A MODFLOW Field-scale Model to Estimate Ditch Blocking Impact on Peat Water Table

Drainage is the main cause of a lower water table in peatlands, resulting in high greenhouse gas emissions. To combat this issue, the water table in peatlands must be raised, one most obvious way is by elevating the water level on the ditches. This practice has been implemented in many countries with extensive peatlands, such as Germany, Finland, Indonesia, Malaysia, etc. One question arises: how much is the water table raised in the peatland body after ditch water table was elevated?

To estimate the impact of elevating the water table in ditches on the field water table, we developed a field-scale model using MODFLOW6 in Python with the FloPy package. A physically-based model was chosen to account for different physical properties of peat and the underlying sediment layer, as well as topography and climate settings. The model was tested on a fen grassland field underlain by a highly porous sand layer in Gnarrenburger Moor, Northwest Germany. We used nationally available datasets as input, including elevation (DEM with 5m resolution), precipitation, and evapotranspiration data. The field size is 550m x 55m, bordered by ditches on all sides, and was dammed on two sides. A daily transient simulation was performed for 1,023 days from November 2020 to August 2023, and the model was calibrated using observational data.

The calibrated model results show an RMSE of 10 cm and a bias of 3 cm compared to observed water levels. We assessed the impact of ditch blocking by creating scenarios with and without ditch blocking. We found that by raising the water table in the ditches by an average of 31 cm (November 2021 – August 2022) and 30 cm (November 2022 – August 2023), the water table at the observation point was raised by 7 cm and 11 cm, respectively. For the entire field, the model estimate average water table raise by 20 cm (from -50 cm to -30 cm) and 23 cm (from -46 cm to -23 cm). If we only consider water table to calculate CO₂ emission, this corresponds to CO2 emission reductions of 5.21 tCO₂ ha^-1yr^-1 and 12 tCO₂ ha^-1yr^-1. Sensitivity analysis, conducted by adjusting calibrated parameters by ±5%, shows that the ditch water table is the most important factor influencing the field water table.

MODFLOW only considers saturated flow, thus minimizing the requirement for parameters. This model requires only saturated hydraulic conductivity (vertical and horizontal), specific yield, riverbed conductivity, and initial head for transient simulation. In this study, all parameters were unknown and therefore optimized. Despite this simplification, the model successfully simulates the observed water table.

The model was developed solely in a Python environment, utilizing open-source software and nationally available data, making it transferable to other sites with minimal modification. The intention is to apply the model to more rewetted agricultural peatland sites in Germany, as raising the water table in peatland drainage has become a Federal Government program.