Reconstructing Daily Weather in the Early 19th Century: New Insights from Data Assimilation

Reconstructing historical weather and climate at the daily scale during the early 19th century (1806–1821) is crucial for understanding variability and predictability in this data-scarce period. This study assimilates additional historical pressure and temperature observations into the 20th Century Reanalysis Version 3 (20CRv3) to improve the reliability of daily reconstructions over Europe. 

We use state-of-the-art data assimilation methods, with the Ensemble Kalman Filter as the primary framework for integrating historical series. Alternative techniques, including the Ensemble Square Root Filter and the Iterative Ensemble Smoother, are also investigated to study their performance in capturing daily weather. 

We assess the impact of these methods using evaluation metrics, probabilistic measures, and comparisons to independent observations. Our results reduce the ensemble spread and uncertainty of 20CRv3, providing insights into internal variability at the daily scale and short-term climate dynamics.