Effects of freeze-thaw cycles on the spatial distribution of soil total nitrogen using a geographically weighted regression kriging method

Mapping Topsoil Total Nitrogen Using Random Forest and Modified Regression Kriging in Agricultural Areas of Central China

Accurate understanding of spatial distribution and variability of soil total nitrogen (TN) is critical for the site-specific nitrogen management. Based on 4337 newly obtained soil observations and 33 covariates, this study applied the random forest (RF) algorithm and modified regression kriging (RF combined with residual kriging: RFK, hereafter) model to spatially predict and map topsoil TN content in agricultural areas of Henan Province, central China. According to the RFK prediction, topsoil TN content ranged from 0.52 to 1.81 g kg^-1, and the farmland with the topsoil TN contents of 1.00-1.23 g kg^-1 and 0.80-1.23 g kg^-1 accounted for 48.2% and 81.2% of the total farmland area, respectively. Spatially, the topsoil TN in the study area was generally higher in the west and lower in the east. By using the Boruta variable selection algorithm, soil organic matter (SOM) and available potassium contents in topsoil, nitrogen deposition, average annual precipitation, livestock discharges, and topsoil pH were identified as the main factors driving the spatial distribution and variation of soil TN in the study area. The RF and RFK models used showed the expected performance and achieved acceptable TN prediction accuracy. In comparison, RFK performed slightly better than the RF model. The R² and RMSE achieved by the RFK model were improved by 4.5% and 4.5%, respectively, compared with that by the RF model. However, the results suggest that RFK was inferior to the RF model in quantifying prediction uncertainty and thus may have a slight disadvantage in model reliability.

Mapping the Spatial Heterogeneity of Anthropogenic Soil Nitrogen Net Replenishment Based on Soil Loss: A Coastal Case in the Yellow River Delta, China

To explore the spatial heterogeneity of nitrogen supply from human activities to soil in coastal areas, we established a soil nitrogen net replenishment index (A-SNNRI). We applied the Revised Universal Soil Loss Equation (RUSLE) model for soil loss risk calculation and geostatistical analysis for process simulation. A case study in the Yellow River Delta (YRD) showed that the A-SNNRI worked well. During the summer crop-growing season, population and land use presented significant influences on the soil total nitrogen (STN) status. Urban villages and arable land both had the largest summary STN and variety. There was a negative correlation between STN change and soil loss. The east coast held both the largest A-SNNRIs and soil loss risks. There were significant positive correlations between A-SNNRIs and population and GDP. Therefore, to control and reduce soil-source nitrogen exports in the YRD, we need to reduce nitrogen emissions from urban villages, agriculture, industry, and aquaculture and determine the main risk locations along the east coast and in the main city.

The Role of Remote Sensing Data and Methods in a Modern Approach to Fertilization in Precision Agriculture

The precision fertilization system is the basis for upgrading conventional intensive agricultural production, while achieving both high and quality yields and minimizing the negative impacts on the environment. This research aims to present the application of both conventional and modern prediction methods in precision fertilization by integrating agronomic components with the spatial component of interpolation and machine learning. While conventional methods were a cornerstone of soil prediction in the past decades, new challenges to process larger and more complex data have reduced their viability in the present. Their disadvantages of lower prediction accuracy, lack of robustness regarding the properties of input soil sample values and requirements for extensive cost- and time-expensive soil sampling were addressed. Specific conventional (ordinary kriging, inverse distance weighted) and modern machine learning methods (random forest, support vector machine, artificial neural networks, decision trees) were evaluated according to their popularity in relevant studies indexed in the Web of Science Core Collection over the past decade. As a shift towards increased prediction accuracy and computational efficiency, an overview of state-of-the-art remote sensing methods for improving precise fertilization was completed, with the accent on open-data and global satellite missions. State-of-the-art remote sensing techniques allowed hybrid interpolation to predict the sampled data supported by remote sensing data such as high-resolution multispectral, thermal and radar satellite or unmanned aerial vehicle (UAV)-based imagery in the analyzed studies. The representative overview of conventional and modern approaches to precision fertilization was performed based on 121 samples with phosphorous pentoxide (P2O5) and potassium oxide (K2O) in a common agricultural parcel in Croatia. It visually and quantitatively confirmed the superior prediction accuracy and retained local heterogeneity of the modern approach. The research concludes that remote sensing data and methods have a significant role in improving fertilization in precision agriculture today and will be increasingly important in the future.

Ingestion of GNSS-Derived ZTD and PWV for Spatial Interpolation of PM2.5 Concentration in Central and Southern China

With the increasing application of global navigation satellite system (GNSS) technology in the field of meteorology, satellite-derived zenith tropospheric delay (ZTD) and precipitable water vapor (PWV) data have been used to explore the spatial coverage pattern of PM_2.5 concentrations. In this study, the PM_2.5 concentration data obtained from 340 PM_2.5 ground stations in south-central China were used to analyze the variation patterns of PM_2.5 in south-central China at different time periods, and six PM_2.5 interpolation models were developed in the region. The spatial and temporal PM_2.5 variation patterns in central and southern China were analyzed from the perspectives of time series variations and spatial distribution characteristics, and six types of interpolation models were established in central and southern China. (1) Through correlation analysis, and exploratory regression and geographical detector methods, the correlation analysis of PM_2.5-related variables showed that the GNSS-derived PWV and ZTD were negatively correlated with PM_2.5, and that their significances and contributions to the spatial analysis were good. (2) Three types of suitable variable combinations were selected for modeling through a collinearity diagnosis, and six types of models (geographically weighted regression (GWR), geographically weighted regression kriging (GWRK), geographically weighted regression—empirical bayesian kriging (GWR-EBK), multiscale geographically weighted regression (MGWR), multiscale geographically weighted regression kriging (MGWRK), and multiscale geographically weighted regression—empirical bayesian kriging (MGWR-EBK)) were constructed. The overall R² of the GWR-EBK model construction was the best (annual: 0.962, winter: 0.966, spring: 0.926, summer: 0.873, and autumn: 0.908), and the interpolation accuracy of the GWR-EBK model constructed by inputting ZTD was the best overall, with an average RMSE of 3.22 μg/m³ recorded, while the GWR-EBK model constructed by inputting PWV had the highest interpolation accuracy in winter, with an RMSE of 4.5 μg/m³ recorded; these values were 2.17% and 4.26% higher than the RMSE values of the other two types of models (ZTD and temperature) in winter, respectively. (3) The introduction of the empirical Bayesian kriging method to interpolate the residuals of the models (GWR and MGWR) and to then correct the original interpolation results of the models was the most effective, and the accuracy improvement percentage was better than that of the ordinary kriging method. The average improvement ratios of the GWRK and GWR-EBK models compared with that of the GWR model were 5.04% and 14.74%, respectively, and the average improvement ratios of the MGWRK and MGWR-EBK models compared with that of the MGWR model were 2.79% and 12.66%, respectively. (4) Elevation intervals and provinces were classified, and the influence of the elevation and the spatial distribution of the plane on the accuracy of the PM_2.5 regional model was discussed. The experiments showed that the accuracy of the constructed regional model decreased as the elevation increased. The accuracies of the models in representing Henan, Hubei and Hunan provinces were lower than those of the models in representing Guangdong and Guangxi provinces.