atmosphcre-cryosphere-hydrosphere energy exchanges and can result 
in accelerated ice formation or ice ablation, depending on the time of 
year (Gudmandsen, 1985) and region. In addition, wind induced ice 
floe motion creates a surface stress gradient which modifies the 
ocean mixed layer. This may be the cause of water upwelling 
observed at ice margins (Carsey and Zwally, 1986; Gudmandsen, 
1985). These upwelling areas bring nutrient rich deep water, to the 
euphotic zone, where photosynthetic processes make use of organic 
carbon in the production of phytoplankton. Proliferation at this 
level of the food chain stimulates production at all other levels, 
making these areas biologically important. Where ice motion is 
inhibited, a sea ice cover can act to reduce wind/wave surface stress 
which, in turn, dampens surface turbulent exchange with the 
atmosphere (OIES, 1988). 
It has been proposed (Bums, 1990) that, since SAR is sensitive to 
microscale surface roughness (particularly during the winter season), 
there may be a relationship between the atmospheric drag over the 
ice surface, and roughness of that surface measured by a Synthetic 
Aperture Radar. Anderson (1987) summarized various observations 
of the relationship between floe size, concentration, and ice 
roughness with atmospheric drag. From observations made during 
the Marginal Ice Zone Experiment (MIZEX) Anderson found that, 
for a neutral atmospheric stratification, the coefficient of drag (Ctf) 
increases as the ice concentration increases and that, if concentration 
is held constant, Cj is a function of flow size and ice roughness. 
Bums (1990) found that in areas of relatively smooth pack ice, the 
backscattering information contained within SAR was at a scale 
inappropriate to infer the atmospheric drag coefficient. The presence 
of ridges in pack ice, and the sensitivity of SAR to the frequency 
and height of ridges is presented as an area requiring future research. 
However, SAR provided good estimates of Qj within a Marginal Ice 
Zone (MIZ). Within SIMS we intend to validate the type and range 
of roughness elements which can be detected in SAR imagery. 
These measurements will be coupled with in situ observations of 
wind profiles. 
Snow cover is essentially transparent in the winter season at 
microwavelengths (Kim et al., 1985). There is, however, evidence 
to suggest that snow cover plays a role in determining SAR 
backscatter by altering the thermal structure at the ice/snow 
interface. It has also been suggested that during the fall season the 
highly saline snow cover, created by the destructive metamorphosis 
of frost flowers, has a strong influence on the backscatter from SAR 
imaging system (Drinkwater and Crocker, 1988). 
4. SEA ICE INFORMATION SYSTEM 
The primary objectives of SIMS is to conduct research on the 
physical processes active in the floating ice regime and to develop 
methods by which these variables can be extracted from various 
remote sensing imagery. The development of appropriate data 
management and analysis tools is integral to the success of this 
project. In the remainder of this paper we concentrate on the design 
of an advanced geographic information system, the Sea Ice 
Information System (SIIS), which is being developed to assimilate 
all of our data sets into a single, coherent package. 
The primary objectives of SIIS is to provide a computer based 
framework by which the various datasets from SIMS can be 
efficiently manipulated and to develop a methodology by which 
historical data sets can be integrated with current measurements for 
monitoring climate change. Two principle tasks are required within 
the framework of SIIS: 
• Provide a flexible research environment whereby in situ 
observations can be linked with concurrent remote sensing 
imagery and where synergistic image datasets can be integrated for 
reduction to Sea Ice geophysical or climatological variables. 
• Provide a research environment whereby Sea Ice geophysical or 
climatological variables, extracted from remote sensing imagery, 
can be incorporated into climate modelling studies of the Arctic 
floating ice regime at both local and regional scales. 
3.4 Sea Ice-Cloud Interactions 
Very little information is available regarding cloud cover and sea ice 
extent, though there have been correlations between ice extent and 
cloudiness (Crane and Barry, 1984). The problem is a difficult one 
due to the compensatory effects clouds have on the shortwave and 
longwave radiation balances; relative domination of either flux is 
sensitive to the season and the nature of the underlying sea ice 
surface (Shine and Crane, 1984) and cloud character. Maltese et al. 
(1984) found that the simultaneous presence of ice-albedo and cloud 
amount-temperature feedbacks greatly amplifies the magnitude of a 
climatic response to an extreme forcing. Barry et al. (1984), Crane 
and Barry (1984), Shine et al. (1984), Shine and Crane (1984), and 
Shine and Henderson-Sellers (1984) provide detailed insights into 
this issue. 
The spectral albedo and thermal fluxes are altered by changing cloud 
cover. The former by decreases in radiation (all wavelengths) 
incident on the ice surface, the later by re-radiation to the surface 
from the warmer cloud surface. Herman (1980) reports values for 
net longwave radiation at the surface of -60 W m"^ versus -32 W m' 
2 for cloud-free versus cloud covered conditions in January for a test 
area in the central Arctic (Bany 1983). 
Thermal infrared wavelength imagery coupled with SAR images of 
the ice surface will allow identification of this cloud-over-ice 
condition. Generalizations about the radiation balance over the ice 
surface can be inferred by determining relationships between short 
wave and long wave components associated with different ice 
surfaces in overcast conditions. 
3.5 Snow Cover 
Sea ice snow cover is important due to its thermodynamic effects 
(Carsey, 1984) and its control on surface albedo (Scharfen et al., 
1987). (The role of snow cover albedo has been discussed in the 
Albedo sub-section). The timing of snow melt initiation varies, 
causing significant impacts on the Arctic Basin heat and mass 
balance (due to the strong control exerted by surface albedo) which 
could ultimately affect the long term stability of the ice (Fletcher, 
1966; Barry, 1985, in Robinson et al., 1987). To compound the 
problem, SAR signatures of sea ice are also strongly controlled by 
the onset of melt. The dielectric constant of the snow/ice matrix 
changes dramatically as the amount of free water increases. This 
causes a shift in the backscattering characteristics as recorded by 
Synthetic Aperture Radar (Livingstone et al., 1987). 
When complete, the Sea Ice Information System (SIIS) will be a 
flexible research utility for the analysis of sea ice as an indicator of 
climate change and variability. It will have the capability to 
integrate data from in situ measurements with those acquired by 
space-borne sensors into a single operation; it will be able to 
compare current observations with trends derived from historical 
charts and tables. Within the design objectives we intend to make 
SIIS a multi-format, multi-temporal geographic information system 
(Figure 2). 
Remote Sensing | 
In Situ h 
HlSTÔRICAL 1 
IMAGERY 
Measurements | 
ICE CHARTS 1 
ÎUREMEK^^J 
Data 
¡d® 0©(§ 
^ANALYSEsjJ 
VALIDATION 
Classification 
Figure 2: Sea Ice Information System 
4.1 Multi-Format Data Processing 
Many scientists make use of a number of different systems in a 
single project to take advantage of processing capabilities that some 
systems have and others do not. In practice, this sharing of spatial 
data among several different systems is difficult because of natural 
incompatibilities in the techniques used to model our spatial 
environment and because of designed incompatibilities in the 
analysis systems we use. In the first case, differences in data 
structures (i.e. vector vs. raster) and formats (e.g. row-order vs. 
quadtree raster or topologically encoded vs. spaghetti vector) present 
very real barriers even to experienced users. We should be free to 
devote all of our analysis time on a system to analyzing our data; 
we shouldn't be concerned with their formats. Secondly, most 
system developers see their processing system as the ultimate 
destination during analysis. Consequently, facilities for importing 
data are much more common than facilities for exporting data. 
CLIMATE 
MODELS 
Statistical 
Analyses 
109