The papers in this supplementary issue of Arctic emerged from the First International Circumpolar Symposium on Remote Sensing of Arctic Environments held in Yellowknife, Northwest Territories, 1-3 May 1990. From 1987 to 1990 personnel from the territorial government, the federal government and the private sector were trained in remote sensing techniques and applications as part of a remote sensing technology transfer program operating in the Northwest Territories. Demonstration projects were undertaken in a variety of fields and the results showed that remote sensing can be an important tool in natural resource management in the North. In order to share the knowledge and experience gained from this program and to exchange information on other programs in the circumpolar regions, a symposium was organized. This was the first symposium to deal specifically with remote sensing applications in northern environments, and the results proved to be of interest to scientists, scholars, and professionals involved in renewable and non-renewable resource management. It provided a forum for the exchange of current applied international research, the presentation of new technologies, and the advancement of international cooperation in the circumpolar regions of the world. The seven plenary sessions focused on the application of remotely sensed data to resource monitoring and management and included facilities and programs, remote sensing techniques, oceanography, hydrology, snow and ice, wildlife and wildlife habitat, geology, forestry and vegetation, and radar remote sensing. Over 80 scientists from Greenland, Norway, Denmark, the United States, Great Britain, and Canada attended the three-day symposium, which was jointly sponsored by the Department of Renewable Resources, Government of the Northwest Territories, and the Canada Centre for Remote Sensing, Energy, Mines and Resources Canada. ... This was the first of, we hope, many symposia at which northern scientists will be able to share their knowledge of the uses of remotely sensed data and geographic information systems.
The science plan for the Alaska SAR Facility (ASF) focuses on earth surface characteristics that are of interest within the overall concept of global change and that show significant regional, seasonal and interannual variations resulting in changes in the strength of their radar returns. The polar oceans, with the continuous motion and deformation of the pack ice and the changes in the surface state of the surrounding open seas, offer excellent opportunities for such research. Because such studies require both frequent and detailed analysis of Synthetic Aperture Radar (SAR) data, a Geophysical Processor System (GPS) has been developed to speed the extraction of useful geophysical information from SAR data sets. The system will initially produce three main types of products: (a) sets of ice motion vectors obtained by automated computer tracking of identifiable ice floes on sequential images, (b) the areal extent and location of several different ice types and open water and (c) a characterization of the wave state in ice-free regions as well as within the ice in the marginal ice zone at locations where significant wave penetration occurs. Details of these analysis procedures are described. Initially the GPS is planned to process 10 image pairs/day for ice motion, 20 images/day for ice type variations and 1 image/day for wave information, with a total estimated processing time of 13 hours. A variety of projects plan to utilize the SAR data stream in studies of ice, lead and polynya dynamics and thermodynamics. A common feature of these research programs will be attempts to provide, via the coupling of the SAR data with ice property and ice dynamics models, improved estimates of the heat and mass fluxes into both the atmosphere and the ocean as affected by the characteristics of the ice cover.
There is ample evidence of the need to expand the North's natural resources information base. It is equally important that this information be made available and accessible to the people of the North. Remote sensing and geographic information systems technologies will be important tools in this effort. To fully meet these objectives, however, significant efforts will be required in the area of training. This paper considers the unique training challenges that will have to be met in the North. The authors discuss a training strategy for remote sensing and geographic information systems (GIS) in northern Canada. This strategy attempts to make maximum use of existing training resources and recognizes the practical need to develop operational skills rapidly enough to fuel the broader effort to supply timely and reliable resource information over all of the Northwest Territories. The strategy stresses the special needs of native peoples to participate fully in the development process. To achieve this, the authors draw on the experience of the Canada Centre for Remote Sensing (CCRS) Technology Transfer Program and Arctic College.
Satellite images and digital evaluation models were analyzed to interpret and quantify vegetation communities and active geomorphic surfaces in a mountainous area in southwest Yukon, Canada. High levels of discrimination were determined for the digital satellite and terrain data when compared to field studies and aerial photo interpretation of basic biophysical units, specific vegetation cover types and geomorphic process categories. The agreement between field identification of a site and discriminant analysis of that site using the digital data as discriminating variables ranged from 60 to 85% and contained improvements of up to 20% when topographic data such as slope angle and incidence value or aspect were added to spectral discriminant functions. Active geomorphic surfaces were grouped successfully into process categories such as landslides, debris flows, solifluction and talus sorting. Visual interpretation of the changes in the landscape detected using Landsat Thematic Mapper imagery from 1985 and SPOT HRV MLA imagery in 1989 were attributed to (1) running water in the alluvial deposits and organic terrain, (2) different water levels in the river, delta and floodplain, and (3) a general trend of class change from wet to dry throughout the study area. This latter change may be a result of imaging the same class (e.g., alpine tundra) under more senescent (i.e., brown) conditions and may be explained with reference to a warming/drying trend in the intervening years.
Monitoring the water bodies of the Mackenzie Delta by remote sensing methods
De Lisle, D.
Arctic, v. 44, suppl. 1, 1991, p. 21-28, ill. (some col.), map
ASTIS record 31323
In the Mackenzie Delta, Northwest Territories, the thousands of lakes, ponds, channels and waterways, connected in an apparently chaotic manner, present a major logistical problem for collecting information regarding the nature of this complex hydrologic system. The use of satellite images gives an economical and synoptic view of this isolated region, while special analysis techniques simplify the environmental appraisal. The use of (mathematical) morphological analysis of the surface waters imaged by the satellite allowed the authors to distinguish all water bodies, even when they were at the limit of the spatial resolution of the sensor. The technique further permitted the classification of these water bodies by their inter- and intra-connectivity. Another technique, termed chromaticity analysis, allows for the removal of atmospheric differences among images, which in turn enables the use of surface calibration data from one date to be used on images of other dates. This method was also used to generate quantitative maps of suspended sediment concentration levels. Together, these techniques permit the assessment of the hydrologic flow (or its hindrance) of sediment and nutrients for the sustenance of aquatic flora and fauna. They further supply a method for the mapping of access routes by water craft to all parts of the Delta.
The Ice Branch of Environment Canada's Atmospheric Environment Service is responsible for providing information about ice conditions in Canada's offshore areas. Principal clients include the Canadian Coast Guard, commercial shipping companies, the oil and gas industry and fishermen, who all require accurate information on ice distribution in near real-time. In order to provide this information, the Ice Branch employs a Challenger jet aircraft equipped with dual wide-swath synthetic aperture radars and a Dash-7 equipped with a real aperture side-looking airborne radar. These aircraft image approximately 100 million sq km annually. Radar image data from these aircraft are downlinked during flight via an S-band telemetry link to Coast Guard icebreakers and to satellite stations in the Ice Reconnaissance Data Network, which relays the data to the Ice Centre in Ottawa. There, the data are integrated with remotely sensed data from satellites and are used as the basis for ice analysis charts produced and distributed by radio facsimile on a daily basis.
A major limitation for salmon (Salmo salar L.) farming in arctic environments is the low winter temperatures influencing the salmon's growth rates, mortality and quality. A detailed knowledge of the sea temperature variations in a region can help to avoid the establishment of fish farms in areas that are less suitable. In order to supply local fish farmers and planning authorities with such information, a satellite survey of sea surface temperatures in a late winter situation was conducted in northern Norway. Landsat Thematic Mapper data were calibrated with in situ measurements. The relationship between sea surface temperatures and other factors in the physical environment was visualized in a very comprehensive way. Temperature zones were found to be consistent with information in literature and of relevance to the fish farming industry. New, potentially suitable sites for fish farming could be indicated in many areas where no historical data were available.
Remote sensing of permafrost by ground-penetrating radar at two airports in Arctic Canada
Arctic, v. 44, suppl. 1, 1991, p. 40-48, ill., map
ASTIS record 31326
Over a cycle of seasons, ground-penetrating radar studies were carried out at Inuvik and Rankin Inlet, Northwest Territories. The two airports are part of the Forward Operating Location (FOL) program of the Department of National Defence and have been slated for significant upgrading of runways, taxiways and parking aprons. This provided a good opportunity to investigate the extent of permafrost and its seasonal variation at two locations distributed over a wide geographic area. The study method involved specific and repeated traverses with a Pulse EKKO III ground-penetrating radar unit. The survey was successful in locating massive ice bodies, imaging several existing problem areas beneath runways and mapping the seasonal depth of thaw in permafrost. The study results imply that future monitoring at FOL sites should be continued in light of suggested ground stability problems due to global warming.
SPOT satellite data were used to detect and map muskox habitat on Devon Island, N.W.T. Muskox habitat in the Canadian High Arctic is restricted to small islands of productive sedge meadow isolated within a matrix of sparsely vegetated polar desert. On Devon Island, muskox herds move among small lowlands on the northeast coast adjacent to Jones Sound in response to the seasonal availability of sedge-dominated habitat. Comparisons between the enhanced satellite images and species composition, plant cover, and standing crop on these lowlands showed that sedge meadows were spectrally distinct from the dwarf shrub/heath and cushion plant/lichen-moss types on beach ridges and rock outcrops, indicating that spectral data can be used to identify critical forage habitat for muskoxen in the High Arctic. The hummocky sedge/moss meadow and the less productive frost boil sedge/moss meadow types could be separated from each other on the enhanced imagery. The satellite data were simplified using a supervised classification to document the type and areal cover of muskox habitat along the northeast coast of Devon Island. The spatially isolated hummocky sedge/moss and frost boil sedge/moss meadows occupied only 3% (16.73 sq/km) and 6% (32.84 sq/km) respectively of a total land area of 549.38 sq/km.
Calibration of aerial thermal infrared imagery for walrus population assessment
Arctic, v. 44, suppl. 1, 1991, p. 58-65, ill., maps
ASTIS record 31328
Concurrent aerial photography and emitted thermal infrared (10.6 µm) imagery were acquired over walrus hauled out on sea ice in Foxe Basin, Northwest Territories, Canada. Digital thermal infrared data from a Forward Looking Infrared (FLIR) imager provides a method for estimating walrus numbers, since the objects (walrus groups) are considerably warmer than the background (ocean and sea ice). Coincident photographic counts and thermal infrared pixel counts are regressed by means of a least squares linear regression and an estimate of group size predicted from the number of pixels represented by each group. The results indicate that analog thermal imagery provides an effective means for obtaining a stratification variable that can subsequently be used in survey design. The FLIR walrus estimation approach is evaluated in the context of replicability, both through the physical mechanics and within a limited range of environmental conditions. Precision of FLIR estimated walrus counts is evaluated relative to the precision of photo counts. Walrus numbers extracted from the digital thermal imagery at a sample swath of 1778 m are as precise as those obtained from three independent photo counts at a sample swath of 686 m. In this configuration the FLIR provides a 160% increase in the sampling (area) fraction. Based on these results, we recommend a stratified approach to estimating walrus abundance using a thermal infrared sensor coupled with visual and photographic censusing techniques. We conclude with recommendations for continued analysis of this infrared censusing technique.
Detection and classification of muskox habitat on Banks Island, Northwest Territories, Canada, using landsat thematic mapper data
Arctic, v. 44, suppl. 1, 1991, p. 66-74, ill., maps
ASTIS record 31329
The feasibility of using Landsat Thematic Mapper data for mapping muskox summer habitat was tested on northern Banks Island, Northwest Territories. Digital image enhancement and classification techniques were examined to determine if summer foraging habitats could be detected and mapped using Thematic Mapper imagery. Interpretations of the satellite data were verified in the field during the summers of 1988 and 1989. The most important summer foraging habitats for muskoxen included the wet sedge meadow, graminoid tundra and graminoid/dwarf shrub tundra cover types. These lowland habitats were generally distinguishable on enhanced colour images and were easily differentiated from upland areas. The most suitable colour composite for differentiating muskox summer habitats was the near-infrared (band 4), shortwave infrared (band 5) and red (band 3) spectral bands displayed in red, green and blue respectively. Upland cover types, including dwarf shrub tundra, hummocky tundra and dwarf shrub/lichen barrens, were more difficult to differentiate because of spectral variability resulting from differences in plant cover and site characteristics. The classified image had an overall accuracy of 88%. The summer habitats of particular importance to muskoxen had classification accuracies of 84-89%. Detection of important foraging habitats on Thematic Mapper imagery is attributable to the spectral distinctiveness of wet graminoid communities and the high spectral sensitivity and spatial resolution of the infrared sensors, which allow detection of differences in surface moisture and vegetation physiognomy.
An assessment of bison habitat in the Mills/Mink lakes area, Northwest Territories, using landsat thematic mapper data
Arctic, v. 44, suppl. 1, 1991, p. 75-80, ill., maps
ASTIS record 31330
Landsat 5 Thematic Mapper data were used to classify and map wood bison (Bison bison athabascae) habitat in an area of 3383.5 sq/km in the vicinity of Mills and Mink lakes, Northwest Territories. Digital image analysis techniques were used to enhance and classify satellite data acquired 31 August 1986 to determine the distribution and extent of forage habitats. Visual interpretation of cover types was carried out using a radiometric enhancement, consisting of a power contrast stretch, applied to spectral bands 5 (shortwave), 4 (near infrared), and 3 (red) and displayed as a colour composite image (red/green/blue). A DIPIX Technologies ARIES III image analysis system was used to perform an unsupervised classification of 16 spectral classes using a maximum likelihood classifier. The resultant thematic map was used during the summer of 1988, when aerial and ground surveys of the study area were carried out. Colour infrared photographs (1:20,000 scale) obtained in 1988 were also used as part of the reference data set to assist in selecting training areas for use in a supervised classification. Ground data were also collected during the summer of 1989 to increase the number of training areas and provide test sites for an accuracy assessment. In the final habitat classification, the ten physiognomic cover types were water, aspen forest, coniferous/mixed forest, sedge meadow, willow savanna, hordeum meadow, wetland complex, peat plateau/burned meadow, burned peat plateau, and burn-aspen regeneration. The Thematic Mapper data was geometrically corrected and smoothed using a post-classification filter. Results of the digital image analysis were produced as hard copy thematic maps and tabular summaries. Overall accuracy of the classification was 91% with the two important winter forage habitats - sedge meadow and willow savanna - having accuracies of 95% and 75% respectively. These forage habitats collectively represent 4.6% of the study area, or 153.3 sq/km. Digital image analysis of Landsat Thematic Mapper data proved to be an effective and cost efficient method of mapping bison habitat in a large area of the boreal forest.
Airborne, narrow swath, C-band synthetic aperture radar (SAR) imagery, obtained from the Yellowknife-Hearne Lake area, essentially reflects the geomorphology or landforms of the region. These in turn can be readily related to specific lithologies, rock masses, structure and cultural features. Terrain analysis using textural and tonal (brightness) characteristics of the radar images along with drainage and lakeshore characteristics permitted definition of several lithologic classes: Granite terrain type-1, generally the brightest (lightest) area, has a "coarse" mottled signature, reflecting the hummocky surface characteristic of granites in this area. Metasedimentary terrain is typified by an intermediate tone, a thinly laminated texture reflecting bedding and angular shorelines of some lakes. Metavolcanic terrain is subordinate in area and lacks well-defined textural or tonal characteristics. It is most easily recognized as parallel ridges with little or no curvature. The city of Yellowknife is readily identifiable by its bright signature and rectangular pattern or texture. Lineaments, recognized by the alignment of rivers and shorelines, are greatly enhanced by bright radar reflections from northerly facing cliffs and radar shadow (zero signal return) of southerly facing cliffs. Several major structural lineaments in the area, known from aeromagnetic and geological maps (diabase dykes, faulted contacts, etc.) are readily apparent in the SAR imagery, as are numerous extensions and subsidiary lineaments. Circumstantial evidence suggests that post-Precambrian and neotectonic activity may be related to lineaments.
The application of the Normalized Difference Vegetation Index (NDVI) for monitoring natural vegetation and biomass production has been evaluated for a sheep farming area in southern Greenland. Field measurements of spectral reflectance data during the growing season have been used to make a calibration between NOAA AVHRR NDVIs and aboveground vegetation quantities. The potential biomass production is estimated as the product of mean NDVI and the length of the growing season. Lowest-order atmospheric as well as geometric corrections were carried out on the satellite data before the seasonal and regional variations were correlated with climate and water balance. Agriculture in southern Greenland started when Eric the Red came from Iceland around 982 A.D., and the Norse era ended approximately 500 years later because of climatic change, extensive overgrazing and soil erosion. Modern sheep farming started in 1924, but the threats to sheep breeding and the environment are the same today as during the Norse era. The satellite-based monitoring has proved to be a useful tool to avoid overgrazing, which in this foehn-affected area easily implies soil erosion. It is a quick and low-cost method, and in combination with meteorological and soil water data it is possible to forecast the dry biomass production at the beginning of each growing season. This facilitates agricultural management and planning of the potential breeding capacity in this vulnerable marginal environment.
An ARIES III image Analysis System was used to manipulate digitized Landsat 5 Thematic Mapper (TM) imagery of the Yellowknife Volcanic belt (YVB) to demonstrate geological applications of this technology in a boreal region characterized by subdued topography, open-canopy forest cover and about 50% outcrop exposure. The YVB was selected for study because detailed geological mapping is available against which computer-generated imagery could be compared and evaluated. The study area encompasses a northerly trending belt of principally mafic volcanic rocks (Kam Group) flanked to the west by a multi-phase batholith (Western Granodiorite Complex) and to the east by turbiditic sediments (Burwash Formation) and granitic stocks. Intermediate volcanic rocks (Banting Group) and a narrow belt of conglomerates (Jackson Lake Formation) separate the Kam Group from the Burwash Formation. Initially, images from each of the six TM channels were analyzed, and it was found that channels 7 (short-wave infrared), 3 (red) and 2 (green) yield the best discrimination of geological features. Subsequently, various computer-generated enhancement programs were applied to these channels to intensify geological discrimination, and the resulting images were combined to produce a number of false-colour composite images that were compared to published geological maps. On the best of these "satellite geological maps" at least three phases of igneous intrusions can be identified, as well as the boundaries of the Kam Group and a number of important structural discontinuities. The adjacent Banting Group volcanics and Burwash Formation sediments could not be chromatically separated but are sometimes distinguishable using textural criteria. Most notably, certain computer-enhanced images highlight two zones of alteration within the Kam Group that have not previously been noted on published maps. These alteration zones appear to be spatially associated with a number of gold showings and one significant gold deposit. It is concluded that computer manipulation of satellite imagery can add worthwhile information to the geological database of areas that have been mapped at a detailed scale of 1:10,000.
Measuring climatic state variables from SAR images of sea ice : the SIMS SAR validation site in Lancaster Sound
Arctic, v. 44, suppl. 1, 1991, p. 108-121, ill., maps
ASTIS record 31334
In this paper we consider three aspects of arctic sea ice climate studies: a review of climate processes related to a seasonal sea ice cover, a review of how climate state variables related to the reviewed processes may be measured using remote sensing data, and an analysis of the relative utility of tonal versus textural clarification of synthetic aperture radar (SAR) images of sea ice. The two reviews are presented in the context of a field validation experiment called the Seasonal Sea Ice Monitoring Site (SIMS) being conducted in Lancaster Sound/Barrow Strait, Northwest Territories (N.W.T.) The analysis of tone versus texture as a means of extracting ice type information from SAR images of sea ice is conducted using X-band, HH polarized data from the SIMS '90 field experiment and data collected over Mould Bay, N.W.T.
Imaging radars have been in use in the Canadian Arctic for over 20 years. Initially the use was sporadic, as the relatively new, declassified technology in the form of real aperture side-looking airborne radars (SLAR) was flown and the results studied. This situation existed until the late 1970s, when the use of two types of imaging radars became more widespread. The Atmospheric Environment Service (AES) introduced the Motorola APS-94 SLAR for use on regular reconnaissance flights, while the Canada Centre for Remote Sensing (CCRS) introduced the CV-580 X-L Synthetic Aperture Radar (SAR) for periodic missions in the Arctic for research in support of ice studies and shipping. As demand for ice information increased in support of offshore drilling in the Beaufort Sea and navigation in the Eastern Arctic and along the east coast, more systems were brought on line. AES added two more SLARs to their reconnaissance efforts in the early '80s, while Intera developed a digital SLAR and two digital SAR systems, STAR-1 and STAR-2. As part of a multi-year program to support AES's ice reconnaissance mandate in the Arctic and east coast areas, Intera has developed a dual-sided SAR in a jet aircraft for high-resolution, large-area coverage. Imaging radar, with its all-weather, day/night and cloud-penetrating capability, has proved to be the almost ideal sensor for many arctic applications. In support of offshore drilling in Alaska and Canada, large areas were flown to obtain up-to-date information for use in navigation and forecasting ice conditions. Real-time SAR and SLAR data can be downlinked to ships navigating in ice-infested waters to aid officers in determining the safest, most efficient and economical routing through the ice. Research into ice properties and signatures has improved our knowledge and understanding of the ice, which covers a large part of Canada's territorial waters for much of the year.
The use of AVHRR thermal infrared imagery to determine sea ice thickness within the Chukchi polynya
Arctic, v. 44, suppl. 1, 1991, p. 130-139, ill., map
ASTIS record 31336
Sea ice thickness changes over a nine-day period are determined for the Chukchi Polynya using Maykut's (1986) and Kuhn et al.'s (1975) theoretical predictive models. The models relate ice thickness to sea ice surface temperature, air temperature, wind speed, and sea water temperature. Sea ice surface temperatures are derived from AVHRR imagery and meteorological observations are taken from the synoptic weather station at Barrow, Alaska. The Maykut equation yields results that appear to be realistic for the ice thickness distribution within the polynya at the beginning stages of polynya formation. Ice thickness calculations at the later stages of polynya formation are partially invalidated by the movement of large floes to the oldest part of the polynya in response to a wind from the northeast. Such a major disturbance on the surface of the polynya complicates the identification of the type and thickness of ice that is forming. These results offer encouragement for the prospects of future field studies to validate and refine the technique and for the extension of the technique to calculation of heat transfer and salt rejection within the Chukchi Polynya and other polynyas.
Surface roughness at the centimetre and millimetre scale is an important factor governing radar backscatter, especially in the case of warm (>-5 degrees C) or highly saline sea ice types. Quantitative measurements of surface roughness are required as input to backscatter models. Several field techniques have been used to quantitatively measure the surface roughness of sea ice. These techniques usually posses at least one of the following obstacles: difficult field operation, expense, poor accuracies or arduous data processing. A prototype portable field instrument called the Surface Roughness Meter has been designed to measure micro-scale surface roughness. The instrument provides measurements of two surface roughness parameters, root mean square height and correlation length. The instrument consists of a 35 mm camera and a flash mounted on a platform. The system illuminates and photographs a rectangle of known size on the surface from a fixed height. The negatives are digitized and the root mean square height and correlation length are calculated and recorded using a PC-based image analysis system in the laboratory. The first sea ice application for the instrument was the Labrador Ice Margin Experiment (LIMEX) 1989. The instrument was used to measure surface roughness of first-year deformed pack ice. The resulting data from LIMEX '89 were digitized and surface roughness statistics were computed using a PC image analysis system. LIMEX '89 Surface Roughness Meter data compared favourably to roughness statistics obtained from LIMEX '87.
Derivation of snow water equivalent in boreal forests using microwave radiometry
Arctic, v. 44, suppl. 1, 1991, p. 147-152, ill., maps
ASTIS record 31338
Efforts have been made by several investigators to produce a reliable global microwave snow algorithm to estimate snow depth or snow water equivalent (snow volume) and snow extent. Complications arise when trying to apply a global algorithm to specific regions where the climate, snowpack structure and vegetation vary. In forest regions, the microwave emission from dense coniferous forests may overwhelm the emission from the underlying snow-covered ground. As a result, algorithms employing microwave data tend to underestimate snow depths. Preliminary results indicate that the amount of underestimation can be minimized when the fraction of forest cover can be accounted for and used as an additional input in microwave algorithms. In the boreal forest of Saskatchewan, the standard error between the measured and the estimated snow water equivalent was reduced from 2.7 to 2.1 cm by using a generalized snow retrieval algorithm that includes the percentage of forest cover. However, perhaps as much as 25% of the boreal forest of North America and Eurasia is too dense to enable satisfactory snow water equivalent determination to be made using passive microwave techniques alone.
Multichannel airborne SAR data were collected over northern Manitoba in April 1989 and January 1990. During the week of the SAR flights, several reconnaissance helicopter flights were undertaken, and ground calibration sites were visited to collect ice, snow, and water data. A total of six SAR image passes were flown in April 1989 and seven in January 1990, in order to collect a data set with numerous incidence angle, frequency, polarization, and look direction combinations. The data have been qualitatively assessed, with specific emphasis on C-band horizontally polarized imagery - the proposed SAR configuration for Radarsat. Results of the analysis have shown that airborne SAR can be used to identify various freshwater ice features, such as juxtaposition ice, refrozen slush, river ice runs, and lake ice. Open water leads were also successfully identified. A careful interpretation of the airborne SAR imagery in conjunction with the ground truth data has shown that the unusually bright returns characterizing the Burntwood River and the west portion of Split Lake were caused by a layer of refrozen slush that was generated during the initial formation of the ice cover. Although the results reported here focused exclusively on a qualitative analysis of C-HH data, preliminary analysis of the digital data suggests that changes in frequency and polarization produce measurable differences and can be used to develop classification algorithms for freshwater ice.
AVHRR imagery has been used to document the sites of 22 polynyas in the Bering and Chukchi seas. Two principal classes of polynyas have been identified that tend to be negatively correlated: (1) persistent polynyas, which are present much of the time and form off south- and west-facing coasts, and (2) north coast polynyas, which are only occasionally open and form off north-facing coasts. Median extent values characterizing 17 of these polynyas for six years have been compiled for the winter and spring months, and the results of attempts to correlate these values with synoptic meteorological measurements are reported. These attempts were not very successful, suggesting that more sophisticated approaches to the problem are required. Other factors, such as currents, may play a principal role in determining actual polynya extent.