Ensemble learning for spatial interpolation of soil potassium content based on environmental information

Area of Science:

• Soil Science
• Geostatistics
• Environmental Modeling

Background:

• Spatial interpolation is crucial for creating continuous soil property maps from discrete samples.
• Accurate soil property mapping requires integrating diverse environmental factors.

Purpose of the Study:

• To introduce and evaluate an ensemble learning spatial interpolation (EL-SP) method for soil properties.
• To improve the accuracy and detail of soil potassium content interpolation using ancillary environmental data.

Main Methods:

• Calculated trend values for soil potassium from measured data.
• Simulated residuals using an ensemble learning model incorporating soil type, geology, land use, and slope.
• Applied the EL-SP method for spatial interpolation of soil potassium content.

Main Results:

• The EL-SP method demonstrated lower mean absolute error and root mean square error than comparative methods.
• EL-SP maps provided more detailed local information and accurate spatial patterns of soil potassium.
• The method effectively captured abrupt boundary information for soil potassium content.

Conclusions:

• The EL-SP method enhances the accuracy and reasonableness of soil property spatial interpolation.
• Integrating ensemble learning with environmental information significantly improves interpolation results.
• EL-SP offers a valuable tool for detailed soil potassium content mapping and analysis.