Journal Article

Combining Stochastic Parameterized Reduced‐Order Models With Machine Learning for Data Assimilation and Uncertainty Quantification With Partial Observations

Chen, Nan, Mou, Changhong, Smith, Leslie M.

2023

Overview

A hybrid data assimilation algorithm is developed for complex dynamical systems with partial observations. The method starts with applying a spectral decomposition to the entire spatiotemporal fields, followed by creating a machine learning model that builds a nonlinear map between the coefficients of observed and unobserved state variables for each spectral mode. A cheap low‐order nonlinear stochastic parameterized extended Kalman filter (SPEKF) model is employed as the forecast model in the ensemble Kalman filter to deal with each mode associated with the observed variables. The resulting ensemble members are then fed into the machine learning model to create an ensemble of the corresponding unobserved variables. In addition to the ensemble spread, the training residual in the machine learning‐induced nonlinear map is further incorporated into the state estimation, advancing the diagnostic quantification of the posterior uncertainty. The hybrid data assimilation algorithm is applied to a precipitating quasi‐geostrophic (PQG) model, which includes the effects of water vapor, clouds, and rainfall beyond the classical two‐level QG model. The complicated nonlinearities in the PQG equations prevent traditional methods from building simple and accurate reduced‐order forecast models. In contrast, the SPEKF forecast model is skillful in recovering the intermittent observed states, and the machine learning model effectively estimates the chaotic unobserved signals. Utilizing the calibrated SPEKF and machine learning models under a moderate cloud fraction, the resulting hybrid data assimilation remains reasonably accurate when applied to other geophysical scenarios with nearly clear skies or relatively heavy rainfall, implying the robustness of the algorithm for extrapolation. Plain Language Summary Data assimilation of complex turbulent systems by only observing a subset of state variables is a notoriously challenging topic. In this paper, a hybrid data assimilation method is developed to efficiently estimate the model states. The method exploits cheap stochastic parameterized processes as surrogate models for filtering the observed state variables, significantly reducing the computational cost. It also uses machine learning to build a nonlinear map between observed and unobserved state variables for each spectral mode, which advances the computation of the ensemble members of the unobserved states. The hybrid data assimilation algorithm is applied to a precipitating quasi‐geostrophic (PQG) model, which includes the effects of water vapor, clouds, and rainfall beyond the classical two‐level QG model. Utilizing the calibrated models under a moderate cloud fraction, the resulting hybrid data assimilation remains reasonably accurate when applied to other geophysical scenarios with nearly clear skies or relatively heavy rainfall, implying the robustness of the algorithm for extrapolation. Key Points Cheap stochastic parameterized reduced‐order models are designed as efficient forecast models in the ensemble Kalman filter Machine learning is used to map each ensemble member from the observed variables to the corresponding unobserved variables The training residual in the machine learning‐induced nonlinear map is incorporated to advance the quantification of posterior uncertainty

• Publisher: John Wiley & Sons, Inc,American Geophysical Union (AGU)
• Subject: Algorithms / Clear sky / Coefficients / Data assimilation / Data collection / Decomposition / Dynamical systems / Forecasting models / Heavy rainfall / Kalman filters / Machine learning / Modelling / precipitating quasi‐geostrophic model / Precipitation / stochastic parameterized reduced‐order model / Uncertainty / Water vapor / Water vapour