Measuring atmospheric ice mass from the Arctic Weather Satellite

Licentiate thesis, 2026

Atmospheric ice, consisting of various sizes of ice particles in clouds or precipitation, plays critical roles in weather and climate processes. Yet, large uncertainties remain in atmospheric ice estimates and the understanding of cloud formation processes, limiting our ability to predict severe weather and the effects of a changing climate.

Satellites carrying specialized cloud profiling radars provide the most accurate global measurements. However, their coverage is limited to a swath of roughly 1 km. Passive instruments provide much larger spatial coverage, yet, for these, a gap in sensitivity to ice particles exists in the sub-millimetre region (frequencies above 300 GHz) between microwave and infrared wavelengths. Simulations in the 1990s and 2000s established the potential of sub-millimetre observations, demonstrating that such measurements are sensitive to ice particles that account for a significant fraction of atmospheric ice mass.

In 2024, the Arctic Weather Satellite (AWS) was launched with novel 325 GHz channels, making it the first satellite to provide regular earth-facing measurements of the sub-millimetre band. In 2029, the EPS-Sterna programme is set to launch a constellation of six AWS-like satellites. Another space-borne instrument, the Ice Cloud Imager, is due to launch in 2026, with channels at 183, 325, 448, and 664 GHz. Together, these missions will provide a wealth of long-term sub-millimetre data with high spatial and temporal coverage.

This thesis asses the utility of these measurements from AWS and develops ice property estimates from them. Radiative transfer simulations of AWS observations match statistical distributions from real observations over ocean surfaces, confirming that current simulators already show ability in handling the new sub-millimetre channels, essential for operational utilization of these channels in weather centres and for accurate ice property retrievals.

The main contribution is the development of sub-millimetre-informed ice property estimates from AWS. The retrieval method estimates frozen water path, and associated mass-weighted mean altitude and particle size. Statistical comparisons with radar/lidar-based datasets show matching local and zonal means. This has resulted in a public data product of sub-millimetre-based ice mass property estimates, the first of its kind.