57A057
Antarctic ice-edge detection using Ku-band scatterometer
data
Penelope Wagner, Cathleen Geiger, Tracy DeLiberty, David G.
Long, Stephen F. Ackley, Hongjie Xie
Corresponding author: Penelope Wagner –
penelope@udel.edu
Sea-ice edge detection is an integral component of both
climate models and forecast projections, as well as ice charting analysis for
navigational purposes. Scatterometers were originally designed to measure the
radar cross section (σ°) and dielectric properties of the ocean
surface to measure near-surface winds. However, they have been found also to
have a high sensitivity for monitoring ice edge and ice extent compared with the
National Aeronautics and Space Administration Team (NT) 30% ice edge derived
from passive data and Radarsat-1 (SAR) products. Previous studies using passive
microwave signatures and the NT algorithm produce ice-edge estimates with errors
up to 15% as compared with in situ measurements and SAR products are ineffective
in accurately detecting the ice edge based on overlapping signatures between
thin sea ice and open water. The Brigham Young University (BYU) Microwave Earth
Remote Sensing (MERS) Laboratory developed a QuikSCAT (QSCAT) dataset containing
the total archive from 19 July 1999 to 31 December 2008 at 2.225 and 4.450 km
resolutions. Their dataset implements the BYU Scatterometer Image Reconstruction
(SIR) algorithm for the purpose of resolution enhancement. The algorithm outputs
a binary sea-ice mask for ice/ocean detection. Statistical analysis has been
used to define a threshold to validate the QSCAT/Seawinds scatterometer
enhanced-resolution data for the Arctic ice edge based on comparisons with SAR
data and passive microwave, but not for the Antarctic ice edge. Though the QSCAT
dataset with the SIR algorithm improves the ability of the scatterometer
instrument to distinguish ice/ocean signatures by averaging various satellite
passes for 1 day, it has yet to be validated with more reliable and/or in situ
measurements. The National Ice Center (NIC) has generated ice-edge charts from
1972 to the present using a combination of various passive and active microwave
and visible satellite products. The accuracy of the NIC charts is validated with
previous studies that have paired samples of charts with ship-based observations
and/or visible satellite products. This study directly and statistically
compares ice extent between QSCAT ice-edge and NIC daily ice-edge charts using
the full QSCAT archive at both product resolutions for the entire Southern
Ocean. The long-term goal is to provide validation for scatterometer instruments
for systematic application in operational and logistical ice-charting
capabilities. The development of such a product for operational centers
optimizes applications for near-real-time forecasting and archiving of long-term
records for climate modeling. A manuscript is in preparation for the Annals
volume of this IGS Symposium based on this work.
Go Back