Dynamics of topsoil carbon and nitrogen along a tropical forest–cropland chronosequence: Evidence from stable isotope analysis and spectroscopy

Abstract

Analyses of δ13C and δ15N abundances and near infrared reflectance spectroscopy (NIRS) were used to evaluate the changes in SOC and total nitrogen (TN) content along a forest–cropland chronosequence in the margins of Kakamega forest in Kenya. 300 topsoil samples were collected from 50 paired forest–cropland plots cultivated between 17 and 60 years. Changes in δ13C values between forest and cropland soils were used to model the dynamics of forest-derived and maize-derived C. Mean residence times (MRTs) of forest-derived C in bulk topsoil samples was calculated based on changes in δ13C with time since cultivation. An ordinal soil fertility classification developed using NIRS was evaluated against SOC and TN concentrations and δ13C and δ15N abundances in the topsoil. Mean δ13C isotopic ratios increased from 24.3 ± 0.2‰ in forest to −16.3 ± 0.4‰ under cropland. Similarly, mean δ15N isotopic ratios increased from 5.9 ± 0.1‰ in the forest sites to 6.8 ± 0.1‰ in soils cultivated sites. δ13C and δ15N enrichment, low levels SOC and TN concentrations were observed in soils that were designated as low fertility based on ordinal fertility classes defined by NIRS. SOC content declined from 7.27 kg C m−2 in forest soils to 2.67 kg C m−2 in soils cultivated for 60 years. A nonlinear regression model predicted there was an accretion of maize-derived C4 carbon, attaining equilibrium at circa 4 kg C m−2 after circa 70 years. The model also predicted that after about 38 years, maize-derived C4 carbon was the predominant input-source of bulk topsoil SOC. The minimum, maximum and mean MRTs values for forest-derived SOC were 19, 149 and 60 years, respectively. The results of this study demonstrates that δ15N and δ13C values offer a robust and direct technique for stable isotope techniques for understanding the dynamics of SOC and TN after conversion of forest soils to cropland.