How seasons are found
Every season on OpenPhenology comes from the same published method, applied the same way to every species and place.
Method seasonal-5 · Updated 12 October 2026
What a season shows
A season shows when people recorded a species in a place. It is not a migration, flowering or emergence date. It is not corrected for how many people were out recording, and it says nothing about how common a species is or whether it is absent.
Which records count
- iNaturalist observations that reached Research Grade, and records from GBIF datasets identified by experts.
- Records shared under CC0, CC BY 4.0 or CC BY-NC 4.0, observed on one exact day in the last 10 complete years, 2016 to 2025. The window moves forward one year each new year. Older records are kept but not counted.
- Wild organisms only, identified to species. Captive and cultivated ones are left out.
- A record counts toward a county only when its location is precise enough to place it there.
- Locations of sensitive species are protected before anything is calculated.
When there is a result
A species has a result in a place once it has at least 20 records there from at least 5 different years. With fewer, the page says there is not enough evidence yet. A county with too few records can still be compared with Ohio's result.
Finding the season
Records are counted by week, with every complete year from 2016 pooled. Records from the current year are kept but not counted until the year ends, so every week has the same years behind it. The early, middle and late weeks are the weeks by which 5%, 50% and 95% of the records had been made.
For species recorded through the winter, the year is counted from the quietest stretch of weeks, so one winter season is not split across New Year. To check how certain the weeks are, the records are resampled 2,000 times; the same records always give the same result.
When no season is shown
The result is still shown, without season weeks, when any of these applies:
| Reason | When | Why |
|---|---|---|
| Recorded all year | The span from early to late is more than 39 weeks | There is no single season to describe |
| Two seasons | The weekly records have two clear, separate peaks | One early-to-late window would put the middle week in the gap |
| Mostly one year | One year holds more than 40% of the records | One year would stand in for all of them |
| Too uncertain | Any of the three weeks is uncertain by more than 6 weeks | The weeks are too uncertain to show |
Nearby labels
For a chosen county and week, each species gets the first label that applies:
| Label | When |
|---|---|
| Commonly reported this time of year | Recorded in the county within a week of this one in at least half of its years |
| Seasonally reported | The week falls between the county's early and late weeks |
| Reported in your area | The species has a county result |
| Within its mapped range | A published range map covers the county; shown with its source |
A label describes past records or a published range. It never says a species is present, absent or expected in a yard.
Recording effort
Beside each result we show how many records of all species were made in that place each week, so you can see when recording itself was high or low. Results are not adjusted by it.
Versions
This is method seasonal-5. Any change to a number or rule here makes a new version, and each release recalculates every result with one version. The full rules are public.