Harvest & availability

Dated harvest windows

Analysis by OlioWise · Calculated 4 Oct 2026 · Snapshot 0cc247d1e3d3

Across 618 dated harvest releases from 388 oils, northern oils typically pick around 1 Nov and southern oils around 15 May. 239 are treated as day-precise, 364 as month-precise, and 15 as range-precise. Year-to-year movement is shown only for groves with consecutive dated seasons, not as a country headline.

Shop-available dates are first observed producer-shop availability, not mill calendars. Browse oils that disclose a harvest date.

States with enough dated harvests

Shown where at least five dated harvests exist. Named regions remain too thin for a country-style later chip.

Later than this grove’s history

Each row is one oil whose latest dated season is more than 14 days later than that oil’s own earlier harvest dates. This is the honest year-to-year instrument at current occupancy. Day-precise pairs are listed first.

Methodology and limitations

Methodology updated 17 September 2026. The calculation date above is when the catalogue was analysed, not when the fruit was picked. The snapshot identifies the calculation’s input observations.

• Scope: 3614 release records from 1409 oils were considered. A reported harvest day, or a harvest_window, is one observation. The calendar uses a window’s inclusive start and end; the existing summary statistics use its midpoint. Legacy harvest_date values without harvest_date_basis: reported_day may have unverified day precision. Shop availability uses date_available as first observed producer-shop availability. 0 harvest dates were dropped as unparseable, year-mismatched, or missing a hemisphere.
• Season: Southern dates use the calendar year of harvest (YYYY-SH). Northern dates wrap: August through the following July is one season (YYYY/YY-NH), so October 2024 and January 2025 are the same northern season. Day-in-season is counted from 1 August (north) or 1 January (south) so medians do not wrap 31 December. Calendar year alone is never treated as a measured harvest date.
• Precision: A sourced harvest day is day-precise. A reported_range is range-precise; a named month or interpreted phrase is month-precise. Legacy dates on the first of a month are treated as month-precise; other legacy days are day-precise. Typical windows publish both the all-dated median and, where n allows, the day-precise median. A 1 November cluster is a month bin, not a phenology reading.
• Typical windows: Country and hemisphere medians and interquartile ranges require at least five dated harvests. Countries below eight dated harvests are marked as not enough for a later-than-typical call.
• Later than typical: A country or state chip is shown only when the latest season and the earlier baseline each have at least eight day-precise dates, the median is more than 14 days later or sits above the baseline 75th percentile, and at least five of the same oils appear in both samples. Otherwise the cell is “—” or “not enough dated harvests”, never an unchanged-harvest claim from month-pinned zeros. Mix changes are not promoted as climate or weather.
• Grove panel: 147 oils have harvest dates in two or more seasons. A grove is listed when its latest dated season is more than 14 days later than that oil’s own earlier harvest dates. Day-precise pairs are separated from rows that use a first-of-month date.
• Availability: Mill-to-shop lag uses the same release’s harvest date and first-available date, discarding negative or ≥400-day gaps. Missing date_sold_out is not treated as still in stock. Current producer_inventory is a live range flag and is not a year series on this page.
• Calendar: The week-by-week grid uses fixed seven-day calendar bins from 1 January; the final bin contains two days. In the combined view each harvest period distributes one unit across its days, so a longer range does not count more than a reported day.
• Limits: This is a selected catalogue of producer-direct oils with uneven harvest disclosure, not a representative crop calendar. The mix of oils changes each year. Outliers whose harvest year differs from the release year by more than two years are excluded. See Trust and Evidence and transparent harvest dating.

Corrections: Report a discrepancy with the oil URL, release year, harvest date and supporting source. Corrected catalogue records flow into the next calculation. Refer to this page with its calculation date and snapshot when citing a finding.