Nonparametric Techniques for Predicting Soil Bulk Density of Tropical Rainforest Topsoils in Rwanda

A study on Nyungwe National Park, Rwanda, reveals that organic carbon (OC) content alone is a robust predictor of soil bulk density (BD) in tropical rainforest topsoils. Using nonparametric techniques—k-nearest neighbor (k-NN) and boosted regression tree (BRT)—researchers found these methods performed comparably to traditional pedotransfer functions but outperformed those developed for non-tropical or non-forest soils. The findings highlight the critical role of OC data in soil fertility assessments, yet underscore its scarcity in African tropical forests. High-throughput infrared spectroscopy is proposed as a solution to fill this data gap. This approach challenges conventional assumptions about BD prediction and offers a scalable method for biogeochemical modeling.

Ghehi, N. G., Nemes, A., Verdoodt, A., Van Ranst, E., Cornelis, W. M., & Boeckx, P. (2012). Nonparametric Techniques for Predicting Soil Bulk Density of Tropical Rainforest Topsoils in Rwanda. Soil Science Society of America Journal, 76(4), 1172–1183. Portico. https://doi.org/10.2136/sssaj2011.0330