A strategy for quality and uncertainty assessment of long-term eddy-covariance measurements

Eddy-covariance measurements are performed at several hundred sites all over the world on a long-term basis. The gathered data are used to characterise ecosystem exchanges of trace gases, water and energy and to validate or constrain process-based models. There is an increasing demand on standardised and comprehensive quality flagging and uncertainty quantification of these fluxes. In this paper, we review established quality assessment procedures and present a comprehensive newly composed strategy emphasising tests on high-frequency raw data, expanding existing tests on statistics, fluxes and corrections, plus quantification of errors. Moreover, representativity of fluxes is checked by footprint analysis. This strategy is applied within the recently launched TERENO network of ecosystem observatories, and its robustness is demonstrated for data acquired with different measurement set-ups. Four test data sets from TERENO and one data set from CarboEurope-IP were subjected to this quality assessment. The presented strategy is compared with established quality assessment schemes, and it is demonstrated that unrealistic fluxes are now efficiently excluded while retaining the largest possible amount of high quality data. Additionally, the algorithms applied provide comprehensive, reproducible, qualitative and quantitative uncertainty estimates for users of eddy-covariance flux data.