Modelling soil organic carbon turnover in improved fallows in eastern Zambia using the RothC-26.3 model

Abstract

Scarcity of simple and reliable methods of estimating soil organic carbon (SOC) turnover and lack of data from long-term experiments make it difficult to estimate attainable soil C sequestration in tropical improved fallows. Testing and validating existing and widely used SOC models would help to determine attainable C storage in fallows. The Rothamsted C (RothC) model, therefore, was tested using empirical data from improved fallows at Msekera in eastern Zambia. This study (i) determined the effects of nitrogen fixing tree (NFT) species on aboveground organic C inputs to the soil and SOC stocks, (ii) estimated annual net organic C inputs to the soil using the RothC, and (iii) tested the performance of RothC model using empirical data from improved fallows. Soil samples (0–20 cm) were collected from coppicing and non-coppicing fallow experiments in October 2002 for determination of SOC by LECO CHN-1000 analyser. Data on surface litter, maize and weed biomasses, and on weather, were supplied by the Zambia/ICRAF Agroforestry Project. Measured SOC stocks to 20 cm depth ranged from 32.2 to 37.8 t ha −1 in coppicing fallows and 29.5 to 30.1 t ha −1 in non-coppicing fallows compared to 22.2–26.2 t ha −1 in maize monoculture systems. Coppicing fallows accumulated more SOC (680–1150 g m −2 year −1) than non-coppicing fallows (410–789 g m −2 year −1). While treatments with NFTs accumulated more SOC than NFT-free systems, SOC stocks increased with increasing tree biomass production and tree rotation. For food security and C sequestration, coppicing fallows are a potentially viable option. The RothC-26.3 model calculates the effect of annual above- and below-ground plant residue inputs to the soil on total organic C, microbial biomass, and radiocarbon age of the soil over a period ranging from a few years to centuries. As plant residue inputs from roots during plant growth are rarely known, the model is most often run in ‘inverse’ mode to generate total annual plant residue inputs from known soil, site, and weather data. The model, run in reverse, estimated the annual net organic C inputs required to maintain SOC stocks. Estimates ranged from 2.8 to 6.1 t ha −1 in coppicing fallows, 2.2–5.7 t ha −1 in non-coppicing fallows, and from 1.4 to 2.7 t ha −1 in controls. Modelled inputs comprising above- and below-ground organic residues in fallows were 12–104% greater than measured above-ground inputs alone. The model provided a good fit to empirical SOC data in fertilized maize monoculture, and in coppicing and non-coppicing fallows. Modelled inputs for Leucaena, Gliricidia, Senna, Sesbania, and Cajanus closely matched plant C input values estimated in separate studies, suggesting that RothC is giving reasonable simulations of soil C changes under improved fallow conditions in Zambia. However, the DPM/RPM ratio for plant C inputs in fallows was increased from 0.25 to 1.10 to suit their biodegradability characteristics. The RothC model can be used to calculate annual organic C inputs and SOC stocks in improved fallows provided suitable DPM:RPM ratios are used.