Predicted long-term effects of decomposition of leaf litter from Pinus taeda, Eucalyptus cloeziana and deciduous miombo trees on soil carbon stocks

Abstract

Decomposition rate of leaf litter in native miombo and in Pinus taeda and Eucalyptus cloeziana plantations was measured in the Inhamacari forest research area, Manica province, Mozambique. The Q-model and a simple first-order exponential decomposition model (K-model) were used to predict long-term accumulation of carbon (C) in the litter layer and in soil 34 years after plantation establishment and at steady-state. Measured litterfall from the study plots was used to estimate litter inputs for prediction of C stocks. The results were compared with measured C stocks when stand age in the first-rotation plantation stands was approx. 34 years.At the end of the one-year decomposition study, about half or less of the initial litter mass remained. Leaf litter of the miombo species Brachystegia spiciformis decomposed faster than leaf litter of non-native Eucalyptus cloeziana and Pinus taeda, and a representative mixture of leaf litters of dominating miombo trees decomposed faster than P. taeda, but not E. cloeziana.Measurements indicated higher soil temperatures, and thus higher microbial growth rate, in miombo than in plantation stands. Estimates of initial leaf litter quality indicated higher decomposability of mixed litter substrates of miombo stands than P. taeda needles.Assuming current measured litterfall, the accumulated C stocks predicted by the Q-model were higher than measured C stocks in the litter layer. A major cause of the discrepancy between predicted and measured C stocks in the litter layer of miombo stands was probably frequent ground fires and probably also the influence of termites and other modes of transport into the soil profile.Previously measured increase in C stocks in the soil (0–50 cm) and the litter layer of first-rotation stands of P. taeda and E. cloeziana plantations were much larger than the Q-model predicted soil C stocks at steady-state, indicating that this increase cannot be explained by leaf litter input only. Keywords: Brachystegia spiciformis, Forest degradation, Litterfall, Mozambique, Q-model, Uapaca kirkiana