Abstract:Remote sensing data play a critical role in improving numerical weather prediction (NWP). However, the physical principles of radiation dictate that data voids frequently exist in physical space (e.g., subcloud area for satellite infrared radiance or no-precipitation region for radar reflectivity). Such data gaps impair the accuracy of initial conditions derived from data assimilation (DA), which has a negative impact on NWP. We use the barotropic vorticity equation to demonstrate the potential of deep learning augmented data assimilation (DDA), which involves reconstructing spatially complete pseudo-observation fields from incomplete observations and using them for DA. By training a convolutional autoencoder (CAE) with a long simulation at a coarse "forecast" resolution (T63), we obtained a deep learning approximation of the "reconstruction operator," which maps spatially incomplete observations to a model state with full spatial coverage and resolution. The CAE was applied to an incomplete streamfunction observation (similar to 30% missing) from a high-resolution benchmark simulation and demonstrated satisfactory reconstruction performance, even when only very sparse (1/16 of T63 grid density) observations were used as input. When only spatially incomplete observations are used, the analysis fields obtained from ensemble square root filter (EnSRF) assimilation exhibit significant error. However, in DDA, when EnSRF takes in the combined data from the incomplete observations and CAE reconstruction, analysis error reduces significantly. Such gains are more pronounced with sparse observation and small ensemble size because the DDA analysis is much less sensitive to observation density and ensemble size than the conventional DA analysis, which is based solely on incomplete observations. Significance StatementData assimilation plays a critical role in improving the skills of modern numerical weather prediction by establishing accurate initial conditions. However, unobservable regions are common in observation data, particularly those derived from remote sensing. The nonlinear relationship between data from observable regions and the physical state of unobservable regions may impede DA efficiency. As a result, we propose that deep learning be used to improve data assimilation in such cases by reconstructing a spatially complete first guess of the physical state with deep learning and then applying data assimilation to the reconstructed field. Such deep learning augmentation is found effective in improving the accuracy of data assimilation, especially for sparse observation and small ensemble size.

Reconstructing the Tropical Pacific Upper Ocean using Online Data Assimilation with a Deep Learning model

Data-driven Global Ocean Modeling for Seasonal to Decadal Prediction

ENSO‐Related Precursor Pathways of Interannual Thermal Anomalies Identified Using a Transformer‐Based Deep Learning Model in the Tropical Pacific

Enhancing interpretability and generalizability of deep learning-based emulator in three-dimensional lake hydrodynamics using Koopman operator and transfer learning: Demonstrated on the example of lake Zurich

Predicting temporal and spatial 4-D ocean temperature using satellite data based on a novel deep learning model

Deep Learning–Based Prediction of Kuroshio Path South of Japan

A Novel Deep Learning Approach for Data Assimilation of Complex Hydrological Systems

Enhanced forecasting of chlorophyll-a concentration in coastal waters through integration of Fourier analysis and Transformer networks

Prediction of long lead monthly three-dimensional ocean temperature using time series gridded Argo data and a deep learning method

Deep Learning Augmented Data Assimilation: Reconstructing Missing Information with Convolutional Autoencoders

Deep Learning Improves Reconstruction of Ocean Vertical Velocity

Deep Learning‐Based Precipitation Simulation for Tropical Cyclones, Mesoscale Convective Systems, and Atmospheric Rivers in East Asia

CAS-Canglong: A skillful 3D Transformer model for sub-seasonal to seasonal global sea surface temperature prediction

A comparison of data-driven approaches to build low-dimensional ocean models

Exploring the use of machine learning to parameterize vertical mixing in the ocean surface boundary layer

Reconstructing Historical Climate Fields With Deep Learning

Reconstructing 3-D Thermohaline Structures for Mesoscale Eddies Using Satellite Observations and Deep Learning

A Hybrid Deep-Learning Model for El Niño Southern Oscillation in the Low-Data Regime

A New DRP-4DVar-based Coupled Data Assimilation System for Decadal Predictions Using a Fast Online Localization Technique

Assessments of Data-Driven Deep Learning Models on One-Month Predictions of Pan-Arctic Sea Ice Thickness