Centre for Earth Observation Science



Doctor of Philosophy (Ph.D.) Thesis




On the estimation of physical roughness of a marginal sea ice zone using remote sensing




The surface roughness of both open water and sea ice cover of the marginal ice zone (MIZ) in the Arctic Ocean change as a function of space and time. The MIZ roughness controls many aspects of mass, gas, and energy fluxes across the ocean-sea ice-atmosphere (OSA) interface, all of which are currently being impacted by a changing climate. The rapid reduction of sea ice in the Arctic in the past few decades has resulted in a MIZ consisting of variable sea ice roughness that necessitates improved methods for observations using ice-based, shipborne, airborne, and spaceborne platforms. This thesis is an attempt to provide insight into improved techniques for the detection and classification of various MIZ roughnesses in the southern Beaufort Sea using state-of-the-art in situ and satellite-based microwave remote sensing methods. The analysis of variance (ANOVA) of the polarimetric backscattering coefficients of sea ice yields statistically significant, separable ice classes. The ANOVA of the root-mean-square height and brightness temperatures (37 and 89 GHz) of sea ice do not yield any ice classes. Polarimetric coherences and ratios at C-band (5.5 GHz) have shown potential in discriminating sea ice roughness. A proposed two-dimensional (2D) backscattering model of surface roughness (by incorporating deviation in the orientation (i.e. the ice slopes) in azimuth and range direction) further shows the dependence of circular coherence, a discriminator of roughness, on both the surface roughness and sea ice dielectric properties. The thesis provides a new 2D formula for the relationship between sea ice slopes in azimuth and range direction.

Microwave brightness temperature of open water is significantly correlated with wave height but not with the wind speed, having the strongest correlations for the horizontal polarization channel at both 37 and 89 GHz. Analysis of AMSR-E brightness temperature at 89 GHz and root-mean-square height (spring to melt onset) shows a significant correlation between the two, for spatial scale of 1-4 km. This thesis provides a modified formula for the relationship between non-dimensional form of energy and wave age at wind speeds 0-10 m/s. The brightness temperature (April-June) of sea ice at horizontal polarization of 89 GHz is found to decrease with increasing physical roughness, and is attributed to the dominant contributions from rapidly varying thermodynamic properties of snow-covered sea ice during the melt season. I also found that the changes in sea ice surface dielectric properties occur much faster temporally than those detected by satellite sensor making it difficult to interpret ice signatures at sub-pixel level (< 5.4 km). A combined analysis of physical roughness, active polarimetry and passive microwave emission of MIZ at compatible spatial and temporal scales has led to improved understanding of the behavior of the Arctic MIZ.




Dr. D. G. Barber

Professor, Department of Environment and Geography, University of Manitoba, Winnipeg, Canada


Copyright 2014 ©Mukesh Gupta