Assessment and Prediction of Greenhouse Gas Fluxes with Experimental and Machine Learning Techniques (ArGEnT)
Land use changes are important drivers of anthropogenic climate change by altering greenhouse gas (GHG) uptake and storage capacities of an ecosystem. Hence it is of great importance to quantify the net carbon exchange between ecosystems and the atmosphere. The eddy covariance technique is the most direct way of measuring GHG fluxes, however, it is very sensitive to the correct experimental setup and ambient conditions and provides only point measurements from a sparse network of stations. The ArGEnT project had two objectives in order to address these shortcomings: 1) the effect of low pass filtering on eddies and the performance of different correction methods were examined at a deforested and now natural regrowth area at the TERENO site Wüstebach/Eifel with two EC stations at different heights and 2) measurements from this site were combined with EC measurements of other land cover types in the Rur catchment area and remote sensing data to train a machine learning model and upscale carbon dioxide fluxes for the whole catchment at high spatiotemporal resolution.
Department of Geography RWTH Aachen, Department of Physical Geography and Climatology:
Research Centre Jülich: Dr. Alexander Graf, Marius Schmidt