Meeting Documents

Rectifying and Understanding Significant Biases in Passive-Microwave-Derived Antarctic Sea Ice Concentration (Invited)

Abstract

Algorithms that invert sea ice concentration (SIC) from brightness temperature retrievals by passive microwave (PM) satellites are critical and long-standing tools for understanding polar change. Yet in Antarctica, differing algorithms and sensors lead to discrepancies in total sea ice coverage that can reach as high as 2 million km² or up to 40% of total sea ice area. Locally, PM-derived SIC differences average 21% between two classes of commonly-used algorithms, and occur consistently in regions of mixed to compact SIC (60-100%). Given the role of SIC in driving air-sea exchange, and the widespread use of these PM-derived datasets as observational data in reanalysis products, this can lead to significant differences in the expected evolution of the atmospheric and oceanic boundary layers.

We show that this problem can be rectified through the use of the linear ice fraction (LIF), a complementary SIC product derived from the ICESat-2 altimetry. LIF is able to sense thinner sea ice that is challenging to observe in some PM bands and fracture features that are below the scale of a typical PM swath footprint. PM-algorithmic biases are highly consistent and can be readily parameterized as a function of the PM signature alone, allowing for a new, merged Antarctic SIC product that incorporates this small-scale information. We use this newly-derived SIC product as a new constraint in the ECCOV4r4 ocean-sea ice state estimate and explore the impact on ocean heat uptake and Antarctic bottom water formation compared to previously-used PM-SIC data.