Journal of Geophysical Research Atmospheres · 1983 · 155 citations · 35 references
Natural SciencesEarth SciencesClimatologyCryosphere StudiesRemote SensingCryospheric Remote SensingMicrowave MeasurementsRemote Sensing Of IceSatellite SensorsSea Ice Remote SensingThermal Remote SensingEmissivityMicrowave RadiometryMultiyear Sea Ice Snow CoverPassive Microwave Remote SensingPassive Microwave Sea Ice AnalysisPassive Microwave Sea Ice RetrievalsPassive Microwave SensingPassive Microwave Sensors
Quantifying emissivity variability within clusters using multispectral analysis could enable classification of ice age and provide insights into ice thickness and surface characteristics. The study uses near‑simultaneous THIR and SMMR images from Nimbus 7 to globally assess sea‑ice emissivities. Across fall, winter, and spring, sea‑ice emissivities vary only about 3 % for both first‑year and multiyear ice, but during summer snow melt the 37 GHz emissivity of multiyear ice rises ~30 % (dropping to <5 % at 6.6 GHz); cluster analysis shows substantial variability at 37 GHz and one‑third as much at 18 GHz, with nonlinear relationships between ice ages, yet distinct clusters correspond to radiometrically different ice types.
Microwave sea ice emissivities are investigated on a global scale by using near‐simultaneous images from the temperature‐humidity infrared radiometer (THIR) and the dual‐polarization scanning multichannel microwave radiometer (SMMR), both onboard the Nimbus 7 satellite. The emissivities in several study areas in the Arctic region are observed to be approximately constant during a 9‐month period within the fall, winter, and spring months for both first‐year and multiyear ice, with a standard deviation of about 3%. During the onset of summer, when the snow cover starts to melt, emissivity increases of about 30% are observed at 37 GHz for multiyear ice, with the effect decreasing to less than 5% at the 6.6‐GHz channels. Multichannel cluster analysis over very large study areas during winter shows considerable variability in emissivities of well‐defined clusters of consolidated ice at 37 GHz and about one third as much at 18 GHz. Some nonlinearities in the functional relationship of the emissivities in the consolidated ice regions between first‐year and multiyear ice are also evident, especially when data from one frequency is plotted against that of another. However, the cluster analysis technique separates the various emissivity data points into distinct clusters corresponding to radiometrically different ice types. If the variability within each cluster is quantified and taken into account by using multispectral analysis, in addition to improving the capability of classifying ice by age, information about ice thickness and surface characteristics might also be obtainable.
35
Remote Sensing for the Environment
Wolter Arnberg · AMBIO · 2003 · 311 citations
Natural Sciences, Earth Sciences, Environmental Sciences +10