Bill Corner

BSc (Aberdeen, 1992), MPhil (Cambridge, 1994)

The relationship between remotely sensed and ground truth data for terrestrial snow and ice surfaces.

Abstract

The purpose of this study is to determine the relationship between remotely sensed data and ground truth data for snow and ice surfaces. By attempting to correlate the relationship between these two data sets for a test site, it is hoped that the degree of ambiguity present when interpreting an image from a site where no ground truth is available will be decreased.

The test site selected for this study is a small glacier in Okstindan, Northern Norway, and is a field site used in the European Multi-spectral Airborne Campaign 1994/945 (EMAC 94/95) The remotely sensed data will be obtained from an airborne, fully polarimetric C-band SAR – the EMISAR. Three passes have been flown over the glacier, during spring, the early summer and late summer of 1995. Ground data was collected concurrently with each overflight enabling a direct temporal correlation between the two data set to be made.

Various factors that will affect the radar return signal have been anticipated, and are outlined below:

Topographic Distortion Effects
Radiometric Distortion Effects
Platform Instabilities
Radar Waves/Surface Interaction
SAR Band Width and Polarity
Structure of the Snow and Ice Facies.

As yet the relative importance of each of these factors is unknown, and therefore any decision as to which will be investigated has yet to be reached. However, it is certain that the two distortions present in a SAR image – geometric and radiometric – and certain important snow pack parameters – liquid water content, grain size and surface roughness – will have to be accounted for in this study. Figure 1.1 illustrates, in a flow chart format, the methodology to be used in the study, with data inputs, processes, analytical techniques to be implemented and probable result shown.

To this end the contribution of glacier/snow field topography and structure to the backscatter coefficient in a SAR image can be assessed, potentially allowing a more accurate interpretation of the radar image. The results of this study can then be used to interpret radar images of glacial areas that have no ground truth data available, potentially increasing the accuracy of their interpretation.