On the need for uncertainty assessment of long-term eddy-covariance measurements

Eddy-covariance measurements are routinely performed in a worldwide long-term observation network of ecosystem exchange of trace gases, water and energy. The data obtained are needed to validate or constrain process-based models and for evaluating ecosystem budgets. There is a strong demand for harmonized and comprehensive quality flagging and uncertainty quantification to assure comparability of data sets from different sites. We review established quality assessment procedures and suggest a newly composed strategy comprising tests on high-frequency raw data, tests on statistics, fluxes and corrections. Additionally, we quantify different types of errors on the calculated fluxes. This strategy will be applied within the recently launched TERENO network of ecosystem observatories. Five test data sets from TERENO and CarboEurope-IP were subjected to the specific quality assessment scheme. These data sets include two different sonic types, open- and closed-path instruments, tall and low vegetation, flat and complex terrain. We show the robustness and applicability of the scheme to data acquired with the different measurement set-ups which demonstrate coherences between established flagging schemes and newly added error determination. This uncertainty assessment for each flux value represents an indispensable value for modelers as well as for budgeting fluxes.