Barley Straw Decomposition in the Field: A Comparison of Models

Abstract

Models were applied to results from a field experiment on the decomposition of barley straw, incubated 10—15 cm below the soil surface in a barley field. Litter bags were sampled 14 times during a 2—yr period to follow the dynamics of total mass and chemical components, e.g., water—solubles and total N. Zero— and first—order regression models were fitted to total mass, with and without adjustment for ambient temperature. R 2, adjusted for the number of model parameters, was used for comparisons of model fits. The first—order model showed a good fit (R 2 = 0.9875) if days with mean soil temperatures @<0°C were excluded. An improved fit was obtained using a temperature correction with Q 1 0 = 1.21 (R 2 = 0.9913). A one—compartment simulation model, using temperature and moisture as driving variables, showed a further improved fit (R 2 = 0.9952) and a best Q 1 0 = 1.78. Parallel and consecutive first—order models with two components did not improve the overall fit (R 2 = 0.9896), but the initial loss of water—solubles coincided well with the predicted initial loss from the labile fraction. To describe the dynamics of selected chemical components, a four—compartment simulation model, including decomposition product formation, was fitted to total mass, water—solubles and total nitrogen. The observed dynamics of these components were well reproduced by the model. Influences of experimental and statistical techniques on interpretations of model results are discussed.