Simulating trends in soil organic carbon in long-term experiments using the NCSOIL and NCSWAP models

Abstract

NCSOIL simulates C and N transformations in 4 soil organic pools: Pool I labile, Pool I resistant, Pool II, and Pool III, with half-lives of 2, 17, 115 days and about 150 years, respectively. Pool I labile and Pool I resistant represent the microbial biomass. Pool I and Pool II represent the potentially mineralizable N, or the biologically active soil organic matter. The sum of Pools I, Pool II, and Pool III — the soil organic matter - corresponds to the total organic matter minus residues. Each residue is described by 2 pools. NCSOIL is a stand-alone model. It is also a module of NCSWAP, a larger model which encompasses the soil-water-air-plant system. A number of systems and treatments, including the Rothamsted nitrate treatment and the Calhoun tracer C data were simulated. The initial level of Pool II and the decay rate constant of Pool III were calibrated on the basis of measured total soil organic matter and above-ground production. Simulated data were sensitive to above-ground production as it controlled residues input to soil. Model performance, based on total soil organic matter only, is discussed elsewhere in this issue. Most decay rate constants for Pool III ranged from 1.0E — 5 to 3.0E — 5 d −1. Rate constants outside this range were associated with peculiarities of the soil or agronomic practices. Levels of biologically active organic matter (Pool I plus Pool II) in the top soil layers ranged from 4 to 108 μg N g -1. They were consistent with those reported for the potentially mineralizable nitrogen and reflected the agronomic practice and soil fertility level better than did the total soil organic matter. The simulated biologically active organic matter presented a 1 year periodic cycle. In the future, a major challenge in modelling studies will be to free simulations from the calibration process and to devise experimental methods which will provide initial values relevant to the dynamic requirements of the model.