Remote Sensing
Breakpoint detection
An algorithm family that finds abrupt shifts in a satellite time series, dating disturbance to within days or weeks.
Breakpoint methods fit a running signal and identify the moment its behaviour changes. Applied to dense time series, they can date a clearing, a burn or a flood far more precisely than comparing two annual composites.
The gain is temporal precision, which matters for attributing loss to a period, to a project boundary, and to whether it occurred before or after crediting began.
Frequently asked questions
Do breakpoints identify the cause?
No. They date and locate change; interpretation of cause needs imagery review and context.
What data suits this method?
Dense, well-calibrated series such as Sentinel-2 or harmonised Landsat.