Deep learning for downward longwave radiative flux forecasts in the Arctic

Kim, Dae-Hui; Kim, Hyun Mee

doi:10.1016/j.eswa.2022.118547

Cited by 6 publications

(1 citation statement)

References 64 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead of directly solving intractable formulations like Naiver-Stokes or other prognostic equations for ocean modeling, a data-driven surrogate model is trained using the substantial amounts of historical training data available via numerical models or raw observations [12]. The use of observation assimilated models to train deep learning surrogates has been seen multiple times using both HYCOM [13] [14] and ERA5 [15] [16][17] models.…”

Section: Related Workmentioning

confidence: 99%

Forecasting Buoy Observations Using Physics-Informed Neural Networks

Schmidt,

Pokhrel,

Abdelguerfi

et al. 2024

Preprint

View full text Add to dashboard Cite

Methodologies inspired by physics-informed neural networks (PINNs) were used to forecast observations recorded by stationary ocean buoys. We combined buoy observations with numerical models to train surrogate deep learning networks that performed better than with either data alone. Numerical model outputs were collected from two sources for training and regularization: the hybrid circulation ocean model and the fifth ECMWF reanalysis experiment. A hyperparameter determines the ratio of observational and modeled data to be used in the training procedure, so we conducted a grid search to find the most performant ratio. Overall, the technique improved the general forecast performance compared with nonregularized models. Under specific circumstances, the regularization mechanism enabled the PINN models to be more accurate than the numerical models. This demonstrates the utility of combining various climate models and sensor observations to improve surrogate modeling.

show abstract

Section: Related Workmentioning

confidence: 99%

Forecasting Buoy Observations Using Physics-Informed Neural Networks

Schmidt,

Pokhrel,

Abdelguerfi

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

Adjoint‐based observation impact on meteorological forecast errors in the Arctic

Kim,

Kim

2024

Quart J Royal Meteoro Soc

View full text Add to dashboard Cite

Accurate weather forecasting using numerical weather prediction in the Arctic remains a challenge owing to the complex atmospheric dynamics and relatively sparse observations of the region. Nevertheless, the impact of observations used for data assimilation (DA) on forecast errors (observation impacts, OIs) in the Arctic has not been fully explored. This study evaluated the OIs on 24 hr forecast errors in the Arctic by the adjoint‐based forecast sensitivity OI (FSOI) method using the Polar Weather Research and Forecasting (WRF) model, three‐dimensional variational DA, and adjoint model of WRFPLUS. The time‐averaged FSOI during August 2018 was analyzed in terms of the type, variable, satellite channel, altitude, and horizontal location of the observations. The results showed that the total FSOI was greatest for aircraft observations (AIRCRAFT), followed by sonde observations, scatterometer sea‐surface winds, advanced microwave sounding unit‐A radiance data from each satellite, and surface synoptic observations. The impact of AIRCRAFT in reducing forecast errors was greater in the Arctic than in other regions of the globe. The total FSOI of the advanced microwave sounding unit‐A radiance data from all four satellites was the greatest among the total FSOI of the individual observation types. For both AIRCRAFT and sonde observations, the FSOI per observation increased at higher latitudes, indicating that forecast errors in the Arctic could be reduced when upper atmospheric in‐situ observations at higher latitudes are used for DA. The beneficial ratio of the Global Positioning System precipitable water in the Arctic was greater than that in the midlatitudes with relatively moist atmospheres that cause greater uncertainties in converting zenith wet delay. The results of this study can contribute to configuring an optimal observing system in the Arctic by presenting the relative importance of observations for reducing forecast errors.

show abstract

On the feasibility of an ensemble multi-fidelity neural network for fast data assimilation for subsurface flow in porous media

Wang,

Yan

2025

Expert Systems with Applications

View full text Add to dashboard Cite

Deep learning for downward longwave radiative flux forecasts in the Arctic

Cited by 6 publications

References 64 publications

Forecasting Buoy Observations Using Physics-Informed Neural Networks

Forecasting Buoy Observations Using Physics-Informed Neural Networks

Adjoint‐based observation impact on meteorological forecast errors in the Arctic

On the feasibility of an ensemble multi-fidelity neural network for fast data assimilation for subsurface flow in porous media

Contact Info

Product

Resources

About