Methodology:

The eddy-covariance sea-air methane flux data was collected on the RRS James Clark Ross during cruises JR18004, JR18005, JR18006, JR18007, JR19001, JR19002, and during periods 15/11/2020 - 30/11/2020 and 02/12/2020 - 23/03/2021. The data processing procedure is described: The raw high-frequency wind and CH4 data are processed to calculate fluxes in 20 minute intervals. Related statistics are used for quality control. Linear detrending is employed to identify turbulent fluctuations (w ' and c '). Ship motion is corrected by applying a complementary filtering method using Euler angles which eliminates apparent winds caused by ship movements. The motion-corrected winds are decorrelated against ship motion and double-rotated to account for the wind streamline over the ship. The resulting vertical wind velocity (w) is used to calculate flux. The processed wind data is used to compute variables such as friction velocity and sensible heat flux. CH4 data is despiked, decorrelated against analyser cell pressure and temperature, and decorrelated against factors related to ship motion. A lag (of a few seconds) between CH4 and wind data is addressed using a puff system or the maximum covariance method. The inlet tube, particle filter, and dryer introduce attenuation in CH4 flux signals, which is compensated for in the calculations. Horizontal CH4 fluxes and other statistics are computed for quality control. Finally, the computed fluxes undergo filtering to exclude non-ideal ship manoeuvres or violations of the EC technique requirements. Additionally, the CH4 mixing ratio data was filtered based on wind direction to remove data corresponding to wind coming from behind the ship to remove sources of pollution from the ship stack. 20 minute intervals that pass the QA/QC (Quality Assurance/Quality Control) are further averaged to 2 hours to reduce noise.

Data collection:

Data was collected using a eddy-covariance system installed on the JCR. The system was located at the top of the foremast (approximately 20 m a.m.s.l). The system comprises of a Picarro G2311-f greenhouse gas analyser, a three-dimensional sonic anemometer (Metek Inc., Sonic-3 Scientific) and a motion sensor (initially Systron Donner Motionpak II, which was replaced by 169 a Life Performance-Research LPMS-RS232AL2 in April 2019). Python 3 is used to analyse the data and create the dataset.

Data quality:

QA/QC filters used are:

Basic CH4 flux filtering: same filtering for wind direction and ship movement as for CO2 flux (Dong et al, 2021), then filter data based on these stats within 20min window:

CH4(range) < 0.02ppm

CH4(sdev) < 0.0014ppm

abs(CH4(slope)) < 0.02ppm/hr

Finally remove flux outside of 5 sdev and average to 2hr intervals (retaining 2hr periods with at least 4 good flux segments).

Missing flux data is due to quality control filtering/instrument not collecting data.