NGC Catalog
CLASSIC
Welcome Guest
Models
PhysicsNeMo Checkpoints: AFNO_DX_FI-V1-ERA5

PhysicsNeMo Checkpoints: AFNO_DX_FI-V1-ERA5

For downloads and more information, please view on a desktop device.
Logo for PhysicsNeMo Checkpoints: AFNO_DX_FI-V1-ERA5
Description
This deep learning model, based on the Adaptive Fourier Neural Operator framework (AFNO), interpolates a prognostic forecast model to a shorter time-step size (by default from 6 h to 1 h).
Publisher
NVIDIA
Latest Version
v0.1.0
Modified
April 4, 2025
Size
1004.12 MB

PhysicsNeMo Checkpoints: AFNO_DX_FI-V1-ERA5 (ERA5 AFNO Forecast Interpolation Diagnostic)

Description:

This is a forecast interpolation diagnostic model checkpoint. This deep learning model, based on the Adaptive Fourier Neural Operator framework (AFNO), interpolates a prognostic forecast model to a shorter time-step size (by default from 6 h to 1 h) and operates on 0.25 degree lat-lon equirectangular grid with 73 variables.

This model interpolates the following variables surface variables: "u10m", "v10m", "u100m", "v100m", "t2m", "sp", "msl", "tcwv" and the variables u, v, t, z, q at pressure levels 1000, 925, 850, 700, 600, 500, 400, 300, 250, 200, 150, 100, 50.

For training recipes, see NVIDIA PhysicsNeMo. For inference, see NVIDIA Earth2Studio.

This model is ready for commercial use.

License/Terms of Use:

Use of this model is governed by the NVIDIA Community Model License. By pulling and using this model, you accept the terms and conditions of this license.

Reference(s):

Modulated Adaptive Fourier Neural Operators for Temporal Interpolation of Weather Forecasts

Model Architecture:

Architecture Type: This model uses an Adaptive Fourier Neural Operator (AFNO) architecture.
Network Architecture: AFNO

Input:

Input Type(s):

  • Tensor (73 atmospheric variables on a 0.25 degree grid.)
  • DateTime (NumPy Array)

Input Format(s): PyTorch Tensor / NumPy array
Input Parameters:

  • Five Dimensional (5D) (batch, time, variable, latitude, longitude)
  • DateTime (1D)

Other Properties Related to Input:

  • Input latitude/longitude grid: 0.25 degree resolution regular latitude/longitude grid (720 x 1440 pixels).
  • Input state weather variables: u10m, v10m, u100m, v100m, t2m, sp, msl tcwv, u1000, u925, u850, u700, u600, u500, u400, u300, u250, u200, u150, u100, u50, v1000, v925, v850, v700, v600, v500, v400, v300, v250, v200, v150, v100, v50, t1000, t925, t850, t700, t600, t500, t400, t300, t250, t200, t150, t100, t50, z1000, z925, z850, z700, z600, z500, z400, z300, z250, z200, z150, z100, z50, q1000, q925, q850, q700, q600, q500, q400, q300, q250, q200, q150, q100, q50

For lexicon information, review the ERA5 Lexicon at Earth2Studio. u, v, t, z and q refer to eastward and northward wind components, temperature, geopotential, and specific humidity (respectively).

Output:

Output Type(s):

  • Tensor (1-hourly interpolated forecasts of a 73-channel 6-hour forecast).
  • DateTime (NumPy Array)

Output Format: Pytorch Tensors
Output Parameters: Five Dimensional (5D) (batch, time, variable, latitude, longitude)
Other Properties Related to Output:

  • Output latitude/longitude grid: 0.25 degree resolution regular latitude/longitude grid (720 x 1440 pixels).
  • Output state weather variables: u10m, v10m, u100m, v100m, t2m, sp, msl tcwv, u1000, u925, u850, u700, u600, u500, u400, u300, u250, u200, u150, u100, u50, v1000, v925, v850, v700, v600, v500, v400, v300, v250, v200, v150, v100, v50, t1000, t925, t850, t700, t600, t500, t400, t300, t250, t200, t150, t100, t50, z1000, z925, z850, z700, z600, z500, z400, z300, z250, z200, z150, z100, z50, q1000, q925, q850, q700, q600, q500, q400, q300, q250, q200, q150, q100, q50

For lexicon information, review the ERA5 Lexicon at Earth2Studio. u, v, t, z and q refer to eastward and northward wind components, temperature, geopotential, and specific humidity (respectively).

Software Integration

Runtime Engine(s): PyTorch
Supported Hardware Microarchitecture Compatibility:

  • NVIDIA Ampere
  • NVIDIA Hopper
  • NVIDIA Turing

Supported Operating System(s):

  • Linux

Model Version(s):

Model Version: v1

Training, Testing, and Evaluation Datasets:

Training Dataset:

Link: ERA5

** Data Collection Method by dataset

  • Automatic/Sensors

** Labeling Method by dataset

  • Automatic/Sensors

Properties: ERA5 data for the period January 1980 - December 2016. ERA5 provides hourly estimates of various atmospheric, land, and oceanic climate variables. The data covers the Earth on a 30km grid and resolves the atmosphere at 137 levels.

For more information about ERA5 and a list of known and resolved issues, see the associated documentation and reference publication which provides more detailed information about the quality of the ERA5 reanalysis. The ERA5 reanalysis, and therefore any model trained using it as a data source, is an imperfect representation of the state of global weather, with correlated errors as a function of region and time. Furthermore, the limited spatial and vertical resolution of the dataset means that physical behavior near or below that resolution scale are not accurately resolvable. Review ECMWF's publications for further information about the quality of its forecasts.

Testing Dataset:

Link: ERA5

** Data Collection Method by dataset

  • Automatic/Sensors

** Labeling Method by dataset

  • Automatic/Sensors

Properties: ERA5 data for the date range of January 2017 - December 2018. ERA5 provides hourly estimates of various atmospheric, land, and oceanic climate variables. The data covers the Earth on a 30km grid and resolves the atmosphere at 137 levels.

For more information about ERA5 and a list of known and resolved issues, see the associated documentation and reference publication which provides more detailed information about the quality of the ERA5 reanalysis. The ERA5 reanalysis, and therefore any model trained using it as a data source, is an imperfect representation of the state of global weather, with correlated errors as a function of region and time. Furthermore, the limited spatial and vertical resolution of the dataset means that physical behavior near or below that resolution scale are not accurately resolvable. Review ECMWF's publications for further information about the quality of its forecasts.

Evaluation Dataset:

Link: ERA5

** Data Collection Method by dataset

  • Automatic/Sensors

** Labeling Method by dataset

  • Automatic/Sensors

Properties: ERA5 data for the date range of January 2019 - December 2020. ERA5 provides hourly estimates of various atmospheric, land, and oceanic climate variables. The data covers the Earth on a 30km grid and resolves the atmosphere at 137 levels.

For more information about ERA5 and a list of known and resolved issues, see the associated documentation and reference publication which provides more detailed information about the quality of the ERA5 reanalysis. The ERA5 reanalysis, and therefore any model trained using it as a data source, is an imperfect representation of the state of global weather, with correlated errors as a function of region and time. Furthermore, the limited spatial and vertical resolution of the dataset means that physical behavior near or below that resolution scale are not accurately resolvable. Review ECMWF's publications for further information about the quality of its forecasts.

Inference:

Engine: PyTorch
Test Hardware:

  • A100
  • H100
  • L40S

Ethical Considerations:

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

For more detailed information on ethical considerations for this model, please see the Model Card++ Explainability, Bias, Safety & Security, and Privacy Subcards [Insert Link to Model Card++ here].

Please report security vulnerabilities or NVIDIA AI Concerns here.