juliensimon/dst-index

Dst Geomagnetic Storm Index (Hourly)

Credit: NASA

Part of a dataset collection on Hugging Face.

Dataset description

Hourly Disturbance Storm Time (Dst) index from WDC Kyoto -- the standard measure of geomagnetic storm intensity since 1957.

The Dst index measures the strength of the ring current -- a toroidal electric current flowing in the magnetosphere. During geomagnetic storms, the ring current intensifies and Dst drops sharply (e.g. -100 to -500 nT for major storms). This index is the primary metric used by satellite operators and power grid managers to assess storm severity.

Dst complements the Kp/Ap indices by providing hourly resolution vs. 3-hourly/daily. The ring current is a toroidal band of 10-200 keV ions (primarily H+ and O+) trapped in the inner magnetosphere at geocentric distances of 3-8 Earth radii. During quiet times, the ring current produces a modest depression of the surface magnetic field (Dst around -20 to +10 nT). When a CME or high-speed stream arrives with sustained southward interplanetary magnetic field (Bz < 0), enhanced convection electric fields inject fresh particles from the plasma sheet into the ring current, causing Dst to plunge rapidly during the storm main phase.

Dst is derived from the horizontal field component (H) at four low-latitude magnetometer stations: Hermanus (South Africa), Kakioka (Japan), Honolulu (Hawaii), and San Juan (Puerto Rico). The hourly cadence resolves the storm main phase (typically 6-12 hours of rapid decrease) and the recovery phase (1-7 days of gradual return to baseline). The Dst index has direct applications in satellite operations: empirical models relate Dst excursions to increased satellite surface charging, single-event upset rates in electronics, and thermospheric density enhancements that accelerate orbital decay.

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

Schema

Column	Type	Description	Sample	Null %
datetime	datetime64[us]	Hour-averaged timestamp (UTC); Dst has been reported at 1-hour cadence since 1957	1957-01-01 00:00:00	0.0%
dst_nt	float64	Disturbance Storm Time index in nT -- measures ring current injection at mid-latitudes; negative values indicate ring current enhancement; quiet: -20 to 0 nT, minor storm: -30 to -50 nT, moderate: -50 to -100 nT, intense: < -100 nT, extreme: < -250 nT	11.0	0.0%
daily_mean_nt	float64	24-hour mean Dst (nT) for the calendar day containing this record; smooths over short-duration spikes to represent overall storm level	-30.0	67.5%
quality	str	Data quality flag: 'final' (definitive WDC-processed values), 'provisional' (recent months, pending full processing), or 'realtime' (near-real-time, subject to revision)	final	0.0%
is_storm	bool	True if Dst <= -50 nT, the conventional threshold for a geomagnetic storm	False	0.0%
storm_intensity	category	Derived storm severity: 'quiet' (> -50 nT), 'weak' (-50 to -100 nT), 'moderate' (-100 to -250 nT), 'intense' (-250 to -500 nT), 'super' (< -500 nT, extremely rare; Carrington 1859 estimated at -850 nT)	quiet	0.0%

Quick stats

• 589,960 hourly readings (1957-01-01 to 2026-05-13)
• 25,961 storm hours (Dst <= -50 nT)
• Deepest storm: -406.0 nT on 2024-05-11 02:00
• Data quality: 547,368 final, 39,408 provisional, 3,184 realtime

Usage

from datasets import load_dataset

ds = load_dataset("juliensimon/dst-index", split="train")
df = ds.to_pandas()

from datasets import load_dataset

ds = load_dataset("juliensimon/dst-index", split="train")
df = ds.to_pandas()

# Major storms (Dst < -100)
major = df[df["dst_nt"] < -100].sort_values("dst_nt")

# Storm frequency by year
import matplotlib.pyplot as plt

df["year"] = df["datetime"].dt.year
storms_per_year = df[df["is_storm"]].groupby("year").size()
storms_per_year.plot.bar(figsize=(12, 4))
plt.title("Geomagnetic Storm Hours per Year")
plt.ylabel("Hours with Dst <= -50 nT")
plt.tight_layout()
plt.show()

# Dst time series around a specific storm
storm = df[(df["datetime"] >= "2024-05-10") & (df["datetime"] <= "2024-05-15")]
storm.plot(x="datetime", y="dst_nt", title="May 2024 Storm")
plt.show()

Data source

https://wdc.kugi.kyoto-u.ac.jp/dstdir/

Update schedule

Daily at 13:00 UTC

Related datasets

• juliensimon/space-weather-indices

• juliensimon/solar-flare-events

• juliensimon/geomagnetic-kp-index

• juliensimon/ae-index

• juliensimon/celestrak-sw

• juliensimon/donki

• juliensimon/f107

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{dst_index,
  title = {Dst Geomagnetic Storm Index (Hourly)},
  author = {juliensimon},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/dst-index},
  publisher = {Hugging Face}
}

License

CC-BY-4.0