A major thrust of the current research program is to study the linkages between hydrographic and meteorological conditions and the ice-formation and sediment-entrainment regimes in what has been identified as an important source area for sediment export by sea ice in the eastern Laptev and western East Siberian Seas (see also page on Sediment entrainment and export). In attempting to achieve this goal, however, we are facing two major problems that continue to affect the study of sediment transport by sea ice in the Arctic. First, it has been established that sediment entrainment and export are highly variable and patchy processes both in time and space, rendering systematic studies based on field observations difficult. Furthermore, the critical period of fall freeze-up when the bulk of the entrainment occurs is characterized by extremely adverse environmental conditions (high wind speeds, poor visibility, cloud cover) that hamper both ground- and satellite-based observations of the ice-growth and entrainment process.

We are addressing this problem by employing a combination of satellite-remote sensing techniques (mostly passive microwave, SSM/I, and visible/IR-range AVHRR with ancillary SAR data) and integrating the data products derived from these with results from a large-scale numerical ice-ocean model. As a first step in tracking ice formation and export over the shallow shelves, we are deriving the sea-ice distribution from 85 GHz-channel passive microwave (SSM/I) data over the entire study area. This work required adaptation and extension of ice-edge extraction algorithms and we are currently in the process of publishing results of this work. The end product of this work are maps such as those shown in Figure 1 below, indicating the temporal progression of fall freeze-up across the ice-free shelf during the months of September to November for the year 1995. Also shown is the surface hydrography for October 1995 as derived from ship-board measurements (Figure 2).

In a second step, which is currently underway, we are utilizing this technique to derive the interannual variability of ice formation. This work will also be extended to determine ice export based on sea-ice dislocation vectors obtained from sequential SSM/I 85 GHz scenes. As an example of this work, Figure 3 shows the first principal component (PC1, as derived from the eigenfields of the data set) of the entire freeze-up pattern data set derived for the years 1987 to 1998 (explaining 66% of the total variance contained within the data set). While the first PC corresponds roughly to the long-term annual mean, the lesser PCs are indicative of more complex spatial freeze-up patterns, in particular in the vicinity of the New Siberian Islands and in the areas affected by river runoff.

The third step which is currently underway is to identify corresponding patterns in the oceanographic and atmospheric forcing data that are of relevance to ice formation and sediment entrainment (surface water temperature and salinity, sea-level air pressure, wind velocity and direction, air temperature, sea surface height). These data have been derived from hydrographic measurements and from a large-scale ice-ocean model simulation (A. Proshutinsky).

A final, fourth step upon which we have started work is to link these processes to sediment entrainment. This will be achieved by mapping the extent of different classes of sediment-laden ice in the study region from multi-spectral AVHRR data in a technique developed by J. Kolatschek (see also paper referred to in Section on "Sediment entrainment and export"). HRPT data over the study area have been obtained from the NOAA Satellite Active Archive and the Geophysical Institute's HRPT receiving station. These scenes will be classified through supervised and unsupervised maximum-likelihood classification techniques (channels 1 and 2) to map the distribution of sediment-laden fast (and possibly drifting) ice.