Post-Rewetting Carbon Dynamics of a Temperate Peatland: Eddy Covariance–Based CO2 and CH4 Fluxes from Himmelmoor, Germany

Peatland rewetting has been widely adopted in Germany as a climate-change mitigation strategy by reducing CO₂ emissions from peat decomposition. However, rewetted peatlands may simultaneously act as sources of CH₄ with varying strength. Quantifying their net greenhouse gas exchange remains challenging due to landscape heterogeneity and continuously changing vegetation, water table, and surface conditions following the rewetting. Therefore, sustained monitoring during and after the rewetting process is necessary.

In this study, we measured the fluxes of CO₂ and CH₄ at a rewetted peatland, Himmelmoor, in Northern Germany from 2015 to 2018, extending previous work conducted from 2012 to 2014. The site is characterized by heterogeneous terrain resulting from peat extraction and phased rewetting, which has been carried out alongside extraction activities since 2004. The objective is to provide updated greenhouse gas (GHG) estimates to assess how carbon dynamics have evolved in the rewetted section. Eddy covariance (EC) measurements of CO₂ and CH₄ fluxes were conducted at the center of the former mining area, allowing flux contributions from multiple surface classes, including rewetted area, vegetation strips, and bare soils to be captured within the source area of a single EC tower.

For CO₂ fluxes, we apply and compare a wind-direction-based mechanistic source partitioning approach with a footprint-based source partitioning method to estimate flux contributions from multiple surface classes within the EC footprint, including the rewetted area, vegetated peat strips, and bare (non-rewetted) peat surfaces. For CH₄ fluxes, a machine-learning framework based on an ensemble of multilayer perceptrons is used for gap filling and flux modeling of different surface classes. The model is driven by meteorological variables, optimally lagged predictors identified via cross-correlation, fuzzy seasonal and diurnal time variables, and class contributions derived from footprint analysis. 

The processed fluxes are compared with previously published EC measurements from 2012 to 2014. Preliminary results based on tower view fluxes (without flux partitioning and gap filling) show that the mean CO₂ flux during the summer months (July–September) ranged from −0.68 to −0.71 µmol m^-2 s^-1 for the year 2016 to 2018. In comparison, the mean summer CO₂ flux in 2012 was 0.67 µmol m^-2 s^-1, indicating a substantial shift from a net CO₂ source to a net CO₂ sink. In contrast, mean winter (January, November, and December) CO₂ fluxes for 2016 and 2017 were 0.70 and 0.66 µmol m^-2 s^-1, respectively, which are comparable to the 2012 value (0.66 µmol m^-2 s^-1).  For CH₄, the mean summer flux increased from 45 nmol m^-2 s^-1 in 2015 to 70 nmol m^-2 s^-1 in 2018, compared to 40 nmol m^-2 s^-1 in 2012, indicating a substantial increase in CH₄ emissions following rewetting during the summer. Overall, the study suggests that rewetting reduced CO₂ emissions while increasing CH₄ emissions, providing new insights into the long-term impacts of peatland rewetting on climate and into the processing of EC flux data in heterogeneous landscapes.