Improving Seasonal Ice Forecasts Using a Coupled Lake-Ice System

Session: 36. - Improving Model Predictions through Coupled System and Data Assimilation

James Kessler, University of Michigan, [email protected]
Ayumi Fujisaki, [email protected]
Jia Wang, NOAA, GLERL, [email protected]
Philip Chu, NOAA/GLERL, [email protected]

Abstract

@page { margin: 0.79in } p { margin-bottom: 0.1in; line-height: 120% }

Seasonal ice cover in the Great Lakes plays a key role in regional weather, commerce, and society at large but is difficult to predict due to it's high year-to-year variability. For example, annual maximum ice cover (AMIC) for all five lakes is frequently as low as 30% or as high as 80% (data: National Ice Center). This high variability is likely due to the many physical processes and state variables related to ice formation which span atmosphere, lake and ice domains. Therefore, a coupled modeling system is essential to improving the accuracy of seasonal ice cover predictions.

In this study, the Finite Volume Community Ocean Model (FVCOM) is two-way coupled with the Unstructured Grid–Community Ice Code (UG–CICE) to re-forecast seasonal ice cover for both 2017 and 2018.  Atmospheric variables from components of the North American Multi-Model Ensemble are used as surface boundary conditions but two-way atmosphere-lake coupling is also investigated as a future possibility.  The results are compared to satellite and in-situ observations to assess skill and identify ways to improve future projections.

1. Keyword
ice

2. Keyword
lake model

3. Keyword
atmosphere-lake interaction

4. Additional Keyword
seasonal forecast

5. Additional Keyword
ice forecast