Seawinds-Derived Antarctic Sea-Ice Extent ----------------------------------------- OVERVIEW -------- This data set consists of files of ice masked SIRF images of the Antarctic as well as the lat,lon extent of Antarctic sea ice. Each file represents an "average" ice extent over a 1 day period. The ASCII files contain latitude/longitude pairs which represent the contour points of the estimated sea ice edge (including edited changes to the edge when they exist). Each line entry in the file consists of two values: a longitude and a lattitude. The longitude values range from -180 to +180. Multiple contours are separated by a '0 0' entry. IMAGE RECONSTRUCTION METHODOLOGY -------------------------------- The ice extent maps were made from enhanced resolution backscatter images created by the Microwave Earth Remote Sensing (MERS) Laboratory at Brigham Young University (BYU) using the Scatterometer Image Reconstruction with Filtering (SIRF) resolution enhancement algorithm. The multivariate SIRF algorithm is a non-linear resolution enhancement algorithm based on modified algebraic reconstruction and maximum entropy techniques. The algorithm is described in detail in: Long, D.G., P.J. Hardin, and P.T. Whiting, "Resolution Enhancement of Spaceborne Scatterometer Data," IEEE Trans. Geoscience Remote Sens., Vol. 31, No. 3, pp. 700-715, May 1993. The SIRF algorithm has been successfully applied to SASS and NSCAT measurements to study tropical vegetation and glacial ice. Variants of SIRF has been successfully applied to ERS-1 scatterometer and various radiometers (SSM/I and SMMR). For scatterometers, the multivariate form of the SIRF algorithm models the dependence of sigma-0 on incidence angle as sigma-0 (in dB) = A + B * (Inc Ang - 40 deg) over the incidence angle range of 15 to 60 deg. The output of the SIRF algorithm is images of the A and B coefficients. A represents the "incidence angle normalized sigma-0" (effectively the sigma-0 value at 40 deg incidence angle). The units of A are dB. The B coefficient describes the incidence angle dependence of sigma-0 and has the units of dB/deg. SEA ICE MAPPING METHODOLOGY --------------------------- The polarization ratio (AV-AH in dB), AH, and the v- and h-pol sigma-0 estimate error standard deviations (VV and VH) were used to perform the sea ice/ocean discrimination. The ice/ocean modes of the bimodal 4-dimensional (AV-AH,AH,VV,VH) distribution were separated using linear and quadratic techniques. The linear segmentation was used to obtain an estimate of the necessary cluster centroids and covariance matrices required to implement a Maximum Likelihood (ML) discrimination. Two ML classification iterations results in an estimate of the spatial distribution of sea ice and open ocean. Residual misclassification noise was reduced using binary image processing techniques such as region growing, erosion, and dilation resulting which constrain the ice edge from exhibiting unprobably ice growth or retreat in localized regions. The final product is a binary image depicting the extent of the sea ice pack. It was found that when a similar algorithm was used with NSCAT data, the resulting edge closely matched the NSIDC SSM/I-derived 30% ice concentration edge. The lat/lon values in the ASCII edge files were obtained by computing the lat/lon location of each pixel along the edge of the binary ice mask. For a more detailed description of the NSCAT ice extent algorithm see: Remund, Q.P., Long, D.G., "Automated Antarctic Ice Edge Detection Using NSCAT Data," Proc. IGARSS'97, pp. 1841-1843, Singapore, 4-8 August, 1997. Remund, Q.P., Long, D.G., "Sea-Ice Extent Mapping Using Ku-Band Scatterometer Data," Journal of Geophysical Research, vol. 104, no. C5, pp. 11515-11527, May 1999. DATA USE -------- (c) Data is copyright 1999 by David G. Long at Brigham Young University. Any use of these images should include a citation to appropriate paper(s) below and the following acknowledgement: "Ice extent images (or maps) are courtesy of David G. Long at Brigham Young University generated from data obtained from the PO.DAAC." Remund, Q.P., Long, D.G., "Sea-Ice Extent Mapping Using Ku-Band Scatterometer Data," Journal of Geophysical Research, vol. 104, no. C5, pp. 11515-11527, May 1999. Long, D.G., P.J. Hardin, and P.T. Whiting, "Resolution Enhancement of Spaceborne Scatterometer Data," IEEE Trans. Geoscience Remote Sens., Vol. 31, No. 3, pp. 700-715, May 1993 Please send copies of all papers, reports, or presentations using this data to the JPL PO.DAAC and David Long. Use of this data is contingent on meeting these requirements. ============================================================================== Dr. David G. Long long@ee.byu.edu Associate Professor voice: (801) 378-4383 Electrical and Computer Engineering Department fax: (801) 378-6586 459 Clyde Building Brigham Young University Provo, Utah 84602 BYU Electrical and Computer Engineering home page: http://www.ee.byu.edu/ BYU Microwave Earth Remote Sensing (MERS) Laboratory home page: http://www.ee.byu.edu/ee/mers/ or http://mers.byu.edu/ ============================================================================== Last revised: 22 Oct. 1999