juliensimon/solar-eclipse-catalog

Five Millennium Catalog of Solar Eclipses

Part of the Astronomy Datasets and Solar System Datasets collections on Hugging Face.

Static dataset -- uploaded once.

Complete catalog of 11,898 solar eclipses spanning five millennia (-1999 to 3000), computed by Fred Espenak as part of NASA's Five Millennium Canon of Solar Eclipses.

Dataset description

A solar eclipse occurs when the Moon passes between Earth and the Sun, casting its shadow on Earth's surface. The geometry of each eclipse depends on the Moon's orbital elements at the moment of conjunction, producing four distinct types: total (Moon fully covers the Sun), annular (Moon appears smaller than the Sun, leaving a bright ring), hybrid (transitions between total and annular along the eclipse path), and partial (Moon only partially obscures the Sun).

This catalog is derived from Fred Espenak's Five Millennium Canon of Solar Eclipses, a monumental computational effort that uses Besselian elements and the polynomial expressions of Chapront, Chapront-Touze, and Francou for lunar and solar coordinates to predict every solar eclipse from -1999 to +3000. The calculations account for the secular acceleration of the Moon, the variable rotation of the Earth (via Delta-T extrapolations), and the irregular lunar limb profile.

The gamma parameter measures how close the Moon's shadow axis passes to Earth's center — values near zero produce central eclipses at low latitudes, while values exceeding roughly 0.9972 produce partial eclipses. Eclipse magnitude gives the fraction of the Sun's diameter covered at greatest eclipse. The Saros number groups eclipses into families that repeat every 18 years 11 days 8 hours — each Saros series produces 70-80 eclipses over roughly 1,300 years, cycling through partial, total/annular, and back to partial phases.

The dataset has 3,049 total eclipses, 3,755 annular eclipses, 502 hybrid eclipses, and 3,875 partial eclipses.

Schema

Column	Type	Description
catalog_number	int64	Sequential catalog number (1 to 11,898)
date	string	Date of greatest eclipse (YYYY-MM-DD, negative years for BCE)
year	int64	Calendar year (negative for BCE)
td_of_greatest_eclipse	string	Time of greatest eclipse (Terrestrial Dynamical Time)
delta_t	float64	Delta-T: difference TDT minus UT in seconds
luna_number	int64	Lunation number (Brown's series)
saros_number	int64	Saros series number
eclipse_type	string	Eclipse type code: T (total), A (annular), H (hybrid), P (partial)
eclipse_type_name	string	Full name of eclipse type
is_total	bool	True if eclipse is total
is_annular	bool	True if eclipse is annular
gamma	float64	Distance of Moon shadow axis from Earth center
magnitude	float64	Eclipse magnitude (fraction of Sun diameter covered)
latitude	float64	Latitude of greatest eclipse (degrees, + N / - S)
longitude	float64	Longitude of greatest eclipse (degrees, + E / - W)
sun_altitude	float64	Sun altitude at greatest eclipse (degrees)
path_width_km	float64	Width of central eclipse path (km, null for partial)
central_duration	string	Duration of central eclipse (e.g. "04m57s", null for partial)
century	int64	Derived century (year / 100, truncated)

Quick stats

• 11,898 solar eclipses (-1999 to 3000)
• 3,049 total, 3,755 annular, 502 hybrid, 3,875 partial
• Average: ~4.7 eclipses per year

Usage

from datasets import load_dataset

ds = load_dataset("juliensimon/solar-eclipse-catalog", split="train")
df = ds.to_pandas()

# Filter by eclipse type
total = df[df["is_total"]]
print(f"{len(total):,} total solar eclipses across 5 millennia")

# Eclipses per century
by_century = df.groupby("century")["eclipse_type"].value_counts().unstack(fill_value=0)
print(by_century.tail(10))

# Total eclipses visible from a region (e.g. Europe: lat 35-70, lon -10 to 40)
europe_total = df[
    (df["is_total"]) &
    (df["latitude"].between(35, 70)) &
    (df["longitude"].between(-10, 40)) &
    (df["year"].between(2000, 2100))
]
print(f"Total eclipses over Europe (2000-2100): {len(europe_total)}")

# Saros series analysis
saros = df.groupby("saros_number").agg(
    count=("catalog_number", "size"),
    first_year=("year", "min"),
    last_year=("year", "max"),
).sort_values("count", ascending=False)
print(saros.head(10))

Data source

Five Millennium Canon of Solar Eclipses: -1999 to +3000 by Fred Espenak (NASA/GSFC). Besselian elements CSV export.

Update schedule

Static dataset -- uploaded once. The underlying catalog covers -1999 to +3000 and is not expected to change.

Related datasets

• silso-sunspot-number -- Daily sunspot numbers from SILSO
• iers-earth-orientation -- Earth orientation parameters (UT1-UTC, polar motion)
• lunar-craters-robbins -- Lunar crater database

Pipeline

Source code: juliensimon/space-datasets

Support

If you find this dataset useful, please give it a ❤️ on the dataset page and share feedback in the Community tab! Also consider giving a ⭐️ to the space-datasets repo.

Citation

@dataset{solar_eclipse_catalog,
  author = {Simon, Julien},
  title = {Five Millennium Catalog of Solar Eclipses},
  year = {2026},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/juliensimon/solar-eclipse-catalog},
  note = {Based on Fred Espenak's Five Millennium Canon of Solar Eclipses (NASA/GSFC)}
}

License

CC-BY-4.0