10 Years of Quality Controlled Surface Weather Measurements from the NSF Ocean Observatories Initiative (OOI) Coastal Endurance Array

10-minute and hourly measurements of surface weather observations collected by the OOI Coastal Endurance Array using the ASIMET meteorological sensor suite (METBK) measuring (among other variables) the sea surface temperature, salinity and wind speed. This data has been re-processed from source data available on the OOI Gold Copy THREDDS catalog. The reprocessing workflow consists of:

• Loading all available raw METBK data sources.
• Applying automated and human-in-the-loop (HITL) quality control flags to remove values marked as 'fail'.
• Calculating precipitation rates.
• Replacing missing CT data with recovered instrument data where available.
• Despiking the conductivity measurements and calculating the sea surface salinity.
• Correcting wind data for magnetic declination.
• Resampling the 1-minute data to 10-minute and hourly averages.
• Concatenating all deployments into final datasets.

Additionally, the wind records in these datasets are corrected for underspeeding caused by a identified firmware issue noted in the OOI Data Explorer annotations.  Based on comparisons to the OOI direct covariance flux wind measurements, the wind records presented in these datasets are adjusted upwards by a gain factor of 1.1236.

In this version of the dataset (v1.1.0), only the sea surface temperature, salinity and wind speed data has been fully QC'd.

Data is organized in both NetCDF and Zarr files (in Zarr ZipStores) per mooring location differentiated by the file contents:

In the 'per_deployment' files, data from overlapping deployments is made available for researchers wishing to assess the data on a per deployment basis (e.g., cross-compare overlapping deployments and discrete samples, assess sensor drift, develop correction curves, etc). The deployment number is included in the data record and the data has been sorted by deployment number and then time. The data record in the per deployment files is 'gappy' in time; if no data was available from any of the sensors (e.g., the unit was inoperable), then that timestamp was not included. A potential downside to these files is the time record is not monotonic, which may cause some issues depending on how the data is analyzed. These files are only available in NetCDF format.

The hourly 'per_deployment' data files are used in the processing of the PCO2A data available in 10.5281/zenodo.17883448.

For those wishing to work with monotonic time records (with no gaps in time), the overlapping deployment data has been averaged, the deployment variable has been dropped and missing time stamps have been added to create a 'continuous' record. These 'continuous' files are available in both NetCDF and Zarr format (as Zarr ZipStores) and span from the beginning of the first deployment for that mooring through to the end of the available data record.

In all cases, missing data is filled with NaN's.

Example code (in both Python and Matlab) is included to assist users in accessing the data and exploring the contents. See ooi_metbk_example_python.ipynb and/or ooi_metbk_example_matlab.mlx. The example code files are included for download and a GitHub Gist version of the examples is also available via the links.