| Abstract: | The FLUXNET2015 dataset provides ecosystem-scale data on carbon dioxide, water, and energy exchange between the biosphere and the atmosphere and other measurements from 212 sites around the globe. It contains over 1500 site-years (up and including year 2014) of flux measurements. The sites are independently managed and operated, data are voluntarily contributed by site investigators often through regional networks, which collaborate in FLUXNET to create globally representative datasets. The data collection process at the sites encompasses different steps from the selection of the ecosystem and location, to the equipment choices and setup, measurements collection and fluxes calculation. At the network level, all data were quality controlled and processed using uniform methods and algorithms, improving the FLUXNET2015 dataset consistency and inter-comparability across sites. This network-level effort was coordinated by the teams of the European Fluxes Database Cluster and the AmeriFlux Management Project. The implementation of these processing methods is available in a open-source set of codes called ONEFlux, which can be run as a data pipeline. The data collection is already being used in a number of applications, including plant physiology studies, validation of remote sensing datasets, and evaluation of ecosystem, land-surface, and earth system models. At its original release time, FLUXNET2015 included derived-data products, such as gap-filled micrometeorological and flux data, carbon dioxide fluxes partitioned into ecosystem respiration and photosynthetic uptake, and estimation of uncertainties from the measurement technique and data processing steps. The new release, presented here, adds data products for many applications including key information about sonic anemometer and gas analyser models, canopy height, and measurement depths of the multi-layer soil variables (temperature and water content). In addition, with this release the global eddy covariance data are for the first time distributed under a Creative Commons Attribution (CC-BY 4.0) license. Not all 212 sites could be distributed under this more open license, so only 206 sites are included in this paper. This paper details the enhanced dataset and the processing methods, now made available as open-source codes, making the dataset more accessible, transparent, and reproducible. |