Simulation of carbon pool changes in woodlots in eastern Zambia using the CO2FIX model

Agroforestry systems have the potential to contribute significantly to climate change mitigation and adaptation. However, data on tree and soil organic carbon (SOC) pools for most agroforestry systems are lacking because reliable methods for estimating ecosystem carbon (C) pools are scarce. This study quantified the effects of five Leucaena species (L. leucocephala, L. macrophylla, L. diversifolia, L. collinsii and L. pulverulenta) on vegetal and soil C stocks and on mean annual increment (MAI) in aboveground tree C stocks. Specifically, it tested the validity of the CO2FIX model using empirical data from 7?year-old woodlots at Msekera, Zambia, and assessed the impact of converting a degraded agricultural ecosystem to woodlots on C stocks. Measured above- and below-ground tree C stocks and MAI of aboveground biomass differed significantly among the Leucaena species. Measured stem and total aboveground tree C stocks in seven-year old woodlots ranged from 17.1 to 29.2 and from 24.5 to 55.9?Mg?ha^?1, respectively. Measured SOC stocks at 0?C200?cm depth in Leucaena stands ranged from 106.9 (L. diversifolia) to 186.0?Mg?ha^?1 (L. leucocephala). Modeled stem and branch C stocks closely matched measured stocks, but the soil module of CO2FIX did not predict the soil C. The soil C data are inconclusive at this stage. We recommend that a fractionation and a soil aggregate hierarchy study backed by C dating is carried out to explain soil C dynamics in these soils. However, the model can be used only for estimating changes in aboveground tree C stocks in woodlots until soil C module is proven to predict SOC stocks.