Using 50 years of soil radiocarbon data to identify optimal approaches for estimating soil carbon residence times

In 1959, Athol Rafter began a substantial programme of systematically monitoring the flow of C-14 produced by atmospheric thermonuclear tests through organic matter in New Zealand soils under stable land use. A database of 500 soil radiocarbon measurements spanning 50 years has now been compiled, and is used here to identify optimal approaches for soil C-cycle studies. Our results confirm the potential of C-14 to determine residence times, by estimating the amount of 'bomb C-14' incorporated. High-resolution time series confirm this approach is appropriate, and emphasise that residence times can be calculated routinely with two or more time points as little as 10 years apart. This approach is generally robust to the key assumptions that can create large errors when single time-point C-14 measurements are modelled. The three most critical assumptions relate to: (1) the distribution of turnover times, and particularly the proportion of old C ('passive fraction'), (2) the lag time between photosynthesis and C entering the modelled pool, (3) changes in the rates of C input. When carrying out approaches using robust assumptions on time-series samples, multiple soil layers can be aggregated using a mixing equation. Where good archived samples are available, AMS measurements can develop useful understanding for calibrating models of the soil C cycle at regional to continental scales with sample numbers on the order of hundreds rather than thousands. Sample preparation laboratories and AMS facilities can play an important role in coordinating the efficient delivery of robust calculated residence times for soil carbon. (C) 2012 Elsevier B.V. All rights reserved.