IEEE TRANSACTIONS ON GEOSCIENCE ELECTRONICS, VOL. GE-17, NO. 3, JULY 1979 Surface-Based Scatterometer Results of Arctic Sea Ice ROBERT G. ONSTOTT, STUDENT MEMBER, IEEE, RICHARD K. MOORE, FELLOW, IEEE, AND W. F. WEEKS Abstract-Radar backscatter measurements were made of shorefast sea ice near Point Barrow, AK, in May 1977, with a surface-based FM- CW scatterometer that swept from 1-2 GHz and from 8.5-17.5 GHz. The 1-2 GHz measurements showed that thick first-year and multiyear ice cannot be distinguished at 10-70° incidence angles, but that un- deformed sea ice can be discriminated from pressure ridges and lake ice. Results also indicate that frequencies between 8-18 GHz have the ability to discriminate between thick first-year, multiyear, and lake ice. Cross polarization was found to be a better discriminator than like polarization. In addition, at these latter frequencies the differential scattering cross section a0 was found to have an approximately linearly increasing frequency response. I. INTRODUCTION T he fact that radar is especially well suited as an operational remote sensing tool for the polar oceans, because of its all-weather, day-night capability, and its ability to sense large expanses of terrain, has been recognized for years. For example, side-looking airborne radar (SLAR) imagers have proven to be useful aids in sea ice reconnaissance and mapping. Neverthe- less, development of the full potential for the application of radar-to-polar oceanography is hindered by a general lack of fundamental knowledge of the interaction of microwaves with the various types of sea ice. Clearly, to obtain the greatest utility from such sensors, careful study must be made of the physical phenomenon of radar return from sea ice. This knowledge should then be used in selecting the optimum param- eters for future designs of polar reconnaissance and mapping radars. The potential application of SLAR to sea ice research and reconnaissance was recognized in 1962, when the U.S. Army Cold Regions Research and Engineering Laboratory utilized a U.S. Air Force AN/APQ-56 K-band SLAR in sea ice experi- ments conducted during April and August [1]. This study in- dicated that various important sea ice features could be identified on the K-band imagery, and that the SLAR system provided the capability of obtaining good-resolution image maps of large expanses of ice, independent of the incident light and weather conditions. Since that time, radar has been used extensively to monitor sea ice by Rouse [2], Johnson and Farmer [3], Glushkov and Manuscript received January 15, 1979; revised May 1, 1979. This work was supported by the Office of Naval Research under Contract N00014-74-C-1 105. R. G. Onstott and R. K. Moore are with the Remote Sensing Lab- oratory, University of Kansas Center for Research, Inc., KS 66045. W. F. Weeks is with the U.S. Army Cold Regions Research and En- gineering Laboratory, Hanover, NH 03755. Komarov [4], Ketchum and Tooma [5], Parashar et al. [6], Dunbar [7], Dunbar and Weeks [8], Gray et al. [9], and Ketchum [10]. These studies have primarily been made at 2- and 3-cm wavelengths, although a few studies were made at 25-cm wavelengths. Most SLAR flights have used angles of incidence near grazing. These studies tended to agree that with some reservations, the ability to detect and interpret ice age, ice drift, ice surface topography such as pressure ridges and fractures, and to evaluate sea ice conditions, was very good. The first attempt to expand the knowledge of the parametric response of radar systems sensing sea ice began in 1967 when the National Aeronautics and Space Administration (NASA), Navy Oceanographic Office, U.S. Army Cold Regions Research and Engineering Laboratory, Arctic Institute of North America, and the University of Kansas, Lawrence, undertook a study of the ability of a 2.25-cm wavelength (13.3-GHz) scatterometer (a calibrated instrument designed to measure the differential scattering coefficient a' as a function of incidence angle [ 11 ] ) to identify different ice types [2]. The results demonstrated that a vertically polarized scatterometer operating at this fre- quency could indeed be used to discriminate ice types. For instance, multiyear returns were higher than thick first-year ice returns at all angles of incidence. In April 1970, another joint experiment was conducted by NASA, the Naval Oceanographic Office, and the University of Kansas, in which systematic radar backscatter measurements of sea ice were made at 400 MHz (HH, W, VH, and HV polarizations; H = horizontal, V = vertical; the first letter identifies transmitted polarization and the second letter identifies received polarization) and 13.3 GHz (W polariza- tion). The results showed that multiyear ice gave the strongest return at 13.3 GHz, while first-year ice and open water gave the strongest returns at 400 MHz [6]. Discrimination of ice from open water was possible with both frequencies. Four polarization 1 6.5-GHz radar imagery were also analyzed. The results were consistent with scatterometer results in that 0-18-cm, 18-90-cm, and 90-cm ice and open water were discriminable. Cross polarization appeared to be better in discriminating sea ice types, but problems with the images prevented a firm conclusion. During the winter and spring of 1975-1976, measurements were made with a multipolarized 13.3-GHz scatterometer in conjunction with the AIDJEX and Beaufort Sea offshore programs [9], [15]. Results showed systematic changes in microwave backscatter which strongly correlated with gross ice type categories. Multiyear ice showed significantly higher 0018-9413/79/0700-0078$00.75 © 1979 IEEE 78