MATLAB code was developed to calculate the EC fluxes. Several data treatment procedures were tested; most of them were adapted from commonly used calculation procedures for terrestrial eddy covariance measurements (e.g. Baldocchi et al. 2003). The flux was determined over periods of 0.25 h. Data were averaged to 8Hz for flux calculations due to the ~ 10 Hz sample output refresh rate of the O2 optode meter. The fluctuating components of the vertical velocity, O2 and H+ (Eq. 1), were determined by Reynolds decomposition with the means determined by a running average window of 5 minutes, which was identified to be the optimal time period for maintaining a constant flux signal while removing non-turbulent fluctuations (McGinnis et al. 2008).
Inherent in the high-frequency ADV velocity data were anomalous spikes that contaminated the EC signal. These velocity data spikes were replaced with interpolated data using existing MATLAB de-spiking procedures described by Goring and Nikora (2002) that resulted in the interpolation of <1 % of the vertical velocities used to calculate the flux. The ECHOES H+ and O2 signals were compared to the water column data sonde measurements to verify that the sensors had not malfunctioned by confirming that the real-time correlation between sensors and data sonde matched that of the overall correlation (i.e. the calibration curves). All signals were then examined manually to remove any further data spikes that were due to malfunction, fouling, or debris contacting the sensors. A GoPro camera (HERO2, GoPro, USA) on the ECHOES frame helped identify fouling of the sensors. Due to the difficulty of accurately leveling the instrument in the field, and the resulting potential to bias the vertical velocity measurements, a planar rotation was used based on the methods described by Lorke et al. (2013) that uses an average rotation angle for the different current directions.
The benthic chamber fluxes were calculated from the change in O2, H+, DIC and TA concentrations in the chambers through time, using the known volume and sediment surface area of the chamber. Corrections for the removal of water for DIC and TA samples were included to account for the influx of ambient water, utilizing the measured water column conditions.
BCO-DMO Processing:
- added conventional header with dataset name, PI name, version date, reference information
- renamed parameters to BCO-DMO standard
- reformatted date