Eddy covariance fluxes of carbon dioxide and methane from the Herring River in Wellfleet, MA (ver 2.0, June 2022)

An eddy covariance system was originally installed at the Herring River site in 2017 (2017-2020 data not available at time of this publication) to collect ecosystem scale measurements of CO2 and CH4 fluxes across the land surface (tower is part of the Ameriflux network code US-HRP). A tripod tower was equipped with a LI-7500A open path CO2/H2O analyzer (LI-COR, Inc., Lincoln, NE, USA) and a LI-7700 open path CH4 analyzer (LI-COR, Inc., Lincoln, NE, USA), along with a CSAT-3 3D sonic anemometer (Campbell Scientific, Logan, UT, USA). Instruments are deployed at 4.4 m height above the ground surface, along with a suite of meteorological sensors that measure air temperature and humidity (Onset HOBO Temperature/RH Smart Sensor (S-THB-M00x), Bourne, MA, USA), incoming photosynthetically active radiation (PAR; Onset HOBO Photosynthetically Active Radiation Smart Sensor S-LIA-M003, Bourne, MA, USA), and precipitation (Onset HOBO Rain Gauge Smart Sensor S-Rgx-M002, Bourne, MA, USA). Anemometer and gas analyzer data were collected at a frequency of 10 Hz and ancillary meteorological data were collected every five minutes. Data were downloaded every 3-4 weeks as compressed .ghg files from the anemometer and gas analyzers and as .csv files for the ancillary meterological sensors.

Half hourly fluxes of CO2 and CH4 were computed from the covariation of high frequency vertical wind direction and CO2 and CH4 concentration using the EddyPro software package (Version 7.0.6, LICOR, Inc., Lincoln, NE, USA), as well as latent heat exchange and friction velocity. Within software specifications, raw high frequency data were despiked and a double coordinate rotation was applied to anemometer measurements along with a flux correction for the effects of humidity on sonic air temperature, high-frequency and low-frequency spectral attenuations, and air density fluctuations coupled to instrument-related sensible heat flux in the winter. Computed fluxes were exported to .csv files. We report fluxes of carbon dioxide as net ecosystem exchange (NEE), where negative values for both NEE and CH4 indicate net uptake of gas by the land surface and positive values indicate net loss of gas to the atmosphere. We removed low quality measurements (ranked as less than or equal to 1 quality flag per EddyPro based on Mauder and Foken, 2006), measurements made during and immediately following precipitation events, and any measurement collected during site maintenance or known instrument malfunction. We also removed measurements made during low‐turbulence conditions with potential advective losses and when friction velocity, ustar, was below a site specific threshold of 0.118 meters per second calculated in the R-based software package REddyProc (Version 1.2.2, Wutzler and others 2018). Instrument malfunction caused a gap in some meteorological data from late-January through mid-April 2021 and are represented as blanks in the data. The .csv files containing the meterological data and the .csv file containing the computed fluxes were matched by datetime and merged in R.
Wutzler, T., Lucas-Moffat, A., Migliavacca, M., Knauer, J., Sickel, K., Šigut, L., Menzer, O., and Reichstein, M., 2018, Basic and extensible post-processing of eddy covariance flux data with REddyProc: Biogeosciences Discussions, February, 1–39. https://doi.org/10.5194/bg-2018-56.
Mauder, M., and Foken, T., 2006, Impact of post-field data processing on eddy covariance flux estimates and energy balance closure: Meteorologische Zeitschrift, vol. 15, p. 597-609

Latitude, longitude and elevation of the eddy covariance tower was measured with a Trimble Real-Time Kinematic (RTK) GPS. Accuracy was 2-3 cm according to instrument corrections. Data files were exported from the instrument via a .csv file and appended to the data file of calculated fluxes via R.

Measurements were continued as described above from May 2021 to May 2022. All new data were processed as described above and 17308 records were appended to the original dataset (17568 records) bringing the total dataset to 34876 records and revising the end date of the dataset from May 23, 2021 to May 18, 2022. In addition, newly available data was added to blank columns in the last 71 rows of the original dataset, but the original dataset is otherwise unchanged and no existing data values were modified. The metadata was updated to reflect the new ranges of values.