boosted regression tree

Digital soil mapping plays an important role in upgrading the knowledge of soil survey in line with the advances in the spatial data of infrastructure development. The main aim of this study was to provide a digital map of the soil family classes using the random forest (RF) models and boosting regression tree (BRT) in a semi-arid region of Ilam province. Environmental covariates were extracted from a digital elevation model with 30 m spatial resolution, using the SAGAGIS7.3 software. In this study area, 46 soil profiles were dug and sampled; after physico-chemical analysis, the soils were classified based on key to soil taxonomy (2014). In the studied area, three orders were recognized: Mollisols, Inceptisols, and Entisols. Based on the results of the environmental covariate data mining with variance inflation factor (VIF), some parameters including DEM, standard height and terrain ruggedness index were the most important variables. The best spatial prediction of soil classes belonged to Fine, carbonatic, thermic, Typic Haploxerolls. Also, the results showed that RF and BRT models had an overall accuracy and of 0.80, 0.64 and Kappa index 0.70, 0.55, respectively. Therefore, the RF method could serve as a reliable and accurate method to provide a reasonable prediction with a low sampling density.

This study was conducted in order to digital mapping soil classes according to WRB system was conducted by boosted regression tree (BRT) method on about 60.000 hectares of Hirmand county lands. 108 soil profiles were dug and soil profiles were sampled and described based on WRB system. Twenty environmental covariates were used as estimators for soil mapping including terrain attributes and remote sensing covariates. Results showed that the young soil covered the study area and mostly influenced by flood sediment which classified as Fluvisol and Cambisol groups and Solonchak group in salt -affected area. The variable importance showed that the environmental attributes such as Multi-resolution Valley Bottom Flatness Index (MrVBF), Valley Depth, Convergence Index, Catchment Area and Salinity Index (NDSI) had the highest importance among all covariates for two levels of WRB prediction. The validation results showed that the BRT model could predict WRB1 and WRB2 levels with overall accuracy of 47 and 25%, respectively, and also from WRB1 to WRB2 levels the accuracy decreased. In low relief area and young soil that the low soils variability, digital soil mapping approach could be useful, efficient, and fast technique to produce and predict soil classes map.

Digital soil mapping includes soils, spatial prediction and their properties based on the relationship with covariates. This study was designed for digital soil mapping using binary logistic regression and boosted regression tree in Zarand region of Kerman. A stratified sampling scheme was adopted for the 90,000 ha area based on which, 123 soil profiles were described. In both approaches, the occurrence of relevant diagnostic horizons was first mapped, and subsequently, various maps were combined for a pixel-wise classification by combining the presence or absence of diagnostic horizons. Covariates included a geomorphology map, terrain attributes and remote sensing indices. Among the predictors, geomorphology map was identified as an important tool for digital soil mapping approaches as it helped increase the prediction accuracy. After geomorphic surfaces, the terrain attributes were identified as the most effective auxiliary parameters in predicting the diagnostic horizons. The methods predicted high probability of salic horizon in playa landform, gypsic horizon in gypsiferous hills and calcic horizon in alluvial fans. Both models predicted Calcigypsids with very low reliability and accuracy, while prediction of Haplosalids and Haplogypsids was carried out with high accuracy.