Carbon cycling of subsurface organic matter recorded in speleothem 14C records: Maximizing bomb-peak model fidelity

From 1955 to 1963, atmospheric testing of nuclear weapons caused a significant rise in atmospheric radiocarbon activity. This bomb peak has been used to calculate turnover rates of organic carbon in soils and other recent sedimentary deposits. Some speleothems contain precise and independently dated records of radiocarbon activity. These records can be used to understand, through inverse modeling, the processes and rates of turnover of subsurface organic carbon in karst regions. This approach is complicated, however, by the contribution of radioactively dead carbon to the stalagmite by the dissolution of host-rock limestone and/or by the respiration of relatively old soil organic matter. Previously published inverse models of the radiocarbon bomb peak in speleothems constrain the dead carbon proportion (DCP, in percent) by comparing measurements of speleothem radiocarbon activity from before the onset of the bomb peak to measurements of coeval atmospheric radiocarbon. This approach precludes modeling of speleothems that began growing after the onset of atmospheric nuclear weapons testing in 1955. Here, we advance the inverse modeling framework to calculate DCP using the entire length of the speleothem record, allowing for the modeling of speleothems that began growing after the initiation of atmospheric nuclear weapons testing. We test the sensitivity and resolution of this model and find that it can precisely resolve the turnover times and relative contributions of subsurface organic matter pools with residence times of less than a decade. The model fails to resolve turnover times or relative contributions of organic matter pools on millennial or greater timescales. These results also hold for the previously published models from which the current model is derived. We find that imprecise estimates of slow-turnover carbon add significant uncertainty to the calculated average age of respired carbon, which is a common metric of subsurface carbon cycling. The high precision and resolution attainable between sub-decadal carbon pools will allow researchers to differentiate the (sub-)annual pool, which is likely dominated by root/rhizosphere respiration, from the 2- to 10-year pools, which are likely dominated by microbial decomposition of labile organic carbon. The high precision attainable in fast-turnover pools also suggests that when there are multiple viable chronological interpretations for the same speleothem, bomb peak models could be used to help select which chronology is most likely to be accurate. This is important for high-resolution (sub-annual) speleothem climate records, where even single-year chronological offsets can result in misleading calibrations to the instrumental record. (C) 2018 Elsevier Ltd. All rights reserved.