Geomorphometric segmentation of complex slope elements for detailed digital soil mapping in southeast Brazil

Abstract

Hillslope elements have considerable potential in predicting soil properties and types in the landscape, making them likely to be a useful basis for detailed soil mapping. The goal of this research was to apply a previously developed digital hillslope position (DHP) model, calibrate it as needed to a Brazilian landscape, and test its utility as a basis for identification of detailed soil map units. The study area covers 2500 ha and is located on the border between the municipalities of Piracicaba and Santa Bárbara d'Oeste, São Paulo state, Brazil. A digital elevation model, with spatial resolution of 5 m, was used to obtain slope gradient, profile curvature and relative elevation with different analysis scales. Hierarchical rules for these digital terrain derivatives were used to segment the landscape into hillslope positions. The user-calibrated hillslope position model was verified against local experience by identifying the hillslope position in the field and comparing it with the model classification using the Kappa statistic and a confusion matrix. Soil samples were collected across multiple hillslopes with different lithologies. The samples were analyzed for chemical composition and soil particle size separates. The measured soil properties were assessed for statistical significance by variance analysis among hillslope position, parent material, and the interaction between the two. Student's t-tests were performed iteratively across each hillslope position within a given parent material to identify specifically which soil properties were significantly different among the hillslope position map units. Variance analysis of soil samples located within the respective parent material map units identified significant differences for all soil properties measured, but only for some soil properties when categorized by DHP. Focusing on the parent material with a sufficient quantity of samples, there was always at least one hillslope position that was significantly different from the others for each soil property. Because each of these map units presented a significant difference in at least one soil property, they are useful for detailed soil mapping.

A Relief Dependent Evaluation of Digital Elevation Models on Different Scales for Northern Chile

“…Furthermore, in the field of hydrology DEMs are required for stream network analysis [16,17] and ground-water flow modelling [18]. Terrain features derived by DEMs are also used as a predictor for digital soil mapping [19][20][21]. Additionally, DEMs are crucial for ecological analysis, such as vegetation and plant distribution research [22][23][24].…”

The Influence of Slope Positions on the Recovery Response of Compacted Soil Properties and Enzyme Activity in an Oriental Beech Stand in the Hyrcanian Forests, Iran

“…Because of this relocation of fine particles, the highest sand content was observed in the S treatments, while the highest contents of silt and clay particles were found in the TS treatment. Reasons for this have been described by Marques et al [45], stating that the S treatment was exposed more to the erosion processes than the TS and BS treatments, which concentrated the sand particles following removal of finer particles.…”

Methods of verification of soils prediction maps: a case study from Chernivtsi region, Ukraine

“…Then it allows getting precisely those proportions between the training pixels and the total sample size, which is conditioned by the conditions of the planned experiment, for example, 5%, 10%, 15% or other needed proportions. The implementation of such calculations is done using the written script on Python in the GRASS GIS environment (Cherlinka 2017c To evaluate the quality of the model, the Cohen's kappa index was used (Landis and Koch 1977, Foody 2004, Li and Zhang 2007, Grinand et al 2008, Congalton and Green 2008, Kuhn 2008, Hengl 2009, Malone et al 2016, Marques al. 2018, because kappa statistics is a common measure of classification accuracy.…”

“…The general idea underlying the application of such methods is using the reference points of the landscapes with linked taxa associated with them (Lagacherie et al 2001). Digital elevation model (DEM) is the main source of predictors in such a simulation as many geomorphometric parameters can be derived of the DEM and used as proxies of soil parameters (Kempen et al 2009, Hengl et al 2017, Marques al. 2018).…”