Modeling Soil Organic Carbon Turnover in Four Temperate Forests Based on Radiocarbon Measurements of Heterotrophic Respiration and Soil Organic Carbon

Soils of temperate forests store significant amounts of soil organic matter and are considered to be net sinks of atmospheric CO₂. Soil organic carbon (SOC) dynamics have been studied using the Δ¹⁴C signature of bulk SOC or different SOC fractions as observational constraints in SOC models. Further, the Δ¹⁴C signature of CO₂ evolved during the incubation of soil and roots has been widely used together with Δ¹⁴C of total soil respiration to partition soil respiration into heterotrophic respiration (Rh) and root respiration. However, these data have rarely been used together as observational constraints to determine SOC turnover times. Here, we present a multiple constraints approach, where we used SOC stock and its Δ¹⁴C signature, and heterotrophic respiration and its Δ¹⁴C signature to estimate SOC turnover times of a simple serial two-pool model via Bayesian optimization. We used data from four temperate forest ecosystems in Germany and the USA with different disturbance and management histories from selective logging to afforestation in the late 19th and early 20th century. The Δ¹⁴C signature of the atmosphere with its prominent bomb peak was used as a proxy for the Δ¹⁴C signature of aboveground and belowground litterfall. The Δ¹⁴C signature of litterfall was lagged behind the atmospheric signal to account for the period between photosynthetic fixation of carbon and its addition to SOC pools. We showed that the combined use of Δ¹⁴C measurements of Rh and SOC stocks helped to better constrain turnover times of the fast pool (primarily by Δ¹⁴C of Rh) and the slow pool (primarily by Δ¹⁴C of SOC). In particular, by introducing two additional parameters that describe the deviation from steady state of the fast and slow cycling pool for both SOC and SO¹⁴C, we were able to demonstrate that we cannot maintain the often used steady-state assumption of SOC models in general. Furthermore, a new transport version of our model, including SOC transport via bioturbation and with the liquid phase as dissolved organic carbon, is able to make full use of ¹⁴C measurements throughout a soil profile. Here, the additional parameters that describe a potential deviation from steady state are allowed to vary with depth. The transport version helped to disentangle the relative importance of advection, bioturbation and root litterfall as carbon inputs to different soil horizons. Overall our results suggest that using Δ¹⁴C data from more than one carbon pool or flux helps to better constrain SOC models. In addition the modeling approach might be able to better design future measurement campaigns with regard to sampling depth and time because the transport version shows how the bomb peak propagates through the soil profile with time.