Datasets:

nasa-ibm-ai4science
/

Surya-bench-solarwind

license: mit
size_categories:
  - 100K<n<1M
tags:
  - solar-wind
  - omni
  - wind-speed
  - space-weather
pretty_name: Solar Wind Forecasting Dataset

Solar Wind Forecasting Dataset

Dataset Summary

This dataset provides hourly solar wind plasma and interplanetary magnetic field (IMF) parameters at L1, derived from NASA’s OMNI dataset. The primary forecasting target is the solar wind speed (V), while additional parameters are included for completeness:

Solar wind speed (V)
IMF Bx (GSE)
IMF By (GSM)
IMF Bz (GSM)
Proton number density (N)

The dataset is structured for machine learning experiments, particularly in this collection for evaluating foundation model efficiency for space weather forecasting tasks. Temporal coverage spans from May 13, 2010 to December 31, 2024, with splits designed for consistent training, validation, and testing across years.

Supported Tasks and Applications

time-series forecasting: Predicting future solar wind speed from past observations.
regression: Using solar wind parameters for supervised forecasting tasks.

Dataset Structure

Data Files

train.csv: Instances from Feb 15 to Dec 31 in each year between 2010–2019
validation.csv: Instances from Jan 15–31 of each year between 2010–2019
test.csv: All instances from each year between 2020–2024
leaky_validation.csv: Instances from Jan 1–14 and Feb 1-14 of each year between 2010–2019

Features

Each row contains hourly measurements with the following fields:

V: Solar wind speed (km/s) [primary target]
Bx_gse, By_gsm, Bz_gsm: IMF components (nT)
N: Proton number density (cm⁻³)

Dataset Details

Field	Description
Temporal Coverage	May 13, 2010 – Dec 31, 2024
File Format	CSV (.csv)
Data Shape	(1, 5) per timestep
Number of Time Steps	119,225
Cadence	1 hour
Forecast Target	Solar wind speed (`V`)

Authors

Vishal Upendran, [email protected]