Estimation of the soil organic carbon (SOC) content is of utmost importance in understanding the chemical, physical, and biological functions of the soil. This study proposes machine learning algorithms of support vector machines (SVM), artificial neural networks (ANN), regression tree, random forest (RF), extreme gradient boosting (XGBoost), and conventional deep neural network (DNN) for advancing prediction models of SOC. Models are trained with 1879 composite surface soil samples, and 105 auxiliary data as predictors. The genetic algorithm is used as a feature selection approach to identify effective variables. The results indicate that precipitation is the most important predictor driving 14.9% of SOC spatial variability followed by the normalized difference vegetation index (12.5%), day temperature index of moderate resolution imaging spectroradiometer (10.6%), multiresolution valley bottom flatness (8.7%) and land use (8.2%), respectively. Based on 10-fold cross-validation, the DNN model reported as a superior algorithm with the lowest prediction error and uncertainty. In terms of accuracy, DNN yielded a mean absolute error of 0.59%, a root mean squared error of 0.75%, a coefficient of determination of 0.65, and Lin’s concordance correlation coefficient of 0.83. The SOC content was the highest in udic soil moisture regime class with mean values of 3.71%, followed by the aquic (2.45%) and xeric (2.10%) classes, respectively. Soils in dense forestlands had the highest SOC contents, whereas soils of younger geological age and alluvial fans had lower SOC. The proposed DNN (hidden layers = 7, and size = 50) is a promising algorithm for handling large numbers of auxiliary data at a province-scale, and due to its flexible structure and the ability to extract more information from the auxiliary data surrounding the sampled observations, it had high accuracy for the prediction of the SOC base-line map and minimal uncertainty.

Soil texture and particle size fractions (PSFs) are a critical characteristic of soil that influences most physical, chemical, and biological properties of soil; furthermore, reliable spatial predictions of PSFs are crucial for agro-ecological modeling. Here, series of hybridized artificial neural network (ANN) models with bio-inspired metaheuristic optimization algorithms such as a genetic algorithm (GA-ANN), particle swarm optimization (PSO-ANN), bat (BAT-ANN), and monarch butterfly optimization (MBO-ANN) algorithms, were built for predicting PSFs for the Mazandaran Province of northern Iran. In total, 1595 composite surficial soil samples were collected, and 64 environmental covariates derived from terrain, climatic, remotely sensed, and categorical datasets were used as predictors. Models were tested using a repeated 10-fold nested cross-validation approach. The results indicate that the hybridized ANN methods were far superior to the reference approach using ANN with a backpropagation training algorithm (BP-ANN). Furthermore, the MBO-ANN approach was consistently determined to be the best approach and yielded the lowest error and uncertainty. The MBO-ANN model improved the predictions in terms of RMSE by 20% for clay, 10% for silt, and 24% for sand when compared to BP-ANN. The physiographical units, soil types, geology maps, rainfall, and temperature were the most important predictors of PSFs, followed by the terrain and remotely sensed data. This study demonstrates the effectiveness of bio-inspired algorithms for improving ANN models. The outputs of this study will support and inform sustainable soil management practices, agro-ecological modeling, and hydrological modeling for the Mazandaran Province of Iran.

Citrus spp. are one of the most important commercial crops with global marketing potential in the world, as in Iran. A soil management zone (MZ) as an appropriate approach is necessary to achieve sustainable production, along with improving soil management and increasing economic benefits in the commercial citrus plantations of northern Iran. As the first report, the biological and terrain attributes along with the physicochemical properties (57 soil samples, 0–30 cm) were used for MZ delineation using the integration of principal component analysis (PCA) and the fuzzy c-means clustering methods. An economic analysis based on the MZ results was also performed to determine the changes in each MZ using a relative cost (RC) value. The high correlation between soil properties and terrain attributes and the considerable spatial variation of these factors in the study area call for site-specific nutrient management. The optimal number of MZs was six and there was a significant heterogeneity variation among different MZs. The ranking of the MZs were MZ5 > MZ2 > MZ6 > MZ1 > MZ3 > MZ4 based on higher soil quality and lower costs per tree. The MZ4, MZ3, MZ1, MZ6, and MZ2 required 34.4, 30.6, 29.4, 9.77, and 9.44% more costs than MZ5 (as reference MZ) for achieving similar productivity, respectively. Therefore, this simple and cost-effective approach could be an initial step to utilize fertilizers site-specifically for data-scarce areas and reduce the soil property variability within the delineated MZs, which is fundamental for precision agriculture management.

Soil provides a key interface between the atmosphere and the lithosphere and plays an important role in food production, ecosystem services, and biodiversity. Recently, demands for applying machine learning (ML) methods to improve the knowledge and understanding of soil behavior have increased. While real-world datasets are inherently imbalanced, ML models overestimate the majority classes and underestimate the minority ones. The aim of this study was to investigate the effects of imbalance in training data on the performance of a random forest model (RF). The original dataset (imbalanced) included 6100 soil texture data from the surface layer of agricultural fields in northern Iran. A synthetic resampling approach using the synthetic minority oversampling technique (SMOTE) was employed to make a balanced dataset from the original data. Bioclimatic and remotely sensed data, distance, and terrain attributes were used as environmental covariates to model and map soil textural classes. Results showed that based on mean minimal depth (MMD), when imbalanced data was used, distance and annual mean precipitation were important, but when balanced data were employed, terrain attributes and remotely sensed data played a key role in predicting soil texture. Balanced data also improved the accuracies from 44% to 59% and 0.30 to 0.52 with regard to the overall accuracy and kappa values, respectively. Similar increasing trends were observed for the recall and F-scores. It is concluded that, in modeling soil texture classes using RF models through a digital soil mapping approach, data should be balanced before modeling.

Estimation of the soil organic carbon content is of utmost importance in understanding the chemical, physical, and biological functions of the soil. This study proposes machine learning algorithms of support vector machines, artificial neural networks, regression tree, random forest, extreme gradient boosting, and conventional deep neural network for advancing prediction models of SOC. Models are trained with 1879 composite surface soil samples, and 105 auxiliary data as predictors. The genetic algorithm is used as a feature selection approach to identify effective variables. The results indicate that precipitation is the most important predictor driving 15 percent of SOC spatial variability followed by the normalized difference vegetation index, day temperature index of moderate resolution imaging spectroradiometer, multiresolution valley bottom flatness and land use, respectively. Based on 10 fold cross validation, the DNN model reported as a superior algorithm with the lowest prediction error and uncertainty. In terms of accuracy, DNN yielded a mean absolute error of 59 percent, a root mean squared error of 75 percent, a coefficient of determination of 0.65, and Lins concordance correlation coefficient of 0.83. The SOC content was the highest in udic soil moisture regime class with mean values of 4 percent, followed by the aquic and xeric classes, respectively. Soils in dense forestlands had the highest SOC contents, whereas soils of younger geological age and alluvial fans had lower SOC. The proposed DNN is a promising algorithm for handling large numbers of auxiliary data at a province scale, and due to its flexible structure and the ability to extract more information from the auxiliary data surrounding the sampled observations, it had high accuracy for the prediction of the SOC baseline map and minimal uncertainty.