juliensimon/goes-xray-flux

GOES Solar X-Ray Flux (1-Minute)

Credit: NASA/SDO

Part of a dataset collection on Hugging Face.

Dataset description

Solar soft X-ray flux from the GOES X-Ray Sensor (XRS), the operational backbone of solar flare monitoring. Updated daily from NOAA SWPC, growing incrementally at 1-minute cadence.

The GOES (Geostationary Operational Environmental Satellite) X-Ray Sensor measures the Sun's soft X-ray irradiance in two wavelength bands: a "short" 0.05-0.4 nm band and a "long" 0.1-0.8 nm band. The long band is the international standard for solar flare classification: the A/B/C/M/X scale is defined directly from its peak flux, with each letter marking a tenfold increase in intensity. A C1.0 flare corresponds to 1e-6 W/m^2, an M1.0 to 1e-5 W/m^2, and an X1.0 to 1e-4 W/m^2.

X-ray flux is the earliest and most direct space-weather signature of a solar flare. Because soft X-rays travel at the speed of light, they arrive ~8 minutes after the flare and immediately ionize Earth's dayside ionosphere, causing sudden ionospheric disturbances and shortwave (HF) radio blackouts. Operationally, the GOES XRS time series is what space-weather forecasters watch in real time to issue flare alerts and R-scale radio-blackout warnings. This dataset complements event-level flare catalogs by preserving the underlying continuous flux from which those events are derived.

This dataset is suitable for time-series forecasting, tabular classification tasks.

Schema

Column	Type	Description	Sample	Null %
datetime	datetime64[us, UTC]	Observation timestamp (UTC), 1-minute cadence. The GOES X-Ray Sensor (XRS) integrates over each minute.	2026-06-16 09:57:00+00:00	0.0%
satellite	Int64	GOES satellite number providing the measurement (e.g. 16, 18). NOAA designates one spacecraft as the primary X-ray source; the number can change when the primary is reassigned.	18	0.0%
flux_short	float64	Solar X-ray irradiance in the 0.05-0.4 nm ('short') band, in W/m^2. Electron-contamination corrected science-quality flux. The harder short band responds to the hottest flare plasma and rises earlier in impulsive events.	8.71441940830664e-09	0.0%
flux_long	float64	Solar X-ray irradiance in the 0.1-0.8 nm ('long') band, in W/m^2. This is the band used for the standard solar flare classification. Electron-contamination corrected science-quality flux.	5.565992751144222e-07	0.0%
flare_class	string	Solar flare magnitude class derived from the long-band peak flux: A (<1e-7 W/m^2), B (1e-7-1e-6), C (1e-6-1e-5), M (1e-5-1e-4), X (>=1e-4). Each letter is a 10x step. M- and X-class flares can drive radio blackouts and radiation storms.	B	0.0%

Quick stats

• 115,733 1-minute readings (2026-06-16 to 2026-09-04)
• Peak long-band flux: 1.31e-04 W/m^2
• 1,531 minutes at M-class or above, 8 at X-class

Usage

from datasets import load_dataset

ds = load_dataset("juliensimon/goes-xray-flux", split="train")
df = ds.to_pandas()

from datasets import load_dataset
import matplotlib.pyplot as plt

ds = load_dataset("juliensimon/goes-xray-flux", split="train")
df = ds.to_pandas().sort_values("datetime")

# Classic GOES X-ray plot: long-band flux on a log scale with flare-class bands
fig, ax = plt.subplots(figsize=(12, 4))
ax.plot(df["datetime"], df["flux_long"], linewidth=0.5)
ax.set_yscale("log")
for level, label in [(1e-6, "C"), (1e-5, "M"), (1e-4, "X")]:
    ax.axhline(level, color="red", linestyle="--", alpha=0.4)
    ax.text(df["datetime"].iloc[0], level, f" {label}", va="bottom", color="red")
ax.set_ylabel("0.1-0.8 nm flux (W/m^2)")
ax.set_title("GOES Solar X-Ray Flux")
plt.tight_layout()
plt.show()

# Largest flares in the record
print(df.nlargest(10, "flux_long")[["datetime", "flux_long", "flare_class"]])

Data source

https://www.swpc.noaa.gov/products/goes-x-ray-flux

Update schedule

Daily at 16:10 UTC

Related datasets

• juliensimon/solar-flare-events

• juliensimon/solar-radio-bursts

• juliensimon/donki-space-weather-events

• juliensimon/f107-solar-flux

• juliensimon/silso-sunspot-number

If you find this dataset useful, please consider giving it a like on Hugging Face. It helps others discover it.

About the author

Created by Julien Simon — AI Operating Partner at Fortino Capital. Part of the Space Datasets collection.

Citation

@dataset{goes_xray_flux,
  title = {GOES Solar X-Ray Flux (1-Minute)},
  author = {juliensimon},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/goes-xray-flux},
  publisher = {Hugging Face}
}

License

CC-BY-4.0