Provenance & quality
Raw data were pre-processed to check for the following, spike detection, coordinate rotation, frequency response correction, cross correlation and WPL (Webb, Pearman, and Leuning) correction. Half-hourly data were gap-filled using the code published on GitHub and subjected to season-specific u* filtering using the ReddyProc R package v1.2. Cospectra analysis was performed using Hamming window. Spikes in the data were excluded using a method based on Tukey’s fence. Data were supplied to the EIDC as comma seperated value (.csv) files.