Caelus3,462 checksGitHubv0.25.0MIT
Instrument I-04

First visibility

When a star or planet passes near the Sun, solar glare hides it for days or months. Select a body, location, and year to compute every first and last visibility event across the year and inspect the twilight margin or crescent score behind each date.

INBody, place, year

139 CE (Julian calendar). Years use astronomical numbering (year 0 is 1 BCE), with east longitude positive. Extinction is V-band magnitude loss per unit air mass: 0.25 for a clear sea-level sky, around 0.15 at high desert sites, and 0.35 or higher in haze. Censorinus implies that Sirius rose on the Egyptian New Year in 139 CE. The manuscripts give 21 July, usually emended to 20 July.

OUTFirst and last visibility
0eventsSirius from Memphis, 139 CE
caelus 0.25.0, embedded tier
Not seen0
Evening last0
Morning first0
JanFebMarAprMayJunJulAugSepOctNovDec

Each marker shows a nominal event date, while the shaded band behind it spans the earliest (clearer sky) to latest (hazier sky) dates. The top row highlights intervals when the body is lost in solar glare.

TWLThrough the twilightchoose an event

Searching.

TBLThe eventstimes in UT, at the moment of best visibility

Searching.

Alt and Sun give the body and solar altitudes at peak visibility. Margin is magnitudes above the visual threshold. Arcus is the implied arcus visionis (solar depression when the body crossed the horizon).

TSThe call
import { Engine, heliacal, julianDay, calendarDate } from "caelus";
import { embeddedData } from "caelus/data-embedded";

const engine = new Engine(embeddedData);
const start = julianDay(139, 1, 1, 0, 0, 0, "julian");

const ev = heliacal(engine, "Sirius", start, 29.85, 31.25);

calendarDate(ev.jd, "julian");
// the next one: search again from ev.jd + 1
RETWhat it returnsUT Julian Days
null
M-1Visibility model

For each dawn or dusk, the solver steps every two minutes as the body clears or approaches the horizon and compares its extinguished magnitude against the naked-eye contrast threshold. Background sky brightness combines night-sky glow, twilight scattering, and moonlight using B. E. Schaefer's published equations. Lunar crescents use B. D. Yallop's empirical q criterion instead.

M-2Atmospheric window

Because local atmospheric haze on a historical date cannot be known exactly, each event is solved at the selected extinction and at ±0.05 magnitudes per air mass. The resulting window spans the clearer-sky and hazier-sky dates (typically two to three days for bright stars, and up to three weeks for slow Mars conjunctions). Downstream analysis should use the full window rather than a single nominal date.

M-3Limits

Calculations assume unaided naked-eye vision over a flat horizon at the observer's elevation with uniform seasonal extinction. Real-world sightings vary by several days with weather and observer acuity. Before roughly 800 BCE, ΔT clock uncertainty can shift a marginal lunar crescent date by one day. Dates prior to 15 October 1582 use the Julian calendar.