl−, Na+, Mg2+, total soluble solids (TDS), and electrical conductivity (EC), but not with SO42−, Ca2+, K+, or pH, and was positively correlated with all other variables, while the remaining variables, except for pH, were strongly positively correlated with each other. PRF water chemistry was controlled by silicate and evaporite mineral weathering and evaporation processes, resulting in high groundwater TDS, EC, and a major ion content, while POF major groundwater ions were regulated by mineral weathering and flood recharge. We demonstrated the high accuracy of remote sensing inversion, confirming this as a reliable method for evaluating groundwater chemistry. The results of the study help to reshape and predict the history of the regional hydrogeological environment and hydrogeochemical development, and provide a theoretical basis for assessing the rational use of local water resources and protecting the ecological environment.

A novel seasonal index-based machine learning approach for air pollution forecasting

Novel machine learning models (MLMs) using the seasonal indexing approach that captures the variation in air quality caused due to meteorological changes have been used to provide short-term, real-time forecasts of PM_2.5 concentration for one of the most polluted air quality control regions (AQCR) in the capital city of Delhi. Two MLMs-multi-linear regression and random forest-have been developed for using time series data for 1-h and 24-h average PM_2.5 concentration. Short-term, real-time forecasts have been made using the developed models. Various model performance evaluation indices indicate satisfactory model performance. R² values for the hourly and daily models varied between 0.95 and 0.72 and between 0.76 and 0.68 for the 1st to 5th h/day, respectively. The lagged values of PM_2.5 concentration (persistence) and the hourly and daily indices are the most influential variables for the forecasts for immediate time steps. In contrast, seasonal indices become more important with the forecasting time horizon. The developed models can be used for making short-term, real-time air quality forecasts and issuing a warning when the pollution levels go beyond acceptable limits.

Forecasts of tropospheric ozone in the Metropolitan Area of Rio de Janeiro based on missing data imputation and multivariate calibration techniques

Multivariate calibration based on partial least squares, random forest, and support vector machine methods, combined with the MissForest imputation algorithm, was used to understand the interaction between ozone and nitrogen oxides, carbon monoxide, wind speed, solar radiation, temperature, relative humidity, and others, the data of which were collected by air quality monitoring stations in the metropolitan area of Rio de Janeiro in four distinct sites between, 2014 and, 2018. These techniques provide an easy and feasible way of modeling and analyzing air pollutants and can be used when coupled with other methods. The results showed that random forest and support vector machine chemometric techniques can be used in modeling and predicting tropospheric ozone concentrations, with a coefficient of determination for making predictions up to 0.92, a root-mean square error of calibration between 4.66 and 27.15 µg m^-3, and a root-mean square error of prediction between 4.17 and 22.45 µg m^-3, depending on the air quality monitoring stations and season.

Assessing model performance via the most limiting environmental driver in two differently stressed pine stands

Climate change will impact forest productivity worldwide. Forecasting the magnitude of such impact, with multiple environmental stressors changing simultaneously, is only possible with the help of process-based models. In order to assess their performance, such models require careful evaluation against measurements. However, direct comparison of model outputs against observational data is often not reliable, as models may provide the right answers due to the wrong reasons. This would severely hinder forecasting abilities under unprecedented climate conditions. Here, we present a methodology for model assessment, which supplements the traditional output-to-observation model validation. It evaluates model performance through its ability to reproduce observed seasonal changes of the most limiting environmental driver (MLED) for a given process, here daily gross primary productivity (GPP). We analyzed seasonal changes of the MLED for GPP in two contrasting pine forests, the Mediterranean Pinus halepensis Mill. Yatir (Israel) and the boreal Pinus sylvestris L. Hyytiälä (Finland) from three years of eddy-covariance flux data. Then, we simulated the same period with a state-of-the-art process-based simulation model (LandscapeDNDC). Finally, we assessed if the model was able to reproduce both GPP observations and MLED seasonality. We found that the model reproduced the seasonality of GPP in both stands, but it was slightly overestimated without site-specific fine-tuning. Interestingly, although LandscapeDNDC properly captured the main MLED in Hyytiälä (temperature) and in Yatir (soil water availability), it failed to reproduce high-temperature and high-vapor pressure limitations of GPP in Yatir during spring and summer. We deduced that the most likely reason for this divergence is an incomplete description of stomatal behavior. In summary, this study validates the MLED approach as a model evaluation tool, and opens up new possibilities for model improvement.

Prediction of Soil Organic Carbon under Different Land Use Types Using Sentinel-1/-2 Data in a Small Watershed

Soil organic carbon (SOC) is a key property for evaluating soil quality. SOC is thus an important parameter of agricultural soils and needs to be regularly monitored. The aim of this study is to explore the potential of synthetic aperture radar (SAR) satellite imagery (Sentinel-1), optical satellite imagery (Sentinel-2), and digital elevation model (DEM) data to estimate the SOC content under different land use types. The extreme gradient boosting (XGboost) algorithm was used to predict the SOC content and evaluate the importance of feature variables under different land use types. For this purpose, 290 topsoil samples were collected and 49 features were derived from remote sensing images and DEM. Feature selection was carried out to prevent data redundancy. Coefficient of determination (R2), mean absolute error (MAE), mean squared error (MSE), percent root mean squared error (%RMSE), ratio of performance to interquartile range (RPIQ), and corrected akaike information criterion (AICc) were employed for evaluating model performance. The results showed that Sentinel-1 and Sentinel-2 data were both important for the prediction of SOC and the prediction accuracy of the model differed with land use types. Among them, the prediction accuracy of this model is the best for orchard (R2 = 0.86 and MSE = 0.004%), good for dry land (R2 = 0.74 and MSE = 0.008%) and paddy field (R2 = 0.66 and MSE = 0.009%). The prediction model of SOC content is effective and can provide support for the application of remote sensing data to soil property monitoring.

Quantifying the influencing factors and multi-factor interactions affecting cadmium accumulation in limestone-derived agricultural soil using random forest (RF) approach

Cadmium (Cd) is a highly toxic heavy metal that occurs widely in the environment and poses extensive threats to human health, animals, and plants. This study aims to identify and apportion multi-source and multi-phase Cd pollution from natural and anthropogenic inputs using ensemble models that include random forest (RF) in agricultural soils on Karst areas. The contributions of natural and anthropogenic factors to Cd accumulation were quantitatively assessed using the RF machine learning method. The results revealed that the main influencing factors were pH, organic carbon (C_org), and elevation. Moreover, the interaction effects of pH and C_org on distance and elevation were also quantified and visualised. It is observed that pH and C_org had stronger effects on soil Cd concentration than that of distance when pH > 7.02 and C_org > 1.53. In other words, higher Cd content in the soil along roadways may be caused by the interaction of distance, pH and C_org, with pH and C_org playing the dominant role in our case. Moreover, the maximum contribution of a single factor, elevation, to Cd concentration was about 0.13 mg/kg, and its interactions reached 1.082 mg/kg and 0.83 mg/kg, respectively, when combined with pH and C_org at 194.0 m. However, with increasing elevation, pH and C_org gradually took over the leading roles. This result not only gives us a quantitative understanding of the relationship between the factors that affect soil cadmium accumulation, but also provides an accurate method for source apportionment of heavy metals in soil.

Machine learning for predicting greenhouse gas emissions from agricultural soils

Machine learning (ML) models are increasingly used to study complex environmental phenomena with high variability in time and space. In this study, the potential of exploiting three categories of ML regression models, including classical regression, shallow learning and deep learning for predicting soil greenhouse gas (GHG) emissions from an agricultural field was explored. Carbon dioxide (CO₂) and nitrous oxide (N₂O) fluxes, as well as various environmental, agronomic and soil data were measured at the site over a five-year period in Quebec, Canada. The rigorous analysis, which included statistical comparison and cross-validation for the prediction of CO₂ and N₂O fluxes, confirmed that the LSTM model performed the best among the considered ML models with the highest R coefficient and the lowest root mean squared error (RMSE) values (R = 0.87 and RMSE = 30.3 mg·m^-2·hr^-1 for CO₂ flux prediction and R = 0.86 and RMSE = 0.19 mg·m^-2·hr^-1 for N₂O flux prediction). The predictive performances of LSTM were more accurate than those simulated in a previous study conducted by a biophysical-based Root Zone Water Quality Model (RZWQM2). The classical regression models (namely RF, SVM and LASSO) satisfactorily simulated cyclical and seasonal variations of CO₂ fluxes (R = 0.75, 0.71 and 0.68, respectively); however, they failed to reasonably predict the peak values of N₂O fluxes (R < 0.25). Shallow ML was found to be less effective in predicting GHG fluxes than other considered ML models (R < 0.7 for CO₂ flux and R < 0.3 for estimating N₂O fluxes) and was the most sensitive to hyperparameter tuning. Based on this comprehensive comparison study, it was elicited that the LSTM model can be employed successfully in simulating GHG emissions from agricultural soils, providing a new perspective on the application of machine learning modeling for predicting GHG emissions to the environment.

Chemometrics for environmental monitoring: a review

Environmental monitoring is necessary to ensure the overall health and conservation of an ecosystem. However, ecosystems (e.g. air, water, soil), are complex, involving numerous processes (both native and external), inputs, contaminants, and living organisms. As such, monitoring an environmental system is not a trivial task. The data obtained from natural systems is often multifaceted and convoluted, as a multitude of inputs can be intertwined within the matrix of the information obtained as part of a study. This means that trends and important results can be easily overlooked by conventional and single dimensional data analysis protocols. Recently, chemometric methods have emerged as a powerful method for maximizing the details contained within a chemical data set. Specifically, chemometrics refers to the use of mathematical and statistical analysis methods to evaluate chemical data, beyond univariant analysis. This type of analysis can provide a quantitative description of environmental measurements, while also having the capacity to reveal previously overlooked trends in data sets. Applying chemometrics to environmental data allows us to identify and describe the inter-relationship of environmental drivers, sources of contamination, and their potential impact upon the environment. This review aims to provide a detailed understanding of chemometric techniques, how they are currently used in environmental monitoring, and how these techniques can be used to improve current practices. An enhanced ability to monitor environmental conditions and to predict trends would be greatly beneficial to government and research agencies in their ability to develop environmental policies and analytical procedures.

Spatially Explicit Mapping of Historical Population Density with Random Forest Regression: A Case Study of Gansu Province, China, in 1820 and 2000

This study established a random forest regression model (RFRM) using terrain factors, climatic and river factors, distances to the capitals of provinces, prefectures (Fu, in Chinese Pinyin), and counties as independent variables to predict the population density. Then, using the RFRM, we explicitly reconstructed the spatial distribution of the population density of Gansu Province, China, in 1820 and 2000, at a resolution of 10 by 10 km. By comparing the explicit reconstruction with census data at the township level from 2000, we found that the RFRM-based approach mostly reproduced the spatial variability in the population density, with a determination coefficient (R2) of 0.82, a positive reduction of error (RE, 0.72) and a coefficient of efficiency (CE) of 0.65. The RFRM-based reconstructions show that the population of Gansu Province in 1820 was mostly distributed in the Lanzhou, Gongchang, Pingliang, Qinzhou, Qingyang, and Ningxia prefecture. The macro-spatial pattern of the population density in 2000 kept approximately similar with that in 1820. However, fine differences could be found. The 79.92% of the population growth of Gansu Province from 1820 to 2000 occurred in areas lower than 2500 m. As a result, the population weighting in the areas above 2500 m was ~9% in 1820 while it was greater than 14% in 2000. Moreover, in comparison to 1820, the population density intensified in Lanzhou, Xining, Yinchuan, Baiyin, Linxia, and Tianshui, while it weakened in Gongchang, Qingyang, Ganzhou, and Suzhou.

Seasonal variability of forest sensitivity to heat and drought stresses: A synthesis based on carbon fluxes from North American forest ecosystems

Climate extremes such as heat waves and droughts are projected to occur more frequently with increasing temperature and an intensified hydrological cycle. It is important to understand and quantify how forest carbon fluxes respond to heat and drought stress. In this study, we developed a series of daily indices of sensitivity to heat and drought stress as indicated by air temperature (T_a ) and evaporative fraction (EF). Using normalized daily carbon fluxes from the FLUXNET Network for 34 forest sites in North America, the seasonal pattern of sensitivities of net ecosystem productivity (NEP), gross ecosystem productivity (GEP) and ecosystem respiration (RE) in response to T_a and EF anomalies were compared for different forest types. The results showed that warm temperatures in spring had a positive effect on NEP in conifer forests but a negative impact in deciduous forests. GEP in conifer forests increased with higher temperature anomalies in spring but decreased in summer. The drought-induced decrease in NEP, which mostly occurred in the deciduous forests, was mostly driven by the reduction in GEP. In conifer forests, drought had a similar dampening effect on both GEP and RE, therefore leading to a neutral NEP response. The NEP sensitivity to T_a anomalies increased with increasing mean annual temperature. Drier sites were less sensitive to drought stress in summer. Natural forests with older stand age tended to be more resilient to the climate stresses compared to managed younger forests. The results of the Classification and Regression Tree analysis showed that seasons and ecosystem productivity were the most powerful variables in explaining the variation of forest sensitivity to heat and drought stress. Our results implied that the magnitude and direction of carbon flux changes in response to climate extremes are highly dependent on the seasonal dynamics of forests and the timing of the climate extremes.

Tackling environmental challenges in pollution controls using artificial intelligence: A review

This review presents the developments in artificial intelligence technologies for environmental pollution controls. A number of AI approaches, which start with the reliable mapping of nonlinear behavior between inputs and outputs in chemical and biological processes in terms of prediction models to the emerging optimization and control algorithms that study the pollutants removal processes and intelligent control systems, have been developed for environmental clean-ups. The characteristics, advantages and limitations of AI methods, including single and hybrid AI methods, were overviewed. Hybrid AI methods exhibited synergistic effects, but with computational heaviness. The up-to-date review summarizes i) Various artificial neural networks employed in wastewater degradation process for the prediction of removal efficiency of pollutants and the search of optimizing experimental conditions; ii) Evaluation of fuzzy logic used for intelligent control of aerobic stage of wastewater treatment process; iii) AI-aided soft-sensors for precisely on-line/off-line estimation of hard-to-measure parameters in wastewater treatment plants; iv) Single and hybrid AI methods applied to estimate pollutants concentrations and design monitoring and early-warning systems for both aquatic and atmospheric environments; v) AI modelings of short-term, mid-term and long-term solid waste generations, and various ANNs for solid waste recycling and reduction. Finally, the future challenges of AI-based models employed in the environmental fields are discussed and proposed.

Quantitative Inversion of Fixed Carbon Content in Coal Gangue by Thermal Infrared Spectral Data

Fixed carbon content is an important factor in measuring the carbon content of gangue, which is important for monitoring the spontaneous combustion of gangue and reusing coal gangue resources. Although traditional measurement methods of fixed carbon content, such as chemical tests, can achieve high accuracy, meeting the actual needs of mines via these tests is difficult because the measurement process is time consuming and costly and requires professional input. In this paper, we obtained the thermal infrared spectrum of coal gangue and developed a new spectral index to achieve the automated quantification of fixed carbon content. Thermal infrared spectroscopy analyses of 42 gangue and three coal samples were performed using a Turbo FT thermal infrared spectrometer. Then, the ratio index (RI), difference index (DI) and normalized difference index (NDI) were defined based on the spectral characteristics. The correlation coefficient between the spectral index and the thermal infrared spectrum was calculated, and a regression model was established by selecting the optimal spectral DI. The model prediction results were verified by a ten times 5-fold cross-validation method. The results showed that the mean error of the proposed method is 5.00%, and the root mean square error is 6.70. For comparison, the fixed carbon content was further predicted by another four methods, according to the spectral depth H, spectral area A, the random forest and support vector machine algorithms. The predicted accuracy calculated by the proposed method was the best among the five methods. Therefore, this model can be applied to predict the fixed carbon content of coal gangue in coal mines and can help guide mine safety and environmental protection, and it presents the advantages of being economic, rapid and efficient.

A Comparative Assessment of Geostatistical, Machine Learning, and Hybrid Approaches for Mapping Topsoil Organic Carbon Content

Accurate digital soil mapping (DSM) of soil organic carbon (SOC) is still a challenging subject because of its spatial variability and dependency. This study is aimed at comparing six typical methods in three types of DSM techniques for SOC mapping in an area surrounding Changchun in Northeast China. The methods include ordinary kriging (OK) and geographically weighted regression (GWR) from geostatistics, support vector machines for regression (SVR) and artificial neural networks (ANN) from machine learning, and geographically weighted regression kriging (GWRK) and artificial neural networks kriging (ANNK) from hybrid approaches. The hybrid approaches, in particular, integrated the GWR from geostatistics and ANN from machine learning with the estimation of residuals by ordinary kriging, respectively. Environmental variables, including soil properties, climatic, topographic, and remote sensing data, were used for modeling. The mapping results of SOC content from different models were validated by independent testing data based on values of the mean error, root mean squared error and coefficient of determination. The prediction maps depicted spatial variation and patterns of SOC content of the study area. The results showed the accuracy ranking of the compared methods in decreasing order was ANNK, SVR, ANN, GWRK, OK, and GWR. Two-step hybrid approaches performed better than the corresponding individual models, and non-linear models performed better than the linear models. When considering the uncertainty and efficiency, ML and two-step approach are more suitable than geostatistics in regional landscapes with the high heterogeneity. The study concludes that ANNK is a promising approach for mapping SOC content at a local scale.

Spatial analysis of total dissolved solids in Dezful Aquifer: Comparison between universal and fixed rank kriging

The spatial structure and auto-correlation of total dissolved solids (TDS) in an aquifer located in southwestern part of Iran were investigated by both Moran's index and variography. Since the feature of interest was non-stationary so, conventional methods of spatial analysis were not applicable and Universal kriging (UK) as a common method for spatial prediction of features with a spatial trend along with a novel geostatistical method known as fixed rank kriging (FRK) were utilized in this respect. The results of Moran's index were consistent with that of spatial analysis by geostatistical methods indicating the dominance of spatial clusters within the extent of study area. The spatial analysis by FRK was more efficient than that of its UK counterpart however the performance of UK was reasonable enough, as well. A variable selection by random forest (RF) was applied on eleven other water quality parameters that were the main constituents of TDS to identify the main parameters influencing the observed variability of TDS. It was turned out that RF is a viable method for variable selection in the realm of environmental sciences.

Chlorophyll a predictability and relative importance of factors governing lake phytoplankton at different timescales

Assessing the key drivers of eutrophication in lakes and reservoirs has long been a challenge, and many studies have developed empirical models for predicting the relative importance of these drivers. However, the relative roles of various parameters might differ not only spatially (between regions or localities) but also at a temporal scale. In this study, the relative roles of total phosphorus, total nitrogen, ammonia, wind speed and water temperature were selected as potential drivers of phytoplankton biomass by using chlorophyll a as a proxy for biomass. A generalized additive model (GAM) and a random forest model (RF) were developed to assess the predictability of chlorophyll a and the relative importance of various predictors driving algal blooms at different timescales in a freshwater lake. The results showed that the daily datasets yielded better predictability than the monthly datasets. In addition, at a daily scale, water temperature was a more important predictor of chlorophyll a than nutrients, and the importance of phosphorus was comparable to that of nitrogen. In contrast, at a monthly scale, nutrients are more important predictors than water temperature and phosphorus is a better predictor than nitrogen. This study indicates that the drivers of phytoplankton fluctuations vary at different timescales and that timescale has an influence on the relative roles of nitrogen and phosphorus limitation in lakes, which suggests that the temporal scale should be considered when explaining phytoplankton fluctuations. Moreover, this study provides a reference for the monitoring of phytoplankton fluctuations and for understanding the mechanisms underlying phytoplankton fluctuations at different timescales.

River ice breakup timing prediction through stacking multi-type model trees

River ice breakup is an annual event with ecological and economic significance in the Northern Hemisphere. Breakup timing forecasting is critical for supporting emergency responses to river-ice related flooding. Little attention has been paid to applications of the classification and regression tree (CART) and M5 models as well as the stacking ensemble of multiple types of model trees to river ice forecasting problem. Thus, a framework of stacking ensemble tree models (SETM) is proposed, which consists of multiple types of model trees in a two-level structure: base and ensemble models. The Athabasca River at Fort McMurray is selected as the study area because the Athabasca River is the largest unregulated river in Alberta, Canada and ice jams frequently occur in the vicinity of Fort McMurray. To facilitate the comparison of models, the historical data in the past 36 years is collected and the leave-one-out cross validation method is employed. The results show that, the indicators influencing or corresponding with the breakup timing can be categorized as temperature and water flow conditions just before breakup (in March), during freeze-up (in last November and last December) and during middle winter (in January). The performance of optimal CART and M5 models are almost the same but the M5 model does simplify the tree structure. Although their performance can be further improved by the SETM framework, the structure of the base models can facilitate explicit explanations of the relations between indicators and the breakup date. In terms of validation performance (RMSE_avg), the optimal ensemble model is the simple average method, which improves upon the two optimal base models (CART and M5) by 13.1% and 13.2%, respectively.

Spatiotemporal distribution and national measurement of the global carbonate carbon sink

The magnitudes, spatial distributions and contributions to global carbon budget of the global carbonate carbon sink (CCS) still remain uncertain, allowing the problem of national measurement of CCS remain unresolved which will directly influence the fairness of global carbon markets and emission trading. Here, based on high spatiotemporal resolution ecological, meteorological raster data and chemical field monitoring data, combining highly reliable machine learning algorithm with the thermodynamic dissolution equilibrium model, we estimated the new CCS of 0.89 ± 0.23 petagrams of carbon per year (Pg C yr^-1), amounting to 74.50% of global net forest sink and accounting for 28.75% of terrestrial sinks or 46.81% of the missing sink. Our measurement for 142 nations of CCS showed that Russia, Canada, China and the USA contribute over half of the global CCS. We also presented the first global fluxes maps of the CCS with spatial resolution of 0.05°, exhibiting two peaks in equatorial regions (10°S to 10°N) and low latitudes (10°N to 35°N) in Northern Hemisphere. By contrast, there are no peaks in Southern Hemisphere. The greatest average carbon sink flux (CCSF), i.e., 2.12 tC ha^-1 yr^-1, for 2000 to 2014 was contributed by tropical rainforest climate near the equator, and the smallest average CCSF was presented in tropical arid zones, showing a magnitude of 0.26 tC ha^-1 yr^-1. This research estimated the magnitudes, spatial distributions, variations and contributions to the global carbon budget of the CCS in a higher spatiotemporal representativeness and expandability way, which, via multiple mechanisms, introduced an important sink in the terrestrial carbon sink system and the global missing sink and that can help us further reveal and support our understanding of global rock weathering carbon sequestration, terrestrial carbon sink system and global carbon cycle dynamics which make our understanding of global change more comprehensive.

Influence of Sampling Point Discretization on the Regional Variability of Soil Organic Carbon in the Red Soil Region, China

Research on the regional variability of soil organic carbon (SOC) has focused mostly on the influence of the number of soil sampling points and interpolation methods. Little attention has typically been paid to the influence of sampling point discretization. Based on dense soil sampling points in the red soil area of Southern China, we obtained four sample discretization levels by a resampling operation. Then, regional SOC distributions were obtained at four levels by two interpolation methods: ordinary Kriging (OK) and Kriging combined with land use information (LuK). To evaluate the influence of sample discretization on revealing SOC variability, we compared the interpolation accuracies at four discretization levels with uniformly distributed validation points. The results demonstrated that the spatial distribution patterns of SOC were roughly similar, but the contour details in some local areas were different at the various discretization levels. Moreover, the predicted mean absolute errors (MAE) and root mean square errors (RMSE) of the two Kriging methods all rose with an increase in discretization. From the lowest to the largest discretization level, the MAEs of OK and LuK rose from 4.47 and 3.02 g kg−1 to 5.46 and 3.54 g kg−1, and the RMSEs rose from 5.13 and 3.95 g kg−1 to 5.76 and 4.76 g kg−1, respectively. Though the trend of prediction errors varied with discretization levels, the interpolation accuracies of the two Kriging methods were both influenced by the sample discretization level. Furthermore, the spatial interpolation uncertainty of OK was more sensitive to the discretization level than that of the LuK method. Therefore, when the spatial distribution of SOC is predicted using Kriging methods based on the same sample quantity, the more uniformly distributed sampling points are, the more accurate the spatial prediction accuracy of SOC will be, and vice versa. The results of this study can act as a useful reference for evaluating the uncertainty of SOC spatial interpolation and making a soil sampling scheme in the red soil region of China.

A Machine Learning Approach to the Residential Relocation Distance of Households in the Seoul Metropolitan Region

This study aimed to evaluate the applicability of a machine learning approach to the description of residential mobility patterns of households in the Seoul metropolitan region (SMR). The spatial range and temporal scope of the empirical study were set to 2015 to review the most recent residential mobility patterns in the SMR. The analysis data used in this study included the Internal Migration Statistics microdata provided by the Microdata Integrated Service of Statistics Korea. We analysed the residential relocation distance of households in the SMR using machine learning techniques, such as ordinary least squares regression and decision tree regression. The results of this study showed that a decision tree model can be more advantageous than ordinary least squares regression in terms of explanatory power and estimation of moving distance. A large number of residential movements are mainly related to the accessibility to employment markets and some household characteristics. The shortest movements occur when households with two or more members move into densely populated districts. In contrast, job-based residential movements are relatively farther. Furthermore, we derived knowledge on residential relocation distance, which can provide significant information for the urban management of metropolitan residential districts and the construction of reasonable housing policies.

Simulating spatial distribution of coastal soil carbon content using a comprehensive land surface factor system based on remote sensing

Surface soil carbon content (SCC) in coastal area is affected by complex factors, and revealing the SCC spatial distribution is considerably significant for judging the quantity of stored carbon and identifying the driving factors of SCC variation. A comprehensive land surface factor system (CLSFS) was established; it utilized the ecological significances of remote sensing data and included four-class factors, namely, spectrum information, ecological indices, spatial location, and land cover. Different simulation algorithms, including single-factor regression (SFR), multiple-factor regression (MFR), partial least squares regression (PLSR), and back propagation neural network (BPNN), were adopted to conduct the surface (0-30cm) SCC mapping in the Yellow River Delta in China, and a 10-fold cross validation approach was used to validate the uncertainty and accuracy of the algorithms. The results indicated that the mean simulated standard deviations were all <0.5g/kg and thus showed a low uncertainty; the mean root mean squared errors based on the simulated and measured SCC were 3.88g/kg (SFR), 3.85g/kg (PLSR), 3.67g/kg (MFR), and 2.78g/kg (BPNN) with the BPNN exhibiting a high accuracy compared to similar studies. The mean SCC was 17.40g/kg in the Yellow River Delta with distinct spatial heterogeneity; in general, the SCC in the alongshore regions, except for estuaries, was low, and that in the west of the study area was high. The mean SCCs in farmland (18.31g/kg) and wetland vegetation (17.98g/kg) were higher than those in water area (16.07g/kg), saltern (15.61g/kg), and bare land (14.71g/kg). Land-sea interaction and human activity jointly affected the SCC spatial distribution. The CLSFS was proven to have good applicability, and can be widely used in simulating the SCC spatial distribution in coastal areas.