Depth distribution of soil organic carbon in an Oxisol under different land uses: Stratification indices and multifractal analysis

- Soil organic carbon (SOC) was measured at 2 cm intervals until 130 cm depth.
- SOC profiles were obtained under no-tilled sugarcane (5 and 10 years) and native forest.
- Stratification ratios were highest under continuous cropping during 10 year.
- SOC profiles exhibited quasi-multifractal or multifractal behaviour.
- Highest scaling heterogeneity was for SOC profiles after 10 years untilled sugarcane.

Understanding the depth distributions of soil organic carbon (SOC) content in a soil profile has become increasingly important to examine land use effects on soil carbon sequestration and soil quality. We addressed the variability of SOC in deep soil profiles under different land uses using the multifractal approach. In addition, classical indices such as proportion of SOC stored at the upper soil horizon and stratification ratios (SR) were used to evaluate organic matter accumulation near to the soil surface. Six soil profiles belonging to a Rhodic Hapludox were sampled at two sites with different textures under three land uses: native forest (NF) and sugar cane managed as continuous cropping, during either 5 (SuCa 5) or 10 (SuCa 10) years. Sixty-five soil samples per profile were analyzed for SOC. Soil carbon storage at the surface horizons and stratification ratios were significantly higher (P < 0.05) under SuCa 10 than under SuCa 5 and NF. Vertical distributions of SOC exhibited various degrees of scaling heterogeneity, so that those under SuCa 10 clearly showed multifractal behavior, while those under SuCa 5 and NF behave as quasi monofractal. In all the SOC depth distributions, the amplitude of the Rényi and singularity spectra was higher for positive than for negative q moments. Multifractal parameters estimated for the most positive moments were negatively correlated to SR, while those for the most negative q moments were best correlated to the proportion of SOC stored at the soil surface. Our results suggest that single scale analysis may not be sufficient to fully characterize vertical variability of SOC content.