Student or educator? Sign up with your school email and get 6 months of Pro free.

The DOI matches its Crossref record; 27 of the 28 references it cites were found in a registry or an index; none of the 38 sources its claims cite is checked for support yet.

Measurement of the Earth's rotation: 720 BC to AD2015

Open access
Stephenson, F. R., Morrison, L. V., Hohenkerk, C. Y. · 2016 · Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences

Abstract

New compilations of records of ancient and medieval eclipses in the period 720 BC to AD 1600, and of lunar occultations of stars in AD 1600–2015, are analysed to investigate variations in the Earth’s rate of rotation. It is found that the rate of rotation departs from uniformity, such that the change in the length of the mean solar day (lod) increases at an average rate of +1.8 ms per century. This is significantly less than the rate predicted on the basis of tidal friction, which is +2.3 ms per century. Besides this linear change in the lod, there are fluctuations about this trend on time scales of decades to centuries. A power spectral density analysis of fluctuations in the range 2–30 years follows a power law with exponent −1.3, and there is evidence of increased power at a period of 6 years. There is some indication of an oscillation in the lod with a period of roughly 1500 years. Our measurements of the Earth’s rotation for the period 720 BC to AD 2015 set firm boundaries for future work on post-glacial rebound and core–mantle coupling which are invoked to explain the departures from tidal friction.

Overview

CoChat Summary

TL;DR

New compilations of eclipse and occultation records from 720 BC to AD 2015 show Earth’s rotation is not uniform: the length of the mean solar day increases by +1.78±0.03 ms per century, less than the +2.3±0.1 ms per century expected from tidal friction alone. The residual non-tidal acceleration is +1.5±0.4×10−22 rad s−2, with decade-to-millennial fluctuations including a possible ~1500-year oscillation and a 2–30 year power-law spectrum.

Key findings

  • The long-term average rate of change in the length of the mean solar day (lod) is +1.78±0.03 ms cy−1 (abstract rounds to +1.8 ms per century), compared with +2.3±0.1 ms cy−1 predicted by tidal friction.
  • The best-fitting parabola corresponds to a rotational acceleration of −4.7±0.1×10−22 rad s−2, significantly less algebraically than the tidal-friction expectation of −6.2±0.4×10−22 rad s−2; the non-tidal component is +1.5±0.4×10−22 rad s−2.
  • A power spectral density analysis of lod fluctuations in 1800–2015 follows a power law with exponent −1.3 over periods of 2–30 years and shows increased power near a period of 6 years.
  • There is some indication of an lod oscillation with a period of roughly 1500 years; the measurements are presented as firm boundaries for future work on post-glacial rebound and core–mantle coupling.

Methods

  • Data: ancient and medieval eclipse records from 720 BC to AD 1600 (Babylon, China, Greece, Arab dominions, medieval Europe; timed and untimed), plus lunar occultations of stars from AD 1600–2015; figure 10 reports 478,843 occultation observations. TAI−UT1 data for 1962–2015 were used only as a comparative control.
  • ΔT = TT−UT was computed using the authors’ eclipse software, JPL DE430 and analytical ephemeris j=2 with lunar tidal acceleration −26.00′′ cy−2 before AD 1600 and −25.82′′ cy−2 after AD 1600; observations were weighted and outliers rejected.
  • Main analysis: parabolic fit for −720 to 2015, cubic spline fits with knots reflecting data density, loess local polynomial regression after 1600, and adaptive sine multitaper power spectral density analysis of lod after 1800.

Limitations

  • Paper states: the ~1500-year oscillation should be treated guardedly; it may represent the envelope of stochastic processes rather than a true periodic change, the periodic model fits poorly around 1500, and the apparent period is driven partly by spline knot spacing. Extrapolation beyond the dataset depends on the reality of that oscillation and on the assumption that tidal friction is essentially unchanged from −2000 to 2500; both are called conjectural.
  • Paper states: before 1600, timed observations are sparse and relatively inaccurate, especially in −100 to 800, and critical untimed eclipses are few; discrepant observations such as −600, −382, and 1605 were rejected. Text makes evident: surviving records are fragmentary (e.g., fewer than 10% of original Babylonian astronomical records are extant), and accuracy varies by culture and epoch.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1098/rspa.2016.0404

Cites

27 of 27 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1. Difference in longitude between JPL lunar ephemeris DE430 and the analytical ephemeris designated j=2 incorporating a value of −26′′ cy−2 for the Moon’s tidal acceleration. The dashed line is the average difference of about 10′′ in longitude, which produces a systematic difference of about 20s in ΔT.
Figure 1. Difference in longitude between JPL lunar ephemeris DE430 and the analytical ephemeris designated j=2 incorporating a value of −26′′ cy−2 for the Moon’s tidal acceleration. The dashed line is the average difference of about 10′′ in longitude, which produces a systematic difference of about 20s in ΔT.CC BY · PubMed Central
Figure 2. Values of ΔT for all the Babylonian-timed observations −720 to −9 listed in the electronic supplementary material, tables S1–S4. (The conversion factor for deriving the sidereal rotational displacement angle of the Greenwich meridian, Δθ, measured in radians, from ΔT, measured in seconds of mean solar time, is −7.29×10−5 ΔT.) The observation in parenthesis was not used in fitting curves to the data (see electronic supplementary material section S4, −666).
Figure 2. Values of ΔT for all the Babylonian-timed observations −720 to −9 listed in the electronic supplementary material, tables S1–S4. (The conversion factor for deriving the sidereal rotational displacement angle of the Greenwich meridian, Δθ, measured in radians, from ΔT, measured in seconds of mean solar time, is −7.29×10−5 ΔT.) The observation in parenthesis was not used in fitting curves to the data (see electronic supplementary material section S4, −666).CC BY · PubMed Central
Figure 3. Babylonian observations −720 to −65 with weights greater than 2 after −560, which is indicated by a vertical dotted line. The observations before −560 are intrinsically less accurate (see §3b(i)). The two observations in brackets around −200 were treated as outliers: an observation at −666 is intrinsically doubtful.
Figure 3. Babylonian observations −720 to −65 with weights greater than 2 after −560, which is indicated by a vertical dotted line. The observations before −560 are intrinsically less accurate (see §3b(i)). The two observations in brackets around −200 were treated as outliers: an observation at −666 is intrinsically doubtful.CC BY · PubMed Central
Figure 4. Values of ΔT for Chinese lunar eclipse timings 434–1280 listed in the electronic supplementary material, table S5. The observations are less accurate before 900 (indicated by a vertical dotted line). Observations in brackets were treated as outliers.
Figure 4. Values of ΔT for Chinese lunar eclipse timings 434–1280 listed in the electronic supplementary material, table S5. The observations are less accurate before 900 (indicated by a vertical dotted line). Observations in brackets were treated as outliers.CC BY · PubMed Central
Figure 5. Values of ΔT for Chinese solar eclipse timings 586–1277 listed in the electronic supplementary material, table S6. The observations are less accurate before 900 (indicated by a vertical dotted line). The observation in brackets was treated as an outlier.
Figure 5. Values of ΔT for Chinese solar eclipse timings 586–1277 listed in the electronic supplementary material, table S6. The observations are less accurate before 900 (indicated by a vertical dotted line). The observation in brackets was treated as an outlier.CC BY · PubMed Central
Figure 6. Values of ΔT for Greek lunar and solar eclipse timings −200 to 364, listed in the electronic supplementary material, table S7. The observation in brackets was treated as an outlier.
Figure 6. Values of ΔT for Greek lunar and solar eclipse timings −200 to 364, listed in the electronic supplementary material, table S7. The observation in brackets was treated as an outlier.CC BY · PubMed Central
Figure 7. Values of ΔT for Arab solar eclipse timings 829–1004 listed in the electronic supplementary material, table S8. The observation in brackets was treated as an outlier.
Figure 7. Values of ΔT for Arab solar eclipse timings 829–1004 listed in the electronic supplementary material, table S8. The observation in brackets was treated as an outlier.CC BY · PubMed Central
Figure 8. Values of ΔT for Arab lunar eclipse timings 854–1019 listed in the electronic supplementary material, table S9.
Figure 8. Values of ΔT for Arab lunar eclipse timings 854–1019 listed in the electronic supplementary material, table S9.CC BY · PubMed Central
Figure 9. Results for ΔT for collected timed observations −720 to 1280 and the untimed total solar eclipse of 1567. The dotted red curve is the parabola given by equation (4.1). The black curve is the spline curve described in §4b.The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction. The observations in brackets were treated as outliers, apart from a Babylonian observation in −666 which is intrinsically doubtful.
Figure 9. Results for ΔT for collected timed observations −720 to 1280 and the untimed total solar eclipse of 1567. The dotted red curve is the parabola given by equation (4.1). The black curve is the spline curve described in §4b.The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction. The observations in brackets were treated as outliers, apart from a Babylonian observation in −666 which is intrinsically doubtful.CC BY · PubMed Central
Figure 10. Results for ΔT for timed data 1623–2015: lunar occultations of stars (478 843 observations); lunar occultation of Jupiter;fourth contacts of solar eclipses (1623–1670). The untimed total solar eclipse of 1567 is also plotted. A sample of up to 10 observations in any one year are plotted to avoid saturation, which otherwise would give a false impression of the scatter of the data. The black curve is the spline curve described in §4b.
Figure 10. Results for ΔT for timed data 1623–2015: lunar occultations of stars (478 843 observations); lunar occultation of Jupiter;fourth contacts of solar eclipses (1623–1670). The untimed total solar eclipse of 1567 is also plotted. A sample of up to 10 observations in any one year are plotted to avoid saturation, which otherwise would give a false impression of the scatter of the data. The black curve is the spline curve described in §4b.CC BY · PubMed Central
Figure 11. Solution space for ΔT: total/annular eclipses −708 to 454. The dotted red curve is the parabola given by equation (4.1).The black curve is the spline curve described in §4b. The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction.
Figure 11. Solution space for ΔT: total/annular eclipses −708 to 454. The dotted red curve is the parabola given by equation (4.1).The black curve is the spline curve described in §4b. The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction.CC BY · PubMed Central
Figure 12. Solution space for ΔT: total/annular eclipses 454–1567. The dotted red curve is the parabola given by equation (4.1). The black curve is the spline curve described in §4b.The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction.
Figure 12. Solution space for ΔT: total/annular eclipses 454–1567. The dotted red curve is the parabola given by equation (4.1). The black curve is the spline curve described in §4b.The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction.CC BY · PubMed Central
Figure 13. Solution space for ΔT: −719 to 360, large partial solar eclipses; solar and lunar eclipses rose or set eclipsed; estimates of degree of obscuration of lunar and solar eclipses at rising or setting. The dotted red curve is the parabola given by equation (4.1). The black curve is the spline curve described in §4b.The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction.
Figure 13. Solution space for ΔT: −719 to 360, large partial solar eclipses; solar and lunar eclipses rose or set eclipsed; estimates of degree of obscuration of lunar and solar eclipses at rising or setting. The dotted red curve is the parabola given by equation (4.1). The black curve is the spline curve described in §4b.The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction.CC BY · PubMed Central
Figure 14. Solution space for ΔT: the dotted red curve is the parabola given by equation (4.1). The black curve is the spline curve described in §4b.The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction.
Figure 14. Solution space for ΔT: the dotted red curve is the parabola given by equation (4.1). The black curve is the spline curve described in §4b.The grey curve is the parabola (equation (1.5)), predicted on the basis of tidal friction.CC BY · PubMed Central
Figure 15. Plot of the residuals δΔT with respect to the parabola (4.1) represented as a straight line. The key to the symbols is as in figures 9–14. The black curve is the spline fit discussed in §4b.
Figure 15. Plot of the residuals δΔT with respect to the parabola (4.1) represented as a straight line. The key to the symbols is as in figures 9–14. The black curve is the spline fit discussed in §4b.CC BY · PubMed Central
Figure 16. lod 1960–2015 derived from lunar occultations: black curve from cubic splines with knots at 3 year intervals; green curve from fitting by loess with smoothing parameter q=0.08. Lod 1962–2015 from IERS data smoothed by 2 year moving average. (The conversion factor for deriving the change in the rotational velocity Δω in rad s−1 from the lod in ms is −0.843×10−12 lod.)
Figure 16. lod 1960–2015 derived from lunar occultations: black curve from cubic splines with knots at 3 year intervals; green curve from fitting by loess with smoothing parameter q=0.08. Lod 1962–2015 from IERS data smoothed by 2 year moving average. (The conversion factor for deriving the change in the rotational velocity Δω in rad s−1 from the lod in ms is −0.843×10−12 lod.)CC BY · PubMed Central
Figure 17. lod 1985–2015 with atmosphere angular momentum subtracted.
Figure 17. lod 1985–2015 with atmosphere angular momentum subtracted.CC BY · PubMed Central
Figure 18. lod −2000 to 2500. The dotted red line is the average measured rate of change in the lod, +1.78±0.03 ms cy−1, which is equivalent to an acceleration of −4.7±0.1×10−22 rad s−2. The shaded grey area shows the change expected on the basis of tidal friction, +2.3±0.1 ms cy−1, equivalent to −6.2±0.4× 10−22 rad s−2. The black curve is the slope on the spline fit shown in figures 9 and 10. The green-dashed curve is the extrapolation of the oscillation (equation (5.1)).
Figure 18. lod −2000 to 2500. The dotted red line is the average measured rate of change in the lod, +1.78±0.03 ms cy−1, which is equivalent to an acceleration of −4.7±0.1×10−22 rad s−2. The shaded grey area shows the change expected on the basis of tidal friction, +2.3±0.1 ms cy−1, equivalent to −6.2±0.4× 10−22 rad s−2. The black curve is the slope on the spline fit shown in figures 9 and 10. The green-dashed curve is the extrapolation of the oscillation (equation (5.1)).CC BY · PubMed Central
Figure 19. lod 1700–2015 derived from lunar occultations. The black curve is the slope on the spline curve displayed in figure 10. The green curve is the half-yearly values derived from the loess smoothing after 1800 as described in §5b. The red curve is the IERS data as displayed in figure 16.
Figure 19. lod 1700–2015 derived from lunar occultations. The black curve is the slope on the spline curve displayed in figure 10. The green curve is the half-yearly values derived from the loess smoothing after 1800 as described in §5b. The red curve is the IERS data as displayed in figure 16.CC BY · PubMed Central
Figure 20. Power spectrum density (PSD) of half-yearly lod values in the years 1800–2015 obtained from the slope on local polynomial regression of occultation ΔT values.
Figure 20. Power spectrum density (PSD) of half-yearly lod values in the years 1800–2015 obtained from the slope on local polynomial regression of occultation ΔT values.CC BY · PubMed Central

Verification

Claims the paper makes

0 Valid · 4 Not valid · 28 Not checked · 113 Own results and interpretations

75 more claims

Create a free account to see all 145 claims and to verify the 35 references still open.

See all claims — free account
  • It may state two separate claims as one. 50
  • The quote only partly states this claim. 3
  • Probably not a substantial claim. 21

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Citations

Cites

Research works printed but not linked: 5

  1. Parker RA, Dubberstein WH. 1956 Babylonian chronology: 626 BC–AD 75. Providence, RI: Brown University Press.
  2. Hunger H. 2001 Lunar and planetary texts in astronomical diaries and related texts from Babylonia, vol. V. Vienna, Austria: Austrian Academy of Science.
  3. Hsueh C, Ou-yang I. 1956 A Sino-Western calendar for two thousand years. Beijing, China: San-lien Shu-tien.
  4. Rome A. 1950 The calculation of an eclipse of the Sun according to Theon of Alexandria. In Proc. of the Int. Congress of Mathematicians Cambridge MA 30 August–6 September 1950 pp. 209–219. Providence RI: American Mathematical Society.
  5. Newcomb S. 1895 Tables of the Sun. Astr. Pap. Amer. Ephem. 6, part 1.

Cited by

See all 122

Metadata

Authors
Stephenson, F. R., Morrison, L. V., & Hohenkerk, C. Y.
Affiliations
  • Durham University (GB)
  • Pevensey, East Sussex, UK
  • HM Nautical Almanac Office, UK Hydrographic Office, Taunton TA1 2DN, UK
Corresponding author
Stephenson, F. R.
Venue
Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences
Published
2016-12
Volume
472
Issue
2196
Pages
20160404-20160404
ISSN
  • 1364-5021 (ISSN-L)
  • 1471-2946
OpenAlex
W2559792849
Cited by
122
References
28
Citations by year
2026: 17 · 2025: 19 · 2024: 17 · 2023: 14 · 2022: 9 · 2021: 14 · 2020: 9 · 2019: 6 · 2018: 9 · 2017: 8
FWCI
2.30
Type
journal article
Language
en
Open access
Licence
  • cc-by
  • royalsociety.org/journals/ethics-policies/data-sharing-mining/ (tdm)
Locations
  • Proceedings of the Royal Society A Mathematical Physical and Engineering Sciences (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, other-oa)
  • PubMed (publishedVersion)
Funders
National Institute of Standards and Technology
Topics
  • Physical Sciences › Physics and Astronomy › Astronomy and Astrophysics › Historical Astronomy and Related Studies
  • Physical Sciences › Physics and Astronomy › Astronomy and Astrophysics › History and Developments in Astronomy
  • Physical Sciences › Earth and Planetary Sciences › Geophysics › Geophysics and Gravity Measurements
Keywords
Earth rotation, length of day, historical eclipse records, lunar occultations, tidal friction, postglacial rebound, core-mantle interaction, long-term variability, power spectral density

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-17, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-17 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.