A Model Intercomparison Analysis for Controls on C Accumulation in North American Peatlands

Abstract

Peatland biogeochemical processes have not been adequately represented in existing earth system models, which might have biased the quantification of Arctic carbon‐climate feedbacks. We revise the Peatland Terrestrial Ecosystem Model (PTEM) by incorporating additional peatland biogeochemical processes. The revised PTEM is evaluated by comparing with Holocene Peatland Model (HPM) in simulating peat physical and biogeochemical dynamics in three North American peatlands: a permafrost‐free fen site, a permafrost‐free bog site and a permafrost bog site. Peatland carbon dynamics are simulated from peat initiation to 1990 and then to year 2300. Model responses to the changes in temperature and precipitation are analyzed to identify key processes affecting peatland carbon accumulation rates. We find that the net C balance is sensitive to water table depth and nutrient availability. Future simulations to year 2300 are conducted with both models under RCP 2.6, RCP 4.5, and RCP 8.5. PTEM predicts these peatlands to be C sources or weaker C sinks when insufficient precipitation suppresses soil moisture and thereby net N mineralization and net primary production, while HPM predicts the same when drier climate leads to increasing water table depth. Our results highlight the importance of water balance and C‐N feedback on peatland C dynamics. With a warmer climate, these peatlands could become a weaker C sink or a source under drier conditions, otherwise a larger C sink if wetter. Improved understanding to peatland processes can help future quantification of peatland C dynamics in the boreal and Arctic regions.

Modeling Carbon Accumulation and Permafrost Dynamics of Northern Peatlands Since the Holocene

“…In this study, a process-based model, the Peatland Terrestrial Ecosystem Model (PTEM), designed to simulate peatland ecosystem dynamics, is used. Although PTEM has been applied to various northern peatland sites and the simulated C accumulation processes is consistent with the core data (Zhao, Zhuang, Treat, & Frolking, 2022), simulating northern peatlands dynamics and permafrost conditions in the Holocene at the pan-Arctic scale with the existing version still has a couple of challenges. First is the current peatland extent is highly uncertain.…”

“…In PTEM 2.0, we assume that, when the soil is acidic, the methane production rate increases with pH values. Soil pH declines gradually from 6.5 to 4.2 as fens transition to bogs, and the decrease in methane production is usually simultaneous (Zhao, Zhuang, Treat, et al., 2022). The PTEM 2.0 fen‐bog transition at the site‐level happens when peat thickness exceeds a certain threshold determined by peat core profiles.…”

“…There are two reasons that the correlation with Hugelius et al (2020) is higher than that with the other two data sets. First, the run-on parameters were calculated based on the peat thickness data in Hugelius et al (2020), and run-on influences WTD and thereby decomposition in PTEM (Zhao, Zhuang, Treat, et al, 2022). Second, the soil C in Hugelius et al ( 2020) was derived by a machine learning approach and the training samples include data from thousands of peat cores.…”

“…Finally, the existing PTEM 2.0 is designed for site‐level simulation. With no run‐on and runoff considered, the variation of water and nutrient status (i.e., fen/bog status) were identified from exiting peat cores (Zhao, Zhuang, Treat, et al., 2022). There is no core information for most of the pan‐Arctic region.…”

“…There is no core information for most of the pan‐Arctic region. We thus address this issue by developing a new version PTEM 2.1 considering grid‐specific water run‐on and run‐off, and their impacts on peatland nutrient supply in PTEM 2.0 (Zhao, Zhuang, Treat, et al., 2022).…”