Evaluation of modelling approaches for predicting the spatial distribution of soil organic carbon stocks at the national scale

TL;DR

This study compares boosted regression trees and geostatistical models for predicting soil organic carbon distribution at the national scale, highlighting when combining methods improves accuracy and when standalone models suffice.

Contribution

It evaluates the effectiveness of combining boosted regression trees with geostatistics for SOC prediction, providing guidance on model choice based on predictor variables and data characteristics.

Findings

Geostatistical residual modeling improves BRT predictions with limited predictors.

Standalone BRT models are adequate with multiple predictors and no local autocorrelation.

Model performance depends on predictor number and spatial autocorrelation presence.

Abstract

Soil organic carbon (SOC) plays a major role in the global carbon budget. It can act as a source or a sink of atmospheric carbon, thereby possibly influencing the course of climate change. Improving the tools that model the spatial distributions of SOC stocks at national scales is a priority, both for monitoring changes in SOC and as an input for global carbon cycles studies. In this paper, we compare and evaluate two recent and promising modelling approaches. First, we considered several increasingly complex boosted regression trees (BRT), a convenient and efficient multiple regression model from the statistical learning field. Further, we considered a robust geostatistical approach coupled to the BRT models. Testing the different approaches was performed on the dataset from the French Soil Monitoring Network, with a consistent cross-validation procedure. We showed that when a limited number of predictors were included in the BRT model, the standalone BRT predictions were significantly improved by robust geostatistical modelling of the residuals. However, when data for several SOC drivers were included, the standalone BRT model predictions were not significantly improved by geostatistical modelling. Therefore, in this latter situation, the BRT predictions might be considered adequate without the need for geostatistical modelling, provided that i) care is exercised in model fitting and validating, and ii) the dataset does not allow for modelling of local spatial autocorrelations, as is the case for many national systematic sampling schemes.