From raw data to standardized, fully corrected, quality ensured eddy covariance flux data: the ICOS Ecosystem processing pipeline

The eddy covariance is a micrometeorological technique which allows for the estimation of the net fluxes of gases and energy between the atmosphere and an ecosystem. To estimate the net balance, the required input data are high frequency measurements (e.g. 10 or 20 Hz) of wind speed and gas concentration or amount of energy, plus lower frequency measurements (e.g. 1s to 30min ) of some meteorological variables and gas concentration vertical gradients below the measuring point. From these measurements, through a set of processing algorithms and corrections, continuous time series of fluxes are obtained which can be used to, e.g.,  estimate the net ecosystem exchange, as input/validation for modelling purposes, or for eco-physiological analyses. Although the fundamental processing steps and corrections are well established, there is still a discrete margin of subjectivity in the choice of specific operations and corrections which leads to different results even starting from the same set of measured data. The ICOS Ecosystem infrastructure consists of a network of eddy covariance stations equipped with high-level standardized instrumentation, whose data are processed centrally by means of a fully standardized and documented processing pipeline. This allows to obtain robust and consistent datasets, along with sets of metadata (e.g. instruments characteristics and location) and ancillary variables (e.g. meteorological and biometric) that help their interpretation and ensure their traceability and reproducibility. The description of the full processing pipeline is the aim of this contribution. All the data and metadata produced by the ICOS Ecosystem Thematic Centre (ETC) are freely available through the ICOS Carbon Portal as well as the processing codes are available in the ICOS ETC GitHub repository.