Spatial water table level modelling with multi-sensor unmanned aerial vehicle data in boreal aapa mires

Peatlands have been degrading globally, which is increasing pressure on restoration measures and monitoring. New monitoring methods are needed because traditional methods are timeconsuming, typically lack a spatial aspect, and are sometimes even impossible to execute in practice. Remote sensing has been implemented to monitor hydrological patterns and restoration impacts, but there is a lack of studies that combine multi-sensor ultra-high-resolution data to assess the spatial patterns of hydrology in peatlands. We combine optical, thermal, and topographic unmanned aerial vehicle data to spatially model the water table level (WTL) in unditched open peatlands in northern Finland suffering from adjacent drainage. We predict the WTL with a linear regression model with a moderate fit and accuracy (R2 = 0.69, RMSE = 3.85 cm) and construct maps to assess the spatial success of restoration. We demonstrate that thermal-optical trapezoid-based wetness models and optical bands are strongly correlated with the WTL, but topography-based wetness indices do not. We suggest that the developed method could be used for quantitative restoration assessment, but before-after restoration imagery is required to verify our findings.

Automated summary

This study used unmanned aerial vehicle data and a linear regression model to spatially model the water table level in peatlands in northern Finland, demonstrating the potential for assessing the spatial success of restoration using multi-sensor ultra-high-resolution data.