Quantifying carbon sequestration in a conventionally tilled crop rotation study in southwestern Saskatchewan

Abstract

In this study we used results from 10 cropping systems in a 37-yr field experiment being conducted on a medium-textured Orthic Brown Chernozem in semiarid southwestern Saskatchewan, in which soil organic carbon (SOC) had been sampled in 7 different years, to quantify trends and changes in SOC in the 0- to 15-cm depth. We tested the effectiveness of three models: Century, the Introductory Carbon Balance Model (ICBM), and the Campbell model to predict the measured values. The 10 cropping systems allowed us to assess the influence of cropping frequency, fertilization and crop type on SOC and, because growing season weather was distinctly more humid in the final 12 yr of the study, we were able to assess the impact of weather. In this soil on which a fallow-wheat (Triticum aestivum L.) (F-W) rotation was maintained for the previous 60 yr, SOC remained fairly constant under normal weather for the first 20 yr of the study for the systems that were frequently fallowed, except for fallow-fall rye (Secale cereale L.)- wheat (F-Rye-W). In contrast, in the final 12 yr, SOC increased in all systems in response to increased C inputs from crop residues associated with improved precipitation. SOC gains were greater for well-fertilized extended crop rotations such as continuous wheat (Cont W) and wheat-lentil (Lens culinaris Medikus) (W-Lent) and the F-Rye-W systems receiving N and P than for the F-W, F-W-W, F-Flax (Linum usitatissimum L.)-W (F-Flx-W) receiving N and P, and Cont W receiving only P. SOC was also greater for well- fertilized than for poorly fertilized systems. The ICBM and Campbell models performed well in simulating SOC trends, partly because they used measured grain yields to estimate C inputs. However, the Century model was less effective in its simulation of SOC especially for the fallow-containing systems due to its difficulty in estimating spring soil water and crop yields. We showed how grain yields can be used, together with coefficients of conversion of C inputs from crop residues to SOC, to estimate SOC changes. Using these relationships, and assuming the coefficient of conversion of C inputs to SOC sequestered is 15% for well-fertilized extended rotations, or 10% for well-fertilized frequently fallowed spring-seeded systems, one can make reasonable first estimates of the impact of management on C sequestration in degraded soils of the semiarid prairies. Key words: ICBM model, Century model, Campbell model, soil organic C, N and P fertilizer, cropping frequency