Horizontal section through MRI data
(corresponding vertical section shown below).

To learn more about the segregation and 3-D distribution of pores in sea ice, conventional, destructive analysis techniques do not offer as much insight as one would wish. In part this is due to the standard sea-ice sample storage and analysis procedure which involves sample cooling to low temperatures (typically -25 to -30 C) prior to sample analysis.

In collaboration with the Alfred Wegener Institute in Bremerhaven, Germany, we have attempted to adapt a powerful, non-destructive method, magnetic resonance imaging (MRI), to the study of sea ice under variable temperatures in a specially designed enclosure. Moreover, we have taken the greatest possible care to maintain the samples at or near the in-situ temperatures throughout the sample storage and analysis process. This was facilitated by the convenient locations of Barrow and the University of Alaska Fairbanks that allow rapid sample transfer in insulated sampling containers (and Dewar vessels for smaller samples) at close to in-situ temperatures.

Vertical section through MRI data
(corresponding to horizontal section shown above).

Enlargement of pores from vertical slice shown above.

Studies of artificial sea ice provide considerable insight into the 3-D arrangement of pores in sea ice and furthermore allow us to scrutinize the pore/ice-matrix morphologies that are associated with the segregation of pore space into isolated pockets.

More important, warming of samples from in-situ temperatures to higher values allows an assessment of the increase in pore size and shape without interfering with the sample. Image analysis of these data sets indicates that pores typically increase in vertical elongation rather than horizontal cross-section upon warming and that they furthermore link up into connected, larger pores associated with a decrease in pore number. This finding is consistent with the notions discussed above. However, the issue of scale plays an important role here, since MRI in this case provides a resolution of at best around 60 to 100 µm. To delve into the scale of individual ice organisms and in particular bacteria, we have to resort to microscopy, though in this case microscopy with a twist, that allows us to locate and study sea-ice organisms under in-situ temperatures by adapting fluorescent staining methods for this low-temperature purpose (this is the second area where we have had to struggle, though eventually successfully, with the methodological problems that are an integral part of very-low-temperature, high-salinity sea-ice research).

This image shows an overview and a close-up of brine layers and pockets occurring both within and between ice crystals in one of our samples. We have developed an image-analysis and classification scheme that allows us to derive the size and distribution parameters for roughly a dozen of different pore/inclusion features in the ice and furthermore helps in mapping the distribution of organisms within this matrix. As visible in the close-up, intragranular inclusions even at a microscopic scale pinch off at lower temperatures, with the brine veins eventually displaced by ice bridges. Intergranular inclusions (grain-boundary brine sheets/tubes), however, persist even to very low temperatures. In part, this is explained by phase chemistry and thermodynamics with impurities depressing the freezing point to low temperatures. However, as evident from the pore size distributions (widths in the range of few micrometers at low temperatures), part of this unfrozen water is also maintained by dispersive, interfacial forces. The latter as well as natural temperature variations, are likely to induce significant brine volume fluxes through the connected pore volume persisting at very low temperatures.

These images show an example of the successful adaptation of fluorescent staining techniques to low-temperature applications down to between -15 and -25 C. Digital imaging systems help in locating and identifying bacteria, diatoms and other inclusions. Only fluorescent staining allows the actual localization and identification of bacteria in such a case.

Furthermore, our research has shown that aggregates of sedimentary particles are typically concentrated along grain boundaries or grain triple-point junctions, possibly as a result of fluid flow induced as discussed above.