EAT v1.0.0: a 1D test bed for physical–biogeochemical data assimilation in natural waters

Abstract

Abstract. Data assimilation (DA) in marine and freshwater systems combines numerical models and observations to deliver the best possible characterization of a waterbody's physical and biogeochemical state. DA underpins the widely used 3D ocean state reanalyses and forecasts produced operationally by, e.g., the Copernicus Marine Service. The use of DA in natural waters is an active field of research, but testing new developments in realistic setting can be challenging as operational DA systems are demanding in terms of computational resources and technical skill. There is a need for test beds that are sufficiently realistic but also efficient to run and easy to operate. Here, we present the Ensemble and Assimilation Tool (EAT), a flexible and extensible software package that enables data assimilation of physical and biogeochemical variables in a one-dimensional water column. EAT builds on established open-source components for hydrodynamics (GOTM), biogeochemistry (FABM), and data assimilation (PDAF). It is easy to install and operate and is flexible through support for user-written plugins. EAT is well suited to explore and advance the state of the art in DA in natural waters thanks to its support for (1) strongly and weakly coupled data assimilation, (2) observations describing any prognostic and diagnostic element of the physical–biogeochemical model, and (3) the estimation of biogeochemical parameters. Its range of capabilities is demonstrated with three applications: ensemble-based coupled physical–biogeochemical assimilation, the use of variational methods (3D-Var) to assimilate sea surface chlorophyll, and the estimation of biogeochemical parameters.