Engineering topics
Publications and source records attributed to Grenfell, T. C..
Recent advances in forward modeling of the electromagnetic scattering properties of sea ice are presented. The results have direct relevance to microwave remote sensing, and serve at the basis for inverse algorithms for reconstructing the physical properties of sea ice from scattering data.
The objective of the present work is to characterize the temporal evolution of the electromagnetic signatures of sea ice from initial formation through the development of first-year ice on the basis of the temporal variations in the physical properties of the ice.
Projections for global warming indicate that the polar regions will likely provide both a larger magnitude and earlier indiation of climate variability and change that will the more temperate regions of the planet.
Explore the source record for details and available documents.
To invesigate effects on polarimetric backscatter of sea ice grown under diurnal cycling conditions, we carried out an experiment inJanuary 1994 at the outdoor Geophysical Research Facility in the Cold Regions Research and Engineering Laboratory.
The results of multispectral passive microwave observations (6.7 to 90-GHz) are presented from the cruises of the FS Polarstern in the Weddell Sea from July to December 1986. This paper includes primarily the analysis of radiometric observations taken at ice station sites. Averaged emissivity spectra for first-year (FY) ice were relatively constant throughout the experiment and were not statistically different from FY ice signatures in the Arctic. Detailed ice characterization was carried out at each site to compare the microwave signatures of the ice with the physical properties. Absorption optical depths of FY ice were found to be sufficiently high that only the structure in the upper portions of the ice contributed significantly to interstation emissivity variations. The emissivities at 90-GHz, e(90), had the greatest variance. Both e(90) at vertical polarization and GR(sub e)(90, 18.7)(defined as (e(sub V)(90)-e(sub V)(18.7))/e(sub V)(90 + e(sub V)(18.7)) depended on the scattering optical depth which is a function of the snow grain diameter and layer thickness. The variance showed a latitude dependence and is probably due to an increase in the strength of snow metamorphism nearer the northern edge of the ice pack. The contribution of variations of near-surface brine volume to the emissivity was not significant over the range of values encountered at the station sites. Emissivity spectra are presented for a range of thin ice types. Unsupervised principal component analysis produced three significant eigenvectors and showed a separation among four different surface types: open water, thin ice, FY ice, and FY ice with a thick snow cover. A comparison with SMMR satellite data showed that average ice concentrations derived from the ship's ice watch log were consistent with the satellite concentrations. The surface based emissivities for FY ice were also compared with emissivities calculated from scanning multichannel microwave radiometer (SMMR) satellite radiances. Best agreement was found at 6.7 and 10-GHz, while at 18 and 37-GHz, SMMR emissivities were slightly lower than surface based results. For the three lower frequencies agreement was found within a confidence limit of 95% and for 37-GHz within about 90%.
Results are presented on surface-based multifrequency passive microwave observations of multiyear (MY) sea ice in the eastern Arctic Basin, the Beaufort Sea, the Canadian archipelago, and the northern Greenland Sea. The analyses of these data show that the magnitude of the spectral gradient of emissivity is directly related to the existence and the thickness of a decomposed surface ice layer with very high porosity. Spectra for melt ponds with a frozen surface layer closely resembled those of lake ice and showed a positive spectral gradient. The variance among emissivity spectra for MY ice was caused primarily by the distributions of melt ponds and by the presence of significant amounts of scattering inhomogeneities in the snow and the upper 20-30-cm layer of the ice.
Surface-based passive microwave observations of simulated pancake ice and of a first-year pressure ridge are described. The observed emissivity spectra are very similar to those observed on several field experiments in the polar regions. Certain spectral features such as a maximum emissivity at 37 GHz for pancake ice and a polarization ratio very close to zero for ridged ice are distinct from those of new, young, and first-year ice. They suggest possible signatures which may help to characterize these ice types using passive microwave techniques.
Surface-based active and passive microwave measurements were made in conjunction with ice property measurements for several distinct ice types in the Fram Strait during March and April 1987. Synthesis aperture radar imagery downlinked from an aircraft was used to select study sites. The surface-based radar scattering cross section and emissivity spectra generally support previously inferred qualitative relationships between ice types, exhibiting expected separation between young, first-year and multiyear ice. Gradient ratios, calculated for both active and passive data, appear to allow clear separation of ice types when used jointly. Surface flooding of multiyear floes, resulting from excessive loading and perhaps wave action, causes both active and passive signatures to resemble those of first-year ice. This effect could possibly cause estimates of ice type percentages in the marginal ice zone to be in error when derived from aircraft- or satellite-born sensors.
Results are presented on the microwave radiative characteristics of Antarctic sea ice measured during the 1986 Winter Weddell Sea Project with a set of portable radiometers. Radiometer measurements at 6, 10, 18, 37, and 90 GHz in vertical and horizontal polarizations were supplemented by near-simultaneous measurements of the ice physical characteristics (including ice thickness, salinity, temperature, snow cover, and density) made during two cruises, lasting 3 months each. Measurements were obtained on various types of sea ice over a large portion of the Weddell-Sea ice cover, including four transects across the entire ice pack. Data analysis shows a large variability in the multispectral microwave emissivities of different ice types, especially at 90 GHz, demonstrating a strong potential of the use of the 90-GHz channel, in combination with lower-frequency channels, for detailed characterizations of the ice cover.
Microwave radiometers observed artificial sea ice formation. During initial ice growth, interference fringe effects can occur both from changing ice thickness and snow or frost which may prove useful in interpreting layer thicknesses. A simulation of multiyear ice was attempted by allowing the ice to desalinate, but the resulting emissivity spectrum was more characteristic of lake ice. A multiyearlike spectrum was obtained when rubble was deposited on the ice surface.
Emissivities at frequencies from 5 to 94 GHz and backscatter at frequencies from 1 to 17 GHz were measured from sea ice in Fram Strait during the marginal Ice Zone Experiment in June and July of 1983 and 1984. The ice observed was primarily multiyear; the remainder, first-year ice, was often deformed. Results from this active and passive microwave study include the description of the evolution of the sea ice during early summer and midsummer; the absorption properties of summer snow; the interrelationship between ice thickness and the state and thickness of snow; and the modulation of the microwave signature, especially at the highest frequencies, by the freezing of the upper few centimeters of the ice.
Microwave brightness temperatures of new, young, and optically opaque sea ice grown in a large tank were obtained in the course of a joint microwave experiment at CRREL in Hanover, NH, during the winters of 1983-1984 and 1984-1985. Dual-polarized observations were taken at frequencies of 10, 18, 37, and 90 GHz over a range of incidence angles, and the concurrent temperature and ice thickness were obtained. Emissivity and degree of polarization were observed in detail during the early stages of ice growth and variations were found indicating that the ice became optically opaque at 10 GHz for ice thickness between 30 and 50 mm. Cluster plots of the data show six distinguishable surface types: optically opaque bare ice, thin ice (less than 15 mm), roughened ice, ridged ice, rotting wet ice, and open water.
Brightness temperatures of sea ice was measured during the Marginal Ice Zone Experiment (MIZEX) field experiments in February in the Bering Sea and in June/July in the northern Greenland Sea. In the Bering Sea thin growing sea ice types from black ice to 40 cm thick snow covered floes were investigated. Brightness temperatures increase with ice thickness up to 10 cm from values of 100 K for open water to as high as 250 K (e = 0.97) for thick ice, and a moderate dependence on snow thickness is found. In the Greenland Sea thick first and multiyear (FY, MY) ice types were studied. Brightness temperatures vary, depending on the daily melt-freeze cycle superimposed on the seasonal warming, ranging from near blackbody values for melting conditions to multiyear-like spectra when the surface layers refroze. The melt season was sufficiently advanced, however, that FY and MY ice could not be differentiated radiometrically.
Microwave radiometric temperature T(B) profiles of first-year sea ice were obtained along 70- to 100-m traverses, with sled-mounted radiometers at 10, 18.7, 33.6, and 37 GHz and an effective spot size of 30 cm. Measurements of T(B) as a function of nadir angle were obtained at selected sites along the traverses. Snow and ice properties were recorded and correlated with the T(B) measurements to infer the effect of snow cover and ice conditions on the radiometric temperature. T(B) correlated positively with the brine volume profile in the ice at several sites, suggesting that brine volume has a strong effect on T(B) under these conditions. An overall statistical comparison of snow thickness with T(B), when compared with previously published models, suggests that the effect of snow cover on the microwave transmission coefficient of the snow/ice interface may be an important contribution to the radiometric temperature at these frequencies. A model is proposed to explain the data.
It has been shown that passive microwave imagery obtained from satellite-borne sensors provides an important basis for the study of the polar regions. Because of the optical thinness of high-latitude clouds at microwave frequencies, radiometry can provide all-weather all-time observing capability. However, in order to clarify observational uncertainties and investigate the information content of passive microwave imagery, detailed ground-based observations are needed. Multifrequency data are also required to utilize the strong spectral dependence of both the dielectric properties of liquid water and volume scattering. The present investigation has the aim to provide information of the considered type for the calibration and interpretation of satellite observations of the Arctic during the summer season. Attention is given to instruments and calibration, the field program and the state of the ice cover, and the results.
The remote sensing techniques utilized in the Marginal Ice Zone Experiment (MIZEX) to study the physical characteristics and geophysical processes of the Fram Strait Region of the Greenland Sea are described. The studies, which utilized satellites, aircraft, helicopters, and ship and ground-based remote sensors, focused on the use of microwave remote sensors. Results indicate that remote sensors can provide marginal ice zone characteristics which include ice edge and ice boundary locations, ice types and concentration, ice deformation, ice kinematics, gravity waves and swell (in the water and the ice), location of internal wave fields, location of eddies and current boundaries, surface currents and sea surface winds.