# Optical time-series change detection
*analysis . methodologies*

Flags surface change by testing each new optical observation against a per-pixel statistical baseline built from the historical time-series, accumulating evidence across dates before confirming a change. Underlies near-real-time deforestation and disturbance alert systems.

## Specifications
- **family**: Change detection
- **entity type**: methodology
- **last verified date**: 2026-06-04
- **verified by**: agency-doc
- **claim status**: unclaimed
- **subtype**: analysis
- **attributes**: {"family":"Change detection","kind":"analysis","summary":"Flags surface change by testing each new optical observation against a per-pixel statistical baseline built from the historical time-series, accumulating evidence across dates before confirming a change. Underlies near-real-time deforestation and disturbance alert systems."}

## Editorial
Optical time-series change detection answers the question: has this surface location changed, and when? Rather than comparing two single-date images, the method builds a per-pixel statistical baseline from the historical time-series of optical reflectance, then tests each new observation against that baseline. Change is not flagged from a single anomalous scene but confirmed only when evidence accumulates across multiple dates, reducing false alarms from cloud contamination and sensor noise.

Two algorithm variants have been deployed in near-real-time deforestation and disturbance alert systems. The GLAD-L alert system (Hansen et al. 2016) applies bagged decision trees to Landsat red, near-infrared, and shortwave-infrared temporal metrics, confirming a disturbance when at least 2 of 4 consecutive observations within a 180-day window cross the change threshold.[^hansen-2016-irl] The GLAD-S2 extension applies decision trees to Sentinel-2 bands 2, 3, 4, 8, 11, and 12 with a 4-tier confidence scheme (single, low, medium, high).[^gfw-alert-integration-2024] The RADD alert system (Reiche et al. 2021) takes a different approach: it uses Bayesian conditional probability of forest disturbance, fitting a per-pixel Gaussian mixture model as the temporal baseline and reporting low-confidence detections above 85% and high-confidence detections above 97.5% probability on Sentinel-1 C-band 10m imagery.[^reiche-2021-erl]

The method applies to any surface monitored by a dense optical or SAR time-series and is not limited to forest. Known failure modes include persistent cloud cover that prevents baseline accumulation, seasonal phenological cycles misread as change, and rapid re-vegetation that obscures confirmed disturbances within the confirmation window.

## Sources
- [hansen-2016-irl] | Humid tropical forest disturbance alerts using Landsat data | https://iopscience.iop.org/article/10.1088/1748-9326/11/3/034008 | tier=peer-reviewed | accessed=2026-06-04 | author=Hansen et al.
  (2016) Primary GLAD-L algorithm paper: bagged decision trees on Landsat red, NIR, SWIR temporal metrics; confirmation rule 2/4 observations within 180d
- [reiche-2021-erl] | Forest disturbance alerts for the Congo Basin using Sentinel-1 | https://iopscience.iop.org/article/10.1088/1748-9326/abd0a8 | tier=peer-reviewed | accessed=2026-06-04 | author=Reiche et al.
  (2021) RADD: Bayesian conditional probability of forest disturbance; per-pixel Gaussian mixture baseline; low-confidence >85%, high-confidence >97.5%; Sentinel-1 C-band 10m
- [gfw-alert-integration-2024] | Integrating satellite-based forest disturbance alerts improves detection timeliness and confidence | https://iopscience.iop.org/article/10.1088/1748-9326/ad2d82 | tier=peer-reviewed | accessed=2026-06-04 | author=GFW consortium
  (2024) Describes GLAD-L, GLAD-S2, RADD side-by-side; GLAD-S2 uses 4-tier confidence (single, low, medium, high) via decision trees on Sentinel-2 bands 2,3,4,8,11,12

---
Source: https://eo-atlas.org/methodologies/optical-timeseries-change-detection
Maintainer: SpectraWorks B.V. (CC-BY 4.0)