Nonlinear Ensemble Filtering with Diffusion Models: Application to the Surface Quasigeostrophic Dynamics

The intersection between classical data assimilation methods and novel machine learning techniques has attracted significant interest in recent years. Here, we explore another promising solution in which diffusion models are used to formulate a robust nonlinear ensemble filter for sequential data assimilation. Unlike standard machine learning methods, the proposed ensemble score filter (EnSF) is completely training free and can efficiently generate a set of analysis ensemble members. In this study, we apply the EnSF to a surface quasigeostrophic model and compare its performance against the popular local ensemble transform Kalman filter (LETKF), which makes Gaussian assumptions in the analysis step. Numerical tests demonstrate that EnSF maintains stable performance in the absence of localization and for a variety of experimental settings. We find that while LETKF maintains optimal performance in the case of linear observations of the entire state and a perfect model, EnSF shows improvements over LETKF when nonlinear observations are assimilated and the system is subject to unexpected model errors. A spectral decomposition of the analysis results in this nonlinear observation regime shows that the largest improvements over LETKF occur at large scales (small wavenumbers), where LETKF lacks sufficient ensemble spread. Overall, this initial application of EnSF to a geophysical model of intermediate complexity motivates further development of the algorithm for more realistic problems. Significance Statement It is hard to overstate the crucial role that the field of artificial intelligence (AI) has played in the weather and climate enterprise. Considerable progress in the last few years has come from a particular area known as generative AI (GenAI). In this work, we borrow ideas from a popular GenAI technique known as diffusion models to explore a new data assimilation (DA) method which can extract useful information from complex (nonlinear) observation types like satellite data. We implement an efficient (training free) version of the resulting algorithm [ensemble score filter (EnSF)] in a benchmark model of geophysical turbulence. In the case of assimilating nonlinear observations and dealing with unknown model errors, EnSF demonstrates improvements against a popular DA method used in operations [local ensemble transform Kalman filter (LETKF)]. These improvements in EnSF occur without any additional fine-tuning efforts or the application of standard regularization techniques such as localization.