Perseids

The Perseids tolerate a mediocre sky better than most: r = 2.2 is on the bright-rich side, so each magnitude of sky glow costs you a factor of 2.2 rather than the 3.2 of the delta-Aquariids. Combined with a ZHR of 100 and a radiant that reaches 77° from 45°N, a suburban garden on a clear peak night still produces a meaningful count. The case for travelling is strongest in a dark-Moon year — a dark site multiplies a dark-Moon Perseid night, but it cannot rescue a moonlit one.

The radiant is at declination +57.9°. The highest a radiant ever gets is 90° minus the difference between your latitude and its declination, so south of 32.1°S the radiant never rises at all. Sydney at 33.9°S and Cape Town at 33.9°S are both inside that band: the Perseid radiant is under the horizon for the entire night, every night, and the modelled rate is exactly zero rather than merely low. The eta-Aquariids in May are the southern hemisphere's equivalent event.

The Geminids have the higher ZHR (150 against 100) and a radiant at declination +32.3° that suits mid-northern latitudes well. The Perseids counter with r = 2.2 against 2.6, meaning a richer share of bright meteors, and with August weather rather than December weather. The honest answer is that the Geminids are the stronger shower and the Perseids are the more likely one to actually happen for you. In a year where both peaks fall under a dark Moon, take the Geminids and dress properly.

2026 and 2029. The rest are compromised, and 2027 and 2030 land near full Moon. 2026 is the outlier of the decade: the peak falls within hours of new Moon, which removes the single factor you cannot do anything about.

The window runs 17 July – 24 August — 39 days — so yes, the shower is genuinely active for weeks and the nights immediately around the peak still produce well. That is what makes the Perseids forgiving of cloud: if the peak night is overcast, the following night is usually still worth going out for. Compare the Ursids, whose window is nine days and whose ZHR is 10.

Not on its own. What matters is whether the Moon is above your horizon during the dark hours, not merely how lit it is — a half Moon that sets before midnight leaves the best hours untouched. This site applies the penalty from the Moon's altitude and illuminated fraction hour by hour for your coordinates rather than from the phase alone, which is why the verdict can differ between two places on the same night. Scale matters too: the Perseid population index is 2.2, so even the model’s worst-case 2.0-magnitude Moon penalty is a factor of 4.8, and a tabulated ZHR of 100 has room to absorb it. A shower with a base rate in single figures does not.

Judged only on how much of the Moon is lit at the peak instant: 2026 — 13 August 2026 (Moon 0% lit); 2027 — 13 August 2027 (Moon 87% lit); 2028 — 12 August 2028 (Moon 59% lit); 2029 — 12 August 2029 (Moon 10% lit); 2030 — 13 August 2030 (Moon 100% lit). The dark-Moon years are 2026 and 2029; the washed-out ones are 2027 and 2030. Moon illumination is the only one of the four factors that is knowable years ahead — cloud cover is not, and neither is whether you will be somewhere dark.

Because 100 is the zenithal hourly rate: what one observer would count under a magnitude 6.5 sky with the radiant straight overhead. Two corrections pull it down before you ever look up. Rate scales with the sine of the radiant altitude, so a radiant 30° up delivers half of what it delivers at the zenith. And the population index for this shower is r = 2.2, which means each magnitude of sky brightness you lose costs you a factor of 2.2 in the count. A suburban sky two magnitudes shallower than the reference therefore divides the rate by about 4.8. Real counts are typically 30–50% under the modelled figure even after those corrections.