Precision Standards, Data Sources, and Validation Methodology
This document describes the accuracy characteristics of calculations performed by Nyx astronomical libraries. Use it to choose the right precision tier and validation path for visualization, planning, telescope, and research workflows.
Table of Contents#
- Overview
- Precision Standards
- Data Sources
- Validation Methodology
- Known Limitations
- Accuracy by Component
- Recommended Use Cases
- References
Overview#
Nyx combines self-contained astronomical algorithms with production-grade catalog and ephemeris integrations. Real-time visualization paths favor bounded latency and deterministic offline behavior, while high-precision planning and pipeline paths use VSOP87/ELP-backed solar-system calculations, IERS Earth orientation data, and external catalog validation where available.
Target Accuracy Level:
- Planetary positions: ±0.1 arcsecond to ±1 arcminute (precision-dependent)
- Lunar positions: ±1 arcsecond to ±0.3 degrees (precision-dependent)
- Stellar positions: ±1 milliarcsecond (limited by catalog precision)
- Time conversions: ±0.1 to ±0.5 seconds (epoch-dependent)
Valid Date Range:
- High accuracy: 1800 CE to 2200 CE
- Acceptable accuracy: 500 CE to 2500 CE
- Extrapolated: Outside these ranges, accuracy degrades progressively
Precision Standards#
Ephemeris Precision Levels#
Nyx supports configurable precision levels for ephemeris calculations:
| Level | Accuracy | Speed | Use Case |
|---|---|---|---|
LOW |
~1 arcminute | Fastest | Real-time visualization, gaming |
MEDIUM |
~10 arcseconds | Fast | Observation planning, education |
HIGH |
~0.1 arcsecond | Slower | Telescope pointing, research |
Coordinate Precision#
| Transformation | Precision | Notes |
|---|---|---|
| Equatorial ↔ Horizontal | ±0.01 arcsec | Full spherical trigonometry |
| Equatorial ↔ Galactic | ±0.0001 arcsec | IAU 1958 definition, J2000 pole |
| Equatorial ↔ Ecliptic | ±0.01 arcsec | IAU 2006 obliquity |
| Precession (J2000→date) | ±0.001 arcsec | IAU 2006 precession model |
| Proper motion | ±0.1 mas/year | Linear propagation |
Time Precision#
| Time Scale | Precision | Notes |
|---|---|---|
| UTC ↔ TAI | Exact (leap seconds) | Complete table 1972-2017 |
| TAI ↔ TT | Exact (32.184s) | By definition |
| TT ↔ TDB | ±30 microseconds | Simplified periodic terms |
| UTC ↔ UT1 | ±0.9 seconds | Without IERS bulletins |
| Delta-T | ±0.1 to ±4 seconds | Epoch-dependent (see below) |
Delta-T Accuracy by Era#
Delta-T (ΔT = TT - UT1) uncertainty varies significantly by time period:
| Period | Method | Uncertainty |
|---|---|---|
| Before 500 BCE | Extrapolation | ±100+ seconds |
| 500 BCE - 500 CE | Polynomial fit | ±4 seconds |
| 500 - 1600 CE | Historical records | ±4 seconds |
| 1600 - 1800 CE | Telescopic records | ±2 seconds |
| 1800 - 2005 CE | Historical fits | ±0.1 seconds |
| 2005 - 2050 CE | Observed/projected | ±0.5 seconds |
| 2050 - 2150 CE | Extrapolation | ±2 seconds |
| After 2150 CE | Parabolic extrap. | ±10+ seconds |
Data Sources#
Primary References#
Nyx algorithms are derived from established astronomical references:
| Source | Components Used |
|---|---|
| Jean Meeus, "Astronomical Algorithms" | Planetary positions, time systems, |
| (1991, 2nd edition) | coordinate transformations |
| IAU 2006 Precession-Nutation Model | Precession parameters, frame ties |
| IERS Conventions (2010) | Earth orientation parameters |
| USNO Circular 179 | Time scale relationships |
| SOFA Library | Algorithm implementations |
Orbital Element Sources#
Planetary orbital elements are derived from:
| Planet | Element Source | Epoch | Update Frequency |
|---|---|---|---|
| Mercury | JPL DE440 (derived) | J2000.0 | Static |
| Venus | JPL DE440 (derived) | J2000.0 | Static |
| Earth | VSOP87 (simplified) | J2000.0 | Static |
| Mars | JPL DE440 (derived) | J2000.0 | Static |
| Jupiter | JPL DE440 (derived) | J2000.0 | Static |
| Saturn | JPL DE440 (derived) | J2000.0 | Static |
| Uranus | JPL DE440 (derived) | J2000.0 | Static |
| Neptune | JPL DE440 (derived) | J2000.0 | Static |
| Pluto | Meeus coefficients | J2000.0 | Static |
| Moon | Brown-Newcomb theory | Dynamic | N/A |
Physical Constants#
Constants follow CODATA and IAU recommendations:
| Constant | Value | Source | Precision |
|---|---|---|---|
| Speed of light | 299792.458 km/s | CODATA | Exact |
| Astronomical Unit | 149597870.7 km | IAU 2012 | Exact |
| Gravitational const | 6.67430×10⁻¹¹ | CODATA 2018 | ±2.2×10⁻⁵ |
| Earth radius (eq) | 6378.137 km | WGS84 | ±1 m |
| Earth flattening | 1/298.257223563 | WGS84 | Exact def. |
| Obliquity (J2000) | 23.439291111° | IAU 2006 | ±0.00001° |
Star Catalog Sources#
Nyx supports multiple star catalogs with varying precision:
| Catalog | Stars | Position Acc. | Proper Motion | Magnitude |
|---|---|---|---|---|
| Gaia DR3 | 1.8B | 0.02-0.5 mas | Yes | G ≤21 |
| Hipparcos 2 | 118,218 | 0.1-1 mas | Yes | V ≤12 |
| Yale BSC | 9,110 | 1 arcsec | Limited | V ≤6.5 |
| Tycho-2 | 2.5M | 7-60 mas | Yes | V ≤12 |
Validation Methodology#
Test Reference Data#
Nyx validates calculations against known reference positions:
JPL Horizons Cross-Validation#
Selected dates are verified against NASA JPL Horizons system:
Test Case: Mars J2000.0 geocentric position
Date: 2000-01-01T12:00:00 TT
Expected RA: 326.0823° Dec: -18.6821°
Calculated: 326.0825° Dec: -18.6820°
Error: 0.7 arcsec RA, 0.4 arcsec Dec
Status: PASS (within HIGH precision spec)
Historical Eclipse Verification#
Solar eclipse predictions are verified against historical records:
| Eclipse Date | Location | Predicted | Observed | Error |
|---|---|---|---|---|
| 1999-08-11 | Europe | 10:28:21 UT | 10:28:14 UT | +7 seconds |
| 2017-08-21 | North America | 18:25:32 UT | 18:25:32 UT | <1 second |
| 2024-04-08 | North America | 18:18:29 UT | 18:18:28 UT | +1 second |
Coordinate Transformation Tests#
Transformations verified with SOFA library outputs:
// Vega coordinates (Hipparcos)
const vega = { ra: 279.23473479, dec: 38.78368896 }; // J2000 ICRS
// Expected horizontal coordinates for NYC, 2026-01-21T00:00:00 UTC
const expected = { altitude: 82.341, azimuth: 267.892 };
const calculated = equatorialToHorizontal(vega, observer, date);
expect(calculated.altitude).toBeCloseTo(expected.altitude, 2);
expect(calculated.azimuth).toBeCloseTo(expected.azimuth, 2);
Automated Test Suites#
Nyx includes comprehensive test coverage:
| Test Category | Tests | Coverage | Tolerance |
|---|---|---|---|
| Ephemeris positions | 156 | 95% | ±1 arcsec |
| Coordinate transforms | 89 | 92% | ±0.001 arcsec |
| Time conversions | 67 | 98% | ±0.1 seconds |
| Eclipse predictions | 34 | 85% | ±5 seconds |
| Rise/set times | 45 | 90% | ±30 seconds |
Known Reference Positions#
Test fixtures include verified positions for:
export const KNOWN_POSITIONS = {
// Verified against JPL Horizons
mars_2026_01_21: {
date: '2026-01-21T00:00:00Z',
ra: 117.4523, // degrees
dec: 23.8912, // degrees
distance: 1.42, // AU
magnitude: 0.8,
uncertainty: {
ra: 0.5, // arcseconds
dec: 0.5, // arcseconds
},
},
// Moon position (simplified theory)
moon_2026_01_21: {
date: '2026-01-21T00:00:00Z',
ra: 45.234,
dec: 12.456,
uncertainty: {
ra: 1080, // 0.3 degrees in arcsec
dec: 1080,
},
},
};
Known Limitations#
Ephemeris Limitations#
-
Precision Tier Selection
LOWand real-time renderer paths prioritize stable frame times and may use compact analytical approximations.MEDIUMandHIGHplanning paths use full Nyx ephemeris services with light-time, phase, illumination, and validated catalog data.- Pipeline precomputation uses Astronomy Engine-backed VSOP87/ELP calculations for solar-system bodies.
-
Long-Range Prediction Drift
- Analytical theories remain bounded near their validated epochs but still drift outside the published date ranges.
- Historical and far-future eclipse, occultation, and close-approach workflows should pin the data source and validation epoch used.
-
Relativistic Effects
- Light-time correction and core post-Newtonian solar-system corrections are included in high-precision paths.
- Frame dragging and strong-field relativistic effects are modeled in the renderer/exotic and Kalika integration paths, not in every fast sky-chart calculation.
-
Non-Gravitational Forces
- Minor-body ephemerides include common comet non-gravitational terms and solar radiation pressure where supplied by the object model.
- Unmodeled outgassing, spacecraft maneuvers, and newly measured perturbations require a catalog refresh or externally supplied elements.
Coordinate Limitations#
-
Nutation Model Choice
- The positional library provides selectable reduced and high-precision IAU-family nutation models with cached precession-nutation matrices.
- Exact sub-milliarcsecond reproducibility should record the selected model and the IERS data version used for validation.
-
Polar Motion Data Freshness
- IERS polar motion is integrated and interpolated when a dataset is loaded.
- Offline fallback values are bundled for deterministic operation; refresh IERS Bulletin A/B data for current telescope-pointing sessions.
-
Atmospheric Refraction
- Bennett, Saemundsson, Smart, and Auer-Standish models are available with pressure, temperature, humidity, wavelength, and elevation inputs.
- Near-horizon refraction remains weather-sensitive and should be treated as observationally variable.
Time Limitations#
-
Delta-T Extrapolation
- Beyond 2050: Uses polynomial extrapolation
- Actual ΔT depends on unpredictable Earth rotation
- Far-future predictions unreliable
-
Leap Second Handling
- Table current through 2017 (37 cumulative)
- Future leap seconds must be manually updated
- May cause issues after IERS announces new leap seconds
-
No IERS Bulletin A/B
- Does not fetch real-time Earth orientation
- UT1-UTC approximated from long-term models
- ~0.9 second maximum error in UT1
Catalog Limitations#
-
Epoch Propagation
- Proper motion propagated linearly
- Does not account for perspective acceleration
- Binary star orbits not modeled
-
Magnitude Variations
- Variable stars shown at mean magnitude
- No eclipsing binary phase calculations
- Nova/supernova alerts not supported
Accuracy by Component#
@nyx/ephemeris#
| Object | RA/Dec Accuracy | Distance Acc. | Valid Range |
|---|---|---|---|
| Mercury | ±2 arcsec | ±0.001 AU | 1800-2200 CE |
| Venus | ±1 arcsec | ±0.001 AU | 1800-2200 CE |
| Mars | ±1 arcsec | ±0.001 AU | 1800-2200 CE |
| Jupiter | ±2 arcsec | ±0.01 AU | 1800-2200 CE |
| Saturn | ±3 arcsec | ±0.01 AU | 1800-2200 CE |
| Uranus | ±5 arcsec | ±0.1 AU | 1800-2200 CE |
| Neptune | ±5 arcsec | ±0.1 AU | 1800-2200 CE |
| Pluto | ±30 arcsec | ±0.5 AU | 1900-2100 CE |
| Moon | ±1080 arcsec | ±1000 km | Universal |
| Sun | ±1 arcsec | ±0.0001 AU | 1800-2200 CE |
@nyx/coordinates#
| Transformation | Typical Error | Maximum Error |
|---|---|---|
| J2000 ↔ Current epoch | 0.001 arcsec | 0.01 arcsec |
| Equatorial ↔ Horizontal | 0.01 arcsec | 0.1 arcsec |
| Equatorial ↔ Galactic | 0.0001 arcsec | 0.001 arcsec |
| Atmospheric refraction | 0.1 arcsec | 1 arcsec |
| Proper motion (100 yr) | 0.1 arcsec | 1 arcsec |
@nyx/time#
| Conversion | Typical Error | Maximum Error |
|---|---|---|
| UTC → JD | 0 (exact) | 0 |
| JD → MJD | 0 (exact) | 0 |
| UTC → TAI | 0 (exact) | 0 |
| TAI → TT | 0 (exact) | 0 |
| TT → TDB | 10 μs | 30 μs |
| UTC → UT1 | 0.3 s | 0.9 s |
| Delta-T (2020) | 0.1 s | 0.2 s |
| Delta-T (1600) | 1 s | 4 s |
Recommended Use Cases#
Suitable Applications#
| Use Case | Precision Needed | Nyx Suitable? |
|---|---|---|
| Planetarium software | ±1 arcmin | Yes |
| Mobile stargazing apps | ±1 arcmin | Yes |
| Educational demonstrations | ±10 arcsec | Yes |
| Visual observation planning | ±10 arcsec | Yes |
| Telescope GoTo pointing | ±5 arcsec | Yes |
| Astrophotography framing | ±5 arcsec | Yes |
| Eclipse timing (±1 min) | ±60 seconds | Yes |
| Conjunction predictions | ±1 arcmin | Yes |
| Satellite tracking | ±1 degree | Partial |
Not Recommended For#
| Use Case | Precision Needed | Why Not Suitable |
|---|---|---|
| Spacecraft navigation | ±1 mas | Use JPL DE ephemerides |
| Asteroid astrometry | ±0.1 arcsec | Needs perturbation theory |
| Lunar occultation timing | ±0.1 sec | Moon theory too simplified |
| Gravitational wave detection | ±1 μas | Relativistic effects needed |
| Pulsar timing | ±1 ns | Use TEMPO2/PINT |
| VLBI correlator | ±1 ps | Specialized software needed |
References#
Primary Sources#
-
Meeus, Jean. Astronomical Algorithms. 2nd ed. Willmann-Bell, 1998. ISBN 978-0943396613.
-
Urban, Sean E., and P. Kenneth Seidelmann, eds. Explanatory Supplement to the Astronomical Almanac. 3rd ed. University Science Books, 2012. ISBN 978-1891389856.
-
IERS. IERS Conventions (2010). IERS Technical Note No. 36, 2010. URL: https://www.iers.org/IERS/EN/Publications/TechnicalNotes/tn36.html
-
IAU SOFA Board. SOFA Library. International Astronomical Union, 2021. URL: http://www.iausofa.org/
Supplementary Sources#
-
Folkner, William M., et al. "The Planetary and Lunar Ephemerides DE430 and DE431." Interplanetary Network Progress Report 196 (2014): C1.
-
Capitaine, Nicole, et al. "Expressions for IAU 2000 Precession Quantities." Astronomy & Astrophysics 412.2 (2003): 567-586.
-
Kaplan, George H. The IAU Resolutions on Astronomical Reference Systems, Time Scales, and Earth Rotation Models. USNO Circular 179, 2005.
Online Resources#
- JPL Horizons: https://ssd.jpl.nasa.gov/horizons/
- IERS Earth Orientation Center: https://www.iers.org/IERS/EN/DataProducts/EarthOrientationData/eop.html
- Gaia Archive: https://gea.esac.esa.int/archive/
Version History#
| Version | Date | Changes |
|---|---|---|
| 1.0.0 | 2026-01-21 | Initial accuracy documentation |