This study links two models, one that simulates changes in snow microstructure and one that recovers microstructure properties from measurements of snow reflectance. An energy and mass transfer model, SNTHERM.89, was used to calculate snow grain growth. Grain-sizes from the model and measurements of grain bond areas provided estimates of the surface-to-volume ratio of the bulk snow, which were transformed to geometrically-equivalent sphere sizes. An inversion technique based on a discrete-ordinate model of the directional reflectance recovered optically-equivalent sphere sizes from reflectance measurements at 1.075 μm. The predictions of equivalent sphere sizes from the snow model and the recovered optical sphere sizes from the inversion method were compared with stereological measurements from snow sections. The geometrically-equivalent and optically-equivalent grain-sizes showed good agreement with each other and with stereological measurements from snow a few days old. The predictions of the reflectance inversion method also compared favorably with geometrically-equivalent grain-sizes measured from a melt-freeze surface crust. This investigation showed the potential for fully coupling snow property simulations with models to predict the spectral reflectance of snow.

We have recently completed an analysis that examines in detail the spatial and temporal variations in global sea-ice coverage from 26 October 1978, through 20 August 1987. The sea-icemeasurements we analyzed are derived from data collected by a multispectral, dual-polarized, constant incidence-angle microwave imager, the Scanning Multichannel Microwave Radiometer (SMMR) on board the NASA Nimbus 7 satellite. The characteristics of the SMMR have permitted a more accurate calculation of total sea-ice concentrations (fraction of ocean area covered by sea ice) than earlier single-channel instruments and, for the first time, a determination of both multiyear sea-ice concentrations and physical temperatures of the sea-ice pack. An estimate of the SMMR wintertime total ice concentration accuracy of ± 7% in both hemispheres has been obtained. As this is an improvement over the estimated accuracies of previous microwave sensors, we are able to present improved calculations of the sea-ice extents (areas enclosed by the 15% ice concentration boundaries), sea-ice concentrations, and open-water areas within the ice margins. This analysis will be published in a book, Arctic and Antarctic sea ice, 1978–1987: satellite passive microwave observations and analysis, due for publication in1992. Some highlights from the analysis are presented in this paper.

Microwave brightness temperatures from snowpacks are simulated with a multiple-scattering model using observed hydrometeorological variables at three target areas on the Canadian plains. Comparison of model microwave emissions with those observed from the Nimbus 7 satellite allows the derivation of the snowpack properties of grain-size and microwave absorption. A simulated time series of microwave brightness temperature is produced for the winter season of 1884—85, for assessing the utility of multi-temporal satellite observations for snowpack monitoring.

Scanning Multichannel Microwave Radiometer (SMMR) data are compared with the snow characteristics measured during several traverses over Antarctica including average density over 0–2 m, 10 m temperature, grain-size distributions and number of strata. First a local study of specified areas and then a global study show some strong dependence between satellite data and ground data. The polarization ratios are affected by the stratification of snow (number and nature of layers); large polarization ratios correspond to strong stratification, mainly for the lower frequencies. Because low density induces strong stratification there is also a dependence between polarization ratios and density even if the density is not the critical factor. The gradient ratios are linked to the grain-sizes as determined from detailed stratigraphies; the larger the grain-size, the smaller is the gradient ratio between 6.6 and 18 GHz.

Photogrammetric analysis of Eliot Glacier, Mount Hood, Oregon indicates an average thinning of the ablation area of 4.2 m during 1984–89. Thinning is generally more marked near the equilibrium line, where debris cover is absent, than on the lowest part of the glacier, where superglacial debris thickness is greater than 1.0 m. The pattern of glacier thickness change is partly related to the reduction of ablation with increasing debris thickness, though flow kinematics must also be important to charge in ice thickness.

The application of passive microwave radiometry to the remote sensing of snow properties is based on the ratio of emitted to scattered portions of the upwelling radiation. Increased scattering is indicative of increased snow amount, i.e. the number of snow grains present. However, scattering is also directly proportional to snow grain-size for a given snow amount. Current snow cover retrieval algorithms produce inaccurate results when snow grain-sizes are unusually large. Therefore, it is necessary to characterize snow grain-size on a regional scale (and perhaps local scale in extreme situations) in order to adjust passive microwave algorithms. Preliminary analysis indicates that: (1) algorithms are not as sensitive to the presence of large grain-sizes as the initial theory had indicated; (2) standard deviation of grain-size diameters throughout the total snow cover may often be less than 0.5 mm, thus average grain-size data may often serve to characterize the detailed stratigraphy of the total snow cover; (3) conditions in subfreezing snow which produce grain-sizes that greatly exceed a mean diameter value of 1–2 mm result from snow cover/climate relationships which can be modelled/monitored on a regional scale. A preliminary method is investigated for selecting snow retrieval algorithms according to prevailing regional-scale grain-size.

The velocity field of the confluence area of two large ice stream tributaries forming Ice Stream D in West Antarctica is studied using sequential Landsat images. Sequential satellite image analysis allows for a very high density of velocity measurements, based on computer-measured displacements of features such as crevasses, crevasse scars, and ice mounds recognizable in both images. Automated displacement measurement of these features results in a detailed map of surface velocities from which surface-horizontal strain-rate fields can be calculated. Correlations between the surface morphology, the velocity field, and the strain-rate field of Ice Stream D reveal a number of important characteristics of ice stream flow:

• the characteristic flowband appearance of streaming ice is present at velocities from below 100 m a−1 to above 350 m a−1;

• in the upstream areas, there appears to be no sharp transition between “sheet” flow, typical of the surrounding ice sheet, and “streaming” flow;

• the fastest moving portions of the ice stream are nearly devoid of surface topography undulations;

• the confluence area is characterized by acceleration of the ice in the slower tributary as it impinges on faster-moving ice, and by highly convergent flow. Velocity in the faster-moving tributary changes little, and there is no persistent evidence of the shear margins of the joined tributaries downstream of the confluence.

This study demonstrates that sequential satellite image analysis, coupled with computer-determined displacement measurements, can provide accurate velocity and strain-rate information on a regional scale rapidly and cost-effectively. Such data sets are required for modelling ice sheet evolution, and for monitoring any changes in ice flow within the ice streams.

Formation of a surface-hoar/depth-hoar complex at the GISP2 site in central Greenland was correlated with large changes in Special Sensor Microwave/Imager (SSM/I) brightness-temperature data. Pass-averaged SSM/I brightness-temperature data over a 1/2° latitude by 1° longitude cell for the 19 and 37 GHz, vertically (V) and horizontally (Η) polarized bands were manipulated to yield differential (V-Η) trends which clearly show a gradual decline as the hoar formation caused a progressively rougher surface with progressively lower density. The hoar episode ended as snowfall, and high winds buried and destroyed the surface-hoar layer and caused rapid V-Η increases in ≈ 1 day. Comparison of the different trends with changes in the field-monitored variables and theoretical values suggest that the V-Η trends are sensitive primarily to changes in surface roughness, and secondarily to near-surface density changes. Consistent expression of trends in microwave brightness temperature over 35 adjacent study cells indicates that this technique may provide a remote-sensing signature capable of defining the timing and spatial extent of surface- and depth-hoar formation in central Greenland.

A time series of satellite passive microwave observations by the Special Sensor Microwave/Imager (SSM/I) instrument, of the Antarctic marginal ice shelves, reveals sharp transitions in the radiometric brightness temperatures at 19 and 37 GHz. These transitions, sometimes occurring over the period of a single day, indicate the onset of surface melting of the ice. Monitoring the day of the onset of melting and the permanent re-freezing will be a measure of rising air temperatures as a result of global warming. The response from a stable ice shelf, the Ronne, where no appreciable melting occurs, is compared with observations from the marginal ice shelves on Antarctic Peninsula. It is observed that the spectral gradient (Tb19 - Tb37) varies as a function of time on a clear annual cycle. It is shown that variations in the spectral gradient arise due both to the temperature profile within the snow and the dielectric effects of a wet surface layer.

The precision of radar altimetry above an ice sheet can improve glaciological studies such as mass balance surveys or ice-sheet flow models, the first by comparing altimetric data at different times (see this issue), the second by testing or constraining models with data. This paper is a first step towards the latter. From a precise topography deduced by inversion of altimetric data (Remy and others, 1989), we calculate ice-flow direction, balance velocity and basal shear stress. The rheological parameters involved in the relation linking velocity, stress and temperature are then derived by least-squares regression. Ice flow is well represented by setting the Glen parameter, n to 1 ± 0.25 and the activation energy as 70 ± 10 kJ mol−1.

With the development of coupled atmosphere–ocean models for the polar seas, there will be a great need of surface radiation budget data over partially ice-covered sea surfaces of the marginal ice zones. This paper presents an attempt to retrieve the surface radiation budget components over the Fram Strait area from operational satellite data, i.e. AVHRR visible and infrared radiances and SSM/I passive microwave brightness temperatures. The cloud optical thickness, which is the main modulator for the incoming solar flux, is retrieved from AVHRR visible radiances through a radiative transfer model, assuming surface conditions deduced from the SSM/I ice concentrations. The cloud base emissivity, required for the downwelling infrared flux computation, is linked to the optical thickness through the liquid water path. The results presented show a good agreement with field measurements and little sensitivity to the cloud and aerosol properties extracted from the literature rather than from the satellite data. Infrared fluxes retrievals would however require a better knowledge of the atmosphere temperature profile and cloud base altitude.

For the years 1980–88, we use satellite and meteorological data from the Chukchi and Laptev Seas to estimate the volume flux into the Chukchi Sea through the Bering Strait. We assume that the two seas have approximately the same amount of stored heat at the onset of freezing, then calculate the heat loss from both seas during the period from the onset of freezing to the formation of a solid ice cover. Because the Chukchi Sea generally takes 4–6 weeks longer to freeze over than the Laptev Sea, more heat is lost to the atmosphere from the Chukchi Sea than from the Laptev. On the assumption that the only source of this additional heat from the Chukchi is the Bering Strait inflow, and using published values for the mean temperature of the Bering Strait water, we calculate that the flow through the Bering Strait lies in the range of 0.5–1.5 Sverdrups, which is comparable to values cited in the literature. This procedure suggests a technique for determination of the interannual variability of the Bering Strait inflow.

Wilkins Ice Shelf has an area of 16000 km2 and lies off the west coast of the Antarctic Peninsula bounded by Alexander, Latady, Charcot and Rothschild islands. Several ice shelves, including Wilkins, exist close to a climatic limit of viability. The recent disintegration of the neighbouring Wordie Ice Shelf has been linked to atmopsheric warming observed on the Antarctic Peninsula. The limit of ice-shelf viability thus appears to have migrated south. Should this continue, the question arises; how long will Wilkins Ice Shelf survive?

Compared with the other ice shelves on the Antarctic Peninsula, few surface glaciological data have been collected on Wilkins Ice Shelf. We compare, contrast and combine a variety of remotely sensed data: the recently declassified GEOSAT Geodetic Mission altimetry, Landsat MSS and TM imagery, and radio-echo sounding data (RES), to study its structure and mass balance regime.

We find that this shelf has an unusual mass balance regime and relies heavily for sustenance on in situ accumulation. Its response to a continued atmospheric warming may be significantly different from that of Wordie Ice Shelf. Wordie Ice Shelf was fed by several dynamic outlet glaciers which accelerated the disintegration process when the ice shelf fractured. Wilkins Ice Shelf by contrast is almost stagnant and is expected to respond by normal calving at the ice front. Changes in the accumulation rate or basal melt-rate may, however, dominate any dynamic effect. Over the last two decades the ice front positions have remained stable.

With the use of time series of AVHRR imagery we examine a non-typical case of ice motion and lead pattern formation in the Beaufort Sea during a period of persistent southwesterly to southerly winds during 31 March-14 April 1991. Our goal is to suggest sea-ice rheologies through the use of satellite imagery. Ice motions were estimated by making a negative of the first image in a time sequence and combining it with a second image using a logical AND operation. Thus leads for the first day are white and those for the second day are black. For the period 31 March-5 April with westerly winds there was a slip line oriented southwest-northeast across the central Beaufort Sea and another slip line parallel to it 150 km to the north. Relative motion was 0.015 m s−1 across the slip lines. With southerly winds during 9-14 April the entire field of leads was advected to the north. Lead patterns are organized on a hierarchy of scales and the scales are a result of the particular magnitude of forcing by the atmosphere. Angles between diamond-shaped lead features with spacings of both 150 km and 10-20 km are approximately 30°, which as Erlingsson showed corresponds to an angle of internal friction of 15°. We provide an argument for the scale-invariant nature of the lead patterns and suggest that it may be possible to parameterize small-scale (less than tens of km) lead patterns.

The release of heat from sea-ice leads is an important component of the heat budget in the Arctic, but the impact of leads on regional scale climate is difficult to assess without information on their distribution in both space and time. Remote sensing of leads using satellite data, specifically AVHRR thermal and Landsat visible imagery, is examined with respect to one lead parameter: lead width. Atmospheric effects are illustrated through the concept of thermal contrast transmittance, where the brightness temperature contrast between leads of various ice thicknesses and the surrounding multi-year ice is simulated using a radiative transfer model. Calculations are made as a function of aerosol, ice crystal precipitation, and cirrus cloud optical depths. For example, at ice crystal optical depths of more than about 1.5 under mean January conditions in the central Arctic, the brightness temperature differences between 2 m and 5 cm thick ice are similar to the ice temperature variability so that there would be insufficient contrast to distinguish a lead from the surrounding ice. The geometrical aspects of the sensor are also simulated by degrading Landsat data so that the effect of sensor field-of-view on retrieved lead width statistics can be assessed. Large leads tend to “grow” with increased pixel size while small leads disappear. Changes in lead width and orientation distributions can readily be seen.

Passive microwave-brightness temperatures over the Greenland ice sheet are examined during the melt season in order to develop a technique for determining surface-melt occurrences. Time series of Special Sensor Microwave/ Imager (SSM/I) data are examined for three locations on the ice sheet, two of which are known to experience melt. These two sites demonstrate a rapid increase in brightness temperatures in late spring to early summer, a prolonged period of elevated brightness temperatures during the summer, and a rapid decrease in brightness temperatures during late summer. This increase in brightness temperatures is associated with surface snow melting. An objective technique is developed to extract melt occurrences from the brightness-temperature time series. Of the two sites with summer melt, the site at the lower elevation had a longer period between the initial and final melt days and had more total days classified as melt during 1988 and 1989. The technique is then applied to the entire Greenland ice sheet for the first major surface-melt event of 1989. The melt-zone signal is mapped from late May to early June to demonstrate the advance and subsequent retreat of one “melt wave”. The use of such a technique to determine melt duration and extent for multiple years may provide an indication of climate change.

The Thwaites Glacier/iceberg tongue complex has been a significant feature of the Antarctic coastline for at least 50 years. In 1986, major changes began to occur in this area. Fast ice melted and several icebergs calved from the base of the iceberg tongue and the terminus of Thwaites Glacier. The iceberg tongue rotated to an east-west orientation and drifted westward. Between 1986 and 1992, a total of 140 km of drift has occurred.

Remote digital velocity measurements were made on Thwaites Glacier using sequential Landsat images to try to determine if changes in velocity had occurred in conjunction with the changes in ice position. Measurements on Landsat images from 1972 and 1984 revealed an average annual velocity of 2.62 km a−1 during that time period, and measurements on images from 1984 and 1990 an average annual velocity of 2.84kma−1, an increase of more than 8%. Estimates made on Thwaites Glacier by others using earlier data contained enough uncertainty that it was impossible to determine a velocity profile over a longer period of time. Examination of the morphology of the glacier/iceberg tongue showed no evidence of surge activity.

For a pilot study of change detection of small tropical mountain glaciers from SPOT satellite, Mount Kenya was chosen because a map at scale 1 : 5000 in a local coordinate system, based on aerial photography in September 1987, provided a control for mapping of horizontal ice extent, and because the precipitous relief offered a particular challenge for technique development. A terrain model in compatible map projection was constructed from the topography of the 1987 map. A SPOT subscene of February 1988 was geocoded and terrain corrected to conform with the terrain model and identifiable terrain features. Another 26 control points established on the ground were entered in the digital array. The pixels were re-sampled to a nominal resolution of 10 m. This digital map was output at scale 1 : 5000. The glacier boundaries and the location of the terrain control points agreed with the September 1987 map within the mapping accuracy commensurate with the 1 : 5000 scale. This satellite-based mapping provides an internally consistent reference for determining the changes of Mount Kenya’s glaciers from February 1988 to a SPOT sampling intended for the mid-1990s. This pilot study further indicates the feasibility of glacier change detection on various other high mountains of the tropics.

It is well known that the interpretation of high resolution (<100 m) visible and near infrared (e.g. Landsat) imagery of large ice masses is hindered by the uniform reflectivity of snow, ice and cloud surfaces. Such interpretation is at present largely performed manually, but there is a good prospect that it could be automated by the incorporation of image texture. This paper describes preliminary work towards the identification of the most appropriate texture technique, or combination of techniques, and assesses the likely performance of such methods.

Different textures are identified with different types of surface cover, and the use of these differences to classify images is investigated. Specifically, we compare a traditional texture measure, the Grey Level Co-occurrence Matrix (GLCM), with a modification of a relatively new technique, fractional Brownian motion (FBM). These two methods are applied to three Landsat MSS images of the Nordaustlandet ice cap, Svalbard. The classification accuracy, computation time and memory required, advantages and limitations of the two methods are compared. The GLCM technique appears to be able to distinguish three groups of image classes, namely dry snow, wet snow, and melt features, ablation areas or cloud cover. The FBM technique is computationally more efficient, and though it performs in general less well than the GLCM technique it gives better discrimination of cloud cover.

Curvilinear flow stripes appear on aerial photographs and satellite imagery of polar ice. On Byrd Glacier, Antarctica, flow stripes are especially prominent and can be detected on AVHRR imagery down to the ice shelf margin.

Aerial photographs of Byrd Glacier (Brecher, 1986) are used to determine photogrammetrically the relief associated with flow stripes on two transverse profiles separated by a distance of 65 km. Two kinds of stripes are found, topographic and textural flow stripes. Topographic flow stripes are associated with a ridge-trough topography with double amplitudes of 7 m to 45 m and slopes of 1–7 %. The valleys of the topographic flow stripes appear bright in the photographs and correspond to snow-covered areas, while ridges appear in general as dark stripes that correspond to bare ice areas with abundant crevasses. Textural flow stripes correspond to bands of distinct crevasse pattern which are not associated with topography.

The photogrammetric information is correlated with the brightness pattern across flow stripes that appears on Advanced Very High Resolution Radiometer (AVHRR) data. Flow stripes decay rapidly on the AVHRR imagery over the first 40 km downstream from Byrd Glacier and some persist for nearly 400 km to the front of the Ross Ice Shelf.

Velocity was measured across a transect on the upstream photogrammetric profile. There are no velocity discontinuities across the boundaries of flow stripes, which shows that lateral shear between flow stripes is not a valid mechanism for explaining their persistence.