Carbon sequestration efficiency in paddy soil and upland soil under long-term fertilization in southern China

Soil organic carbon (SOC) stock can be improved through the return of crop residues. However, the efficiency of C sequestered in soil (i.e., ΔSOC/ΔC input) might differ among crop systems. In this paper, we investigated the C input and SOC stabilization in paddy soil and upland soil under different long-term fertilization practices. Our objectives were to determine (i) the response of SOC stock to C input under different fertilization practices, and (ii) C sequestration efficiency in the two contrasting agroecosystems. The long-term fertilization experiment in paddy soil started in 1981, while the adjacent upland soil experiment commenced in 1986. Each experiment consists of 9 treatments: CK (no fertilization), N, P, K, NP, NK, NPK, 2NPK (double dose), and NPKOM (NPK plus organic manure). Physical SOC fractions (cPOM, silt + clay_f, fPOM, iPOM_m, silt + clay_m) were isolated by sieving, dispersion, and density flotation. Fertilization increased crop yield and C input, but it did not change the quality of SOC as revealed by CPMAS-13C NMR spectra. During the period of the experiment, SOC stock was improved by 6.7–15.3 Mg ha−1 in paddy soil for all fertilization practices, while in upland soil the CK, N, P, K, and NP fertilizations reduced the SOC by 1.2–3.8 Mg C ha−1 and the other four fertilizations increased it by 0.5–7.4 Mg C ha−1. The change in SOC was mainly ascribed to the POM fraction, whereas the two silt + clay sized fractions were independent of fertilization practice except the NPKOM treatment. At a given C input, the C sequestration efficiency was greater in paddy soil than in upland soil, which may be attributed to lower microbial activity but greater chemical (i.e., oxalate-soluble Fe) and physical stabilizations (i.e., soil structure) in paddy field. Our results indicate that paddy soil may sequester more SOC, with higher efficiency, than upland soil does.