Reporting
Relative abundance (RAI) and trap-nights
You usually can't count individual animals of an unmarked species on camera. Relative abundance, how often a species turns up corrected for how long you watched, is the standard way around that. It's a useful index, as long as you know exactly what it does and doesn't measure.
Most wildlife on a camera trap can't be told apart individual by individual — one red fox looks like the next. So instead of counting animals, camera studies count how often a species is detected and correct that for survey effort. The result is a Relative Abundance Index (RAI): a single number per species that lets you compare the same species across time or places on a level footing.
How the index is calculated
The platform builds RAI in three steps, per camera and per species:
- Independent events. Raw detections are grouped into events: a new event starts only when at least 30 minutes have passed since the last detection of that species. This stops one animal lingering in front of the lens from counting as dozens of observations.
- Effort in trap-nights. Effort is the length of the deployment — the registered deployment period where it exists, otherwise the span from the first to the last upload — measured in calendar days.
- The index. RAI = events × (100 / trap-nights), reported per species to one decimal. In words: independent events per 100 trap-nights.
The assumptions behind the number
RAI is a convention, and every convention carries assumptions. Four are worth keeping in mind when you read one:
- The 30-minute independence rule is applied identically to every species — a house mouse and a red deer are treated the same way. It's a practical convention, not a law of nature.
- Effort is calendar time, not working time. A trap-night is a night on the calendar the camera was deployed, whether or not it fired.
- Every detection that wasn't rejected counts — including ones the AI labeled on its own, without a person confirming them. Your review discipline flows straight into the index.
- The per-100-trap-nights denominator is just a scaling choice to get readable numbers; it doesn't change what's being compared.
What RAI can — and can't — tell you
This is the part that matters most. A RAI value blends three different things together: how many animals are actually there (abundance), how easy that species is to detect (detectability), and how it moves through the landscape (behavior). Because those are baked in, RAI is:
- Valid for the same species over time or between comparable sites. If detectability and method stay roughly constant, a rising or falling RAI is a real signal worth reporting.
- Not valid as a comparison between species. A higher RAI for wild boar than for pine marten does not mean there are more boar — boar are simply bulkier, warmer, and more detectable. Ranking species by RAI reads as a head-count, and it isn't one.
Treat RAI as a trend line for one species, not a leaderboard across species. That single distinction prevents most of the ways the number gets misread.
Why the effort figure has to be honest
RAI is only as trustworthy as the trap-night figure underneath it, because effort sits in the denominator. If a camera fails halfway through a period but the outage still counts as effort, the index can be pulled down sharply — a camera that goes dark for half its deployment, with the dead time still counted, can underestimate RAI by a factor of three. As a matter of method, effort should reflect the time a camera was genuinely able to record, which is why it's worth reading RAI alongside camera health: a quiet camera and an empty forest look identical in the raw numbers, and only uptime tells them apart.
RAI is one measure, not the whole picture. Alongside relative abundance, camera-trap monitoring produces survey effort in trap-nights and activity patterns across the day — together the standard measures ecological monitoring runs on. See how they feed reporting in camera traps for Natura 2000 monitoring, and how detections become monitoring evidence on the Product page.
