Permafrost—perennially frozen soil—covers 15% of the Northern Hemisphere. Temperatures below zero allow for the preservation and accumulation of organic matter (OM) within the soil. OM accumulation over millennia has resulted in permafrost storing the largest pool of terrestrial carbon (C) on Earth. As global temperatures rise, thawing permafrost exposes this stored OM to microbial decomposition, potentially leading to the release of carbon dioxide (CO2) and methane (CH4) into the atmosphere. As a result, permafrost could shift from a long-term C sink to a C source. Thawing permafrost drives significant shifts in landscape structure and site hydrology, which further influence C cycling, especially in ice-rich permafrost. To investigate these changes and future C releases, the incubation approach has proven particularly effective. It has helped identify several abiotic and biotic factors that regulate OM decomposition during permafrost thaw. However, most of these studies relied on simplified hydrological conditions, often neglecting important hydrochemical properties, such as localized redox dynamics, that are critical to understanding C releases. In my thesis, I investigated how changes in hydrology influence C cycling during permafrost thaw sequences. I focused on three sites where hydrology plays a considerable role in soil conditions: a permafrost peatland in Finland, two coastal permafrost features near Utqiaġvik, Alaska, and deltaic landforms in the Lena Delta, Russia. Specifically, I assessed (1) microbial–C interactions in palsa mires in response to inundation from abrupt thaw; (2) the C response to rapid hydrological changes across permafrost landscapes with different landscape histories; and (3) the representativeness of ex situ incubations for assessing C responses to hydrological changes during permafrost thaw. To address these research gaps, I incorporated site-specific in situ hydrological shifts from the three study areas and quantified CO2 and CH4 releases through multi-scale ex situ incubations (vial and mesocosm). To better replicate field conditions during permafrost thaw, I designed mesocosm incubations that preserve in situ soil structure while allowing for controlled laboratory conditions and continuous greenhouse gas (GHG) measurements. Finally, I combined GHG measurements with microbial analyses to link biogeochemical changes with microbial dynamics. The palsa thaw simulation revealed that CO2 emissions were enhanced under abrupt permafrost thaw, while no CH4 emissions were detected. In combination with microbial analyses, I attributed the increase in CO2 primarily to the mobilization of less degraded OM within the soil column and a general increase in microbial diversity, rather than a shift in metabolic pathways. Metagenomic results also revealed potential C–nitrogen cycle interactions, particularly denitrification, likely driven by increased soil moisture. In contrast, methanogenesis was strongly inhibited despite flooded conditions, likely due to alternative electron acceptors in permafrost soils and delayed anaerobic onset. Across three distinct landforms, the results of this thesis highlight that thaw-induced shifts in redox conditions more strongly influenced CH4 production, whereas CO2 responses were more sensitive to site-specific landscape history. Pre-thaw hydrological conditions shaped by landscape position influenced microbial communities, leading to high micro-scale variability in CH4 emissions in both the coastal permafrost and the Lena Delta. These patterns were further influenced by soil type, as OM degradation in the permafrost layer appeared to contribute less to total C release in peat soils than in mineral soils, likely due to limited gas diffusion under saturated conditions. Finally, I found that CH4 release from thawing permafrost was considerably lower in mesocosm incubations than in traditional vial incubations. I attribute this difference mainly to the closed system in vial incubations, which precludes the utilization of other available electron acceptors and CH4 oxidation within the soil column. Nevertheless, CH4 responses to redox changes, as simulated in this thesis, were consistent across both mesocosm and vial incubations. Conversely, CO2 responses to treatment were less consistent between scales. Mesocosm incubations appeared to better replicate in situ CO2 emissions, as they capture the relationship between CO2 release and OM quality across all soil horizons. Vial incubations, by contrast, rely on a single soil horizon, which may introduce bias and underestimate CO2 release, especially following flooding events. Collectively, these findings suggest that local hydrological shifts must be evaluated in the context of specific characteristics across permafrost landscapes. In a wetter Arctic, short-term CH4 emissions from permanently saturated soils underlain by ice-rich permafrost may be overestimated, while CO2 emissions may be underestimated. Conversely, permafrost areas subject to periodic flooding could rapidly adapt to saturated conditions and become significant CH4 sources, despite typically covering relatively small areas. These findings demonstrate that mesocosm incubations provide a robust approach for addressing critical research gaps, such as C dynamics under extreme precipitation events, and offer valuable potential for improving estimates of GHG emissions from permafrost soils under a warming climate.
Mélissa Laurent (Thu,) studied this question.