Operational definition of recurrence statistics

How the statistics shown in the "Recurrence by zone" panel and the full report page are computed.

Method

The algorithm divides the whole map into cells of ~50m × 50m (latitude-corrected so they stay square). Each raw thermal is assigned to a single cell according to its coordinates. For each zone, we keep the cells whose centroid falls inside the polygon.

Fields

n_total: total number of raw thermals observed in the zone, all cumulated (5 seasons).
n_unique_locations: number of distinct 50m cells touched at least once.
distribution["N"] = X%: percentage of thermals (not cells!) that lie inside a cell containing at least N thermals.

Note: proximity is measured via a 50m grid, not a circle. Two thermals 80m apart can end up in the same cell (if close to the boundary), and two thermals 30m apart can end up in different cells. It is a discrete approximation of "within 50m" - good in practice but not exact.

Interactive map Full report (charts)