A warm anomaly becomes a marine heatwave only after it is compared with a location- and season-specific baseline and persists long enough to meet a stated event rule.
A widely used framework defines a marine heatwave from daily temperature exceeding a seasonally varying 90th-percentile threshold for at least five consecutive days. The climatology period, temporal smoothing, spatial cell and treatment of short gaps must be recorded because they can change event boundaries.
This diagram is editorial and contains no measured temperatures. Five consecutive “above” labels illustrate the event rule; they do not estimate an event at a real location.
A temperature event can be mapped from physical data, but bleaching, mortality, range shifts or fishery losses require biological observations matched in space and time. Species differ in acclimatization, life stage and exposure. A surface product may also miss subsurface refuges or heating.
Use the Marine Heatwave Tracker to learn the interface, then replace its illustrative values with a documented dataset. Record product version, coordinates, depth, baseline dates, percentile, event rule and retrieval date before comparing locations.
Definitions and dataset documentation were checked on 10 October 2026. Threshold choices and product resolution must accompany any reported event.
Ocean field notebook / discoveries
A warm anomaly becomes a marine heatwave only after it is compared with a location- and season-specific baseline and persists long enough to meet a stated event rule.
A threshold depends on the reference period and local seasonality, so duration, intensity and baseline must be read together. Relate that field focus to “Event detection is a threshold calculation” before moving to “Teaching threshold example.”
Pair field observations made with consistent sampling effort with regional records that expose spatial variation. Record depth, location, date, units and sampling effort so a change in conditions is not mistaken for a change in the ocean.
Compare “Teaching threshold example” with “Detection does not measure impact.” Decide whether the pattern represents exposure, response or recovery, and state the observation period needed to tell them apart.