Testing CERES model predictions of N release from legume cover crop residue

Abstract

Nitrogen release into soil from a legume cover crop (LCC) is affected by soil biotic and abiotic factors, quality of the LCC incorporated, and management practices. This study was conducted to test the ability of the CERES soil-process submodels to predict N release from an LCC, prior to growth of a subsequent crop. Field data were collected from 1995 to 1997 following spring incorporation of a mixed stand of woollypod vetch and field pea either early (C/N = 10) or late (C/N = 17) and in unfertilized fallow control plots (no-LCC). Soil temperature and water content by layer were predicted reasonably well with a few exceptions, although temperature errors of even a few degrees may be biologically significant. Inorganic soil N content during the fallow period with no-LCC and after late LCC incorporation was also realistically predicted. With early LCC incorporation, however, the N-transformation submodel substantially underestimated flushes in inorganic soil N immediately after rainfall in 1996 and irrigation in 1997. Second, the nitrification capacity factor, which should simulate a lag between a rapid increase in ammonium concentrations and slower nitrate increase after incorporation of LCC residues, performed poorly. Simulated N dynamics after incorporation of the maize residues (C/N ratio of 80) was also compared with field observations. A measured decrease in inorganic soil N contents was not correctly predicted. These discrepancies indicate the presence of some significant limitations in the soil-process submodels.