Presented on November 12, 2025, by

Zilu Meng
PhD student
Department of Atmospheric and Climate Science
University of Washington

ABSTRACT:

Reconstructing past climate is crucial for understanding the dynamics of our climate system and improving climate projections. This presentation introduces the first seasonal-resolution reanalysis of the last millennium (LMR-Seasonal), which combines an ocean-atmosphere-sea-ice coupled linear inverse model with paleoclimate proxy data. This “online” data assimilation approach leverages the predictive skill of a dynamically consistent model to propagate information across seasons, with specific focus on variables such as ocean temperature, sea ice concentration and thickness. Instrumental verification reveals robust skill in reconstructing surface temperatures, even in seasons and regions where proxy coverage is sparse. We demonstrate that LMR-Seasonal outperforms other paleoclimate data assimilation products in reproducing recent instrumental records. Results reveal the reconstruction’s ability to capture the seasonal evolution of El Niño events, seasonal temperature trends consistent with orbital forcing, and the distinct climatic shifts associated with the Medieval Climate Anomaly and Little Ice Age. We will highlight how this seasonally resolved dataset contributes to our understanding of past climate variability, opening avenues for further research within oceanography, earth sciences, and atmospheric sciences.