Digital Soil Mapping by Machine Learning Techniques

Articles in Press

Authors

1 MSc Student of Soil Science and engineering Department, Urmia University

2 Faculty member m Department of Soil Science, Urmia University

3 Shahid Bakeri High Education Center of Miandoab, Urmia University

4 Soil Science and engineering department, Urmia University

Abstract

Background and Objectives: The use of geospatial techniques for mapping soils is broadly covered by the term “digital soil mapping”. Soil maps have considerable significance as basic maps in many environmental and natural resources studies.
Methodology: According to semi-detailed soil, survey and using stratified random sampling method, 50 pedons and 50 augers with approximate distance of 1000 m were excavated, described and soil samples were taken from different genetic horizons. namely boosted regression tree (BRT), random forest (RF), artificial neural networks (ANNs) and multinomia l logistic regression (MLR) were used to test the predictive power for mapping the soil classes.
Findings: Results showed that the different models had the same ability for prediction of the soil classes across all taxonomic levels but a considerable decreasing trend was observed for their accuracy at subgroup and family levels. The terrain attributes were the most important auxiliary information to predict the soil classes up to the family level. It is noticeable that the artificial neural network model has a good accuracy up to the great group level through the acceptable level of overall accuracy (i.e., 75 %), hence it has a high degree of uncertainty.
Conclusion: Terrain attributes were the main predictors among different studied auxiliary information. The accuracy of the estimations with more observations is recommended to give a better understanding about the performance of DSM approach over low-relief areas.

Keywords

Main Subjects

Water and Soil Science

Articles in Press, Accepted Manuscript
Available Online from 04 August 2024

Files

Share

How to cite

Statistics