High resolution automated water table mapping using UAV LiDAR terrain data and near real time water logger measurements

The water table depth (WTD) is one of the main drivers for greenhouse gas (GHG) emissions in peatlands. Peatlands act as long term carbon sinks when the water table remains close to the soil surface, whereas drainage typically leads to substantial GHG emissions. Hence, detailed spatial water table information with high temporal resolution is a crucial requirement for evaluating restoration success and supporting practical management decisions. Combining continuous in situ WTD measurements with remote sensing approaches can open an innovative way of deriving hydrological information for applied peatland management und restoration. However, few studies combine near real time in situ loggers with remote sensing to generate spatially continuous and frequently updated water table products for operational monitoring.

Here we present an end-to-end workflow for automated water table mapping, that couples continuously transmitted in situ water level measurements and high-resolution terrain information by UAV LiDAR with immediate data processing. A network of water level loggers transmits measurements every 30 minutes automatically to a central database using the public cellular network. Additionally, UAV LiDAR point clouds are acquired twice a year using a Zenmuse L2 sensor, suitable for the use in densely vegetated areas. Ground points are classified using the cloth simulation filter to minimize residual canopy artefacts to generate a 0.25 m Digital Terrain Model (DTM) to capture peatland microrelief. The generated DTM ist then validated against RTK GNSS ground truth points to quantify the vertical accuracy and ensure the reliability of the LiDAR data. The WTD is computed by referencing gauge water levels to the DTM. Spatially continuous water level maps are produced using a regression kriging approach that exploits the strong dependence of water level on relative surface elevation. Model performance is evaluated by using leave-one-out cross-validation across the gauge network, allowing a direct comparison of LiDAR based against public DTM based WTD and quantifying uncertainty in derived water level maps.

Initial results suggest that the LiDAR based DTM reduces elevation bias compared to public DTMs and yields more consistent interpolated WTD dynamics, especially in heterogeneous areas. Furthermore, by linking near real time logger data with high resolution DTMs, the workflow enables a reproducible, automated delivery of regularly updated WTD maps for operational use to support water management, restoration planning and the establishment of paludiculture. In addition, these products provide hydrological inputs for proxy based GHG emission assessment, e.g. supporting GEST (Greenhouse Gas Emission Site Types) approach for baseline emission estimates in strongly drained areas and for tracking changes during the transition after rewetting. The output can contribute to monitoring, reporting and verification of workflows, supporting certification schemes such as voluntary carbon crediting.