No abstract

No abstract

No abstract

No abstract

Understanding the spatial distribution of soil organic carbon (SOC) content over different climatic regions will enhance our knowledge of carbon gains and losses due to climatic change. However, little is known about the SOC content in the contrasting arid and sub-humid regions of Iran, whose complex SOC–landscape relationships pose a challenge to spatial analysis. Machine learning (ML) models with a digital soil mapping framework can solve such complex relationships. Current research focusses on ensemble ML models to increase the accuracy of prediction. The usual ensemble method is boosting or weighted averaging. This study proposes a novel ensemble technique: the stacking of multiple ML models through a meta-learning model. In addition, we tested the ensemble through rescanning the covariate space to maximize the prediction accuracy. We first applied six state-of-the-art ML models (i.e., Cubist, random forests (RF), extreme gradient boosting (XGBoost), classical artificial neural network models (ANN), neural network ensemble based on model averaging (AvNNet), and deep learning neural networks (DNN)) to predict and map the spatial distribution of SOC content at six soil depth intervals for both regions. In addition, the stacking of multiple ML models through a meta-learning model with/without rescanning the covariate space were tested and applied to maximize the prediction accuracy. Out of six ML models, the DNN resulted in the best modeling accuracies, followed by RF, XGBoost, AvNNet, ANN, and Cubist. Importantly, the stacking of models indicated a significant improvement in the prediction of SOC content, especially when combined with rescanning the covariate space. For instance, the RMSE values for SOC content prediction of the upper 0–5 cm of the soil profiles of the arid site and the sub-humid site by the proposed stacking approaches were 17% and 9% respectively, less than that obtained by the DNN models—the best individual model. This indicates that rescanning the original covariate space by a meta-learning model can extract more information and improve the SOC content prediction accuracy. Overall, our results suggest that the stacking of diverse sets of models could be used to more accurately estimate the spatial distribution of SOC content in different climatic regions.

The most important properties affecting the soil loss and runoff were investigated, and the effects of land use on the soil properties, together with the erodibility indices in a semiarid zone, central Iran, were evaluated. The locations of 100 positions were acquired by cLHS and 0-5-cm surface soil layer samples were used for laboratory analyses from the Borujen Region, Chaharmahal-Va-Bakhtiari Province, central Iran. To measure in situ runoff and soil erodibility of three different land uses comprising dryland, irrigated farming, and rangeland, a portable rainfall simulator was used. The results showed that the high variations (coefficient of variation, CV) were obtained for electrical conductivity (EC), mean weight diameter (MWD), soil organic carbon (SOC), and soil erodibility indices including runoff volume, soil loss, and sediment concentration (CV ~ 43.6-77.4%). Soil erodibility indices showed positive and significant correlations with bulk density and negative correlations with SOC, MWD, clay content, and soil shear strength in the area under investigation. The values of runoff in the dryland, irrigated farming, and rangeland were found 1.5, 28.9, and 58.7 cm; soil loss in the dryland, irrigated farming, and rangeland were observed 0.25, 2.96, and 76.8 g; and the amount of sediment concentration in the dryland, irrigated farming, and rangeland were found 0.01, 0.11, and 0.15 g cm. It is suggested that further investigations should be carried out on soil erodibility and the potential of sediment yield in various land uses with varying topography and soil properties in semiarid regions of Iran facing the high risk of soil loss.

No abstract

No abstract