Peatlands contain one-third of soil carbon (C), mostly buried in deep, saturated anoxic zones (catotelm). The response of catotelm C to climate forcing is uncertain, because prior experiments have focused on surface warming. We show that deep peat heating of a 2 m-thick peat column results in an exponential increase in CH4 emissions. However, this response is due solely to surface processes and not degradation of catotelm peat. Incubations show that only the top 20–30 cm of peat from experimental plots have higher CH4 production rates at elevated temperatures. Radiocarbon analyses demonstrate that CH4 and CO2 are produced primarily from decomposition of surface-derived modern photosynthate, not catotelm C. There are no differences in microbial abundances, dissolved organic matter concentrations or degradative enzyme activities among treatments. These results suggest that although surface peat will respond to increasing temperature, the large reservoir of catotelm C is stable under current anoxic conditions.

No abstract

Abstract. Radiocarbon is a critical constraint on our estimates of the timescales of soil carbon cycling that can aid in identifying mechanisms of carbon stabilization and destabilization and improve the forecast of soil carbon response to management or environmental change. Despite the wealth of soil radiocarbon data that have been reported over the past 75 years, the ability to apply these data to global-scale questions is limited by our capacity to synthesize and compare measurements generated using a variety of methods. Here, we present the International Soil Radiocarbon Database (ISRaD; http://soilradiocarbon.org, last access: 16 December 2019), an open-source archive of soil data that include reported measurements from bulk soils, distinct soil carbon pools isolated in the laboratory by a variety of soil fractionation methods, samples of soil gas or water collected interstitially from within an intact soil profile, CO2 gas isolated from laboratory soil incubations, and fluxes collected in situ from a soil profile. The core of ISRaD is a relational database structured around individual datasets (entries) and organized hierarchically to report soil radiocarbon data, measured at different physical and temporal scales as well as other soil or environmental properties that may also be measured and may assist with interpretation and context. Anyone may contribute their own data to the database by entering it into the ISRaD template and subjecting it to quality assurance protocols. ISRaD can be accessed through (1) a web-based interface, (2) an R package (ISRaD), or (3) direct access to code and data through the GitHub repository, which hosts both code and data. The design of ISRaD allows for participants to become directly involved in the management, design, and application of ISRaD data. The synthesized dataset is available in two forms: the original data as reported by the authors of the datasets and an enhanced dataset that includes ancillary geospatial data calculated within the ISRaD framework. ISRaD also provides data management tools in the ISRaD-R package that provide a starting point for data analysis; as an open-source project, the broader soil community is invited and encouraged to add data, tools, and ideas for improvement. As a whole, ISRaD provides resources to aid our evaluation of soil dynamics across a range of spatial and temporal scales. The ISRaD v1.0 dataset is archived and freely available at https://doi.org/10.5281/zenodo.2613911 (Lawrence et al., 2019).

We are conducting a large-scale, long-term climate change response experiment in an ombrotrophic peat bog in Minnesota to evaluate the effects of warming and elevated CO 2 on ecosystem processes using empirical and modeling approaches. To better frame future assessments of peatland responses to climate change, we characterized and compared spatial vs. temporal variation in measured C cycle processes and their environmental drivers. We also conducted a sensitivity analysis of a peatland C model to identify how variation in ecosystem parameters contributes to model prediction uncertainty. High spatial variability in C cycle processes resulted in the inability to determine if the bog was a C source or sink, as the 95% confidence interval ranged from a source of 50 g C m -2 yr -1 to a sink of 67 g C m -2 yr -1 . Model sensitivity analysis also identified that spatial variation in tree and shrub photosynthesis, allocation characteristics, and maintenance respiration all contributed to large variations in the pretreatment estimates of net C balance. Variation in ecosystem processes can be more thoroughly characterized if more measurements are collected for parameters that are highly variable over space and time, and especially if those measurements encompass environmental gradients that may be driving the spatial and temporal variation (e.g., hummock vs. hollow microtopographies, and wet vs. dry years). Together, the coupled modeling and empirical approaches indicate that variability in C cycle processes and their drivers must be taken into account when interpreting the significance of experimental warming and elevated CO 2 treatments. Core Ideas• We compared spatial vs. temporal variation in C cycle processes and drivers in a bog.• The bog was indistinguishable as a C source or sink because of high spatial variation.• Sensitive C cycle parameters in the model differed under ambient vs. warming scenarios.• Characterizing pretreatment variability is necessary when interpreting warming effects.Abbreviations: ELM_SPRUCE; SPRUCE-specific version of the Energy Exascale Earth System model; ANPP, aboveground net primary production; CI, confidence interval; [CO 2 ], CO 2 concentration; DOC, dissolved organic C; flnr, fraction of leaf N in ribulose-1,5-biphosphate carboxylase/

Forest soils represent a significant pool for carbon sequestration and storage, but the factors controlling soil carbon cycling are not well constrained. We compared soil carbon dynamics at five broadleaf forests in the Eastern US that vary in climate, soil type, and soil ecology: two sites at the University of Michigan Biological Station (MI-Coarse, sandy; MI-Fine, loamy); Bartlett Experimental Forest (NH-BF); Harvard Forest (MA-HF); and Baskett Wildlife Recreation and Education Area (MO-OZ). We quantified soil carbon stocks and measured bulk soil radiocarbon to at least 60 cm depth. We determined surface (0-15 cm) soil carbon distribution and turnover times in free light (unprotected), occluded light (intra-aggregate), and dense (mineral-associated) soil fractions. Total soil carbon stocks ranged from 55 ± 4 to 229 ± 42 Mg C ha -1 and were lowest at MI-Coarse and MO-OZ and highest at MI-Fine and NH-BF. Differences in climate only partly explained differences in soil organic matter 14 C and mean turnover times, which were 75-260 year for free-light fractions, 70-625 year for occluded-light fractions, and 90-480 year for dense fractions. Turnover times were shortest at the warmest site, but longest at the northeastern sites (NH-BF and MA-HF), rather than the coldest sites (MI-Coarse and MI-Fine). Soil texture, mineralogy, drainage, and macrofaunal activity may be at least as important as climate in determining soil carbon dynamics in temperate broadleaf forests.