Abstract

CO2 emissions from preindustrial land use change (LUC) are subject to large uncertainties with model-based estimates ranging from 60 to 360 GtC (Olofsson and Hickler, 2008; Pongratz et al., 2009; Kaplan et al., 2011; Stocker et al., 2011). Thus, early anthropogenic impacts rose to significance between 7-3 kyr BP depending on reconstruction and may have altered the natural carbon© cycle and climate states to a degree that would lend support for the definition of a correspondingly early onset of the Anthropocene. However, the reconstructed parallel evolution of atmospheric CO2 and its 13C-signature indicate only 36+/-37 Gt loss of terrestrial C during the last 5 millennia (Elsig et al., 2009). It has been argued that this is the result of compensating effects of large LUC emissions and C sequestration in northern peatlands, which is estimated to be on the same order as upper-end estimates of preindustrial LUC (Ruddiman and Ellis, 2009).

Here, we combine updated observation-based and model-based reconstructions of peat C buildup (∆Cpeat) and model-based LUC emission estimates for a range of recently published reconstructions (Kaplan et al., 2009; Klein Goldewijk and Verburg, 2013) and accounting for changing land management regimes over time and space. Using the independent constraint on the total terrestrial C budget from ice core measurements of CO2 and d13C (∆Ctot), we assess the compatibility of different LUC scenarios with ∆Ctot and ∆Cpeat.

This reveals that large LUC emissions required to explain the observed CO2 rise between 7 and 5 kyr BP cannot be reconciled with ∆Ctot and ∆Cpeat unless a large additional terrestrial sink is invoked. Furthermore, this analysis points to the importance of other, non-anthropogenic impacts for explaining the ~150 Gt terrestrial C source between 5 and 2 kyr BP, where scenarios suggest emissions of only 20-50 GtC. More highly resolved ice core (Bauska et al., 2015) and peat C balance data (Charman et al., 2013) covering the last millennium further reveals that only extreme assumptions on the extent of post-Columbian reforestation in the Americas can close the C budget between 1500 and 1650 CE and that upper-end scenarios of preindustrial LUC are incompatible with the C budget between 1760 and 1920 CE.