Location: Dale Bumpers Small Farms Research CenterTitle: Tree-based techniques to predict soil units
|PINHEIRO, HELENA - Universidade Federal Do Rio De Janeiro|
|DOS ANJOS, LUCIA - Universidade Federal Do Rio De Janeiro|
|JUNIOR, WALDIR - Embrapa|
|CHAGAS, CESAR - Embrapa|
Submitted to: Soil Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/24/2017
Publication Date: 6/1/2017
Citation: Pinheiro, H.S., Owens, P.R., Dos Anjos, L.H., Junior, W.D., Chagas, C.D. 2017. Tree-based techniques to predict soil units. Soil Research. 55:788-798. https://doi.org/10.1071/SR16060.
Interpretive Summary: Many of the regions of the world lack soil maps and information to manage soil resourcces. This research focused on a method to create soil untis based on a mathamatecal approach using terrain as a way to predict soils. This method was found to be a useful tool to predict soil units which function in a similar manner. Future applications of this tool will allow land managers to make better decsions about land-use for long-term sustainability.
Technical Abstract: Quantitative soil-landscape models offer a method for conducting soil surveys that employs statistical tools to predict natural patterns in the occurrence of particular map units across a landscape. The goal of this work was to predict soil units in a watershed with wide variation in landscape conditions. The approach relied on a modeling of soil forming factors in order to understand the variability of the landscape components in the region. Models were generated for landscape attributes related to pedogenesis, specifically Elevation, Slope, Curvature, Compound Topographic Index, Euclidean Distance from Stream Networks, Landforms Map, Clay Minerals Index, Iron Oxide Index, and Normalized Difference Vegetation Index, along with an existing Geology Map. The soil classification was adapted from the World Reference Base System for Soil Resources; and the predominant soil taxonomic orders observed were Ferrasols, Acrisols, Gleysols, Cambisols, Fluvisols, and Regosols. The algorithms used to predict the soil units were based on Decision Tree (DT) and Random Forest (RF) methods. The criteria applied to evaluate the models’ performance were statistical indices, coherence between predicted units and the legacy map; and accuracy checks based on control samples. The best performing model was found to be the RF algorithm, with resulting statistical indices considered excellent (Overall= 0.966, Kappa = 0.962). The accuracy of the map as determined by control points was 67.89%, with a Kappa value of 61.39%.