AMLCS-DA: A data assimilation package in Python for Atmospheric General Circulation Models

Abstract

This paper introduces AMLCS-DA, a Python package designed to perform sequential Data Assimilation (DA) on Atmospheric General Circulation Models (AT-GCM). The package provides implementations of well-known sequential ensemble-based methods. The default forecast step relies on the AT-GCM SPEEDY. Users can define various configurations for assimilation steps, including the density of observational networks, background error correlation structures, and inflation factors. The package also allows for analyzing errors in various ways, such as time evolution and statistics across pressure levels. Additionally, the package enables the testing of new methods in realistic operational scenarios and comparing their performance with well-known DA formulations.

Keywords

• Ensemble-based data assimilation
• AT-GCM
• Python