Veröffentlichungen aus dem Verbundvorhaben KliMoBay

Estimation of peatland evapotranspiration by means of remote sensing

Verena Kuch 2021, Department für Georgraphie, Lehr- und Forschungseinheit Hydrologie und Fernerkundung, Ludwig-Maximilians-Universität München, Masterarbeit

Emissions from peatlands are controlled by the water table. Next to precipitation the evapotranspiration (ET) is the main control of the water table level. In Bavaria, peatlands get more attention regarding the change from release of greenhouse gas emissions to an uptake.
In the framework of the KliMoBay (Klimaschutz- und Anpassungspotenziale in Mooren Bayerns) project at the chair of geography and geographic remote sensing, three different methods for modelling ET were tested in this study. The Triangle Method, the DATTUTDUT (Deriving Atmosphere Turbulent Transport Useful To Dummies Using Temperature) model and the TSEB (Two Source Energy Balance) method was used to retrieve ET in a Bavarian bog. The Schechenfilz in the south of the Starnberger See is a semi-natural bog and provides measured atmosphere data from an Eddy Covariance tower. All methods use the land surface energy balance and land surface temperatures (LST) to calculate ET.

The DATTUTDUTT model only needs LST information for the calculation but can be supported with measured data. In this study, the Triangle Method and TSEB-method dependent on measured data. Thereby, the Triangle Method uses a relationship between LST and NDVI. The TSEB-method is the most complex model and calculates ET separately for every pixel. ET was calculated for the timespan 2015 to 2020 at 12 am at the Eddy Covariance station.
The DATT-model has the best results (RMSD = 0.11 W/m²), when calculating with a measured value for the net radiation. The independent calculation of ET without ground measurement presented good results until 2019. Then the findings got negative, and an analysis could not correct the miscalculation. The Triangle Method has an overall RMSD of 0.26 W/m² and gives the best results when the relationship of LST and NDVI fits a triangle between wet and dry edge of the existing data points. Tests showed that an improvement of the triangle shape also improves the results. With a RMSD of 0.3 W/m² the TSEB-method presented the lowest accuracy. Rough estimated input data for the complicated calculation downgraded the results.
The method would need a large data collection during the hole timespan in the peatland. Overall, especially the DATT-model and Triangle Method showed high potential for spatially distributed ET maps over peatlands and is the first step to calculate ET independently from ground measurements for future spatially distributed maps for hole Bavaria.