Groundwater level control as GHG emission reduction option tested using eddy covariance for peatland in the Netherlands

In this work, the annual CO₂ and N₂O balance and analysis of the drivers for fertilized grassland on peat for a dairy farm under three groundwater level control options will be presented. This experiment is conducted in the context of the Dutch NOBV project (National Research program on GHG for peatland areas) as part of the Dutch Climate Agreement which has a chapter to reduce GHG emissions from peatland areas by 1 MtCO₂-eq annually.

The groundwater level control options applied are the conventional ditch water level control system, nowadays often being replaced by drainage and ditch level control, and finally drainage and pressure control. Several fields of the Zegveld experimental farm are divided in three segments, each which a control option applied and a such allow study under comparable conditions.

Two years of GHG fluxes are reported and measured using one closed-path Aerodyne system switching inlet line each half hour using three small towers equipped with a Gill sonic anemometer at 1.75 m height in the middle of the largest elongated farm field. The eddy-covariance method was used to calculate half-hourly fluxes. The location and low measurement height maximize the representation of the field in the flux measured from all wind directions. CO₂ fluxes showed uptake during the day and respiration during nighttime. After harvesting and grazing, emission fluxes prevailed. N₂O peaks coincided well with agricultural management, e.g. grazing and fertilization, but also biometeorological factors like water temperature and ground water table influenced N₂O emissions. From October 2023 to March 2024, N₂O emissions were close to zero due to prolonged precipitation resulting in a shallow water level across all fields inhibiting production of N₂O in the subsurface layer. Due to the high contribution of the field to the footprint compared to surrounding ditches, CH₄ fluxes didn’t correlate with any in-field measured parameters driving the hypothesis that ditches appear to be the main source of CH₄.

Flux loss corrections based on an empirical approach using measured ogives. Gap-filling of the N₂O and CO₂ fluxes was done using the gradient boosted regression trees XGBoost. The gap-filling process utilized a comprehensive set of predictors to enhance the accuracy and reliability of flux measurements. A comparison with an open-path CO₂ system installed at one location showed a reasonable agreement with CO₂ fluxes of the closed-path measurement system.