McMurdo Sound sea ice properties and temperatures

Metadata for ice temperatures and ice core properties collected in McMurdo
Sound, Antarctica as part of the project "Measurements and Improved
Parameterization of the Thermal Conductivity and Heat Flow through First-Year
Sea Ice", OPP-0126007*
---

The polar oceans' sea-ice cover strongly modifies ocean-atmosphere heat
transfer. Most important, the ice cover thermally insulates the ocean, with the
sea-ice thermal conductivity determining the magnitude of the heat flow for a
given ice temperature gradient. Despite its importance (second only to ice
albedo), large-scale sea-ice and climate models currently include overly
simplistic parameterizations of the ice thermal conductivity that are likely to
contribute significantly to errors in estimating ice production rates. As part
of this project and in collaboration with colleagues from Victoria University
Wellington in New Zealand, we have completed a set of field measurements in
McMurdo Sound, Antarctica, and Barrow, Alaska, from which the thermal
conductivity of first-year sea ice has been derived as a function of ice
microstructure, temperature (and temperature gradients), salinity and other
environmental parameters. In order to arrive at in-situ, non-destructive
estimates of brine volume and salinity, capacitance probes that provide data
from which the complex permittivity of sea ice at 50 MHz can be derived have
also been frozen in.

Based on the permittivity data and laboratory experiments, we have been able to
link dielectric properties and in-situ brine volume fractions and salinities.
Dielectric mixture modeling and inversion of permittivity data thus allow for
non-destructive monitoring of in-situ salinity and thermal property evolution.
Analysis of thermistor string data from field sites in Barrow and McMurdo has
allowed for a derivation of ice thermal conductivity. In conjunction with the
permittivity data and microstructural studies the relative importance of
diffusive and convective heat transfer in first-year sea ice has been assessed.
Finally, the information gathered as part of this project has been synthesized
and led to an improved parameterization scheme of thermal conductivity in
large-scale sea ice and climate models.

Data provided here have been collected as part of the project "Measurements and
Improved Parameterization of the Thermal Conductivity and Heat Flow through
First-Year Sea Ice", OPP-0126007* and include measurements of temperature and
various ice properties at selected sites in first-year and multiyear sea ice in
McMurdo Sound, Antarctica in the years 2002-2004. Data from earlier
installations of thermistor chains for measurements of ice temperature carried
out by the New Zealand team have also been included. Data files are in
Microsoft Excel format, with individual worksheets for specific cores or
temperature data sets. Detailed information and comments on data sampling
location etc. are provided in the files.

Ice core data (Filename: MMcoredata_xx.xls):

Sea-ice cores were obtained with a 10-cm diameter corer and ice thickness was
measured at the site with a tape measure placed alongside the entire length of
core. On-site, ice temperature was measured, typically at 5 to 10 cm intervals,
by drilling holes (ca. 5 mm) into the center of the core and inserting a
temperature probe (precision 0.05 K, accuracy <0.1 K) while covering the core
with a light/thermal shield. The core was then sawed into 5 to 10 cm segments
and core segments were transferred into sealed plastic containers to minimize
loss of brine. Salinity of samples was determined with a YSI Model 30
Conductivity sonde (measurement error <0.02 or <1 % of the bulk salinity,
whichever is larger).

Stable isotope (delta-18-O) measurements were carried out at the Alaska Stable
Isotope Facility, University of Alaska Fairbanks (Dr. M. Wooller).
Stable-isotope measurements were performed on a Delta+XL Mass Spectrometer
(H2-CO equilibration, measured against VSMOW) at a precision of better than 0.5
ppt (with standard deviation between replicate runs of three samples below 0.4
ppt).

Ice temperature data (Filename: XX_tempdata_yyyy_uuu[].xls):

Here yyyy is year, and uuu is university; [ ] denotes additional qualifiers in
some cases. There are descriptive headers in each file.

All measurements were made in landfast first-year ice, except in 2003, where
‘MY’ in filenames denotes deployment in multiyear ice. Temperature measurements
were carried out with thermistors (Omega #44031), wired into a Campbell CR10
data logger (precision better than 0.06 K). An important aspect of the
experimental set-up is the mounting of the thermistors. The thermistor arrays
(5-10 cm spacing) consisted of a separate wire duct, embedded in a
polycarbonate-polyethylene matrix of thermal conductivity somewhat lower than
that of ice, with the actual thermistors (embedded in a small glass bead)
mounted approximately 5 cm away from the duct on a thin support in direct
contact with the ice matrix. Temperatures were calculated from resistivity data
based on resistivity-temperature dependence relationships provided by the
manufacturer for individual thermistors. Also, where noted in data files,
thermistors were calibrated in an ice bath at 0.00 &#730;C before deployment,
with a one-point correction of the resultant temperature offset made (typically
0.1 K or less). Positions of thermsitors are recorded in meters, with ice
surface at time of deploymeny z = 0 and positive values into ice. Time is
recorded as Julian Day of year.

As noted in the relevant data file headers, some VUW measurements made use of
YSI 55031 thermistors and custom-built loggers (with a temperature precision of
better than 1 x 10-4 K). VUW arrays were of a very different design, and
thermistors were housed inside thin-walled stainless steel tubes with total
thermal conductance closely matched to the ice. A comparison of the different
designs of the VUW and UAF arrays is made in Pringle et al. (2007 submitted).