USGS T-3 enhanced thermal data from T-3 Ice Island, 1963-73

This process step gives an overview of the acquisition of the original data. This is copied from details given in this data release in metadata for dataset: "USGS T-3 original thermal gradient, thermal conductivity, and heat flow data from T-3 ice island, 1963-73."
Details about the operations to acquire the data are given in Lachenbruch and others (2019)--see cross-reference. Summary information about procedures is also provided in the other cross-referenced publications, particularly the Journal of Geophysical Research paper by Ruppel and others (2019). The heat flow measurements were made once a day or once every several days (occasionally two per day) during USGS occupation of the T-3 ice island. The measurements were made through a permanent hole in the ice (moonpool) maintained under the protection of a hut near the Colby Bay portion of the T-3 ice island. The apparatus consisted of a corer with attached thermistors (four thermistors during the first few years of the operation and then five thermistors on a slightly longer core barrel during subsequent years). The thermistor closest to the bottom of the core barrel was designated as #1. Thermistor accuracy was 0.001 degrees Celsius based on Lachenbruch and others(2019).
The corer with attached thermistors was lowered to the seafloor on a winch system. The probe remained in the bottom sediments long enough for the frictional heat pulse associated with sediment penetration to decay to the conductive part of the curve so that an equilibrium temperature could be determined for each thermistor during post-analysis. Thermal conductivity was not measured in situ in the seafloor. The probe was recovered with sediments in the corer. Far more sediment cores were recovered than heat flow measurements taken (for example, refer to the information in the University of Wisconsin NCEI dataset at https://doi.org/10.7289/v51834g7 for selected information about USGS T-3 sediment cores). It is unknown whether heat flow measurements were not attempted on every coring run or whether the heat flow probe or recording instrumentation suffered frequent problems. The consistent recovery of sediment cores indicates that the probe penetrated the sediments at many locations, meaning that a heat flow measurement was in theory possible at those locations as well.
After completion of a measurement, data would have been analyzed to determine an equilibrium temperature for each thermistor. These equilibrium temperature data have been lost, as have the bottom water temperature data recorded during the surveys and information about the penetration depth of the probe. For the Journal of Geophysical Research paper by Ruppel and others (2019)--see cross-reference, attempts were made to use the lengths of recovered cores as an indicator of the probe's penetration depth to assist in determination of equilibrium thermal gradients, but this analysis was unsuccessful. The reported data from the heat flow measurements are thermal gradients between adjacent thermistors. For a 4-thermistor penetration with all 4 thermistors operational, the dataset provides thermal gradient values between thermistors 1 and 2, 2 and 3, and 3 and 4. If one thermistor were not functioning (for example, thermistor #2), the reported gradients would be between thermistors 1 and 3 and 3 and 4.
Thermal conductivity values were measured on the laboratory benchtop on recovered sediment cores using needle probes. The reported values correspond to core sections between each of the working thermistors. For example, if a gradient is reported for the interval between thermistors 2 and 3, then the dataset gives a corresponding thermal conductivity determination in that interval from the corresponding core. The original thermal conductivity data have been lost, but it is suspected that far more thermal conductivity measurements were completed than are compiled or averaged to generate the values in the dataset. Ruppel and others (2019)--see cross-reference-- analyzes some of the thermal conductivity data in this dataset. Heat flow was calculated by multiplying average thermal gradient and average thermal conductivity. Uncertainties are reported on a subset of the values as a percentage of the heat flow. A maximum heat flow uncertainty is also provided by for heat flow values to which a percentage deviation was assigned. Process date likely spanned the duration of the experiment and is reported below as the last day of 1974, which is the approximate date of the USGS report provided in Lachenbruch and others (2019)--see cross-reference

This process step details how the original table was digitized and quality checked. The original table was available only as a PDF scan of Table 2 from Lachenbruch and others (2019). In 2016, Carolyn Ruppel manually keyed most of this table. Subsequently, technical staff (B. Clark and M. Arsenault) at the USGS Woods Hole Coastal and Marine Science Center conducted optical character recognition on parts of the scanned file. The two inputs were combined, and then Deborah Hutchinson conducted quality control on the resulting table. This process identified entries that were partially illegible in the original PDF scan of the report, positions that had been improperly keyed, and one position (station FL-352) that did not plot on the T-3 navigation track obtained through independent means and released with this dataset as "T-3 Ice Island One Hour Navigation: May 14, 1962 to September 15, 1974." Values were corrected in the table based on this analysis. It is expected that some errors likely remain in the transfer of the original PDF scan to the digital data file. It is not believed that any of the possible errors affect the subsequent outcome or analysis in any significant way.

Original data were averaged, converted, and updated: Latitudes and longitudes were transformed from degree-minute to decimal degree format. Average thermal gradients were determined by averaging reported interval thermal gradients over the number of interval thermal gradients per station. Average thermal conductivity was determined in the same manner using the interval thermal conductivities. Heat flow was converted from Heat Flow Units (HFU) to modern SI units (mW/m^2) by multiplication by the factor of 41.8 mW/m^2 per HFU. The reported maximum deviation of heat flow measurements was similarly converted to HFU. Uncertainties expressed as a percentage were multiplied by the converted (SI units) heat flow to generate an error bar for use in plotting. Where percentage uncertainties were not reported, a value of 10% was adopted to generate the appropriate error bar. The heat flow value was divided by the average gradient to generate an inferred thermal conductivity used by the original reference (Lachenbruch and others, 2019--see cross-reference).

Using Esri ArcGIS 10.2, the latitude and longitude of the heat flow stations were merged with the Arctic Ocean physiographic province polygons from Jakobsson, M., Hypsometry and volume of the Arctic Ocean and its constituent seas, Geochem. Geophys. Geosyst., 3(1), doi:10.1029/2001GC000302, 2002 and a split added between Alpha and Mendeleev Ridges. The polygon shapefiles were provided by Martin Jakobsson to Carolyn Ruppel. This produced a mapping of physiographic province for each heat flow station. In addition, the positions of the heat flow stations were merged with surficial geology from Boggild, K., D. C. Mosher, C. Gebhardt, M. Jakobsson, L. A. Mayer, K. Hogan, and D. Dove (2018), Surficial geology of the Amerasian Basin from subbottom profiler data, Open File Rep. 8465, Geological Survey of Canada, Ottawa, Canada. The surficial geology GIS was provided to the U.S. Geological Survey by David Mosher, co-author on the cross-referenced Journal of Geophysical Research (Ruppel and others, 2019) paper. Dates were added to the dataset first by using known dates of piston cores retrieved from the Arctic Ocean simultaneously with the acquisition of the heat flow data. The dates are given in the NCEI dataset from University of Wisconsin (2016)--see cross-reference. All successful heat flow stations were presumed to have yielded piston cores since thermal conductivity data are reported for all stations. However, not all the cores may have been sent to University of Wisconsin and are therefore not in the database. When core dates were missing, ArcGIS 10.2 was used to find the date on which the T-3 drift path was closest to a particular heat flow station using the 1-hour navigation provided as part of this release in "T-3 Ice Island One Hour Navigation: May 14, 1962 to September 15, 1974." This method assumes that the position recorded for the heat flow station was where the heat flow was actually measured in the seafloor. However, the position could be recorded when the corer entered the water, and the heat flow station actually recorded when it penetrates the bottom, up to hours later. Also note that this method of merging time-stamped navigation with the positional information for the heat flow data assumes that the heat flow probe entered the seafloor vertically beneath the location of the surface infrastructure on T-3 Ice Island. In modern heat flow measurements, the heat flow probe is only rarely directly beneath the vessel when heat flow is recorded. For stations at which neither method for assigning dates was possible, the date was interpolated based on known dates of preceding/subsequent stations. The lengths of cores recovered at each station were added to the file using information from the University of Wisconsin (2016) NCEI release and from a data file provided by Dr. Dennis Darby of Old Dominion University.

Text data flags were added as columns in the Excel file to replace color coding used in the working copy of the enhanced data file.

Added keywords section with USGS persistent identifier as theme keyword.