juliensimon/orbital-fragmentation-events

Orbital Fragmentation Events

Part of the [Orbital Mechanics Datasets](https://huggingface.co/collections/juliensimon/orbital-mechanics-datasets-69c24caca4ab3934c9856994) collection on Hugging Face.

Catalog of 1,073 orbital fragmentation events derived from the NORAD Satellite Catalog (SATCAT) via CelesTrak. A fragmentation event is identified as any launch that produced 4 or more cataloged debris objects, indicating an in-orbit breakup caused by explosions, collisions, anomalous events, or deliberate destruction.

Dataset description

Every significant breakup event in Earth orbit since 1960 is captured in this dataset. When a satellite or rocket body fragments -- whether from a propulsion failure, accidental collision, deliberate destruction (e.g., anti-satellite tests), or unexplained anomaly -- it produces tracked debris objects cataloged by the 18th Space Defense Squadron. This dataset aggregates those debris back to their parent launch, identifying the primary spacecraft or rocket body involved and computing debris statistics. It is inspired by the methodology used in NASA's "History of On-Orbit Satellite Fragmentations" technical report series published by the Orbital Debris Program Office.

Schema

Quick stats

• —1,073 fragmentation events spanning 1960 to 2026
• —32,567 total cataloged debris, 11,920 still on orbit
• —Orbit distribution: LEO (947), MEO (65), GEO (47), unknown (8), HEO (6)
• —Top countries: CIS (618), US (267), PRC (99), JPN (22), FR (13)

Most prolific breakup events

Usage

from datasets import load_dataset

ds = load_dataset("juliensimon/orbital-fragmentation-events", split="train")
df = ds.to_pandas()

# Most prolific breakups
df.nlargest(10, "debris_cataloged")[["parent_name", "debris_cataloged", "debris_on_orbit"]]

# Events still polluting orbit (>90% debris remaining)
active_pollution = df[df["decay_pct"] < 10].sort_values("debris_on_orbit", ascending=False)

# Breakups by orbit type
df.groupby("orbit_type")["debris_cataloged"].sum()

# Events by decade
df["decade"] = (df["launch_year"] // 10) * 10
df.groupby("decade")["parent_norad_id"].count()

# Country breakdown
df.groupby("country_code")["debris_cataloged"].agg(["count", "sum"]).sort_values("sum", ascending=False).head(10)

Data source

Derived from the CelesTrak SATCAT, which mirrors the official US Space Command catalog maintained by the 18th Space Defense Squadron. Fragmentation events are identified by grouping cataloged debris objects by their international designator prefix (launch ID) and filtering for launches with 4+ debris pieces.

For authoritative event-by-event analysis including assessed causes, see NASA's History of On-Orbit Satellite Fragmentations report series.

Related datasets

• —reentry-events -- Atmospheric reentry catalog
• —space-track-satcat -- Full NORAD satellite catalog
• —active-satellites -- Currently operational spacecraft

Pipeline

Source code: juliensimon/space-datasets

Citation

@dataset{fragmentation_events,
  author = {Simon, Julien},
  title = {Orbital Fragmentation Events},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/juliensimon/orbital-fragmentation-events},
  note = {Derived from NORAD SATCAT via CelesTrak (Dr. T.S. Kelso)}
}

License

CelesTrak Usage Policy