How the forecasts are computed

The inputs

Each of the eleven major annual showers is described by seven published parameters: its radiant position (right ascension and declination), the solar longitude of its peak, its activity window, its peak zenithal hourly rate (ZHR), its entry speed, its population index, and its parent body. These come from the International Meteor Organization's annual calendar and the IAU Meteor Data Center. They change only when the science does.

The ephemeris

Positions of the Sun, the Moon and each radiant are computed with astronomy-engine (version 2.1.19), an open-source library whose positions agree with NASA JPL's to within about an arcminute. A shower's peak is the instant the Sun reaches the shower's peak solar longitude in that year, so the date drifts by up to a day between years and the hour changes every time; we solve for it rather than repeat a fixed date.

"Darkness" means the Sun is more than 18° below the horizon (astronomical night). At high latitudes in summer the Sun never gets that low, and the pages say so: the window falls back to nautical twilight (12°) and the rates are discounted for the brighter sky.

Activity across the window

Between the start of a shower's activity and its peak, the rate rises exponentially, and after the peak it decays the same way, reaching about 2% of the peak ZHR at the window's edges. Short windows therefore give sharp peaks (the Quadrantids) and long ones gentle plateaus (the Taurids). This is a smooth approximation of the IMO's activity profiles. Real showers have structure it does not capture, outbursts above all.

From ZHR to what you will count

ZHR is defined for an ideal case. The rate an observer actually sees is lower, by the standard relation

rate = ZHR × sin(h) ÷ r^{(6.5 − LM)}

where h is the radiant's altitude, r the population index (2.5 here) and LM the faintest star visible to the naked eye. We use LM 6.3 for a dark rural site, 5.3 for suburbs and 4.3 for a city. A Moon above the horizon lowers LM further, by up to 2.5 magnitudes at a dark site when it is full. Rates are shown as a range of roughly ±40% around the estimate, because counts from a single observer scatter that much from hour to hour.

The random "sporadic" background is modeled as 8 meteors an hour under a dark sky, from no particular direction. It is included in tonight's totals, which is why a night with no active shower still shows a few meteors an hour.

Moon conditions

The calendar and shower pages grade moonlight during the hours it matters: three hours centered on the moment the radiant is highest in darkness on the peak night, evaluated at 40° N (or 34° S for the southern showers). The grade combines how much of the Moon is lit with how much of those three hours it spends above the horizon, so a bright Moon that has set by then counts for nothing.

Clouds

Cloud cover is the one input that is not astronomy. On the home page it is fetched in your browser from Open-Meteo for the place you chose, and used to pick the best viewing window. Everything else on that page would read the same under any weather.

What the numbers do not promise

• Outbursts and storms. Some years a shower far exceeds its normal ZHR. The model shows a normal year.
• Your exact horizon. Trees, buildings and haze near the horizon take a real share of what you could see.
• Local light pollution. You choose the sky class; the site cannot measure your sky.
• The weather beyond a few days. Cloud cover is only shown for tonight.

How often it is recomputed

The site is rebuilt every day and each build recomputes every date, rate and moon grade, but a page only changes when its content does: the home page's tonight answer daily, a shower page when its next apparition begins. If something looks wrong, write to support@ilsur.me. A wrong number is easier to fix when you include the page and your location.