Limiting land subsidence of an island polder with a clay - peat subsurface

Abstract

For hundreds of years, ditch water levels in Dutch agricultural peatlands have been lowered to increase the loading capacity of farming parcels. This lowering results in groundwater levels in the middle of the parcels that fall a few decimeters under the ditch levels. When the groundwater level in a peat soil is lowered, newly uncovered peat is exposed to oxygen and begins to oxidise. This process emits greenhouse gasses (GHG), releases nutrients; like nitrogen and phosphorus that stimulate eutrophication, dries out the top subsurface layer and causes the land to subside. This study focuses on an island polder in Warmond, the Zwanburgerpolder, with a clay - peat subsurface. Half of the Zwanburgerpolder belongs to the Eenzaamheid, a biological cheese farm with the ambition to transition to a regenerative cheese farm. One of the key objectives to achieve this regenerative goal is to limit the GHG emitted from the farm. An analysis of the current (ground)water system of the Zwanburgerpolder was done which focused on limiting land subsidence by raising the groundwater level and thereby reducing the emission of GHG, improving (ground)water quality and stimulating biodiversity. First, four research components were examined through field experiments and literature studies: 1) water quantity, 2) water quality, 3) GHG emissions and 4) land surface displacements. The obtained results were used to understand how the Zwanburgerpolder works and identify the relationships between the four components. Then, two groundwater models were developed to represent the current groundwater level and flows in the Zwanburgerpolder: one in iMOD and one in FlexPDE. The iMOD model was used to get a visual representation of the groundwater level variations across the island and to see the groundwater response to precipitation and evaporation. The FlexPDE model was used to get an overview of the phreatic groundwater level drop between two ditches over the summer. To find the most suitable way to limit land subsidence in the Zwanburgerpolder, the models were adjusted to represent possible future situations. Instead of using the collected climate data of 2021 as input parameter, the projected climate data of 2065 was used. The future models consisted of a base scenario, in which no changes were made to the polder compared to the current situation and adapted scenarios, in which the following five water management measures were tested: 1. Adding ditches, 2. Installing horizontal drains, 3. Temporarily inundating parcels, 4. Installing vertical drains and 5. Raising the summer ditch water level. Temporarily inundating parcels performed the best quantitatively and qualitatively during the comparison of the measures. However, this measure is agriculturally unfavourable because as a consequence, parcels cannot be used for grazing during a significantly long period of time. For the Eenzaamheid, where regenerative practices and agricultural capacity take center stage, the recommendation for limiting land subsidence is to combine and adjust two measures: temporarily inundating parcels and adding ditches. Gutters that currently run down the center of the parcels should be enlarged and inundated during the summer period, by siphoning water from the boezem. Inundation can take place during the whole summer, since grazing is still possible alongside the gutters. It is recommended to start off by only applying the measures in the most critical parcels in order to use it as a testing ground to check the possibly negative effects of the measures besides the desired positive effects of limiting land subsidence, reducing GHG emissions and improving the (ground)water quality. Further, the results discussed in this report provide an interesting addition to peatland subsidence studies. They were obtained through field experiments done on a much smaller budget than other studies done so far with expensive measurement setups. The recommendation towards such peatland subsidence studies is therefore to apply a large network of lower cost measurement setups, instead of a few costly ones, in order to get an extensive representation of the behaviour of different Dutch peatlands. The main limitations in this study are the uncertainties linked to the parameters used to build the iMOD and FlexPDE models and the time constraint on the field experiments. The parameter uncertainties mean that the final model outputs fall within a certain error range. To minimize the error range, a model sensitivity and uncertainty analysis should be done. Without the field work time constraint, it would have been possible to identify the seasonal patterns in the groundwater level fluctuations compared to the other water bodies and in the land surface displacements.